Methods and apparatus for streamlining sample testing in one or more diagnostic analyzers

Abstract

An efficient load plan generating method for one or more diagnostic analyzers. The method involves collecting work order data for work orders on samples of a customer by the one or more diagnostic analyzers over a selected timeframe, separating the work order data into primary and secondary tests, and applying sequence mining to the separated work order data to find a probability of ordering secondary test(s) if a primary test(s) had been ordered. One or more secondary tests can be selected based on the probability and a load plan for the one or more analyzers can be generated by a work load planner based on the selected secondary test(s). The load plan, test menus, and inventories can include tests currently ordered and secondary test(s) anticipated to be ordered based on the probability that the selected secondary test(s) will be ordered in the future. Numerous other embodiments are provided.

Description

METHODS AND APPARATUS FOR STREAMLINING SAMPLE TESTING IN ONEOR MORE DIAGNOSTIC ANALYZERSCROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. Provisional Patent Application No. 63 / 513,544, entitled "METHODS AND APPARATUS FOR STREAMLINING SAMPLE TESTING IN ONE OR MORE DIAGNOSTIC ANALYZERS" filed July 13, 2023, the disclosure of which is hereby incorporated by reference in its entirety for all purposes .FIELD

[0002] The present application relates to diagnostic testing using one or more diagnostic analyzers, and more particularly to methods and apparatus for generating efficient load plans for such one or more diagnostic analyzers.BACKGROUND

[0003] In-vitro diagnostic laboratories perform tests to quantify various components or analytes contained in samples obtained from the human body. Samples may include whole blood, plasma or serum, urine, saliva, semen, mucus, interstitial fluid, cerebrospinal fluids, or other bodily fluids or eluted tissue samples, or the like. Such diagnostic laboratories can include one or more diagnostic analyzers, some with the same test menus (duplicate menus) and some with different test menus (i.e. , test capabilities) . Handling the sample work load, i.e., the number and type of tests assigned to each diagnostic analyzer is a perpetual challenge, especially in mid-volume and high-volume in-vitro diagnostic laboratories.

[0004] Accordingly, improved methods and apparatus adapted to provide efficient load plans for samples in one or more diagnostic analyzers are desired.SUMMARY

[0005] In some embodiments herein, a load plan generatingmethod for one or more diagnostic analyzers is provided . The method comprises collecting work order data for tests performed on samples by one or more diagnostic analyzers over a f irst timeframe , separating work order data into primary tests and secondary tests for each sample and forming separated work order data , and operating on the separated work order data f rom the one or more diagnostic analyzers to find a probability p± of one or more secondary tests being ordered when one or more primary tests have been ordered over the first timeframe .

[0006] In some embodiments provided herein, a load plan generating apparatus is provided . The load plan generating apparatus comprises one or more diagnostic analyzers that have actually performed work orders for tests on samples over a first timeframe , a load plan server conf igured to receive work order data on the work orders for tests on sample s actually performed over the first timef rame , the load plan server further conf igured to : separate work order data into primary tests and secondary tests for each of the samples and form separated work order data , and apply a sequence mining to the separated work order data to f ind a probability pi of one or more secondary te sts being ordered when one or more primary tests are ordered .

[0007] Other features and a spects of the present invention will become more fully apparent from the following detailed des cription , the appended claims , and the accompanying drawings .BRIEF DESCRIPTION OF THE DRAWING

[0008] The drawings , described below , are for illustrative purposes only, and are not necessarily drawn to scale . The drawings are not intended to limit the s cope of the disclosure in any way .

[0009] FIG . 1 illustrates a schematic view of an exampleload plan generating apparatus in accordance with embodiments provided herein.

[0010] FIG. 2A illustrates a schematic view of portion of a load plan generating apparatus illustrating the data flow in accordance with embodiments provided herein.

[0011] FIG. 2B illustrates a graphical plot of the frequency of secondary tests given that a primary test was ordered in accordance with embodiments provided herein.

[0012] FIG. 3 is a flowchart of an example method of generating the load plan for one or more diagnostic analyzers in accordance with embodiments provided herein.

[0013] FIG. 4 is a flowchart of an example method of using the load plan generating apparatus to streamline sample testing in one or more diagnostic analyzers in accordance with embodiments of the present disclosure.DETAILED DESCRIPTION

[0014] Independent of the grammatical term usage, individuals with male, female or other gender identities are included within the term.

[0015] Up to now, it has been difficult to estimate test orders that may be received over a defined work period. Thus, setting up a work load plan to include a proper test menu and analyzer inventory of testing supplies needed for any one or more diagnostic analyzers and / or diagnostic laboratories over that defined work period has been challenging. Failing to properly pre-populate a test menu with tests as well as with inventories used for such tests can produce inefficiencies because certain work orders for certain tests that may not be present in the current test menu may be received. Thus, the menu and associated inventories may need to be modified to accommodate such tests, and this is inefficient.

