Household Firearm Ownership and Firearm Mortality.

Importance: Causal associations between household firearm ownership rates (HFRs) and firearm mortality rates are not well understood. Objective: To assess the population-level temporal sequencing of firearm death rates and HFRs. Design, Setting, and Participants: This cohort study used autoregressive cross-lagged models to analyze HFRs, firearm suicide rates, and firearm homicide rates in the US from 1990 to 2018. The suicide analyses included 16 demographic subgroups of adults, defined by study year, state, sex, race and ethnicity, marital status, and urbanicity. The homicide analyses consisted of adult subgroups living in urban or rural areas. Data analysis was conducted from March to December 2023. Exposures: Firearm mortality rates and HFRs. Main Outcomes and Measures: Firearm homicide and suicide rates with HFRs as the exposure, and HFR with mortality as the exposure. Results: A total of 10 416 observations of 16 demographic subgroups by state and 2-year periods were included in the suicide analyses, while 1302 observations from 2 demographic subgroups by state and 2-year period were included in the homicide analysis. At baseline, the mean (SD) rate per 100 000 population across strata was 7.46 (7.21) for firearm suicides and 3.32 (2.13) for firearm homicides. The mean (SD) baseline HFR was 36.9% (20.2%) for firearm suicides and 36.9% (14.8%) for firearm homicides. Higher HFR preceded increases in suicide rates: demographic strata with equal firearm suicide rates but which differ by 18.6 percentage points on HFR (1 SD) would be expected to have firearm suicide rates that diverged by 0.19 (95% CI, 0.15-0.23) deaths per 100 000 population per period. With these differences accumulated over 8 years, firearm suicide rates in subgroups with the highest decile HFR would be expected to have 1.93 (95% CI, 1.64-2.36) more suicides per 100 000 population than strata with lowest decile HFR, a difference of 25.7% of the overall firearm suicide rate in 2018 and 2019. Firearm suicide rates had a smaller magnitude of association with subsequent changes in HFR: strata with equal HFRs but which differ by 1 SD in firearm suicide rates had minimal subsequent change in HFRs (-0.02 [95% CI, -0.04 to 0.01] percentage points). A 1-SD difference in HFRs was associated with little difference in next-period overall firearm homicides rates (0.03 [95% CI, -0.02 to 0.08] per 100 000 population), but a 1-SD difference in homicide rates was associated with a decrease in HFR (-0.09 [95% CI, -0.16 to -0.04] percentage points). Conclusions and Relevance: This cohort study found an association between high HFRs and subsequent increases in rates of firearm suicide. In contrast, higher firearm homicide rates preceded decreases in HFRs. By demonstrating the temporal sequencing of firearm ownership and mortality, this study may help to rule out some theories of why gun ownership and firearm mortality are associated at the population level.

Revisiting diabetes risk of olanzapine versus aripiprazole in serious mental illness care.

BACKGROUND: Exposure to second-generation antipsychotics (SGAs) carries a risk of type 2 diabetes, but questions remain about the diabetogenic effects of SGAs. AIMS: To assess the diabetes risk associated with two frequently used SGAs. METHOD: This was a retrospective cohort study of adults with schizophrenia, bipolar I disorder or severe major depressive disorder (MDD) exposed during 2008-2013 to continuous monotherapy with aripiprazole or olanzapine for up to 24 months, with no pre-period exposure to other antipsychotics. Newly diagnosed type 2 diabetes was quantified with targeted minimum loss-based estimation; risk was summarised as the restricted mean survival time (RMST), the average number of diabetes-free months. Sensitivity analyses were used to evaluate potential confounding by indication. RESULTS: Aripiprazole-treated patients had fewer diabetes-free months compared with olanzapine-treated patients. RMSTs were longer in olanzapine-treated patients, by 0.25 months [95% CI: 0.14, 0.36], 0.16 months [0.02, 0.31] and 0.22 months [0.01, 0.44] among patients with schizophrenia, bipolar I disorder and severe MDD, respectively. Although some sensitivity analyses suggest a risk of unobserved confounding, E-values indicate that this risk is not severe. CONCLUSIONS: Using robust methods and accounting for exposure duration effects, we found a slightly higher risk of type 2 diabetes associated with aripiprazole compared with olanzapine monotherapy regardless of diagnosis. If this result was subject to unmeasured selection despite our methods, it would suggest clinician success in identifying olanzapine candidates with low diabetes risk. Confirmatory research is needed, but this insight suggests a potentially larger role for olanzapine in the treatment of well-selected patients, particularly for those with schizophrenia, given the drug's effectiveness advantage among them.

