Intelligent pre-inquiry form generation method and system for children tumor

Abstract

The invention provides an intelligent pre-inquiry form generation method and system for a child tumor, and relates to the technical field of intelligent medical inquiry, and the method comprises the steps: firstly recognizing the type of a child tumor pre-inquiry scene, and extracting scene demand features to form a set; and obtaining multi-source input data such as a parent preliminary symptom description text, a child past health record and an examination result sheet, and performing association processing. And then adjusting child tumor large model parameter configuration based on the scene demand feature set, and inputting an associated data set to generate an inquiry entry semantic prototype. According to parent expression habit sample optimization expression logic, a preliminary pre-inquiry form is formed, and parent feedback is collected to form a multi-dimensional feedback data set. And finally, inputting the feedback data set into the model to generate a form field optimization scheme and a preliminary diagnosis suggestion draft, and adjusting the preliminary pre-inquiry form to obtain a final child tumor intelligent pre-inquiry form containing a preliminary diagnosis direction. According to the invention, the scene adaptability and accuracy of pre-inquiry are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent medical diagnosis, in particular to an intelligent pre-diagnosis form generation method and system for children's tumors. BACKGROUND

[0002] In the field of children's tumor treatment, pre-diagnosis is a key preliminary link in the disease diagnosis and treatment process. Currently, children's tumor pre-diagnosis mainly relies on medical personnel to collect information through face-to-face communication or parents filling out simple questionnaires. However, there are differences in different pre-diagnosis scenarios. Initial diagnosis pre-diagnosis needs to collect children's basic information, symptom manifestations, etc.; follow-up scenarios focus on understanding disease progression, treatment effect, etc.; and special symptom screening scenarios mainly ask in-depth questions about specific symptoms. The existing pre-diagnosis method cannot be flexibly adjusted according to different scenarios, and cannot accurately meet the diversified needs of different scenarios.

[0003] At the same time, when parents describe children's symptoms, the expression methods are various and may be inaccurate and incomplete, and the data sources such as children's past health records and examination result sheets are scattered, lacking effective association and integration, resulting in low information utilization efficiency. In addition, the existing pre-diagnosis form is of fixed format, without fully considering the expression habits of parents, with many professional terms, which is difficult for parents to understand, and the form content and diagnosis suggestions cannot be optimized in time according to the feedback of parents, affecting the accuracy and efficiency of pre-diagnosis, and being not conducive to the early detection and precise treatment of children's tumors. SUMMARY

[0004] In view of the above-mentioned problems, in combination with the first aspect of the present application, the present application embodiment provides an intelligent pre-diagnosis form generation method for children's tumors, which comprises:

[0005] Recognizing the scene type of children's tumor pre-diagnosis, extracting the scene demand characteristics corresponding to each scene type to form a scene demand characteristic set, the scene type including initial diagnosis pre-diagnosis scene, follow-up scene and special symptom screening scene;

[0006] Obtaining multi-source input data related to pre-diagnosis, and performing association processing on the multi-source input data to obtain an associated data set, the multi-source input data including parents' preliminary symptom description text, children's past health records and parents' uploaded examination result sheets;

[0007] Adjusting the parameter configuration of the children's tumor large model based on the scene demand characteristic set to obtain an adjusted children's tumor large model adapted to the scene, the parameter configuration including case report form structure generation parameters, examination result association weights and preliminary diagnosis suggestion generation thresholds;

[0008] The association data set is input into the adjusted child tumor large model to generate an inquiry entry semantic prototype containing case report form standard fields, and the inquiry entry semantic prototype is associated with examination result verification items and preliminary diagnosis direction prompts;

[0009] The expression logic of the inquiry entry semantic prototype is optimized according to the parent expression habit sample, professional case report form fields are converted into expressions understandable by parents, and associated guiding sentences of symptoms and examination results are supplemented to form a preliminary pre-diagnosis form. The preliminary pre-diagnosis form is pushed to the parent interaction terminal, and the content of the form filled by the parents, the supplementary symptom description and the supplementary explanation of the examination results are collected, and the form filling time, the field modification times and the skipping behavior of the parents are recorded to form a multi-dimensional feedback data set;

[0010] The multi-dimensional feedback data set is input into the adjusted child tumor large model to analyze the correlation between the form content and the examination results and the causes of the parent interaction obstacles, generate a form field optimization scheme and a preliminary diagnosis suggestion draft, adjust the field order and expression content of the preliminary pre-diagnosis form based on the form field optimization scheme, and supplement diagnostic related verification items in combination with the preliminary diagnosis suggestion draft to obtain a final intelligent pre-diagnosis form of child tumor containing a preliminary diagnosis direction.

[0011] In still another aspect, the embodiments of the present application also provide an intelligent pre-diagnosis form generation system for child tumor, comprising:

[0012] A processor; a machine readable storage medium for storing machine executable instructions of the processor; wherein the processor is configured to execute the machine executable instructions to perform the intelligent pre-diagnosis form generation method for child tumor described above.

[0013] In still another aspect, the embodiments of the present application also provide a computer program product, the computer program product comprising machine executable instructions stored in a computer readable storage medium, and the processor of the intelligent pre-diagnosis form generation system for child tumor reads the machine executable instructions from the computer readable storage medium, and the processor executes the machine executable instructions, so that the intelligent pre-diagnosis form generation system for child tumor executes the intelligent pre-diagnosis form generation method for child tumor described above.

[0014] Based on the above, by identifying different scenario types for pediatric oncology pre-consultation and extracting corresponding scenario requirement features to form a set, the focus of pre-consultation in each scenario can be identified, effectively improving the scenario adaptability of pre-consultation. Acquiring and associating multi-source input data can integrate scattered information, improve data utilization efficiency, and make the information used in pre-consultation more comprehensive and accurate. Adjusting the parameter configuration of the pediatric oncology large-scale model based on the scenario requirement feature set yields a model adapted to the scenario, generating semantic prototypes of consultation items that better meet scenario requirements, enhancing the targeting of pre-consultation. Converting professional terms into language easily understood by parents and supplementing with related guiding statements forms a preliminary pre-consultation form, improving the convenience and accuracy of parents filling out the form. Collecting multi-dimensional feedback datasets and inputting them into the model for analysis allows for continuous optimization of form fields and generation of preliminary diagnostic suggestions. Ultimately, a final intelligent pre-consultation form for pediatric oncology containing preliminary diagnostic directions is obtained, achieving dynamic optimization of the pre-consultation form and accurate generation of diagnostic suggestions, improving the efficiency, accuracy, and intelligence level of pediatric oncology pre-consultation. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the execution flow of the intelligent pre-diagnosis form generation method for pediatric tumors provided in an embodiment of the present invention.

[0016] Figure 2 This is a schematic diagram of exemplary hardware and software components of the intelligent pre-diagnosis form generation system for pediatric tumors provided in an embodiment of the present invention. Detailed Implementation

[0017] The present invention will now be described in detail with reference to the accompanying drawings. Figure 1 This is a flowchart illustrating a method for generating intelligent pre-consultation forms for pediatric tumors according to an embodiment of the present invention. The following is a detailed description of this method for generating intelligent pre-consultation forms for pediatric tumors.

[0018] Step S110: Identify the scenario types for pre-diagnosis of pediatric tumors, extract the scenario requirement features corresponding to each scenario type, and form a set of scenario requirement features.

[0019] This example illustrates the scenario of a three-year-old child suspected of having neuroblastoma during their initial pre-diagnosis consultation. In this scenario, the parents bring the child for their first visit, and the pre-diagnosis system is used to comprehensively collect information to assist in the initial diagnosis.

[0020] Step S111: Obtain the scene identifier selected by the parent through the scene selection entry on the parent interaction terminal, or identify the scene tendency based on the initial input of the parent, and determine the current pre-consultation scene type.

[0021] Parents see scenario selection options on the interactive homepage, including "Initial Consultation / Pre-consultation," "Follow-up Visit," and "Specific Symptom Screening." Clicking the "Initial Consultation / Pre-consultation" option confirms the current scenario as such. If parents do not select this option directly but instead enter "My child first discovered a lump in their abdomen" in the text box, the system analyzes phrases like "first time discovered" to identify the initial consultation scenario.

[0022] Step S112: If the scenario type is the initial consultation and pre-consultation scenario, extract the demand features of the initial consultation and pre-consultation scenario, including the demand for complete medical history collection, the demand for basic physiological indicator coverage, the demand for preliminary verification of examination results, and the demand for preliminary diagnosis guidance. Extract the core focus content corresponding to each demand feature and record it.

[0023] After determining the scenario as an initial pre-consultation, the core focus for collecting a complete medical history includes the child's gestational age at birth, delivery method, birth weight, height and weight gain during growth, teething and language / motor development timing, any history of recurrent infections, allergies, or surgeries, and any family history of tumors or hereditary diseases. The core focus for covering basic physiological indicators includes vital signs such as body temperature, heart rate, respiratory rate, blood pressure, and blood oxygen saturation, as well as current weight, height, and head circumference. The initial verification of examination results focuses on the completeness of the reports uploaded by the parents, the presence of any abnormal indicators, and the qualifications of the examination institution. The initial diagnostic guidance involves suggesting possible tumor types based on existing information, recommending further examinations, and using language that will not cause undue panic among parents. All of the above core focus points should be recorded in detail.

[0024] Step S113: If the scenario type is a follow-up visit scenario, extract the requirement features of the follow-up visit scenario, including the requirement to record changes in symptoms after treatment, the requirement to compare recent test results, the requirement to generate follow-up conclusions, and the requirement to suggest the next follow-up time. Extract and record the time range limit corresponding to each requirement feature.

[0025] When the scenario is a follow-up visit, the time frame for recording changes in post-treatment symptoms is limited to the period from the last visit to this follow-up visit, and changes in the frequency, duration, and severity of symptoms must be recorded. The time frame for comparing recent test results is to compare test reports from the past few months with the baseline report. The time frame for generating follow-up conclusions is the day of this follow-up visit, generating a phased assessment based on current symptoms and test results. The time frame for suggesting the next follow-up time is determined based on the treatment stage; for example, a follow-up visit may be suggested two weeks later during maintenance treatment, and three months later during the recovery period. Record the time frame limits corresponding to each requirement feature.

[0026] Step S114: If the scenario type is a specific symptom screening scenario, extract the demand features of the specific symptom screening scenario, including the demand for detailed description of the target symptom, the demand for examination of the symptom-related sites, the demand for inquiry of specific tumor indicators, and the demand for screening conclusion prompts. Extract and record the tumor type association range corresponding to the target symptom.

[0027] For a specific symptom screening scenario, let's assume the target symptom is unexplained fever accompanied by joint pain. The detailed description of the target symptom should include the onset time, fever pattern, highest temperature, method and effectiveness of fever reduction, and the location, nature, and degree of limitation of movement of the joint pain. The required related examinations should include bone marrow aspiration, joint ultrasound, and whole-body PET-CT. Inquiries about specific tumor markers should cover indicators such as complete blood count, inflammatory markers, and tumor markers. The screening conclusions should be limited to tumor types highly correlated with fever and joint pain. By querying a medical knowledge base, the range of tumor types corresponding to the target symptom should be extracted and recorded.

