Assessing racial bias in healthcare predictive models: Practical lessons from an empirical evaluation of 30-day hospital readmission models.

OBJECTIVE: Despite increased availability of methodologies to identify algorithmic bias, the operationalization of bias evaluation for healthcare predictive models is still limited. Therefore, this study proposes a process for bias evaluation through an empirical assessment of common hospital readmission models. The process includes selecting bias measures, interpretation, determining disparity impact and potential mitigations. METHODS: This retrospective analysis evaluated racial bias of four common models predicting 30-day unplanned readmission (i.e., LACE Index, HOSPITAL Score, and the CMS readmission measure applied as is and retrained). The models were assessed using 2.4 million adult inpatient discharges in Maryland from 2016 to 2019. Fairness metrics that are model-agnostic, easy to compute, and interpretable were implemented and apprised to select the most appropriate bias measures. The impact of changing model's risk thresholds on these measures was further assessed to guide the selection of optimal thresholds to control and mitigate bias. RESULTS: Four bias measures were selected for the predictive task: zero-one-loss difference, false negative rate (FNR) parity, false positive rate (FPR) parity, and generalized entropy index. Based on these measures, the HOSPITAL score and the retrained CMS measure demonstrated the lowest racial bias. White patients showed a higher FNR while Black patients resulted in a higher FPR and zero-one-loss. As the models' risk threshold changed, trade-offs between models' fairness and overall performance were observed, and the assessment showed all models' default thresholds were reasonable for balancing accuracy and bias. CONCLUSIONS: This study proposes an Applied Framework to Assess Fairness of Predictive Models (AFAFPM) and demonstrates the process using 30-day hospital readmission model as the example. It suggests the feasibility of applying algorithmic bias assessment to determine optimized risk thresholds so that predictive models can be used more equitably and accurately. It is evident that a combination of qualitative and quantitative methods and a multidisciplinary team are necessary to identify, understand and respond to algorithm bias in real-world healthcare settings. Users should also apply multiple bias measures to ensure a more comprehensive, tailored, and balanced view. The results of bias measures, however, must be interpreted with caution and consider the larger operational, clinical, and policy context.

Making Science Computable Using Evidence-Based Medicine on Fast Healthcare Interoperability Resources: Standards Development Project.

BACKGROUND: Evidence-based medicine (EBM) has the potential to improve health outcomes, but EBM has not been widely integrated into the systems used for research or clinical decision-making. There has not been a scalable and reusable computer-readable standard for distributing research results and synthesized evidence among creators, implementers, and the ultimate users of that evidence. Evidence that is more rapidly updated, synthesized, disseminated, and implemented would improve both the delivery of EBM and evidence-based health care policy. OBJECTIVE: This study aimed to introduce the EBM on Fast Healthcare Interoperability Resources (FHIR) project (EBMonFHIR), which is extending the methods and infrastructure of Health Level Seven (HL7) FHIR to provide an interoperability standard for the electronic exchange of health-related scientific knowledge. METHODS: As an ongoing process, the project creates and refines FHIR resources to represent evidence from clinical studies and syntheses of those studies and develops tools to assist with the creation and visualization of FHIR resources. RESULTS: The EBMonFHIR project created FHIR resources (ie, ArtifactAssessment, Citation, Evidence, EvidenceReport, and EvidenceVariable) for representing evidence. The COVID-19 Knowledge Accelerator (COKA) project, now Health Evidence Knowledge Accelerator (HEvKA), took this work further and created FHIR resources that express EvidenceReport, Citation, and ArtifactAssessment concepts. The group is (1) continually refining FHIR resources to support the representation of EBM; (2) developing controlled terminology related to EBM (ie, study design, statistic type, statistical model, and risk of bias); and (3) developing tools to facilitate the visualization and data entry of EBM information into FHIR resources, including human-readable interfaces and JSON viewers. CONCLUSIONS: EBMonFHIR resources in conjunction with other FHIR resources can support relaying EBM components in a manner that is interoperable and consumable by downstream tools and health information technology systems to support the users of evidence.

Trends and opportunities in computable clinical phenotyping: A scoping review.

