Related Experiment Video
Updated: May 29, 2026

The Participant-Reported Implementation Update and Score (PRIUS): A Novel Method for Capturing Implementation-Related Data Over Time
Published on: February 19, 2021
Measuring the effect of policy interventions at the population level: some methodological concerns
Marco D Huesch1, Truls Østbye, Michael K Ong
1Community & Family Medicine, Duke University School of Medicine, Durham, NC, USA. m.huesch@duke.edu
This study on New Jersey's smoking ban found that policy evaluations need clear causal relationships and testing methods. Results show sensitivity to different statistical models and inference modes for health outcomes.
Area of Science:
- Health Services Research
- Public Health Policy
- Biostatistics
Background:
- Health policy evaluations often lack clarity in causal relationships and parameter identification.
- Methodological challenges can hinder accurate estimation of intervention effects on population health outcomes.
Purpose of the Study:
- To examine the impact of the New Jersey statewide smoking ban on acute myocardial infarctions, strokes, lower limb fractures, and asthma exacerbations.
- To discuss identification options and assess the sensitivity of response function estimates to various model specifications and inference modes.
Main Methods:
- Utilized interrupted time series analysis to evaluate statewide admission and emergency room encounter rates before and after the smoking ban.
- Employed rolling Chow tests to detect structural breaks and assess model misspecification.
Main Results:
- Estimates of the health policy response function were sensitive to different specifications of stochastic and intervention components.
- Different modes of statistical inference yielded varying results, highlighting the impact of methodological choices.
Conclusions:
- Clear definition of causal relationships and robust identification strategies are crucial for health policy evaluations.
- Model misspecification can significantly affect the estimated impact of public health interventions, necessitating careful validation.
Related Concept Videos
Analysis of Population Pharmacokinetic Data
Strategies for Assessing and Addressing Confounding
Confounding can be addressed at both the design phase of a study and through analytical methods after data...
Mechanistic Models: Compartment Models in Individual and Population Analysis
Types of Biopharmaceutical Studies: Controlled and Non-Controlled Approaches
Non-controlled studies, commonly employed for initial exploration, lack a control group, rendering them susceptible to biases and external influences. In contrast, controlled...
Measurement of Bioavailability: Pharmacodynamic Methods
Regression Toward the Mean