Developer Perspectives on Potential Harms of Machine Learning Predictive Analytics in Health Care: Qualitative Analysis.

BACKGROUND: Machine learning predictive analytics (MLPA) is increasingly used in health care to reduce costs and improve efficacy; it also has the potential to harm patients and trust in health care. Academic and regulatory leaders have proposed a variety of principles and guidelines to address the challenges of evaluating the safety of machine learning-based software in the health care context, but accepted practices do not yet exist. However, there appears to be a shift toward process-based regulatory paradigms that rely heavily on self-regulation. At the same time, little research has examined the perspectives about the harms of MLPA developers themselves, whose role will be essential in overcoming the "principles-to-practice" gap. OBJECTIVE: The objective of this study was to understand how MLPA developers of health care products perceived the potential harms of those products and their responses to recognized harms. METHODS: We interviewed 40 individuals who were developing MLPA tools for health care at 15 US-based organizations, including data scientists, software engineers, and those with mid- and high-level management roles. These 15 organizations were selected to represent a range of organizational types and sizes from the 106 that we previously identified. We asked developers about their perspectives on the potential harms of their work, factors that influence these harms, and their role in mitigation. We used standard qualitative analysis of transcribed interviews to identify themes in the data. RESULTS: We found that MLPA developers recognized a range of potential harms of MLPA to individuals, social groups, and the health care system, such as issues of privacy, bias, and system disruption. They also identified drivers of these harms related to the characteristics of machine learning and specific to the health care and commercial contexts in which the products are developed. MLPA developers also described strategies to respond to these drivers and potentially mitigate the harms. Opportunities included balancing algorithm performance goals with potential harms, emphasizing iterative integration of health care expertise, and fostering shared company values. However, their recognition of their own responsibility to address potential harms varied widely. CONCLUSIONS: Even though MLPA developers recognized that their products can harm patients, public, and even health systems, robust procedures to assess the potential for harms and the need for mitigation do not exist. Our findings suggest that, to the extent that new oversight paradigms rely on self-regulation, they will face serious challenges if harms are driven by features that developers consider inescapable in health care and business environments. Furthermore, effective self-regulation will require MLPA developers to accept responsibility for safety and efficacy and know how to act accordingly. Our results suggest that, at the very least, substantial education will be necessary to fill the "principles-to-practice" gap.

Genomics Research with Undiagnosed Children: Ethical Challenges at the Boundaries of Research and Clinical Care.

OBJECTIVE: To explore the perspectives of parents of undiagnosed children enrolled in genomic diagnosis research regarding their motivations for enrolling their children, their understanding of the potential burdens and benefits, and the extent to which their experiences ultimately aligned with or diverged from their original expectations. STUDY DESIGN: In-depth interviews were conducted with parents, audio-recorded and transcribed. A structured codebook was applied to each transcript, after which iterative memoing was used to identify themes. RESULTS: Fifty-four parents participated, including 17 (31.5%) whose child received a diagnosis through research. Themes describing parents' expectations and experiences of genomic diagnosis research included (1) the extent to which parents' motivations for participation focused on their hope that it would directly benefit their child, (2) the ways in which parents' frustrations regarding the research process confused the dual clinical and research goals of their participation, and (3) the limited clinical benefits parents ultimately experienced for their children. CONCLUSIONS: Our results suggest that parents of undiagnosed children seeking enrollment in genomic diagnosis research are at risk of a form of therapeutic misconception-in this case, diagnostic misconception. These findings indicate the need to examine the processes and procedures associated with this research to communicate appropriately and balance the potential burdens and benefits of study participation.

Demographic Characteristics Associated With Perceptions of Personal Utility in Genetic and Genomic Testing: A Systematic Review.