Identifying patient cohorts meeting the criteria of specific phenotypes is essential in biomedicine and particularly timely in precision medicine. Many research groups deliver pipelines that automatically retrieve and analyze data elements from one or more sources to automate this task and deliver high-performing computable phenotypes. We applied a systematic approach based on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines to conduct a thorough scoping review on computable clinical phenotyping. Five databases were searched using a query that combined the concepts of automation, clinical context, and phenotyping. Subsequently, four reviewers screened 7960 records (after removing over 4000 duplicates) and selected 139 that satisfied the inclusion criteria. This dataset was analyzed to extract information on target use cases, data-related topics, phenotyping methodologies, evaluation strategies, and portability of developed solutions. Most studies supported patient cohort selection without discussing the application to specific use cases, such as precision medicine. Electronic Health Records were the primary source in 87.1 % (N = 121) of all studies, and International Classification of Diseases codes were heavily used in 55.4 % (N = 77) of all studies, however, only 25.9 % (N = 36) of the records described compliance with a common data model. In terms of the presented methods, traditional Machine Learning (ML) was the dominant method, often combined with natural language processing and other approaches, while external validation and portability of computable phenotypes were pursued in many cases. These findings revealed that defining target use cases precisely, moving away from sole ML strategies, and evaluating the proposed solutions in the real setting are essential opportunities for future work. There is also momentum and an emerging need for computable phenotyping to support clinical and epidemiological research and precision medicine.

Electronic Health Record-Based Recruitment and Retention and Mobile Health App Usage: Multisite Cohort Study.

BACKGROUND: To address the obesity epidemic, there is a need for novel paradigms, including those that address the timing of eating and sleep in relation to circadian rhythms. Electronic health records (EHRs) are an efficient way to identify potentially eligible participants for health research studies. Mobile health (mHealth) apps offer available and convenient data collection of health behaviors, such as timing of eating and sleep. OBJECTIVE: The aim of this descriptive analysis was to report on recruitment, retention, and app use from a 6-month cohort study using a mobile app called Daily24. METHODS: Using an EHR query, adult patients from three health care systems in the PaTH clinical research network were identified as potentially eligible, invited electronically to participate, and instructed to download and use the Daily24 mobile app, which focuses on eating and sleep timing. Online surveys were completed at baseline and 4 months. We described app use and identified predictors of app use, defined as 1 or more days of use, versus nonuse and usage categories (ie, immediate, consistent, and sustained) using multivariate regression analyses. RESULTS: Of 70,661 patients who were sent research invitations, 1021 (1.44%) completed electronic consent forms and online baseline surveys; 4 withdrew, leaving a total of 1017 participants in the analytic sample. A total of 53.79% (n=547) of the participants were app users and, of those, 75.3% (n=412), 50.1% (n=274), and 25.4% (n=139) were immediate, consistent, and sustained users, respectively. Median app use was 28 (IQR 7-75) days over 6 months. Younger age, White race, higher educational level, higher income, having no children younger than 18 years, and having used 1 to 5 health apps significantly predicted app use (vs nonuse) in adjusted models. Older age and lower BMI predicted early, consistent, and sustained use. About half (532/1017, 52.31%) of the participants completed the 4-month online surveys. A total of 33.5% (183/547), 29.3% (157/536), and 27.1% (143/527) of app users were still using the app for at least 2 days per month during months 4, 5, and 6 of the study, respectively. CONCLUSIONS: EHR recruitment offers an efficient (ie, high reach, low touch, and minimal participant burden) approach to recruiting participants from health care settings into mHealth research. Efforts to recruit and retain less engaged subgroups are needed to collect more generalizable data. Additionally, future app iterations should include more evidence-based features to increase participant use.

What influences the "when" of eating and sleeping?A qualitative interview study.