[0028] Step S115: Provide a textual description of the requirements for each scenario type and indicate the priority order of the requirements.

[0029] The four key requirements for the initial consultation and pre-diagnosis scenario are described in text. The requirement for a complete medical history is described as "comprehensively collecting health-related information from birth to the present, providing basic data support for diagnosis"; the requirement for coverage of basic physiological indicators is described as "obtaining the child's current vital signs and growth and development indicators to assess overall health status"; the requirement for preliminary verification of examination results is described as "verifying the validity of the examination reports uploaded by parents and the correlation between abnormal indicators and symptoms"; and the requirement for preliminary diagnostic guidance is described as "providing possible diagnostic directions and suggestions for further examinations without causing panic among parents." Following clinical diagnostic logic, the requirements for a complete medical history and coverage of basic physiological indicators are marked as high priority, the requirement for preliminary verification of examination results as medium priority, and the requirement for preliminary diagnostic guidance as low priority.

[0030] Step S116: Organize the requirement feature descriptions of different scenario types according to the format of scenario type, requirement feature, priority, and core content, remove duplicate requirement feature descriptions between different scenarios, and form a set of scenario requirement features.

[0031] The requirement feature descriptions for each scenario type are organized according to the format of scenario type, requirement characteristics, priority, and core content. For example, in the initial consultation and pre-diagnosis scenario, the requirement for complete medical history collection corresponds to the appropriate priority and core content. By comparing the requirement feature descriptions of different scenarios, duplicate content is removed. For example, if "verification of examination results" appears in both the initial consultation and special screening scenarios, the feature description of the higher-priority scenario is retained, ultimately forming a set of scenario requirement features.

[0032] Step S120: Obtain multi-source input data related to pre-consultation, perform correlation processing on the multi-source input data, and obtain a correlation data set.

[0033] In the aforementioned initial diagnosis scenario, the multi-source input data includes preliminary symptom description text entered by parents, such as "the child has had recurrent abdominal pain for the past month, a mass can be felt in the lower right abdomen, and weight loss," the child's past health records, such as full-term vaginal delivery at birth, no history of asphyxia, previous hospitalization for bronchitis, timely vaccination, and abdominal ultrasound reports and blood routine test results uploaded by parents from other hospitals.

[0034] Step S121: Collect the preliminary symptom description text entered by parents through the text input box on the parent interaction terminal, format the preliminary symptom description text, remove preset symbols and repetitive expressions, and retain the description of the symptom occurrence time, manifestation, duration and factors that relieve or aggravate the symptoms.

[0035] Parents enter the following text into the input box: "My child has had stomach pain for over a month, on the right side of his belly button. He cries when pressed, has recently lost his appetite, and has lost weight. He can't sleep at night because of the pain, but feels better in the morning?". The text is formatted by removing pre-defined symbols such as "!" and "?". Duplicate expressions are identified and removed; for example, between "stomach pain" and "pain on the right side of the belly button," the latter is retained. The extracted information shows that the symptoms have been present for over a month, manifesting as right lower abdominal tenderness, decreased appetite, and weight loss. The symptoms are persistent and worsen at night, relieved in the morning, and aggravated by pressure.

[0036] Step S122: Obtain the child's past health records through the history import interface of the interactive terminal, sort the child's past health records in chronological order, and extract the key health events in each child's past health records. The child's past health records include past medical records, treatment plan records, and past examination reports.

[0037] Parents authorize access to their child's past health records via the history import interface on the interactive platform. The records include: a hospitalization at a community hospital for acute bronchitis, with treatment including anti-infection medication and nebulized inhalation; a physical examination at a pediatric clinic, showing normal height and weight, and normal blood and urine tests; and an emergency room visit for febrile seizures, with a normal electroencephalogram (EEG). These records are then sorted chronologically, and key health events are extracted, such as "hospitalization for acute bronchitis with anti-infection medication and nebulized inhalation at a certain time."

[0038] Step S123: Receive the examination result sheets uploaded by parents through the file upload portal of the interactive terminal, parse the text content and numerical data of each examination result sheet, and extract indicator information related to the diagnosis of childhood tumors. The examination result sheets include imaging examination reports, laboratory test reports and pathology examination reports.

[0039] In the initial diagnosis scenario of suspected neuroblastoma, the parents uploaded three test results: an abdominal ultrasound report from another hospital (PDF format), a blood routine test report (JPG format), and a pathology consultation report from another hospital (TXT format).

[0040] For example, step S1231: Receive the examination result sheet uploaded by the parent through the file upload portal of the parent interaction terminal, identify the format of the uploaded examination result sheet, and distinguish between imaging examination reports, laboratory test reports and pathology examination reports. The examination result sheet includes imaging examination reports, laboratory test reports and pathology examination reports. Supported examination result sheet formats include image format, portable document format and text format.

[0041] After receiving the file, the file upload module first reads the file's binary header information: the abdominal ultrasound report is identified as a portable document using the "%PDF-" identifier; the blood routine test report is identified as an image format using the "FFD8FF" file header; and the pathology consultation report is identified as a text format using ASCII character stream detection. Then, the medical document classification model is invoked, combining the keywords "ultrasound," "blood routine," and "pathology" in the filenames to classify the three reports as imaging examination reports, laboratory test reports, and pathology examination reports, respectively. The classification results are stored in the file metadata for subsequent parsing process routing.

[0042] Step S1232: If the examination result sheet is an imaging examination report, extract the text content of the imaging examination report through optical character recognition technology, locate the examination site, imaging manifestations, and diagnostic opinion fields in the text, extract the specific organ or tissue name corresponding to the examination site, extract the lesion description content in the imaging manifestations, and extract the preliminary judgment result in the diagnostic opinion.

[0043] For abdominal ultrasound reports (imaging examination reports), the parsing engine calls the PDF text extraction interface to obtain the text layer, and the scanned pages use optical character recognition technology (medical-specific OCR model). The BERT named entity recognition model is used to locate fields: from "Examination location: entire abdomen (focusing on the adrenal region)", the specific organ name "adrenal region (right)" is extracted; from "Imaging findings: a hypoechoic mass is seen in the right adrenal region, with indistinct borders, heterogeneous internal echoes, and abundant blood flow signals", the lesion description (location, echo characteristics, borders, internal structure, blood flow signals) is extracted; from "Diagnosis: neuroblastoma is suspected, enhanced CT is recommended", the preliminary judgment result "neuroblastoma is suspected" is extracted. The extracted information is stored in a "examination type-location-features-conclusion" structure.

[0044] Step S1233: If the test result sheet is a laboratory test report, extract the test item name, test result value, reference range, and unit fields from the laboratory test report, and filter out the test items related to the diagnosis of childhood tumors; for the filtered test items, record the test result value, reference range, and unit of each test item, mark the test items whose test result values ​​exceed the reference range as abnormal indicators, and record the direction of the abnormality.

[0045] After image format correction, the complete blood count (CBC) report is used to extract four columns of data: "Test Item, Result, Unit, and Reference Range" using a table-based detection algorithm. The report is then compared against a core pediatric tumor marker library (containing 58 indicators such as NSE and LDH) to identify relevant items: hemoglobin (result, unit, reference range), platelet count (normal), white blood cell count (normal), and lactate dehydrogenase (above the upper limit of the reference range). A numerical comparison algorithm identifies decreased hemoglobin and increased lactate dehydrogenase as abnormal indicators, recording the direction of the abnormality. Normal indicators are marked as "Normal." The results are stored in a structure of "Test Item - Result - Unit - Reference Range - Abnormal Marker."

[0046] Step S1234: If the examination result is a pathology report, extract the specimen type, pathological diagnosis, and immunohistochemical result fields from the pathology report, record the tissue source corresponding to the specimen type, extract the conclusive description in the pathological diagnosis, and extract the positive and negative indicators in the immunohistochemical results.

[0047] The pathology consultation report (pathology examination report) uses regular expressions to locate the field: "Specimen type: Right adrenal region puncture biopsy tissue" extracts the tissue source "adrenal region (right)"; "Pathological diagnosis: Small round blue cell tumor, combined with immunohistochemistry consistent with neuroblastoma" extracts the conclusive description "neuroblastoma"; "Immunohistochemistry: Syn(+), CgA(+), NSE(+), Ki-67 (70%+), LCA(-), CK(-)" is split by semicolons to identify "(+)" as positive indicators (Syn, CgA, NSE, Ki-67) and "(-)" as negative indicators (LCA, CK), with the positive percentage of Ki-67 recorded simultaneously. The extracted indicators are compared with the neuroblastoma marker library to verify consistency.

[0048] Step S1235: Organize the examination site, lesion description, and diagnosis opinion extracted from the imaging examination report, the tumor-related test items, values, and abnormal markers extracted from the laboratory test report, and the specimen type, pathological diagnosis, and immunohistochemical results extracted from the pathology examination report according to the report type, extracted fields, and specific content format.

[0049] The integration module organizes data according to a three-level structure: "Report Type - Extracted Fields - Specific Content". Imaging reports correspond to "Examination Location: Adrenal Region (Right); Lesion Description: Hypoechoic mass, indistinct borders, rich blood flow; Diagnosis: Possible neuroblastoma". Laboratory test reports correspond to "Tumor-related Items: Lactate dehydrogenase (elevated), Hemoglobin (decreased)". Pathological examination reports correspond to "Specimen Type: Adrenal Region puncture tissue; Pathological Diagnosis: Neuroblastoma; Immunohistochemistry: Syn (+), CgA (+), NSE (+), Ki-67 (70%+)". Data is stored in structured JSON format, with each report type as the top-level key and extracted fields as secondary keys.

[0050] Step S1236: Compare the organized information with the common childhood tumor indicator database, supplement the tumor association descriptions corresponding to the indicators, and form single indicator data of the examination results.

[0051] The database of common childhood tumor markers was accessed, and the compiled markers were compared with entries in the database: elevated lactate dehydrogenase was associated with "common abnormalities in solid tumors such as neuroblastoma and lymphoma"; immunohistochemistry Syn(+) and CgA(+) were associated with "positive neuroendocrine tumor markers, supporting the diagnosis of neuroblastoma"; Ki-67 (70%+) was associated with "high cell proliferation activity, suggesting a high degree of tumor malignancy". After supplementing the association explanations, the single-marker data of the test results, including the original extracted information and tumor association explanations, were completely stored in the database for subsequent association analysis.

[0052] Step S124: Associate the symptom keywords in the preliminary symptom description text with key health events in the child's past health records, identify the correlation between the current symptoms and past health events, extract the correlation results and record them.

[0053] Keywords such as "abdominal pain," "abdominal mass," and "weight loss" were extracted from the initial symptom description and mapped to UMLS standard terminology. Past health events in the child included "hospitalization for bronchitis at 6 months," "febrile seizures at 1 year," and "normal routine checkups." Cosine similarity scores calculated using the symptom-disease association strength algorithm were all below the threshold (0.6), indicating no direct association. Time-series analysis showed no abdominal abnormalities in the most recent checkup (3 months ago), suggesting a new onset of symptoms. The association result was recorded as "Current symptoms have no significant association with past health events; recent new abnormality," and stored in the association analysis results table.

