Disaggregating health differences and health disparities in clinical care

Item: Disaggregating health differences and health disparities in clinical care
Rating: 4
Author: PostSingularity Institute

by Paula Strassle PhD

Life & Natural Systems Medicine & Health MBNPublic health and preventive medicine MBNDSocial medicine

🎙 Paula Strassle PhD 👥 884 📅 March 20, 2026 ⏱ 39 min 👁 83 📄 original study 🧭 2026-08-16

Available in: English (current) Français

Keywords

health disparitieshealth differencesobserved-to-expected ratiomachine learningperipheral arterial disease

Summary

In this seminar, Dr. Paula Strassle, an epidemiologist at the University of Maryland, presents her research on disaggregating health differences and health disparities in clinical care. She begins by sharing her career trajectory, emphasizing that research interests often evolve in a non-linear fashion. She then defines health disparities according to NIMHD and Healthy People 2030, highlighting the importance of reference groups and the distinction between allowable and unallowable differences. She introduces a novel analytic approach that uses observed-to-expected ratios and machine learning to separate total differences into allowable (biological/clinical) and unallowable (social/environmental) components, and further into explained and unexplained portions. The method is applied to major lower limb amputation (MLLA) in patients with peripheral arterial disease (PAD) using HCUP state inpatient databases from five states. The analysis includes over 1.5 million hospitalizations and uses LASSO regression to predict amputation risk, then calculates O/E ratios stratified by race/ethnicity and rurality. The results indicate that clinical factors do not fully explain disparities, and a significant portion remains unexplained, suggesting potential implicit bias. The approach aims to avoid problematic reference groups and to provide interpretable results for clinical audiences. The presentation concludes with a discussion of limitations and future directions.

197 words

Critical Evaluation

Value of the Information & Strength of the Argument

The talk provides valuable insights into the methodological challenges of measuring health disparities. The speaker clearly explains the conceptual distinction between health differences and disparities, and the importance of reference groups. The proposed method is innovative, using machine learning to adjust for allowable factors and to quantify unexplained disparities. The argumentation is logical and well-structured, with a concrete example (amputation rates) that illustrates the concepts. The speaker acknowledges limitations and potential biases, which strengthens the credibility of the approach. However, the presentation is a seminar, so some technical details are simplified, and the results are not fully presented. The value lies in the methodological contribution and the call for more nuanced analysis in health disparities research.

Scientific Rigor, Source Quality, Title Accuracy

The presentation demonstrates scientific rigor through the use of a peer-reviewed methodology paper (published in Epidemiology) and the use of large, multi-state administrative databases (HCUP). The speaker cites relevant definitions from NIMHD and Healthy People 2030. The title accurately reflects the content. The talk is well-organized and the methods are described transparently. However, as a seminar, it does not provide full citations for all claims, and the results are not presented in detail. The speaker’s background and affiliation add credibility. Overall, the scientific rigor is high, but the format limits the depth of source documentation.

226 words

Title / Content Match

The title accurately reflects the content, focusing on disaggregating health differences and disparities in clinical care, with a specific example from amputation rates.

Quality & Reliability

8/10

The presentation is based on a peer-reviewed methodology paper published in Epidemiology, with transparent description of data sources and analytic approach. The speaker is a trained epidemiologist with relevant experience. However, the talk is a seminar presentation, not a full paper, and some details are simplified.

Key Moments

Markers derived by PSI from the transcript: the creator did not define chapters.

Cited Sources

NIMHD Health Disparities Definition — Definition of health disparities as differences in health outcomes among disadvantaged populations.
Healthy People 2030 — Benchmarks and definitions for health disparities.
HCUP State Inpatient Databases — Data source for the analysis.

Concurring Sources

NIMHD Health Disparities Framework — Provides the definition of health disparities used in the talk.
Healthy People 2030 — Defines health disparities and benchmarks.

Contribution & Novelties

The presentation introduces a novel analytic framework for disaggregating health disparities by separating total differences into allowable and unallowable components using machine learning and observed-to-expected ratios. This approach avoids problematic reference groups and allows for intersectional analysis. The method is applied to a real-world clinical outcome (amputation) and provides a way to quantify the potential role of implicit bias. This is a significant contribution to health disparities research methodology.

Pour aller plus loin :

Health Disparities: Concepts and Measurement — Overview of health disparities measurement.
Machine Learning in Epidemiology — Discussion of ML applications in public health.
Implicit Bias in Health Care — Review of implicit bias and its impact on health disparities.

113 words

Radar Profile

The radar profile shows high scores in quantity and quality of information, with moderate technical level and high reliability. This indicates a well-balanced presentation that is both informative and credible, with a strong methodological focus.

Reliability 8/10

💬 No comments were provided for analysis.