Computer-assisted abstraction for reporting of quality measures

Abstract

Methods and systems are disclosed for tracking quality measures in abstracted documents. Embodiments include, determining, based on the abstracted content, a quality measure category, obtaining a quality measure definition corresponding to a quality measure included in the determined quality measure category, determining, based on keywords corresponding to the criterion, whether a portion of the abstracted content satisfies a criterion, recording, in association with the criterion, a reference to the portion of the abstracted content that satisfies the criterion; and selectively generating, a report including the query corresponding to the criterion, a query response, and the portion of the abstracted content satisfying the criterion.

Description

FIELD

The present disclosure relates generally to computer-assisted abstraction and coding of information.

BACKGROUND

Medical insurers tie reimbursements of healthcare providers to their reporting of quality measures. Quality measures are metrics that represent the healthcare providers' adherence to clinical guidelines and indicate whether appropriate care has been given to patients. For example, a hospital treating a patient for an acute myocardial infarction (i.e., heart attack) may follow a guideline requiring the patient be given an aspirin at the time the patient is discharged. If the hospital cannot provide evidence to the patient's insurer that this quality measure was satisfied, the insurer can penalize the hospital.

Identifying and reporting quality measures is a time-consuming and expensive requirement that falls outside the primary role of healthcare providers. As a consequence, providers can be inconsistent and inaccurate in reporting quality measures. However, by effectively tracking and reporting quality measures, the healthcare providers may improve their quality of care and increase fees covered by insurers.

Transcription services support healthcare providers by transcribing information and linking predefined medical terms to the transcribed content. For example, a transcription service may encode documents with identifiers of treatments, medications and procedures described in a patient's medical reports. Thus, transcription services are in a position to support the healthcare providers' tracking and reporting of quality measures. However, transcription services often do not have access to all the information required to fully satisfy reporting requirements. For instance, the service provider may lack access to lab reports generated by third-party providers of lab analyses.

SUMMARY

Exemplary embodiments disclosed herein identify and report quality measures criteria that are satisfied in documents, as well as quality measure criteria that are not included in documents. In some embodiments, where documents are processed in concurrence with an encounter, alerts of pending quality measure criteria that have not been satisfied can be provided to the service provider.

Some exemplary embodiments provide methods for tracking quality measures in one or more documents that are respectively associated with content abstracted from the documents. These methods include, determining, based on the abstracted content of a document, a quality measure category; obtaining a quality measure definition, including at least one quality measure criterion and keywords and queries corresponding to the criterion, corresponding to a quality measure in the determined quality measure category; determining, based on the keywords, whether a portion of the abstracted content satisfies the criterion; recording a reference to the portion of the abstracted content that satisfies the criterion in association with the criterion; and selectively generating, based on the quality measure definition, a report including the query corresponding to the criterion, a query response, and the portion of the abstracted content satisfying the criterion.

Other exemplary embodiments provide a system having a processor and a computer-readable information storage medium that stores computer-readable instruction for tracking quality measures in one or more documents. The instructions, when executed by the processor control the system to determine, based on content abstracted from a document, a quality measure category, obtain a quality measure definition corresponding to a quality measure included in the determined quality measure category, said definition including at least one quality measure criterion, keywords corresponding to the criterion and queries corresponding to the criterion, determine, based on the keywords corresponding to the criterion, whether a portion of the abstracted content satisfies the criterion, record, in association with the criterion, a reference to the portion of the abstracted content that satisfies the criterion, and selectively generating, based on the quality measure definition, a report including the query corresponding to the criterion, a query response, and the portion of the abstracted content satisfying the criterion.

In some of the embodiments, the quality measure definition includes a plurality of criteria for the quality measure and respective sets of queries and keywords corresponding to each criterion.

In some of the embodiments, the report includes a plurality of criteria for the quality measure and a respective sets of queries, query responses, and portions of abstracted content satisfying the corresponding criterion.

In some of the embodiments, generating the report includes, determining that no information was located in the documents satisfying one or more criteria, determining that a time frame for performing the one or more criteria has not expired, and generating an alert for the one or more determined criteria that have not expired.

In some of the embodiments, the documents belong to a collection of documents that correspond to a single patient encounter.

In some of the embodiments, the extracted content is extracted based on a lexicon of medical terminology.

In some of the embodiments, the extracted content is extracted using a natural language processing engine.

In some of the embodiments, the extracted content is linked to medical codes.

In some of the embodiments, determining that the portion of the abstracted content satisfies the criterion includes, providing an abstractor an interactive computer-user interface presenting the query, the criterion and the portion of the abstracted content, and receiving a response to the query from the abstractor affirming that the portion satisfies the criterion.

