Elsevier

Journal of Clinical Epidemiology

Volume 89, September 2017, Pages 4-11
Journal of Clinical Epidemiology

Series: Quasi-Experimental Study Designs
Quasi-experimental study designs series—paper 1: introduction: two historical lineages

https://doi.org/10.1016/j.jclinepi.2017.02.020Get rights and content

Abstract

Objectives

The objective of this study was to contrast the historical development of experiments and quasi-experiments and provide the motivation for a journal series on quasi-experimental designs in health research.

Study Design and Setting

A short historical narrative, with concrete examples, and arguments based on an understanding of the practice of health research and evidence synthesis.

Results

Health research has played a key role in developing today's gold standard for causal inference—the randomized controlled multiply blinded trial. Historically, allocation approaches developed from convenience and purposive allocation to alternate and, finally, to random allocation. This development was motivated both by concerns for manipulation in allocation as well as statistical and theoretical developments demonstrating the power of randomization in creating counterfactuals for causal inference. In contrast to the sequential development of experiments, quasi-experiments originated at very different points in time, from very different scientific perspectives, and with frequent and long interruptions in their methodological development. Health researchers have only recently started to recognize the value of quasi-experiments for generating novel insights on causal relationships.

Conclusion

While quasi-experiments are unlikely to replace experiments in generating the efficacy and safety evidence required for clinical guidelines and regulatory approval of medical technologies, quasi-experiments can play an important role in establishing the effectiveness of health care practice, programs, and policies. The papers in this series describe and discuss a range of important issues in utilizing quasi-experimental designs for primary research and quasi-experimental results for evidence synthesis.

Section snippets

Background

The quest to understand causal relationships permeates the history of health research [1], [2]. While the randomized controlled trial has become the mainstay for clinical efficacy and safety testing, quasi-experiments offer important alternatives and additional opportunities for causal inferences about health and health care. Quasi-experiments can generate effect size estimates that can come close in causal strength to those obtained in controlled trials because—like trials—they can control for

Historical origins of quasi-experiments

One very broad definition of an experiment is a study, in which a researcher intervenes in the “natural” processes, to establish the causal effects of a treatment—in contrast, a quasi-experiment can then be identified as any study, in which the causal effects of a treatment are established without a researcher's intervention. In different sciences, different types of experiments have developed, which can lead to strong causal inferences through complete control of confounding. In the laboratory

A series on quasi-experiments in health research

While the historical origins and developments of different quasi-experiments are highly varied, the designs share important similarities. These similarities—in potential uses, in strengths and limitations, in the processes that typically generate the necessary data, etc.—become apparent when quasi-experiments are discussed together as a category of study designs that is distinct from experiments, on the one hand, and “nonexperiments,” on the other hand. The distinctness as a category of study

References (81)

  • T. Bärnighausen et al.

    Quasi-experimental study designs series - Paper 7: assessing the assumptions

    J Clin Epidemiol

    (2017)
  • A.M. Aloe et al.

    Quasi-experimental study designs series-paper 9: collecting data from quasi-experimental studies

    J Clin Epidemiol

    (2017)
  • B.J. Becker et al.

    Quasi-experimental study designs series-paper 10: synthesizing evidence for effects collected from quasi-experimental studies presents surmountable challenges

    J Clin Epidemiol

    (2017)
  • J.N. Lavis et al.

    Quasi-experimental study designs series - Paper 11: Supporting the production and use of health systems research syntheses that draw on quasi-experimental study designs

    J Clin Epidemiol

    (2017)
  • P.C. Rockers et al.

    Quasi-experimental study designs series-paper 12: strengthening global capacity for evidence synthesis of quasi-experimental health systems research

    J Clin Epidemiol

    (2017)
  • T. Bärnighausen et al.

    Quasi-experimental study designs series - paper 1: history and introduction

    J Clin Epidemiol

    (2017)
  • J. Glanville et al.

    Quasi-experimental study designs series-paper 8: identifying quasi-experimental studies to inform systematic reviews

    J Clin Epidemiol

    (2017)
  • P. Vineis

    Causality in epidemiology

    Soz Praventivmed

    (2003)
  • A. Morabia

    Epidemiological causality

    Hist Philos Life Sci

    (2005)

  • USAID evaluation policy – evaluation: learning from experience

    (2011)
  • S. Bennett

    Building the field of health policy and systems research: an Agenda for action

    PLoS Med

    (2011)
  • X. Bosch-Capblanch

    Guidance for evidence-Informed policies about health systems: rationale for and challenges of guidance development

    PLoS Med

    (2012)
  • K. Baicker

    The Oregon experiment–effects of Medicaid on clinical outcomes

    N Engl J Med

    (2013)
  • A. Finkelstein et al.