Timing of eating relative to sleep and endogenous circadian rhythm impacts weight and cardiometabolic health. We used qualitative methods to explore what influences the "when" of eating and sleeping. We conducted 37 one-on-one semi-structured interviews among participants with a body mass index (BMI) ≥ 25 kg/m2 recruited from three internal medicine clinics affiliated with an urban academic hospital. Participants (70.3% Female; 51.4% White; Age range: 21-83 years old) completed measures of social jetlag, physical activity, eating habits, and mobile application use and participated in interviews following a guide developed by the study team. Responses were recorded, transcribed and coded sequentially by two trained researchers using editing-style analysis to identify themes. We identified two main themes, each with subthemes: 1) influences on the "when" of eating and sleeping, with subthemes including social jetlag and being overscheduled, and 2) contextualizing beliefs and perceptions about the "when" of eating and sleeping, with subthemes including perceived recommendations for timing of eating and sleeping, and alignment of behaviors with perceived recommendations. Many participants noted being more flexible in their eating and sleeping times on work-free vs work days. The themes this study identified should be considered when designing interventions that influence the timing of eating and sleeping for weight management.

Patient Recruitment Into a Multicenter Clinical Cohort Linking Electronic Health Records From 5 Health Systems: Cross-sectional Analysis.

BACKGROUND: There is growing interest in identifying and recruiting research participants from health systems using electronic health records (EHRs). However, few studies have described the practical aspects of the recruitment process or compared electronic recruitment methods to in-person recruitment, particularly across health systems. OBJECTIVE: The objective of this study was to describe the steps and efficiency of the recruitment process and participant characteristics by recruitment strategy. METHODS: EHR-based eligibility criteria included being an adult patient engaged in outpatient primary or bariatric surgery care at one of 5 health systems in the PaTH Clinical Research Network and having ≥2 weight measurements and 1 height measurement recorded in their EHR within the last 5 years. Recruitment strategies varied by site and included one or more of the following methods: (1) in-person recruitment by study staff from clinical sites, (2) US postal mail recruitment letters, (3) secure email, and (4) direct EHR recruitment through secure patient web portals. We used descriptive statistics to evaluate participant characteristics and proportion of patients recruited (ie, efficiency) by modality. RESULTS: The total number of eligible patients from the 5 health systems was 5,051,187. Of these, 40,048 (0.8%) were invited to enter an EHR-based cohort study and 1085 were enrolled. Recruitment efficiency was highest for in-person recruitment (33.5%), followed by electronic messaging (2.9%), including email (2.9%) and EHR patient portal messages (2.9%). Overall, 779 (65.7%) patients were enrolled through electronic messaging, which also showed greater rates of recruitment of Black patients compared with the other strategies. CONCLUSIONS: We recruited a total of 1085 patients from primary care and bariatric surgery settings using 4 recruitment strategies. The recruitment efficiency was 2.9% for email and EHR patient portals, with the majority of participants recruited electronically. This study can inform the design of future research studies using EHR-based recruitment.

A method for measuring the effect of certified electronic health record technology on childhood immunization status scores among Medicaid managed care network providers.

OBJECTIVE: To provide a methodology for estimating the effect of U.S.-based Certified Electronic Health Records Technology (CEHRT) implemented by primary care physicians (PCPs) on a Healthcare Effectiveness Data and Information Set (HEDIS) measure for childhood immunization delivery. MATERIALS AND METHODS: This study integrates multiple health care administrative data sources from 2010 through 2014, analyzed through an interrupted time series design and a hierarchical Bayesian model. We compared managed care physicians using CEHRT to propensity-score matched comparisons from network physicians who did not adopt CEHRT. Inclusion criteria for physicians using CEHRT included attesting to the Childhood Immunization Status clinical quality measure in addition to meeting "Meaningful Use" (MU) during calendar year 2013. We used a first-presence patient attribution approach to develop provider-specific immunization scores. RESULTS: We evaluated 147 providers using CEHRT, with 147 propensity-score matched providers selected from a pool of 1253 PCPs practicing in Maryland. The estimate for change in odds of increasing immunization rates due to CEHRT was 1.2 (95% credible set, 0.88-1.73). DISCUSSION: We created a method for estimating immunization quality scores using Bayesian modeling. Our approach required linking separate administrative data sets, constructing a propensity-score matched cohort, and using first-presence, claims-based childhood visit information for patient attribution. In the absence of integrated data sets and precise and accurate patient attribution, this is a reusable method for researchers and health system administrators to estimate the impact of health information technology on individual, provider-level, process-based, though outcomes-focused, quality measures. CONCLUSION: This research has provided evidence for using Bayesian analysis of propensity-score matched provider populations to estimate the impact of CEHRT on outcomes-based quality measures such as childhood immunization delivery.