[0054] Step S125: Associate the symptom location in the preliminary symptom description text with the indicator information in the examination result sheet, match the examination indicators corresponding to the symptom location, extract the associated indicator name and value, and record them.

[0055] A mapping relationship was constructed using a human anatomical location ontology database. The "right lower abdomen" was matched with the "right adrenal gland region" using a spatial topology algorithm, with an overlap of 0.82 (above the threshold of 0.7), thus determining the associated location. Corresponding indicators for the right adrenal gland region were extracted: imaging findings showed "hypoechoic mass (5cm × 4cm)" and "abundant blood flow signal"; laboratory tests showed "elevated lactate dehydrogenase" and "decreased hemoglobin"; and pathological examination diagnosed "neuroblastoma". The association results were stored in a "symptom location - examination location - matching degree - indicator list" format, where the indicator list included name, value, unit, and abnormal markers.

[0056] Step S126: Correlate key health events in the child's past health records with recent indicators in the examination results sheet, analyze the trend of indicator changes after treatment, extract the trend analysis results and record them.

[0057] Laboratory data from the past year were extracted, with a focus on comparing similar indicators: lactate dehydrogenase was normal three months ago, but is currently elevated by more than 50%. The rate of increase, combined with the absence of any specific treatment history, was determined to be related to a space-occupying lesion; hemoglobin decreased from the lower limit of normal to a slight decrease, which, combined with weight loss, suggests chronic wasting. The trend analysis results were recorded as "significantly elevated lactate dehydrogenase (compared to baseline), slightly decreased hemoglobin, associated with a space-occupying lesion in the right adrenal region," and a trend curve was generated and stored.

[0058] Step S127: Integrate the correlation results between symptoms and past records, the correlation results between symptoms and examination indicators, and the trend analysis results between past records and recent indicators. At the same time, attach the original preliminary symptom description text, the sorted past health records of children, and the parsed single-indicator data of examination results to form a correlation data set.

[0059] The associated dataset uses a nested JSON structure: the top layer contains "Raw Data" and "Association Analysis" modules; "Raw Data" stores formatted symptom descriptions, past health records in reverse chronological order, and parsed examination indicators; "Association Analysis" includes symptom-past records, symptom location-examination indicators, and indicator trend analysis results. Data is associated with pre-consultation session IDs via UUIDs, and after integrity verification and supplementation of missing metadata, it is stored in a distributed database.

[0060] Step S130: Adjust the parameter configuration of the pediatric tumor big model based on the set of scenario requirements features to obtain an adjusted pediatric tumor big model adapted to the scenario.

[0061] Based on the set of scenario requirements for initial diagnosis and pre-diagnosis, the parameters of the large-scale pediatric tumor model are adjusted to adapt it to the scenario.

[0062] Step S131: Extract the requirement features corresponding to the current scenario type from the scenario requirement feature set, and determine the adjustment direction of each requirement feature for the model parameters.

[0063] From the set of scenario requirements, the following requirements for the initial consultation and pre-diagnosis scenario are extracted: complete medical history collection, basic physiological indicator coverage, preliminary verification of examination results, and preliminary diagnostic guidance. Based on the correspondence between requirements and model parameters, it is determined that the complete medical history collection requirement corresponds to adjusting the field completeness parameter in the case report form structure generation parameters, with the direction of maximizing coverage; the basic physiological indicator coverage requirement corresponds to adjusting the field type parameter, with the direction of including vital signs and growth indicators; the preliminary verification of examination results requirement corresponds to adjusting the correlation weight of examination results, with the direction of increasing the correlation weight between imaging examinations and symptoms; and the preliminary diagnostic guidance requirement corresponds to adjusting the threshold for generating preliminary diagnostic suggestions, with the direction of lowering the threshold to allow the generation of preliminary directions based on partial data.

[0064] Step S132: If the current scenario is an initial consultation and pre-consultation scenario, set the complete case report form field generation mode for the form structure generation parameters, and set the form to include basic information field, medical history field, symptom field, examination result field and preliminary diagnosis prompt field. At the same time, adjust the field generation order parameter so that the fields are arranged according to the preset logic.

[0065] In the initial consultation and pre-diagnosis scenario, the form structure generation parameters of the case report form are set to full mode. The form includes basic information fields, such as the child's name, gender, and date of birth; medical history fields, such as birth history and growth and development history; symptom fields, such as current main symptoms and accompanying symptoms; examination result fields, such as imaging examination and laboratory test results; and preliminary diagnosis suggestion fields, such as possible diagnostic directions and suggested examinations. The field generation order is adjusted to basic information, medical history, symptoms, examination results, and preliminary diagnosis suggestion, so that the fields are arranged according to this preset logic.

[0066] Step S133: For the weight parameters associated with the examination results, if the scenario requirement features include the requirement for preliminary verification of examination results, adjust the weight values ​​associated with the examination result indicators and the consultation items to values ​​that meet the requirements for preliminary verification of examination results, so that the generated consultation items are preferentially associated with the uploaded examination result single indicator.

[0067] Since the initial consultation scenario requires preliminary verification of examination results, the weight of the association between imaging examination indicators and consultation items is increased. This allows the generated consultation items to be preferentially associated with uploaded imaging examination indicators such as abdominal ultrasound reports, facilitating preliminary verification of examination results.

[0068] Step S134: Generate threshold parameters for preliminary diagnostic suggestions. Based on the preliminary diagnostic direction guidance requirements in the scenario, set a threshold for generating diagnostic suggestions that meets the preliminary diagnostic direction guidance requirements. This allows preliminary diagnostic direction prompts to be generated after obtaining basic symptoms and preliminary examination results. At the same time, the scope of the diagnostic suggestions is limited to avoid absolute statements.

[0069] Based on the need for preliminary diagnostic guidance, the threshold for generating preliminary diagnostic suggestions is set to a low level, so that preliminary diagnostic guidance can be generated after obtaining basic symptoms and preliminary examination results. At the same time, the wording of diagnostic suggestions is restricted, avoiding the use of absolute terms such as "confirmed as," and instead using expressions such as "possibly" or "cannot be ruled out."

[0070] Step S135: If the current scenario is a follow-up visit, adjust the form structure generation parameters of the case report form to the follow-up-specific mode, set the form to retain the fields of previous treatment records and current symptom changes, add the fields of treatment effect evaluation and next follow-up plan, and adjust the field generation order parameters so that the fields are arranged according to the preset logic.

[0071] When the scenario is a follow-up visit, adjust the form structure generation parameters of the case report form to a follow-up-specific mode. The form retains the previous treatment record field, recording previous treatment plans, medication dosages, etc., and the current symptom change field, recording changes in symptoms compared to the last follow-up visit. Add a treatment effectiveness evaluation field to assess the treatment's effect, and a next follow-up plan field to determine the time and examinations for the next follow-up visit. Adjust the field generation order to: previous treatment record, current symptom change, treatment effectiveness evaluation, and next follow-up plan.

[0072] Step S136: Adjust the correlation weight parameter of the examination results, adjust the comparison weight of the recent examination results sheet and the previous examination results sheet to a value that meets the requirements of the examination results comparison, so that the generated consultation items include the content of asking about the reasons for the changes in the examination results; adjust the threshold parameter for generating the preliminary diagnosis suggestion to a value that meets the requirements of complete data, and generate follow-up conclusion suggestions only after obtaining the comparison data of post-treatment symptoms and examination results sheets.

[0073] In follow-up visits, increase the weight of comparing recent and previous test results, ensuring that the generated consultation entries include inquiries about the reasons for changes in test results. Adjust the threshold for generating preliminary diagnostic suggestions to a higher value, so that follow-up conclusion suggestions are only generated after obtaining data comparing post-treatment symptoms and test results.

[0074] Step S137: Input the adjusted case report form structure generation parameters, examination result association weights, and preliminary diagnostic suggestion generation thresholds into the parameter adjustment interface of the pediatric tumor big data model.

[0075] Input the adjusted case report form structure generation parameters, examination result association weights, and preliminary diagnostic suggestion generation thresholds into the parameter adjustment interface of the pediatric tumor model to complete the model parameter configuration.

[0076] Step S138: Call the model adaptability test program, input test data that matches the current scenario, and obtain the output test questions and draft diagnostic suggestions.

[0077] The model adaptability testing program is invoked, and test data matching the initial diagnosis scenario is input into the adjusted pediatric tumor model to obtain the output test consultation items and draft diagnostic suggestions.

[0078] Step S139: Compare the fit between the test output results and the characteristics of the scenario requirements. Check whether the fields of the case report form cover the scenario requirements, whether the correlation of the result sheet meets the weight settings, and whether the diagnostic suggestions meet the threshold requirements.

[0079] Compare the degree of fit between the consultation items and draft diagnostic suggestions obtained from the comparative test and the needs of the initial consultation scenario. Check whether the fields of the case report form cover the requirements such as complete medical history collection and basic physiological indicators. Check whether the association of the result sheet meets the set weight and whether the diagnostic suggestions meet the generation threshold requirements.

[0080] Step S1310: If the fit does not meet expectations, readjust the corresponding parameters and test again until the test output fully matches the requirements of the scenario, and obtain the adjusted pediatric tumor large model adapted to the scenario.

[0081] If the test output does not match the expected characteristics of the scenario, analyze the reasons for the mismatch, readjust the corresponding model parameters, such as the integrity parameters of the case report table fields and the correlation weight of the examination results, and then test again. Repeat this process until the test output fully matches the characteristics of the scenario and obtain an adjusted large model of pediatric tumors adapted to the initial diagnosis scenario.

[0082] Step S140: Input the associated data set into the adjusted pediatric tumor big model to generate a semantic prototype of the consultation item containing standard fields of the case report form. The semantic prototype of the consultation item is associated with the examination result verification item and the preliminary diagnosis direction prompt.

[0083] The aforementioned associated data set is input into the adapted pediatric tumor model. Based on the information in the associated data set, the model generates a semantic prototype of the consultation item containing standard fields of the case report form. This semantic prototype is associated with examination result verification items, such as verification of the size of the mass in the abdominal ultrasound report, and preliminary diagnostic direction suggestions, such as suggesting that it may be neuroblastoma.

[0084] Step S141: Extract the set of symptom keywords from the preliminary symptom description text, the sorted key events in the child's past health records, the parsed single-indicator data of the examination results, and the correlation results between the data from the associated data set, and organize them into input data groups according to data type.

[0085] Extract symptom keyword sets, such as "abdominal pain" and "weight loss," from the associated dataset; sorted key events in the child's past health records, such as "hospitalization for acute bronchitis" and "emergency treatment for febrile seizures"; parsed single-indicator data of examination results, such as the size of lesions on abdominal ultrasound and the lactate dehydrogenase value in blood routine tests; and the correlation results between the data. Classify and organize these data into input data groups according to data type so that they can be input into the model for processing.

[0086] Step S142: Input the input data group into the field mapping layer of the adjusted pediatric tumor large model case report form, generate parameters based on the case report form structure, match the content in the input data group with the fields of the standard case report form, and determine the field types of the case report form to be generated.

