Task-based relative observation bloodline physical examination reexamination state control method and system

Abstract

The present application relates to physical examination data processing technical field, disclose a kind of physical examination reexamination state control method and system based on task relative observation blood relationship.The method analyzes the electronic physical examination data carrier of same examinee, forms the appearance record carrying delivery domain mark and evidence nature;Original collection mark is alias merged, and observation collection root and standard observation event are established;Review suggestion is anchored to task source observation event, and stable task identification, window lower bound and deadline are generated;The historical source relationship of candidate observation event is formed to form source closure, and three-value evidence gate is executed to collection identity independence and non-pure back reference;According to time classification and preset priority, determine unique task state and output longitudinal recheck clues.The present application can inhibit repeat event, repeat task, task source pure back reference performance and task time boundary drift, improve the replayability and checkability of task state control.

Description

Technical Field

[0001] This invention relates to the fields of electronic medical examination data processing, graph data processing, and computer status control technology, and in particular to a method and system for controlling the status of medical examination re-examination based on task-relative observation of bloodline.

[0002] More specifically, the present invention relates to a technology for deterministically processing occurrence records, actual collection identities, standardized observation events, review tasks, and historical source relationships in heterogeneous electronic medical examination data carriers, and thereby controlling the status of review tasks. Background Technology

[0003] In the context of health checkup data management, the examination data of the same examinee may be stored in structured records, resource-based data, tables, text reports, image reports, or other electronic health checkup data carriers. Different carriers may not only contain data generated in the current batch of health checkups, but may also carry previous observation results through historical fields, comparison content, copied fragments, or nested objects. Therefore, the reappearance of an observation in a certain carrier does not necessarily indicate that a new actual data collection action has occurred.

[0004] Existing longitudinal data organization schemes typically merge records based on examination items, observation time, text content, carrier identifier, or carrier reception time, and determine whether a record belongs to current or historical data based on the carrier or field it belongs to. Existing review task tracking schemes usually generate tasks from review suggestion texts, and determine the task start and end times using report issuance time, carrier generation time, carrier reception time, or task creation time, and then update the task status based on subsequent similar project records.

[0005] However, the delivery location in the electronic medical examination data carrier and the actual source of the observation results belong to different dimensions. Records located in the historical delivery domain may correspond to independent subsequent acquisitions, while records located in the current delivery domain may simply be references, copies, or derivations of the task source observation content. If the task fulfillment is judged solely based on the delivery domain, text similarity, or carrier time, it is easy to mistake historical projections for new acquisitions or to miss independent subsequent observations packaged in the past. In addition, different systems may use different identifiers for the same actual acquisition, and the same original identifier may not be able to directly prove the acquisition identity due to source conflicts, making it difficult to reliably establish standardized observation events based solely on string deduplication.

[0006] Therefore, the existing technology has at least the following limitations: First, it lacks a control mechanism that uses the task source observation events of a specific review task as a reference to jointly judge the actual collection identity and historical source path of candidate observation events, which can easily lead to the duplication of standardized observation events or the self-fulfillment of task source content; Second, the review task is easily affected by repeated delivery, late supplementary recording, structural changes and carrier time changes of electronic physical examination data carriers, which can lead to the repeated creation of tasks or changes to the task time anchor point, lower window limit and deadline of existing tasks, making it difficult to determine the task status for replay and verification. Summary of the Invention

[0007] The technical problem to be solved by this invention is to provide a method and system for controlling the status of physical examination re-examination based on the relative observation lineage of the task, so as to distinguish the delivery location of the record and the actual collection source of the corresponding observation result, identify independent subsequent observations and pure back-reference of the task source, suppress the repetition of standardized observation events and task self-fulfillment; at the same time, it ensures that the stable task identifier and time boundary of the re-examination task do not drift with repeated delivery of the carrier or changes in the carrier time, thereby obtaining a unique task status that can be replayed and verified and longitudinal verification clues.

[0008] To address the aforementioned technical problems, this invention provides a method for controlling the status of physical examination follow-up based on task-relative observational lineage. The task-relative observational lineage refers to the relationship between a candidate observation event and the follow-up task in terms of task source observation events, data acquisition identity, and historical source path. The method is executed by an electronic device and includes:

[0009] By analyzing the electronic medical examination data carrier of the same examinee, the occurrence records of the delivery domain markers and evidentiary nature can be obtained;

[0010] Alias ​​merging is performed on the original acquisition identifiers of the records that appear, and observation acquisition roots and standard observation events are established according to acquisition identity equivalence classes;

[0011] The review recommendations in the recorded occurrences are anchored to the task source observation events, and a stable task identifier, lower window bound, and deadline for the review task are generated according to the task time anchor point and performance offset interval.

[0012] Referring to the task source observation event, the historical source relationship of the candidate observation event is traversed to form a source closure. Three-valued evidence gating is performed on the independent collection identity and non-pure backreference. The candidate observation event is a standard observation event selected from the standard observation events that initially matches or has an uncertain matching result with the review target range of the review task. The non-pure backreference means that the result source of the candidate observation event does not only refer back to the task source observation event.

[0013] Candidate observation events that pass the three-valued evidence gating are classified according to the lower bound of the window and the deadline. The unique task status is determined according to the preset priority based on the current calculation time, and the longitudinal verification clues are output.

[0014] Further, referring to the task source observation events, the historical source relationships of candidate observation events are traversed to form a source closure. Three-valued evidence gating is performed on independent collection identities and non-pure backreferences, including: determining the observation collection root associated with the task source observation events through collection attribution relationships as the task source observation collection root, wherein the collection attribution relationship indicates that the canonical observation event belongs to the observation collection root; selecting canonical observation events from the canonical observation events that preliminarily match or whose matching result is uncertain with the review target range of the review task as candidate observation events, and the selection is not restricted by the delivery domain marker of the corresponding occurrence record; and aggregating the candidate observations supported by actual collection assertion relationships or the historical source relationships. All occurrence records of the event result are deduplicated and sorted according to their occurrence record identifiers to form a candidate support occurrence record set. Starting from each occurrence record in the candidate support occurrence record set, the historical source relations of type reference, copy, or observation derivation are recursively traversed along the result source direction, while the historical source relations of type comparison are not traversed. The occurrence records and relations obtained from the traversal are collected to form the source closure. Terminal occurrence records that have no historical source relations that can be traversed further, are supported by actual acquisition assertion relations, and whose acquisition attribution relation association status is verified or corroborated as observation acquisition roots are identified as non-derived actual acquisition sources. Only terminal occurrence records that are reached in every traversal path and have no endless circular dependencies are identified. When the source closure is a non-empty complete source closure, and at least one observation root of the non-derived actual acquisition source is independent of the task source observation root, the non-pure backreference component is determined to be true only when the source closure is a non-empty complete source closure, there are no unclassifiable mixed terminals, and each terminal occurrence record constitutes the non-derived actual acquisition source, and at least one observation root of the non-derived actual acquisition source is independent of the task source observation root; when all non-derived actual acquisition sources are associated with the task source observation root and there are no independent non-derived actual acquisition sources, the non-pure backreference component is determined to be false and the corresponding candidate observation event is determined to be a task source pure backreference, wherein the mixed terminal is a node reached during traversal that neither constitutes the non-derived actual acquisition source nor can be determined to be merely a derived intermediate node. The terminal records an event; except in cases where the result is determined to be true or false, the non-pure reference component is determined to be uncertain; when the candidate observation event and the task source observation event are each associated with exactly one observation acquisition root in a verified or corroborated state, and the observation acquisition root associated with the candidate observation event and the task source observation acquisition root belong to different acquisition identity equivalence classes, the acquisition identity independent component is determined to be true; when they belong to the same acquisition identity equivalence class, it is determined to be false; in other cases, it is determined to be uncertain; when all gate control components are true, the gate control process is determined to pass; when any gate control component is false, the gate control process is determined to fail; when there are no false gate control components and at least one gate control component cannot be determined, the gate control process is determined to be uncertain.

[0015] Furthermore, terminal records where there are no further traversable historical source relationships, are supported by actual acquisition assertion relationships, and whose acquisition attribution relationship association status is verified or corroborated are identified as non-derived actual acquisition sources. This includes: obtaining strong acquisition identifiers from combinations of inspection execution identifiers, sample identifiers and completed inspection application identifiers, image inspection identifiers, and inspection equipment output identifiers, and forming acquisition identifier atoms that include at least identifier type, issuing system identifier, identifier value, and validity period. The strong acquisition identifier is an acquisition identifier generated by the execution system after actual acquisition and uniquely points to the actual acquisition action. When the same acquisition identity equivalence class contains at least one strong acquisition identifier verified and valid by the issuing system, and there are no mutually exclusive strong acquisition identifiers, the status of the corresponding observation acquisition root is set to verified. Only when there is a cross-system deterministic mapping, identifier replacement record, equipment migration record, or verified same inspection... During instance mapping, an aliasing relationship satisfying reflexivity, symmetry, and transitivity is established between two acquisition identifier atoms, and acquisition identifier atoms that are aliases to each other are classified into the same acquisition identity equivalence class. When the strong acquisition identifier is missing, acquisition features are formed based on at least three of the following: stable identifier of the inspected object, acquisition time, physical examination items, examined body part, side, examination modality, sample type, device identifier, and execution source. When the acquisition features meet the preset consistency conditions and there are no mutually exclusive fields, the status of the corresponding observation acquisition root is set to verified; otherwise, it is set to source pending confirmation. When the two original acquisition identifiers are different and there is no aliasing relationship, the corresponding observation acquisition roots are not considered to be independent. Instead, based on the non-conflicting information of the inspected object, acquisition time, physical examination items, examination modality, device identifier, and execution source, the observation acquisition roots corresponding to the two original acquisition identifiers are determined to belong to different acquisition identity equivalence classes or to have a source pending confirmation.

[0016] Furthermore, the review recommendations in the occurrence records are anchored to the task source observation event, and a stable task identifier, window lower bound, and deadline for the review task are generated according to the task time anchor point and the fulfillment offset interval. This includes: splitting the review recommendations containing multiple review targets into atomic review tasks, each with a task source observation event supported by an actual acquisition assertion relationship and a clearly defined scope of the review target; and generating a different review recommendation source occurrence record identifier for each atomic review task based on the corresponding occurrence record identifier, the source sub-position within the carrier corresponding to each review target, and the atomic sequence number; and, according to the task time anchor point type, selecting from the observation effective time interval of the task source observation event, the effective time interval of the review recommendation, or the observation... The task time anchor is determined from the later values ​​of each boundary of the effective time interval and the effective time interval of the review suggestion. The task time anchor is a left-closed, right-open, non-empty interval whose lower bound is strictly earlier than its upper bound. If any of the task time anchor type, the task time anchor, or the performance offset interval cannot be determined, or if a valid non-empty task time anchor cannot be formed, the corresponding field is set to a preset time confirmation mark that remains unchanged across data structure versions, and a time uncertainty record is generated. Only when the task time anchor type, the task time anchor, and the performance offset interval are not the time confirmation mark, is the left-closed, right-open suggested performance time interval determined based on the task time anchor and the performance offset interval. In the meantime, the lower bound of the suggested performance time interval is defined as the lower bound of the window, and the upper bound of the suggested performance time interval is defined as the deadline. If a valid, non-empty suggested performance time interval cannot be formed, the lower bound of the window and the deadline are set as time-pending confirmation markers, generating a time-uncertain record, and time classification and state comparisons dependent on the deadline are prohibited. Based on the stability identifier of the inspected object, the identifier of the task source observation event, the identifier of the observation acquisition root associated with the task source observation event, the normalized review target range, the normalized review suggestion type, the task time anchor type, the task time anchor, and the performance offset interval, a stable task marker is generated. The system identifies the electronic medical examination data carrier identifier, carrier generation time, carrier reception time, report issuance time, and medical examination batch time as audit attributes, but does not use them as input for the stable task identifier. When a new review suggestion is traced back to the source of the existing review suggestion through the historical source relationship and a record identifier appears, or when the stable task identifier generated by the new review suggestion is the same as the stable task identifier of the existing review task, the new review suggestion is associated with the existing review task without recreating the review task, and the task time anchor point, the lower boundary of the window, and the deadline of the existing review task are not changed by the carrier generation time, carrier reception time, or report issuance time of the new electronic medical examination data carrier.

[0017] Further, the stable task identifier is generated based on the stability identifier of the inspected object, the task source observation event identifier, the identifier of the observation collection root associated with the task source observation event, the normalized review target range, the normalized review suggestion type, the task time anchor type, the task time anchor, and the performance offset interval. The electronic medical examination data carrier identifier, carrier generation time, carrier reception time, report issuance time, and medical examination batch time are saved as audit attributes and not as inputs to the stable task identifier. This includes: performing unique normalized encoding on each input field of the stable task identifier according to a preset field order and concatenating them according to the preset field order to obtain a stable task identifier field tuple; implementing the stable task identifier based on the stable task identifier field tuple using any one of a deterministic digest, composite primary key, or unique index; and further comparing the stable task identifier when the deterministic digest is used and the digests of two review tasks are the same. The task identifier field tuple is defined to exclude summary collisions; the identifier of the observation collection root associated with the task source observation event is generated by normalized sorting and deterministic encoding of the original collection identifier in the collection identity equivalence class or the verified collection feature tuple, wherein the verified collection feature tuple is a feature tuple obtained by verifying the collection identity based on the collection features, and the collection features are formed based on at least three of the following: the stability identifier of the examined object, collection time, physical examination items, examined parts, lateral, examination modality, sample type, equipment identifier, and execution source; the task source observation event identifier is generated by deterministic encoding of the stability identifier of the examined object, the identifier of the observation collection root associated with the task source observation event, physical examination items, examined parts, lateral, examination modality, sample type, and the observation effective time interval and remains immutable; repeated delivery of the electronic physical examination data carrier does not regenerate the task source observation event identifier and the stability task identifier.

