A Framework for Considering the Value of Race and Ethnicity in Estimating Disease Risk

Background: Accounting for race and ethnicity in estimating disease risk may improve the accuracy of predictions but may also encourage a racialized view of medicine. Objective: To present a decision analytic framework for considering the potential benefits of race-aware over race-unaware risk predictions, using cardiovascular disease, breast cancer, and lung cancer as case studies. Design: Cross-sectional study. Setting: NHANES (National Health and Nutrition Examination Survey), 2011 to 2018, and NLST (National Lung Screening Trial), 2002 to 2004. Patients: U.S. adults. Measurements: Starting with risk predictions from clinically recommended race-aware models, the researchers generated race-unaware predictions via statistical marginalization. They then estimated the utility gains of the race-aware over the race-unaware models, based on a simple utility function that assumes constant costs of screening and constant benefits of disease detection. Results: The race-unaware predictions were substantially miscalibrated across racial and ethnic groups compared with the race-aware predictions as the benchmark. However, the clinical net benefit at the population level of race-aware predictions over race-unaware predictions was smaller than expected. This result stems from 2 empirical patterns: First, across all 3 diseases, 95% or more of individuals would receive the same decision regardless of whether race and ethnicity are included in risk models; second, for those who receive different decisions, the net benefit of screening or treatment is relatively small because these patients have disease risks close to the decision threshold (that is, the theoretical “point of indifference”). When used to inform rationing, race-aware models may have a more substantial net benefit. Limitations: For illustrative purposes, the race-aware models were assumed to yield accurate estimates of risk given the input variables. The researchers used a simplified approach to generate race-unaware risk predictions from the race-aware models and a simple utility function to compare models. Conclusion: The analysis highlights the importance of foregrounding changes in decisions and utility when evaluating the potential benefit of using race and ethnicity to estimate disease risk. Primary Funding Source: The Greenwall Foundation.

Does Racial Bias Play a Role in Internal Medicine Resident Knowledge Evaluations?

Comparative Effectiveness and Safety of Apixaban, Rivaroxaban, and Warfarin in Patients With Cirrhosis and Atrial Fibrillation: A Nationwide Cohort Study: Annals of Internal Medicine: Vol 177, No 8

Background: Apixaban, rivaroxaban, and warfarin have shown benefit for preventing major ischemic events, albeit with increased bleeding risk, among patients in the general population with atrial fibrillation (AF). However, data are scarce in patients with cirrhosis and AF. Objective: To compare the effectiveness and safety of apixaban versus rivaroxaban and versus warfarin in patients with cirrhosis and AF. Design: Population-based cohort study. Setting: Two U.S. claims data sets (Medicare and Optum’s de-identified Clinformatics Data Mart Database [2013 to 2022]). Participants: 1:1 propensity score (PS)–matched patients with cirrhosis and nonvalvular AF initiating use of apixaban, rivaroxaban, or warfarin. Measurements: Primary outcomes included ischemic stroke or systemic embolism and major hemorrhage (intracranial hemorrhage or major gastrointestinal bleeding). Database-specific and pooled PS-matched rate differences (RDs) per 1000 person-years (PY) and Cox proportional hazard ratios (HRs) with 95% CIs were estimated, controlling for 104 preexposure covariates. Results: Rivaroxaban initiators had significantly higher rates of major hemorrhagic events than apixaban initiators (RD, 33.1 per 1000 PY [95% CI, 12.9 to 53.2 per 1000 PY]; HR, 1.47 [CI, 1.11 to 1.94]) but no significant differences in rates of ischemic events or death. Consistently higher rates of major hemorrhage were found with rivaroxaban across subgroup and sensitivity analyses. Warfarin initiators also had significantly higher rates of major hemorrhage than apixaban initiators (RD, 26.1 per 1000 PY [CI, 6.8 to 45.3 per 1000 PY]; HR, 1.38 [CI, 1.03 to 1.84]), particularly hemorrhagic stroke (RD, 9.7 per 1000 PY [CI, 2.2 to 17.2 per 1000 PY]; HR, 2.85 [CI, 1.24 to 6.59]). Limitation: Nonrandomized treatment selection. Conclusion: Among patients with cirrhosis and nonvalvular AF, initiators of rivaroxaban versus apixaban had significantly higher rates of major hemorrhage and similar rates of ischemic events and death. Initiation of warfarin versus apixaban also contributed to significantly higher rates of major hemorrhagic events, including hemorrhagic stroke. Primary Funding Source: National Institutes of Health.