[0087] The input data set is fed into the model's case report form field mapping layer. Based on the previously set case report form structure, parameters are generated. The content in the input data set is matched with the standard case report form fields to determine the types of fields to be generated, such as basic information fields and medical history fields.

[0088] Step S143: For each determined field type of the case report form, generate a basic expression framework, which includes field names, prompts for the inquiry content, and input format hints.

[0089] For each identified field type in the case report form, a basic description framework is generated. For example, the basic description framework for the basic information field includes the field name "Child's Name", the prompt "Please enter the child's name", and the input format prompt "Please enter the Chinese name".

[0090] Step S144: Based on the correlation weight of the examination results, identify the abnormal items of the single indicator data of the examination results in the input data group, generate the corresponding examination result verification items for the abnormal indicators, and add them to the basic expression framework of the corresponding case report table fields.

[0091] Based on the correlation weight of the examination results, identify abnormal items in the single indicator data of the examination results in the input data group, such as the abnormal indicator of elevated lactate dehydrogenase in blood routine test. Generate a verification item for the examination result, such as "Your uploaded blood routine report shows that lactate dehydrogenase is higher than the normal range. Does your child have any other discomfort?", and add the verification item to the basic expression framework of the corresponding case report table field.

[0092] Step S145: Based on the preliminary diagnostic suggestion generation threshold, analyze the correlation between symptom keywords, children's past health records and single indicators of examination results in the input data group. If the correlation results reach the diagnostic suggestion generation threshold, generate a preliminary diagnostic direction prompt and add the prompt content to the corresponding case report table field according to the preset format.

[0093] Based on the preliminary diagnostic suggestions, a threshold is generated, and the correlation between symptom keywords "abdominal pain" and "weight loss", children's past health records, and single indicators of examination results such as right adrenal gland lesions and elevated lactate dehydrogenase is analyzed. When the correlation results reach the set threshold, a preliminary diagnostic direction suggestion is generated, such as "considering the symptoms and examination results, neuroblastoma is a possibility", and then added to the corresponding case report form field according to the preset format.

[0094] Step S146: Semantically integrate the generated content, which includes the basic expression framework of the case report form fields, the verification items of the examination results, and the preliminary diagnosis direction prompts. Remove expressions that are irrelevant to the current scenario requirements from the semantically integrated content, and arrange the integrated content according to the order of the case report form fields. Each field corresponds to a complete expression containing basic inquiries, verification items, and prompts, forming a semantic prototype of the consultation item.

[0095] The generated content, including the basic framework of case report form fields, verification items for examination results, and preliminary diagnostic guidance, is semantically integrated to make the expression more coherent. Expressions irrelevant to the initial consultation scenario, such as treatment effect evaluation content related to follow-up visits, are removed. The integrated content is arranged according to the order of case report form fields, with each field forming a complete expression containing basic questions, verification items, and prompts, thus forming the semantic prototype of consultation items.

[0096] Step S150: Optimize the semantic prototype expression logic of the consultation items based on the parents' expression habits sample, convert the professional case report form fields into expressions that are easy for parents to understand, supplement the guiding statements that relate symptoms and examination results, form a preliminary pre-consultation form, push the preliminary pre-consultation form to the parents' interactive terminal, collect the form content filled in by parents, supplementary symptom descriptions and supplementary explanations of examination results, and record the parents' form filling time, number of field modifications and skipping behavior, and integrate them to form a multi-dimensional feedback dataset.

[0097] For example, in the initial diagnosis scenario of suspected neuroblastoma, after optimizing the semantic prototype of the consultation items, a preliminary pre-consultation form containing 23 items is generated and pushed to parents through the form display interface on the parent's interactive terminal. The form adopts a paginated loading mode, displaying 5 items per page, with "Previous Page / Next Page" navigation buttons and "Save Draft" and "Submit" function keys at the bottom. Parents start filling in the form from the "Basic Information" page, and complete the content of the four modules "Medical History Collection", "Symptom Description", "Examination Result Confirmation", and "Preliminary Diagnosis Guidance" in sequence. The system records the filling behavior data in real time, and finally integrates it into a multi-dimensional feedback dataset containing form content, supplementary explanations, and interactive behaviors.

[0098] Step S151: Collect expression habit samples from parents with different backgrounds, and divide them into symptom description sample group, terminology understanding sample group and examination result feedback sample group according to the expression scenario. The expression habit samples include the expression text of parents describing children's symptoms, understanding medical terminology and providing feedback on examination results.

[0099] For example, a multi-center clinical survey was used to collect samples of children's communication habits. The inclusion criteria were parents who had visited a child's cancer clinic in the past three years. Stratified sampling was conducted based on region (East / Central / West), education level (primary school to undergraduate and above), and type of cancer (solid tumor / hematologic malignancy), resulting in 1200 valid samples. The symptom description sample group included 400 parents' natural language descriptions of symptoms such as "abdominal pain," "fever," and "weight loss," such as "My child's stomach hurts in waves, like being pricked with needles," and "The fever always starts in the afternoon, and the fever goes down a little after taking antipyretics." The terminology comprehension sample group included 400 parents' feedback on their understanding of professional terms such as "neuroblastoma" and "lactate dehydrogenase," such as "I thought 'space-occupying lesion' meant a cyst," and "I don't understand what 'LDH' means." The examination result feedback sample group included 400 parents' interpretations of ultrasound and CT reports, such as "The ultrasound showed a 5-centimeter object, and the doctor said it might be a tumor," and "Anything with an upward-pointing arrow on the blood test report is abnormal, right?" After all samples are de-identified, they are stored in the corpus in the format of "sample ID-scene type-original text-annotated terminology".

[0100] Step S152: Analyze the text of the symptom description sample group, count the common words that parents use when describing symptoms, and form a symptom common word comparison table; analyze the text of the terminology understanding sample group, extract the names of professional case report form fields that are difficult for parents to understand and their corresponding common-sounding conversion methods, and form a case report form field common-sounding conversion table; analyze the text of the examination result feedback sample group, summarize the common expression logic used by parents when providing examination results, and form an examination result guidance expression template.

[0101] The symptom description sample group analysis used the term frequency-inverse document frequency (TF-IDF) algorithm to statistically obtain the top 50 common words such as "stomach ache" (corresponding to "abdominal pain"), "hot body" (corresponding to "fever"), and "lost weight" (corresponding to "weight loss"). A symptom common word comparison table was constructed, including fields of "professional term - common word - frequency of use - regional differences", such as "vomiting - vomited - 89.2% - the proportion of 'dry retching' in the North is 12%". The terminology comprehension sample group identified high-difficulty terms through perplexity calculation. For 32 professional fields such as "neuron-specific enolase" and "mesenchymal tumor", a three-level conversion structure of "full name - abbreviation - common explanation" was used to form a common-sound conversion table of case report table fields, such as "lactate dehydrogenase - LDH - an enzyme in the body that helps cells metabolize energy, elevated levels may indicate cell damage". The sample group of test results feedback used a sequence pattern mining algorithm to summarize the typical expression logic of "test type + abnormal manifestation + parental concern", such as "We had an ultrasound and they said there was something on the kidney. We are very afraid that it might be cancer". Based on this, a test result guidance expression template was designed, which includes three elements: "confirmation of test items + interpretation of key indicators + emotional guidance". It is divided into three sub-templates: "imaging examination", "laboratory test" and "pathological examination".

[0102] Step S153: Extract the professional case report table field names from the semantic prototype of the consultation item, and replace them with common expressions according to the common expression conversion table of case report table fields; extract the symptom-related expressions from the semantic prototype of the consultation item, and replace the professional symptom terms with common expressions according to the common expression conversion table of symptoms.

[0103] Thirty-eight professional field names and 27 symptom terms were extracted from the semantic prototype of the consultation entries and then simplified. Fields in the case report form, such as "family history of cancer," were replaced with "has anyone in your family had cancer?" and "past surgical history" with "has had any open surgery before?". Symptom terms, such as "persistent dull pain," were replaced with "a persistent, dull ache," and "intermittent fever" with "fever that comes and goes, sometimes high and sometimes low." A bidirectional maximum matching algorithm was used to ensure accurate terminology localization during the replacement process. Long sentences containing multiple professional terms were segmented for conversion; for example, "the child has unexplained weight loss and loss of appetite" was converted to "the child has lost weight for no apparent reason and has no appetite." The mapping relationship between the terms before and after conversion is stored in a version control table, supporting backtracking queries.

[0104] Step S154: In the semantic prototype of the consultation item, add related guiding statements before and after the description of the verification item of the inspection result, referring to the template for guiding description of the inspection result.

[0105] For verification items related to examination results, relevant content is added to the preceding introductory statement and the following follow-up question. For example, for the item "Confirmation of Ultrasound Examination Results," the preceding introductory statement is "Your uploaded abdominal ultrasound report shows an abnormality in the right adrenal region (in layman's terms: there's something growing on the kidney). Please confirm the following information," and the following follow-up question is "If you have any questions about the report, you can write down your questions here, and the doctor will answer them in a subsequent consultation." Referring to the template for guiding statements of examination results, imaging examination items should focus on adding guidance on "plain location of the examination site," such as "The 'right adrenal region' in the report is roughly located on the right side of the child's abdomen, slightly above the waist"; laboratory test items should focus on adding guidance on "significance of abnormal indicators," such as "'Elevated lactate dehydrogenase' may indicate cell damage and needs to be considered in conjunction with other examinations"; and pathology examination items should focus on adding guidance on "diagnostic relevance," such as "The 'small round blue cell tumor' mentioned in the pathology report is a type of tumor that requires further diagnosis." All guiding statements should be limited to 20 characters in length and should be in bold to distinguish them from the main content of the item.

[0106] Step S155: Perform a logical coherence check on the optimized description. Use a field logic order verification tool to compare the field order of the simplified description with the standard logic of the case report form. Use the description conflict detection module to identify and correct any conflicts between the guiding statements, verification items, and diagnostic prompts.

[0107] The system invokes a field logic order validation tool, loads the standardized case report form logic template developed by the International Society for Pediatric Oncology (SIOP), and compares the optimized form field order with it. For example, the standard logic requires "birth history" to precede "growth and development history." The tool detects that the current form places "growth and development history" first, automatically marks it as having an "abnormal logical order," and suggests adjustments. The order of "symptom duration" and "symptom relief factors" is deemed "logically consistent" because it conforms to the "time-feature" description logic. The statement conflict detection module employs a dual-track detection approach: the rule engine pre-sets 12 conflict rules, such as "the guiding statement must not contain diagnostic conclusions" and "verification items must be strongly correlated with the examination results." If it finds an absolute statement like "this result may be cancer," it automatically replaces it with "this result requires further examination to determine its nature." The deep learning model, based on the BERT conflict detection model, identifies implicit conflicts between the guiding statement "a child's weight loss is definitely due to malnutrition" and the subsequent "tumors may cause weight loss" in the "weight loss" entry, suggesting a change to "a child's weight loss may have multiple causes; let's understand them step by step."

[0108] Step S156: Obtain the level of understanding of the optimized description by parent representatives from different backgrounds and the questions they raised. Adjust the description again for the questions, and arrange all the adjusted items in the order of the fields in the optimized case report form. Each item includes basic questions in plain language, related guiding statements, examination result verification items and preliminary diagnosis direction prompts to form a preliminary pre-consultation form.

