Examining the patient journey: Case studies using linked health insurance claims and clinical data

Transcription

1 Examining the patient journey: Case studies using linked health insurance claims and clinical data Daniel M. Huse, MA March 2017 White Paper

2 The availability of electronic health data has virtually exploded in the past decade. Various healthcare stakeholders, including life sciences companies, medical service providers and health insurers, view these data as a potentially rich asset for achieving the triple aim of better patient care experience, improved medical care quality and reduced cost per capita. The purpose of this white paper is to describe the approach of Truven Health Analytics, part of the IBM Watson Health business, to enriching our claims data through collaborations with other sources of patient-level clinical information. The value of linking data Electronic healthcare databases have been amassed by organizations with varying roles in the healthcare ecosystem. As such, each database reflects the mission of its creator and owner. For example, health insurers may capture data for the purpose of reimbursing patients and providers for the costs incurred due to illness and associated medical care. Healthcare service providers may maintain records of the medical care provided to their patients, including presenting symptoms, diagnostic evaluations and treatment plans. Life sciences companies may collect and store information about the uses and outcomes of their products; first in the context of clinical trials to support regulatory approval, then after marketing to track each new product s uptake in the market and to fulfill any post-marketing risk assessment monitoring commitments they may have made to regulators at the time of approval. To truly understand the big picture of population health and identify gaps with respect to achieving the triple aim, we believe it is necessary to combine all these perspectives. Ideally, a healthcare research database should have de-identified patient-level data that combine clinical and economic outcomes of each episode of medical care. Health insurance claims data have become the most widely used data resource for examining the patient journey, including diagnosis, treatment and outcomes. Such data are so commonplace that it may be easy to forget that they potentially would not be very useful without the ability to link each patient s claims from all their care providers, including hospitals, medical practices and pharmacies. Truven Health provides such a linkage to its health plan clients who pay claims for care. De-identified, linked and longitudinal claims histories for individual plan members are the core of the Truven Health MarketScan Research Databases. 1 EXAMINING THE PATIENT JOURNEY: CASE STUDIES USING LINKED HEALTH INSURANCE CLAIMS AND CLINICAL DATA

3 In addition to paid claims, many health plans and self-insured employers have other sources of information about their enrollees or employees, such as absence and disability data, workers compensation data, dental claims, laboratory results and health risk assessments (HRAs). Truven Health offers tools and knowledge to link such information to medical claims and thereby help data owners generate more insightful analytic reports than with claims alone. MarketScan linked data At the center of Truven Health for 40+ years has been the creation of large, patientlevel data repositories for our payer, provider and government clients. For payers, the foundation is provided by comprehensive histories of insurance claims linked to patient demographics and plan eligibility. The key to this effort is the ability to link information from multiple sources (adjudicated and paid medical claims, adjudicated and paid pharmacy claims, and member eligibility records). Many payers find such record linkage to be challenging and therefore look for outside assistance from Truven Health. Payers may also have other information about their members that they want to link to claims, in an attempt to better understand the determinants of medical spending. For example, some health plans obtain the results of clinical laboratory tests that they pay for, such as blood and urine chemistries and Pap smears. These results can be used in conjunction with the claims to help evaluate the quality of care. Self-insured employers are increasingly investing in employee wellness programs that may include personal HRAs. These can be linked to claims to see how medical spending differs according to employees self-reported health risks (such as body weight, blood pressure and tobacco use). Employers often want to include the cost of time lost from work in their calculation of the cost burden of employee illness. To that end, they may opt to create linkages from claims to records of paid time off from payroll systems and claims for short- and long-term disability payments and workers compensation. Truven Health builds client-specific databases for our payer (employer and health plan) customers that link all of the types of data they need to manage their healthcare costs. Many of these customers ask Truven Health to include their data in the MarketScan Databases, which have been de-identified and certified by the Health Insurance Portability and Accountability Act (HIPAA) for use in healthcare research. In the next section, we provide some examples of how Truven Health uses its data linkage knowledge to serve other customers, particularly in the life sciences industry. EXAMINING THE PATIENT JOURNEY: CASE STUDIES USING LINKED HEALTH INSURANCE CLAIMS AND CLINICAL DATA 2