Importance: The expansion of genetic and genomic testing in health care has led to recognition that these tests provide personal as well as clinical utility to patients and families. However, available systematic reviews on this topic have not reported the demographic backgrounds of participants in studies of personal utility, leaving generalizability unclear. Objective: To determine the demographic characteristics of participants in studies examining the personal utility of genetic and genomic testing in health care. Evidence Review: For this systematic review, we utilized and updated the results of a highly cited 2017 systematic review on the personal utility of genetics and genomics, which identified relevant articles published between January 1, 2003, and August 4, 2016. We also used the original methods to update this bibliography with literature published subsequently up to January 1, 2022. Studies were screened for eligibility by 2 independent reviewers. Eligible studies reported empirical data on the perspectives of patients, family members, and/or the general public in the US on the personal utility of any type of health-related genetic or genomic test. We utilized a standardized codebook to extract study and participant characteristics. We summarized demographic characteristics descriptively across all studies and by subgroup based on study and participant characteristics. Findings: We included 52 studies with 13 251 eligible participants. Sex or gender was the most frequently reported demographic characteristic (48 studies [92.3%]), followed by race and ethnicity (40 studies [76.9%]), education (38 studies [73.1%]), and income (26 studies [50.0%]). Across studies, participants disproportionately were women or female (mean [SD], 70.8% [20.5%]), were White (mean [SD], 76.1% [22.0%]), had a college degree or higher (mean [SD], 64.5% [19.9%]), and reported income above the US median (mean [SD], 67.4% [19.2%]). Examination of subgroups of results by study and participant characteristics evidenced only small shifts in demographic characteristics. Conclusions and Relevance: This systematic review examined the demographic characteristics of individual participants in studies of the personal utility of health-related genetic and genomic testing in the US. The results suggest that participants in these studies were disproportionately White, college-educated women with above-average income. Understanding the perspectives of more diverse individuals regarding the personal utility of genetic and genomic testing may inform barriers to research recruitment and uptake of clinical testing in currently underrepresented populations.

Rare Disease, Advocacy and Justice: Intersecting Disparities in Research and Clinical Care.

Rare genetic diseases collectively impact millions of individuals in the United States. These patients and their families share many challenges including delayed diagnosis, lack of knowledgeable providers, and limited economic incentives to develop new therapies for small patient groups. As such, rare disease patients and families often must rely on advocacy, including both self-advocacy to access clinical care and public advocacy to advance research. However, these demands raise serious concerns for equity, as both care and research for a given disease can depend on the education, financial resources, and social capital available to the patients in a given community. In this article, we utilize three case examples to illustrate ethical challenges at the intersection of rare diseases, advocacy and justice, including how reliance on advocacy in rare disease may drive unintended consequences for equity. We conclude with a discussion of opportunities for diverse stakeholders to begin to address these challenges.

Patient Perspectives of Health System Barriers to Accessing Care for Hidradenitis Suppurativa: A Qualitative Study.

Importance: Patient-perceived barriers to hidradenitis suppurativa (HS) care are poorly understood. Understanding health care barriers is a critical first step toward improving care for this population. Objective: To characterize the health care experiences of people living with HS, including perceived barriers and facilitators to health care access, and to elucidate potential associations among these barriers and facilitators, health care access, and disease activity. Design, Setting, and Participants: In this qualitative study, an inductive thematic analysis was conducted on 45 in-depth, 60- to 90-minute semistructured interviews of 45 people with HS from diverse sociodemographic backgrounds that took place between March and April 2020. Individuals were eligible if they could speak English, were 18 years or older, and were diagnosed with HS. A diagnosis of HS was confirmed through physician diagnosis or through self-reported, affirmative response to the validated screening question, "Do you experience boils in your armpits or groin that recur at least every six months?" Main Outcomes and Measures: Interviews were audio recorded and transcribed verbatim. A modified grounded theory approach was used to develop the codebook, which investigators used for inductive thematic analysis. Results: Among the 45 participants included, the median (IQR) age was 37 (16) years, 33 (73%) were female, and 22 (49%) were White. There were 6 interrelated themes associated with participant-perceived barriers to accessing HS care: (1) bidirectional associations of disease activity and employment, (2) association of employment with health care coverage, (3) association of health care coverage with costs and perceived access to care, (4) association of costs with access to patient-centered care, (5) health care professional attitudes and knowledge influence patient-centered care and perceived access to care and disease activity, and (6) health system characteristics influence patient-centered care and associated costs, perceived access to care, and disease activity. Conclusions and Relevance: This qualitative study highlights themes that generate a conceptual model for understanding barriers that may act synergistically to limit health care access and influence disease activity. The disease activity of HS may be reduced when cycle elements are optimized. This study also highlights areas for future investigations and potential systems-level changes to improve access to patient-centered HS care.

Not in my AI: Moral engagement and disengagement in health care AI development.