[0016] One way of understanding test work order patterns inaccordance with embodiments of this disclosure is by using past test work order data from one or more existing analyzers of a particular customer or a customer laboratory to guide the menu setup for future test work load planning . Diagnostic analyzer suppliers can have multiple customers ( e . g . , hundreds or even thousands ) and each customer may have one or more diagnostic analyzers resident at their respective diagnostic laboratories . In some embodiments described herein, one or more diagnostic analyzers can be connected to a load plan server, which may be located remotely from the customer' s laboratory in some embodiments , such as at an analyzer manufacturer ' s location or in the cloud . Optionally, the load plan server can be networked and acces sed locally, i . e . , in a customer ' s laboratory or facility .

[0017] Customer te st work order data may be collected and analyzed over a selected timef rame in the pa st in order to better understand customer test ordering patterns and offer future solutions for the customer' s one or more diagnostic analyzers and / or diagnostic laboratories . The future solutions provided by embodiments herein may help guide inventory ordering and / or more proactively set future test menus for the one or more analyzers . Such work load data inputs from the one or more customer diagnostic analyzers or laboratories may be proces sed by a reflex module , which may be part of the load plan server to arrive at a number of ref lex tests that are statistically likely to be run over a future test period . A ref lex test is a secondary test that is statistically probable to be requested given a current test order has been received by an analyzer .

[0018] These and other embodiments of the present dis closure are de scribed with reference to FIGs . 1-4 herein .

[0019] As is shown in FIGs . 1-2 , and according to embodiments herein , the reflex module 118 is a software moduleof a load plan generating apparatus 100 that uses the work order data 104 collected from one or more diagnostic analyzers 102A-102N over a selected timeframe 118T (e.g., a first timeframe of 2 weeks to 2 months of past work order data 104) , where N is a finite integer (e.g. , 5 or 45) . This work order data 104 is processed by the reflex module 118 of a load plan server 108 to provide a statistical compilation of the analyzer' s (e.g. , analyzer 1 of LAB 1) test orders that has occurred in the past (over the selected timeframe 118T) . The selected timeframe 118T (e.g. , first timeframe) can be selected by an operator or set automatically such as by being coded into the program of the reflex module 118, for example. In some embodiments, the selecting of the first timeframe can be for 2 months or less. In other embodiments, the selecting the first timeframe to be from 2 weeks to 6 weeks. Other suitable timeframes may be used.

[0020] Work order data 104 is data on completed tests that have been ordered over the selected timeframe 118T ( (e.g. , first timeframe) . Work order data 104 may include non-stat tests, but may include stat tests and / or any special tests, but need not include them if they are too sporadic.

[0021] Work order data 104 can include sample information, such as sample ID 104ID, test timestamp 104TS (time of test order) , and customer identification information, test type 104TT ordered for each sample, and / or analyzer ID 104A. For example, a sample ID 104ID may be AB 1254974, where AB can be a customer identifier assigned from AA to ZZ, and the number 1254974 may be a sample number from 1 to 10, 000,000, for example. Other suitable means for sample ID 104ID including costumer ID may be used. Test type 104TT may be an abbreviation for the test ordered over the selected timeframe 118T (e.g., first timeframe) , such as GLU for glucose test, ALB for albumin test, and the like. Analyzer ID 102A caninclude any suitable identifier enabling identification of each diagnostic analyzer 102A-102N. Analyzer ID 102A can include whether a particular analyzer is running a clinical chemistry (CC) test or an immunoassay (IA) test. For example, analyzer ID 102A can be All standing for analyzer 1 running an IA test on a sample, or A1C standing for analyzer 1 of a customer running an CC test on a sample, for example.

[0022] The stored work order data 104 can be communicated between the one or more diagnostic analyzers 102A-102N and the load plan server 108 through any suitable type of network 106. The work order data 104 may be stored in a memory 112 of the load plan server 108, for example. In one embodiment, the network 106 can be a local area network (LAN) , wherein the load plan server 108 may be local to a particular customer laboratory (i.e. , LAB 1) . Thus, work order data 104 may be received by the load plan server 108 from a laboratory information system (LIS) coupled with the particular laboratory (e.g. , LAB 1) and one or more diagnostic analyzers 102A-102N associated therewith, a hospital information system (HIS) , or even from workstation ( s ) of the one or more diagnostic analyzers (e.g. , 102A-102B) .

[0023] Optionally, in another embodiment, the load plan server 108 may be a remote server, such as resident at an analyzer manufacturer's location or otherwise provided in the cloud storage. In this remote server case, the network 106 may be a WAN, VPN, or other network communicating over the internet or a wide area network. It should be recognized that load plan server 108 may include work order data 104 from multiple customers and thus the storage thereof may be partitioned by customer if remotely located. In either the local or remote cases, work order data 104 may be communicated via any suitable type of communication interface 110.

