NZ743565A - Health management system with multidimensional performance representation - Google Patents

Abstract

With the trend of comprehensive healthcare systems, has come an increased volume of data that is overwhelming and often paralyzing. Prior approaches use extensive sets of reports or big data techniques for identifying patterns. The prior approaches have had limited success in solving operational healthcare problems. A health management system includes a processor, a searchable multi-dimensional data representation of the performance of an entire health care delivery system accessible by the processor, in which the performance of every healthcare provider, including downstream providers, that are delivering services is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of quality outcomes and resource utilization measures wherein the performance matrix has multiple dimensions, and a memory device coupled to the processor and having a program stored thereon for execution by the processor to perform operations. The operations include creating the multi-dimensional data representation to obtain performance measures of a selected healthcare provider and accessing the multi-dimensional data representation to obtain performance measures of the selected healthcare provider. By including analytic points in a unique new multidimensional data structure, querying the data structure is more efficient for a search engine to analyze and derive actual performance results and determination of actions to improve healthcare provider performance. ealthcare problems. A health management system includes a processor, a searchable multi-dimensional data representation of the performance of an entire health care delivery system accessible by the processor, in which the performance of every healthcare provider, including downstream providers, that are delivering services is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of quality outcomes and resource utilization measures wherein the performance matrix has multiple dimensions, and a memory device coupled to the processor and having a program stored thereon for execution by the processor to perform operations. The operations include creating the multi-dimensional data representation to obtain performance measures of a selected healthcare provider and accessing the multi-dimensional data representation to obtain performance measures of the selected healthcare provider. By including analytic points in a unique new multidimensional data structure, querying the data structure is more efficient for a search engine to analyze and derive actual performance results and determination of actions to improve healthcare provider performance.

Description

Health Management System with Multidimensional Performance Representation Related Application This application claims priority to United States Provisional Application serial number ,024 (entitled REAL TIME POPULATION HEALTH MANAGEMENT, filed December 22, 2016) and to United States Provisional Application serial number 62/270,735 (entitled HEALTHCARE SYSTEM PERFORMANCE MATRIX AND SEARCH ENGINE, filed December 22, 2016), both of which are incorporated herein by reference.

Background The implementation of electronic health record systems has increased the volume of data available for healthcare management to the point that it can be overwhelming and often paralyzing.

Attempts to find a solution to healthcare management improvement have tended to go in one oftwo extremes. The first approach is to provide extensive sets of structured comparative s that the user must search through in order to draw any conclusions and to develop an action plan. The second ch is to use “big data” techniques to search through the vast amounts of data to identify patterns and insights.

While the big data approach holds great e, actual examples of real world operational healthcare problems that have been solved by this ch have been very limited. Furthermore, there is a 2O fiindamental difference between identifying a pattern and ultimately finding a solution to the issue identified by the n.

Summary A health management system includes a processor, a searchable multi-dimensional data representation of the performance of an entire health care ry system accessible by the processor, in which the performance of every healthcare provider, including downstream providers, that are delivering services is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of y es and resource ation measures wherein the performance matrix has multiple dimensions including individual health care providers, sites of 3O service, quality outcomes and resource use measures, type of ts, time periods covered, geographic location of provider and patient, and the patient’s payer, and a memory device coupled to the processor and having a program stored thereon for execution by the processor to perform operations. The operations include creating the multi-dimensional data representation to obtain performance measures of a selected care provider and ing the multi-dimensional data representation to obtain performance measures of the selected healthcare provider.

A non-transitory e readable storage device has instructions for execution by a processor of the machine to perform accessing payer data for multiple ers in a health care delivery system, conforming the accessed payer data to a standard format, populating, based on the accessed payer data, a multi—dimensional data representation of the performance of an entire health care ry system accessible by the processor, in which the performance of every healthcare provider, including ream providers, that are delivering services is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of quality outcomes and resource utilization measures wherein the performance matrix has multiple dimensions including individual health care providers, sites of service, quality outcomes and resource use measures, type of ts, time periods covered, geographic on of provider and patient and the patient’s payer, creating the multi-dimensional data representation to obtain performance measures of a ed healthcare er, and accessing the multi-dimensional data representation to obtain performance measures ofthe selected healthcare provider.

A health management system includes a searchable multi-dimensional data representation of the performance of an entire health care delivery system accessible by one or more sors, in which the performance of every healthcare provider, including downstream providers, that are delivering services, is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of y outcomes and resource utilization measures, a memory device d to the processor and having a program stored thereon for execution by the one or more processors to perform operations. The ions include creating the multi-dimensional data representation to obtain performance measures of a selected healthcare er and accessing the multi- dimensional data representation to obtain performance es ofthe selected care provider.

Brief Description ofthe Drawings is a block diagram representation of a system for integrating information from multiple health care delivery systems to provide a data matrix that is searchable Via a search engine according to an example embodiment. is a block perspective representation of a three dimensional version of the performance matrix ing to an example embodiment. is a block schematic flow diagram illustrating population of analytic points in the performance matrix according to an e embodiment. is a block diagram of a health management system that includes a real time tion health management tool according to an example embodiment. is a block diagram of a circuitry adaptable to perform one or more methods and processors with memory according to an example embodiment.

Detailed Description In the following description, nce is made to the accompanying drawings that form a part hereof, and in which is shown by way of illustration specific ments which may be practiced.

These embodiments are described in sufficient detail to enable those skilled in the art to practice the invention, and it is to be understood that other ments may be utilized and that structural, logical and electrical changes may be made t departing from the scope of the present invention. The following description of example embodiments is, therefore, not to be taken in a limited sense, and the scope ofthe present invention is defined by the appended claims.