Bioinformatics for medical students: a 5-year experience using OMIM® in medical student education.

PURPOSE: Given advances in genomic medicine, medical students need increased confidence in clinical genetics skills to address multiple genetic conditions. After success of first-year medical school instruction in the Online Mendelian Inheritance in Man (OMIM®) database, we report the impact on gaining confidence in broad clinical genetics skills in 5 subsequent years. METHODS: We collected 5 years of successive pre- and postintervention survey based self-assessments on medical student use of genetic medicine information resources and confidence in genetic medicine skills. To assess retention of confidence in these skills, we administered a follow-up survey to students after 1-2 years of clinical rotations. RESULTS: We found a consistent, statistically significant increase in students' confidence in clinical genetics skills after the first-year OMIM educational session, with confidence retention above baseline up to 2 years after the educational exposure. Skills include ability to generate a differential diagnosis for genetic conditions, share information with patients and families, and find accurate information on genetic conditions. The majority agreed that increased use of OMIM will better prepare students to achieve these skills. CONCLUSION: Integration of the OMIM database in first-year education is an effective instructional tool that may provide a lasting increase in confidence in clinical genetics skills.

Understanding Primary Care Patients' Self-weighing Habits: Cohort Analysis from the PaTH Clinical Data Research Network.

BACKGROUND: Greater than 60% of adults have overweight or obesity. Self-weighing is an effective weight loss and weight maintenance tool. However, little is known about self-weighing habits among the primary care patient population. Our objective was to examine the frequency of patient-reported self-weighing, and to evaluate the associations of self-weighing with demographic characteristics and self-monitoring behaviors. METHODS: We conducted an analysis of survey data collected as part of the PaTH Clinical Data Research Network, which recruited a cohort of 1,021 primary care patients at 4 academic medical centers. Patients of all body mass index (BMI) categories were included. RESULTS: Response rate of 6-month survey was 727 (71%). The mean age was 56 years, and most were female (68%), White (78%), college graduates (66%), and employed/retired (85%). The mean BMI was 30.2 kg/m2, 80% of participants had a BMI â‰§ 25 kg/m2. Of patients with BMI â‰§ 25 kg/m2, 35% of participants self-weighed weekly and 23% daily. Participants who reported self-weighing at least weekly were more likely to be older (59 vs 54 years, p < 0.01), married (p = 0.01), college graduates (p = 0.03), White (p < 0.01), and employed vs disabled/unemployed (p < 0.01). Patients who self-weighed daily had a lower BMI (29 kg/m2 vs 31 kg/m2, p = 0.04). Patients who tracked exercise or food intake were more likely to self-weigh daily (p < 0.01), as were patients wanting to lose or maintain weight (p < 0.01). CONCLUSIONS: Despite its potential for primary and secondary obesity prevention, only 35% of primary care patients with overweight or obesity engage in self-weighing weekly and less than a quarter (23%) self-weigh daily. Socioeconomic status appears to be a factor influencing regular self-weighing in this population, potentially contributing to greater health disparities in obesity rates. Patients who self-weighed daily had a lower BMI, suggesting that it may play a role in primary prevention of obesity. More work is needed to explore self-weighing among patients.

Association of sex, age and education level with patient reported outcomes in atrial fibrillation.