[0024] Memory 112 may be any suitable type of memory, suchas, but not limited to, one or more of a volatile memory, a non-volatile memory, or combinations thereof. Volatile memory may include, but is not limited to, a static random access memory (SRAM) , or a dynamic random access memory (DRAM) . Nonvolatile memory may include, but is not limited to, an electrically programmable read-only memory (EPROM) , an electrically erasable programmable read-only memory (EEPROM) , a flash memory, etc. Memory 112 may have a plurality of instructions stored therein that, when executed by the processor 114, perform various actions specified by one or more of the stored plurality of instructions, including the reflex module 118 and work load planner 119.

[0025] The load plan server 108 may be used to generate and provide load plans 126 for the one or more diagnostic analyzers 102A-102N or laboratories (e.g., LAB 1 - LAB N) . Load plans 126 are plans for set up of the test menus and thus the associated inventories for one or more diagnostic analyzers 102A-102N. The load plans 126 are based both on current work order data 130 received through the communication interface 110 as well as selected reflex tests based on the analysis taking place in the reflex module 118 and selections made thereof. Thus, the load plans 120 for the analyzers(e.g., diagnostic analyzers 102A-102N) facilitate readiness for not only actual orders received as primary tests, but for one or more secondary tests that are statistically probable to be ordered based on the analysis of the past work order data 104 and selected either manually or automatically.

[0026] According to the embodiments of the method, the collected work order data 104 over the selected timeframe 118T (e.g., first timeframe) can separated. For example, the reflex module 118 can separate the work orders into primary tests performed on the customer samples and secondary tests performed on the customer samples. The collected work orderdata 104 is separated by diagnostic analyzers (102A-102N) as well, if more than one analyzer is providing data. As used herein, secondary tests are defined as one or more tests ordered within a selected PS time limit 118L after the order of a primary test on a sample of a particular customer.

[0027] In more detail, and according to embodiments, a reflex sequence (RS) is carried out by the reflex module 118 and may be defined as follows:(RS) = [{Ptest(l, AnalyzerTypeC) , Ptest (2, AnalyzerTypel) }, {Stest(l, AnalyzerTypeC) , Stest (2, AnalyzerTypel ) . . . } ] where Ptest denotes the primary tests and Stest denotes the secondary tests, and AnalyzerTypeC and AnalyzerTypel are the analyzer type used for each ordered test of an diagnostic analyzers 102A-102N. Analyzer type can include a designation as either a CC test (Type C) or an IA test (Type I) , or for example. Other test types may be used, as well. For example, depending on the type of tests ordered for the customer sample, there could be one or more Ptests, and one or more Stests associated with the one or more Ptests.

[0028] As discussed above, the work order data 104 including primary tests and secondary tests are first separated by a separating function 118P, wherein primary tests are separated from secondary tests in memory 112. Next the separated work order data over the selected timeframe 118T (e.g., first timeframe) is operated on by a sequence mining function 118M of the reflex module 118. By applying the sequence mining function 118M on the separated work order data within the reflex module 118, a frequent reflex-sequence for a given diagnostic analyzer 102A-102N and / or diagnostic laboratory (e.g. , LAB 1) can be obtained over the selected timeframe 118T (e.g. , first timeframe) .

[0029] A PrefixSpan algorithm may be used for theseparating function 118P as well as the sequence mining function 118M of the reflex module 118. The PrefixSpan algorithm is an open-source sequential pattern mining technique and operates on a sequence database to explore prefix projection in sequential pattern mining. The PrefixSpan algorithm operates to mine a complete set of patterns, but greatly reduces the efforts of candidate subsequence generation. Moreover, prefix-pro ection substantially reduces the size of projected databases and leads to more efficient processing .

[0030] The PrefixSpan algorithm uses a level-wise approach to find all the sequences and their frequencies in the data. The database is passed multiple times to the algorithm and in each iteration; it removes all the non-frequent item sets or subsequences. This removal is based on a threshold set called the support. Only those item sets are kept whose occurrence frequency is greater than the minimum support value. After the first pass, the algorithm finds the frequent item sets of length 1 that are called 1-sequences. All 1-sequences with frequency greater than the minimum support are then input to the next pass. These are the candidate (or prefixes) for 2- sequences. At the end of this pass, 2-sequences and their frequencies are generated. These 2-sequences form the candidates for 3-sequences after considering the minimum support. The algorithm recursively runs until no more frequent item sets are found. To reduce the time for computation, the number of items or pattern length can be set . The output of the algorithm is the frequency of occurrence of the frequent item sets.

[0031] In application to the present method and apparatus, the output of the reflex module 118 is the frequency of occurrence of the primary and secondary test combinations. In other words, the reflex module 118 provides the frequency ofoccurrence of the secondary test(s) given the combination of primary tests. These frequencies are then converted to probabilities by the reflex module 118 by dividing the frequencies by the number of samples with at least one secondary test.

[0032] According to the reflex module 118, the minimum support value 118S and PS time limit 118L, in the primary and secondary test combination can be set. The minimum support value 118S and the PS time limit 118L could be set by the customer or operator, but a value of support of 1% and PS time limit of 5 are good as default. A minimum support value 118S of 1% indicates the combination is present in at least 1% of the total samples. The output of the reflex module 118 is a frequency of occurrence of the secondary test(s) given the one or more primary tests (usually a combination of primary tests) . These frequencies are then converted to probabilities by dividing the frequency by the number of samples with at least one secondary test.

