Recent scrutiny of artificial intelligence (AI)-based facial recognition software has renewed concerns about the un-intendedeffectsofAIonsocialbiasandinequity.Academic and government officials have raised concerns over racial and gender bias in several AI-based technologies, including internet search engines and algorithms to predict risk ofcriminalbehavior.CompanieslikeIBMandMicrosofthave made public commitments to "de-bias" their technologies, whereas Amazon mounted a public campaign criticizing such research. As AI applications gain traction in medicine, cliniciansandhealthsystemleadershaveraisedsimilarconcerns over automating and propagating existing biases. 1 But is AI the problem? Or can it be part of the solution? Whilepotentiallyinadvertentlycontributingtobias,AItechnologies, when used responsibly, may also help counteract the risk of bias in unique ways. Using AI to identify bias in health care may help identify interventions that could help correct biased clinician decision-making and possibly reduce health disparities.
IMPORTANCE Medicare launched the mandatory Comprehensive Care for Joint Replacement bundled payment model in 67 urban areas for approximately 800 hospitals following its experience in the voluntary Acute Care Episodes (ACE) and Bundled Payments for Care Improvement (BPCI) demonstration projects. Little information from ACE and BPCI exists to guide hospitals in redesigning care for mandatory joint replacement bundles. OBJECTIVE To analyze changes in quality, internal hospital costs, and postacute care (PAC) spending for lower extremity joint replacement bundled payment episodes encompassing hospitalization and 30 days of PAC. DESIGN, SETTING, AND PARTICIPANTS This observational study followed 3942 total patients with lower extremity joint replacement at Baptist Health System (BHS), which participated in ACE and BPCI. EXPOSURES Lower extremity joint replacement surgery under bundled payment at BHS. MAIN OUTCOMES AND MEASURES Average Medicare payments per episode, readmissions, emergency department visits, prolonged length of stay, and hospital savings from changes in internal hospital costs and PAC spending. RESULTS Overall, 3942 patients (mean [SD] age, 72.4 [8.4] years) from BHS were observed.
Key PointsQuestionCan machine learning algorithms identify oncology patients at risk of short-term mortality to inform timely conversations between patients and physicians regrading serious illness?FindingsIn this cohort study of 26 525 patients seen in oncology practices within a large academic health system, machine learning algorithms accurately identified patients at high risk of 6-month mortality with good discrimination and positive predictive value. When the gradient boosting algorithm was applied in real time, most patients who were classified as having high risk were deemed appropriate by oncology clinicians for a conversation regarding serious illness.MeaningIn this study, machine learning algorithms accurately identified patients with cancer who were at risk of 6-month mortality, suggesting that these models could facilitate more timely conversations between patients and physicians regarding goals and values.
Behavioral economics provides insights about the development of effective incentives for physicians to deliver high-value care. It suggests that the structure and delivery of incentives can shape behavior, as can thoughtful design of the decision-making environment. This article discusses several principles of behavioral economics, including inertia, loss aversion, choice overload, and relative social ranking. Whereas these principles have been applied to motivate personal health decisions, retirement planning, and savings behavior, they have been largely ignored in the design of physician incentive programs. Applying these principles to physician incentives can improve their effectiveness through better alignment with performance goals. Anecdotal examples of successful incentive programs that apply behavioral economics principles are provided, even as the authors recognize that its application to the design of physician incentives is largely untested, and many outstanding questions exist. Application and rigorous evaluation of infrastructure changes and incentives are needed to design payment systems that incentivize high-quality, cost-conscious care.
Objective
To evaluate the prevalence of seven social factors using physician notes as compared to claims and structured electronic health records (EHRs) data and the resulting association with 30‐day readmissions.
Study Setting
A multihospital academic health system in southeastern Massachusetts.
Study Design
An observational study of 49,319 patients with cardiovascular disease admitted from January 1, 2011, to December 31, 2013, using multivariable logistic regression to adjust for patient characteristics.
Data Collection/Extraction Methods
All‐payer claims, EHR data, and physician notes extracted from a centralized clinical registry.
Principal Findings
All seven social characteristics were identified at the highest rates in physician notes. For example, we identified 14,872 patient admissions with poor social support in physician notes, increasing the prevalence from 0.4 percent using ICD‐9 codes and structured EHR data to 16.0 percent. Compared to an 18.6 percent baseline readmission rate, risk‐adjusted analysis showed higher readmission risk for patients with housing instability (readmission rate 24.5 percent; p < .001), depression (20.6 percent; p < .001), drug abuse (20.2 percent; p = .01), and poor social support (20.0 percent; p = .01).
Conclusions
The seven social risk factors studied are substantially more prevalent than represented in administrative data. Automated methods for analyzing physician notes may enable better identification of patients with social needs.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.