BACKGROUND: In atrial fibrillation (AF), there are known sex and sociodemographic disparities in clinical outcomes such as stroke. We investigate whether disparities also exist with respect to patient-reported outcomes. We explored the association of sex, age, and education level with patient-reported outcomes (AF-related quality of life, symptom severity, and emotional and functional status). METHODS: The PaTH AF cohort study recruited participants (N = 953) with an AF diagnosis and age ≥ 18 years across 4 academic medical centers. We performed longitudinal multiple regression with random effects to determine if individual characteristics were associated with patient-reported outcomes. RESULTS: Women reported poorer functional status (ß - 2.23, 95% CI: -3.52, - 0.94) and AF-related quality of life (ß - 4.12, 95% CI: -8.10, - 0.14), and higher symptoms of anxiety (ß 2.08, 95% CI: 0.76, 3.40), depression (ß 1.44, 95% CI: 0.25, 2.63), and AF (ß 0.29, 95% CI: 0.08, 0.50). Individuals < 60 years were significantly (p < 0.05) more likely to report higher symptoms of depression, anxiety, and AF, and poorer AF-related quality of life. Lack of college education was associated with reporting higher symptoms of AF (ß 0.42, 95% CI: 0.17, 0.68), anxiety (ß 1.86, 95% CI: 0.26, 3.45), and depression (ß 1.11, 95% CI: 0.15, 2.38), and lower AF-related quality of life (ß - 4.41, 95% CI: -8.25, - 0.57) and functional status. CONCLUSION: Women, younger adults, and individuals with lower levels of education reported comparatively poor patient-reported outcomes. These findings highlight the importance of understanding why individuals experience AF differently based on certain characteristics.

Developing a risk-based composite neurologic outcome for a trial of hydroxyurea in young children with sickle cell disease.

BACKGROUND: Studies of interventions to prevent the many neurological complications of sickle cell disease must take into account multiple outcomes of variable severity, with limited sample size. The goals of the studies presented were to use investigator preferences across outcomes to determine an attitude-based weighting of relevant clinical outcomes and to establish a valid composite outcome for a clinical trial. METHODS: In Study 1, investigators were surveyed about their practice regarding hydroxyurea therapy and opinions about outcomes for the "Hydroxyurea to Prevent the Central Nervous System Complications of Sickle Cell Disease Trial" (HU Prevent), and their minimally acceptable relative risk reduction for the two outcome components, motor and neurocognitive deficits. In Study 2, HU Prevent investigators provided overall weights for these two components. In Study 3, they provided more granular rankings, ratings, and maximum number acceptable to harm. A weighted composite outcome, the Stroke Consequences Risk Score, was constructed that incorporates the major neurologic complications of sickle cell disease. The Stroke Consequences Risk Score represents the 3-year risk of suffering the adverse consequences of stroke. In Study 4, the results of the Optimizing Primary Stroke Prevention in Sickle Cell Anemia (STOP2) and Silent Infarct Transfusion Trials were reanalyzed in light of the composite outcome. RESULTS: In total, 22 to 27 investigators participated per study. In Study 1, across three samplings between 2009 and 2015, the average minimally acceptable relative risk reduction ranged from 0.36 to 0.50, at or below the target effect size of 0.50. In 2015, 21 (91%) reported that a placebo-controlled trial is reasonable; 23 (100%), that it is ethical; and 22 (96%), that they would change their practice, if the results of the trial were positive. In Studies 2 and 3, the weight elicited for a cognitive decline (of 10 IQ points) from the overall assessment was 0.67 (and for motor deficit, the complementary 0.33); from ranking, 0.6; from rating, 0.58; and from maximal number acceptable to harm, 0.5. Using data from two major clinical trials, Study 4 demonstrated the same conclusions as the original trials using the Stroke Consequences Risk Score, with smaller p-values for both reanalyses. An assessment of acceptability was performed as well. CONCLUSION: This set of studies provides the rationale, justification, and validation for the use of a weighted composite outcome and confirms the need for the phase III HU Prevent study. Surveys of investigators in multi-center studies can provide the basis of clinically meaningful outcomes that foster the translation of study results into practice while increasing the efficiency of a study.

Controlled trial of transfusions for silent cerebral infarcts in sickle cell anemia.