In some of the embodiments, the report indicates criteria in the definition for which no abstracted content satisfying the criteria was included in the collection of documents.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a system block diagram illustrating an exemplary environment in which embodiments of the disclosed system may operate;

FIG. 2 is a functional block diagram illustrating the exemplary system;

FIG. 3A is a flowchart illustrating an exemplary process for generating quality measures reports;

FIG. 3B is a flowchart illustrating an exemplary process for determining quality measures information included in the quality measures report;

FIG. 3C is a flowchart illustrating an exemplary process for presenting a computer-user interface for determining quality measures information;

FIG. 4 illustrates exemplary quality measures information;

FIG. 5 illustrates an exemplary quality measures report; and

FIG. 6 illustrates an exemplary a graphical user interface.

DETAILED DESCRIPTION

FIG. 1 is a system block diagram illustrating an exemplary environment 100 in which embodiments the disclosed system may operate. For the sake of illustration, the figures are discussed in terms of an example system in which patient information is received from a hospital. The disclosed embodiments are not limited to this example and may be applied to other systems and services.

The environment 100 may include a provider 114, a host system 110, an abstractor 118 and an end-user 120. The provider 114 can be any individual or entity that provides document information to the host. For instance, the provider 114 can be an employee of a doctor's office, law firm or an insurance company that provides dictation for transcription. Alternatively, the provider 114 may be a company, a hospital, a law firm, an insurer or any other such entity that generates documents.

The provider 114 may provide document information to the host via a provider device, which can be any device for capturing information from the provider. For instance, the provider device 115 may be a dictation machine, a telephone, a personal computer (e.g., desktop or laptop), a handheld recording device, a smart phone, or personal digital assistant or any other device that allows user to dictate, store and transmit audio/video files and documents for transmission to the host.

The document information is a set of information that, together, comprises a document or a number of documents. The document information may be structured or unstructured information. In some cases, the document information is text written or typed on paper. In other cases, document information is a dictation provided to the host in real-time (e.g., via a telephone or other information network) or from an audiovisual recording. In still other cases, the document information may be provided from an electronic record (e.g. a plain text document, a hypertext markup language document, an extensible markup language document or a word processor document).

The host can be a device or system for receiving, storing and processing document information and documents received from the provider. In addition, the host can extract the content of the received document information and documents 130 and generate reports using their content.

The host is linked with the provider, the abstractor and/or the end user via one or more communication channels (not shown). The communication channels may be wired or wireless connections. In some instances, the communication channels can be a direct link such as an analog, a serial or a parallel interface. In other instances, the communication channels can be a shared, public, private, or peer-to-peer network, encompassing any wide or local area network such as an extranet, an intranet, the Internet, a Local Area Network (LAN), a Wide Area Network (WAN), a virtual private network (VPN), a voice over internet packet network (VoIP), a public switched telephone network (PSTN), an Integrated Services Digital Network (ISDN), or any other form of wired or wireless communication network.

The host can be implemented as one or more computer systems including, for example, a personal computer, a minicomputer, a microprocessor, a server, a workstation, a mainframe or a similar computing platform. The host can include a controller 112 and a data storage device 116. In addition, while not illustrated, the controller 112 can include one or more processors, computer-readable memory (e.g., read-only memory and random access memory), in addition to other components such as a clock, a communication interface, a data bus, an input/output device, a user-input device and a display device.

The computer-readable data storage device 116 may include any hardware, software, firmware or combination thereof that stores and retrieves information, including computer-readable program instructions and data. The data storage device 116 may be, for instance, a semiconductor, magnetic or optical-based information storage/retrieval device (e.g., flash memory, hard disk drive, CD-ROM, flash RAM). Although the data storage device 116 is depicted as a single element, the device may comprise any additional number of storage media. Although the controller 112 and the data storage device 116 are shown as being within the host system 110, this location is merely exemplary. The controller 112 and the data storage device 116 can be physically located inside or outside of the host system 110. For instance, the data storage device 116 can be configured as a network accessible storage device 116 located remotely from controller 112.

The abstractor can be one or more individuals, software systems, computer systems or a combination thereof for reviewing abstracted data for quality measures. For instance, the abstractor can be a nurse, a clinical document specialist, a physician or an insurer. In FIG. 1, only one abstractor is shown for illustrative purposes. However, embodiments may have different numbers of abstractors in a similar arrangement.

The abstractor terminal 119 can be a data processing device such as a remote terminal, personal computer or network workstation for presenting a computer-user interface. The abstractor terminal may include a processor, a data storage device and stored program instructions that control the abstractor terminal to receive and display information for the abstractor. In some embodiments, the abstractor terminal may emulate the function of a terminal and allow concurrent use of local programs and access to a remote terminal host system.

The end user may be one or more individuals, practices, organizations, businesses or government agencies that receive quality measures reports from the host system 110. For instance, the end-user 120 may be a hospital that uses the reports in requesting reimbursement from an insurer or an insurer that reviews a healthcare provider's compliance with clinical guidelines.

