Paper Session 6
AbstractAsk a Question?
The Effect of Differential Victim Crime Reporting on Predictive Policing Systems
Police departments around the world have been experimenting with forms of place-based data-driven proactive policing for over two decades. Modern incarnations of such systems are commonly known as hot spot predictive policing. These systems predict where future crime is likely to concentrate such that police can allocate patrols to these areas and deter crime before it occurs. Previous research on fairness in predictive policing has concentrated on the feedback loops which occur when models are trained on discovered crime data, but has limited implications for models trained on victim crime reporting data. We demonstrate how differential victim crime reporting rates across geographical areas can lead to oUCTome disparities in common crime hot spot prediction models. Our analysis is based on a simulation patterned after district-level victimization and crime reporting survey data for Bogotá, Colombia. Our results suggest that differential crime reporting rates can lead to a displacement of predicted hotspots from high crime but low reporting areas to high or medium crime and high reporting areas. This may lead to misallocations both in the form of over-policing and under-policing.
Algorithmic Fairness in Predicting Opioid Use Disorder using Machine Learning
There has been recent interest by payers, health care systems, and researchers in the development of machine learning and artificial intelligence models that predict an individual’s probability of developing opioid use disorder. The scores generated by these algorithms can be used by physicians to tailor the prescribing of opioids for the treatment of pain, reducing or foregoing prescribing to individuals deemed to be at high risk, or increasing prescribing for patients deemed to be at low risk. This paper constructs a machine learning algorithm to predict opioid use disorder risk using commercially available claims data similar to those utilized in the development of proprietary opioid use disorder prediction algorithms. We study risk scores generated by the machine learning model in a setting with quasi-experimental variation in the likelihood that doctors prescribe opioids, generated by changes in the legal structure for monitoring physician prescribing. We find that machine-predicted risk scores do not appear to correlate at all with the individual-specific heterogeneous treatment effect of receiving opioids. The paper identifies a new source of algorithmic unfairness in machine learning applications for health care and precision medicine, arising from the researcher's choice of objective function. While precision medicine should guide physician treatment decisions based on the heterogeneous causal impact of a course of treatment for an individual, allocating treatments to individuals receiving the most benefit and recommending caution for those most likely to experience harmful side effects, ML models in health care are often trained on proxies like individual baseline risk, and are not necessarily informative in deciding who would most benefit, or be harmed, by a course of treatment.
A Bayesian Model of Cash Bail Decisions
The use of cash bail as a mechanism for detaining defendants pre-trial is an often-criticized system that many have argued violates the presumption of ``innocent until proven guilty.'' Many studies have sought to understand both the long-term effects of cash bail's use and the disparate rate of cash bail assignments along demographic lines (race, gender, etc). However, such work is often susceptible to problems of infra-marginality -- that the data we observe can only describe average outcomes, and not the outcomes associated with the marginal decision. In this work, we address this problem by creating a hierarchical Bayesian model of cash bail assignments. Specifically, our approach models cash bail decisions as a probabilistic process whereby judges balance the relative costs of assigning cash bail with the cost of defendants potentially skipping court dates, and where these skip probabilities are estimated based upon features of the individual case. We then use Monte Carlo inference to sample the distribution over these costs for different magistrates and across different races. We fit this model to a data set we have collected of over 50,000 court cases in the Allegheny and Philadelphia counties in Pennsylvania. Our analysis of 50 separate judges shows that they are uniformly more likely to assign cash bail to black defendants than to white defendants, even given identical likelihood of skipping a court appearance. This analysis raises further questions about the equity of the practice of cash bail, irrespective of its underlying legal justification.