[0109] Fifteen parent representatives from different backgrounds (three tiers each of region, education level, and tumor type, with five representatives from each tier) were selected for a cognitive test. The "think aloud" method was used to record the reading process: parents read each item aloud and explained their understanding, while the duration of pauses (>3 seconds was considered a comprehension obstacle) and points of doubt were recorded. The test revealed that even after "right adrenal gland region lesion" was changed to "something grew on the kidney," six parents still misunderstood it as "growing inside the kidney." Further adjustments were made to "something grew on the right side of the abdomen, slightly above the waist (medically called a space-occupying lesion)." Similarly, after "Ki-67 index 70%" was changed to "high cell proliferation activity," eight parents could not understand it. The change was further clarified as "the higher this index, the faster the malignant cells inside grow." Based on feedback, 12 wording changes were made, and the final preliminary pre-consultation form contains 4 modules and 23 items. Each item consists of four parts: "Basic Inquiry (plain language) + Related Guidance (blue text) + Verification Item (editable text box / radio button) + Diagnostic Tips (gray text annotation)". For example, item S15-08: "When your child has abdominal pain, do they have any of the following symptoms? (Multiple selections allowed) [Basic Inquiry] → The ultrasound report you uploaded shows a mass in the right abdomen. Pain in this location is usually accompanied by these symptoms. [Related Guidance] → □ Sweating profusely from the pain □ Not wanting to walk □ Vomiting □ Other (please fill in) [Verification Item] → These symptoms help doctors determine whether the mass is compressing surrounding tissues. [Diagnostic Tips]".

[0110] Step S157: Load the items of the preliminary pre-consultation form into the form display interface of the parent interaction terminal in the optimized order. Each item has an independent filling area, which includes a text input box, an option selection button, and a supplementary explanation upload entry.

[0111] The form interface adopts a responsive design, adapting to mobile phones (portrait / landscape) and tablets. The entry area is laid out in three rows: "Problem Description Area - Input Control Area - Supplementary Instructions Area". The height of the text input box is set according to the expected input length. For example, "Symptom Duration" is set as a single-line input (maxlength=20), and "Detailed Symptom Description" is set as a multi-line text box (rows=4, maxlength=500). Voice input is also supported (long press the microphone button to speak, and it will be automatically transcribed into text). The option selection buttons are divided into single-choice (circular buttons) and multiple-choice (square checkboxes). For example, "Abdominal Pain Nature" is a single-choice group (□Persistent □Intermittent □Paroxysmal), and "Accompanying Symptoms" is a multiple-choice group (□Fever □Vomiting □Fatigue). The supplementary instructions upload entry is set as a "+Upload Image / Record" icon button. Clicking it will bring up a file selection menu, supporting the upload of JPG / PNG format images (≤10MB) or MP3 format recordings (≤60 seconds). The upload progress is displayed as a percentage in real time, and a thumbnail / audio waveform preview is generated upon completion. All input fields are marked as required (*). If no field is filled in, the submit button will be grayed out and a message will be displayed saying "Please complete the required fields marked in red".

[0112] Step S158: When the parent clicks the "Start Filling" button on the form display interface, the form filling recording program on the interactive terminal is started to record the total start time of the form filling.

[0113] After a parent clicks the "Start Filling" button (blue gradient background, white text, located at the top center of the form homepage), the front-end JavaScript triggers the `startRecording()` function. It retrieves the current timestamp (accurate to milliseconds) using `Date.now()`, stores it as the total start time in `localStorage`, and simultaneously sends a WebSocket message to the back-end, recording "User ID - Form ID - Start Time" in the form entry log table. The form entry recording program uses a dual-thread design: the main thread handles UI interactions (input / selection / upload), while the secondary thread uses `requestAnimationFrame()` to listen for page switching, input box focus / focus events, etc., ensuring that time recording is not affected by UI blocking. If the parent closes the page midway, the previous entry progress is automatically resumed upon reopening, and the total start time remains based on the timestamp of the first click of "Start Filling," avoiding duplicate timing.

[0114] Step S159: For each item, when a parent enters the filling area for that item, record the start time of filling that item. When the parent clicks the next or skip button, record the end time of filling that item. Calculate the difference between the end time and the start time of filling, and use it as the filling time for that item.

[0115] Each entry has an independent entry timer. When a parent enters the entry's visible area by clicking or scrolling (i.e., the top of the entry is ≤200px from the top of the viewport), the `entryFocus()` event is triggered, recording the entry start time (accurate to milliseconds). When the parent clicks the "Next" or "Skip" button, the `entryBlur()` event is triggered, recording the entry end time. Entry time = Entry end time - Entry start time. If the entry time is <3 seconds (quickly skipped), the system automatically marks it as "may not have been carefully read"; if the entry time is >300 seconds (5 minutes), it is marked as "difficult to complete". For example, entry S15-03, "What was the child's birth weight?", started at time 1692512345678 and ended at time 1692512346890, with a completion time of 1212 milliseconds (approximately 1.2 seconds), and is marked as "Quick Completion". Entry S15-17, "What questions do you have about the preliminary diagnosis of 'possible neuroblastoma'?", started at time 1692512456789 and ended at time 1692512678901, with a completion time of 222112 milliseconds (approximately 3.7 minutes), and is marked as "Difficult to Complete". All time data is stored in the format "Entry ID - Start Time - End Time - Duration - Marker".

[0116] Step S1510: When a parent modifies the input content in the fill-in area and submits it, record the number of modification operations. After each modification, save the content before and after the modification to form a modification record for that entry.

[0117] Input controls are bound to input events to listen for modification behavior. Text input boxes are debounced (triggered after 300ms of no input), while option buttons are triggered upon clicking. Each modification generates a modification record object: {"Modification ID": uuid, "Item ID": "S15-05", "Content before modification": "38.5℃", "Content after modification": "39.2℃", "Modification time": 1692512345678, "Modification type": "Number adjustment"}. For text input boxes, a Levenshtein distance-based difference comparison algorithm is used to record added, deleted, and modified content, such as "originally wrote 'started hurting yesterday,' later changed to 'started hurting 3 days ago'"; for option buttons, changes in selection status are recorded, such as "from '□ Fever' unselected to selected". The number of modifications is accumulated; multiple undoings / redos of the same content are considered a single modification. All modification records are stored locally in IndexedDB and uploaded to the server upon form submission for analysis of which content parents are uncertain about.

[0118] Step S1511: When a parent clicks the skip button, the skipping behavior of the item is recorded. At the same time, a supplementary explanation window pops up, prompting the parent to select the reason for skipping. The reasons for skipping include information loss, irrelevance of content, and difficulty in understanding expression. Parents are allowed to enter supplementary explanations, and the reasons for skipping and the supplementary explanations are recorded.

[0119] A "Skip" button (gray text, hidden by default, displayed when the mouse hovers / touches) is placed on the right side of each entry. Clicking it triggers the `skipEntry()` function: First, it records "Entry ID - Skip Time - Current Progress" in the behavior log; then, a modal window pops up (semi-transparent black background, white foreground border, centered), titled "Why skip this question?", with three radio buttons in the content area: "A. I don't understand the question (comprehension difficulties)", "B. I don't know how to answer (lack of information)", and "C. This question is irrelevant to my child's situation (content irrelevance)". Each option is followed by a "Please provide additional explanation" text box (optional). After the parent selects a reason, they click the "OK" button to close the window. The system records "Skip Reason Code (A / B / C) - Additional Explanation Text - Window Duration", such as "B - My child has never had surgery, I don't know how to fill this out - 15 seconds". If the parent closes the window without selecting a reason, it is marked as "D. Other Reasons" by default, and the additional explanation is left blank. Skipped entries are displayed as a gray "Skipped" label on the form page, allowing parents to return and fill in the form again.

[0120] Step S1512: When the parent completes all entries or clicks the submit button, collect the form content entered by the parent in each entry area, including text input and option selection results.

[0121] After the parent clicks the "Submit" button (green background, located at the bottom of the last page of the form), the front-end performs form validation: checking whether all required fields are filled (except for skipped items), whether the text input box content meets the format requirements (e.g., a mobile phone number is 11 digits), and whether the uploaded file is complete. After successful validation, all entered content is collected through a FormData object: text input box content is stored as "item ID-value" key-value pairs, such as "height=105cm"; option selection results are stored as "item ID-selected value array", such as "painCharacteristics=['paroxysmal','colic']"; unfilled optional fields are marked as "null", and skipped items are marked as "skipped". The form content is serialized in JSON format and sent to the back-end via an HTTPS POST request. The request header includes "Content-Type:application / json" and "Authorization:Bearertoken". Upon receiving the data, the back-end performs data integrity verification (e.g., required fields are not empty). If the verification passes, it returns "Submission Successful"; otherwise, it returns the specific error location (e.g., "Please enter the child's birth weight").

[0122] Step S1513: Collect additional symptom descriptions and supplementary explanations of the test results uploaded by parents through the supplementary explanation upload portal.

[0123] The supplementary information upload portal is displayed at the bottom of each entry as a "+Supplementary Information" link. Clicking it opens the file upload panel, supporting two types of supplementary information: text supplementation (within 200 characters), such as "When the child is in pain, he / she covers his / her right abdomen with his / her hand"; and multimedia supplementation (images / audio recordings), such as uploading photos of the child's facial expressions when in pain or audio recordings describing the symptoms. Uploaded files are compressed (image resolution ≤ 1920×1080, audio sampling rate 44.1kHz, bitrate 128kbps) and transmitted to the object storage service via a chunked upload algorithm (5MB per chunk), returning a unique resource identifier (URI). Supplementary information is stored in association with the corresponding entry ID. For example, the supplementary information for entry S15-09 "Specific location of abdominal pain" is "Uploaded a video of the child pointing out the location of the pain (URI: / uploads / 20230815 / xxx.mp4)". All supplementary content generates an associated index table when the form is submitted to ensure traceability.

[0124] Step S1514: Record the total end time of form filling, and calculate the difference between the total end time and the total start time as the total form filling time.

[0125] After the parent clicks the "Submit" button and the server returns a successful response, the front-end obtains the total end timestamp using `Date.now()` and calculates the total filling time as: Total End Time - Total Start Time. For example, if the start time is 1692512000000 and the end time is 1692512360000, the total time is 360000 milliseconds (60 minutes). If the parent fills out the form multiple times (saving a draft midway), the total time is the sum of the times for each filling out session. For example, if the first filling takes 20 minutes and the second takes 40 minutes, the total time is 60 minutes. The average filling time is calculated by dividing the total filling time by the number of items in the form (23 items): Total Time / Number of Items. This is used to assess the overall complexity of the form. If the average filling time is greater than 60 seconds per item, a suggestion is made to simplify the wording.

[0126] Step S1515: Link the filling time, number of modifications, skipped behavior, skipped reason, and supplementary explanation for each item with the form content filled in by the parents, supplementary symptom descriptions, and supplementary explanations for the test results, according to the item number, filled information, behavior records, and supplementary content format, and integrate all related data to form a multi-dimensional feedback dataset.