Disparities in Tuberculosis Incidence by Race and Ethnicity Among the U.S.-Born Population in the United States, 2011 to 2021: An Analysis of National Disease Registry Data: Annals of Internal Medicine: Vol 177, No 4

Background: Elevated tuberculosis (TB) incidence rates have recently been reported for racial/ethnic minority populations in the United States. Tracking such disparities is important for assessing progress toward national health equity goals and implementing change. Objective: To quantify trends in racial/ethnic disparities in TB incidence among U.S.-born persons. Design: Time-series analysis of national TB registry data for 2011 to 2021. Setting: United States. Participants: U.S.-born persons stratified by race/ethnicity. Measurements: TB incidence rates, incidence rate differences, and incidence rate ratios compared with non-Hispanic White persons; excess TB cases (calculated from incidence rate differences); and the index of disparity. Analyses were stratified by sex and by attribution of TB disease to recent transmission and were adjusted for age, year, and state of residence. Results: In analyses of TB incidence rates for each racial/ethnic population compared with non-Hispanic White persons, incidence rate ratios were as high as 14.2 (95% CI, 13.0 to 15.5) among American Indian or Alaska Native (AI/AN) females. Relative disparities were greater for females, younger persons, and TB attributed to recent transmission. Absolute disparities were greater for males. Excess TB cases in 2011 to 2021 represented 69% (CI, 66% to 71%) and 62% (CI, 60% to 64%) of total cases for females and males, respectively. No evidence was found to indicate that incidence rate ratios decreased over time, and most relative disparity measures showed small, statistically nonsignificant increases. Limitation: Analyses assumed complete TB case diagnosis and self-report of race/ethnicity and were not adjusted for medical comorbidities or social determinants of health. Conclusion: There are persistent disparities in TB incidence by race/ethnicity. Relative disparities were greater for AI/AN persons, females, and younger persons, and absolute disparities were greater for males. Eliminating these disparities could reduce overall TB incidence by more than 60% among the U.S.-born population. Primary Funding Source: Centers for Disease Control and Prevention.

The New Proposed U.S. Preventive Services Task Force Recommendation on Breast Cancer Screening for Women in Their 40s: Should the Message Change?

School Mask Mandates and COVID-19: The Challenge of Using Difference-in-Differences Analysis of Observational Data to Estimate the Effectiveness of a Public Health Intervention

Background: There are considerable challenges when using difference-in-differences (DiD) analysis of ecological data to estimate the effectiveness of public health interventions in rapidly changing situations. Objective: To discuss the shortcomings of DiD methodology for the estimation of the effects of public health interventions using ecological data. Design: As an example, the authors consider an analysis that used DiD methodology and reported a causal reduction in COVID-19 cases due to the maintenance of school mask mandates. They did alternate analyses using various control groups to assess the robustness of the prior analysis. Setting: School districts in the greater Boston area and Massachusetts during the 2021-to-2022 academic year. Participants: Students and school staff. Measurements: Changes in COVID-19 case rates in districts that did and did not lift mask mandates. Results: Important potential confounders rendered DiD methodology inappropriate for causal inference, including prior immunity, temporal variation in rates of infection, and changes in testing practices. The racial composition and income of intervention and control groups also differed substantially. Compared with maintaining the mask requirement, dropping the requirement was associated with anywhere from an increase of 5.64 cases (95% CI, 3.00 to 8.29 cases) per 1000 persons to a decrease of 2.74 cases (CI, 0.63 to 4.85 cases) per 1000 persons, depending on choice of control group and whether students or staff were examined. Limitation: Ecological data were used; detailed data on all potential confounders were unavailable. Conclusion: Alternate analyses yielded estimates consistent with a wide range of both negative and positive associations in COVID-19 case rates after removal of mask mandates. The findings highlight the challenges of using DiD analysis of ecological data to estimate the effectiveness of interventions in divergent intervention and control groups during rapidly changing circumstances. Primary Funding Source: None.