Machine learning predictive analytics (MLPA) are utilized increasingly in health care, but can pose harms to patients, clinicians, health systems, and the public. The dynamic nature of this technology creates unique challenges to evaluating safety and efficacy and minimizing harms. In response, regulators have proposed an approach that would shift more responsibility to MLPA developers for mitigating potential harms. To be effective, this approach requires MLPA developers to recognize, accept, and act on responsibility for mitigating harms. In interviews of 40 MLPA developers of health care applications in the United States, we found that a subset of ML developers made statements reflecting moral disengagement, representing several different potential rationales that could create distance between personal accountability and harms. However, we also found a different subset of ML developers who expressed recognition of their role in creating potential hazards, the moral weight of their design decisions, and a sense of responsibility for mitigating harms. We also found evidence of moral conflict and uncertainty about responsibility for averting harms as an individual developer working in a company. These findings suggest possible facilitators and barriers to the development of ethical ML that could act through encouragement of moral engagement or discouragement of moral disengagement. Regulatory approaches that depend on the ability of ML developers to recognize, accept, and act on responsibility for mitigating harms might have limited success without education and guidance for ML developers about the extent of their responsibilities and how to implement them.

Perspectives of Rare Disease Social Media Group Participants on Engaging With Genetic Counselors: Mixed Methods Study.

BACKGROUND: Social media provides a potential avenue for genetic counselors to address gaps in access to reliable genetics information for rare disease communities. However, only limited research has examined patient and family attitudes toward engaging with genetic counselors through social media. OBJECTIVE: Our study assessed the attitudes of members of rare disease social media groups toward engaging with genetic counselors through social media, characteristics associated with greater interest, and the benefits and potential pitfalls of various approaches to such engagement. METHODS: We conducted a mixed methods survey of patients and family members recruited from a systematic sample of rare disease Facebook groups. Patient characteristics and their associations with interest in engagement with genetic counselors were evaluated using univariate and bivariate statistics. Responses to open-ended questions were analyzed using thematic content analysis. RESULTS: In total, 1053 individuals from 103 rare disease groups participated. The median overall interest in engaging with genetic counselors on social media was moderately high at 7.0 (IQR 4.0-9.0, range 0-10). No past experience with a genetic counselor was associated with greater interest in engaging with one through social media (µ=6.5 vs 6.0, P=.04). Participants expressed greatest interest (median 9.0, IQR 5.0-10.0) in engagement models allowing direct communication with genetic counselors, which was corroborated by the majority (n=399, 61.3%) of individuals who responded to open-ended questions explicitly stating their interest in 1-on-1 interactions. When asked what forms of support they would request from genetic counselors through social media, participants desired individualized support and information about how to access services. However, participants also expressed concerns regarding privacy and confidentiality. CONCLUSIONS: Patients and family members in rare disease social media groups appear interested in engaging with genetic counselors through social media, particularly for individualized support. This form of engagement on social media is not meant to replace the current structure and content of genetic counseling (GC) services, but genetic counselors could more actively use social media as a communication tool to address gaps in knowledge and awareness about genetics services and gaps in accessible patient information. Although encouraging, concerns regarding privacy and feasibility require further consideration, pointing to the need for professional guidelines in this area.

Beyond race: Recruitment of diverse participants in clinical genomics research for rare disease.

Purpose: Despite recent attention to increasing diversity in clinical genomics research, researchers still struggle to recruit participants from varied sociodemographic backgrounds. We examined the experiences of parents from diverse backgrounds with enrolling their children in clinical genomics research on rare diseases. We explored the barriers and facilitators parents encountered and possible impacts of sociodemographic factors on their access to research. Methods: We utilized semi-structured interviews with parents of children participating in the Undiagnosed Diseases Network. Interview data were analyzed using comparative content analysis. Results: We interviewed 13 Hispanic, 11 non-Hispanic White, four Asian, and two biracial parents. Participants discussed different pathways to clinical genomics research for rare disease as well as how sociodemographic factors shaped families' access. Themes focused on variation in: 1) reliance on providers to access research; 2) cultural norms around health communication; 3) the role of social capital in streamlining access; and 4) the importance of language-concordant research engagement. Conclusion: Our findings suggest that variables beyond race/ethnicity may influence access in clinical genomics research. Future efforts to diversify research participation should consider utilizing varied recruitment strategies to reach participants with diverse sociodemographic characteristics.

Perceived utility and disutility of genomic sequencing for pediatric patients: Perspectives from parents with diverse sociodemographic characteristics.

Given the limited therapeutic options for most rare diseases diagnosed through genomic sequencing (GS) and the proportion of patients who remain undiagnosed even after GS, it is important to characterize a broader range of benefits and potential harms of GS from the perspectives of families with diverse sociodemographic characteristics. We recruited parents of children enrolled in the Undiagnosed Diseases Network. Parents completed an in-depth interview, and we conducted a comparative content analysis of the data. Parents (n = 30) were demographically diverse, with 43.3% identifying as Hispanic, 33.3% primarily Spanish-speaking, and widely variable household income and education. Parents reported minimal changes in their child's health status following GS but did report a range of other forms of perceived utility, including improvements in their child's healthcare management and access, in their own psychological well-being, and in disease-specific social connections and research opportunities. Parents who received a diagnosis more frequently perceived utility across all domains; however, disutility also was reported by both those with and without a diagnosis. Impacts depended on multiple mediating factors, including parents' underlying expectations and beliefs, family sociodemographic characteristics, individual disease characteristics, and prior healthcare access. Our study suggests that the perceived utility of GS varies widely among parents and may depend on multiple individual, sociodemographic, and contextual factors that are relevant for pre- and post-GS counseling, for value assessment of GS, and for policymaking related to access to new genomic technologies.