[0033] An example frequent reflex sequence generated by the reflex module 118, which is based on the operation of the separating function 118P and sequence mining function 118M is shown graphically in FIG. 2B. FIG. 2B illustrates graphically, in a first column 235, various combinations of primary tests culled from the work order data 104 over the selected timeframe 118T (e.g. , first timeframe) by the separating function 118P. A second column 240 is a listing of the frequent secondary tests determined after operation of the sequence mining function 118M of the reflex module 118. By way of example, operation of the reflex module 118 has determined that four tests are secondary tests that have met a minimum frequency when operating on the separated work order data, namely secondary tests; ALP 2c, CO2 c, NA, and K tests. The third column 245 lists the frequency of occurrence of thesecondary tests for the given combinations of primary tests from the first column 235. The first collection of connector lines 250 show the various primary test and secondary test combinations as determined by the sequence mining function 118M that meet the minimum frequency.

[0034] According to the operation of the sequence mining function 118M, let {RS (l,pi) , RS ( 2 , p2) . . . RS ( i , p2) } , wherein p2through p2represents a reflex probability that a particular reflex sequence RS will occur within the separated work order data over the selected timeframe 118T. That is, given that one or more primary tests were ordered in the work order data 104, Pi is the probability of also ordering one or more secondary tests. For example, if a first primary test is ordered in the work order data 104, then there is a finite probability (or possibly zero) that secondary test will be ordered for that customer sample after the selected PS time limit 118L. Thus, there may be a finite probability for each primary / secondary test combination found in the work order data 104 over the selected timeframe 118T.

[0035] In some embodiments, to limit the selectable options to a reasonable number, the sequence mining function 118M may only report frequencies (or probabilities) for secondary tests that meet a minimum support value 118S. Minimum support value 118S may be preset by an operator, or optionally automatically set via coding into the sequence mining function 118M. For example, the minimum support value 118S to be considered may be set at 1%, thus any secondary occurrences that occur in combination with less than 1% of a primary test combination is not considered or reported, as it is not considered frequent. Once the reflex sequence (RS) is generated by the sequence mining function 118M of the reflex module 118, one or more of the secondary tests with sufficient probabilities p2through p2can be selected. The second connector lines 255 connect to thethird column 245 to indicate the frequency of the secondary tests in combination with the various primary tests of the first column 235.

[0036] In some embodiments, the one or more secondary tests (e.g., from second column 240) having sufficient probabilities Pi can be manually selected, such as by an operator. They may be selected by the use of a populated drop-down menu, through clicking on a box or other indicator in a display of the probabilities for the secondary tests, or by inputting, in any manner, a minimum probability threshold PT 118MP. In other embodiments, the one or more secondary tests having sufficient probabilities pi can be selected automatically, such as selecting, within the sequence mining function 118M or example, selection one or more secondary tests that have probability p± above a minimum probability threshold PT 118MP that can be preset and either manually input or otherwise coded into the software.

[0037] The selected one or more secondary tests (referred to herein as "selected reflex tests 128") may be provided to the work load planner 119. Work load planner 119 is a computer program that operates to define a work load plan 126 for the one or more analyzers 102A-102N. In particular, a work load plan 126 including the test menu(s) and test inventory ( ies ) for the one or more analyzers 102A-102N or one or more diagnostic laboratories (e.g. , LAB 1 - LAB N) over a selected work period 129 can be sent out and communicated by the work load planner 119 to the one or more analyzers 102A-102N or one or more diagnostic laboratories LAB 1 - LAB N. The selected work period 129 is a period of time over which the current work order data 130 received is to be considered.

[0038] The current work order data 130 can, in some embodiments, come directly from a hospital information system (HIS) of a customer. The HIS can be a computer system wheremedical professionals enter current test orders and receive test results thereof. In other embodiments, the current work order data 130 can optionally come from a laboratory information system (LIS) of a customer. The LIS is a computer system that decides where (which diagnostic analyzers 102A- 102N) each test included in the current work order data 130 of a particular customer will be routed to, i.e., to what diagnostic analyzers 102A-102N, based on a load plan 126. However, the selected work period 129 can generally be less time than the selected timeframe 118T over which the past work order data 104 is considered and analyzed. For example, the selected work period 129 may be a shift or part of a day, a day, a week, or several weeks, or possibly longer if the overall test composition doesn't change appreciably.

[0039] Therefore, if one or more primary tests are actually ordered in a selected work period 129 from current work order data 130, then one or more selected reflex test (s) 128 selected as the output from the reflex module 118 can be included in the load plan 126 for the one or more diagnostic analyzers 102A-102N in addition to the primary test orders actually received over the selected work period 129 from the LIS or directly from the HIS.