4 Case studies: Linking MarketScan claims data to other clinical databases In this section, we summarize some completed data linkages that illustrate the value of adding clinical data to the claims histories already present in the MarketScan Databases. These include linkages to oncology electronic medical record (EMR) data and some clients in-house data on their customers. Linkage to oncology EMR data can improve ascertainment of adverse events (AEs) Truven Health obtained data from a specialty EMR system for oncology practices and linked the patient-level medical records to health insurance claims for the same patients found in the MarketScan Databases. Data on approximately two million oncology patients who received treatment for a wide range of cancer types from 2004 to 2014 are captured in the linked data set. To complete a recent study, Truven Health researchers took advantage of this linkage to compare the reporting of treatment-related AEs in medical records and insurance claims. A cohort of 278 patients with multiple myeloma whose treatment was initiated between 2011 and 2014 was selected. Five categories of AEs were identified based on ICD-9-CM diagnosis codes in the patients insurance claims and medical records: neutropenia, thrombocytopenia, venous thromboembolism (VTE), peripheral neuropathy and diarrhea. The results* for patients first line of therapy were as follows: Neutropenia, thrombocytopenia and peripheral neuropathy were recorded less often in the claims data compared to the EMR data (neutropenia 9.4 versus 11.5 percent; thrombocytopenia 6.5 versus 11.2 percent; peripheral neuropathy 4.0 versus 6.5 percent). VTE and diarrhea were recorded more often in claims compared to the EMR (VTE 9.0 versus 3.6 percent; diarrhea 7.2 versus 4.3 percent). Overall more patients were identified as having the selected AEs in the combined claims and EMR data than either alone: 18.7 percent neutropenia, 15.8 percent thrombocytopenia, 9.7 percent VTE, 9.0 percent peripheral neuropathy and 10.8 percent diarrhea. * Irwin DE, Varker H, Princic N, Farr A. Comparison of treatment-related adverse events recorded in administrative claims data with those recorded in electronic medical records for multiple myeloma patients. Truven Health Analytics Research Brief. October Linkage to in-house customer data can help manufacturers study patient outcomes Many life science innovators maintain in-house databases of patients who use their products. Truven Health can link these private databases to MarketScan claims data for the purpose of studying patient outcomes. The following examples cover a range of medical technologies. Clinical laboratory testing A clinical laboratory has a proprietary diagnostic algorithm for evaluating patient compliance with pain treatment, including taking their medication as prescribed and not taking other medications that interact adversely with the prescribed treatment. This laboratory company maintains a database of all claims they 3 EXAMINING THE PATIENT JOURNEY: CASE STUDIES USING LINKED HEALTH INSURANCE CLAIMS AND CLINICAL DATA