Opportunities and pitfalls of social media research in rare genetic diseases: a systematic review.

PURPOSE: Social media may be particularly valuable in research in rare genetic diseases because of the low numbers of patients and the rare disease community's robust online presence. The goal of this systematic review was to understand how social media is currently used in rare disease research and the characteristics of the participants in these studies. METHODS: We conducted a systematic review of six databases to identify studies published in English between January 2004 and November 2020, of which 120 met inclusion criteria. RESULTS: Most studies were observational (n = 114, 95.0%) and cross-sectional (n = 107, 89.2%), and more than half (n = 69, 57.5%) utilized only surveys. Only 101 rare diseases were included across all studies. Participant demographics, when reported, were predominantly female (70.1% ± 22.5%) and white (85.0% ± 11.0%) adult patients and caregivers. CONCLUSION: Despite its potential benefits in rare disease research, the use of social media is still methodologically limited and the participants reached may not be representative of the rare disease population by gender, race, age, or rare disease type. As scholars explore using social media for rare disease research, careful attention should be paid to representativeness when studying this diverse patient community.

A Typology of Existing Machine Learning-Based Predictive Analytic Tools Focused on Reducing Costs and Improving Quality in Health Care: Systematic Search and Content Analysis.

BACKGROUND: Considerable effort has been devoted to the development of artificial intelligence, including machine learning-based predictive analytics (MLPA) for use in health care settings. The growth of MLPA could be fueled by payment reforms that hold health care organizations responsible for providing high-quality, cost-effective care. Policy analysts, ethicists, and computer scientists have identified unique ethical and regulatory challenges from the use of MLPA in health care. However, little is known about the types of MLPA health care products available on the market today or their stated goals. OBJECTIVE: This study aims to better characterize available MLPA health care products, identifying and characterizing claims about products recently or currently in use in US health care settings that are marketed as tools to improve health care efficiency by improving quality of care while reducing costs. METHODS: We conducted systematic database searches of relevant business news and academic research to identify MLPA products for health care efficiency meeting our inclusion and exclusion criteria. We used content analysis to generate MLPA product categories and characterize the organizations marketing the products. RESULTS: We identified 106 products and characterized them based on publicly available information in terms of the types of predictions made and the size, type, and clinical training of the leadership of the companies marketing them. We identified 5 categories of predictions made by MLPA products based on publicly available product marketing materials: disease onset and progression, treatment, cost and utilization, admissions and readmissions, and decompensation and adverse events. CONCLUSIONS: Our findings provide a foundational reference to inform the analysis of specific ethical and regulatory challenges arising from the use of MLPA to improve health care efficiency.

"Doctors can read about it, they can know about it, but they've never lived with it": How parents use social media throughout the diagnostic odyssey.

Parents of children with undiagnosed conditions struggle to obtain information about how to treat and support their children. It can be particularly challenging to find communities and other parents who share their experiences and can provide emotional and informational support. This study sought to characterize how parents use social media, both throughout the diagnostic odyssey and post-diagnosis, to meet their informational, social, and emotional support needs. We conducted qualitative semi-structured interviews with 14 parents from the Stanford site of the Undiagnosed Diseases Network (UDN), including five whose children had received a diagnosis through study participation. Interview recordings were analyzed using inductive, team-based coding and thematic analysis based in grounded theory using Dedoose qualitative analysis software. Through this process, we identified four key themes related to social media use. First, parents struggled to find the "right" community, often seeking out groups of similar patients based on symptoms or similar conditions. Second, though they found much valuable information through social media about caring for their child, they also struggled to interpret the relevance of the information to their own child's condition. Third, the social support and access to other patients' and families' lived experiences were described as both highly valued and emotionally challenging, particularly in the case of poor outcomes for similar families. Finally, parents expressed the need to balance concerns about their child's privacy with the value of transparency and data sharing for diagnosis. Our results suggest that the needs and experiences of undiagnosed patients and families differ from those with diagnosed diseases and highlight the need for support in best utilizing social media resources at different stages of the diagnostic odyssey.

