Meta-analysis of survival curve data using distributed health data networks: application to hip arthroplasty studies of the International Consortium of Orthopaedic Registries

Journal Research Synthesis Methods
Authors Cafri, Guy; Banerjee, Samprit; Sedrakyan, Art; Paxton, Elizabeth; Furnes, Ove; Graves, Stephen; Marinac-Dabic, Danica
Year Published 2015
Link to article


The motivating example for this paper comes from a distributed health data network, the International Consortium of Orthopaedic Registries (ICOR), which aims to examine risk factors for orthopedic device failure for registries around the world. Unfortunately, regulatory, privacy, and propriety concerns made sharing of raw data impossible, even if de-identified. Therefore, this article describes an approach to extraction and analysis of aggregate time-to-event data from ICOR. Data extraction is based on obtaining a survival probability and variance estimate for each unique combination of the explanatory variables at each distinct event time for each registry. The extraction procedure allows for a great deal of flexibility; models can be specified after the data have been collected, for example, modeling of interaction effects and selection of subgroups of patients based on their values on the explanatory variables. Our analysis models are adapted from models presented elsewhere – but allowing for censoring in the calculation of the correlation between serial survival probabilities and using the square root of the covariance matrix to transform the data to avoid computational problems in model estimation. Simulations and a real-data example are provided with strengths and limitations of the approach discussed.

Methods for multiple treatment comparisons – Annual Meeting Short Course

Half-day course: Sept 30, 2015, 12:00 -4:00 pm

Target audience: MS- and PhD-level quantitative researchers

Registration is now open – Course is FREE to all


Laura Hatfield, PhD
Department of Health Care Policy – Harvard Medical School

Sherri Rose, PhD
Department of Health Care Policy – Harvard Medical School


Randomized 2-arms trials remain the primary source of information establishing medical device effectiveness. Yet in many clinical settings, clinicians and patients choose among multiple possible medical devices. Comparative effectiveness research addresses this gap. To use observational data, we apply causal inference techniques to address confounding. In this course, we present modern statistical techniques that produce consistent evidence to support clinical decision making from among multiple medical device treatment options.

Learning objectives

At the end of the course, participants will be able to

  1. Describe a clinical decision process in terms of the applicable patient population, set of treatment options, treatment assignment mechanism (including all relevant confounders, both observed and unobserved), and clinical outcomes
  2. Identify inferential targets relevant to the desired clinical decision process
  3. Specify causal and statistical assumptions required for the inferential target to be valid
  4. Choose an estimation method that is feasible given the available data, decision process, and required assumptions


  • Essential methods elements of clinical decision problems
  • Two clinical examples
    • claims data for implantable cardiac devices
    • registry data for stents
  • Inferential targets
    • all pairwise comparisons (device effect on “treated” or average device effect)
    • posterior rankings of device effects
    • marginal structural model
  • Causal and statistical assumptions
  • Estimation methods
    • balancing scores
    • device effect estimates

Instructor Bios

Laura Hatfield, PhD is an Assistant Professor of Health Care Policy, with a specialty in Biostatistics. Dr. Hatfield received her BS in genetics from Iowa State University and her PhD in biostatistics from the University of Minnesota. Her research focuses on trade-offs and relationships among health outcomes. In particular, she develops and applies statistical methods that incorporate multiple sources of information and relationships among outcomes. Dr. Hatfield has particular expertise in Bayesian hierarchical modeling and has taught short courses in industry, government, and professional society meetings.

Sherri Rose, PhD is an Assistant Professor in the Department of Health Care Policy at Harvard Medical School. Dr. Rose received her BS in statistics from The George Washington University and her PhD in biostatistics from the University of California, Berkeley. Broadly, Dr. Rose’s methodological research focuses on nonparametric estimation in causal inference and machine learning for prediction.  Within health policy, Dr. Rose works on risk adjustment, health care program impact evaluation, and comparative effectiveness research. Dr. Rose has taught short courses on robust estimation in causal inference and comparative effectiveness for varied audiences.