BACKGROUND: Silent cerebral infarcts are the most common neurologic injury in children with sickle cell anemia and are associated with the recurrence of an infarct (stroke or silent cerebral infarct). We tested the hypothesis that the incidence of the recurrence of an infarct would be lower among children who underwent regular blood-transfusion therapy than among those who received standard care. METHODS: In this randomized, single-blind clinical trial, we randomly assigned children with sickle cell anemia to receive regular blood transfusions (transfusion group) or standard care (observation group). Participants were between 5 and 15 years of age, with no history of stroke and with one or more silent cerebral infarcts on magnetic resonance imaging and a neurologic examination showing no abnormalities corresponding to these lesions. The primary end point was the recurrence of an infarct, defined as a stroke or a new or enlarged silent cerebral infarct. RESULTS: A total of 196 children (mean age, 10 years) were randomly assigned to the observation or transfusion group and were followed for a median of 3 years. In the transfusion group, 6 of 99 children (6%) had an end-point event (1 had a stroke, and 5 had new or enlarged silent cerebral infarcts). In the observation group, 14 of 97 children (14%) had an end-point event (7 had strokes, and 7 had new or enlarged silent cerebral infarcts). The incidence of the primary end point in the transfusion and observation groups was 2.0 and 4.8 events, respectively, per 100 years at risk, corresponding to an incidence rate ratio of 0.41 (95% confidence interval, 0.12 to 0.99; P=0.04). CONCLUSIONS: Regular blood-transfusion therapy significantly reduced the incidence of the recurrence of cerebral infarct in children with sickle cell anemia. (Funded by the National Institute of Neurological Disorders and Stroke and others; Silent Cerebral Infarct Multi-Center Clinical Trial ClinicalTrials.gov number, NCT00072761, and Current Controlled Trials number, ISRCTN52713285.).

Preemptive Bone Marrow Transplantation for FANCD1/BRCA2.

Children with biallelic mutations in FANCD1/BRCA2 are at uniquely high risks of leukemia and solid tumors. Preemptive bone marrow transplantation (PE-BMT) has been proposed to avoid the development of leukemia, but empirical study of PE-BMT is unlikely because of the rarity of these children and the unknown benefit of PE-BMT. We used survival analysis to estimate the risks of leukemia and the expected survival if leukemia could be eliminated by curative PE-BMT. We used the results in a decision analysis model to explore the plausibility of PE-BMT for children with variable ages at diagnosis and risks of transplantation-related mortality. For example, PE-BMT at 1 year of age with a 10% risk of transplantation-related mortality increased the mean survival by 1.7 years. The greatest benefit was for patients diagnosed between 1 and 3 years of age, after which the benefit of PE-BMT decreased with age at diagnosis, and the risk of death from solid tumors constituted a relatively greater burden of mortality. Our methods may be used to model survival for other hematologic disorders with limited empirical data and a pressing need for clinical guidance.

Network motifs that recur across species, including gene regulatory and protein-protein interaction networks.

Cellular molecules interact in complex ways, giving rise to a cell's functional outcomes. Conscientious efforts have been made in recent years to better characterize these patterns of interactions. It has been learned that many of these interactions can be represented abstractly as a network and within a network there in many instances are network motifs. Network motifs are subgraphs that are statistically overrepresented within networks. To date, specific network motifs have been experimentally identified across various species and also within specific, intracellular networks; however, motifs that recur across species and major network types have not been systematically characterized. We reason that recurring network motifs could potentially have important implications and applications for toxicology and, in particular, toxicity testing. Therefore, the goal of this study was to determine the set of intracellular, network motifs found to recur across species of both gene regulatory and protein-protein interaction networks. We report the recurrence of 13 intracellular, network motifs across species. Ten recurring motifs were found across both protein-protein interaction networks and gene regulatory networks. The significant pair motif was found to recur only in gene regulatory networks. The diamond and one-way cycle reversible step motifs were found to recur only in protein-protein interaction networks. This study is the first formal review of recurring, intracellular network motifs across species. Within toxicology, combining our understanding of recurring motifs with mechanism and mode of action knowledge could result in more robust and efficient toxicity testing models. We are sure that our results will support research in applying network motifs to toxicity testing.

The Ontology of Clinical Research (OCRe): an informatics foundation for the science of clinical research.