The functions or algorithms described herein may be implemented in software in one embodiment. The software may consist ofcomputer executable instructions stored on computer readable media or computer readable storage device such as one or more non-transitory memories or other type of hardware based storage devices, either local or networked. Further, such functions correspond to modules, which may be software, hardware, firmware or any combination thereof. Multiple functions may be performed in one or more s as desired, and the embodiments described are merely examples. The software may be executed on a digital signal processor, ASIC, microprocessor, or other type of processor operating on a computer system, such as a personal computer, server or other computer system, turning such er system into a ically programmed e.

[0013] The rapidly accelerating trend toward provider idation and the on of provider based comprehensive health systems and payment reforms focus on payment bundles such as capitation has created the need for effective population health management. Simultaneously, the implementation of electronic health record systems has increased the volume of data ble to the point that it can be overwhelming and often paralyzing. Attempts to find a solution have tended to go in one oftwo es.

The first approach is to provide extensive sets of structured comparative reports that the user must search through in order to draw any sions and to develop an action plan. The second approach is to use “big data” techniques to search through the vast amounts of data to identify patterns and ts. While the big data approach holds great promise, actual examples of real world ional healthcare problems that have been solved by this approach have been very limited. Furthermore, there is a fimdamental difference between identifying a pattern and ultimately g a solution to the issue fied by the pattern. is a block diagram representation of a system 100 for integrating information from multiple health care delivery systems 105 to provide a data matrix 110 that evaluates performance and is searchable via a search engine 115. The health care delivery systems 105 may be coupled via a network 120 to a system 125 for integration and pre-processing of the data from such health care delivery systems 105 into the matrix 110. System 125 may also be a health care delivery system and include health care data which is also integrated into matrix 110.

In one embodiment, the data matrix is implemented as a performance matrix that is a searchable multi-dimensional data representation of the performance of an entire health care delivery system in which the performance of every healthcare provider who is delivering services is distilled down to a clinically credible measure of actual versus expected performance across a comprehensive set of y es (readmission rate, complication rate, etc.) and resource use measures (hospital length of stay, ceutical expenditures, etc.). The performance matrix may have multiple ions including, but not limited to, individual health care providers, quality es and resource use measures, type of patients, time periods covered, and the patient’s payer.

An example representation of a three dimensional version of the performance matrix is shown in a perspective block diagram form in at 200. The entation may be thought of as a database schema rating an overall data base structure comprising multiple analytic points, where each analytic point, also referred to as a cell, incudes actual and expected results of provider performance.

In some embodiments, there may be trillions of such analytic points which are in a form that makes it more efficient for a search engine to analyze and derive actual performance results, as well as show areas of performance that are below expected, and why such performance is adversely affected. Such results allow communication of the performance as well as actions that can be taken to improve performance, such as using a different lab for diagnostics, or a ent post operation discharge care facility.

The performance matrix 200 represents a new approach that allows the cost and quality performance of an entire health delivery system to be simultaneously evaluated. The perfomiance matrix distills key performance data into an integrated data representation that is able allowing the identification of succinct and prioritized information that is clinically credible and at a level of specificity that is actionable and can lead to sustainable behavior changes that lower cost and improve quality.

The perfomiance matrix 200 may be thought of as an integrated data representation that allows the cost and y performance of an entire health delivery system to be simultaneously 2O evaluated across a multitude of performance measures across all sites of service and providers. The mance matrix distills key performance information into a succinct data representation that is searchable allowing for the identification of ation that is at a level of specificity that is actionable and can lead to sustainable behavior changes that lower cost and improve quality.

Matrix 200 includes several dimensions that intersect to form the analytic . A providers dimension 210 es hospitals 212, nursing homes 214, home health care 216, specialists 218, and ians 220. A patients dimension 230 includes procedures 232, disease cohorts 234, episodes 236, and population 238. A performance dimension 240 is broken into a resources portion 242 and outcomes 244. Resources 242 includes length of stay 246, laboratory 248, pharmacy 250, and radiology 252. Outcomes 244 includes ssions 254, complications 256, emergency room visits 258, and mortality 260.

At its most basic level, excess cost is due to either high unit production cost or an excess volume of services. High or inefficient unit tion cost is typically the result of an inability to manage the level of inputs or site of service selection. An excess volume of services is often the result of poor y since more services will generally be needed to treat the problems caused by the poor quality. To facilitate the development of an action plan to address poor performance, the poor mance needs to be attributed to specific e categories and ic providers. The performance matrix 200 provides a means of simultaneously evaluating performance across the entire healthcare delivery system. The performance matrix 200, in one embodiment, is a cross tabular representation ofthe performance of the healthcare delivery system across multiple performance dimensions as previously mentioned, including Providers or sites of service (hospitals, physicians, specialists, g homes, etc.) Efficiency performance measures (unit itures per hospitalization and outpatient visit, per enrollee annual expenditures, expenditures by cost categories such as a laboratory, etc) Quality performance measures (excess complications, excess readmissions, excess emergency room visits, under-utilization of outpatient mental health services, etc) Site of service substitution (Over use of skilled nursing facilities versus home health, over utilization of the ncy versus office based primary care, etc) Expenditure type (total cost of care, dual cost categories such a laboratory, etc.).

Expenditure types are only applicable to expenditure performance measures.

Patient Categories (disease cohort such as patients with diabetes, types of encounters such as patients admitted for an appendectomy, etc) Population ts (total population, disease cohorts, etc.) Time period (month, year) Payer (Medicare, Medicaid, commercial insurance company A, insurance company B, etc.) Geographic location (location of patient, location of site of service, urban/rural, census region, etc) Individual provider (physician, specialist, hospital, etc) Thus, the mance matrix has an evaluation of every provider in the care delivery system on every performance measure for every type of expenditure for every population segment for every time period, across a wide range of attributes such as payer and phic region. For e, the performance matrix includes detailed identification of poor performance such as specifying that the high per patient tion expenditures for a primary care ian were due to the high pharmaceutical use by the specialists to whom the primary care physician is referring diabetic patients.

