Guiding Principles to Address the Impact of Algorithm Bias on Racial and Ethnic Disparities in Health and Health Care.

Importance: Health care algorithms are used for diagnosis, treatment, prognosis, risk stratification, and allocation of resources. Bias in the development and use of algorithms can lead to worse outcomes for racial and ethnic minoritized groups and other historically marginalized populations such as individuals with lower income. Objective: To provide a conceptual framework and guiding principles for mitigating and preventing bias in health care algorithms to promote health and health care equity. Evidence Review: The Agency for Healthcare Research and Quality and the National Institute for Minority Health and Health Disparities convened a diverse panel of experts to review evidence, hear from stakeholders, and receive community feedback. Findings: The panel developed a conceptual framework to apply guiding principles across an algorithm's life cycle, centering health and health care equity for patients and communities as the goal, within the wider context of structural racism and discrimination. Multiple stakeholders can mitigate and prevent bias at each phase of the algorithm life cycle, including problem formulation (phase 1); data selection, assessment, and management (phase 2); algorithm development, training, and validation (phase 3); deployment and integration of algorithms in intended settings (phase 4); and algorithm monitoring, maintenance, updating, or deimplementation (phase 5). Five principles should guide these efforts: (1) promote health and health care equity during all phases of the health care algorithm life cycle; (2) ensure health care algorithms and their use are transparent and explainable; (3) authentically engage patients and communities during all phases of the health care algorithm life cycle and earn trustworthiness; (4) explicitly identify health care algorithmic fairness issues and trade-offs; and (5) establish accountability for equity and fairness in outcomes from health care algorithms. Conclusions and Relevance: Multiple stakeholders must partner to create systems, processes, regulations, incentives, standards, and policies to mitigate and prevent algorithmic bias. Reforms should implement guiding principles that support promotion of health and health care equity in all phases of the algorithm life cycle as well as transparency and explainability, authentic community engagement and ethical partnerships, explicit identification of fairness issues and trade-offs, and accountability for equity and fairness.

Green Carts (mobile produce vendors) in the Bronx--optimally positioned to meet neighborhood fruit-and-vegetable needs?

Poor access to fresh produce likely contributes to disparities in obesity and diet-related diseases in the Bronx. New York City's Green Cart program is a partial response to the problem. We evaluated this program (permitting street vendors to sell fresh produce) by canvassing the Bronx for carts, interviewing vendors, and analyzing their locations and food offerings. Green Carts were clustered in areas of probable high pedestrian traffic, covering only about 57% of needy areas by liberal estimates. Some carts sold outside allowed boundaries; a few sold sugary snacks. Vendor locations and their food offerings suggest possible areas for program improvement.

Leveraging Social Media to Promote Evidence-Based Continuing Medical Education.

IMPORTANCE: New dissemination methods are needed to engage physicians in evidence-based continuing medical education (CME). OBJECTIVE: To examine the effectiveness of social media in engaging physicians in non-industry-sponsored CME. DESIGN: We tested the effect of different media platforms (e-mail, Facebook, paid Facebook and Twitter), CME topics, and different "hooks" (e.g., Q&A, clinical pearl and best evidence) on driving clicks to a landing site featuring non-industry sponsored CME. We modelled the effects of social media platform, CME topic, and hook using negative binomial regression on clicks to a single landing site. We used clicks to landing site adjusted for exposure and message number to calculate rate ratios. To understand how physicians interact with CME content on social media, we also conducted interviews with 10 physicians. SETTING: The National Physicians Alliance (NPA) membership. PARTICIPANTS: NPA e-mail recipients, Facebook followers and friends, and Twitter followers. MAIN OUTCOMES AND MEASURES: Clicks to the NPA's CME landing site. RESULTS: On average, 4,544 recipients received each message. Messages generated a total of 592 clicks to the landing site, for a rate of 5.4 clicks per 1000 recipients exposed. There were 5.4 clicks from e-mail, 11.9 clicks from Facebook, 5.5 clicks from paid Facebook, and 6.9 clicks from Twitter to the landing site for 1000 physicians exposed to each of 4 selected CME modules. A Facebook post generated 2.3x as many clicks to the landing site as did an e-mail after controlling for participant exposure, hook type and CME topic (p<0.001). Twitter posts (p = 0.13) and paid Facebook posts (p = 0.06) were not statistically different from e-mail in generating clicks to the landing site. Use of different hooks to engage physicians had no impact on clicks to the landing site. Interviews with physicians suggested that social media might not be a preferred vehicle for disseminating CME. CONCLUSIONS: Social media has a modest impact on driving traffic to evidence-based CME options. Facebook had a superior effect on driving physician web traffic to evidence-based CME compared to other social media platforms and email.

Exploring family and social context through the electronic health record: Physicians' experiences.

INTRODUCTION: There is ample evidence that social and familial context significantly impacts health. However, family and social history templates typically used in clinical practice exclude prompts to explore important contextual information, such as family dynamics, health beliefs, housing, and neighborhood environment. METHOD: At the Residency Program in Social Medicine at Montefiore Medical Center/Albert Einstein College of Medicine in Bronx, NY, we developed and piloted an expanded family and social information (FSI) template in our electronic health record (EHR) system. After 10 physicians used the FSI template during routine clinical practice, we conducted and qualitatively analyzed transcripts of semistructured interviews to assess their experiences. RESULTS: The major themes of physician's experiences using the FSI template included: expanded thinking (promoted thinking and discussion about contextual information), relevancy to care (highlighted important life events, helpful resource for future care or for team-based care), fragmentation (patient narrative difficult to categorize into discrete sections), and abstract phrasing (prompts too abstract and/or too complex to explore well). Some minor themes (themes that were repeated often, but not consistent in all interviews) were also noted: practice demands, educational purpose, and wish list. DISCUSSION: Within our small sample, we found that an expanded family and social information template broadened physicians' clinical thinking during routine visits. We also found that the structure of the EHR and practice demands create significant barriers to exploring a patient's narrative effectively. Themes identified in this research will inform national efforts to prioritize a contextual and biopsychosocial perspective in EHR systems. (PsycINFO Database Record