[0040] Below is a detailed description of one embodiment of a method to obtain the reflex sequence (RS) . The method can comprise I through III:I. A new column can be computed (let' s call it test category) from the initial test timestamp 104TS of a sample from a customer, wherein test_category is a determination that the test is a primary test or a secondary test. In this example, the selected PS time limit 118L (demarcation between primary tests and secondary tests) may be set to be one hour, which means any test ordered up to and including the selected PS timelimit 118L of one hour from the initial test timestamp104TS for the sample is considered a primary test and anything after the selected PS time limit 118L of one hour is considered a secondary test. The selected PS time limit 118L can be manually set by an operator or automatically set, such as by programming in separating function 118P of the reflex module 118.II. Based on the test timestamp 104TS for each ordered test for a sample of a customer, work order data 104 over the selected timeframe 118T is then grouped by sample_ID (104ID) , Analyzer_ID (104A) , test type (104TT) , and test category. This gives the reflex sequence RS of primary and secondary tests being ordered. For example, a reflex sequence RS of [ ["Crea_2", "GGT"] , ["CC>2_C"] ] shows that ["Crea_ 2", "GGT"] is the primary test and ["CO2_ c"] is a secondary test for a particular sample ID 104ID and analyzer ID 104A. The Analyzer ID 104A the analyzer type of the ordered test, can be CC or IA.III. As should be understood, the set of work order data 104 over the selected timeframe 118T is huge (e.g. , when selected timeframe 118T for work order data 104 selected for consideration is 1 month, for example) . Thus, a simple search of reflex sequences is not feasible due to the existence of so many primary test and secondary test combinations. Thus, for the search, the method uses the PrefixSpan algorithm that operates to find all the frequent subsequences, i.e., the primary and secondary test subsequences whose occurrence frequency in the set of sequences is no less than a minimum support value 118MS. Minimum support value 118S is the minimum proportion required to be considered a frequent sequential pattern.

[0041] In application, the database comprising the work order data 104 is passed multiple times to the sequence mining function 118M of the reflex module 118 and in each iteration, it removes all the non-frequent item sets (i.e. , combinations of primary and secondary tests) . This removal is based on the threshold set by the minimum support value 118S. Only those item sets are kept whose occurrence frequency is greater than the minimum support value 118S.

[0042] After the first pass, the method finds the frequent item sets of length 1 (the 1-sequences) . All 1-sequences with frequency greater than the minimum support value 118S are then input to the next pass. These are the candidates (or prefixes) for 2-sequences. At the end of this pass, 2-sequences and their frequencies are generated. These 2-sequences form the candidates for 3-sequences after considering the minimum support value 118S.

[0043] The algorithm recursively runs until no more frequent item sets are found. To reduce the time for computation, the number of items, or pattern length 118PL, in the primary and secondary test combination can be set. Values for minimum support value 118S and pattern length 118PL may manually set by the operator or customer, automatically set, or otherwise included / coded in the method. Experimental analysis has found that a minimum support value 118S of 1% and pattern length 118PL of 5 are good, as defaults. A minimum support value 118S of 1% indicates the combination must be present in at least 1% of the total samples from the work order data 104 over the selected timeframe 118T.

[0044] The output of the sequence mining function 118M is the frequency of occurrence of the primary and secondary test combination or in other words it provides the frequency of occurrence of the secondary test(s) given the combination of primary tests as shown in FIG. 2B, for example. Thesefrequencies are then converted to probabilities by the sequence mining function 118M by dividing it with the number of samples from the work order data 104 over the selected timeframe 118T with at least one secondary test. Following this, the selected reflex tests 128 can be forwarded to be considered by the work load planner 119. Thus, the reflex module 118 aids in providing reflex tests to the work load planner 119 that are highly probable of being requested in the future .

[0045] Again referring now to FIG. 1, the load plan generating apparatus 100 is schematically shown. Work order data 104 available concerning tests ordered by a one or more customers over a selected timeframe 118T are obtained from any one or a combination of the one or more diagnostic analyzers 102A-102N, such as from the work stations thereof, from the LIS or HIS of each customer, wherein HIS-l-N and LIS 1-N indicates HIS and LIS from 1 to N customers. The work order data 104 for each customer if handled by one load plan server 108 may be segregated by customer and securely stored in memory 112. This work order data 104 is from a selected timeframe 118T that has passed. The work order data 104 are operated on by the load plan server 108, as discussed above, in order to better understand test ordering patterns by the customers for one or more analyzers (e.g., 102A-102N) or even for one or more diagnostic laboratories (e.g., Lab 1 - Lab N) . Any number of diagnostic analyzers 102A-102N and / or laboratories (e.g., LAB 1 - LAB N) may be providing work order data 104 for each customer. Work order data 104 may include, as discussed above, data such as sample identification (sample ID 104ID) , test type 104TT, test timestamp 104TS, and / or analyzer ID 104A.