Implementations of the performance matrix may be very large, with trillions of analytic points. Each analytic point in the performance matrix contains the following y performance information that is ocessed prior to use: Continuous variables (e.g., expenditures): count, actual average, expected average, test of statistical significance, and binary variables (e.g., readmissions): count, actual rate, expected rate, cost of difference between actual and expected, test of statistical cance.

Thus, each analytic point in the performance matrix contains a pre-processed specific measure of performance expressed as a difference n actual and expected along with the financial impact ofthe difference. The ed values are risk ed to account for differences in case mix. The test of significance provides a determination of whether the observed ence between actual and 2016/068253 expect is meaningfiil (as opposed the result of chance variation). Essentially the mance Matrix creates a data entation that distills all aspects of delivery system performance down to manageable units of comparison and does every possible drill down providing the basis for identifying the source of performance problems.

Each measure of performance has a pre-computed expected value for every analytic point in the performance matrix. There are many ways to e an expected value of a performance measure.

One ofthe most common is indirect rate standardization using an exhaustive and mutually exclusive set of risk groups for risk adjustment. Using indirect rate standardization the expected value in the analytic points in the mance matrix is ed based on the following steps:

[0024] For each risk group (g) for each performance measure (m), a target value (T(g,m)) is established based on the actual ical average value in a reference database.

For service provider (p) for measure (m). the expected value (E(p,m)) is the sum of overall risk groups of the product of the number of patients/enrollees in each risk group (N(p,m,g) times the corresponding target value (T(g,m) divided by the total number of patients/enrollees: E(p,m) = sum over g [N(p,m,g)*T(g,m)] / sum over g N(p,m,g) For service provider (p) for measure (m), the difference between the service er’s actual value and the expected value can be either above expected (negative performance) or below expected (positive performance). Once the Performance Matrix is populated, it is searchable allowing the identification of the sources of poor performance and report the s in a meaningful way that 2O empowers interventions that can lower costs and improve quality.

The performance matrix es distilled mance down to a financial measure of the difference between actual and expected spending. The financial measures in the performance matrix are essentially a measure of relative internal ce use (production efficiency focusing on volume of services and unit cost). An example of identification of performance differences generated Via a search of the performance matrix and presented to the health delivery system is as follows: In the enrolled population ofthe health system there are 1,342 patients with CHF (congestive heart failure) who are incurring annual expenditures of $69,752 which is 32 percent higher than would be expected resulting $21.4 million in annual excess itures. 80 percent of the excess expenditures are concentrated in high severity CHF patients who have multiple comorbid diseases. The high severity severity CHF patients have a potentially table hospital admission rate that is 41 percent higher than expected and a potentially preventable ER Visit rate that is 24 percent higher than would be expected.

Although the inpatient hospital expenditures for high severity CHF patients are consistent with expectations the 30 day post-acute care itures for these patients are 38 percent higher than would be ed. 52 percent of the excess post-acute care for high severity CHF patients are the result of a potentially preventable readmission rate (that is 62 t higher than would be expected. 62 percent of the excess post acute care readmission rate is due readmissions from one nursing home (ElderCare) which has a potentially preventable readmission rate that 88 percent higher than would be expected. 78 percent ofthe patients discharged to this nursing home are for patients discharged by physician James Smith and physician Donald Jones both ofwhom have a disproportionate number of their high severity CHF patients being discharge to a nursing home.

The overarching objective of the performance matrix is to provide a data model that allows the fication of succinct and prioritized information that is at a level of specificity that is actionable.

[0035] is a block schematic flow diagram illustrating population of analytic points in the performance matrix generally at 300. Several sites of service are indicated at 310, 315, and 320 coupled by a network 325 to a healthcare delivery system 330. Sites of service 310, 315, and 320 may be downstream providers which each have their own health care databases with ation regarding patients and services provided, as well as performance data, medical records, and other information.

System 330 has longitudinally integrated delivery system data 335 that represents all ation regarding healthcare services ed by healthcare ers covered by system 330. The data 335 may be gathered from multiple different databases for the delivery system, but provides a consistent interface to that data.

At 340, processing is performed on the data to computer performance measures. Enrollee health status is determined at 345. In one embodiment, the enrollee corresponds to a patient receiving services at delivery system 330 and the various network coupled sites of service. At 350, a risk adjusted expected value for each performance measure is computed. The risk adjusted expected value may include external target mance measure values 355, corresponding to the networked ted sites of service 310, 315, and 320.

[0037] A difference between actual and expected value for each performance measure is calculated at 365 and may include sion factors 370 to convert data from the connected sites of service 310, 315, and 320 that may not be stored using the same schema as data 335, which may be a cal form of data. In some ments, both data 335 and data from the connected sites of service may be converted to a cal form.

In one embodiment, the difference between actual and expected value for each mance measure is a representation of the , such as a financial impact for each mance measure. At 375, the impact from 365 is used to populate each analytic point or cell in the performance matrix 200, resulting in a ted performance matrix 380 ready for use.

In one embodiment, longitudinal historical claims data, such as data from one or more insurance companies (payer) for multiple patients and multiple ers is obtained at 335 from one or more systems. The data obtained may be run through a classification system to obtain a consistent representation of the data at 340, 345 and define what each service corresponding to the claims was. One e fication system includes a 3M Patient Classification System. The data may be used to determine the actual performance at 346. The classification data from classification 340, 345 is also used to generate performance norms for y outcomes and resource use at 355. At 360, the actual and ed performance is compared to generate performance differences by subtracting the actual performance measure from the expected performance. The result is used to determine the financial impact of negative quality outcomes at 365, which may involve aggregating data from multiple patients over multiple providers and other dimensions. This information is then used to populate the performance matrix at 380.