To date, the scientific process for generating, interpreting, and applying knowledge has received less informatics attention than operational processes for conducting clinical studies. The activities of these scientific processes - the science of clinical research - are centered on the study protocol, which is the abstract representation of the scientific design of a clinical study. The Ontology of Clinical Research (OCRe) is an OWL 2 model of the entities and relationships of study design protocols for the purpose of computationally supporting the design and analysis of human studies. OCRe's modeling is independent of any specific study design or clinical domain. It includes a study design typology and a specialized module called ERGO Annotation for capturing the meaning of eligibility criteria. In this paper, we describe the key informatics use cases of each phase of a study's scientific lifecycle, present OCRe and the principles behind its modeling, and describe applications of OCRe and associated technologies to a range of clinical research use cases. OCRe captures the central semantics that underlies the scientific processes of clinical research and can serve as an informatics foundation for supporting the entire range of knowledge activities that constitute the science of clinical research.

Assessing the Effect of Electronic Health Record Data Quality on Identifying Patients with Type 2 Diabetes.

BACKGROUND: Increasing and substantial reliance on Electronic health records (EHR) and data types (i.e., diagnosis (Dx), medication (Rx), laboratory (Lx)) demands assessment of its data quality (DQ) as a fundamental approach; especially since there is need to identify appropriate denominator population with chronic conditions, such as Type-2 Diabetes (T2D), using commonly available computable phenotype definitions (phenotype). OBJECTIVE: To bridge this gap, our study aims to assess how issues of EHR DQ, and variations and robustness (or lack thereof) in phenotypes may have potential impact in identifying denominator population. METHODS: Approximately 208k patients with T2D were included in our study using retrospective EHR data of Johns Hopkins Medical Institution (JHMI) during 2017-2019. Our assessment included 4 published phenotypes, and 1 definition from a panel of experts at Hopkins. We conducted descriptive analyses of demographics (i.e., age, sex, race, ethnicity), healthcare utilization (inpatient and emergency room visits), and average Charlson Comorbidity score of each phenotype. We then used different methods to induce/simulate DQ issues of completeness, accuracy and timeliness separately across each phenotype. For induced data incompleteness, our model randomly dropped Dx, Rx, and Lx codes independently at increments of 10%; for induced data inaccuracy, our model randomly replaced a Dx or Rx code with another code of the same data type and induced 2% incremental change from -100% to +10% in Lx result values; and lastly, for timeliness, data was modeled for induced incremental shift of date records by 30 days up to a year. RESULTS: Less than a quarter (23%) of population overlapped across all phenotypes using EHR. The population identified by each phenotype varied across all combination of data types. Induced incompleteness identified fewer patients with each increment, for e.g., at 100% diagnostic incompleteness, Chronic Conditions Data Warehouse (CCW) phenotype identified zero patients as its phenotypic characteristics included only Dx codes. Induced inaccuracy and timeliness similarly demonstrated variations in performance of each phenotype and therefore, resulting in fewer patients being identified with each incremental change. CONCLUSIONS: We utilized EHR data with Dx, Rx, and Lx data types from a large tertiary hospital system to understand the T2D phenotypic differences and performance. We learned how issues of DQ, using induced DQ methods, may impact identification of the denominator populations upon which clinical (e.g., clinical research and trials, population health evaluations) and financial/operational decisions are made. The novel results from our study may inform in shaping a common T2D computable phenotype definition that can be applicable to clinical informatics, managing chronic conditions, and additional healthcare industry-wide efforts.

Racial Disparities in Diabetes Care and Outcomes for Patients with Visual Impairment: A Descriptive Analysis of the TriNetX Research Network.