Association Between Autoimmune Diseases and Monoclonal Gammopathy of Undetermined Significance: An Analysis From a Population-Based Screening Study: Annals of Internal Medicine: Vol 177, No 6

Background: Monoclonal gammopathy of undetermined significance (MGUS) is a precursor of multiple myeloma (MM) and related conditions. In previous registry-based, retrospective studies, autoimmune diseases have been associated with MGUS. However, these studies were not based on a screened population and are therefore prone to ascertainment bias. Objective: To examine whether MGUS is associated with autoimmune diseases. Design: A cross-sectional study within iStopMM (Iceland Screens, Treats, or Prevents MM), a prospective, population-based screening study of MGUS. Setting: Icelandic population of adults aged 40 years or older. Patients: 75 422 persons screened for MGUS. Measurements: Poisson regression for prevalence ratios (PRs) of MGUS among persons with or without an autoimmune disease, adjusted for age and sex. Results: A total of 10 818 participants had an autoimmune disorder, of whom 599 had MGUS (61 with a prior clinical diagnosis and 538 diagnosed at study screening or evaluation). A diagnosis of an autoimmune disease was not associated with MGUS (PR, 1.05 [95% CI, 0.97 to 1.15]). However, autoimmune disease diagnoses were associated with a prior clinical diagnosis of MGUS (PR, 2.11 [CI, 1.64 to 2.70]). Limitation: Registry data were used to gather information on autoimmune diseases, and the homogeneity of the Icelandic population may limit the generalizability of these results. Conclusion: The study did not find an association between autoimmune disease and MGUS in a systematically screened population. Previous studies not done in systematically screened populations have likely been subject to ascertainment bias. The findings indicate that recommendations to routinely screen patients with autoimmune disease for MGUS may not be warranted. Primary Funding Source: The International Myeloma Foundation and the European Research Council.

Computer-Aided Diagnosis for Leaving Colorectal Polyps In Situ: A Systematic Review and Meta-analysis: Annals of Internal Medicine: Vol 177, No 7

Background: Computer-aided diagnosis (CADx) allows prediction of polyp histology during colonoscopy, which may reduce unnecessary removal of nonneoplastic polyps. However, the potential benefits and harms of CADx are still unclear. Purpose: To quantify the benefit and harm of using CADx in colonoscopy for the optical diagnosis of small (≤5-mm) rectosigmoid polyps. Data Sources: Medline, Embase, and Scopus were searched for articles published before 22 December 2023. Study Selection: Histologically verified diagnostic accuracy studies that evaluated the real-time performance of physicians in predicting neoplastic change of small rectosigmoid polyps without or with CADx assistance during colonoscopy. Data Extraction: The clinical benefit and harm were estimated on the basis of accuracy values of the endoscopist before and after CADx assistance. The certainty of evidence was assessed using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) framework. The outcome measure for benefit was the proportion of polyps predicted to be nonneoplastic that would avoid removal with the use of CADx. The outcome measure for harm was the proportion of neoplastic polyps that would be not resected and left in situ due to an incorrect diagnosis with the use of CADx. Histology served as the reference standard for both outcomes. Data Synthesis: Ten studies, including 3620 patients with 4103 small rectosigmoid polyps, were analyzed. The studies that assessed the performance of CADx alone (9 studies; 3237 polyps) showed a sensitivity of 87.3% (95% CI, 79.2% to 92.5%) and specificity of 88.9% (CI, 81.7% to 93.5%) in predicting neoplastic change. In the studies that compared histology prediction performance before versus after CADx assistance (4 studies; 2503 polyps), there was no difference in the proportion of polyps predicted to be nonneoplastic that would avoid removal (55.4% vs. 58.4%; risk ratio [RR], 1.06 [CI, 0.96 to 1.17]; moderate-certainty evidence) or in the proportion of neoplastic polyps that would be erroneously left in situ (8.2% vs. 7.5%; RR, 0.95 [CI, 0.69 to 1.33]; moderate-certainty evidence). Limitation: The application of optical diagnosis was only simulated, potentially altering the decision-making process of the operator. Conclusion: Computer-aided diagnosis provided no incremental benefit or harm in the management of small rectosigmoid polyps during colonoscopy. Primary Funding Source: European Commission. (PROSPERO: CRD42023402197)

Association of Albuminuria With Chronic Kidney Disease Progression in Persons With Chronic Kidney Disease and Normoalbuminuria: A Cohort Study: Annals of Internal Medicine: Vol 177, No 4