Although the elements are shown in FIG. 1 as separate entities, some or all of these elements can be combined or divided into fewer or greater number of elements at one or more locations. The particular division of functions in FIG. 1 is for illustration only, and different elements may perform one or more of the functions disclosed above.

As shown in FIG. 1, the host system 110 may store computer-executable instructions (e.g., software, firmware, applications, programs, modules, code, portions of code, and combinations thereof) and data (e.g., data compilations, databases, data sets) in the data storage device 116 that, when retrieved and executed by controller 112, configure the host system 110 to perform the functions described herein. The computer-executable instructions can be encoded using any suitable computer programming language such as, C++, JAVA and the SCALA. SCALA is a programming language that supports both object-oriented computing and functional programming. The data may be organized in one or more databases allowing the data to be stored, organized, referenced and selectively retrieved. The databases may be organized in a flat-file, relational or multidimensional structure.

As for computer-executable instructions, the data storage device 116 may include a transcription application 122, an abstraction application 124, a reporting application 126 and a job flow engine 128. Although not shown, the data storage device 116 may include other computer-executable instructions that control host system 110 (e.g., a bootloader, an operating system, control modules and hardware drivers). In addition, the data storage device 116 can store job information 129 (e.g., patient data), document information 130 (e.g., document information, transcribed documents and quality measures reports). Furthermore, the data storage device 116 may store extraction information 132 (e.g., rules, lexicons, medical codes, abstractions), linker information 134 (e.g., rules, terms and codes), abstraction information 136 (e.g., quality measures categories, quality measures definitions) and job queue information 138 (e.g., rules, job flow descriptions.)

The transcription application 122, when executed by controller 112, controls the host system 110 to transcribe document information received in transmissions by the host system 110 from the provider. For instance, the transcription application 122 may convert dictation and/or written documents received from the provider device 115 into documents that are computer-searchable and/or computer-editable. In some cases, the transcription application 122 may use voice recognition software to convert aural dictations into text. In other cases, the transcription application 122 may use optical character recognition (OCR) software to convert written documents into text. Alternatively or additionally, the transcription application 122 may allocate the dictations or documents to human transcribers. In some instances, these transcribers verify the transcriptions performed by the transcription application.

The abstraction application 124, when executed by the controller 112, controls the host system 110 to extract content from the documents and evaluate whether the extracted content satisfies quality measures. As shown in FIG. 1, the abstraction application 124 includes an extractor module, a linker module, an abstractor module and an abstractor interface module. Briefly, the extractor module extracts content from documents, the linker module associates predetermined codes with the extracted content, the abstractor module identifies content corresponding to quality measures and the abstractor interface module presents interactive graphical computer-user interfaces for allowing the abstractor to interact with the information provided by the abstraction application 124.

The reporting application, when executed by the controller 112, controls the host system 110 to generate structured reports based on the transcribed, extracted, coded and/or abstracted information produced from the received document information or documents and stored in the data storage device. The reporting application 126 also generates quality measure reports indicating a correspondence, or lack thereof, between quality measures, query responses associated with the quality measures and information in the documents 130 evincing the quality measures. The reports include quality measures information that may be used for end-user validation, billing and reimbursement of the provider. Furthermore, in some embodiments, the reporting application 126 generates alerts indicating that a pending quality measure has not yet been satisfied.

The job flow engine 128, when executed by the controller 112, controls the host system 110 to process documents based on job queue information 138 and job flow rules. The job flow engine 128 may include multiple program modules for handling and scheduling the job flow of information between the applications and modules illustrated, for example, in FIG. 2.

FIG. 2 is a functional block diagram illustrating the exemplary system. The host system 110 receives document information in an electronic form over a communication channel, such as a telephone network. For example, the host system 110 can receive patient information and diagnoses that have been dictated into provider device 115 after the patient's admission to the hospital for an acute myocardial infarction (AMI). Additionally, the host system 110 can receive other documents generated by the hospital during the patient's admission and preliminary diagnosis (e.g., progress notes, procedure lists, lab results, medical histories, physical examination reports and consultation referrals). The documents may be papers (e.g., facsimiles) or computer-readable files (e.g., text, spreadsheets, images, datasets, multimedia, sound and/or video).

The document information and other documents submitted to the host system 110 by the provider 114 may be identified with a single encounter. The documents identified to an encounter are referred to below as a “document collection.” In some instances, the provider 114 assigns an encounter identifier to the document collection. In addition, the host system 110 may assign an encounter identifier to the document collection based on provider information and/or job information 129 received with the contents of the document collection.