[0018] Furthermore, aliasing and merging are performed on the original acquisition identifiers of the occurrence records. Observational acquisition roots and standardized observation events are established according to acquisition identity equivalence classes. This includes: standardizing the physical examination items, observation values, units of measurement, examined body parts, and effective observation time intervals in the occurrence records; expanding precise times into left-closed, right-open non-empty intervals with a lower bound strictly earlier than the upper bound according to a preset minimum time granularity; expanding times containing only calendar dates into left-closed, right-open intervals from 00:00 on the current day to 00:00 on the next day according to a preset time zone; when a valid non-empty effective observation time interval cannot be formed, a time-determined record is generated, and the effective observation time interval is prohibited from participating in subsequent judgments and time classifications; conflict checking is performed on the original acquisition identifiers, grouping original acquisition identifiers that can jointly identify the same actual acquisition action into the same acquisition identity equivalence class, and assigning each acquisition... An identity equivalence class is used to establish an observation collection root and a deterministic observation collection root identifier. Based on the observation collection root and the standardized physical examination items, a standardized observation event with a deterministic identifier is established. Each standardized observation event is associated with only one observation collection root, but the same observation collection root is allowed to be associated with more than one standardized observation event generated by the same actual collection action. When the original collection identifier intended to be classified into the same standardized observation event points to two or more mutually unequal observation collection roots, the content that can be split according to the record boundary is split into multiple standardized observation events. If splitting is not possible, an observation collection root conflict record is generated. If the observation collection root cannot be uniquely determined, a source unconfirmed record is generated. Standardized observation events whose observation collection root conflict records are not resolved or whose observation collection roots are not confirmed do not form valid performance records.

[0019] Further, candidate observation events that pass the three-valued evidence gating are classified according to the lower bound of the window and the deadline. A unique task status is determined based on a preset priority, combined with the current computation time, and vertical verification clues are output. This includes: performing time classification only on candidate observation events that pass the three-valued evidence gating and whose effective observation time interval and suggested fulfillment time interval are both valid non-empty intervals, wherein the suggested fulfillment time interval is defined by the lower bound of the window and the deadline; when the lower bound of the effective observation time interval of the candidate observation event is not earlier than the lower bound of the window and the upper bound is not later than the deadline, on-time evidence and valid fulfillment records are generated; the lower bound of the effective observation time interval is not... If the deadline is earlier than the specified deadline, overdue evidence and valid performance records are generated; if the observation validity time interval of the candidate observation event strictly crosses the lower bound of the window or the deadline, a time-determined record is generated; if the upper bound of the observation validity time interval is not later than the lower bound of the window, an early candidate record is generated, and no valid performance record is generated in either of the aforementioned two cases; for candidate observation events where the three-valued evidence gating fails, an evidence blocking record containing a reason code is generated; for candidate observation events where the three-valued evidence gating is uncertain, a verification record is generated, and verification records generated due to the inability to determine the observation collection root or the collection identity are marked as source uncertainty type, and verification records generated due to the inability to determine the time sequence are marked as source uncertainty type. Verification records are marked as time-uncertain. Records awaiting verification due to uncertainties in the actual acquisition assertion path, the source closure, or the scope of the review target are marked as incomplete information candidate types. Observational acquisition roots associated with candidate observation events that can change the current earliest fulfillment time classification or task status are determined as decisive candidate observational acquisition roots. A unique task status is determined for each review task according to a preset priority that decreases sequentially: if there are unresolved conflict records involving the task source observation event or that can change the current earliest fulfillment time classification, it is determined as a conflict pending verification status; if there are no unresolved conflict records of the observational acquisition root, but the observations associated with the task source observation event are... When a source or the decisive candidate observation source needs to be confirmed, or when there is a record of uncertain source type to be verified, the status is determined to be "source pending confirmation". When there is no such situation but there is evidence of on-time completion, the status is determined to be "on-time completion". When there is no evidence of on-time completion but there is a record of uncertain time or a record of uncertain time type to be verified, the status is determined to be "uncertain time". When there is no such situation but there is evidence of overdue completion, the status is determined to be "overdue completion". When there is no valid performance record, no record of uncertain source, and no record of uncertain time, and the deadline is not a time pending confirmation marker and the current calculation time is not earlier than the deadline, the status is determined to be "overdue and incomplete".The candidate fulfillment status is determined when the deadline is not a time-pending confirmation marker, the current calculation time is earlier than the deadline, and there are pending verification records of the early candidate record or the candidate type with incomplete information; the open status is determined when none of the aforementioned conditions are met; the current calculation time and the adopted time zone identifier are included in the input digest, and the observation acquisition root identifier, standardized observation event identifier, or stable task identifier are not generated based on the current calculation time, nor are the task time anchor point, the lower bound of the window, or the deadline changed based on the current calculation time; the task status before state transition, the task status after state transition, the current calculation time, the time zone identifier, the input digest, the stable task identifier, the identifier of the candidate observation event, the reason code, the electronic physical examination data carrier identifier corresponding to the evidence and the source location within the carrier, the standardization rule version, the task generation rule version, and the evidence gating rule version are written into the vertical review clue.

[0020] Further, parsing the electronic medical examination data carrier of the same examinee to obtain occurrence records carrying delivery domain markers and evidentiary properties includes: determining the examinee's stable identifier that remains unchanged within the same processing range, and reading the existing medical examination data when corresponding existing medical examination data exists in the storage device, so that the electronic medical examination data carrier itself or together with the existing medical examination data carries medical examination data for at least two observation periods; performing recursive field parsing and source location on the electronic medical examination data carrier, and merging the parsed occurrence records with the occurrence records in the existing medical examination data to obtain multiple occurrence records; each occurrence record is associated with at least the examinee's stable identifier, the electronic medical examination data carrier identifier, and The carrier includes the source location, original content, time information, the original acquisition identifier, historical citation path, and citation depth; the delivery domain marker is used to indicate the current delivery domain, historical delivery domain, or unknown delivery domain; the evidence nature is used to indicate the actual acquisition assertion, citation, copying, comparison, or derivation, and the evidence nature is a multi-valued attribute that allows more than one of these to coexist; the historical citation path and the citation depth are determined by at least one of the following: explicit historical fields, historical object keys, pointers or Uniform Resource Identifiers, past semantics, deterministic content fingerprints, and direct or transmitted historical source relationships; the delivery domain marker is neither a sufficient condition for the three-valued evidence gate to pass nor a sufficient condition for the three-valued evidence gate to fail.

[0021] Furthermore, the method also includes: performing single-item physical examination projection exclusion based on the nature of the evidence, the actual collection assertion relationship, the historical source relationship, the observation collection root, and the physical examination batch corresponding to the electronic physical examination data carrier; distinguishing between the actual observation set of the physical examination batch and the historical projection set embedded in the electronic physical examination data carrier; when generating the actual observation set, not counting the historical projection set as a newly occurring actual collection action of the physical examination batch, while retaining the occurrence record, standardized observation event, observation collection root, and typified relationship corresponding to the historical projection set, wherein the typified relationship is a relationship distinguished by type identifier, including actual collection assertion relationship, collection attribution relationship, and historical source relationship of type reference, copy, comparison, or observation derivation; when the same actual collection action is repeatedly delivered... Only the occurrence record or the historical source relationship is added, without adding the collection identity equivalence class; when the occurrence record located in the historical delivery domain has an original observation valid time interval and an independent observation collection root with a verified or corroborated status, the corresponding normative observation event is not excluded because the occurrence record is located in the historical delivery domain, and the three-valued evidence gating is allowed when all departmental control components are true; when the occurrence record located in the current delivery domain is only referenced back to the task source observation event through the historical source relationship and there is no independent actual collection assertion path, it is prohibited to establish a new observation collection root, normative observation event or valid performance record based on this, and the occurrence record is not used as a new review suggestion source to create a review task, and it is prohibited to reset the task time anchor point, the lower bound of the window and the deadline of the existing review task.

[0022] This invention also provides a physical examination re-examination status control system based on task-relative observation lineage, wherein the task-relative observation lineage refers to the relationship formed between candidate observation events and re-examination tasks in terms of task source observation events, data acquisition identity, and historical source path, including:

[0023] The occurrence record parsing module is used to parse the electronic medical examination data carrier of the same examinee to obtain occurrence records carrying delivery domain markers and evidence nature;

[0024] The data acquisition identity merging module is used to perform alias merging on the original data acquisition identifiers of the occurrence records, and to establish observation acquisition roots and standard observation events according to the data acquisition identity equivalence classes;

[0025] The task anchoring module is used to anchor the review suggestions in the occurrence record to the task source observation event, and generate a stable task identifier, lower window bound and deadline of the review task according to the task time anchor point and the performance offset interval.

[0026] The evidence gating module is used to refer to the task source observation event, traverse the historical source relationships of candidate observation events to form a source closure, and perform three-valued evidence gating on the independent collection identity and non-pure backreference. The candidate observation event is a standard observation event selected from the standard observation events that initially matches or has an uncertain matching result with the review target range of the review task. The non-pure backreference means that the result source of the candidate observation event does not only refer back to the task source observation event.

[0027] The status control module is used to classify candidate observation events that have passed the three-valued evidence gating according to the lower bound of the window and the deadline, determine the unique task status according to the preset priority based on the current calculation time, and output the longitudinal verification clues.

[0028] Because this invention stores the delivery domain marker and evidence nature of the recorded data separately, and establishes equivalence classes of collection identities, observation roots and normative observation events based on the alias relationship and collection characteristics of the original collection identifier, the repeated packaging or repeated delivery of the same actual collection action will no longer form a new collection identity. At the same time, it avoids the mistaken identification of independent collection identities simply because the original collection identifier strings are different, thereby reducing the duplication of normative observation events and misjudgment of sources.

[0029] Because this invention anchors the review recommendations to the task source observation events supported by the actual acquisition assertion relationship, and generates stable task identifiers using the task source observation events, task source observation acquisition roots, normalized review target range, task time anchors, and performance offset intervals, while not using the carrier generation time, carrier reception time, and report issuance time as inputs to the stable task identifiers, it can suppress repeated review tasks and the drift of task time anchors, window lower bounds, and deadlines as carriers are repeatedly delivered.

[0030] Because this invention gathers all occurrence records supporting candidate observation events and recursively traverses the historical source relationships of reference, copy, or observation derivation types to form a source closure, it judges non-pure back-reference components under the condition that the source closure is complete and the terminal source can be confirmed, so that independent successor observations located in the historical delivery domain can participate in fulfillment, while pure back-references of the task source located in the current delivery domain are blocked, thereby reducing the possibility of task self-fulfillment.

[0031] Because this invention uses a three-value processing method of true, false, and uncertain for independent collection identity, non-pure backreference and other gated components, it generates corresponding verification records when the source is missing, source conflicts, circular dependencies, or the collection identity cannot be determined, instead of directly forming a valid performance record, so that incomplete evidence will not be forcibly interpreted as valid performance.

[0032] Because this invention classifies complete intervals according to the lower bound of the window and the deadline, and determines a unique task state according to a fixed priority, while writing the current calculation time, time zone identifier, input summary, reason code, source location and rule version into the vertical verification clue, it can replay the task state under the same input and the same evaluation conditions, and provide verifiable data basis for state changes. Attached Figure Description

[0033] Figure 1 This is a schematic diagram of the overall process of the physical examination re-examination status control method based on task-relative observation of bloodline provided in an embodiment of the present invention;

[0034] Figure 2 This is a schematic diagram showing the correspondence between each processing stage and the overall process in the physical examination follow-up business scenario provided by the embodiments of the present invention;

[0035] Figure 3 This is a schematic diagram of the gating process for relative observation of kinship three-value evidence provided in an embodiment of the present invention;

[0036] Figure 4 This is a schematic diagram of the task anchoring and stable task identifier generation process provided in an embodiment of the present invention;

[0037] Figure 5 This is a schematic diagram of the activity process for parsing occurrence records, collecting and merging identity data, and anchoring review tasks, provided in an embodiment of the present invention.

[0038] Figure 6 This is a schematic diagram of the occurrence record parsing and identity collection merging process provided in an embodiment of the present invention;

[0039] Figure 7 A schematic diagram of the time classification, state control, and vertical review clue output process provided in this embodiment of the invention;

[0040] Figure 8 A schematic diagram of the activity flow for source closure, ternary gating, and state control provided in an embodiment of the present invention;

[0041] Figure 9 This is a schematic diagram of the module structure of the physical examination re-examination status control system based on task-relative observation of bloodline provided in an embodiment of the present invention. Detailed Implementation

[0042] The technical solution of the present invention will be further described below with reference to the accompanying drawings. It should be understood that the embodiments described are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. Where there is no conflict, the technical features of the various embodiments can be combined with each other. All electronic medical examination data carriers, identifiers, time, items, and status data are exemplary data and do not correspond to any specific real object or institution.

[0043] The electronic device referred to in this invention can be a server, a server cluster, a terminal device, or a distributed computing system composed of multiple computing nodes. The electronic device may include a processor, a memory, and a communication interface. The memory stores a computer program, which, when executed by the processor, performs the data parsing, identifier merging, task anchoring, evidence gating, time classification, and state control processing described in the following embodiments.

[0044] Example 1

[0045] like Figure 1 As shown, this embodiment provides a method for controlling the status of physical examination re-examination based on task-relative observation lineage. The task-relative observation lineage refers to the relationship between a candidate observation event and the re-examination task in terms of task source observation events, data acquisition identity, and historical source path. The method is executed by an electronic device and includes the following steps:

[0046] The task-to-observation lineage is not an inherent, isolated attribute of the observed event itself, but rather a relative relationship calculated with the review task as the reference frame and the task source observed event anchored by the review task as the benchmark. It is used to characterize the origin of the candidate observed event's result and whether its collection identity is independent of the task source. Regarding the task source observed event, lineage manifests as the source dependency relationship between the candidate observed event's result source and the task source observed event, i.e., whether there is a path pointing back to the task source observed event when tracing back along the historical source relationship. Regarding the collection identity, lineage manifests as whether the observation collection root associated with the candidate observed event and the task source observation collection root associated with the task source observed event belong to the same collection identity equivalence class. Regarding the historical source path, lineage manifests as the path shape and the occurrence record nature of the path terminal when tracing back level by level along historical source relationships of type citation, copy, comparison, or observation derivation. The same standard observation event can yield different lineage determination results relative to different review tasks. For example, a standard observation event whose source is only a backreference to the task source observation event constitutes a pure task source backreference relative to the current review task, while a standard observation event with independent actual acquisition assertions and independent observation acquisition roots constitutes an independent subsequent observation relative to the same review task. The electronic device distinguishes between independent subsequent observations and pure task source backreferences based on the above three aspects of lineage, thereby identifying true fulfillment and inhibiting task self-fulfillment.

[0047] S1. Analyze the electronic medical examination data carrier of the same examinee to obtain the occurrence records of the delivery domain marker and the nature of the evidence;

[0048] The electronic device first obtains the stability identifier of the examined object and reads one or more electronic medical examination data carriers associated with the stability identifier. The electronic medical examination data carriers can be in the form of structured records, semi-structured documents, resource data, tables, text reports, or combinations thereof. During the parsing process, the source location within the carrier is preserved, allowing the parsing results to be traced back to the original fields, objects, table cells, or character ranges within the carrier.