    Health care policy. Randomize evaluations to improve health care delivery

    Science

    (2015)
  • J. Tollefson

    Revolt of the randomistas

    Nature

    (2015)
  • E.V. Langlois

    Advancing the field of health systems research synthesis

    Syst Rev

    (2015)

  • Social policy experiments in the European Union: examples in member states

    (2011)

  • Testing social policy innovation: primer for training

    (2014)
  • J. Bor

    Regression discontinuity designs in epidemiology: causal inference without randomized trials

    Epidemiology

    (2014)
  • E.J. Tchetgen Tchetgen

    Instrumental variable estimation in a survival context

    Epidemiology

    (2015)
  • A.C. Ericsson et al.

    A brief history of animal modeling

    Mo Med

    (2013)
  • J. Neyman

    On the application of probability theory to agricultural experiments. Essay on principles. Section 9

    Stat Sci

    (1990)
  • D. Rubin

    Estimating causal effects of treatments in randomized and nonrandomized studies

    J Educ Psychol

    (1974)
  • L.E. Bothwell et al.

    The emergence of the randomized, controlled trial

    N Engl J Med

    (2016)
  • I. Chalmers

    The advent of fair treatment allocation schedules in clinical trials during the 19th and early 20th centuries

    J R Soc Med

    (2012)
  • Jamison JC. A controlled history of randomized assignment in social science. Working...
  • C.S. Peirce et al.

    On small differences in sensation

    Mem Natl Acad Sci

    (1885)
  • J. Lind

    A treatise of the scurvy. In three parts. Containing an inquiry into the nature, causes and cure, of that disease. Together with a critical and chronological view of what has been published on the subject

    (1753)
  • J. Fibiger

    Om serumbehandling af difteri

    Hospitalstidende

    (1898)
  • T. Sollmann

    The crucial test of therapeutic evidence

    JAMA

    (1917)

    • Cited by (46)

    • Effects of initiating insulin pump therapy in the real world: A nationwide, register-based study of adults with type 1 diabetes

      2023, Diabetes Research and Clinical Practice
      Citation Excerpt :

      Whereas current observational studies contribute ample evidence of an existing association between IPT and improved health outcomes that complements that of RCTs, causal inference is limited by design [24]. Natural or quasi-experimental research designs are becoming increasingly popular as a means to overcome this limitation and evaluate causal effects of, e.g., health care policies or treatments using real-world, observational data [25,26]. Like the randomisation process in RCTs, the goal of such designs is to control for both observed and unobserved confounding under a potential outcomes framework by exploiting quasi-random variation in routinely collected data [25–27].

    • Systematic Review Reveals Lack of Causal Methodology Applied to Pooled Longitudinal Observational Infectious Disease Studies

      2022, Journal of Clinical Epidemiology
      Citation Excerpt :

      Based on the small number of pooled ID-focused IPD analyses and the lack of modern causal method implementation found in this systematic review, we recommend increased interdisciplinary collaborations between statistical methodologists and ID-focused researchers. We also suggest the development of guidance documents related to the implementation and reporting of causal methods in the analysis of pooled participant-level data from longitudinal observational ID-focused studies which could be similar to the guidance produced for a recent series on quasi-experimental methods in health [57–69]. This study did not require any ethical approval because we used only publicly available data.

    • Epidemiology and development economics two sides of the same coin in impact evaluation

      2022, Journal of Clinical Epidemiology
      Citation Excerpt :

      Evaluation studies conducted by economists frequently use quasi-experimental methods, although randomized trials have become accepted as the golden standard (R. Khandker, Gayatri B., & Hussain A., 2010). Methods used in economic studies such as instrumental variable models (IV), local average treatment effect models, regression discontinuity design are now starting to be considered by epidemiologists [9,10]. Interrupted-time series designs and cohort studies, which are used by epidemiologists, are practically absent in evaluations conducted by economists.

    • Adjusting mean arterial pressure alarms improves the time spent within blood pressure targets in patients with septic shock: A quasi-experimental study

      2021, Australian Critical Care
      Citation Excerpt :

      Nurses were also unaware that MAP control was evaluated both before and after implementation of MAP alarms, limiting the risk of Hawthorne effect. Thus, the comparability of the two groups herein limits usual concerns regarding internal validity of quasi-experimental studies.17 Second, patients were still outside the prescribed target range a quarter of the time, which suggests that, at least in part, the dose of norepinephrine was not timely adjusted.

    • Trials embedded in cohorts, registries, and health care databases are gaining ground

      2020, Journal of Clinical Epidemiology

    • Design and statistical analysis reporting among interrupted time series studies in drug utilization research: a cross-sectional survey

      2024, BMC Medical Research Methodology
    View all citing articles on Scopus
    View full text
    © 2017 Elsevier Inc. All rights reserved.