As used herein, the term “encounter” means an instance of contact by a particular client or patient with a single provider. The encounter can, for example, include events occurring from the time the hospital begins the admission process for a patient to the time the hospital discharges the patient. In other examples, an encounter may be an online consultation, an office visit or a laboratory test. In some embodiments, sub-encounters can occur within the scope of an encounter that allow quality measures to be tracked in greater detail. For example, during the time the exemplary patient is admitted to the hospital, different groups within the hospital may treat the patient (e.g., emergency, cardiology, radiology, neurology, laboratory). The documents generated by each group may be associated with a sub-encounter identifier.

The host system 110 may receive one or more pieces of the document collection for an encounter from the provider 114 while the encounter is ongoing. In some instances, the received document collection are provided to the host system 110 after an encounter is completed (e.g., after an office visit). In other instances, some of the documents in the collection for an encounter are received while an encounter is ongoing and the received portion is processed concurrently with the encounter (e.g., during the period a patient is admitted to a hospital). Reports produced from the concurrently provided documents may be used to generate reminders and/or alerts for pending treatments allowing action to be taken by the provider 114 while the encounter is ongoing. For example, if clinical guidelines for treating a patient suffering from an acute myocardial infarction require a particular test or medications and the hospital has not yet fulfilled one or more of the requirements according to reports in the document collection, the host system 110 may generate a reminder or alert to the hospital. In response to the reminder or alert, the provider 114 may take action to comply with the guideline or to correct the documents to reflect that the provider 114 has already complied with the guideline. Hence, the disclosed system allows the host system 110, such as a transcription service to assist the provider 114 in satisfying guidelines and documenting their actions. Based on these improvements, the provider 114 may receive more revenue from insurers.

The received document collection may be stored in the document database of the data storage device 116 for processing by the transcription application 122 for conversion into a text document. The transcribed documents may be stored with a corresponding document identifier uniquely identifying each document. In addition, the received documents included in a document collection may be stored with a corresponding encounter identifier relating the documents 130 to the current encounter.

The transcribed document, as well as any other documents in the document collection, are submitted by the host system 110 (e.g., by the job engine) to the abstraction application 124. Within the abstraction application 124, the extractor module 205 analyzes the documents to determine a corresponding document type. The document type is a code that indicates a category of a document based on identifying keywords in the document or identifiers generated during transcription. For example, in the case of a patient's documents, keywords identifying a document type can be “discharge summary,” “history and physical consultation,” “laboratory results,” “admission” and the like. The keyword “discharge summary” would identify the document type as a discharge summary.

Based on extraction information 132 corresponding to the document types, an extractor module 205 extracts specific content. The extractor module 205 may extract information from the document corresponding to rules for the document type using natural language processing (NLP). Natural language processing allows for the extraction of free text data that other programs to automate applications. Any of a variety of natural language processing techniques can be employed to perform the extractions. In some embodiments, a “bag of words” methodology can be used. An example of this methodology that is suitable for use in the disclosed embodiments is described in “Natural Language Processing Framework to Assess Clinical Conditions”, published in the Journal of American Medical Informatics Association, Volume 16, Number 4, July/August 2009, written by Ware et al., the content of which is incorporated by reference herein in its entirety.

For example, the extractor module 205 may extract the term “acute myocardial infarction” based on the text of a discharge summary provided to the host 110 from a hospital. The extractor module 205 may parse the text and identify terms that match a predefined set of terms in the extraction information 132. For instance, extractor module 205 can match one or more of the terms “heart attack,” “myocardial infarction” and “coronary” in the extraction information 132 to text in the discharge summary. The extractor module 205 may also identify related (e.g., modifier) terms in proximity of the matched terms, such as “severe,” “acute” or “massive.” Thus, the phrase term “acute myocardial infarction” may be extracted from the discharge summary based on the phrase, “ . . . heart attack was severe.” The extraction module may store a reference between the identified text in the document and the extracted term. The reference may be stored inside the document (e.g., as a markup language code) or outside the document (e.g., in a data file).

A linker module 210 determines codes corresponding to all or part of the extracted information (i.e., extractions). The linker module 210 determines the corresponding codes from the linker information 134 using natural language processing (NLP). NLP linkers are engines that evaluate the extractions to determine corresponding codes. The codes may include words or phrases that correspond to an industry standard coding system such as, SNOMED (systemized nomenclature of medicine), RxNorm, ICD 9 and LOINC (logical observation identifiers names and codes). The linker module 210 can use any natural language processing technique for coding the extractions such as, regular expression (Regex) pattern matching and context evaluation. For example, the linker information 134 may associate the extracted term “acute myocardial infarction” with the SNOMED code “57054XX5” The extractions and corresponding codes are referred to below as “abstracted content.”

An abstractor module 215, based on the abstractor information, determines whether portions of the abstracted content corresponds to one or more quality measures, locates content and/or codes corresponding to the quality measures and determines whether the portions of abstracted content satisfy criteria for the quality measure. Where the content satisfies a criterion, the abstractor module 215 may store a reference between the portions of the abstracted content and the criterion of the quality measure. Where no content satisfies a criterion of a quality measure, the abstractor module 215 indicates so.

