RTI International Methods for Comparative Effectiveness and Health Services Research Kathleen N. Lohr, PhD Distinguished Fellow (Health Services Research) RTI International Research Triangle Park, NC USA AcademyHealth Tuesday June 14, 2011 Seattle Washington RTI International is a trade name of Research Triangle Institute. www.rti.org RTI International CER Methods Research—Gaps and Needs CER defined well by previous speakers Need to be clear, though, that CER Covers more than clinical research – Involves health services research – Can go beyond condition-specific health care to health care delivery and systems – Includes both generation and synthesis of evidence – For CTSAs: 3 “critical needs” with 10 “priority areas” for methods – Today: One priority area from each of the three buckets – Value of information Treatment heterogeneity Translation and dissemination RTI International Value of Information Relatively new concept; multiplicity of subparts Methods research as applied to decisionmaking about CER in CTSAs Methods research needed, also, for decisionmaking about systematic reviews and health technology assessments: e.g., Value of doing a systematic or comparative effectiveness review on one topic vs. another – Value of doing an update -- of a review; of a guideline – Overarching questions Value to whom? – Interactions: CER and systematic review feedback loops? – RTI International Heterogeneity of Treatment Effects As well illustrated and explained: HTE = major challenge For researchers in designing studies, analyzing data – For investigators in conducting systematic reviews or health technology assessments – For clinical professionals and guidelines developers – Numerous approaches to addressing the program and reporting results – but still need to be widely adopted RTI International Heterogeneity Related concepts and methods challenges: Clinical heterogeneity: “variability in study population characteristics, interventions, and outcomes across studies”* – Statistical and/or methodologic heterogeneity: causal relationship (or not) with clinical heterogeneity? – Effect-measure modification: a intervention-disease association that differs according to the level of a factor under investigation – Applicability (external validity; generalizability): “extent to which effects observed in published studies are likely to reflect the expected results when an intervention is applied to a pre-specified population under ‘real-world’ conditions”† – *Gartlehner et al., IJTAHC, forthcoming †Atkins et al., JCE, 2011 RTI International Heterogeneity Any unifying concepts for CER and for guidelines developers – that call for methods research? – PICOTS Patients/populations Interventions Comparators Outcomes Time frames Settings Net benefits: totality of benefits and harms (risks) – Information for informing patient care – Shared/informed decisionmaking Individualized (personalized) care Are “averages” or “subgroup” results ever enough (maybe sometimes?) RTI International Translation and Dissemination Getting stakeholder involvement in research (or systematic reviews) – only a first, early step Relatively easy: professional groups, disease advocacy groups, insurers, payers, health care systems – Not so easy: Patients and patient advocacy groups – Consider role of community-based participatory research – Moving research findings into practice: overcoming the “17-year lag” Conveying the meaning of “uncertain” results, and helping decisionmakers deal with uncertainty Reaching patients/families/consumers: overcoming literacy and numeracy challenges RTI International CER Research Design, Statistics, Analysis AHRQ CER Methods conferences (2006, 2009, 2011) plus Medical Care supplements (2007*, 2010†) Broad range of topics New and expanded designs: randomization techniques; pragmatic/effectiveness trials; practice-based studies; registry studies – Data collection: electronic, computer- or web-based platforms and hardware; EMRs; distributed data networks; administrative to primary data (e.g., patient-reported outcomes), and combinations of these – Established/newer statistical and analytic methods: propensity scores to Bayesian techniques; confounding; missing data – Applications to policy and clinical decisionmaking – * AHRQ Effective Health Care program or DEcIDE sites: http://effectivehealthcare.ahrq.gov/index.cfm † http://journals.lww.com/lww-medicalcare RTI International Patient-Centered Care and CER PCORI and its Methodology Committee is a start Patient-Reported Outcomes: another challenge for: Meeting CER (and CTSA) and health services research agendas – Conducting Systematic reviews and health technology assessments – Compiling clinical practice guidelines – Using routinely in clinical practice – PROs: Concepts, Measures, and Tools Decades of measurement and instruments – Newer tools: PROMIS (NIH): http://www.nihpromis.org/ – Translational: from research to practice* – *Lohr & Zebrack, QOL Research 2009