5 have submitted to insurers for reimbursement for testing provided to the plans beneficiaries. To the extent that those claims were paid by insurers that contribute data to the MarketScan Databases, Truven Health was able to match the proprietary diagnostic services to individuals who had claims for the same services (CPT codes) on the same dates and with the same demographic profile. This linked data set helped researchers examine patients use of pain medication and other medical services after testing, compared with a control group who used other clinical laboratory services. Patients using the proprietary laboratory services had higher levels of service utilization over follow-up periods ranging from six months to three years, including more frequent dispensing of opioid analgesics, more frequent clinical laboratory testing, more frequent office visits and more frequent hospital-based services (both inpatient and outpatient). The study sponsor (the owner of the proprietary diagnostic algorithm) was able to present this information to a national meeting of health plan medical directors. Linkage to registry of implanted medical devices Manufacturers of implantable medical devices are required to keep records of all patients who have received such implants. Truven Health was approached by a manufacturer of implantable cardiac defibrillators and cardiac resynchronization therapy devices to examine healthcare utilization associated with actuation of these devices ( shocks ). The manufacturer is able to monitor all shocks generated by their devices in use and keeps a record of the dates of these events for all patients living with the implants. The manufacturer patient data were matched to MarketScan claims based on the date and type of implant and patient demographics (age, gender and geographic location). A total of 7,705 patients were matched between the two databases, of whom 715 had a total of 1,351 recorded shock events while in the community (that is, not while hospitalized). Of these events, 575 (43 percent) were followed by shock-related healthcare services beginning within seven days, at a mean cost of $5,124. Another 224 (22 percent) of the shock events were followed by healthcare service utilization that did not meet the study definition of shockrelated (diagnosis of dysrhythmia, syncope, palpitations, tachycardia or device malfunction )*. * Turakhia MP, Reynolds MR, Gunnarson CL, Swain AL, Mollenkopf SA, Zweibel SL. Healthcare utilization and expenditures associated with implantable cardioverter defibrillator shocks. Circulation: Cardiovascular Quality and Outcomes. 2013; 6:A3. Manufacturers of implantable medical devices are required by the US Food and Drug Administration (FDA) to keep records of all patients who have received an implant in case of a potential recall or other safety issue. When linked to longitudinal patient-level data, these registries can be an effective tool for investigating the safety and effectiveness of the devices that are tracked. EXAMINING THE PATIENT JOURNEY: CASE STUDIES USING LINKED HEALTH INSURANCE CLAIMS AND CLINICAL DATA 4

6 Summary Patient-level electronic healthcare data are now available from diverse sources, each providing a different view of the patient experience. The data become potentially more valuable for stakeholders in healthcare if they can be linked, to provide comprehensive longitudinal records for individual patients. Truven Health Analytics provides such data linkage and analysis services for both commercial and government healthcare clients. This white paper illustrates what can be accomplished by linking our proprietary MarketScan claims databases to additional data sets held by our clients and other parties. 5 EXAMINING THE PATIENT JOURNEY: CASE STUDIES USING LINKED HEALTH INSURANCE CLAIMS AND CLINICAL DATA

7 EXAMINING THE PATIENT JOURNEY: CASE STUDIES USING LINKED HEALTH INSURANCE CLAIMS AND CLINICAL DATA 6

8 For more information To learn more about this white paper or other health analytics and research services offered by Truven Health Analytics, part of the IBM Watson Health business, please contact us at or visit truvenhealth.com/life-sciences. Corporate Headquarters 100 Phoenix Drive Ann Arbor, MI USA Greater Europe Boulevard Saint Michel, Brussels, Belgium Asia-Pac 50 Raffles Place Singapore Land Tower Level 30 Singapore For local offices worldwide, visit truvenhealth.com/ls-locations. About Truven Health Analytics, part of the IBM Watson Health business Truven Health Analytics, part of the IBM Watson Health business, provides market-leading performance improvement solutions built on data integrity, advanced analytics and domain expertise. For more than 40 years, our insights and solutions have been providing hospitals and clinicians, employers and health plans, state and federal government agencies, life sciences companies and policymakers, the facts they need to make confident decisions that directly affect the health and well-being of people and organizations in the US and around the world. The company was acquired by IBM in 2016 to help form a new business, Watson Health. Watson Health aspires to improve lives and give hope by delivering innovation to address the world s most pressing health challenges through data and cognitive insights. Truven Health Analytics owns some of the most trusted brands in healthcare, such as MarketScan, 100 Top Hospitals, Advantage Suite, Micromedex, Simpler and ActionOI. Truven Health has its principal offices in Ann Arbor, MI, Chicago, IL and Denver, CO. For more information, please visit truvenhealth.com. truvenhealth.com Copyright 2017 Truven Health Analytics, part of the IBM Watson Health business. IBM, the IBM logo and ibm.com are trademarks of IBM Corporation in the United States, other countries or both. Truven Health Analytics and its respective logo are trademarks of Truven Health Analytics in the United States, other countries or both. All other company or product names are registered trademarks or trademarks of their respective companies. A current list of IBM trademarks is available on the web at Copyright and trademark information at ibm.com/legal/copytrade.shtml. LS