In some embodiments, an abstractor 118 confirms whether the located content satisfies the corresponding criteria. An abstractor interface module 220 provides an interactive human-computer interface presenting an association between a quality measures, one or more queries corresponding to the quality measure and one of more portions of the abstracted content from which answers to the queries may be found. By referencing the corresponding portion of the document collection, the abstractor answers the queries associated with the quality measure to confirm that the corresponding criteria of the quality measure were satisfied.

FIG. 3A is a flow chart illustrating an exemplary process for generating quality measures reports. The host system 110 receives document information and/or documents 130 from the provider 114, via the provider device 115 (e.g., a dictation device). (Step 303) For example, the host system 110 can receive document information from a hospital via a personal computer. The received information may be part of one or more reports, including the patient's admission information, progress notes, procedure lists, lab results, history and physical, discharge summary. After being processed by the transcription application 122, the host system 110 can store the transcribed document information in the data storage device 116 along with a unique document identifier and encounter identifier. Different documents from the same encounter can, thus, be stored together and be related with a particular encounter in the data storage device.

The documents are provided (e.g., by the job flow engine 128) to the abstraction application 124. Within the abstraction application 124, the extractor module 205 extracts content from the document. (Step 306) The extractor module 205 searches each document for keywords identifying the document type. Once the document type is identified, the extractor module 205 selects a rule set from the extraction rules corresponding to the document type for each document. The extraction rules specify the information type(s) to extract from a particular document types. For example, an abstraction rule for a History and Physical may specify the extraction of medical problems. Other examples of information that can be extracted from various types of reports include medications, problems, allergies, procedures, laboratory tests or results, quality measures, and adherence to guidelines.

The extractor module 205 extracts information from the documents 130 based on identified keywords. The NLP extractor can identify the format of the document and its corresponding section title headers, from the document type. Alternatively, during the transcription of a document, tags may have been inserted to identify individual section headers. For example, a History and Physical document may contain a “Past Medical History” section title. The words within that section may be represented as an unordered collection of words, disregarding grammar and even word order, and can be searched for any words relating to medical problems. Any word found relating to a medical problem is extracted.

Some extractors can employ a relatively simple set of filtering rules to identify and retrieve desired information. For example, a “History and Physical” document may include one or more of the keywords “AMI,” “infarction,” and “acute,” which may be keywords denoting acute myocardial infarction (i.e., a medical problem). The extractor module 205 can extract “acute myocardial infarction” as a medical problem from the History and Physical document that contains any of these keywords. Other extractors may be based upon filters that are more complex. For instance, an extractor could verify if a guideline for acute myocardial infarction has been followed. An example of this type of filter is described in greater detail in U.S. patent application Ser. No. 12/265,495, the disclosure of which is incorporated herein in its entirety.

The linker module 210 evaluates the extractions to determine codes for each extracted item of information. (Step 309) The extracted content can be mapped to a code using pattern matching and searching algorithms. Associated concepts (e.g., synonyms of the extracted data and medications associated with specific medical problems) are also searched to find a pattern match.

The linker module 210 determines the context of the matched extractions based on evaluating neighboring words or phrases. For example, the phrase “arthritis” may be determined as a match for the extracted term “osteoarthritis.” The context evaluation determines that the term “rheumatoid” precedes “arthritis.” Since rheumatoid arthritis describes a different disease from osteoarthritis, it may be concluded that the matched phrase “arthritis” is not a match. When the linker module 210 identifies a pattern match and context match, a code associated with the match information may be linked to the extractions. The codes can each be a unique numeric code (e.g., 57054XX5 is an associated medical code for the medical problem congestive heart failure).

For the existing document collection, the abstractor module 215 determines quality measures included in the abstracted content (i.e., the extractions and/or codes) of the document collection. (Step 312) The quality measures categories may correspond to types of conditions, treatments or medications. As shown in FIG. 4, Table 404, conditions may include acute myocardial infarction (AMI, 405), stroke (STK, 406) and venous thromboembolism (VTE, 407). Each condition may have a set of predetermined clinical guidelines that specify measures (e.g., treatments, prescriptions, mitigations) that should be performed. Table 408 in FIG. 4 shows an exemplary set of measures and measure codes corresponding to the AMI category. For example, as illustrated in Table 408, a quality measure corresponding to an AMI-XX2 may be “Thrombolytic within 30 minutes of arrival” (409). As described in detail below, quality measure AMI-XX2 is associated with a quality measure definition including criteria and values required to satisfy the quality measure.

