Center for Large Data Research and Data Sharing in Rehabilitation Learning Modules Use of Geographic Variation to Estimate Effectiveness With Nonrandomized Data Use of Geographic Variation to Estimate Effectiveness With Nonrandomized Data Click to view the video in a new window. Notes Recorded at the Comparative Effectiveness Research with Population-Based Data Conference, Baker Institute at Rice University, 2012. Presented by Mary Beth Landrum, PhD, Harvard Medical School. Learning Module Notes Modules Use of Geographic Variation to Estimate Effectiveness with Nonrandomized Data Introduction (0:00 - 8:09)Mary Beth Landrum: Presentation Overview (8:09 – 55:40)The difficulty of comparative effectiveness research in population-based data setsMary Beth Landrum: Q&A (55:40 – 1:11:29 )Speaker Transition Introductions: (1:11:29 – 1:14:45)Matt Mejuski – Health Economist & Health Services Researcher Presentation Overview: (1:14:45 – 1:16:53 )Discuss motivation around quasi-experiments, context & skepticismIncident vs. prevalent user cohortStudy design characteristics to improve the rigorInternal Validity Threats, Strengths of External Validity Principles around Quasi-Experimental study designs: (1:16:53-2:05:10 )Experimental studies/RCTs enable causality statementsMotivation & Value of Quasi-Experimental StudiesIdeal Quasi-Experiment: Internal Validity, HRT ControversyIndirect evidence between RTC & quasi-experimentsConclusions about concordance4 Design choices for rigorous quasi-experimentsThe bottom line for quasi-experiments Q&A (2:05:10 – 2:17:33) Learning Modules