Background: Albuminuria is a major risk factor for chronic kidney disease (CKD) progression, especially when categorized as moderate (30 to 300 mg/g) or severe (>300 mg/g). However, there are limited data on the prognostic value of albuminuria within the normoalbuminuric range (<30 mg/g) in persons with CKD. Objective: To estimate the increase in the cumulative incidence of CKD progression with greater baseline levels of albuminuria among persons with CKD who had normoalbuminuria (<30 mg/g). Design: Multicenter prospective cohort study. Setting: 7 U.S. clinical centers. Participants: 1629 participants meeting criteria from the CRIC (Chronic Renal Insufficiency Cohort) study with CKD (estimated glomerular filtration rate [eGFR], 20 to 70 mL/min/1.73 m2) and urine albumin–creatinine ratio (UACR) less than 30 mg/g. Measurements: Baseline spot urine albumin divided by spot urine creatinine to calculate UACR as the exposure variable. The 10-year adjusted cumulative incidences of CKD progression (composite of 50% eGFR decline or kidney failure [dialysis or kidney transplantation]) from confounder adjusted survival curves using the G-formula. Results: Over a median follow-up of 9.8 years, 182 of 1629 participants experienced CKD progression. The 10-year adjusted cumulative incidences of CKD progression were 8.7% (95% CI, 5.9% to 11.6%), 11.5% (CI, 8.8% to 14.3%), and 19.5% (CI, 15.4% to 23.5%) for UACR levels of 0 to less than 5 mg/g, 5 to less than 15 mg/g, and 15 mg/g or more, respectively. Comparing persons with UACR 15 mg/g or more to those with UACR 5 to less than 15 mg/g and 0 to less than 5 mg/g, the absolute risk differences were 7.9% (CI, 3.0% to 12.7%) and 10.7% (CI, 5.8% to 15.6%), respectively. The 10-year adjusted cumulative incidence increased linearly based on baseline UACR levels. Limitation: UACR was measured once. Conclusion: Persons with CKD and normoalbuminuria (<30 mg/g) had excess risk for CKD progression, which increased in a linear fashion with higher levels of albuminuria. Primary Funding Source: None.

Clinical Outcomes With Electronic Nudges to Increase Influenza Vaccination: A Prespecified Analysis of a Nationwide, Pragmatic, Registry-Based, Randomized Implementation Trial: Annals of Internal Medicine: Vol 177, No 4

Background: In the NUDGE-FLU (Nationwide Utilization of Danish Government Electronic letter system for increasing inFLUenza vaccine uptake) trial, electronic letters incorporating cardiovascular (CV) gain-framing and repeated messaging increased influenza vaccination by approximately 1 percentage point. Objective: To evaluate the effects of the successful nudging interventions on downstream clinical outcomes. Design: Prespecified exploratory analysis of a nationwide randomized implementation trial. (ClinicalTrials.gov: NCT05542004) Setting: The 2022 to 2023 influenza season. Participants: 964 870 Danish citizens aged 65 years or older. Intervention: Usual care or 9 different electronically delivered behavioral nudging letters. Measurements: Cardiovascular, respiratory, and other clinical end points during follow-up from intervention delivery (16 September 2022) through 31 May 2023. Results: The analysis set included 691 820 participants. Hospitalization for pneumonia or influenza occurred in 3354 of 346 327 (1.0%) participants in the usual care group, 396 of 38 586 (1.0%) in the CV gain-framing group (hazard ratio [HR], 1.06 [95% CI, 0.95 to 1.18]; versus usual care), and 403 of 38 231 (1.1%) in the repeated letter group (HR, 1.09 [CI, 0.98 to 1.21]; versus usual care). In the usual care group, 44 682 (12.9%) participants were hospitalized for any cause, compared with 5002 (13.0%) in the CV gain-framing group (HR, 1.00 [CI, 0.97 to 1.03]; versus usual care) and 4965 (13.0%) in the repeated letter group (HR, 1.01 [CI, 0.98 to 1.04]; versus usual care). A total of 6341 (1.8%) participants died in the usual care group, compared with 721 (1.9%) in the CV gain-framing group (HR, 1.02 [CI, 0.94 to 1.10]; versus usual care) and 646 (1.7%) in the repeated letter group (HR, 0.92 [CI, 0.85 to 1.00]; versus usual care). Limitation: Prespecified but exploratory analysis, potential misclassification of events in routinely collected registry data, and results may not be generalizable to other health systems or countries with other racial compositions and/or cultural or societal norms. Conclusion: In a prespecified exploratory analysis, modest increases in influenza vaccination rates seen with electronic nudges did not translate into observable improvements in clinical outcomes. Seasonal influenza vaccination should remain strongly recommended. Primary Funding Source: Sanofi.