Based on the quality measures determined by the abstractor module 215, the reporting module may determine if any quality measure criterion that is not satisfied is still pending. (Step 315) For instance, in the case where AMI-XX2 may be “Thrombolytic within 30 minutes of arrival” (409) the information provided by the abstractor module 215 may indicate that the document collection related to the encounter includes no evidence satisfying the criteria corresponding to this quality measure. If a criterion has not been satisfied and is still pending, the report module generates a reminder or an alert indicating that the quality measure to perform the pending criteria. (Step 318) The determination of whether a reminder or alert is provided, as well as the manner in which it is provided, may be made based on the amount of time the criterion will remain pending. For the exemplary quality measure, “Thrombolytic within 30 minutes of arrival,” the report module may provide an alert if the three-hour period of time for treating the patient with a thrombolytic has not expired.

The reporting module generates scores for the performance of quality measures and/or the failure to do so. (Step 321) The scores may be generated for a single quality measure or for a combination of quality measures. In addition, the scores may be determined with regard to a single patient during a single encounter, a single patient over a number of encounters, or for a number of patients over a number of encounters.

The reporting module may use any type of scoring system. In some embodiments, the reporting module may add or subtract points based on a provider's determined performance of quality measures. Alternatively, the scores may be based on ratio of the number of criteria met for the measure (numerator) to the number of opportunities to measure the criteria (denominator). For instance, in the case of AMI-XX3 (410) definition shown in Table 412 of FIG. 4, the numerator (413) for the score includes the criteria “Discharge Diagnosis of AMI” and “Aspirin at Discharge.” The denominator (414) for the score includes “Discharge Diagnosis of AMI,” “Aspirin at Discharge,” “Not Age 18 or Less” and “Not 120 or Greater Stay.” If an “Discharge Diagnosis of AMI,” “Aspirin at Discharge” were satisfied aspirin was given to patients in each of ten (10) encounters in which guidelines require an aspirin on the patients' discharge, the score for this measure would be 1.0 (i.e., 10/10). If the aspirin were not given one encounter in which it was required, then the score for this measure would be 0.9 (i.e., 9/10). Notably, as shown for example in Table 412 of FIG. 4, some criteria are exclusions. For instance, the AMI-XX3, “Aspirin on Discharge” may be excluded if the patient age is 18 years or less. Thus, this instance would not be included in the denominator. As such, for the aforementioned ten encounters in which one encounter is excluded, the score for this measure may be 1.0 (i.e., 9/9).

The reporting application 126 generates quality measure reports (505) including information determined by the abstraction application 124 and the reporting application. (Step 324) FIG. 5 illustrates an exemplary quality measures report 505 for a single measure category for single encounter. The report indicates the documents associated with the encounter that has been received by the host system 110 (510). The report also includes the queries associated with each criterion (515) and, if located, the terms that evince the quality measure was satisfied (520). In some cases, the particular term evincing the criteria is highlighted within the document portion (525). Furthermore, as discussed in detail below, the report may also include a response to the query received from an abstractor 118 who has confirmed that the information in the document portion satisfies the corresponding criteria (530).

For example, the report illustrated in FIG. 5 includes a section for the measure AMI-XX3, Aspirin at Discharge. The section includes the queries associated with the measure by the quality measure definition (e.g., “Aspirin at Discharge?” and “Contradictions for Aspirin?), the identifier of the documents in the collection where evidence for the criteria was found, the type of document (e.g., “Discharge Summary”) and portions of the documents 130 including the evidence or indicating a lack thereof.

FIG. 3B illustrates an exemplary process for determining quality measures. The abstractor module 215 determines quality categories in the document collection based on the abstracted content. (Step 330) For instance, Table 404 in FIG. 4 illustrates several exemplary measures categories (e.g., AMI, STK, VTE). The abstractor module 215 may evaluate the abstracted content and codes to determine quality measure categories in the documents. Sets of terms and/or rules for identifying each category may be stored in the abstractor information.

Based on the determined categories, the abstractor module 215 obtains quality measures that correspond to the categories. (Step 336) For instance, Table 408 in FIG. 4 shows an exemplary set of measures corresponding to the AMI category. The measures may be obtained from clinical guidelines provided by a standards-setting body, an insurance organization or a government agency.

For the determined measures, the abstractor module 215 obtains corresponding quality measures definitions. (Step 336) The quality measures tracked by the host system 110 may have a corresponding quality measure definition stored in the data storage device 116 that has criteria and having respective keywords and queries corresponding to each criterion. The criterion are information that must be included or excluded for the quality measure to be satisfied. For instance, the exemplary quality measure AMI-XX3, “Aspirin at Discharge,” (410) illustrated in FIG. 4 includes the criteria “discharge diagnosis of AMI,” (415) “aspirin at discharge” (416) and excludes the criteria “age 18 or less” (417) and “stay greater than 120 days (418).”