[0049] An occurrence record is a single, locatable appearance of a specific piece of content within the electronic medical examination data carrier. The delivery field marker describes the location of the occurrence record within the current input structure; it can be the current delivery field, a historical delivery field, or an unknown delivery field. The delivery field only records the area where the content is located; a single occurrence does not necessarily equate to a new actual data collection action. The evidentiary nature describes the source attribute of the content expressed by the occurrence record, which can include one or more of the following: actual data collection assertion, citation, copy, comparison, or derivation. The delivery field marker and evidentiary nature are stored separately to avoid directly inferring whether a new actual data collection action occurred based on the current or historical field where the record is located.

[0050] S2. Perform alias merging on the original acquisition identifiers of the records that appear, and establish observation acquisition roots and standard observation events according to acquisition identity equivalence classes;

[0051] The electronic device extracts the examination execution identifier, sample identifier, test application identifier, imaging examination identifier, examination equipment output identifier, and other original acquisition identifiers from the occurrence records. It then performs format standardization, issuance system verification, and conflict checking on the original acquisition identifiers. Original acquisition identifiers with a basis for determining gender are grouped into the same acquisition identity equivalence class, with each acquisition identity equivalence class corresponding to one observation acquisition root. Whether the text content is the same, whether the carrier number is different, and whether the report generation time is different can only serve as auxiliary information and cannot independently prove that two actual acquisitions are the same or independent; the electronic medical examination data carrier identifier and carrier generation time cannot replace the acquisition identity.

[0052] An observation acquisition root represents the standardized identity of an actual acquisition action. Multiple physical examination items generated from the same actual acquisition action can form multiple standardized observation events, all of which are associated with the same observation acquisition root. When the original acquisition identifier strings are different but lack sufficient independent evidence, the corresponding observation acquisition roots are not directly considered independent. Instead, they are set to have their source pending confirmation or further corroborated by combining non-conflicting acquisition characteristics.

[0053] S3. Anchor the review suggestions in the occurrence record to the task source observation event, and generate a stable task identifier, lower window bound and deadline of the review task according to the task time anchor point and performance offset interval;

[0054] When a review recommendation includes multiple review targets, the electronic device breaks it down into multiple atomic review tasks. Each atomic review task corresponds to a specific review target range and is anchored to a task source observation event supported by actual acquisition assertions. The relative review time in the review recommendation is converted into a left-closed, right-open absolute time interval based on the task time anchor. If the task time anchor type, task time anchor, or fulfillment offset interval cannot be reliably determined, a "pending confirmation" mark is recorded, and the date is not automatically supplemented.

[0055] In one implementation, the stable task identifier can be represented as:

[0056] Equation (1)

[0057] in, Indicates atomic review task Stable task identifier, This represents a deterministic digest function or a unique canonical coding function. Indicates the stability of the tested object. Indicates the source observation event identifier of the task. Indicates the identifier of the root of the task source observation. This indicates the standardized scope of the review objectives. This indicates the standardized type of review recommendation. This indicates the specification code that includes the task time anchor type and the task time anchor. The standard code representing the performance offset range.

[0058] The electronic medical examination data carrier identifier, carrier generation time, carrier reception time, and report issuance time are only stored as audit attributes and are not used as inputs to equation (1). Therefore, when the same review recommendation is repackaged or delivered repeatedly, a new stable task identifier will not be formed simply because of the change in carrier time.

[0059] The task time anchor is a non-empty time interval that is left-closed and right-open. Let the task time anchor be... The performance deviation range is Therefore, the suggested performance time interval can be expressed as:

[0060] Equation (2)

[0061] in, This indicates the suggested timeframe for fulfillment. Indicates the lower bound of the window. Indicates the deadline. and These represent the lower and upper bounds of the task's time anchor point, respectively. and These represent the lower and upper bounds of the performance offset interval, respectively. This indicates a duration addition or calendar addition determined according to the rule version.

[0062] The duration addition method refers to the accumulation of time in fixed-duration units, that is, the duration is converted into days or milliseconds and then added directly to the time. The calendar addition method refers to the accumulation of time field by field according to the Gregorian calendar, that is, the year, month, and day fields are first accumulated according to calendar rules, and then the time field is added. If the accumulated result exceeds the last day of the month, it is truncated to the last day of the month. When the boundary of the performance offset interval is a calendar offset (e.g., six months), calendar addition is used, and when it is a fixed duration (e.g., 180 days), duration addition is used. The specific method is fixed in advance by the rule version and does not change with the current calculation time within the same rule version.

[0063] When the task time anchor type, task time anchor, or performance offset range cannot be determined, the electronic device uses a time confirmation mark that remains unchanged across data structure versions, and does not perform normal time classification based on uncertain time.

[0064] S4. Referring to the task source observation event, traverse the historical source relationships of candidate observation events to form a source closure, and perform three-valued evidence gating on the independent collection identity and non-pure backreference. The candidate observation event is a standard observation event selected from the standard observation events that initially matches or has an uncertain matching result with the review target range of the review task. The non-pure backreference means that the result source of the candidate observation event does not only refer back to the task source observation event.

[0065] The electronic device selects candidate observation events from the standardized observation events that preliminarily match or whose matching result is uncertain with the review target range of the review task. This selection is not limited by the delivery domain marker of the corresponding occurrence record; that is, candidate observation events are not pre-excluded according to the delivery domain marker. Candidate observation events may be selected from the current delivery domain, historical delivery domain, and unknown delivery domain. When the source closure of a candidate observation event contains at least one independent source path that does not refer back to the task source observation event, the result source of the candidate observation event is not solely a reference to the task source observation event, thus constituting a non-pure backreference.

[0066] For each candidate observation event, the electronic device collects all occurrence records that support its result through actual acquisition assertion relations or historical source relations, recursively traversing the historical source relations of reference, copy, and observation-derived types to form a source closure. Comparison relations only express the comparative semantics between two observations and do not indicate that a new result originates from an old result, nor do they serve as the direction of result source for further traversal.

[0067] Let the set of values ​​for each gated component be . ,in To indicate truth, Indicates false. If the uncertainty is indicated, then the gating process can be represented as:

[0068] Equation (3)

[0069] in, Represents candidate observation events Compared to the review task Gating control processing, Describes the three-valued conjunction function. Indicates the actual collected assertion components. This indicates that the collected identity components are independent. Represents time-series components, Indicates the target range components to be reviewed. This indicates non-pure backreference components. All components are... Time-gated processing is passed; any component is The time-gating process failed; no value exists. The components and at least one component is The timing of gating is uncertain.

[0070] The three-valued conjunction function is a predefined three-valued logic conjunction operation. Its output still takes values ​​from a set of 1, 0 and uncertain values, and the result of the operation is independent of the order of the five gate components. Unlike the two-valued logic conjunction, the uncertain value of the component does not change the determining role of the false value of the component in the gate handling. That is, when any gate component is proven to be false, the gate handling will fail regardless of whether the other components can be determined.

[0071] Since gating uses the task source observation event and the task source observation acquisition root as references, the same candidate observation event can have different task-relative observation lineage for different review tasks.

[0072] S5. Classify the candidate observation events that have passed the three-valued evidence gating according to the lower bound of the window and the deadline, determine the unique task status according to the preset priority based on the current calculation time, and output the longitudinal verification clues.

[0073] For candidate observation events that have passed the gated processing and whose time intervals are valid, the electronic equipment classifies them as on-time evidence, overdue evidence, early candidate records, or time-determined records based on the relationship between their valid observation time intervals and the recommended performance time intervals.

[0074] The electronic device sequentially determines the following statuses according to preset priorities: conflict pending verification, source pending confirmation, on-time completion, uncertain time, overdue completion, overdue non-completion, candidate fulfillment, and open status. The first status that meets the criteria is designated as the unique task status. This preset priority order is used to stably output the current task status when multiple facts exist simultaneously in the same review task, and is not the order in which the status is displayed on the interface.

[0075] The current calculation time and time zone identifier are included in the input summary of the state assessment, but are not used to generate the observation acquisition root identifier, normative observation event identifier, or stable task identifier, nor are they used to change the task time anchor point, window lower bound, or deadline. The longitudinal review thread records the task status, candidate evidence, blocking reasons, source location, input summary, and rule version before and after the state migration, enabling the state assessment to be replayed under the same input conditions. The longitudinal review thread continuously retains records of the task source observation events, candidate sources, gating components, time classification, state results, reasons, and relevant audit attributes around the same stable task identifier, for recalculation and tracing after the arrival of the new carrier.

[0076] In terms of data organization, the occurrence record stores the stable identifier of the examined object, its electronic medical examination data carrier, delivery domain, content location, evidence nature, source citation, original acquisition identifier, and observation validity time; the standardized observation event stores the standardized medical examination items, results, units of measurement, examined body parts, observation validity time interval, and uniquely associated observation acquisition root; the atomic review task stores the components of the stable task identifier and the suggested fulfillment time interval; the gating and status record stores each gating component, reason, time classification, current task status, and decisive source. The processing entry receives the set of electronic medical examination data carriers of the same examined object and the current calculation time, and sequentially performs recursive parsing to form the occurrence record, standardize and merge the acquisition identity, split and anchor the review task, select candidate observation events according to the review target range and construct source closures to execute three-valued evidence gating, perform time classification on the candidate observation events that pass the gating, and summarize the task status according to the preset priority. When a new carrier arrives, it is associated with the existing review task with the stable task identifier and recalculated, without resetting the task time boundary. In one implementation, the aforementioned processing stages can be achieved by integrating content splitting nodes, terminology standardization nodes, suggestion and report generation nodes, comparative semantic processing nodes, and process status management nodes into the existing processing flow of the physical examination general review business. The interface form of each node does not constitute a limitation on the processing rules. This implementation does not depend on specific program files, and there are no restrictions on specific database types, programming languages, or deployment methods.

[0077] To facilitate understanding, the following example uses a complete processing procedure to illustrate the execution of each step above. All identifiers, dates, and results are illustrative. Taking a physical examination image review as an example, the stability identifier for the examined subject is P001. Carrier M1 was delivered on February 25, 2025, and belongs to the current delivery domain. It carries record r1, which has the nature of actual acquisition assertion. The original acquisition identifier is PACS-CT-001. It records the chest CT nodule shadow results acquired on February 20, 2025, and the recommendation for a follow-up chest CT within 6 to 7 months. Carrier M2 was delivered on April 10, 2025, and belongs to the current delivery domain. It carries record r2, which has the nature of copying and referencing. It copies r1 to the current region and points the source to r1. Carrier M3 was delivered on October 2, 2025, and belongs to the historical delivery domain. It carries record r3, which has the nature of actual acquisition assertion and comparison. The original acquisition identifier is PACS-CT-019. It records the new results acquired on August 28, 2025, and includes semantics comparing with the previous results. Although r2 is located in the current delivery domain of the new carrier, it is a copy and reference and the source returns to r1; although r3 is located in the historical delivery domain, it contains an independent actual collection assertion and a new original collection identifier. Therefore, the delivery domain only indicates the location of the content and cannot determine the evidentiary value on its own.

[0078] During the acquisition identity merging and task anchoring stage, after completing the format standardization and association verification of PACS-CT-001, an observation acquisition root R1 is formed. The nodule shadow results and review suggestions in r1 are associated with the standard observation event O1. r2 does not have a new actual acquisition identifier and the source path returns to r1, so it does not form a new observation acquisition root independent of R1. After completing the standardization and verification of PACS-CT-019, an observation acquisition root R2 is formed. The review results in r3 form a standard observation event O2 and are associated with R2. In the existing evidence, there is no alias merging relationship between R2 and R1, and R2 can be used as an independent acquisition identity relative to the task source. An atomic review task with the goal of chest CT review is generated from the suggestions in O1. The task source observation event is O1, and the task source observation acquisition root is R1. The stable task identifier is generated according to formula (1). The carrier identifiers and delivery times of M1, M2, and M3 are not included in the stable task identifier. Therefore, the copying suggestion will not create a new review task. O1 only has date precision, and its effective observation time interval is the left-closed and right-open interval from 00:00 on February 20, 2025 to 00:00 on February 21, 2025; the performance offset interval in this example is 6 months to 7 months, and the recommended performance time interval is obtained by formula (2) according to calendar addition, which is the left-closed and right-open interval from 00:00 on August 20, 2025 to 00:00 on September 20, 2025.

[0079] During the source closure and three-valued evidence gating phase, when constructing the source closure for r2, the effective source only returns to the observation acquisition root R1 associated with r1, and there are no new actual acquisition assertions. At least one of the actual acquisition assertion component, the acquisition identity independent component, and the non-pure backreference component is false, so the gating process fails, r2 forms an evidence blocking record, and the task cannot be closed. When constructing the source closure for r3, the comparison with the previous result is a comparative relationship and is not used as a source path for further backtracking. The actual acquisition source is still R2. r3 has an actual acquisition assertion, R2 is independent of R1, and the effective observation time interval of O2 is the left-closed and right-open interval from 00:00 on August 28, 2025 to 00:00 on August 29, 2025, which is within the suggested performance time interval. The re-examination target is within the scope of chest CT review and is not only a backreference task source. All gating components are true, so the gating process passes. Once O2 passes the gate, it forms evidence of on-time completion and a valid performance record. In the absence of higher-priority conflicts or unconfirmed source matters, the task status is determined as on-time completion. Although M3 was delivered on October 2, 2025, the actual collection date falls within the suggested performance time range. The later delivery time does not change the on-time classification; it is only retained as an audit attribute. Vertical review of the clue records stabilizes the task identifier, the source relationship between O1, R1, O2, R2, r2, and r3, each gate component, blocking reason, time classification, decisive source, and final status. Conversely, if after the deadline there are only r2 candidates without any independent candidates passing the gate, and both the task source observation event and the task time anchor point have been determined, the task enters the overdue and incomplete status according to the preset priority. If the actual collection source of the candidate is insufficient for confirmation, the gate is treated as uncertain and a record pending verification or a source unconfirmed status is retained.