In various embodiments, the use ofthe performance matrix may provide for real time population health care management. As the healthcare industry moves towards increasing use of Accountable Care Organizations (ACOs) and the shift to bundled payment (meaning a single payment to cover all aspects of care for a given condition), there is an increased need for tools to actively manage the healthcare of populations of patients across a wider range of settings and contexts. This management extends beyond those times where the patient is an admitted patient or in the provider’s office for a visit to e factors such as but not limited to prescription compliance, tative checkups, tative vaccinations, healthy living activities, and living arrangements such as assisted living centers, etc. Both private and public healthcare payers singly mandate sets of care guidelines and criteria that need to be ed by providers. If they are not followed, providers may not be fully reimbursed for services provided, t care may be adversely affected, and the overall health ofthe t population may be less than optimal.

In many cases, healthcare provider organizations are required to not only manage adherence to such care ines on a per patient level, but also to report their ance at a population level to various payers and government health agencies. Typically, in the industry today this is a time ing process that requires a significant amount of manual effort to complete. Determining whether or not ed care is within appropriate guidelines requires the review of a wide range of data sources including but not limited to the Electronic Health Records, Visit Scheduling information, Lab and Diagnostic reports, Pharmacy data, and even a patient’s own health tracking data. The process of bringing such data sets together for complete review is usually a cumbersome one. Timing of access to data sets, for one thing, can be an issue: not all cases are usually able to be reviewed in time for interventions to correct cases where proper guidelines are not followed as the reviews are often pective to the patient having left the hospital or provider. For the provider organization this can result in costly claims denials or loss of reimbursement, and for the patient it can result in sub-optimal health treatments when, for example, an incorrect site of service is selected, necessary diagnostics are not performed, diagnostics are performed unnecessarily, medications are not filled and used by the t, and so on.

[0042] Many ofthe challenges ated with beginning to manage care in this new way come from data being housed in multiple systems that are not integrated and which span organizational boundaries. A fiill review of patient care from all settings requires knowledge of multiple systems, review of paper documentation, review of visit schedules, development of a longitudinal view of a t and their associated health issues, and then tracking and coordinating that t’s care in accordance with the necessary guidelines across this myriad of systems. is a block m of a health ment system 400 that includes a real time population health management tool 405 to improve an organization’s ability to care for its population of patients while simultaneously ng the manual efforts required to do so and enabling better use of the organization’s resources to focus on the ry ofproper care. The tool in one embodiment is implemented in software for execution on a processor in a local or cloud computing environment.

Tool 405 includes l components, including but not limited to a guideline/rule repository 410, a patient information store 415, natural language processing (NLP) 420, enterprise master person index (EMPI) 425, and ia evaluation logic 430. The tool 405 also has access to a mance matrix 435 and performance matrix search engine 440. The components may execute on the search engine 440, or other local or remote processing resources 445, or a combination thereof.

Guideline/rule repository 410 contains rule sets needed to satisfy a given care protocol, reporting guideline, or compliance standard. These may apply at a particular patient or population level Examples of these include Core Measures, Patient Safety Incidents, Hospital Acquired Conditions/Infections, Preventable Complication or ssion Requirements, Site of Service assignment criteria, criteria in determining patient transportation, t placement, and care criteria for specific disease, condition, or risk cohorts.

[0046] Patient Information Store 415 is a repository that contains the universe of data known about a c patient. It extends beyond just data that is ble in the Electronic Health Record to include information such as scheduled care follow ups, iption refill information, diagnostics ordered, and patient captured data such as glucose monitoring information. The term “Patient Information Store” is a generic term for this collection of data as in y the store may actually be comprised of multiple repositories able to be ed collectively, to assemble the total longitudinal picture of a patient’s health care information. Data elements may be populated via direct interface with structured data from other systems and may be represented in a variety of formats or code sets such as ICD9, ICDlO, SNOMED-CT, LOINC, etc.. Unstructured data in the Patient Information Store may be processed using Natural Language Processing (NLP) to extract clinical facts from text narrative and other unstructured data sources. In one embodiment, the data is aggregated from a variety of care settings, and includes financial data, patient tracked data, and disease specific items. All data elements are represented with unique concept identifiers that are in turn mapped to care guidelines and rules that makes use of particular types of data. The concept identifiers may be combined to construct a longitudinal t problem list and care history, which may be compared to relevant care guidelines for patients based on plan membership, quality reporting guidelines, and other factors.

Natural ge processing (NLP) 420 component is used to extract data, including clinical facts, from semi-structured and unstructured data s. The NLP also maps the clinical facts found in those sources to discrete ts of the data sources needed to evaluate against rules. Also used to facilitate the question/answer process needed to query the longitudinal patient record as updates are made which affect the Coordination of Care document.

Enterprise Master Person Index (EMPI) 425 is used to consolidate data from various systems and sources around a single t record. Includes ability to match patient data from systems using identifiers from systems and other identifying ation such as Date of Birth, Government ID numbers, Insurance Identifiers, etc. Several s provide the ability to match patient data based on multiple, such as 12 or more such pieces of information to provide an assurance that patients are tly identified and their corresponding data is accurate.

[0049] Criteria evaluation logic 430 is used to apply sets of care guidelines and criteria to the data for a particular patient to determine which have been satisfied and which are deficient. ionalizes the Guideline / Rule Repository and the Patient ation Store together to produce data for the system outputs. Compares data for patient being evaluated against es for similar patients (based on available data ts) to offer insights into likely successful care steps. Considers output of tools such as the performance matrix which will inform the evaluation of next care steps for the patient against the current state of the health system’s ability to successfiJlly deliver those steps. Care deficiencies and needed care may be identified and prioritized.