Furthermore, each criterion may be associated with one or more values sets of keywords and/or codes (419-421) that are referenced to determine whether the criterion is satisfied. For example, the criterion “aspirin at discharge” (416) may have a corresponding value set (422) including different synonyms for “aspirin” (e.g., asaphin, bayer, salicylic acid, zorprin) that may be used to determine whether the criteria was satisfied. Of course, the value sets are not limited to terms and may also include diagnosis codes (e.g., ICD-9 or SNOMED), barcodes, images, symbols or other types of information that might be found in a document and relate to the criteria.

Additionally, as shown in Table 412, the quality measure definition may divide the value sets (419-421) into categories that correspond to document types. Thus, the abstractor module 215 may search for the terms in the value sets in the corresponding document type in order to increase the efficiency and accuracy of the results. The abstractor module 215 can find the value “discharge date” more quickly and accurately by looking for the term in a document that belongs to the discharge summary type.

The quality measure definition may also include one or more queries corresponding to each quality measure. (423) In a simple example, the query corresponding to the quality measure “Aspirin at Discharge” (416) may be “Was an aspirin prescribed at discharge?” (423) In other cases, as illustrated in Table 412, queries may be associated with each criterion in a quality measure. For example, Table 412 illustrates a definition corresponding to AMI-XX3, “Aspirin at Discharge” and the definition includes criteria having respective keywords and queries corresponding to the individual criterion for the quality measure.

In some embodiments, the extraction information 132 stores definitions templates for each measure used by the host system 110. In the event a measure category is determined for an encounter, copies of the templates are made and stored in association with the encounter. The abstractor module 215 can determine the measure (e.g., information corresponding to criteria and responses to queries) and can fill-in the instance of the definition for that measure.

The abstractor module 215 determines information in the document collection corresponding to the criteria in the quality measures definitions. (Step 339) For example, as shown in Table 412, for the criterion “aspirin at discharge,” (416) the abstractor module 215 may search for documents including the keyword “aspirin” (and any synonyms, such as asaphin, bayer, salicylic acid, zorprin) and the keyword “discharge” (and any variations thereof). The search may be performed by using RegEx functions based on the terms in the definition. In some embodiments, the search for terms in different value sets may be limited to documents having a type associated with the value set. Information that is located for the search terms may be stored as data elements in the corresponding quality measure definition. For example, the data element “bayer,” corresponding to the “aspirin at discharge” is stored in association with its respective criterion.

The abstractor module 215 determines whether the abstracted content identified by the searches satisfies the corresponding criterion in the definition. (Step 342) This determination may be based on one or more of: keywords, the proximity of the keywords, pattern matching, similarity analysis, the type of document and the linked code. For the content that is determined to satisfy the criteria, references to the content can be recorded in association with the criteria of the quality measure definition. (Step 345) The reference creates a pointer to the abstracted content that supports the determination and links to the document from which the terms were extracted.

FIG. 3C is a flowchart illustrating an exemplary process for presenting a computer-user interface for determining quality measures information. The abstractor interface module 220 provides a user interface to the abstractor 118 via the abstractor terminal presenting the correspondence between queries corresponding to the quality measure criteria, the respective search information in the definition and the portions of the documents. (Step 350) For example, FIG. 6 illustrates an exemplary user-interface 606 displayed on the abstractor terminal 119 and presenting queries corresponding to AMI-XX3, “Aspirin at Discharge” measure to an abstractor for a particular encounter.

The user-interface includes questions in the quality measures definition that have been tracked (612) by the abstractor module 215 based on abstracted content located in the collection of documents. The interface also indicates the questions from the definition that were not tracked (618) due to a lack of any evidence in the document. For each question that is tracked, one or more document portions (624) are displayed via the abstractor terminal 119 that included the evidence considered to satisfy the criteria. In this example, the illustrated document portions from a discharge summary indicate that the patient was discharged with a diagnosis of “AMI” and that “aspirin” was prescribed. The host system 110 obtains query responses corresponding to criteria and identified terms in documents. (Step 353) For example, the user interface provides one or more user-input object (e.g., a radio-button or a checkbox) (630) allowing the abstractor 118 to confirm or reject the document portions determined by the abstractor module 215 to satisfy the query. The host system 110 stores the received query responses in association with respective criteria. (Step 356) Query responses that have been rejected by the abstractor 118 are excluded from the report.

By identifying information in documents supporting quality measures, as described in the exemplary embodiments above, the host 110 can greatly improve the provider's 114 ability to quickly and accurately report quality measures to the end-user 120. Furthermore, by alerting the provider 114 to criteria that have not been satisfied during an encounter, the host 110 enables the provider 114 to take remedial action to fulfill quality measures to may not have been performed or documented. In addition, by providing a user interface in which abstractor's respond to questions affirming that a presented portion of a document satisfies the criteria, the time and effort required of the abstractors is simplified and reduced.