[0127] The multi-dimensional feedback dataset employs a hybrid relational and document-oriented storage structure: a relational database (MySQL) stores structured data, including "Form ID - Item ID - Content Filled In - Filling Time - Number of Modifications - Skip Flag - Total Time"; a document-oriented database (MongoDB) stores unstructured data, including "Item ID - Array of Modification Records - Skip Reason Details - Supplementary Explanation Text - Upload File URI". Both are linked by an index based on the Form ID and Item ID, supporting joint queries. For example, querying item S15-08 yields: "Content Filled In: □ Sweating profusely in pain (selected), □ Don't want to walk (selected); Filling Time: 180 seconds; Number of Modifications: 2 (first selected '□ Vomiting,' then cancelled); Skip Behavior: None; Supplementary Explanation: Uploaded a video of the child in pain (URI:...)". After the dataset is generated, a data quality verification process is triggered to check for anomalies such as empty content without being marked as skipped, or an abnormally high number of modifications (>10). Anomalies are marked and then entered into a manual review queue.

[0128] Step S160: Input the multi-dimensional feedback dataset into the adjusted pediatric tumor big data model, analyze the correlation between form content and examination results, the causes of parental interaction barriers, generate form field optimization schemes and preliminary diagnosis suggestions drafts, adjust the field order and description of the preliminary pre-diagnosis form based on the form field optimization schemes, and supplement diagnosis-related verification items in conjunction with the preliminary diagnosis suggestions drafts to obtain the final intelligent pre-diagnosis form for pediatric tumors that includes preliminary diagnosis directions.

[0129] The multi-dimensional feedback dataset is input into the adjusted pediatric oncology model. The model analyzes the correlation between form content and examination results, such as whether the abdominal pain symptoms filled in by parents are closely related to the location of the mass in the ultrasound report, and the causes of parental interaction barriers, such as multiple revisions of a field due to its complex wording. Based on the analysis results, optimization schemes for form fields are generated, such as adjusting the field order, simplifying the wording, and drafting preliminary diagnostic suggestions, such as recommending further CT examination if neuroblastoma is initially suspected. The preliminary pre-consultation form is adjusted based on the optimized form fields, such as moving frequently revised fields forward and simplifying their wording, and adding diagnostic verification items based on the draft preliminary diagnostic suggestions, such as "whether there is a family history of neuroblastoma," ultimately resulting in an intelligent pre-consultation form for pediatric oncology that includes preliminary diagnostic directions.

[0130] Step S161: Extract the form content filled in by parents, supplementary symptom descriptions, and supplementary explanations of examination results from the multi-dimensional feedback dataset, and compare them with the original examination result sheets and children's past health records in the associated dataset to analyze the correlation between the form content and the examination result sheets.

[0131] The form content filled out by parents was extracted from the multi-dimensional feedback dataset. This included detailed descriptions of abdominal pain, supplementary symptom descriptions such as cold sweats during abdominal pain, and supplementary explanations of examination results such as a hard mass. This content was compared with the original examination results in the associated dataset, such as abdominal ultrasound reports and children's past health records, including previous physical examinations. The correlation between the form content and the examination results was analyzed to determine whether the symptoms reported by parents were related to the abnormal indicators shown in the examination results.

[0132] Step S162: If any symptom entered in the form is not related to the abnormal indicator in the test result sheet, mark the entry as needing to be supplemented with association guidance, and record the direction of the guidance content to be supplemented; if any value entered in the form is inconsistent with the corresponding indicator value in the test result sheet, mark the entry as needing a value verification prompt, and record the verification prompt content to be added.

[0133] During the analysis, if a symptom entered in the form, such as "the child has a cough," is found to be unrelated to an abnormal indicator in the examination results, such as a mass in the right adrenal region, the entry is marked as requiring supplementary guidance, and the guidance content is recorded as "Explain that the cough is unrelated to the current examination results and does not need to be described in detail." If a value entered in the form, such as a parent entering a certain weight for their child, is inconsistent with the weight value recorded in the examination results, the entry is marked as requiring a value verification prompt, and the verification prompt content is recorded as "The weight you entered does not match the weight in the uploaded report. Please verify before entering."

[0134] Step S163: Extract the filling time, number of modifications, skipping behavior and skipping reasons for each item from the multi-dimensional feedback dataset to analyze the causes of parent interaction barriers.

[0135] Extract the time taken to fill out each item from the multi-dimensional feedback dataset, such as the time taken to fill out a certain item, the number of modifications, such as multiple modifications to the content, whether the item was skipped and the reason for skipping, such as "not understanding the meaning of the question". Combine this information to analyze the causes of the interaction barriers encountered by parents when filling out the form.

[0136] Step S164: If the filling time for any item exceeds the preset threshold for filling time and the number of modifications exceeds the preset threshold for modification number, and there is also a comprehension obstacle in the expression of the reason for skipping, record the optimization direction as simplifying the expression structure and splitting long sentences into short sentences; if the reason for skipping any item is a lack of information cognition, record the optimization direction as adding example explanations or reducing the professional level of the inquiry content.

[0137] If the time taken to complete an entry exceeds a preset threshold, and the number of revisions also exceeds the preset threshold, and parents report difficulties understanding the wording of the entry as a reason for skipping it, such as "the sentence is too long and I can't understand it," the optimization direction is to simplify the wording of the entry, breaking long sentences into multiple shorter sentences to make it easier for parents to understand. If the reason for skipping an entry is a lack of information, such as "I don't know what a lymph node is," the optimization direction is to add an example description to the entry, such as "a lymph node is a lump the size of a soybean," or reduce the technicality of the question and ask questions in more colloquial language.

[0138] Step S165: Input the correlation analysis results and the interaction barrier cause analysis results into the adjusted pediatric tumor big data model. Based on the case report form structure, generate parameters and correlation weights with examination results to generate an optimized form field scheme. The optimized form field scheme includes the item numbers to be modified, the modified description content, the added correlation guidance statements or verification prompts, and suggestions for adjusting the item order. At the same time, based on the preliminary diagnosis suggestion, generate a threshold, and combine the form content, supplementary symptom descriptions, supplementary explanations of examination results, and the child's past health records to analyze the comprehensive correlation between symptoms, medical history, and examination results, and generate a draft preliminary diagnosis suggestion. The draft preliminary diagnosis suggestion includes possible tumor types, suggestions for further examinations, preliminary treatment suggestions for current symptoms, and the description includes prompts for further diagnosis by a doctor.

[0139] The results of the correlation analysis, such as which items need additional correlation guidance, and the results of the analysis of the causes of interaction barriers, such as which items need simplified descriptions, are input into the adjusted pediatric tumor model. The model generates parameters based on the case report form structure and the correlation weights of the examination results, generating an optimization plan for the form fields. This plan clearly indicates the item numbers that need modification, such as item 5, the modified description content (e.g., breaking long sentences into shorter ones), the necessary correlation guidance statements or verification prompts (e.g., adding "This symptom is related to the examination result; please describe it in detail"), and suggestions for adjusting the item order (e.g., swapping the order of items 3 and 5). Simultaneously, the model generates a threshold based on preliminary diagnostic suggestions. Combining the form content, supplementary symptom descriptions, supplementary explanations of the examination results, and the child's past health records, it comprehensively analyzes the correlation between symptoms, medical history, and examination results to generate a draft preliminary diagnostic suggestion. This includes possible tumor types, such as neuroblastoma, suggestions for further examinations (e.g., abdominal CT scan, tumor marker testing), preliminary management suggestions for the current symptoms (e.g., avoiding pressing on the abdominal mass), and a prompt in the description stating "The above suggestions require further diagnosis by a doctor."

[0140] Step S166: Verify the rationality of the form field optimization plan and the draft preliminary diagnosis recommendation. Use a plan logic consistency verification tool to compare and correct the form field optimization plan with the core logic of the case report form. Identify and delete absolute statements in the draft diagnosis recommendation through the expression compliance detection module to form the final form field optimization plan and draft preliminary diagnosis recommendation.

[0141] Using a logic consistency verification tool, the optimized form fields are compared with the core logic of the case report form to check for logical inconsistencies, such as whether the adjustment of field order conforms to diagnostic logic. Inconsistencies are corrected. The draft preliminary diagnostic recommendation is checked by the expression compliance detection module to identify and delete absolute statements, such as "definitely neuroblastoma," and changed to "possibly neuroblastoma," thus forming the final optimized form fields and draft preliminary diagnostic recommendation.

[0142] Step S167: Load the preliminary pre-consultation form. According to the item order adjustment suggestions in the form field optimization scheme, rearrange the order of form items. Adjust the items that need to be supplemented with association guidance to the items related to the corresponding examination result sheet indicators, and adjust the items that need numerical verification prompts to the vicinity of the items related to the original examination result sheet.

[0143] Load the preliminary pre-consultation form and rearrange the form items according to the item order adjustment suggestions in the form field optimization plan. Move items requiring supplementary guidance after items related to the corresponding examination result indicators; for example, place symptom items related to ultrasound examination results after ultrasound report indicator items. Move items requiring numerical verification prompts near items related to the original examination result form; for example, place weight numerical verification items near items containing weight examination results.

[0144] Step S168: For items marked as needing supplementary guidance in the optimization plan, replace the description of the item with the modified description in the optimization plan and add the corresponding guidance statement; for items marked as needing numerical verification prompts in the optimization plan, add verification prompts below the entry area of ​​the item; for items marked as exceeding the preset threshold for entry time and modification number, and having comprehension difficulties in the skipping reason description, simplify the description structure according to the optimization direction, break long sentences into short sentences, and remove complex modifiers; for items marked as skipping due to lack of information cognition, add example explanations.

[0145] For items marked as requiring supplementary guidance in the optimization plan, replace their descriptions with the modified descriptions in the optimization plan and add corresponding guidance statements, such as "This symptom is related to the ultrasound report results you uploaded. Please fill in according to the actual situation." For items requiring numerical verification prompts, add verification prompts below their input area, such as "Please verify the corresponding values ​​in your uploaded examination report before filling in." For items where the filling time exceeds the preset threshold, the number of modifications exceeds the preset threshold, and there are difficulties in understanding the description, simplify the description structure according to the optimization direction, break long sentences into short sentences, remove complex modifiers, and make the description more concise and easy to understand. For items skipped due to a lack of information, add examples, such as "For example: Lymph node enlargement manifests as small lumps that can be felt in the neck."

[0146] Step S169: Extract possible tumor type directions and suggestions for further examinations from the draft preliminary diagnostic recommendations. Generate corresponding diagnostic-related verification items for each tumor type direction. Generate corresponding examination guidance items for each suggestion for further examinations. Insert the generated diagnostic-related verification items and examination guidance items into the adjusted form in logical order, after the relevant items on the corresponding symptom or examination result sheet. Use a form logic chain detection tool to perform an overall logical scan of the inserted form and adjust the item order to eliminate logical breakpoints.