Rural Hospitals Experienced More Patient Volume Variability Than Urban Hospitals During The COVID-19 Pandemic, 2020-21.

Fluctuations in patient volume during the COVID-19 pandemic may have been particularly concerning for rural hospitals. We examined hospital discharge data from the Healthcare Cost and Utilization Project State Inpatient Databases to compare data from the COVID-19 pandemic period (March 8, 2020-December 31, 2021) with data from the prepandemic period (January 1, 2017-March 7, 2020). Changes in average daily medical volume at rural hospitals showed a dose-response relationship with community COVID-19 burden, ranging from a 13.2 percent decrease in patient volume in periods of low transmission to a 16.5 percent increase in volume in periods of high transmission. Overall, about 35 percent of rural hospitals experienced fluctuations exceeding 20 percent (in either direction) in average daily total volume, in contrast to only 13 percent of urban hospitals experiencing similar magnitudes of changes. Rural hospitals with a large change in average daily volume were more likely to be smaller, government-owned, and critical access hospitals and to have significantly lower operating margins. Our findings suggest that rural hospitals may have been more vulnerable operationally and financially to volume shifts during the pandemic, which warrants attention because of the potential impact on these hospitals' long-term sustainability.

Mortality for Time-Sensitive Conditions at Urban vs Rural Hospitals During the COVID-19 Pandemic.

Importance: COVID-19 pandemic-related disruptions to the health care system may have resulted in increased mortality for patients with time-sensitive conditions. Objective: To examine whether in-hospital mortality in hospitalizations not related to COVID-19 (non-COVID-19 stays) for time-sensitive conditions changed during the pandemic and how it varied by hospital urban vs rural location. Design, Setting, and Participants: This cohort study was an interrupted time-series analysis to assess in-hospital mortality during the COVID-19 pandemic (March 8, 2020, to December 31, 2021) compared with the prepandemic period (January 1, 2017, to March 7, 2020) overall, by month, and by community COVID-19 transmission level for adult discharges from 3813 US hospitals in the State Inpatient Databases for the Healthcare Cost and Utilization Project. Exposure: The COVID-19 pandemic. Main Outcomes and Measures: The main outcome measure was in-hospital mortality among non-COVID-19 stays for 6 time-sensitive medical conditions: acute myocardial infarction, hip fracture, gastrointestinal hemorrhage, pneumonia, sepsis, and stroke. Entropy weights were used to align patient characteristics in the 2 time periods by age, sex, and comorbidities. Results: There were 18 601 925 hospitalizations; 50.3% of patients were male, 38.5% were aged 18 to 64 years, 45.0% were aged 65 to 84 years, and 16.4% were 85 years or older for the selected time-sensitive medical conditions from 2017 through 2021. The odds of in-hospital mortality for sepsis increased 27% from the prepandemic to the pandemic periods at urban hospitals (odds ratio [OR], 1.27; 95% CI, 1.25-1.29) and 35% at rural hospitals (OR, 1.35; 95% CI, 1.30-1.40). In-hospital mortality for pneumonia had similar increases at urban (OR, 1.48; 95% CI, 1.42-1.54) and rural (OR, 1.46; 95% CI, 1.36-1.57) hospitals. Increases in mortality for these 2 conditions showed a dose-response association with the community COVID-19 level (low vs high COVID-19 burden) for both rural (sepsis: 22% vs 54%; pneumonia: 30% vs 66%) and urban (sepsis: 16% vs 28%; pneumonia: 34% vs 61%) hospitals. The odds of mortality for acute myocardial infarction increased 9% (OR, 1.09; 95% CI, 1.06-1.12) at urban hospitals and was responsive to the community COVID-19 level. There were significant increases in mortality for hip fracture at rural hospitals (OR, 1.32; 95% CI, 1.14-1.53) and for gastrointestinal hemorrhage at urban hospitals (OR, 1.15; 95% CI, 1.09-1.21). No significant change was found in mortality for stroke overall. Conclusions and Relevance: In this cohort study, in-hospital mortality for time-sensitive conditions increased during the COVID-19 pandemic. Mobilizing strategies tailored to the different needs of urban and rural hospitals may help reduce the likelihood of excess deaths during future public health crises.