As disclosed herein, embodiments and features can be implemented through computer hardware and software. Such embodiments can be implemented in various environments such as networked and computing-based environments with one or more users. The present disclosure, however, is not limited to such examples, and embodiments can be implemented with other platforms and in other environments.

Moreover, while illustrative embodiments have been described herein, further embodiments can include equivalent elements, modifications, omissions, combinations (e.g., of aspects across various embodiments), adaptations and/or alterations as would be appreciated by those in the art based on the present disclosure.

Other embodiments of this disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the embodiments of the embodiments disclosed herein. Further, the steps of the disclosed methods can be modified in various manners, including by reordering steps, executing multiple steps concurrently, and/or inserting or deleting steps, without departing from the principles of the disclosed. It is therefore intended that the specification and embodiments be considered as exemplary only.

Claims

1. A method of tracking quality measures in one or more documents, wherein said documents are respectively associated with content abstracted from the documents, said method comprising: determining that the one or more documents belong to a collection of documents that correspond to a single patient encounter;determining, based on the abstracted content, a quality measure category;obtaining a quality measure definition corresponding to a quality measure included in the determined quality measure category, said definition including at least one quality measure criterion, keywords corresponding to the at least one quality measure criterion and queries corresponding to the at least one quality measure criterion;determining, based on the keywords corresponding to the criterion, whether a portion of the abstracted content satisfies the at least one quality measure criterion;determining that no information was located in the documents satisfying the at least one quality measure criteria;selectively generating, based on the quality measure definition, a report including the query corresponding to the criterion, a query response, and indicates criteria in the definition for which no abstracted content satisfying the criteria was included in the collection of documents;determining that the patient encounter is ongoing; andgenerating an alert to take an action for satisfying the at least one quality measure criteria that has not been satisfied, while the encounter is ongoing.

2. The method of claim 1, wherein the quality measure definition includes criteria for the quality measure and respective sets of queries and keywords corresponding to each criterion.

3. The method of claim 1, wherein generating the alert includes: determining that a time frame for performing the one or more criteria has not expired; andgenerating an alert for the one or more determined criteria that have not expired.

4. The method of claim 1, wherein the content abstracted from the documents comprises content abstracted based on a lexicon of medical terminology.

5. The method of claim 1, wherein the content abstracted from the documents comprises content abstracted using a natural language processing engine.

6. The method of claim 1, wherein the content abstracted from the documents is linked to medical codes.

7. The method of claim 1, wherein determining that the portion of the abstracted content does not satisfy the criterion includes: providing, to an abstractor, an interactive computer-user interface presenting the query, the criterion and the portion of the abstracted content; andreceiving a response to the query from the abstractor, said response affirming that the portion does not satisfy the criterion.

8. A system for tracking quality measures in one or more documents, wherein said documents are respectively associated with content abstracted from the documents, comprising: a processor;a display device;a user-input device; anda non-transient computer-readable information storage device having program instructions recorded therein, said program instructions when executed by the processor controlling the system to: determine that the one or more documents belong to a collection of documents that correspond to a single patient encounter;determine, based on the abstracted content, a quality measure category;obtain a quality measure definition corresponding to a quality measure included in the determined quality measure category, said definition including at least one quality measure criterion, keywords corresponding to the at least one quality measure criterion and queries corresponding to the at least one quality measure criterion;determine, based on the keywords corresponding to the criterion, whether a portion of the abstracted content satisfies the at least one quality measure criterion;determine that no information was located in the documents satisfying the at least one quality measure criteria;selectively generate, based on the quality measure definition, a report including the query corresponding to the criterion, a query response, and indicates criteria in the definition for which no abstracted content satisfying the criteria was included in the collection of documents;determine that the patient encounter is ongoing; andgenerate an alert to take an action for satisfying the at least one quality measure criteria that has not been satisfied, while the encounter is ongoing.

9. The system of claim 8, wherein the quality measure definition includes criteria for the quality measure and respective sets of queries and keywords corresponding to each criterion.

10. The system of claim 8, wherein the program instructions controlling the computer to generate the alert include program instructions that control the computer to: determining that a time frame for performing the one or more criteria has not expired; andgenerating an alert for the one or more determined criteria that have not expired.

11. The system of claim 8, wherein the content abstracted from the documents comprises content abstracted based on a lexicon of medical terminology.

12. The system of claim 8, wherein the content abstracted from the documents comprises content abstracted using a natural language processing engine.

13. The system of claim 8, wherein the content abstracted from the documents is linked to medical codes.

14. The system of claim 8, wherein the program instructions controlling the computer to determine that the portion of the abstracted content does not satisfy the criterion include program instructions that control the computer to: provide, to an abstractor, an interactive computer-user interface presenting the query, the criterion and the portion of the abstracted content; andreceive a response to the query from the abstractor, said response affirming that the portion does not satisfy the criterion.