Rachel N. Grob’s research while affiliated with University of Wisconsin–Madison and other places

Introduction As part of building a platform for epidemiological research on diagnostic errors and problems that centers on patients and equity, this paper summarizes the development and analysis of data collected from fielding a survey in a nationally representative U.S. population to explore the prevalence and harm consequences of diagnostic problems or mistakes (referred to here as “diagnostic P&Ms”) by respondent-reported sociodemographic characteristics. Methods We applied narrative elicitation methods to enhance the rigor of implementing a novel survey about diagnostic experiences. We conducted a U.S. population-based survey of a nationally representative sample in 2022–2023, drawn from the NORC AmeriSpeak® panel. We conducted multivariate regression analysis at the household level and in a patient subsample to explore sociodemographic predictors of diagnostic P&Ms and related outcomes in the aftermath. Results The comparative analysis by sociodemographic characteristics estimates prevalence of diagnostic P&Ms, prevalence of persisting harms, rate of respondent-reported perceptions of personal attribute adversely affecting diagnosis, and concern about future diagnostic P&Ms. Outcome estimates ranged from about 4% (concern about future diagnostic P&M) to 38% (at least one P&M in households during the past 4 years). Several sociodemographic groups experienced statistically significant higher levels of risk for these outcomes, with some at greater than twice the odds compared to reference groups—transgender and gender independent individuals (e.g., 5 + −fold odds of expectation of future P&M compared to cis-males), cis-females (e.g., greater than 1.5 odds of persistent physical and emotional harms compared to cis-males), low household income (e.g., twice the likelihood of multiple P&Ms for incomes under $60 K compared to$100 K+ households), younger age (3-fold odds of at least one diagnostic P&M for those under 25 years old compared to those aged 45–54), multiracial individuals (about twice the odds of diagnostic P&Ms compared to non-Hispanic White), and disabled and unable to work full-time (more than twice the likelihood of perceiving that a personal attribute impaired diagnosis compared to those with other work status designations). Discussion This new survey and accompanying data source facilitate an enriched exploration of the patterns of diagnostic disparities and points of leverage through which diagnostic experiences can be made more equitable.

Objective: To evaluate the measurement properties of a set of six items designed to elicit narrative accounts of pediatric inpatient experience. Data sources: Data came from 163 participants recruited from a probability-based online panel of U.S. adults. Participants were family members of a child who had an overnight hospital stay in the past 12 months. Study design: Cross-sectional survey with follow-up phone interviews. Data collection/extraction methods: Participants completed an online (n = 129) or phone (n = 34) survey about their child's hospitalization experience. The survey contained closed-ended items from the Child Hospital Consumer Assessment of Healthcare Providers and Systems (Child HCAHPS) survey, followed by the six narrative items. Approximately 2 weeks after completing the survey, 47 participants additionally completed a one-hour, semi-structured phone interview, the results of which served as a "gold standard" for evaluating the fidelity of narrative responses. Qualitative content analysis was used to code narrative and interview responses for domains of patient experience and actionability. Principal findings: The average narrative was 248 words (SD = 319). Seventy-nine percent of narratives mentioned a topic included in the Child HCAHPS survey; 89% mentioned a topic not covered by that survey; and 75% included at least one detailed description of an actionable event. Overall, there was 66% correspondence between narrative and interview responses. Correspondence was higher on the phone than in the online condition (75% vs. 59%). Conclusions: Narratives elicited from rigorously designed multi-item sets can provide detailed, substantive information about pediatric inpatient experiences that hospitals could use to improve child and family experiences during pediatric hospitalization. They add context to closed-ended survey item responses and provide information about experiences of care important to children and families that are not included in quantitative surveys.