The tool 405 may take a variety of different types of patient health data as input. While the more available data, the more complete the tool’s review and recommendations will be, not all data sources are required for the Tool to provide valuable ck. Types of data that the Tool may make use of include but are not limited to: patient claims data, pharmacy / medication refill data, pre/post hospital care setting data, clinical documents, visit scheduling information, and personal health information d by the patient (e.g. weights, blood pressure, glucose ation, se data).

The tool 405 will initially enable two primary outputs. One output is a Coordination of Care Document 450. As new clinical documents and diagnostic information about a patient becomes available to the tool, the system evaluates the new data against any known care guidelines that apply to the patient based on the t’s existing health conditions. The tool updates any criterion met by the new data and identifies any new deficiencies that may be introduced by the new data. For example, a particular result on one diagnostic test may warrant a next test be conducted, or the completion of one type of follow up or preventative visit will then trigger the next required visit to be determined.

The new data will also be evaluated to determine if it warrants adding the patient to new care guideline groups. Adding the patient to care ine groups may be done automatically by the Tool, either by the Tool itself or by the Tool calling a sub-process in r , or the Tool may flag the record for evaluation by a human reviewer who may add the patient to the new group. This may occur for e if the new incoming data suggests or definitively diagnoses a new disease such as diabetes.

The system will evaluate the known data about the patient against the new care guideline membership as a diabetic, indicate the initial care steps that need to be applied to the patient, and also flag the patient for inclusion in any reporting on the population of diabetic ts.

The Coordination of Care Document 450 is ible by users ofthe system such as providers and Care Managers as needed through a user interface as is commercially available, such as the 360 ass MD user interface provided by 3M Health Information Systems.

Users will also have the ability to request that the system update the record in “real-time” if needed to incorporate newly added data elements and receive immediate feedback on additional care suggestions or necessary steps to take with the patient. This might also occur for example when a patient currently being seen in the Emergency ment needs to be evaluated against criteria for assignment to a particular site of service or against inpatient admission criteria.

The Coordination of Care Document 450 will offer prioritized guidance for necessary care that is informed by ing outcomes of care for patients deemed to be similar to a particular patient based on available data elements within the population. Prioritization will also incorporate feedback from tools such as the 3M Health System mance Matrix, which can assist in prioritizing care options based on current performance of the healthcare delivery system . This guidance may also include querying the al records ofthe population using NLP in addition to structured / coded data — e.g. to generate ad-hoc population information relevant to the current patient based on patient specific characteristics.

In one embodiment, prioritized worklists may be presented for individual patients.

Prioritization may be infomied by outcome data from a population of like ts within populations.

Reports 455 on Extracted Data may also be provided as an output. The system may generate reports on a scheduled basis for es fied by different care guideline groups.

Examples of this would include reporting to national or state quality agencies, compliance with care ols for particular diseases, effectiveness of preventative care measures, rates of ance with prescription medication refills, etc. Automated ing of population care delivered versus care guidelines may be generated.

Care Managers may also see a tized list of patients within their tion in varying states of care that need attention to stay within care guidelines. Examples of this would be: all patients currently admitted within the healthcare system, all patients due for a particular type of follow up visit, call or diagnostic, or patients needing follow up on medication refills. A prioritized worklist may also be generated for an overall population.

Anticipated benefits to users of the system, depending on entation, may include a reduction in manual effort required to do mandated reporting, which would in turn enable cost savings or redeployment of resources to more directly affect patientcare. A r benefit may include an increase of case review for compliance with varying care guidelines from current percentage to 100%. A reduction in denials, reduction in Recovery Audit Contractor (RAC) audits, ion in payment penalties related to: readmissions, hospital acquired ions, patient safety indicators, and lost 2016/068253 reimbursement due to issues such as ect site of service assignment, patients not meeting admission criteria. An Improved ability may be provided to produce prioritized lists of patients at risk for not meeting care guidelines based on specific disease conditions (e.g. diabetes, heart disease) or other criteria.

Yet a further benefit may include an improved y to predict future care needs of population based on a more comprehensive review of tion status. The tool may fiirther provide for integration of population ment into a single workflow within a single system rather than many ate systems. l, a reduction in complexity of care management process may also be provided. is a block schematic diagram of a computer system 500 to implement methods according to example embodiments. All ents need not be used in various embodiments. One example computing device in the form of a computer 500, may include a processing unit 502, memory 503, removable storage 510, and non-removable storage 512. gh the example ing device is illustrated and described as computer 500, the computing device may be in ent forms in different ments. For example, the computing device may instead be a smartphone, a tablet, smartwatch, or other computing device including the same or similar elements as illustrated and described with regard to Devices such as smartphones, tablets, and smartwatches are generally collectively referred to as mobile devices. Further, although the various data storage elements are illustrated as part of the computer 500, the storage may also or alternatively include cloud-based storage ible via a network, such as the et.

Memory 503 may include volatile memory 514 and non—volatile memory 508. Computer 2O 500 may include — or have access to a computing nment that includes — a variety of er- readable media, such as volatile memory 514 and non-volatile memory 508, removable storage 510 and non—removable storage 512. Computer storage includes random access memory (RAM), read only memory (ROM), erasable programmable nly memory (EPROM) & electrically erasable mmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD ROM), Digital Versatile Disks (DVD) or other optical disk storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices capable of storing computer-readable instructions for execution to perform functions described herein.