Age-Related Differences in Experiences With Social Distancing at the Onset of the COVID-19 Pandemic: A Computational and Content Analytic Investigation of Natural Language From a Social Media Survey.

BACKGROUND: As COVID-19 poses different levels of threat to people of different ages, health communication regarding prevention measures such as social distancing and isolation may be strengthened by understanding the unique experiences of various age groups. OBJECTIVE: The aim of this study was to examine how people of different ages (1) experienced the impact of the COVID-19 pandemic and (2) their respective rates and reasons for compliance or noncompliance with social distancing and isolation health guidance. METHODS: We fielded a survey on social media early in the pandemic to examine the emotional impact of COVID-19 and individuals' rates and reasons for noncompliance with public health guidance, using computational and content analytic methods of linguistic analysis. RESULTS: A total of 17,287 participants were surveyed. The majority (n=13,183, 76.3%) were from the United States. Younger (18-31 years), middle-aged (32-44 years and 45-64 years), and older (≥65 years) individuals significantly varied in how they described the impact of COVID-19 on their lives, including their emotional experience, self-focused attention, and topical concerns. Younger individuals were more emotionally negative and self-focused, while middle-aged people were other-focused and concerned with family. The oldest and most at-risk group was most concerned with health-related terms but were lower in anxiety (use of fewer anxiety-related terms) and higher in the use of emotionally positive terms than the other less at-risk age groups. While all groups discussed topics such as acquiring essential supplies, they differentially experienced the impact of school closures and limited social interactions. We also found relatively high rates of noncompliance with COVID-19 prevention measures, such as social distancing and self-isolation, with younger people being more likely to be noncompliant than older people (P<.001). Among the 43.1% (n=7456) of respondents who did not fully comply with health orders, people differed substantially in the reasons they gave for noncompliance. The most common reason for noncompliance was not being able to afford to miss work (n=4273, 57.3%). While work obligations proved challenging for participants across ages, younger people struggled more to find adequate space to self-isolate and manage their mental and physical health; middle-aged people had more concerns regarding childcare; and older people perceived themselves as being able to take sufficient precautions. CONCLUSIONS: Analysis of natural language can provide insight into rapidly developing public health challenges like the COVID-19 pandemic, uncovering individual differences in emotional experiences and health-related behaviors. In this case, our analyses revealed significant differences between different age groups in feelings about and responses to public health orders aimed to mitigate the spread of COVID-19. To improve public compliance with health orders as the pandemic continues, health communication strategies could be made more effective by being tailored to these age-related differences.

The Intersection of Work and Home Challenges Faced by Physician Mothers During the Coronavirus Disease 2019 Pandemic: A Mixed-Methods Analysis.

Objectives: The coronavirus disease 2019 (COVID-19) pandemic has presented extreme challenges for health care workers. This study sought to characterize challenges faced by physician mothers, compare differences in challenges by home and work characteristics, and elicit specific needs and potential solutions. Methods: We conducted a mixed-methods online survey of the Physician Moms Group (PMG) and PMG COVID19 Subgroup on Facebook from April 18th to 29th, 2020. We collected structured data on personal and professional characteristics and qualitative data on home and work concerns. We analyzed qualitative data thematically and used bivariate analyses to evaluate variation in themes by frontline status and children's ages. Results: We included 1,806 participants in analysis and identified 10 key themes. The most frequently identified need/solution was for Community and Government Support (n = 545, 47.1%). When comparing frontline and nonfrontline physicians, those on the frontline more frequently raised concerns about Personal Health and Safety (67.8% vs. 48.4%, p < 0.001), Organizational Communication and Relationships (31.8% vs. 23.8%, p < 0.001), and Family Health and Safety (27.2 vs. 16.6, p < 0.001), while nonfrontline physicians more frequently addressed Patient Care and Safety (56.4% vs. 48.2%, p < 0.001) and Financial/Job Security (33.8% vs. 46.9%, p < 0.001). Participants with an elementary school-aged child more frequently raised concerns about Parenting/Homeschooling (44.0% vs. 31.1%, p < 0.001) and Work/Life Balance (28.4 vs. 13.7, p < 0.001), and participants with a preschool-aged child more frequently addressed Access to Childcare (24.0 vs. 7.7, p < 0.001) and Spouse/Partner Relationships (15.8 vs. 9.5, p < 0.001), when compared to those without children in these age groups. Conclusions: The physician workforce is not homogenous. Health care and government leaders need to understand these diverse challenges in order to meet physicians' professional and family needs during the pandemic.