[0080] To facilitate understanding of the relationship between the above business scenarios and the overall process, such as Figure 2 As shown in the figure, the correspondence between each processing stage and the overall process in the physical examination review business scenario provided in this embodiment is as follows: In the record parsing stage, the electronic physical examination data carriers M1, M2, and M3 respectively obtain the occurrence records r1, r2, and r3; In the identity standardization and merging stage, the original collection identifiers PACS-CT-001 and PACS-CT-019 respectively establish observation collection roots R1 and R2 and standard observation events O1 and O2, and r2 does not form a new observation collection root because its source returns to r1; In the review task anchoring stage, O1 generates an atomic review task q and calculates the suggested performance time interval; In the candidate selection, source closure, and three-value gating stage, the gating of r2 fails and the blocking reason is recorded, and all gating components of r3 are true; In the time classification, status summary, and auditing stage, when the effective time of O2 is within the suggested performance time interval, the task q is determined to be completed on schedule, and the later delivery time is only retained as an audit attribute.

[0081] Example 2

[0082] like Figure 3 As shown, based on Example 1, referring to the task source observation events, the historical source relationships of candidate observation events are traversed to form a source closure. Three-valued evidence gating is performed on the independent acquisition identities and non-pure backreferences, including:

[0083] S4.1. The observation collection root associated with the task source observation event through the collection attribution relationship is determined as the task source observation collection root, wherein the collection attribution relationship is a relationship that indicates that the standard observation event belongs to the observation collection root;

[0084] The task source observation acquisition root serves as the reference object for acquisition identity comparison and pure backreference judgment. Regarding acquisition identity, the independent component of acquisition identity uses the acquisition identity equivalence class to which the task source observation acquisition root belongs as the comparison benchmark, comparing the acquisition acquisition roots associated with candidate observation events with the task source observation acquisition root at the equivalence class level. Regarding pure backreference, the non-pure backreference component uses the task source observation acquisition root as the backreference judgment benchmark, determining whether each non-derived actual acquisition source in the source closure is still associated with the task source observation acquisition root. The task source observation acquisition root of each review task remains unchanged after being determined in the anchoring phase of the review task, and does not change with different candidate observation events, ensuring comparability between the gating judgment results of each candidate observation event under the same review task.

[0085] When a task source observation event is not associated with a unique observation root whose status is verified or corroborated, the corresponding acquisition identity component is determined to be uncertain, rather than arbitrarily selecting one from multiple possible observation roots. Arbitrarily selecting an observation root as a reference would disguise the lack of evidence as a definite conclusion, causing the gating result to change depending on the selection order, thus losing replayability. In this case, the independent acquisition identity component and the backreference comparisons in non-pure backreference components that depend on this reference object cannot be determined to be true or false, but can only be determined to be uncertain.

[0086] S4.2. Select from the standard observation events those that preliminarily match or whose matching result is uncertain with the scope of the review target of the review task as candidate observation events, and the selection is not limited by the delivery domain marker of the corresponding occurrence record;

[0087] The scope of the review target can include examination items, examined body parts, laminar flow, examination modality, sample type, and review purpose. Candidate selection is performed using a three-value method: When there is a hard constraint conflict between the standard observation event and the review target scope, such as different examination items, opposite laminar flow corresponding to the examined body parts, or an examination modality that contradicts the clearly defined review purpose, it is judged as a mismatch, and the corresponding standard observation event is not considered a candidate observation event; when the required attributes of the standard observation event are missing and cannot be completed through a verified mapping, such as the review suggestion only describing the review imaging examination without specifying the examination modality, and there is no empirical mapping in the existing evidence to determine the examination modality, the matching result is judged as uncertain, and the corresponding standard observation event is still retained as a candidate observation event; except for the above situations, when each comparable field of the standard observation event is consistent with or within the allowable range of the review target scope, it is judged as a preliminary match. Standard observation events with uncertain matching results are retained as candidates, allowing boundary cases to be further determined by the review target scope component in the three-value evidence gating, avoiding exclusion conclusions based on incomplete evidence at the candidate selection stage.

[0088] The selection is not limited by the delivery domain marker of the corresponding occurrence record; that is, the selection is performed in the current delivery domain, the historical delivery domain, and the unknown delivery domain. Canonical observation events in the historical delivery domain are not excluded solely because of the packaging location.

[0089] S4.3. Collect all occurrence records that support the candidate observation event results through actual collection assertion relationships or the historical source relationships, remove duplicates according to the occurrence record identifier and sort them in a standardized manner to form a candidate support occurrence record set;

[0090] The standardized sorting can be based on the occurrence record identifier, carrier identifier, and source position within the carrier in that order, ensuring that the set of candidate supporting occurrence records remains stable when input records are rearranged or repeatedly submitted. Multiple supporting records cannot be arbitrarily chosen to replace the complete set.

[0091] The standard sorting is a deterministic sorting whose output order is independent of the input order. It compares the objects to be sorted byte by byte according to their coded byte sequence, or field by field if a preset field priority exists. Objects with the same comparison result are considered as the same object. Performing standard sorting on the same set of objects twice will result in the same arrangement.

[0092] S4.4. Starting from each occurrence record in the candidate support occurrence record set, recursively traverse the historical source relations of type reference, copy or observation derivation along the result source direction, while not traversing the historical source relations of type comparison, and gather the occurrence records and relations obtained by traversal to form the source closure;

[0093] The recursive traversal proceeds along the direction of the result source, starting from the occurrence record that supports the candidate observation event result and pointing to the existing occurrence record that produced the result, opposite to the direction of content delivery or display. During traversal, the electronic device reads historical source relations of type reference, copy, or observation derivation from the historical reference path of the current occurrence record. After reaching the referenced occurrence record, it continues to traverse in the same direction until there are no more historical source relations for the current record to be traversed. Each occurrence record and each historical source relation reached during the traversal is included in the source closure. The two observations connected by the comparison relation are not the result sources of each other, so the comparison relation is not used as a traversal path.

[0094] During recursive traversal, the identifiers of visited occurrence records and their relationships are recorded to prevent repeated traversal of the same path and to avoid infinite traversal. When a visited record is reached again, the electronic device identifies the circular dependency and determines whether there are still exits in the loop that can reach the terminal source, i.e., whether there are unvisited branches in the loop that can reach the terminal occurrence record. If all exits of the loop are within the loop and no path to the terminal occurrence record can be formed, then the loop is a circular dependency without an endpoint. Circular dependencies without an endpoint cannot form a complete source closure. When the source closure is therefore incomplete, the corresponding gating component is determined as indeterminate using a three-valued method, rather than forcibly making a true or false judgment in the absence of a path.

[0095] S4.5. The terminal occurrence record of the observation collection root that does not have a historical source relationship that can be traversed further, is supported by the actual collection assertion relationship, and is associated with the collection attribution relationship as verified or corroborated, is determined as a non-derived actual collection source.

[0096] For a terminal to appear as a record, three conditions must be met simultaneously: First, the historical source traversal has terminated, meaning that there are no historical source relationships of type reference, copy, or observation derivation that can be continued for the terminal to appear as a record, and there are no unvisited source paths to continue the traversal at this record; Second, there is an actual acquisition assertion relationship, meaning that the evidentiary nature of the terminal to appear as a record includes an actual acquisition assertion, or there is an actual acquisition assertion relationship pointing from the terminal to appear as a record to its observation acquisition root, indicating that its result is generated by actual acquisition action rather than derived from existing observations; Third, the state of the corresponding observation acquisition root can be confirmed, meaning that the state of the observation acquisition root associated with the acquisition attribution relationship has been verified or corroborated.

[0097] Terminal records that only have report generation time, similar text content, or reference pointers are not considered non-derived actual acquisition sources. Report generation time only indicates when the report was generated, not the actual acquisition action; text similarity only indicates similar content, not an assertion that the result originates from the similar record; reference pointers only indicate the location of the content, requiring tracing back to the actual acquisition point via the corresponding reference pointer. These three types of records either continue to serve as the starting point for tracing back to their source, or their source is uncertain, causing the corresponding gating component to be determined as uncertain. Non-derived actual acquisition sources are the terminal basis for distinguishing non-pure backreference components. Strict identification can avoid misjudging copied records, paraphrased records, or comparison records as independent acquisitions, and prevent pure backreferences from the task source from being misjudged as non-pure backreferences.

[0098] S4.6. The source closure is determined to be a complete source closure only when the terminal is reached in every traversal path and there is no endless circular dependency;

[0099] For example, if a candidate result is supported by two occurrence records, one path leads to an independent actual source, and the other path terminates at a derived node lacking a source, then the source closure is still determined to be incomplete, and the latter path cannot be ignored to directly determine that the non-pure backreference component is true.

[0100] S4.7. Under the conditions that the source closure is a non-empty complete source closure, there are no unclassifiable mixed terminals, and each terminal occurrence record constitutes the non-derived actual acquisition source, when at least one observation acquisition root of the non-derived actual acquisition source is independent of the task source observation acquisition root, the non-pure backreference component is determined to be true; when all the non-derived actual acquisition sources are associated with the task source observation acquisition root and there are no independent non-derived actual acquisition sources, the non-pure backreference component is determined to be false and the corresponding candidate observation event is determined to be a pure backreference of the task source, wherein the mixed terminal is a terminal occurrence record reached during traversal that neither constitutes the non-derived actual acquisition source nor can be determined to be merely a derived intermediate node;

[0101] In this embodiment, a non-empty source closure means that the source closure contains at least one occurrence record; an empty source closure indicates that the candidate observation event does not have any supporting occurrence records, and the source of the result cannot be determined, so the non-pure backreference component can only be determined as uncertain. A complete source closure means that each traversal path reaches the terminal occurrence record and there is no endless circular dependency; an incomplete closure indicates that there is at least one source path with an unknown destination, and there is a gap in the evidence chain, so the non-pure backreference component can also only be determined as uncertain. There are no unclassifiable mixed terminals, meaning that every terminal occurrence record reached in the traversal can be clearly classified; mixed terminals neither constitute non-derived actual collection sources nor can they be determined to be merely derived intermediate nodes. For example, a terminal occurrence record may lack an actual collection assertion but may still have an unrecorded upstream source. Such dangling terminals indicate that the source chain is not exhaustive, and when mixed terminals exist, the non-pure backreference component can also only be determined as uncertain. Each terminal record constitutes a non-derived actual data collection source, requiring all terminals within the source closure to be verifiable actual data collection sources. If there are derived intermediate nodes treated as terminals, upstream actual data collection sources will be missed, making the determination of true or false unreliable. Observational data collection roots are independent of task source observational data collection roots, meaning that the observational data collection roots associated with non-derived actual data collection sources belong to different data collection identity equivalence classes than the task source observational data collection roots.

[0102] When a candidate observation event has both independent actual acquisition support records and task source replication records, the non-pure backreference component can be determined as true as long as at least one non-derived actual acquisition source in the complete source closure is independent of the task source observation acquisition root. Conversely, when all terminals return the task source observation acquisition root, the candidate observation event constitutes a pure task source backreference. To determine the non-pure backreference component as false and as a pure task source backreference, all non-derived actual acquisition sources must be associated with the task source observation acquisition root and there must be no independent non-derived actual acquisition source. That is, every actual acquisition source traceable in the source closure returns to the task source itself, and the result of the candidate observation event is merely a reproduction of the task source observation result, such as a replication, reference, or repackaging of the task source result, rather than a newly occurring independent acquisition. When the wrapper layer of a candidate observation event declares this acquisition, but all non-derived actual acquisition sources in its source closure are still associated with the task source observation acquisition root, the non-pure backreference component is still determined as false, and the corresponding candidate observation event is still blocked as a pure task source backreference.

[0103] S4.8. Except in cases where it is determined to be true or false, the non-pure backreference component is determined to be indeterminate;

[0104] When the source closure is empty, incomplete, contains unclassifiable mixed terminals, lacks actual acquisition assertions for the terminal, the terminal's observation acquisition root is yet to be confirmed, or there is an untermined circular dependency, the non-pure backreference component can be determined as uncertain.

[0105] S4.9. When the candidate observation event and the task source observation event are each associated with exactly one observation acquisition root in a verified or corroborated state, and the observation acquisition root associated with the candidate observation event and the task source observation acquisition root belong to different acquisition identity equivalence classes, the acquisition identity independent component is determined to be true; when they belong to the same acquisition identity equivalence class, it is determined to be false; in other cases, it is determined to be uncertain.

[0106] This judgment uses the equivalence class of the collected identity rather than the original collected identifier string. When the original collected identifier is different but the collected evidence is insufficient, the independent component of the collected identity is not directly determined as true.

[0107] S4.10. When all gate control components are true, the gate control process is determined to pass; when any gate control component is false, the gate control process is determined to fail; when there are no false gate control components and at least one gate control component cannot be determined, the gate control process is determined to be uncertain.

[0108] A gated component that is uncertain indicates insufficient evidence; it cannot be treated as false nor presumed to be true. Tri-valued gating preserves the difference between missing evidence and negated evidence. Candidate observations with gated failure generate evidence blocking records, while candidate observations with gated uncertainty generate records pending verification; neither directly generates a valid performance record.

[0109] Example 3

[0110] Based on Example 2, the terminal occurrence records of observation collection roots that do not have a historical source relationship that can be further traversed, are supported by actual collection assertion relationships, and whose status is verified or corroborated by collection attribution relationships are determined to be non-derived actual collection sources, including:

[0111] S4.5.1. Obtain a strong acquisition identifier from the combination of the inspection execution identifier, sample identifier and completed inspection application identifier, image inspection identifier and inspection equipment output identifier, and form an acquisition identifier atom that includes at least the identifier type, the issuing system identifier, the identifier value and the validity period, wherein the strong acquisition identifier is an acquisition identifier generated by the execution system after the actual acquisition is completed and can uniquely point to the actual acquisition action;

[0112] In one implementation, the strong acquisition identifier is the original identifier issued by the execution system itself, and the acquisition identifier atom is the smallest data unit formed after the strong acquisition identifier is structurally decomposed. After an electronic device acquires a strong acquisition identifier, it decomposes the strong acquisition identifier into fields such as identifier type, issuing system identifier, identifier value, and validity period according to a unified field structure, forming an acquisition identifier atom. That is, one strong acquisition identifier corresponds to one acquisition identifier atom. The original value of the strong acquisition identifier is stored in the identifier value field of the corresponding acquisition identifier atom. The acquisition identifier atom supplements the strong acquisition identifier with structured information such as type, source, and validity period, so that subsequent issuing system verification, alias relationship establishment, and acquisition identity equivalence classification are all performed at the acquisition identifier atom as a unified operation granularity, without relying on the free format of the original identifier of each execution system.