Computer 500 may include or have access to a computing environment that includes input 506, output 504, and a communication connection 516. Output 504 may include a display device, such as a touchscreen, that also may serve as an input device. The input 506 may include one or more of a touchscreen, touchpad, mouse, keyboard, camera, one or more device-specific buttons, one or more sensors integrated within or coupled via wired or wireless data connections to the computer 500, and other input devices. The computer may operate in a networked environment using a communication connection to connect to one or more remote computers, such as database servers, including cloud based servers and storage. The remote computer may include a personal computer (PC), server, router, network PC, a peer device or other common k node, or the like. The communication connection may include a Local Area Network (LAN), a Wide Area Network (WAN), cellular, WiFi, Bluetooth, or other networks.

Computer-readable instructions stored on a computer-readable e device are executable by the processing unit 502 of the er 500. A hard drive, CD-ROM, and RAM are some es of articles including a non-transitory computer-readable medium such as a storage . The terms computer-readable medium and storage device do not include carrier waves. For example, a computer program 518 may be used to cause processing unit 502 to perform one or more methods or algorithms described herein.

Examples:

[0065] In e 1, a health management system includes a processor, a searchable multi- ional data representation of the performance of an entire health care delivery system accessible by the sor, in which the performance of every healthcare provider, including downstream providers, that are ring services is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of quality outcomes and resource utilization measures wherein the performance matrix has multiple dimensions including individual health care providers, sites of service, quality outcomes and resource use measures, type of patients, time s covered, geographic on of provider and patient, and the patient’s payer, and a memory device coupled to the processor and having a program stored thereon for execution by the processor to perform operations. The operations include creating the multi-dimensional data representation to obtain performance measures of a selected healthcare provider and accessing the multi-dimensional data representation to obtain performance measures of the selected healthcare provider.

Example 2 includes the health management system of example 1 wherein the ally credible measure comprises at least one of ssion rate and complication rate.

Example 3 includes the health management system of any of examples 1-2 wherein the healthcare providers include at least multiple of hospitals, nursing homes, home health care agencies, specialists, and physicians.

Example 4 includes the health management system of any of examples 1-3 wherein the types of patients include at least one of encounters for a procedure, encounters for c or acute disease ment, disease cohorts of patients, episodes of care, and population management.

Example 5 es the health ment system of any of examples 1—4 wherein a performance dimension of the performance matrix is broken into a resources portion and a quality outcomes portion.

Example 6 includes the health management system of example 5 wherein the resource portions includes at least one of length of stay, laboratory, pharmacy, and radiology.

[0071] Example 7 includes the health management system of any of examples 5-6 wherein the outcomes portion includes at least one of ssions, cations, emergency room visits, and mortality.

Example 8 includes the health management system of any of examples 1-7 wherein each analytic point in the performance matrix contains a pre-processed specific measure of performance sed as a ence between actual and ed along with the financial impact of the difference.

Example 9 includes the health management system of example 8 wherein expected values are risk adjusted to account for differences in case mix.

Example 10 includes the health management system of any of examples 8—9 wherein the pre-processed c measure of mance of each analytic point is lculated using indirect rate standardization based on an tive and mutually exclusive set of risk groups for risk adjustmentr Example 11 includes the health ment system of example 10 wherein for each risk group (g) for each performance measure (m), a target value (T(g,m)) is established based on an actual historical average value in a reference database.

Example 12 includes the health management system of example 11 wherein for service provider (p) for measure (m), an expected value (E(p,m)) is the sum of overall risk groups of the product of the number of patients/enrollees in each risk group ,g) times the corresponding target value (T(g,m) divided by the total number of patients/enrollees expressed as: E(p,m) = sum over g [N(p,m,g)*T(g,m)] / sum over g N(p,m,g), and wherein the difference between the service provider’s actual value and the expected value is expressed as above expected (negative performance) or below expected (positive performance).

In example 13, a non-transitory machine readable storage device has ctions for execution by a processor ofthe machine to perform accessing payer data for multiple providers in a health care delivery system, ming the accessed payer data to a standard format, populating, based on the accessed payer data, a multi—dimensional data representation of the mance of an entire health care delivery system accessible by the processor, in which the performance of every healthcare er, including downstream providers, that are delivering services is distilled down to a clinically credible measure of actual versus expected mance at analytic points across a comprehensive set of quality outcomes and resource utilization measures wherein the performance matrix has multiple dimensions ing individual health care providers, sites of service, quality outcomes and resource use measures, type of patients, time periods covered, geographic location of provider and t and the patient’s payer, creating the multi-dimensional data representation to obtain performance measures of a selected healthcare provider, and accessing the multi-dimensional data representation to obtain mance measures of the selected healthcare er.

Example 14 includes the non-transitory machine readable storage device of e 13 wherein the clinically credible e comprises at least one of readmission rate and complication rate.

Example 15 includes the non-transitory machine readable storage device of any of examples 13-14 wherein the healthcare providers include at least multiple of hospitals, nursing homes, home health care agencies, specialists, and physicians.

Example 16 includes the non-transitory machine readable storage device of any of examples 13-15 wherein the types of patients include at least one of encounters for a procedure, encounters for chronic or acute disease management, disease cohorts of patients, episodes of care, and population management. e 17 includes the non-transitory machine readable storage device of any of examples 13-16 wherein a performance dimension of the performance matrix is broken into a resources portion and an quality outcomes portion.

Example 18 includes the non-transitory machine readable storage device of example 17 wherein the resource portions include at least one of length of stay, laboratory, pharmacy, and ogy.

[0083] Example 19 includes the non-transitory machine readable e device of any of examples 17—18 wherein the outcomes portion es at least one of readmissions, complications, emergency room visits, and mortality.

Example 20 includes the non-transitory machine readable storage device of example 13 wherein each ic point in the mance matrix contains a pre-processed specific measure of performance expressed as a difference n actual and expected along with the financial impact of the difference.