Background: This research delves into the confluence of racial disparities and health inequities among individuals with disabilities, with a focus on those contending with both diabetes and visual impairment. Methods: Utilizing data from the TriNetX Research Network, which includes electronic medical records of roughly 115 million patients from 83 anonymous healthcare organizations, this study employs a directed acyclic graph (DAG) to pinpoint confounders and augment interpretation. We identified patients with visual impairments using ICD-10 codes, deliberately excluding diabetes-related ophthalmology complications. Our approach involved multiple race-stratified analyses, comparing co-morbidities like chronic pulmonary disease in visually impaired patients against their counterparts. We assessed healthcare access disparities by examining the frequency of annual visits, instances of two or more A1c measurements, and glomerular filtration rate (GFR) measurements. Additionally, we evaluated diabetes outcomes by comparing the risk ratio of uncontrolled diabetes (A1c > 9.0) and chronic kidney disease in patients with and without visual impairments. Results: The incidence of diabetes was substantially higher (nearly double) in individuals with visual impairments across White, Asian, and African American populations. Higher rates of chronic kidney disease were observed in visually impaired individuals, with a risk ratio of 1.79 for African American, 2.27 for White, and non-significant for the Asian group. A statistically significant difference in the risk ratio for uncontrolled diabetes was found only in the White cohort (0.843). White individuals without visual impairments were more likely to receive two A1c tests, a trend not significant in other racial groups. African Americans with visual impairments had a higher rate of glomerular filtration rate testing. However, White individuals with visual impairments were less likely to undergo GFR testing, indicating a disparity in kidney health monitoring. This pattern of disparity was not observed in the Asian cohort. Conclusions: This study uncovers pronounced disparities in diabetes incidence and management among individuals with visual impairments, particularly among White, Asian, and African American groups. Our DAG analysis illuminates the intricate interplay between SDoH, healthcare access, and frequency of crucial diabetes monitoring practices, highlighting visual impairment as both a medical and social issue.

Autonomous artificial intelligence increases screening and follow-up for diabetic retinopathy in youth: the ACCESS randomized control trial.

Diabetic retinopathy can be prevented with screening and early detection. We hypothesized that autonomous artificial intelligence (AI) diabetic eye exams at the point-of-care would increase diabetic eye exam completion rates in a racially and ethnically diverse youth population. AI for Children's diabetiC Eye ExamS (NCT05131451) is a parallel randomized controlled trial that randomized youth (ages 8-21 years) with type 1 and type 2 diabetes to intervention (autonomous artificial intelligence diabetic eye exam at the point of care), or control (scripted eye care provider referral and education) in an academic pediatric diabetes center. The primary outcome was diabetic eye exam completion rate within 6 months. The secondary outcome was the proportion of participants who completed follow-through with an eye care provider if deemed appropriate. Diabetic eye exam completion rate was significantly higher (100%, 95%CI: 95.5%, 100%) in the intervention group (n = 81) than the control group (n = 83) (22%, 95%CI: 14.2%, 32.4%)(p < 0.001). In the intervention arm, 25/81 participants had an abnormal result, of whom 64% (16/25) completed follow-through with an eye care provider, compared to 22% in the control arm (p < 0.001). Autonomous AI increases diabetic eye exam completion rates in youth with diabetes.

Computable phenotype for diagnostic error: developing the data schema for application of symptom-disease pair analysis of diagnostic error (SPADE).

OBJECTIVES: Diagnostic errors are the leading cause of preventable harm in clinical practice. Implementable tools to quantify and target this problem are needed. To address this gap, we aimed to generalize the Symptom-Disease Pair Analysis of Diagnostic Error (SPADE) framework by developing its computable phenotype and then demonstrated how that schema could be applied in multiple clinical contexts. METHODS: We created an information model for the SPADE processes, then mapped data fields from electronic health records (EHR) and claims data in use to that model to create the SPADE information model (intention) and the SPADE computable phenotype (extension). Later we validated the computable phenotype and tested it in four case studies in three different health systems to demonstrate its utility. RESULTS: We mapped and tested the SPADE computable phenotype in three different sites using four different case studies. We showed that data fields to compute an SPADE base measure are fully available in the EHR Data Warehouse for extraction and can operationalize the SPADE framework from provider and/or insurer perspective, and they could be implemented on numerous health systems for future work in monitor misdiagnosis-related harms. CONCLUSIONS: Data for the SPADE base measure is readily available in EHR and administrative claims. The method of data extraction is potentially universally applicable, and the data extracted is conveniently available within a network system. Further study is needed to validate the computable phenotype across different settings with different data infrastructures.