Countercurrents: The bias of choice

Leveraging data science and machine learning for urban climate adaptation in two major African cities: a HE2AT Center study protocol

INTRODUCTION

African cities, particularly Abidjan and Johannesburg, face challenges of rapid urban growth, informality and strained health services, compounded by increasing temperatures due to climate change. This study aims to understand the complexities of heat-related health impacts in these cities. The objectives are: (1) mapping intraurban heat risk and exposure using health, socioeconomic, climate and satellite imagery data; (2) creating a stratified heat-health forecast model to predict adverse health outcomes; and (3) establishing an early warning system for timely heatwave alerts. The ultimate goal is to foster climate-resilient African cities, protecting disproportionately affected populations from heat hazards.

METHODS AND ANALYSIS

The research will acquire health-related datasets from eligible adult clinical trials or cohort studies conducted in Johannesburg and Abidjan between 2000 and 2022. Additional data will be collected, including socioeconomic, climate datasets and satellite imagery. These resources will aid in mapping heat hazards and quantifying heat-health exposure, the extent of elevated risk and morbidity. Outcomes will be determined using advanced data analysis methods, including statistical evaluation, machine learning and deep learning techniques.

ETHICS AND DISSEMINATION

The study has been approved by the Wits Human Research Ethics Committee (reference no: 220606). Data management will follow approved procedures. The results will be disseminated through workshops, community forums, conferences and publications. Data deposition and curation plans will be established in line with ethical and safety considerations.

A Validation of Methods for the Evaluation of Observational Studies of Screening Mammography: An Exploratory Analysis Based on Simulating Screening Cohorts

Background

The degree of confidence one should place on non-randomised observational trials studies which estimate the benefit of screening depends on the validity of the analytic method employed. As is the case for all observational trials, screening evaluation studies are subject to bias. The objective of this study was to create a simulated data set and to compare four analytic methods in order to identify the method which was the least biased in terms of estimating the underlying hazard ratio.

Methods

We simulated a cohort of 100,000 women who were accorded US national rates of breast cancer incidence and breast cancer mortality over their lifetime. We assigned at random one-half of them to initiate mammography screening between ages 50 and 60. We used four different analytic approaches to estimate the hazard ratio under a null model (true HR = 1.0) and under a protective model (true HR = 0.80). Two models used the entire data set (with and without including mammography as a time-dependent covariate) and two models invoked matching of screened women with unscreened women (with and without excluding of women who had a mammogram after study initiation). For each of the four analytic methods, we compared the observed hazard ratio with the true hazard ratio. We considered an analytic method to be valid if the observed hazard ratio was close to the true hazard ratio.

Results

Two simple analytic methods generated biased results that led to spurious protective effects observed when none was there. The least biased method was based on matching screened and unscreened women and which emulated a randomized trial design, wherein the unexposed control had no mammogram prior to study entry, but she was not excluded or censored if she had a mammogram after the index date.

Conclusion

There is no single ideal method to analyze observational data to evaluate the effectiveness of screening mammography (ie, which generates an unbiased estimates of the underlying hazard ratio) but designs which emulate randomised trials should be promoted.