Example 21 includes the ansitory machine readable e device of example 20 wherein expected values are risk adjusted to t for differences in case mix. e 22 includes the non-transitory machine readable e device of any of 2O examples 20-21 wherein the pre-processed specific measure of performance of each analytic point is pre- calculated using indirect rate standardization based on an exhaustive and ly exclusive set of risk groups for risk adjustment.

Example 23 includes the non-transitory machine readable storage device of example 22 wherein for each risk group (g) for each performance measure (m), a target value (T(g,m)) is established based on an actual historical average value in a reference database.

Example 24 includes the non—transitory machine le storage device of example 23 wherein for e provider (p) for measure (m), an expected value (E(p,m)) is the sum of l risk groups of the product of the number of patients/enrollees in each risk group (N(p,m,g) times the ponding target value (T(g,m) divided by the total number of patients/enrollees expressed as: E(p,m) = sum over g [N(p,m,g)*T(g,m)] / sum over g N(p,m,g), and wherein the difference between the service er’s actual value and the expected value is expressed as above expected (negative performance) or below expected (positive performance).

In example 25, a health management system includes a searchable multi-dimensional data representation of the mance of an entire health care delivery system accessible by one or more processors, in which the performance of every healthcare provider, including downstream providers, that are delivering services, is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a hensive set of quality outcomes and resource utilization measures, a memory device coupled to the processor and having a program stored thereon for execution by the one or more processors to perform operations. The ions include creating the multi- dimensional data representation to obtain performance es of a selected healthcare provider and accessing the multi-dimensional data representation to obtain performance measures of the selected healthcare provider.

Example 26 includes the health management system of example 25 wherein the clinically le measure comprises at least one of readmission rate and complication rate.

Example 27 includes the health management system of any of examples 25-26 wherein the performance matrix has multiple dimensions including individual health care providers, sites of service, quality outcomes and resource use measures, type of patients, time periods covered, geographic location of provider and patient and the patient’s payer, wherein the healthcare providers include at least multiple of hospitals, nursing homes, home health care agencies, specialists, and physicians.

Example 28 includes the health management system of example 27 wherein the types of ts include at least one of encounters for a procedure, encounters for chronic or acute e management, disease s of patients, episodes of care, and population management. e 29 includes the health ment system of any of examples 27-28 wherein a performance dimension of the performance matrix is broken into a resources portion and an outcomes portion.

Example 30 includes the health management system of example 29 wherein the resource portions include at least one of length of stay, laboratory, pharmacy, and radiology.

Example 31 includes the health management system of any of es 29-30 wherein the outcomes portion es at least one of readmissions, complications, emergency room visits, and mortality.

Example 32 includes the health management system of any of examples 25—31 wherein each analytic point in the performance matrix contains a pre-processed specific measure of mance expressed as a difference between actual and expected along with the financial impact of the ence.

Example 33 includes the health management system of example 32 wherein expected values are risk adjusted to t for differences in case mix.

Example 34 includes the health management system of any of es 32-33 wherein the pre-processed specific measure of performance of each analytic point is pre-calculated using indirect rate standardization based on an exhaustive and mutually exclusive set of risk groups for risk adjustment.

Example 35 includes the health management system of example 34 wherein for each risk group (g) for each performance measure (m), a target value (T(g,m)) is established based on an actual historical average value in a reference database.

[00100] Example 36 includes the health management system of e 35 wherein for service provider (p) for measure (m), an expected value (E(p,m)) is the sum of l risk groups of the t of the number of ts/enrollees in each risk group (N(p,m,g) times the corresponding target value (T(g,m) divided by the total number of ts/enrollees expressed as: E(p,m) = sum over g [N(p,m,g)*T(g,m)] ;’ sum over g N(p,m,g), and wherein the difference between the service provider’s actual value and the expected value is expressed as above expected (negative performance) or below expected (positive performance).

Although a few embodiments have been described in detail above, other modifications are le. For example, the logic flows depicted in the figures do not require the ular order shown, or sequential order, to achieve desirable results. Other steps may be provided, or steps may be eliminated, from the described flows, and other components may be added to, or removed from, the described systems. Other ments may be within the scope ofthe following claims. 2016/068253

Claims (15)

1. A health management system comprising: a processor; a searchable multi-dimensional data representation of the performance of an entire health care delivery system accessible by the processor, in which the performance of every healthcare provider, including ream providers, that are delivering services is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of quality outcomes and ce utilization measures wherein the performance matrix has multiple dimensions including individual health care providers, sites of service, quality outcomes and resource use measures, 10 type of patients, time periods covered, geographic location of provider and patient, and the patient’s payer; a memory device coupled to the processor and having a program stored thereon for execution by the sor to perform ions comprising: creating the multi-dimensional data representation to obtain performance measures of a 15 selected healthcare provider; and accessing the multi-dimensional data representation to obtain performance es of the selected healthcare provider.

2. The health management system of claim 1 wherein the clinically credible measure comprises at 20 least one of readmission rate and complication rate.

3. The health management system of claim 1 wherein the care providers include at least multiple of hospitals, nursing homes, home health care agencies, specialists, and physicians. 25

4. The health management system of claim 1 wherein the types of patients include at least one of encounters for a procedure, encounters for c or acute disease management, disease cohorts of ts, es of care, and population ment.

5. The health management system of claim 1 wherein a performance dimension ofthe performance matrix is broken into a resources n and a quality outcomes portion.

6. The health management system of claim 5 wherein the resource ns includes at least one of length of stay, laboratory, pharmacy, and radiology, and wherein the outcomes portion includes at least one of readmissions, complications, emergency room visits, and mortality.

7. The health management system of claim 1 wherein each analytic point in the performance matrix ns a pre-processed specific measure of performance expressed as a difference between actual and expected along with the financial impact of the difference wherein ed values are risk adjusted to account for differences in case mix, and wherein the pre-processed specific measure of performance of each analytic point is pre-calculated using indirect rate rdization based on an exhaustive and mutually exclusive set of risk groups for risk adjustment.