[0113] The strong acquisition identifier comprises identifiers originating from different acquisition stages: the inspection execution identifier is assigned by the execution system for a single inspection execution, directly pointing to the execution action of that inspection; the combination of the sample identifier and the completed inspection application identifier is applied to inspection-related items, with the sample identifier pointing to the collected sample and the completed inspection application identifier indicating that the corresponding inspection application has been completed, the combination uniquely pointing to a completed inspection; the image inspection identifier is generated by the image inspection process, pointing to the acquisition of an image inspection; the inspection equipment output identifier is generated by the inspection equipment when outputting the acquisition results, pointing to a single result output by the equipment. The strong acquisition identifier is generated by the execution system after the actual acquisition is completed, meaning that pre-assigned appointment numbers or incomplete application numbers before acquisition completion do not belong to the strong acquisition identifier; its ability to uniquely point to the actual acquisition action means that the same strong acquisition identifier does not correspond to two different actual acquisitions.

[0114] The acquisition identifier atom can also carry completion status, issuance time, and verification digest. The identifier type in the acquisition identifier atom is used to distinguish the identifiers contained in the strong acquisition identifier and determine the corresponding verification method; the issuance system identifier is used to confirm the issuance source and establish a deterministic mapping in cross-system scenarios; the identifier value is the actual value of the identifier; the validity period is used to limit the time range within which the identifier can be trusted for use; strong acquisition identifiers that have exceeded their validity period are no longer used individually for verification determination; the completion status is used to distinguish between completed and ongoing acquisition; the issuance time records the moment the identifier was generated; and the verification digest is used to verify that the identifier content has not been tampered with since issuance.

[0115] The report number or carrier number primarily identifies the document itself. The same report document can be copied, converted in format, or reissued. The report number changes as the document circulates and does not point to the actual data collection action that generated the report content.

[0116] In the absence of other data collection evidence, it is not considered a strong data collection identifier, and the corresponding original data collection identifier is processed according to the data collection feature formation and corroboration method.

[0117] S4.5.2. When the same collection identity equivalence class contains at least one strong collection identifier that has been verified and valid by the issuing system, and there are no mutually exclusive strong collection identifiers, the state of the corresponding observation collection root is set to verified.

[0118] Mutual exclusion scenarios can include inconsistent subjects, two incompatible execution instance identifiers, sample type conflicts, device output instance conflicts, or clearly incompatible acquisition periods. A root conflict record for observation acquisition is generated when a strong mutual exclusion acquisition identifier exists.

[0119] S4.5.3. Only when there is a cross-system deterministic mapping, identifier replacement record, device migration record, or a verified mapping of the same inspection instance, establish an alias relationship that satisfies reflexivity, symmetry, and transitivity between two acquisition identifier atoms, and classify acquisition identifier atoms that are aliases to each other into the same acquisition identity equivalence class;

[0120] Let the collection identifier atom set be The alias relationship is Then collect the labeled atoms The identity equivalence class to which it belongs can be represented as:

[0121] Equation (4)

[0122] in, express The identity equivalence class to which it belongs, and Indicates the collection of identifier atoms, This indicates the collection of the complete set of atoms that identify the data. This indicates an alias relationship with a definite basis.

[0123] S4.5.4. When the strong acquisition identifier is missing, the acquisition feature is formed based on at least three of the following: the stability identifier of the object under examination, the acquisition time, the physical examination item, the examination site, the side, the examination mode, the sample type, the equipment identifier, and the execution source. When the acquisition feature meets the preset consistency condition and there are no mutually exclusive fields, the status of the corresponding observation acquisition root is set to verified; otherwise, it is set to source pending confirmation.

[0124] In one implementation, the stability identifier of the examined object and the examination items may be consistent, and at least one of the following factors—collection time, examined site, examination modality, sample type, equipment identifier, and execution source—may be selected to form at least three collection features. Preset consistency conditions and allowable errors are fixedly saved in the rule version.

[0125] Mutually exclusive fields are determined as follows: when the stability identifiers of the inspected objects are inconsistent, the inspection modalities are different, or the sample types are different, the corresponding acquisition features constitute mutually exclusive fields; when the acquisition time is inconsistent and exceeds the error allowed by the preset consistency condition, it also constitutes a mutually exclusive field; when any mutually exclusive field exists, regardless of whether the other acquisition features are consistent, the preset consistency condition is not met.

[0126] S4.5.5. When the two original acquisition identifiers are different and there is no alias relationship, the corresponding observation acquisition roots are not considered to be independent of each other. Instead, based on the non-conflicting examination object, acquisition time, examination items, examination mode, equipment identifier and execution source information, the observation acquisition roots corresponding to the two original acquisition identifiers are determined to belong to different acquisition identity equivalence classes or to be determined to have a source pending confirmation.

[0127] For example, if two original collection identifiers are issued by different execution instances, their collection time intervals do not overlap, and their device output instances are different, they can be identified as different collection identity equivalence classes; if only the identifier string is different and other collection information is missing, then they are identified as having a source that needs to be confirmed.

[0128] Example 4

[0129] like Figure 4 As shown, based on Example 1, the review recommendations in the occurrence records are anchored to the task source observation event, and a stable task identifier, lower window bound, and deadline for the review task are generated according to the task time anchor point and performance offset interval, including:

[0130] S3.1. The review proposal containing multiple review targets is split into atomic review tasks, each having a task source observation event supported by an actual acquisition assertion relationship and a clearly defined scope of the review target. Based on the corresponding occurrence record identifier, the source sub-position in the carrier corresponding to each review target, and the atomic number, a different review proposal source occurrence record identifier is generated for each atomic review task.

[0131] For example, when a follow-up recommendation includes two medical examination items, two atomic follow-up tasks can be formed. Even if the two targets are located in the same occurrence record, different follow-up recommendation source occurrence record identifiers can be generated by the source sub-position and atomic number within the carrier, thereby performing task anchoring and subsequent state control separately.

[0132] S3.2. Based on the task time anchor point type, determine the task time anchor point from the observation effective time interval of the observation event of the task source, the review suggestion effective time interval, or the later value of each boundary of the observation effective time interval and the review suggestion effective time interval. The task time anchor point is a left-closed and right-open non-empty interval whose lower bound is strictly earlier than the upper bound.

[0133] The task time anchor type is a predefined enumerated field, including observation anchor type, effective anchor type, and later boundary anchor type. The observation anchor type indicates that the effective observation time interval of the task source observation event is used as the task time anchor point, suitable for situations where the time granularity of the effective observation time interval is reliable and the review rules are based on the actual acquisition time. The effective anchor type indicates that the effective time interval of the review suggestion is used as the task time anchor point, suitable for situations where the review suggestion itself specifies the start and end dates of the suggestion's effectiveness, but the granularity of the effective observation time is not suitable as a benchmark. The later boundary anchor type indicates that the later value between the lower bound of the effective observation time interval and the effective review suggestion time interval, and the later value between the upper bounds of the two intervals, are used to form the task time anchor point. The task time anchor type is pre-specified by the review rule version and does not change with the current calculation time within the same rule version. The effective review suggestion time interval is the time interval from the start time of the suggestion's effectiveness to the end time specified in the review suggestion, expanded into a left-closed, right-open non-empty interval according to the same normalization method as the effective observation time interval.

[0134] When the rules require that a valid performance record not be formed before the review recommendation takes effect, the later value of each boundary between the valid observation time interval of the task source observation event and the effective time interval of the review recommendation can be taken.

[0135] S3.3. If any of the task time anchor type, the task time anchor, or the performance offset interval cannot be determined, or if a valid non-empty task time anchor cannot be formed, the corresponding field is set to a preset time confirmation mark that remains unchanged across data structure versions and a time uncertainty record is generated.

[0136] In this step, if any of the task time anchor type, the task time anchor, or the performance offset interval cannot be determined, the following situations apply: the task time anchor type field is missing or its value does not belong to the preset enumeration; the source time interval corresponding to the selected anchor type is missing, is a time pending confirmation marker, or does not meet the interval condition that the lower bound is strictly earlier than the upper bound; the performance offset interval is missing, its boundary is inverted, or it contains illegal time values; a valid non-empty task time anchor cannot be formed based on the observed valid time interval and the review suggestion effective time interval. When any of the above situations occur, only the corresponding field that cannot be determined is set as a time pending confirmation marker, and the remaining fields that can be reliably determined are still saved according to their standard encoding; the time uncertainty record at least records the field that cannot be determined, the source occurrence record identifier, and the triggering situation, so that the task anchoring can be re-executed based on the deterministic correction basis later.

[0137] The time-pending flag is a reserved enumeration value that does not collide with any valid time encoding. Its encoding does not change with data structure versions to ensure that the same uncertain input can generate the same stable task identifier field tuple. The time-pending flag is used to record situations where the time cannot be reliably determined, and does not automatically pad the date.

[0138] S3.4. Only when the task time anchor type, the task time anchor and the performance offset interval are not the time to be confirmed mark, determine the left-closed and right-open suggested performance time interval according to the task time anchor and the performance offset interval, determine the lower bound of the suggested performance time interval as the lower bound of the window, and determine the upper bound of the suggested performance time interval as the deadline.

[0139] When determining the suggested fulfillment time interval, the lower bound of the task time anchor point and the lower bound of the fulfillment offset interval are added using the duration addition or calendar addition specified in the rule version to obtain the included lower bound of the suggested fulfillment time interval, i.e., the window lower bound. The upper bound of the task time anchor point and the upper bound of the fulfillment offset interval are added using the same addition method to obtain the excluded upper bound of the suggested fulfillment time interval, i.e., the deadline. If the sum of the lower bound of the task time anchor point and the lower bound of the fulfillment offset interval is not earlier than the sum of the upper bound of the task time anchor point and the upper bound of the fulfillment offset interval, a valid non-empty suggested fulfillment time interval cannot be formed. In this case, the above boundary determination is not performed, and the window lower bound and the deadline are set as time pending confirmation markers.

[0140] The left-closed, right-open interval clearly defines the assignment of boundary times. When the lower bound of the candidate observation's valid time interval equals the deadline, it is classified as overdue (not earlier than the deadline) and not simultaneously classified as on-time.

[0141] S3.5. When a valid non-empty suggested performance time interval cannot be formed, the lower bound of the window and the deadline are set as the time pending confirmation marker, a time uncertain record is generated, and time classification and status comparison dependent on the deadline are prohibited.

[0142] This avoids using empty intervals, inverted upper and lower bound intervals, or time values ​​of unknown origin to perform overdue and incomplete status checks.

[0143] S3.6. Generate the stable task identifier based on the stability identifier of the inspected object, the task source observation event identifier, the identifier of the observation collection root associated with the task source observation event, the normalized review target range, the normalized review suggestion type, the task time anchor type, the task time anchor, and the performance offset interval. Save the electronic physical examination data carrier identifier, carrier generation time, carrier reception time, report issuance time, and physical examination batch time as audit attributes instead of as inputs to the stable task identifier.

[0144] The inputs for stable task identifiers all correspond to the normalized semantics of the task. Changes in carrier time will not alter the task anchoring relationship or the suggested fulfillment time interval, unless there is version evidence proving a deterministic correction to the task semantic fields.

[0145] S3.7. When a record identifier appears in the source of a new review suggestion traced back to an existing review task through the historical source relationship, or when the stable task identifier generated by the new review suggestion is the same as the stable task identifier of an existing review task, the new review suggestion is associated with the existing review task without recreating the review task, and the task time anchor, window lower limit, and deadline of the existing review task are not changed by the carrier generation time, carrier reception time, or report issuance time of the new electronic physical examination data carrier.

[0146] The recommendation for repeated reviews is simply to add the source and audit relevance. This ensures that the review process remains stable even when the same report is received, formatted, or repackaged multiple times.

[0147] Example 5

[0148] Based on Example 4, the stable task identifier is generated according to the stability identifier of the inspected object, the task source observation event identifier, the identifier of the observation collection root associated with the task source observation event, the normalized review target range, the normalized review suggestion type, the task time anchor type, the task time anchor, and the performance offset interval. The electronic medical examination data carrier identifier, carrier generation time, carrier reception time, report issuance time, and medical examination batch time are saved as audit attributes but not as inputs to the stable task identifier, including:

[0149] S3.6.1. Perform unique canonical encoding on each input field of the stable task identifier according to the preset field order and concatenate them according to the preset field order to obtain a stable task identifier field tuple;

[0150] A unique, standardized encoding can specify character encoding, null value representation, enumeration encoding, time interval encoding, and field length encoding, avoiding ambiguity caused by concatenating different field combinations. Concatenation refers to sequentially joining the encoded byte sequences of each input field according to a preset field order to form a stable task identifier field tuple. Each input field applies corresponding encoding rules based on its data type: text fields are converted using a specified character encoding, with variable-length text fields having a pre-defined field length encoding; enumeration fields use a fixed-length encoding from a preset enumeration table; time interval fields are encoded with their lower and upper bounds sequentially, including time zone information; missing fields are filled with a pre-defined null value encoding instead of an empty string. The same semantic input yields a byte-by-byte identical stable task identifier field tuple on any electronic device.

[0151] S3.6.2. The stable task identifier is implemented based on the stable task identifier field tuple using any one of a deterministic digest, a composite primary key, or a unique index;

[0152] In other implementations, the complete stable task identifier field tuple can be directly used as a composite primary key or a unique index without using a digest function. When using a deterministic digest, a digest value is calculated for the entire stable task identifier field tuple, and this digest value is used as the stable task identifier. This is suitable for storage methods that store and compare tasks using a single identifier string. When using a composite primary key, all fields of the stable task identifier field tuple are used as the composite primary key of the task storage table in a preset field order; the field tuple itself is the stable task identifier. When using a unique index, a unique index is created for the stable task identifier field tuple, and the index key value is used as the stable task identifier. These three implementations form mutually verifiable task identity checks for identical field tuples; field tuples generated by repeated deliveries under any implementation will all match the same existing review task.

[0153] S3.6.3. When the deterministic digest is used and the digests of the two review tasks are the same, the stable task identifier field tuple is further compared to eliminate digest collisions;

[0154] When the summaries are the same but the complete field tuples are different, the two review tasks are saved as different tasks, and erroneous merging is avoided by appending conflict sequence numbers or indexing the complete field tuples.

[0155] The deterministic digest is implemented by a cryptographic hash function with an output length of not less than 256 bits. This hash function is a deterministic function that does not contain random salt, timestamps, or counters. The same input byte sequence is always mapped to the same digest value. The probability that different input byte sequences are mapped to the same digest value is constrained to a negligible level by the collision resistance of the hash function.