[0046] The test type can be from any of a number of test categories of clinical chemistry tests (CC tests) . CC testscan be tests related to: diabetes, drugs of abuse or toxicity, general chemistry, immunosuppressant drugs, proteins, specimen validity, therapeutic drug monitoring, and the like.Similarly, test type can be from any number of test categories of immunoassays ( IA tests) . For example, the IA tests can be related to: anemia, autoimmune, bone metabolism, cardiac, diabetes, growth, hepatitis, HIV, immunosuppressant drugs, inflammation, metabolic, neurology, oncology, reproductive endocrinology, sepsis, therapeutic drug, thyroid, ToRCH, and the like. Test type may name specific tests from one or more of these categories of CCs or IAS for each sample ID for each customer .

[0047] For example, sample ID DV 1284913 from customer identifier DV may have work order data 104 for CC tests for hemoglobin Ale, oxycodone, albumin, carbon dioxide, prealbumin, creatinine, and phenobarbital, as well as work order data 104 on IA tests for BNP, HBc total, IgE total, PCT, Total T3, and total T4. A full example list of tests types can be undertaken by the ATELLICA® immunoassay & clinical chemical analyzers available from Siemens Healthineers or other CC and / or IA analyzers. A diagnostic analyzer 102A-102N can include over 200 test types, for example, but typically would be set up for (i.e., contain tests type menus and corresponding inventories of reagents or other materials) for 10-50 test types.

[0048] Work order data 104 over the selected timeframe 118T will be categorized as primary tests (i.e. , if within the selected PS time limit 118L of the ordering of the 1sttest on a sample) , and other ones of the test work order data 104 may be categorized as secondary tests (i.e. , conducted outside of the selected PS time limit 118L) . Thus, work order data 104 over the selected timeframe 118T can be separated by the reflex module 118 to form separated work order data, such asdiscussed above. Based on the separated work order data, the load plan generating apparatus 100 can use the sequence mining function 118M of the reflex module 118 to provide options for selected reflex tests 128 that may be made.

[0049] As should now be apparent, selected reflex tests 128 as used herein means secondary tests that, although not yet currently ordered (i.e., not yet in queue in current work order data 130) within the selected work period 129, can be anticipated to be ordered (after a wait period 132) based on a high probability determined by the reflex module and selected (e.g., selected reflex test (s) 128) .

[0050] Therefore, load plans 126 for the one or more analyzers (e.g., 102A-102N) , such as a plurality of analyzers in a customer's diagnostic laboratory (e.g. , LAB 1 - LAB N) , may be developed based on their current work order data 130 over the selected work period 129 as well as historical ordering patterns derived from work order data 104 by the reflex module 118 over the selected timeframe 118T.

[0051] Thus, the load plan generating apparatus 100 comprises one or more diagnostic analyzers (102A-102N) that have performed past work orders for tests on samples of a customer over the selected timeframe 118T (e.g. , first timeframe) . Load plan generating apparatus 100 further comprises a reflex module 118 configured to receive work order data 104 on the work orders for tests on samples ordered and actually performed over the selected timeframe 118T (e.g. , first timeframe) for the customer.

[0052] The reflex module 118 is further configured to: separate work order data into primary tests (Ptests) and secondary tests (Stests) for each of the samples and thus form separated work order data. The amount of data processed can be limited through using minimum support values 118S as discussed herein .

[0053] Reflex module 118 is further configured to apply a sequence mining function 118M to the frequency data to find a probability p± of one or more secondary tests being ordered, given that a primary test has been ordered. Through manual or automated selection inputs to the reflex module 118, the selected reflex tests 128 (selected automatically or by an operator) may be forwarded to the work load planner 119 to be considered when developing the load plans 126 for the one or more diagnostic analyzers 102A-102N.

[0054] FIG. 3 is a flowchart of a load plan generating method 300 for one or more diagnostic analyzers (102A-102N) . The load plan generating method 300 comprises, in block 302 , collecting work order data (e.g., work order data 104) for tests performed on samples by one or more diagnostic analyzers (102A-102N) of a customer over a selected timeframe 118T (e.g., first timeframe) . The collecting may be performed initially by the diagnostic analyzers (102A-102N) , the LIS, and / or the HIS, but ultimately conveyed to and collected in memory 112 of the load plan server 108.

[0055] The load plan generating method 300 comprises, in block 304, separating work order data into primary tests(e.g., Ptests) and secondary tests (e.g. , Stests) for each sample and forming separated work order data.

[0056] The load plan generating method 300 further comprises, in block 306, operating on the separated work order data from the one or more diagnostic analyzers to find a probability p± of one or more secondary tests (Stests) being ordered when one or more primary tests (Ptests) have been ordered over the first timeframe (e.g. , selected (first) timeframe 118T) . The operation is conducted by the sequence mining function (e.g. , sequence mining function 118M) . The sequence mining function includes determining a frequency for the one or more secondary tests (Stest(s) ) that are associatedwith a particular primary test (Ptest) of a sample of a customer .

[0057] The load plan generating method 300 further comprises, in block 308, selecting one or more secondary tests (Stest(s) ) based on the probability pi .