[0147] Extract potential tumor types from the draft preliminary diagnostic recommendations, such as neuroblastoma, and suggest further investigations, such as abdominal CT scans and tumor marker testing. Generate corresponding diagnostic verification items for the neuroblastoma tumor type, such as "Does the child have unexplained fever?". Generate examination guidance items for the abdominal CT scan recommendation, such as "Do you agree to undergo an abdominal CT scan to further clarify the mass?" Insert these diagnostic verification items and examination guidance items into the adjusted form in logical order, after the relevant items on the corresponding symptom or examination result sheet, such as inserting the neuroblastoma verification item after the abdominal pain symptom item. Use a form logic chain detection tool to perform an overall logical scan of the inserted form to check for logical breaks. If the items are not smoothly connected, adjust the item order to eliminate logical breaks.

[0148] Step S1610: Integrate all adjusted entries, added guiding statements and verification entries, and use the core field integrity verification tool to check the integrity of the core fields of the case report form, the related items of the inspection result sheet and the preliminary diagnosis direction prompts in the form, and supplement the missing content.

[0149] Integrate all adjusted entries, added guiding statements, and validation entries. Use a core field integrity checker to check whether the core fields of the case report form, such as basic information, medical history, and symptoms, are complete. Check the related items in the result sheet, such as whether the entries corresponding to each test result are complete and whether the preliminary diagnosis direction prompts are included. Supplement any missing content, such as adding the family medical history field.

[0150] Step S1611: Obtain the review results of medical personnel on the integrated form, match and verify the form content against the clinical guidelines for pediatric oncology pre-diagnosis, evaluate the clinical reference value of the diagnostic verification items, make final adjustments based on the review comments, and obtain the final intelligent pre-diagnosis form for pediatric oncology that includes the preliminary diagnostic direction.

[0151] The integrated form is submitted to medical personnel for review. The medical personnel match and verify the form content against the clinical guidelines for pediatric oncology pre-diagnosis, assessing the clinical reference value of diagnostic-related verification items, such as whether a particular verification item is significant for diagnosing neuroblastoma. Based on the medical personnel's review comments, such as "the need to add guidance items related to tumor marker testing," the form is finalized to obtain the final intelligent pre-diagnosis form for pediatric oncology that includes preliminary diagnostic directions.

[0152] In one exemplary embodiment, an intelligent pre-consultation form generation system for pediatric tumors is provided. This system can be a terminal, server, etc., and its internal structure diagram can be as follows: Figure 2As shown, the intelligent pre-consultation form generation system for pediatric oncology includes a processor, memory, input / output interface, communication interface, display unit, and input device. The processor, memory, and input / output interface are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interface. The processor provides computational and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides the environment for the operation of the operating system and computer programs in the non-volatile storage medium. The input / output interface is used for information exchange between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, near-field communication, or other technologies. When the computer program is executed by the processor, it implements a method for generating intelligent pre-consultation forms for pediatric oncology. The display unit of this intelligent pre-consultation form generation system for pediatric oncology is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of this intelligent pre-consultation form generation system for pediatric oncology can be a touch layer covering the display screen, or buttons, a trackball, or a touchpad set on the outer shell of the intelligent pre-consultation form generation system for pediatric oncology, or an external keyboard, touchpad, or mouse, etc.

[0153] It should be noted that, in order to simplify the description of the present invention and thus help to understand one or more embodiments of the invention, multiple features may sometimes be grouped into one embodiment, drawing or description thereof in the foregoing description of the embodiments of the present invention.

Claims

1. A method for generating intelligent pre-consultation forms for childhood tumors, characterized in that, The method includes: Identify the scenario types for pediatric tumor pre-consultation, extract the scenario demand features corresponding to each scenario type, and form a scenario demand feature set. The scenario types include initial pre-consultation scenario, follow-up visit scenario, and special symptom screening scenario. Acquire multi-source input data related to pre-consultation, perform correlation processing on the multi-source input data to obtain a correlation data set, wherein the multi-source input data includes the parent's preliminary symptom description text, the child's past health records, and the examination result sheets uploaded by the parent; The parameter configuration of the pediatric tumor big model is adjusted based on the set of scenario requirements to obtain an adjusted pediatric tumor big model adapted to the scenario. The parameter configuration includes parameters for generating the case report form structure, the correlation weight of examination results, and the threshold for generating preliminary diagnostic suggestions. Inputting the associated data set into the adjusted pediatric tumor model generates a semantic prototype of the consultation item containing standard fields of the case report form. The semantic prototype of the consultation item is associated with the examination result verification items and the preliminary diagnosis direction prompts. Based on samples of parents' expression habits, the semantic prototype of the consultation items was optimized. The fields of the professional case report form were converted into expressions that are easy for parents to understand. Guiding statements related to symptoms and test results were added to form a preliminary pre-consultation form. The preliminary pre-consultation form was pushed to the parent interaction terminal. The form content filled in by parents, supplementary symptom descriptions and supplementary explanations of test results were collected. At the same time, the form filling time, field modification times and skipping behavior of parents were recorded and integrated to form a multi-dimensional feedback dataset. By inputting the multi-dimensional feedback dataset into the adjusted pediatric oncology model, the correlation between form content and examination results, and the causes of parental interaction barriers are analyzed. An optimization plan for form fields and a draft of preliminary diagnostic suggestions are generated. Based on the optimization plan, the field order and content of the preliminary pre-consultation form are adjusted. Combined with the draft of preliminary diagnostic suggestions, diagnostic-related verification items are added to obtain the final intelligent pre-consultation form for pediatric oncology that includes preliminary diagnostic directions.

2. The method for generating intelligent pre-consultation forms for pediatric tumors according to claim 1, characterized in that, The process involves identifying the scenario types for pre-diagnosis consultation of pediatric tumors, extracting scenario requirement features corresponding to each scenario type, and forming a set of scenario requirement features, including: The current pre-consultation scenario type can be determined by obtaining the scenario identifier selected by the parent through the scenario selection entry on the parent interaction terminal, or by identifying the scenario tendency based on the initial input of the parent. If the scenario type is the initial consultation and pre-consultation scenario, extract the demand characteristics of the initial consultation and pre-consultation scenario, including the demand for complete medical history collection, the demand for basic physiological indicators coverage, the demand for preliminary verification of examination results, and the demand for preliminary diagnosis guidance. Extract the core focus content corresponding to each demand characteristic and record it. If the scenario type is a follow-up visit scenario, extract the demand features of the follow-up visit scenario, including the demand for recording changes in symptoms after treatment, the demand for comparing recent test results, the demand for generating follow-up conclusions, and the demand for suggesting the next follow-up time. Extract and record the time range limit corresponding to each demand feature. If the scenario type is a specific symptom screening scenario, extract the demand characteristics of this specific symptom screening scenario, including the demand for detailed description of the target symptom, the demand for examination of the symptom-related sites, the demand for inquiry of specific tumor indicators, and the demand for screening conclusion prompts. Extract and record the tumor type association range corresponding to the target symptom. Provide a textual description of the requirements for each scenario type, and indicate the priority order of the requirements. The descriptions of requirements for different scenario types are organized according to scenario type, requirement characteristics, priority, and core content. Duplicate descriptions of requirements between different scenarios are removed to form a set of scenario requirement characteristics.

3. The method for generating intelligent pre-consultation forms for pediatric tumors according to claim 1, characterized in that, The process involves acquiring multi-source input data related to pre-consultation, performing correlation processing on the multi-source input data to obtain a correlated data set, including: The text input box on the parent interaction terminal collects the preliminary symptom description text entered by parents. The preliminary symptom description text is formatted, preset symbols and repetitive expressions are removed, and the description of the symptom occurrence time, manifestation, duration and factors that relieve or aggravate the symptoms is retained. The system retrieves children's past health records through the history import interface of the interactive terminal, sorts the children's past health records in chronological order, and extracts key health events from each children's past health records. The children's past health records include past medical records, treatment plan records, and past examination reports. The system receives examination result reports uploaded by parents through the file upload portal on the interactive terminal, parses the text content and numerical data of each examination result report, and extracts indicator information related to the diagnosis of childhood tumors. The examination result reports include imaging examination reports, laboratory test reports, and pathology examination reports. The symptom keywords in the initial symptom description text are associated with key health events in the child's past health records to identify the correlation between the current symptoms and past health events, and the correlation results are extracted and recorded. The symptom location in the preliminary symptom description text is associated with the indicator information in the examination results form. The corresponding examination indicators are matched with the symptom location, and the associated indicator names and values ​​are extracted and recorded. The key health events in the child's past health records are correlated with recent indicators in the examination results, the trend of indicator changes after treatment is analyzed, and the trend analysis results are extracted and recorded. The correlation results between symptoms and past records, the correlation results between symptoms and examination indicators, and the trend analysis results between past records and recent indicators are integrated, along with the original preliminary symptom description text, the sorted past health records of children, and the parsed single-indicator data of examination results, to form a correlation data set.

4. The method for generating intelligent pre-consultation forms for pediatric tumors according to claim 1, characterized in that, The parameter configuration of the pediatric tumor model is adjusted based on the set of scenario requirements features to obtain an adapted pediatric tumor model, including: Extract the requirement features corresponding to the current scenario type from the set of scenario requirement features, and determine the direction of adjustment of model parameters for each requirement feature; If the current scenario is an initial consultation scenario, set the complete case report form field generation mode for the form structure generation parameters, and set the form to include basic information field, medical history field, symptom field, examination result field and preliminary diagnosis prompt field. At the same time, adjust the field generation order parameter so that the fields are arranged according to the preset logic. Regarding the weight parameters associated with the examination results, if the scenario requirements include the requirement for preliminary verification of the examination results, the weight values ​​associated with the examination result indicators and the consultation items will be adjusted to values ​​that meet the requirements for preliminary verification of the examination results, so that the generated consultation items will be associated with the uploaded examination result single indicator first. For the generation of preliminary diagnostic suggestions, a threshold parameter is set according to the preliminary diagnostic direction guidance requirements in the scenario. The threshold for generating diagnostic suggestions that meets the preliminary diagnostic direction guidance requirements is set so that preliminary diagnostic direction prompts can be generated after obtaining basic symptoms and preliminary examination results. At the same time, the scope of the diagnostic suggestions is limited to avoid absolute statements. If the current scenario is a follow-up visit, adjust the form structure generation parameters of the case report form to the follow-up-specific mode, set the form to retain the fields of previous treatment records and current symptom changes, add the fields of treatment effect evaluation and next follow-up plan, and adjust the field generation order parameters so that the fields are arranged according to the preset logic; Adjust the weighting parameters for the correlation of examination results, adjusting the weighting of the comparison between recent and previous examination results to a value that meets the requirements for comparing examination results, so that the generated consultation items include inquiries about the reasons for changes in examination results; adjust the threshold parameter for generating preliminary diagnostic suggestions to a value that meets the requirements for complete data, and only generate follow-up conclusion suggestions after obtaining data comparing post-treatment symptoms with examination results. The adjusted case report form structure generates parameters, examination result association weights, and preliminary diagnostic suggestions are input into the parameter adjustment interface of the pediatric tumor model. Call the model adaptability testing program, input test data that matches the current scenario, and obtain the output test questions and draft diagnostic suggestions; Compare the test output results with the characteristics of the scenario requirements, check whether the fields of the case report form cover the scenario requirements, check whether the association of the result sheet meets the weight settings, and check whether the diagnostic suggestions meet the threshold requirements; If the fit does not meet expectations, readjust the corresponding parameters and test again until the test output fully matches the characteristics of the scenario requirements, thus obtaining an adjusted large model of pediatric tumors adapted to the scenario.