[0156] Furthermore, the appended conflict number refers to appending an incrementing sequence number field, starting from a preset initial value, to the end of field tuples with the same summary but different field tuples, and regenerating a stable task identifier, so that the identifiers of different tasks are different from each other; the complete field tuple index refers to using the complete field tuple, rather than the summary, as the final basis for comparing task identity. Both methods ensure that different review tasks are not incorrectly merged into the same task.

[0157] S3.6.4. The identifier of the observation acquisition root associated with the task source observation event is generated by the original acquisition identifier in the acquisition identity equivalence class or the verified acquisition feature tuple through normalized sorting and deterministic encoding. The verified acquisition feature tuple is a feature tuple obtained by verifying the acquisition identity based on the acquisition features. The acquisition features are formed based on at least three of the following: the stability identifier of the examined object, acquisition time, physical examination items, examined parts, side, examination modality, sample type, equipment identifier, and execution source.

[0158] When the input order of members of an acquisition identity equivalence class changes, the same observation acquisition root identifier is obtained after normalized sorting. Normalized sorting refers to arranging each original acquisition identifier or each acquisition feature tuple within the acquisition identity equivalence class according to a preset sorting rule before performing deterministic encoding, so that the sorting result does not depend on the input order of the members. When an acquisition identity equivalence class is confirmed by a strong acquisition identifier or alias relation, the original acquisition identifier in the equivalence class is used to generate the observation acquisition root identifier; when an acquisition identity equivalence class is confirmed by acquisition feature verification, the verified acquisition feature tuple is used to generate the observation acquisition root identifier.

[0159] S3.6.5. The task source observation event identifier is generated by deterministic encoding of the stability identifier of the examined object, the identifier of the observation acquisition root associated with the task source observation event, the examination items, the examined part, the side, the examination modality, the sample type, and the effective observation time interval, and remains immutable;

[0160] When the same actual data collection action generates multiple physical examination items, each standardized observation event can share the observation collection root identifier, but will have different task source observation event identifiers due to differences in the physical examination items or other event attributes. The deterministic encoding of the task source observation event identifier adopts the same unique standardized encoding method as the stable task identifier field tuple, with each input field encoded in a preset field order and then concatenated. Maintaining immutability means that once the task source observation event identifier is generated, it will not be modified, repeated calculations will yield the same value, and it will not change due to repeated delivery, format conversion, or changes in the reception time of the electronic physical examination data carrier.

[0161] S3.6.6. Repeated delivery of the electronic physical examination data carrier does not regenerate the task source observation event identifier and the stable task identifier.

[0162] Duplicate deliveries can add new occurrence records, carrier inclusion relationships, and historical source relationships, but they do not add new acquisition identity equivalence classes, nor do they change existing identifiers formed by these acquisition identity equivalence classes. "Not adding new acquisition identity equivalence classes" means that the original acquisition identifier carried by the duplicate delivery, after conflict verification, is either incorporated into an existing acquisition identity equivalence class or has an alias relationship established with it, rather than forming a new observation acquisition root. "Existing identifiers remain unchanged" means that the task source observation event identifier, the observation acquisition root identifier, and the stable task identifier all retain their values ​​from when they were first generated. Duplicate deliveries only expand the source and audit information, and do not generate duplicate review tasks or new time boundaries.

[0163] like Figure 5 As shown, in the activity flow provided in this embodiment, the connection relationship between the three stages of record parsing, identity collection and merging, and anchoring of review tasks, and the judgment branches therein are as follows: When the observation collection root cannot be uniquely determined, it is re-verified after splitting according to the record boundary. If it still cannot be determined, an observation collection root conflict record or a source pending confirmation record is formed; when the task time anchor point, performance offset interval, and the obtained suggested performance time interval are incomplete, invalid, or empty, the record time is marked as pending confirmation, and normal time classification and status comparison based on the deadline time are prohibited; when the source records are the same or the stable task identifier is the same, the existing review task is associated without resetting the existing time boundary.

[0164] Example 6

[0165] like Figure 6 As shown, based on Example 1, aliasing is performed on the original acquisition identifiers of the recorded occurrences, and observation acquisition roots and standard observation events are established according to acquisition identity equivalence classes, including:

[0166] S2.1. Perform normalization processing on the physical examination items, observation values, units of measurement, examined parts and effective observation time intervals in the recorded data. Expand the precise time into a left-closed and right-open non-empty interval with the lower bound strictly earlier than the upper bound according to the preset minimum time granularity. Expand the time with only calendar date into a left-closed and right-open interval from 0:00 on the current day to 0:00 on the next day according to the preset time zone.

[0167] Assume only calendar dates are available. Then the corresponding effective observation time interval can be expressed as:

[0168] Equation (5)

[0169] in, This indicates the valid observation time interval corresponding to the date precision. Indicates a valid calendar date. For precise times... It can be based on the minimum time granularity fixed in the rule version. Expand as ;in, This represents the minimum time granularity.

[0170] S2.2. When a valid, non-empty observation time interval cannot be formed, a time-determined record is generated, and the observation time interval is prohibited from participating in subsequent judgments and time classifications;

[0171] Empty intervals, intervals whose lower bound is later than the upper bound, or intervals whose time zone cannot be determined are not used to infer whether candidate observation events are successors to mission source observation events.

[0172] S2.3. Perform conflict verification on the original acquisition identifiers, and classify the original acquisition identifiers that can jointly identify the same actual acquisition action into the same acquisition identity equivalence class, and establish an observation acquisition root and a deterministic observation acquisition root identifier for each acquisition identity equivalence class;

[0173] Conflict checking includes compatibility checks on the subject, collection time, examination items, sample type, examination modality, device identifier, and execution source. Identifiers with a basis for determining gender are merged; merging is not performed when mutually exclusive fields exist.

[0174] S2.4. Establish standardized observation events with deterministic identifiers based on the observation acquisition roots and the standardized physical examination items. Each standardized observation event is associated with only one observation acquisition root, and the same observation acquisition root is allowed to be associated with more than one standardized observation event generated by the same actual acquisition action.

[0175] For example, when the same actual data collection action produces multiple standardized physical examination results, multiple standardized observation events can be established, but each standardized observation event is associated with the same observation collection root.

[0176] S2.5. When the original acquisition identifier of the observation event to be classified into the same standard observation event points to two or more non-equivalent observation acquisition roots, the event that can be split into multiple standard observation events based on the content of the record boundary splitting will be generated. If it cannot be split, an observation acquisition root conflict record will be generated. If the observation acquisition root cannot be uniquely determined, a source unconfirmed record will be generated.

[0177] Splitting can be performed based on field paths, table rows, resource identifiers, or character ranges. When it is impossible to determine which observation root each result belongs to, conflicts are retained instead of arbitrary assignment.

[0178] S2.6. Standard observation events for which the root conflict records of the observation acquisition are not resolved or the root of the observation acquisition is not confirmed do not form valid performance records.

[0179] Observational acquisition root conflicts can be resolved by using new enhanced acquisition identifiers, verified alias mappings, or further splitting of occurrence records. After resolution, the electronic equipment recalculates the affected candidate choices, ternary evidence gating, and mission status.

[0180] Example 7

[0181] like Figure 7 As shown, based on Example 1, candidate observation events that pass the three-valued evidence gating are classified according to the lower bound of the window and the cutoff time. A unique task status is determined according to a preset priority based on the current calculation time, and longitudinal verification clues are output, including:

[0182] S5.1. Time classification is performed only on candidate observation events that pass the three-valued evidence gating and whose effective observation time interval and suggested fulfillment time interval are both valid non-empty intervals, wherein the suggested fulfillment time interval is defined by the lower bound of the window and the deadline.

[0183] The suggested fulfillment time interval is a left-closed, right-open interval. The lower bound of the window constitutes the inclusive lower bound of this interval, and the deadline constitutes the exclusive upper bound of this interval. That is, moments belonging to the lower bound of the window belong to the suggested fulfillment time interval, while moments belonging to the deadline do not belong to the suggested fulfillment time interval. The lower bound of the window is obtained by adding the lower bound of the task time anchor point and the lower bound of the fulfillment offset interval according to the calendar determined by the rule version. The deadline is obtained by adding the upper bound of the task time anchor point and the upper bound of the fulfillment offset interval according to the same calendar. When the task time anchor point or the fulfillment offset interval is a time pending confirmation marker, the lower bound of the window and the deadline are both time pending confirmation markers, and the corresponding candidate observation events do not enter the time classification. This definition method is compatible with the boundary determination of time classification: when the upper bound of the effective observation time interval is exactly equal to the deadline, the latest observation time represented by the upper bound is still earlier than the deadline, so on-time evidence can still be generated; when the lower bound of the effective observation time interval is exactly equal to the deadline, the actual observation only occurred from the deadline, so overdue evidence is generated; when the effective observation time interval strictly crosses the lower bound of the window or the deadline, the actual observation time cannot be uniquely classified into the interval or is overdue, so an uncertain time record is generated instead of a valid performance record.

[0184] Candidate observations that fail or are uncertain under the three-valued evidence gating system are not included in the normal time classification, in order to avoid the formation of a valid performance record simply because the time falls within the recommended performance time range.

[0185] S5.2. When the lower bound of the effective observation time interval of the candidate observation event is not earlier than the lower bound of the window and the upper bound is not later than the deadline, on-time evidence and valid performance record are generated; when the lower bound of the effective observation time interval is not earlier than the deadline, overdue evidence and valid performance record are generated.

[0186] Let the effective observation time interval of the candidate observation event be . The recommended performance period is as follows: ;in, This indicates the valid observation time interval for candidate observation events. and These represent the lower and upper bounds of the interval, respectively. This indicates the suggested timeframe for fulfillment. and These represent the lower bound and the cutoff time of the window for that interval, respectively. and Timely generation of evidence; when Overdue evidence is generated in a timely manner.

[0187] S5.3. When the effective observation time interval of the candidate observation event strictly crosses the lower boundary of the window or the deadline, a time-determined record is generated; when the upper boundary of the effective observation time interval is not later than the lower boundary of the window, an early candidate record is generated; and in neither of the above two cases, a valid performance record is generated.

[0188] Strict boundary crossing indicates that the candidate interval has non-empty portions on both sides of the boundary. Using complete interval classification can avoid forcibly classifying candidates with only date precision that cross the boundary into the on-time or overdue categories.

[0189] S5.4. Generate evidence blocking records containing cause codes for candidate observation events where the three-valued evidence gating fails; generate verification records for candidate observation events where the three-valued evidence gating is uncertain, and mark the verification records generated due to the uncertainty of the observation acquisition root or acquisition identity as source uncertainty type, mark the verification records generated due to the uncertainty of the timing as time uncertainty type, and mark the verification records generated due to the uncertainty of the actual acquisition assertion path, the source closure, or the scope of the review target as information incomplete candidate type;

[0190] Reason codes can include root conflict, source unconfirmed, task source pure backreference, missing actual collection assertion, uncertain time order, before window, crossing deadline, and range mismatch, etc.

[0191] S5.5. The observation collection roots associated with candidate observation events that can change the current earliest fulfillment time classification or task status are identified as decisive candidate observation collection roots, and a unique task status is determined for each review task according to a preset priority that decreases sequentially: if there is an unresolved conflict record involving the task source observation event or an unresolved conflict record that can change the current earliest fulfillment time classification, it is determined to be in a conflict pending verification state; if there is no unresolved conflict record of the observation collection root, but the observation collection root associated with the task source observation event or the decisive candidate observation collection root is pending confirmation, or if there is a pending verification record of the source uncertainty type, it is determined to be in a source pending confirmation state; if there is no such situation but there is evidence of on-time completion, it is determined to be in a timely completion state; if there is no evidence of on-time completion but there is a time uncertainty record or a pending verification record of the time uncertainty type, it is determined to be in a time uncertainty state; if there is no such situation but there is evidence of overdue completion, it is determined to be in a overdue completion state.

[0192] Prioritization ensures that conflicts or uncertainties in their origin are not obscured by subsequent on-time evidence. When both on-time and overdue evidence exist simultaneously, and no higher-priority anomalies are present, on-time completion status takes precedence over overdue completion status. The preset priority order is used to stably output the current task status when multiple facts exist simultaneously in the same review task, and is not the order in which it is displayed on the screen.

[0193] S5.6. If there are no valid performance records, source-unconfirmed records, or time-uncertain records, and the deadline is not a time-unconfirmed marker and the current calculation time is not earlier than the deadline, the status is determined to be overdue and incomplete. If the deadline is not a time-unconfirmed marker, the current calculation time is earlier than the deadline, and there are early candidate records or incomplete candidate records to be verified, the status is determined to be candidate performance. If none of the foregoing conditions are met, the status is determined to be open.

[0194] The current computation time is used only as a state assessment parameter. Under the same data and rule version, using different current computation times for assessment can produce new state assessment results, but will not change the task's own stability identifier and time boundaries.

[0195] S5.7. The current calculation time and the time zone identifier used are included in the input summary. The current calculation time is not used to generate the observation acquisition root identifier, the standard observation event identifier, or the stable task identifier, nor is the current calculation time used to change the task time anchor point, the lower bound of the window, or the deadline.

[0196] The current computation time is passed in by the caller using a time zone identifier and included in the input summary. The system clock is not reread during the state assessment process. The input summary may also include the identifiers of occurrence records involved in the assessment, the identifiers of standardized observation events, the root state of observation acquisition, the version of the task generation rule, and the version of the evidence gating rule.

[0197] The input digest is constructed as follows: all input fields participating in this state assessment are uniquely encoded according to a preset field order and concatenated into a byte sequence according to the preset field order. Then, a deterministic digest is performed on the byte sequence to obtain a fixed-length input digest value. The same set of input fields will produce the same input digest under the same rule version. The input digests will be different when the set of input fields is different or the rule version is different.

[0198] S5.8. Write the task state before state transition, the task state after state transition, the current calculation time, the time zone identifier, the input summary, the stable task identifier, the identifier of the candidate observation event, the cause code, the electronic physical examination data carrier identifier corresponding to the evidence and the source location within the carrier, the normalization rule version, the task generation rule version and the evidence gating rule version into the vertical review clue.

[0199] Vertical review leads are preserved in an append-only manner. When evidence is withdrawn or the rule version changes, the original leads are not overwritten, but a new status assessment record is generated, and the migration relationship between the old and new results is retained.