[0058] In one example, an operator of the load plan server 108 can manually select one or more of the secondary tests (Stest(s) ) of interest to be included in the load plan 126. The selection can be by any suitable mechanism, such as a drop down menu, selection from a graphic, or table the like. In some embodiments, the selecting of one or more secondary tests (Stest(s) ) involves selecting only those secondary tests (Stest(s) ) that have a probability p± above a preset minimum probability threshold PT. This selection may be accomplished by an operator setting the minimum probability threshold PT or the minimum probability threshold PT may be coded into the programming of the reflex module 118.

[0059] Based on the selection, the load plan generating method 300 may comprise, in block 310, configuring the work load planner 119 to include the one or more secondary tests (Stest(s) ) that have been selected. The configuring involves including the selected one or more secondary tests in the load plan 126 for one or more diagnostic analyzers 102A-102N of a customer. The load plan generating method 300 may update the work order data 104 for a new timeframe from time to time. The new timeframe may abut the selected timeframe 118T (e.g. , (first or subsequent) or may overlap with the selected timeframe 118T. For example, the re-collecting and updating of work order data 104 for a selected second timeframe has a start time after a start time of the selected timeframe 118T (e.g., first timeframe) .

[0060] To further illustrate how the selections are used, the following is provided. The work load planner 119 can beconfigured to include the one or more secondary tests selected when there has been an associated primary test ordered over the selected work period 129 of current work order data 130. Then, if an order for the secondary test is actually received after the selected work period 129, the test menu and associated inventory items will already have been set up on the diagnostic analyzers 102A-102N. As was mentioned above, and as shown in block 312, the work order data 104 can be updated for a new timeframe from time to time, so that seasonal changes or other changes in the composition of the primary and secondary testing can be accommodated, for example .

[0061] With reference to FIG. 4, a method of using the selected one or more secondary tests is shown. According to block 412, the method 400 further comprises aliquoting one or more aliquoted samples as placeholders for conducting selected one or more secondary tests provided one or more primary tests have been ordered during a selected work period. Aliquoting a sample, as used herein, means dispensing the sample into a sample reservoir (e.g., into a cuvette or other holder) . In other words, even though one or more secondary tests have not been ordered in the selected work period 129, if one or more primary tests have been ordered and the one or more secondary tests have been selected (based on being selected and of high probability) , then the sample is aliquoted during the selected work period 129.

[0062] Then, in block 414, the method 4400 comprises performing one or more secondary tests on the one or more aliquoted samples if an actual order (s) for the one or more secondary tests have been received selected work period 129. In block 416, the method 400 can comprise discarding the one or more aliquoted samples if no actual orders are received for the one or more secondary tests. Thus, if after a wait period132 , if no orders for the one or more secondary tests are received , then the aliquoted sample s can be discarded . Thus , it should be recognized that the one or more analyzers are ready to perform secondary tests on the aliquoted samples given the very high probability that an order for the one or more selected secondary tests will be eventually received .

[0063] The foregoing description dis closes only example embodiments of the invention; modif ications of the above dis closed apparatus and methods which fall within the scope of the disclosure will be readily apparent to those of ordinary skill in the art . Accordingly, while the pre sent invention has been dis closed in connection with the example embodiments thereof , it should be understood that other embodiments may fall within the s cope of the disclosure , as defined by the claims .NON-LIMITING ILLUSTRATIVE EMBODIMENTS

[0064] Th e following i s a list of non-limiting illustrative embodiments disclosed herein :

[0065] Illustrative embodiment 1 . A load plan generating method for one or more diagnostic analyzers , comprising : collecting work order data for test s performed on samples by one or more diagnostic analyzers over a first timeframe ; separating work order data into primary test s and secondary tests for each sample and forming separated work order data ; and operating on the separated work order data from the one or more diagnostic analyzers to f ind a probability of one or more secondary te sts being ordered when one or more primary tests have been ordered over the first timeframe .

[0066] Illustrative embodiment 2 . The load plan generating method according to the preceding embodiment , compris ing : selecting one or more secondary tests ba sed on the probability .

[0067] Illustrative embodiment 3 . The load plan generating method according to one of the preceding embodiments , comprising : selecting only those of the one or more secondary tests that have a probability above a preset minimum probability threshold PT .

[0068] Illustrative embodiment 4 . The load plan generating method according to one of the preceding embodiments , comprising : configuring a work load planner to include the one or more secondary tests that have been selected .

[0069] Illustrative embodiment 5 . The load plan generating method according to one of the preceding embodiments , wherein the conf iguring of the work load planner to include the one or more secondary te sts if a primary test has been ordered over a selected work period of current work order data .

[0070] Illustrative embodiment 6 . The load plan generating method according to one of the preceding embodiments , wherein the selecting based on the probability allows a user to manually select one or more secondary te sts of intere st .

[0071] Illustrative embodiment 7 . The load plan generating method according to one of the preceding embodiments , comprising re-collecting work order data for second timeframe with a start time after a start time of the first timeframe .

[0072] Illustrative embodiment 8 . The load plan generating method according to one of the preceding embodiments , comprising selecting the first timeframe to be 2 months or les s .

[0073] Illustrative embodiment 9 . The load plan generating method according to one of the preceding embodiments , comprising selecting the first timeframe to be from 2 weeks to 6 weeks .