5. The method for generating intelligent pre-diagnosis forms for pediatric tumors according to claim 1, characterized in that, The process of inputting the associated data set into the adjusted pediatric tumor model to generate semantic prototypes of consultation entries containing standard fields from case report forms includes: Extract the set of symptom keywords from the preliminary symptom description text, the sorted key events in the child's past health records, the parsed single-indicator data of the examination results, and the correlation results between the data from the associated dataset, and organize them into input data groups according to data type. The input data set is input into the field mapping layer of the adjusted pediatric tumor large model case report form. Parameters are generated based on the case report form structure. The content in the input data set is matched with the fields of the standard case report form to determine the types of field to be generated in the case report form. For each defined field type in the case report form, a basic expression framework is generated, which includes the field name, prompts for the inquiry content, and input format suggestions. Based on the correlation weight of the inspection results, identify the abnormal items of the single indicator data of the inspection results in the input data group, generate corresponding inspection result verification items for the abnormal indicators, and add them to the basic expression framework of the corresponding case report table fields; Based on the threshold for generating preliminary diagnostic suggestions, the correlation between symptom keywords, children's past health records and single indicators of examination results in the input data group is analyzed. If the correlation results reach the threshold for generating diagnostic suggestions, a preliminary diagnostic direction prompt is generated, and the prompt content is added to the corresponding case report table field according to the preset format. The generated content, which includes the basic expression framework of the case report form fields, the verification items of the examination results, and the preliminary diagnosis direction prompts, is semantically integrated. Expressions that are irrelevant to the current scenario requirements are removed from the semantically integrated content. The integrated content is arranged in the order of the case report form fields, with each field corresponding to a complete expression containing basic inquiries, verification items, and prompts, forming a semantic prototype of the consultation item.

6. The method for generating intelligent pre-consultation forms for pediatric tumors according to claim 1, characterized in that, The description logic of optimizing the semantic prototype of the consultation items based on parents' expression habits involves converting professional case report form fields into expressions that are easy for parents to understand, supplementing them with guiding statements linking symptoms and examination results, and forming a preliminary pre-consultation form, including: Samples of parents' expression habits from different backgrounds were collected and divided into three groups according to the expression scenario: symptom description sample group, terminology comprehension sample group, and examination result feedback sample group. The expression habit samples included texts of parents describing children's symptoms, understanding medical terminology, and providing feedback on examination results. The text descriptions of the symptom description sample group were analyzed to collect commonly used colloquial terms by parents when describing symptoms, and a colloquial terminology comparison table for symptoms was created. The text descriptions of the terminology comprehension sample group were analyzed to extract the names of professional case report form fields that parents found difficult to understand and their corresponding colloquial translations, and a colloquial translation table for case report form fields was created. The text descriptions of the examination result feedback sample group were analyzed to summarize the common expression logic used by parents when providing examination results, and a template for guiding the expression of examination results was created. Extract the professional case report table field names from the semantic prototype of the consultation items, and replace them with common expressions according to the common expressions conversion table of case report table fields; extract the symptom-related expressions from the semantic prototype of the consultation items, and replace the professional symptom terms with common terms according to the common symptom terminology comparison table. In the semantic prototype of the consultation item, add related guiding statements before and after the description of the verification item of the inspection result, referring to the template for guiding description of the inspection result; The optimized description is checked for logical coherence. A field logic order verification tool is used to compare the field order of the simplified description with the standard logic of the case report form. The description conflict detection module is used to identify and correct any conflicts between the guiding statements, verification items, and diagnostic prompts. We obtained information on the level of understanding of the optimized descriptions and the questions raised by parent representatives from different backgrounds. We then adjusted the wording to address the questions and arranged all the adjusted items in the order of the optimized case report form fields. Each item included basic questions in plain language, related guiding statements, verification items for examination results, and preliminary diagnostic directions, forming a preliminary pre-consultation form.

7. The method for generating intelligent pre-consultation forms for pediatric tumors according to claim 1, characterized in that, The preliminary pre-consultation form is pushed to the parent's interactive interface, collecting the form content filled in by the parent, supplementary symptom descriptions, and supplementary explanations of the examination results. Simultaneously, the time spent filling out the form, the number of times fields were modified, and skipping actions are recorded. These data are integrated to form a multi-dimensional feedback dataset, including: The items from the preliminary pre-consultation form are loaded into the form display interface of the parent interaction terminal in the optimized order. Each item has an independent filling area, which includes a text input box, option selection button and supplementary explanation upload entry. When a parent clicks the "Start Filling" button on the form display screen, the form filling recording program on the interactive terminal is activated to record the total start time of the form filling. For each item, when a parent enters the entry area, the start time of entry is recorded. When the parent clicks the next or skip button, the end time of entry is recorded. The difference between the end time and the start time is calculated as the entry time. When parents modify the input content in the fill-in area and submit it, the number of modification operations is recorded. After each modification, the content before modification and the content after modification are saved to form the modification record of that entry. When a parent clicks the skip button, the skipping behavior for that item is recorded. At the same time, a supplementary explanation window pops up, prompting the parent to select a reason for skipping. The reasons for skipping include missing information, irrelevant content, and difficulty in understanding the expression. Parents are allowed to enter supplementary explanations, and the reasons for skipping and the supplementary explanations are recorded. When parents complete all entries or click the submit button, collect the form content entered by parents in each entry area, including text input and option selection results; We collect additional symptom descriptions and supplementary explanations of the test results uploaded by parents through the supplementary explanation upload portal. Record the total end time of form completion, and calculate the difference between the total end time and the total start time as the total form completion time; The data for each entry, including the time spent filling out the form, number of modifications, skipped behaviors, reasons for skipping, and supplementary explanations, are linked with the form content filled out by parents, supplementary symptom descriptions, and supplementary explanations of the test results, according to the entry number, filled-in information, behavior records, and supplementary content format. All related data are integrated to form a multi-dimensional feedback dataset.

8. The method for generating intelligent pre-consultation forms for pediatric tumors according to claim 1, characterized in that, The process involves inputting a multi-dimensional feedback dataset into an adjusted pediatric tumor model, analyzing the correlation between form content and examination results, identifying the causes of parental interaction barriers, and generating optimized form field schemes and draft preliminary diagnostic suggestions, including: The forms filled out by parents, supplementary symptom descriptions, and supplementary explanations of examination results were extracted from the multi-dimensional feedback dataset and compared with the original examination result sheets and children's past health records in the associated dataset to analyze the correlation between the form content and the examination result sheets. If any symptom entered in the form is not related to the abnormal indicators in the test results, mark the entry as needing to be supplemented with relevant guidance, and record the direction of the guidance content to be supplemented. If the value entered in any item in the form is inconsistent with the corresponding indicator value in the inspection result sheet, mark the item as requiring value verification and record the verification prompt content to be added. The filling time, number of modifications, skipping behavior and reasons for skipping of each item were extracted from the multi-dimensional feedback dataset to analyze the causes of parent interaction barriers. If the time taken to fill out any entry exceeds the preset threshold for filling time and the number of modifications exceeds the preset threshold for the number of modifications, and there is also a problem in understanding the expression in the reason for skipping, the optimization direction is to simplify the expression structure and split long sentences into short sentences. If the reason for skipping any item is a lack of information, the optimization direction is to add example explanations or reduce the professional level of the question content; The correlation analysis results and the interaction barrier cause analysis results are input into the adjusted pediatric tumor big data model. Based on the case report form structure, parameters are generated and the correlation weights of the examination results are generated to generate an optimization scheme for the form fields. The optimization scheme for the form fields includes the item numbers to be modified, the modified description content, the association guidance statements or verification prompts to be added, and suggestions for adjusting the item order. Simultaneously, based on the preliminary diagnostic suggestions, a threshold is generated. Combining the form content, supplementary symptom descriptions, supplementary explanations of the examination results, and the child's past health records, the comprehensive correlation between symptoms, medical history, and examination results is analyzed to generate a draft preliminary diagnostic suggestion. The draft preliminary diagnostic suggestion includes possible tumor types, suggestions for further examinations, and preliminary treatment suggestions for the current symptoms. The description also includes prompts indicating that further diagnosis by a doctor is necessary. The rationality of the form field optimization plan and the draft preliminary diagnostic recommendations were verified. A plan logic consistency verification tool was used to compare and correct the form field optimization plan with the core logic of the case report form. The absolute statements in the draft diagnostic recommendations were identified and deleted by the expression compliance detection module, resulting in the final form field optimization plan and draft preliminary diagnostic recommendations.

9. The method for generating intelligent pre-consultation forms for pediatric tumors according to claim 1, characterized in that, The optimization scheme based on form fields adjusts the field order and description of the preliminary pre-consultation form, and supplements diagnostic-related verification items with the draft preliminary diagnostic suggestions, resulting in a final intelligent pre-consultation form for pediatric oncology that includes preliminary diagnostic directions, including: Load the preliminary pre-diagnosis form, and rearrange the order of the form items according to the item order adjustment suggestions in the form field optimization plan. Move the items that need to be supplemented with association guidance to the items related to the corresponding examination result sheet indicators, and move the items that need numerical verification prompts to the vicinity of the items related to the original examination result sheet. For items marked as needing additional guidance in the optimization plan, replace the description of the item with the modified description in the optimization plan and add the corresponding guidance statements; for items marked as needing numerical verification prompts in the optimization plan, add verification prompts below the entry area of ​​the item; for items marked as exceeding the preset threshold for entry time and modification number, and having comprehension difficulties in the reason for skipping, simplify the description structure according to the optimization direction, break long sentences into short sentences, and remove complex modifiers; for items marked as skipping due to lack of information, add example explanations. Extract possible tumor types and recommendations for further examinations from the draft preliminary diagnostic recommendations, and generate corresponding diagnostic-related validation items for each tumor type. For each item requiring further examination, a corresponding examination guidance item is generated. The generated diagnosis-related verification items and examination guidance items are then inserted into the adjusted form in logical order, after the relevant items on the corresponding symptom or examination result sheet. A form logic chain detection tool is used to perform an overall logical scan of the inserted form, and the item order is adjusted to eliminate logical breakpoints. Integrate all adjusted entries, added guiding statements and verification entries, and use a core field integrity verification tool to check the integrity of the core fields of the case report form, the related items of the inspection result sheet and the preliminary diagnosis direction prompts in the form, and supplement the missing content; The process involves obtaining the review results of medical personnel on the integrated form, matching and verifying the form content against the clinical guidelines for pediatric oncology pre-diagnosis, evaluating the clinical reference value of diagnostic-related verification items, and making final adjustments based on the review comments to obtain the final intelligent pre-diagnosis form for pediatric oncology that includes preliminary diagnostic directions.

10. A smart pre-consultation form generation system for childhood tumors, characterized in that, include: processor; A machine-readable storage medium for storing machine-executable instructions of the processor; The processor is configured to execute the intelligent pre-diagnosis form generation method for pediatric tumors as described in any one of claims 1 to 9 by executing the machine-executable instructions.

Images