Comparative Performance of Three Claims-Based Frailty Measures Among Medicare Beneficiaries.

Frailty is an important predictor of mortality, health care costs and utilization, and health outcomes. Validated measures of frailty are not consistently collected during clinical encounters, making comparisons across populations challenging. However, several claims-based algorithms have been developed to predict frailty and related concepts. This study compares performance of three such algorithms among Medicare beneficiaries. Claims data from 12-month continuous enrollment periods were selected during 2014-2016. Frailty scores, calculated using previously developed algorithms from Faurot, Kim, and RAND, were added to baseline regression models to predict claims-based outcomes measured in the following year. Root mean square error and area under the receiver operating characteristic curve were calculated for each model and outcome combination and tested in subpopulations of interest. Overall, Kim models performed best across most outcomes, metrics, and subpopulations. Kim frailty scores may be used by health systems and researchers for risk adjustment or targeting interventions.

Do Hospital Characteristics Predict Racial-and-Ethnic Disparities in Patient Experience? National Results From the HCAHPS Survey.

OBJECTIVE: Assess whether hospital characteristics associated with better patient experiences overall are also associated with smaller racial-and-ethnic disparities in inpatient experience. BACKGROUND: Hospitals that are smaller, non-profit, and serve high proportions of White patients tend to be high-performing overall, but it is not known whether these hospitals also have smaller racial-and-ethnic disparities in care. RESEARCH DESIGN: We used linear mixed-effect regression models to predict a summary measure that averaged eight Hospital CAHPS (HCAHPS) measures (Nurse Communication, Doctor Communication, Staff Responsiveness, Communication about Medicines, Discharge Information, Care Coordination, Hospital Cleanliness, and Quietness) from patient race-and-ethnicity, hospital characteristics (size, ownership, racial-and-ethnic patient-mix), and interactions of race-and-ethnicity with hospital characteristics. SUBJECTS: Inpatients discharged from 4,365 hospitals in 2021 who completed an HCAHPS survey ( N =2,288,862). RESULTS: While hospitals serving larger proportions of Black and Hispanic patients scored lower on all measures, racial-and-ethnic disparities were generally smaller for Black and Hispanic patients who received care from hospitals serving higher proportions of patients in their racial-and-ethnic group. Experiences overall were better in smaller and non-profit hospitals, but racial-and-ethnic differences were slightly larger. CONCLUSIONS: Large, for-profit hospitals and hospitals serving higher proportions of Black and Hispanic patients tend to be lower performing overall but have smaller disparities in patient experience. High-performing hospitals might look at low-performing hospitals for how to provide less disparate care whereas low-performing hospitals may look to high-performing hospitals for how to improve patient experience overall.

Do stigma and efficacy mediate the association between training and suicide prevention behavior among Army noncommissioned officers?

INTRODUCTION: Rates of suicide in the Active Component of the military have significantly increased since 2010, with particularly high rates among Army service members. One element of the Army's approach to suicide prevention relies on noncommissioned officers (NCOs) as gatekeepers who have regular contact with soldiers. NCOs receive suicide prevention training, but there is limited evidence that such training leads to behavior change. METHODS: We surveyed 2468 Army NCOs participating in leadership development courses to determine (a) if training on suicide prevention and soft skills (e.g., active listening) was associated with gatekeeper behavior and use of soft skills; and (b) whether that association was explained by two potential barriers, stigma and perceptions of efficacy. RESULTS: Both the number of suicide prevention training topics and soft skills trained were associated with increased gatekeeper behavior; these relationships were explained in part by lower stigma and higher efficacy for use of soft skills. The use of interactive training methods and receiving coaching after training were not associated with stigma or efficacy, though both methods were associated with more frequent use of soft skills. CONCLUSION: Results suggest that the content and format of training is important to preparing NCOs to fulfill a gatekeeper role.

A Formal Framework For Incorporating Equity Into Health Care Quality Measurement.