[0074] Illustrat ive embodiment 10 . The load plan generating method according to one of the precedingembodiments , comprising aliquoting one or more aliquoted samples as a placeholder for conducting selected one or more secondary te sts provided one or more primary test s have been ordered during a selected work period .

[0075] Illustrative embodiment 11 . The load plan generating method according to one of the preceding embodiments , comprising discarding the one or more aliquoted samples if no actual order is received for the selected one or more secondary te sts .

[0076] Illustrative embodiment 12 . The load plan generating method according to one of the preceding embodiments , wherein the discarding occurs after a wait period following the selected work period .

[0077] Illustrative embodiment 13 . The load plan generating method according to one of the preceding embodiments , wherein if a work order is actually received for the selected one or more secondary tests , the selected one or more secondary te sts are run on the one or more diagnostic analyzers .

[0078] Illustrative embodiment 14 . The load plan generating method according to one of the preceding embodiments , wherein the separating work order data into primary test s and secondary te sts for each sample is based on a PS time limit .

[0079] Illustrative embodiment 15 . The load plan generating method according to one of the preceding embodiments , wherein the PS time limit i s manually set .Illustrative embodiment 16 . A load plan generating apparatus , comprising : one or more diagnostic analyzers that have actually performed work orders for tests on samples over a first timeframe ; a load plan server conf igured to receive work order data on the work orders for tests on sample s actuallyperformed over the first timef rame ; the load plan server further conf igured to : separate work order data into primary tests and secondary tests for each of the samples and form separated work order data ; and apply sequence mining to the separated work order data to f ind a probability of one or more secondary te sts being ordered when one or more primary tests are ordered .

[0080] Illustrative embodiment 17 . The load plan generating apparatus according to the preceding embodiment , wherein the load plan server i s further conf igured to allow a selection of one or more secondary tests based on the probability .

[0081] Illustrative embodiment 18 . The load plan generating apparatus according to one of the preceding embodiments , wherein the load plan server is operable to configure a work load planner based on the selection of the one or more secondary tests .

Claims

CLAIMSWHAT IS CLAIMED IS:

1. A load plan generating method for one or more diagnostic analyzers, comprising: collecting work order data for tests performed on samples by one or more diagnostic analyzers over a first timeframe; separating work order data into primary tests and secondary tests for each sample and forming separated work order data; and operating on the separated work order data from the one or more diagnostic analyzers to find a probability of one or more secondary tests being ordered when one or more primary tests have been ordered over the first timeframe.

2. The load plan generating method of claim 1, comprising: selecting one or more secondary tests based on the probability .

3. The load plan generating method of claim 2, comprising: selecting only those of the one or more secondary tests that have a probability above a preset minimum probability threshold PT.

4. The load plan generating method of claim 2, comprising: configuring a work load planner to include the one or more secondary tests that have been selected.

5. The load plan generating method of claim 4, wherein the configuring of the work load planner to include the one or more secondary tests if a primary test has been ordered over a selected work period of current work order data.

6. The load plan generating method of claim 2, wherein the selecting based on the probability allows a user to manually select one or more secondary tests of interest.

7. The load plan generating method of claim 1, comprising re-collecting work order data for second timeframe with a start time after a start time of the first timeframe.

8. The load plan generating method of claim 1, comprising selecting the first timeframe to be 2 months or less.

9. The load plan generating method of claim 1, comprising selecting the first timeframe to be from 2 weeks to 6 weeks.

10. The load plan generating method of claim 1, comprising aliquoting one or more aliquoted samples as a placeholder for conducting selected one or more secondary tests provided one or more primary tests have been ordered during a selected work period .

11. The load plan generating method of claim 10, comprising discarding the one or more aliquoted samples if no actual order is received for the selected one or more secondary tests .

12. The load plan generating method of claim 11, wherein the discarding occurs after a wait period following the selected work period.

13. The load plan generating method of claim 10, wherein if a work order is actually received for the selected one or more secondary tests, the selected one or more secondary tests are run on the one or more diagnostic analyzers.14 . The load plan generating method of claim 1 , wherein the separating work order data into primary test s and secondary tests for each sample is based on a PS time limit .15 . The load plan generating method of claim 14 , wherein the PS time limit is manually set .16 . A load plan generating apparatus , comprising : one or more diagnostic analyzers that have actually performed work orders for test s on samples over a first timeframe ; a load plan server configured to receive work order data on the work orders for te sts on samples actually performed over the first timeframe ; the load plan server further configured to : separate work order data into primary tests and secondary tests for each of the samples and form separated work order data ; and apply sequence mining to the separated work order data to find a probability of one or more secondary tests being ordered when one or more primary te sts are ordered .17 . The load plan generating apparatus of claim 16 , wherein the load plan server is further configured to allow a selection of one or more secondary tests based on the probability .18 . The load plan generating apparatus of claim 17 , wherein the load plan server is operable to conf igure a work load planner based on the selection of the one or more secondary tests .

Images