"The human body is a black box"

Abstract

Machine learning technologies are increasingly developed for use in healthcare. While research communities have focused on creating state-of-the-art models, there has been less focus on real world implementation and the associated challenges to fairness, transparency, and accountability that come from actual, situated use. Serious questions remain underexamined regarding how to ethically build models, interpret and explain model output, recognize and account for biases, and minimize disruptions to professional expertise and work cultures. We address this gap in the literature and provide a detailed case study covering the development, implementation, and evaluation of Sepsis Watch, a machine learning-driven tool that assists hospital clinicians in the early diagnosis and treatment of sepsis. Sepsis is a severe infection that can lead to organ failure or death if not treated in time and is the leading cause of inpatient deaths in US hospitals. We, the team that developed and evaluated the tool, discuss our conceptualization of the tool not as a model deployed in the world but instead as a socio-technical system requiring integration into existing social and professional contexts. Rather than focusing solely on model interpretability to ensure fair and accountable machine learning, we point toward four key values and practices that should be considered when developing machine learning to support clinical decision-making: rigorously define the problem in context, build relationships with stakeholders, respect professional discretion, and create ongoing feedback loops with stakeholders. Our work has significant implications for future research regarding mechanisms of institutional accountability and considerations for responsibly designing machine learning systems. Our work underscores the limits of model interpretability as a solution to ensure transparency, accuracy, and accountability in practice. Instead, our work demonstrates other means and goals to achieve FATML values in design and in practice.

Integrating a Machine Learning System Into Clinical Workflows: Qualitative Study (Preprint)

“…We also identified broader strategies to build trust and accountability in machine learning tools, such as educating clinicians on model performance and utility within the local context, respecting professional discretion, and engaging end users early and often throughout program design and implementation [30].…”

“…ED physicians were informed about the program in faculty meetings and via email. Both nurses and physicians were educated on the model's aggregate performance measures relative to other methods, and visualizations of individual patient cases were presented to demonstrate how the model could detect sepsis hours before the clinical diagnosis [30]. A full description of the planning and implementation process can be found elsewhere [32].…”

“…Eliciting perspectives directly from clinicians using such models can identify real-world barriers and facilitators impacting implementation efforts. Despite a growing body of literature on both physician perspectives of sepsis CDS implementation [26][27][28] and sepsis machine learning algorithms, more research is needed to understand clinician views on black box machine learning models that do not explain their predictions in a way that humans can understand [29,30]. To address this gap in the literature, this study identifies factors that affect the integration of a black box sepsis machine learning system into the workflows of frontline clinicians.…”

High Tech, High Risk: Tech Ethics Lessons for the COVID-19 Pandemic Response

“…Recent research demonstrates the importance of recognizing the additional work integrating new tools into the existing practices of workplaces they are intended for requires from all those who interact with them. 67 …”

“…Technologists should ask themselves: if an automated tool for tracking and treating a pandemic is not useful for the most vulnerable, then in what sense is it useful enough to merit investment and justify the effort of integrating it into existing practices? 67 …”