Quality measurement is an important tool for incentivizing improvement in the quality of health care. Most quality measurement efforts do not explicitly target health equity. Although some measurement approaches may intend to realign incentives to focus quality improvement efforts on underserved groups, the extent to which they accomplish this goal is understudied. We posit that tying incentives to approaches on the basis of stratification or disparities may have unintended consequences or limited effects. Such approaches might not reduce existing disparities because addressing one aspect of equity may be in competition with addressing others. We propose equity weighting, a new measurement framework to advance equity on multiple fronts that addresses the shortcomings of existing approaches and explicitly calibrates incentives to align with equity goals. We use colorectal cancer screening data derived from 2017 Medicare claims to illustrate how equity weighting fixes unintended consequences in other methods and how it can be adapted to policy goals.

Antipsychotics and the risk of diabetes and death among adults with serious mental illnesses.

BACKGROUND: Individuals with schizophrenia exposed to second-generation antipsychotics (SGA) have an increased risk for diabetes, with aripiprazole purportedly a safer drug. Less is known about the drugs' mortality risk or whether serious mental illness (SMI) diagnosis or race/ethnicity modify these effects. METHODS: Authors created a retrospective cohort of non-elderly adults with SMI initiating monotherapy with an SGA (olanzapine, quetiapine, risperidone, and ziprasidone, aripiprazole) or haloperidol during 2008-2013. Three-year diabetes incidence or all-cause death risk differences were estimated between each drug and aripiprazole, the comparator, as well as effects within SMI diagnosis and race/ethnicity. Sensitivity analyses evaluated potential confounding by indication. RESULTS: 38 762 adults, 65% White and 55% with schizophrenia, initiated monotherapy, with haloperidol least (6%) and quetiapine most (26·5%) frequent. Three-year mortality was 5% and diabetes incidence 9.3%. Compared with aripiprazole, haloperidol and olanzapine reduced diabetes risk by 1.9 (95% CI 1.2-2.6) percentage points, or a 18.6 percentage point reduction relative to aripiprazole users' unadjusted risk (10.2%), with risperidone having a smaller advantage. Relative to aripiprazole users' unadjusted risk (3.4%), all antipsychotics increased mortality risk by 1.1-2.2 percentage points, representing 32.4-64.7 percentage point increases. Findings within diagnosis and race/ethnicity were generally consistent with overall findings. Only quetiapine's higher mortality risk held in sensitivity analyses. CONCLUSIONS: Haloperidol's, olanzapine's, and risperidone's lower diabetes risks relative to aripiprazole were not robust in sensitivity analyses but quetiapine's higher mortality risk proved robust. Findings expand the evidence on antipsychotics' risks, suggesting a need for caution in the use of quetiapine among individuals with SMI.

How are state telehealth policies associated with services offered by substance use disorder treatment facilities? Evidence from 2019 to 2022.

BACKGROUND: The COVID-19 pandemic led several states to adopt policies permitting the delivery of substance use disorder treatment (SUDT) by telehealth. We assess the impact of state-level telehealth policies in 2020 that specifically permitted audio or audiovisual forms of telehealth offerings among SUDT facilities. PROCEDURE: Cross-sectional analysis of secondary data from between 2019 and 2022. Pre-pandemic, federal law permitted states to allow audiovisual telehealth modes for SUDT to a limited extent. 2020 laws permitted states to allow audio-only modes for the first time and strengthened ability to offer audiovisual modes. We compared national SUDT facility self-reported telehealth offerings in 2020 and beyond to 2019, in states that in 2020 had policies permitting audiovisual and audio only, compared to other states. MAIN FINDINGS: Among outpatient SUDT facilities (n = 5227) present in all four years of our data, the proportion offering telehealth increased from 18% (n = 921) in 2019-26% in 2020, 60% in 2021, and 79% in 2022. We estimate an audiovisual and audio only policy in 2020 was associated with an increase in telehealth offering rates in 2022 of +16.5% points (pp) (95% CI [+10.4,+22.6]) compared to the rates in states with no such listed policy. There was little evidence of an influence on telehealth offering in 2020 (-2.9 pp, CI [-9.0,+3.2]) and 2021 (+0.6 pp, CI [-5.5,+6.7]). CONCLUSIONS: The enactment of state-level telehealth policies that allow audio and audiovisual modalities may have increased SUDT facilities' likelihood of offering telehealth services two years after enactment.

Effects of the COVID-19 early pandemic on delivery outcomes among women with and without COVID-19 at birth.