[0200] like Figure 8 As shown, in the activity flow provided in this embodiment, the execution process and judgment branches of each link of source closure construction, three-valued evidence gating and state control are as follows: when the source closure is incomplete, empty or contains unclassifiable mixed terminals, the corresponding gating component is recorded as uncertain and forms a verification record; when there is a false component in the gating process, an evidence blocking record is formed; when there is no false component and at least one component cannot be determined, a verification record is formed; when all gating components are true, time classification is entered; time classification generates on-time evidence, overdue evidence, time uncertain record or advance candidate record according to the relationship between the observed effective time interval and the suggested performance time interval; state summary determines the conflict pending verification state, source pending confirmation state, on-time completion state, time uncertain state, overdue completion state, overdue non-completion state, candidate performance state and open state in order of preset priority, and outputs state change and vertical review clues.

[0201] Example 8

[0202] Based on Example 1, the electronic medical examination data carrier of the same examinee is analyzed to obtain occurrence records carrying delivery domain markers and evidentiary properties, including:

[0203] S1.1. Determine the stable identifier of the examined object that remains unchanged within the same processing range, and read the existing physical examination data when there is corresponding existing physical examination data in the storage device, so that the electronic physical examination data carrier itself or together with the existing physical examination data carries physical examination data for at least two observation periods;

[0204] The stable identifier for the inspected object can be an anonymous token or an access-controlled internal index. This identifier remains stable within the same processing scope and is stored separately from direct identity information.

[0205] S1.2. Perform recursive field parsing and source location on the electronic physical examination data carrier, and merge the parsed occurrence records with the occurrence records in the existing physical examination data to obtain multiple occurrence records;

[0206] Recursive field parsing is used to handle nested objects, history lists, linked resources, and recurring wrapper structures. Source location stores the field path, object key, table coordinates, or character offset.

[0207] S1.3. Each occurrence record shall be associated with at least the stability identifier of the examined object, the electronic medical examination data carrier identifier, the source location within the carrier, the original content, the time information, the original acquisition identifier, the historical reference path, and the reference depth;

[0208] The occurrence of a record can be identified by the electronic medical examination data carrier identifier, the source location within the carrier, and the original content summary. This identifier is used to identify a single occurrence of content, rather than directly representing an actual data collection action.

[0209] S1.4. The delivery domain marker is used to indicate the current delivery domain, historical delivery domain, or unknown delivery domain. The evidence property is used to indicate the actual collection assertion, reference, copying, comparison, or derivation. The evidence property is a multi-valued attribute that allows more than one of them to coexist.

[0210] For example, a record can simultaneously possess both an actual acquisition assertion and a comparative property, indicating that the record was formed through independent acquisition and contains semantics comparing it with previous observations. The referencing property indicates that the appearing record points back to an existing appearing record via a pointer, object key, or Uniform Resource Identifier; the copying property indicates that the appearing record, in whole or in part, reproduces the content of an existing appearing record and carries a reference path pointing to the source; the comparative property indicates that there is comparative semantics between the appearing record and previous observations but does not constitute a source dependency on previous observations; the derivation property indicates that the content of the appearing record is obtained by calculation or transformation from more than one existing observation, such as the difference or rate of change calculated from the values ​​of two observations. The referencing, copying, and derivation properties constitute the basis for continuing traversal along the direction of the result source, while the comparative property does not constitute this basis.

[0211] S1.5. The historical reference path and the reference depth are determined by at least one of the following: explicit historical field, historical object key, pointer or uniform resource identifier, past semantics, deterministic content fingerprint and direct or transmitted historical source relationship;

[0212] Multiple determination methods can be used in combination. When there is only content similarity but a lack of stable pointers, content fingerprints, or source relationships, it can be retained as a candidate relationship to be verified.

[0213] S1.6. The delivery domain marker is neither a sufficient condition for the three-valued evidence gate to pass nor a sufficient condition for the three-valued evidence gate to fail.

[0214] Therefore, independent successor observations located in the historical delivery domain can participate in the three-valued evidence gating, while the occurrence records of events located in the current delivery domain but only referencing the mission source observation events are blocked.

[0215] Example 9

[0216] Based on Example 1, the method further includes:

[0217] S6. Based on the nature of the evidence, the actual collection assertion relationship, the historical source relationship, the observation collection root and the physical examination batch corresponding to the electronic physical examination data carrier, perform single physical examination projection exclusion to distinguish the actual observation set of the physical examination batch from the historical projection set embedded in the electronic physical examination data carrier;

[0218] The single physical examination projection exclusion is only used to determine whether a record should be included in the new actual observation set of the current physical examination batch, and does not predetermine whether its corresponding standard observation event can be used to perform other re-examination tasks.

[0219] Projection exclusion is performed according to the following criteria: If the evidentiary nature of an occurrence record only includes citation, copy, or derivation, and the examination batch associated with the occurrence record traced back along the source direction is earlier than the current examination batch, the occurrence record is included in the historical projection set; if an occurrence record contains an actual collection assertion and its observation collection root has been verified to belong to the current examination batch, the occurrence record is included in the actual observation set; occurrence records whose attribution cannot be determined are not included in any set, and a source-pending-confirmation record is generated for subsequent verification. Here, typified relationships refer to relationships distinguished by type identifiers, including actual collection assertion relationships, collection attribution relationships, and historical source relationships of type citation, copy, comparison, or observation derivation.

[0220] S7. When generating the actual observation set, the historical projection set is not counted as a newly occurring actual collection action of the physical examination batch. At the same time, the occurrence record, standardized observation event, observation collection root and typified relationship corresponding to the historical projection set are retained. The typified relationship is a relationship distinguished by type identifier, including actual collection assertion relationship, collection attribution relationship and historical source relationship with type reference, copy, comparison or observation derivation.

[0221] Preserving the data and relationships corresponding to historical projections allows independent observations in the historical delivery domain to continue participating in the candidate selection for other tasks according to their original observation validity period.

[0222] S8. When the same actual collection action is delivered repeatedly, only the occurrence record or the historical source relationship is added, and the collection identity equivalence class is not added;

[0223] The determination of duplicate delivery is based on the equivalence class of the data collection identity, rather than solely on the carrier identifier or text content. Carrier format conversion and field rearrangement do not increase the actual number of data collection actions.

[0224] S9. When an occurrence record located in the historical delivery domain has an original valid observation time interval and an independent observation acquisition root with a verified or corroborated status, the corresponding standard observation event shall not be excluded because the occurrence record is located in the historical delivery domain, and the three-valued evidence gating shall be allowed when all gate control components are true.

[0225] For example, a subsequent delivery carrier packages independently occurring follow-up observations within a history field, and the electronic device still performs gating based on its original acquisition identity and the effective observation time, without using the location of the history field as a veto condition.

[0226] S10. When the occurrence record located in the current delivery domain is only referenced back to the task source observation event through the historical source relationship and there is no independent actual collection assertion path, it is prohibited to establish a new observation collection root, standard observation event or valid performance record based on this, and it is prohibited to create a review task as a new review suggestion source based on the occurrence record, and it is prohibited to reset the task time anchor point, the lower bound of the window and the deadline of the existing review task.

[0227] This process blocks the source content of the task from reappearing after being copied, referenced, or derived, and prevents the time attribute of the subsequent delivery carrier from changing the existing review task.

[0228] Example 10

[0229] like Figure 9As shown, this embodiment provides a physical examination re-examination status control system based on task-relative observation lineage. The task-relative observation lineage refers to the relationship between candidate observation events and re-examination tasks in terms of task source observation events, data acquisition identity, and historical source paths, including:

[0230] The occurrence record parsing module is used to parse the electronic medical examination data carrier of the same examinee to obtain occurrence records carrying delivery domain markers and evidence nature;

[0231] The record parsing module can perform tasks such as determining the stability identifier of the examined object, reading existing physical examination data, parsing recursive fields, locating the source within the carrier, extracting historical reference paths, and generating delivery domain markers.

[0232] The data acquisition identity merging module is used to perform alias merging on the original data acquisition identifiers of the occurrence records, and to establish observation acquisition roots and standard observation events according to the data acquisition identity equivalence classes;

[0233] The data collection identity merging module can perform tasks such as atomic generation of data collection identifiers, verification of strong data collection identifiers, establishment of alias relationships, verification of data collection features, generation of conflict records, and generation of standardized observation event identifiers.

[0234] The task anchoring module is used to anchor the review suggestions in the occurrence record to the task source observation event, and generate a stable task identifier, lower window bound and deadline of the review task according to the task time anchor point and the performance offset interval.

[0235] The task anchoring module can perform tasks such as review suggestion splitting, task source observation event determination, task time anchor point generation, time pending confirmation processing, stable task identifier field tuple encoding, and repetitive task association.

[0236] The evidence gating module is used to refer to the task source observation event, traverse the historical source relationships of candidate observation events to form a source closure, and perform three-valued evidence gating on the independent collection identity and non-pure backreference. The candidate observation event is a standard observation event selected from the standard observation events that initially matches or has an uncertain matching result with the review target range of the review task. The non-pure backreference means that the result source of the candidate observation event does not only refer back to the task source observation event.

[0237] The evidence gating module collects candidate support occurrence records, recursively traverses historical source relationships, identifies non-derived actual collection sources, complete source closures, and task source pure backreferences, and outputs gating actions of pass, failure, or uncertainty based on each gating component.

[0238] The status control module is used to classify candidate observation events that have passed the three-valued evidence gating according to the lower bound of the window and the deadline, determine the unique task status according to the preset priority based on the current calculation time, and output the longitudinal verification clues.

[0239] The status control module can output statuses such as on-time completion, overdue completion, overdue non-completion, conflict pending verification, source pending confirmation, uncertain time, candidate performance, or open status, and also save status transition records.

[0240] The modules described above can be implemented by independent program components or by multiple program units within the same processor. In other implementations, multiple modules can be deployed together, or a single module can be divided into multiple sub-modules, as long as they can perform the corresponding data processing functions.

[0241] The present invention can also be implemented by a computer-readable storage medium. The computer-readable storage medium stores a computer program, which, when executed by a processor, causes an electronic device to perform the method described in any of embodiments 1 to 9.

Claims

1. A method for controlling the status of physical examination follow-up based on task-relative observation of bloodline, characterized in that, The task-relative observation lineage refers to the relationship formed between candidate observation events and review tasks in terms of task source observation events, data collection identity, and historical source paths. The method is executed by electronic devices and includes: By analyzing the electronic medical examination data carrier of the same examinee, the occurrence records of the delivery domain markers and evidentiary nature can be obtained; Alias ​​merging is performed on the original acquisition identifiers of the records that appear, and observation acquisition roots and standard observation events are established according to acquisition identity equivalence classes; The review recommendations in the recorded occurrences are anchored to the task source observation events, and a stable task identifier, lower window bound, and deadline for the review task are generated according to the task time anchor point and performance offset interval. Referring to the task source observation event, the historical source relationship of the candidate observation event is traversed to form a source closure. Three-valued evidence gating is performed on the independent collection identity and non-pure backreference. The candidate observation event is a standard observation event selected from the standard observation events that initially matches or has an uncertain matching result with the review target range of the review task. The non-pure backreference means that the result source of the candidate observation event does not only refer back to the task source observation event. Candidate observation events that pass the three-valued evidence gating are classified according to the lower bound of the window and the deadline. The unique task status is determined according to the preset priority based on the current calculation time, and the longitudinal verification clues are output.

2. The method according to claim 1, characterized in that, Referring to the aforementioned task source observation events, the historical source relationships of candidate observation events are traversed to form a source closure. Three-valued evidence gating is performed on the independent acquisition identities and non-pure backreferences, including: The observation collection root associated with the task source observation event through the collection attribution relationship is determined as the task source observation collection root, wherein the collection attribution relationship is a relationship that indicates that the standard observation event belongs to the observation collection root; From the standard observation events, standard observation events that preliminarily match or whose matching results are uncertain with the scope of the review target of the review task are selected as candidate observation events, and the selection is not restricted by the delivery domain marker of the corresponding occurrence record; Collect all occurrence records that support the candidate observation event results through actual collection assertion relationships or the historical source relationships, remove duplicates by occurrence record identifiers and sort them in a standardized manner to form a candidate support occurrence record set; Starting from each occurrence record in the candidate support occurrence record set, recursively traverse the historical source relations of type reference, copy or observation derivation along the result source direction, while not traversing the historical source relations of type comparison, and gather the occurrence records and relations obtained from the traversal to form the source closure; The terminal occurrence record of the observation collection root that does not have a historical source relationship that can be further traversed, is supported by the actual collection assertion relationship, and whose status is verified or corroborated by the collection attribution relationship is determined as a non-derived actual collection source. The source closure is determined to be a complete source closure only when the terminal is reached in every traversal path and there is no endless circular dependency. Only when the source closure is a non-empty complete source closure, there are no unclassifiable mixed terminals, and each terminal occurrence record constitutes the non-derived actual acquisition source, and at least one observation acquisition root of the non-derived actual acquisition source is independent of the task source observation acquisition root, is the non-pure backreference component determined to be true; when all the non-derived actual acquisition sources are associated with the task source observation acquisition root and there are no independent non-derived actual acquisition sources, the non-pure backreference component is determined to be false, and the corresponding candidate observation event is determined to be a pure backreference of the task source. The mixed terminal is a terminal occurrence record reached during traversal that neither constitutes the non-derived actual acquisition source nor can be determined to be merely a derived intermediate node. Except in cases where the result is determined to be true or false, the non-pure backreference component is determined to be indeterminate; When the candidate observation event and the task source observation event are each associated with exactly one observation acquisition root in a verified or corroborated state, and the observation acquisition root associated with the candidate observation event and the task source observation acquisition root belong to different acquisition identity equivalence classes, the acquisition identity independent component is determined to be true; when they belong to the same acquisition identity equivalence class, it is determined to be false; in other cases, it is determined to be uncertain. The gate control process is determined to pass when all gate control components are true; the gate control process is determined to fail when any gate control component is false; and the gate control process is determined to be uncertain when there are no false gate control components and at least one gate control component cannot be determined.

