Outcomes Research in Psoriasis and Psoriatic Arthritis Using Large Databases and Research Networks: A Report from the GRAPPA 2013 Annual Meeting

Electronic medical records (EMR)

EMR have been widely adopted in the past few years and present a remarkable opportunity for clinicians to enter visit information at point of care; laboratory and imaging data are also readily available. However, challenges exist for extracting EMR data for population research: medical records housed on different electronic health record platforms create difficulties for combining data; and despite technological advances such as natural language processing, extracting valid responses from nonstructured data elements is problematic². Despite these challenges, however, real-world data obtained at point of care presents appealing opportunities for generating “large” data to power hypothesis-driven, patient-oriented research.

In the simplest form, a single practitioner through his or her practice generates clinical care data based on individual patient encounters. Alternatively, a single institution can combine patient data from multiple practitioners to create larger datasets that reflect more diverse patient populations and a wider range of clinician practice patterns. In multi-institutional or multicentered datasets, data are combined from several groups that usually share a common EMR platform.

For example, in the United States, several EMR systems exist. To date, EMA is the largest dermatology-specific EMR system in the United States³, whereas Epic⁴ and Cerner⁵ are non-specialty-specific EMR systems used widely in academic institutions and healthcare maintenance organizations. The Veteran’s Affairs⁶ Healthcare System uses a centralized EMR for all its US sites, which makes extraction of EMR data possible for research across large populations. There are different approaches to validating diagnostic codes for psoriasis and PsA, including in-person examination by dermatologists or rheumatologists, chart reviews, and retrospective surveys of practitioner and patients. Validation of diagnostic codes for psoriasis and PsA was performed in the Kaiser Permanente EMR as well as databases from the UK, including the General Practice Research Database and The Health Improvement Network (THIN^7,8,9,10). THIN, for example, allows investigators to obtain additional information such as body surface area affected by psoriasis by surveying the patient’s general practitioner^11,12. Thus, validation of outcomes, examination of data availability, extraction of structured data, and handling of unstructured data are important considerations when using data extracted from EMR.

Research networks and disease registries

While research networks and disease registries differ in scope, mission, and patient population, many were created to evaluate effectiveness and monitor safety of systemic medications among patients with moderate-to-severe psoriasis and/or PsA. Research networks and disease registries often collect granular and relevant data, but their setup and sustainability costs continue to be challenging.

Psonet¹³ is a European system of independent registries that monitors longterm efficacy of systemic medications for treatment of psoriasis¹⁴. Currently participating countries include the UK, Portugal, Spain, France, The Netherlands, Germany, Switzerland, Italy, Denmark, Sweden, Israel, and Australia. Psonet researchers collect data on effectiveness and adverse events related to biologics or other systemic medications in patients with psoriasis and/or PsA. Similarly, the German Psoriasis Registry collects data from patients with psoriasis and PsA to determine the longterm safety, tolerability, and efficacy of biologics and other systemic medications that are used in Germany¹⁵. The Australasian Psoriasis Registry (APR) established an analogous database of patients with psoriasis in Australia, New Zealand, and New Guinea¹⁶. The Dermatology Clinical Effectiveness Research Network in the United States is a system of multiple dermatology practices that conducts clinical research studies of comparative effectiveness to help inform treatment decisions, and currently maintains a database of over 1700 patients with moderate to severe psoriasis evaluated during routine clinical followup care^{17,18,19,20,21,22}.

Members of the International Psoriasis and Arthritis Research Team (IPART), which includes several groups in Canada and the United States, have developed a large database of well-phenotyped patients with PsA and/or psoriasis based on an observational cohort at the University of Toronto. The international collaboration aims to use the IPART database to perform genome-wide association studies, in addition to studying disease progression, severity, and comorbidities²³.

Government-mandated postmarketing registries can provide valuable data regarding longterm medication safety^21,24, and many registries also collect additional information to help inform comorbidities research. For example, the Psoriasis Longitudinal Assessment and Registry is a 10-year study evaluating the longterm safety of the biologic therapies ustekinumab and infliximab²⁴ in 12,000 US patients with moderate to severe plaque psoriasis. Researchers are also collecting other important outcomes data, e.g., patient quality of life, clinical outcomes, and comorbidities.

While a number of other registries and research networks exist that are not specific to psoriasis or PsA, they often include those populations in their mission of evaluating patients receiving biologic treatments or those with rheumatologic diseases. The British Association of Dermatologists Biologic Interventions Registry (BADBIR) houses observational data from a large cohort of patients with psoriasis, rheumatologic conditions, and/or inflammatory bowel diseases being treated with biologics or other systemic medications to determine the longterm safety of biologics in the UK and Ireland²⁵. Another example is the Consortium of Rheumatology Researchers of North America (CORRONA), a registry that houses data from patients with rheumatologic conditions (primarily rheumatoid arthritis, but also PsA) from the United States²⁶.

Other data sources: National healthcare data and surveys, administrative databases, and crowdsourced data

Various regions of the world have different standards for collecting population-based healthcare data. For example, whereas nationalized and centralized databases for healthcare records exist in countries such as Israel²⁷, Denmark²⁸, and Taiwan²⁹, many countries lack such databases. In the United States, the Centers for Medicare and Medicaid Services house centralized data for individuals ≥ 65 years as well as those enrolled in the Medicaid program (the federal healthcare program for those who cannot afford health insurance)³⁰. Other federally funded databases such as the National Ambulatory Medical Care Survey/National Hospital Ambulatory Medical Care Survey³¹ and National Health and Nutrition Examination Survey^32,33,34,35 obtain nationally representative data, but their utility is limited by the number and type of questions specific to psoriasis and PsA. In the UK, practitioners have an incentive to put data into the Clinical Practice Research Datalink to yield more granular data for outcomes assessment and other research purposes³⁶.

Administrative and claims databases such as Truven Health Analytics MarketScan are helpful in providing aggregate data on diagnosis and medication use³⁷. However, those data are derived for the purposes of billing and may be more prone to misclassification that could threaten data validity.

Finally, crowdsourcing-generated data also provide valuable information^{38,39,40,41,42}. In healthcare research, crowdsourcing is the act of outsourcing data collection to patients. Potential advantages include recruitment of larger patient populations, lower cost for data collection, obtaining patient-centered perspectives, and the ability of patients to contribute through the Internet, providing ongoing opportunities for patients from diverse backgrounds to contribute to the collective patient experience.

The opportunities for conducting outcomes research using the various large databases and research networks are becoming more abundant. Aligning the research question with the appropriate database and analytical approach, however, is key to addressing clinically relevant questions appropriately. Although access to the various databases may be limited by cost and procedural challenges, the future for outcomes research in psoriasis and PsA is promising. With continued improvements in healthcare informatics, creation of more psoriasis and psoriatic disease-specific research networks, and advancements in analytical approaches, we will be able to increasingly use real-world observational data to perform quality research that informs clinical decisions and improves patient outcomes.