8. The health management system of claim 7 wherein for each risk group (g) for each performance e (m), a target value (T(g,m)) is established based on an actual historical average value in a reference database, and wherein for service provider (p) for measure (m), an expected value (E(p,m)) is the sum of overall risk groups of the product of the number of patients/enrollees in each risk group 10 (N(p,m,g) times the corresponding target value (T(g,m) divided by the total number of patients/enrollees expressed as: E(p,m) = sum over g [N(p,m,g)*T(g,m)] / sum over g N(p,m,g) and n the difference between the service provider’s actual value and the expected value is sed as above expected (negative performance) or below expected (positive performance).

9. A non-transitory e readable e device having instructions for execution by a processor of the machine to perform: accessing payer data for multiple providers in a health care delivery system, conforming the accessed payer data to a standard format; 20 populating, based on the accessed payer data, a multi-dimensional data representation of the performance of an entire health care ry system accessible by the sor, in which the mance of every healthcare provider, including downstream providers, that are delivering es is distilled down to a clinically credible measure of actual versus expected performance at analytic points across a comprehensive set of quality outcomes and resource utilization measures wherein the 25 mance matrix has multiple dimensions including individual health care providers, sites of service, quality outcomes and resource use measures, type of patients, time periods covered, geographic location of provider and patient and the patient’s payer, creating the multi-dimensional data representation to obtain performance es of a selected healthcare provider; and accessing the multi-dimensional data representation to obtain performance measures of the selected healthcare provider.

10. The non-transitory machine readable storage device of claim 9 wherein the clinically credible e comprises at least one of readmission rate and complication rate, n the healthcare 35 providers include at least multiple of hospitals, nursing homes, home health care es, specialists, and physicians, wherein the types of ts include at least one of encounters for a procedure, encounters for chronic or acute disease management, disease s of patients, episodes of care, and population management, and wherein a performance dimension of the performance matrix is broken into a resources portion and an quality outcomes portion, wherein the resource portions include at least one of length of stay, laboratory, pharmacy, and ogy, wherein the outcomes portion includes at least one of readmissions, complications, emergency room visits, and mortality, and wherein each analytic point in the performance matrix contains a pre-processed specific measure of performance expressed as a ence between actual and expected along with the al impact of the difference, wherein expected values are risk adjusted to account for differences in case mix.

11, The non-transitory machine readable storage device of claim 10 wherein the pre-processed 10 specific measure of mance of each analytic point is pre-calculated using indirect rate rdization based on an exhaustive and mutually exclusive set of risk groups for risk adjustment.

12. The non-transitory machine readable storage device of claim 11 wherein for each risk group (g) for each performance measure (m), a target value (T(g,m)) is established based on an actual historical 15 average value in a reference database, and wherein for service provider (p) for e (m), an expected value )) is the sum of overall risk groups ofthe product ofthe number of patients/enrollees in each risk group (N(p,m,g) times the corresponding target value (T(g,m) divided by the total number of ts/enrollees expressed as: E(p,m) = sum over g [N(p,m,g)*T(g,m)] / sum over g g) 20 and wherein the difference between the service provider’s actual value and the expected value is expressed as above expected (negative performance) or below expected (positive performance).

13. A health ment system comprising: a searchable multi-dimensional data representation of the performance of an entire health care 25 delivery system accessible by one or more processors, in which the performance of every healthcare provider, including downstream providers, that are delivering es, is distilled down to a clinically le measure of actual versus expected performance at analytic points across a comprehensive set of quality outcomes and resource utilization measures; a memory device coupled to the sor and having a m stored thereon for execution by the one or more processors to perform operations comprising: creating the multi-dimensional data entation to obtain performance measures of a selected care provider; and accessing the multi—dimensional data representation to obtain performance measures of the selected healthcare er.

14. The health management system of claim 13 wherein the clinically credible measure comprises at least one of readmission rate and complication rate, wherein the performance matrix has multiple dimensions including dual health care providers, sites of service, quality outcomes and resource use measures, type of patients, time periods covered, geographic location of provider and t and the t’s payer, wherein the healthcare providers e at least multiple of hospitals, nursing homes, home health care agencies, lists, and physicians, wherein the types of patients include at least one of encounters for a procedure, encounters for chronic or acute disease management, disease cohorts of patients, es of care, and population management, wherein a performance ion of the performance matrix is broken into a resources portion and an outcomes portion, wherein the resource portions e at least one of length of stay, laboratory, pharmacy, and radiology, wherein the outcomes portion includes at least one of readmissions, complications, emergency room visits, and mortality.

15. The health management system of claim 13 wherein each analytic point in the performance matrix contains a pre-processed specific measure of mance expressed as a difference between actual and expected along with the financial impact ofthe difference, wherein expected values are risk adjusted to account for differences in case mix, wherein the ocessed specific measure of performance of each 15 analytic point is pre—calculated using indirect rate standardization based on an exhaustive and mutually exclusive set of risk groups for risk ment, wherein for each risk group (g) for each performance measure (In), a target value (T(g,m)) is established based on an actual historical average value in a nce database, and wherein for service provider (p) for measure (m), an ed value (E(p,m)) is the sum of overall risk groups of the product of the number of patients/enrollees in each risk group 20 (N(p,m,g) times the corresponding target value (T(g,m) divided by the total number of patients/enrollees expressed as: E(p,m) = sum over g [N(p,m,g)*T(g,m)] / sum over g N(p,m,g) and wherein the difference between the service er’s actual value and the expected value is expressed as above expected (negative performance) or below expected (positive performance). WO 12851 SYSTEM 125 105 SYSTEM SYSTEM SEARCH ° 105 ENGWE SYSTEM