3. The method according to claim 2, characterized in that, Terminal occurrence records of observation collection roots that do not have a traversable historical source relationship, are supported by actual collection assertion relationships, and whose collection attribution relationship association status is verified or corroborated, are identified as non-derived actual collection sources, including: A strong acquisition identifier is obtained from the combination of the inspection execution identifier, sample identifier and completed inspection application identifier, image inspection identifier and inspection equipment output identifier, and an acquisition identifier atom is formed that includes at least the identifier type, the issuing system identifier, the identifier value and the validity period. The strong acquisition identifier is an acquisition identifier generated by the execution system after the actual acquisition is completed and can uniquely point to the actual acquisition action. When the same collection identity equivalence class contains at least one strong collection identifier that has been verified and valid by the issuing system, and there are no mutually exclusive strong collection identifiers, the state of the corresponding observation collection root is set to verified. Only when there is a cross-system deterministic mapping, identifier replacement record, device migration record, or a verified mapping of the same inspection instance, an alias relationship satisfying reflexivity, symmetry, and transitivity is established between two of the acquisition identifier atoms, and acquisition identifier atoms that are aliases to each other are classified into the same acquisition identity equivalence class; When the strong acquisition identifier is missing, the acquisition feature is formed based on at least three of the following: the stability identifier of the subject under examination, the acquisition time, the physical examination item, the subject under examination, the side, the examination modality, the sample type, the equipment identifier, and the execution source. When the acquisition feature meets the preset consistency condition and there are no mutually exclusive fields, the status of the corresponding observation acquisition root is set to verified; otherwise, it is set to source pending confirmation. When two original acquisition identifiers are different and there is no alias relationship, the corresponding observation acquisition roots are not considered to be independent of each other. Instead, based on the non-conflicting examination object, acquisition time, examination items, examination mode, equipment identifier and execution source information, the observation acquisition roots corresponding to the two original acquisition identifiers are determined to belong to different acquisition identity equivalence classes or to be determined to have sources pending confirmation.

4. The method according to claim 1, characterized in that, The review recommendations in the recorded occurrences are anchored to the task source observation event, and a stable task identifier, lower window bound, and deadline for the review task are generated according to the task time anchor point and performance offset interval, including: The review proposal containing multiple review targets is split into atomic review tasks, each having a task source observation event supported by an actual acquisition assertion relationship and a clearly defined scope of the review target. Based on the corresponding occurrence record identifier, the source sub-position in the carrier corresponding to each review target, and the atomic number, a different review proposal source occurrence record identifier is generated for each atomic review task. Based on the task time anchor point type, the task time anchor point is determined from the observation effective time interval of the observation event of the task source, the effective time interval of the review suggestion, or the later value of each boundary of the observation effective time interval and the effective time interval of the review suggestion. The task time anchor point is a left-closed and right-open non-empty interval whose lower bound is strictly earlier than the upper bound. If any of the task time anchor type, the task time anchor, or the performance offset interval cannot be determined, or if a valid non-empty task time anchor cannot be formed, the corresponding field is set to a preset time confirmation mark that remains unchanged across data structure versions and a time indeterminate record is generated. Only when the task time anchor type, the task time anchor, and the performance offset interval are not the time to be confirmed marker, a suggested performance time interval with left-closed and right-open is determined based on the task time anchor and the performance offset interval. The lower bound of the suggested performance time interval is determined as the lower bound of the window, and the upper bound of the suggested performance time interval is determined as the deadline. If a valid, non-empty suggested performance time interval cannot be formed, the lower bound of the window and the deadline are set as the time pending confirmation marker, a time uncertain record is generated, and time classification and status comparisons dependent on the deadline are prohibited. The stable task identifier is generated based on the stability identifier of the inspected object, the task source observation event identifier, the identifier of the observation and acquisition root associated with the task source observation event, the standardized review target range, the standardized review suggestion type, the task time anchor type, the task time anchor, and the performance offset interval. The electronic physical examination data carrier identifier, carrier generation time, carrier reception time, report issuance time, and physical examination batch time are saved as audit attributes and not as inputs to the stable task identifier. When a record identifier appears in the source of a newly added review suggestion traced back to an existing review task through the historical source relationship, or when the stable task identifier generated by the newly added review suggestion is the same as the stable task identifier of an existing review task, the newly added review suggestion is associated with the existing review task without recreating the review task, and the task time anchor, window lower limit, and deadline of the existing review task are not changed by the carrier generation time, carrier reception time, or report issuance time of the newly added electronic physical examination data carrier.

5. The method according to claim 4, characterized in that, The stable task identifier is generated based on the stability identifier of the inspected object, the task source observation event identifier, the identifier of the observation collection root associated with the task source observation event, the normalized review target range, the normalized review suggestion type, the task time anchor type, the task time anchor, and the performance offset interval. The electronic medical examination data carrier identifier, carrier generation time, carrier reception time, report issuance time, and medical examination batch time are saved as audit attributes but not as inputs to the stable task identifier. Each input field of the stable task identifier is uniquely and canonically encoded according to a preset field order, and then concatenated according to the preset field order to obtain a stable task identifier field tuple. The stable task identifier is implemented based on the stable task identifier field tuple, using any one of a deterministic digest, a composite primary key, or a unique index. When the deterministic digest is used and the digests of the two review tasks are the same, the stable task identifier field tuple is further compared to eliminate digest collisions; The identifier of the observation collection root associated with the task source observation event is generated by the original collection identifier in the collection identity equivalence class or the verified collection feature tuple through normalized sorting and deterministic encoding. The verified collection feature tuple is a feature tuple obtained by verifying the collection identity based on the collection features. The collection features are formed based on at least three of the following: the stability identifier of the examined object, the collection time, the physical examination items, the examined part, the side, the examination modality, the sample type, the equipment identifier, and the execution source. The task source observation event identifier is generated deterministically and remains immutable by the stability identifier of the examined object, the identifier of the observation acquisition root associated with the task source observation event, the examination items, the examined part, the side, the examination modality, the sample type, and the effective observation time interval; Repeated delivery of the electronic physical examination data carrier does not regenerate the task source observation event identifier and the stable task identifier.

6. The method according to claim 1, characterized in that, Alias ​​merging is performed on the original acquisition identifiers of the recorded occurrences, and observation acquisition roots and canonical observation events are established according to acquisition identity equivalence classes, including: The physical examination items, observation values, units of measurement, examined parts and effective observation time intervals in the records are standardized. The precise time is expanded into a left-closed and right-open non-empty interval with a lower bound that is strictly earlier than the upper bound according to the preset minimum time granularity. The time with only calendar date is expanded into a left-closed and right-open interval from 0:00 on the current day to 0:00 on the next day according to the preset time zone. If a valid, non-empty observation time interval cannot be formed, a time-determined record is generated, and the observation time interval is prohibited from participating in subsequent judgments and time classifications. Perform conflict verification on the original acquisition identifiers, and classify the original acquisition identifiers that can jointly identify the same actual acquisition action into the same acquisition identity equivalence class, and establish an observation acquisition root and a deterministic observation acquisition root identifier for each acquisition identity equivalence class; Based on the observation collection root and the standardized physical examination items, a standardized observation event with a deterministic identifier is established. Each standardized observation event is associated with only one observation collection root, and the same observation collection root is allowed to be associated with more than one standardized observation event generated by the same actual collection action. When the original acquisition identifier of the observation event to be classified into the same standard observation event points to two or more mutually unequal observation acquisition roots, it can be split into multiple standard observation events based on the content of the record boundary split. If it cannot be split, an observation acquisition root conflict record is generated. If the observation acquisition root cannot be uniquely determined, a source unconfirmed record is generated. If the conflict record of the observation acquisition root is not resolved or the observation acquisition root is not confirmed, the standard observation event will not form a valid performance record.

7. The method according to claim 1, characterized in that, Candidate observation events that pass the three-valued evidence gating are classified according to the lower bound of the window and the cutoff time. A unique task status is determined based on a preset priority and the current computation time, and longitudinal verification clues are output, including: Time classification is performed only on candidate observation events that pass the three-valued evidence gating and whose effective observation time interval and suggested fulfillment time interval are both valid non-empty intervals, wherein the suggested fulfillment time interval is defined by the lower bound of the window and the deadline. When the lower bound of the effective observation time interval of the candidate observation event is not earlier than the lower bound of the window and the upper bound is not later than the deadline, on-time evidence and valid performance record are generated; when the lower bound of the effective observation time interval is not earlier than the deadline, overdue evidence and valid performance record are generated. When the effective observation time interval of the candidate observation event strictly crosses the lower boundary of the window or the deadline, a time-determined record is generated; when the upper boundary of the effective observation time interval is not later than the lower boundary of the window, an early candidate record is generated; and in neither of the above two cases, a valid performance record is generated. For candidate observation events where the three-valued evidence gating fails, generate evidence blocking records containing cause codes; for candidate observation events where the three-valued evidence gating is uncertain, generate verification records, and mark verification records generated due to the inability to determine the observation acquisition root or acquisition identity as source uncertainty type, mark verification records generated due to the inability to determine the timing as time uncertainty type, and mark verification records generated due to the inability to determine the actual acquisition assertion path, the source closure, or the scope of the review target as information incomplete candidate type; The observation collection roots associated with candidate observation events that can change the current earliest fulfillment time classification or task status are identified as decisive candidate observation collection roots. A unique task status is then determined for each review task according to a predetermined, decreasing priority: If there are unresolved conflict records involving the task source observation event or those capable of changing the current earliest fulfillment time classification, the task is identified as a conflict pending verification status; if there are no unresolved conflict records but the observation collection roots associated with the task source observation event or the decisive candidate observation collection roots are pending confirmation, or if there are pending verification records of uncertain source type, the task is identified as a source pending confirmation status; if none of the aforementioned situations exist but there is evidence of on-time completion, the task is identified as an on-time completion status; if there is no evidence of on-time completion but there are records of uncertain time or pending verification records of uncertain time type, the task is identified as a time uncertain status; if none of the aforementioned situations exist but there is evidence of overdue completion, the task is identified as an overdue completion status. If there are no valid performance records, source-unconfirmed records, or time-uncertain records, and the deadline is not a time-unconfirmed marker and the current calculation time is not earlier than the deadline, the status is determined to be overdue and incomplete. If the deadline is not a time-unconfirmed marker, the current calculation time is earlier than the deadline, and there are early candidate records or incomplete candidate records that need verification, the status is determined to be candidate performance. If none of the foregoing conditions are met, the status is determined to be open. The current calculation time and the time zone identifier used are included in the input summary. The current calculation time is not used to generate the observation acquisition root identifier, the standard observation event identifier, or the stable task identifier, nor is the current calculation time used to change the task time anchor point, the lower bound of the window, or the deadline. The task state before state transition, the task state after state transition, the current calculation time, the time zone identifier, the input summary, the stable task identifier, the identifier of the candidate observation event, the cause code, the electronic medical examination data carrier identifier corresponding to the evidence and the source location within the carrier, the normalization rule version, the task generation rule version and the evidence gating rule version are written into the longitudinal review clue.

8. The method according to claim 1, characterized in that, Analyzing the electronic medical examination data carrier of the same examinee yields records of occurrences carrying delivery domain markers and evidentiary characteristics, including: Determine the stable identifier of the examined object that remains unchanged within the same processing range, and read the existing physical examination data when there is corresponding existing physical examination data in the storage device, so that the electronic physical examination data carrier itself or together with the existing physical examination data carries physical examination data for at least two observation periods; Recursive field parsing and source location are performed on the electronic physical examination data carrier, and the parsed occurrence records are merged with the occurrence records in the existing physical examination data to obtain multiple occurrence records; Each occurrence record is associated with at least the subject's stability identifier, electronic medical examination data carrier identifier, source location within the carrier, original content, time information, the original acquisition identifier, historical reference path, and reference depth; The delivery domain marker is used to indicate the current delivery domain, historical delivery domain, or unknown delivery domain. The evidence property is used to indicate the actual collection assertion, reference, copying, comparison, or derivation. The evidence property is a multi-valued attribute that allows more than one of them to coexist. The historical reference path and the reference depth are determined by at least one of the following: explicit historical fields, historical object keys, pointers or Uniform Resource Identifiers, past semantics, deterministic content fingerprints, and direct or transmitted historical source relationships; The delivery field marker is neither a sufficient condition for the three-valued evidence gate to pass nor a sufficient condition for the three-valued evidence gate to fail.

9. The method according to claim 1, characterized in that, The method further includes: Based on the nature of the evidence, the actual collection assertion relationship, the historical source relationship, the observation collection root, and the physical examination batch corresponding to the electronic physical examination data carrier, single physical examination projection exclusion is performed to distinguish the actual observation set of the physical examination batch from the historical projection set embedded in the electronic physical examination data carrier; When generating the actual observation set, the historical projection set is not counted as a new actual collection action of the physical examination batch. At the same time, the occurrence record, standardized observation event, observation collection root and typified relationship corresponding to the historical projection set are retained. The typified relationship is a relationship distinguished by type identifier, including actual collection assertion relationship, collection attribution relationship and historical source relationship with type reference, copy, comparison or observation derivation. When the same actual data collection action is delivered repeatedly, only the occurrence record or the historical source relationship is added, without adding the data collection identity equivalence class; When an occurrence record located in the historical delivery domain has an original valid observation time interval and an independent observation acquisition root with a verified or corroborated status, the corresponding normative observation event is not excluded because the occurrence record is located in the historical delivery domain, and the three-valued evidence gating is allowed when all gate control components are true; When an occurrence record located in the current delivery domain is only referenced back to the task source observation event through the historical source relationship and there is no independent actual acquisition assertion path, it is prohibited to establish a new observation acquisition root, standard observation event, or valid performance record based on this. It is also prohibited to create a review task using the occurrence record as a new review suggestion source, and it is prohibited to reset the task time anchor point, the lower bound of the window, and the deadline of an existing review task.

10. A health checkup status control system based on task-relative observation of bloodline, characterized in that, The relative observation lineage of the task refers to the relationship formed between the candidate observation event and the review task in terms of the task source observation event, the collection identity, and the historical source path, including: The occurrence record parsing module is used to parse the electronic medical examination data carrier of the same examinee to obtain occurrence records carrying delivery domain markers and evidence nature; The data acquisition identity merging module is used to perform alias merging on the original data acquisition identifiers of the occurrence records, and to establish observation acquisition roots and standard observation events according to the data acquisition identity equivalence classes; The task anchoring module is used to anchor the review suggestions in the occurrence record to the task source observation event, and generate a stable task identifier, lower window bound and deadline of the review task according to the task time anchor point and the performance offset interval. The evidence gating module is used to refer to the task source observation event, traverse the historical source relationships of candidate observation events to form a source closure, and perform three-valued evidence gating on the independent collection identity and non-pure backreference. The candidate observation event is a standard observation event selected from the standard observation events that initially matches or has an uncertain matching result with the review target range of the review task. The non-pure backreference means that the result source of the candidate observation event does not only refer back to the task source observation event. The status control module is used to classify candidate observation events that have passed the three-valued evidence gating according to the lower bound of the window and the deadline, determine the unique task status according to the preset priority based on the current calculation time, and output the longitudinal verification clues.

Images