BACKGROUND: The COVID-19 pandemic may influence delivery outcomes through direct effects of infection or indirect effects of disruptions in prenatal care. We examined early pandemic-related changes in birth outcomes for pregnant women with and without a COVID-19 diagnosis at delivery. METHODS: We compared four delivery outcomes-preterm delivery (PTD), severe maternal morbidity (SMM), stillbirth, and cesarean birth-between 2017 and 2019 (prepandemic) and between April and December 2020 (early-pandemic) using interrupted time series models on 11.8 million deliveries, stratified by COVID-19 infection status at birth with entropy weighting for historical controls, from the Healthcare Cost and Utilization Project across 43 states and the District of Columbia. RESULTS: Relative to 2017-2019, women without COVID-19 at delivery in 2020 had lower odds of PTD (OR = 0.93; 95% CI = 0.92-0.94) and SMM (OR = 0.88; 95% CI = 0.85-0.91) but increased odds of stillbirth (OR = 1.04; 95% CI = 1.01-1.08). Absolute effects were small across race/ethnicity groups. Deliveries with COVID-19 had an excess of each outcome, by factors of 1.07-1.46 for outcomes except SMM at 4.21. The effect for SMM was more pronounced for Asian/Pacific Islander non-Hispanic (API; OR = 10.51; 95% CI = 5.49-20.14) and Hispanic (OR = 5.09; 95% CI = 4.29-6.03) pregnant women than for White non-Hispanic (OR = 3.28; 95% CI = 2.65-4.06) women. DISCUSSION: Decreasing rates of PTD and SMM and increasing rates of stillbirth among deliveries without COVID-19 were small but suggest indirect effects of the pandemic on maternal outcomes. Among pregnant women with COVID-19 at delivery, adverse effects, particularly SMM for API and Hispanic women, underscore the importance of addressing health disparities.

Upstream suicide prevention in the U.S. Army: Noncommissioned officers' perspectives.

The goal of this study was to examine the factors associated with Army noncommissioned officer (NCO) experiences, attitudes, and behaviors in their role of identifying potential suicide risk factors in their fellow soldiers. To better understand the perspectives of NCOs, an anonymous survey was administered to 2,468 Army NCOs. Descriptive statistics and linear regressions were conducted to compare subgroups of NCOs. Most (71%) Army NCOs have received many (11 or more) hours of suicide prevention training, but training in soft skills that may be important for the gatekeeper role was less consistently reported. Active Component soldiers reported greater confidence in their intervention skills (Cohen's d = 0.25) and fewer logistical barriers (e.g., time and space to talk) to intervening with at-risk soldiers (Cohen's d = 0.80) compared to Reserve and National Guard soldiers. Formal coursework in mental health areas like psychology or chaplaincy was associated with a greater level of confidence in intervention skills (Cohen's d = 0.23) and in more frequent intervention behavior (Cohen's d = 0.13). Army NCO trainings should be modified to better equip soldiers with the soft skills (e.g., active listening skills and verbally and nonverbally conveying nonjudgment/acceptance and empathy) needed to have effective conversations with soldiers about suicide risk factors and other sensitive topics. Strategies used within mental health education, which appears to be a strength for NCO gatekeepers, could be used to achieve this goal. Reserve and Guard NCOs may need additional supports and tailored trainings to better fit their operational context. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

Development of a Medicare plan dashboard to promote health equity.

OBJECTIVE: To describe a multistage process of designing and evaluating a dashboard that presents data on how equitably health plans provide care for their members. STUDY DESIGN: We designed a dashboard for presenting summative and finer-grained data to health plans for characterizing how well plans are serving individuals who belong to racial/ethnic minority groups and individuals with low income. The data presented in the dashboard were based on CMS' Health Equity Summary Score (HESS) for Medicare Advantage plans. METHODS: Interviews and listening sessions were conducted with health plan representatives and other stakeholders to assess understanding, perceived usefulness, and interpretability of HESS data. Usability testing was conducted with individuals familiar with quality measurement and reporting to evaluate dashboard design efficiency. RESULTS: Listening session participants understood the purpose of the HESS and expressed a desire for this type of information. Usability testing revealed a need to improve dashboard navigability and to streamline content. CONCLUSIONS: The HESS dashboard is a potentially useful tool for presenting data on health equity to health plans. The multistage process of continual testing and improvement used to develop the dashboard could be a model for targeting and deciding upon quality improvement efforts in the domain of health equity.

Development and Validation of Algorithms to Predict Activity, Mobility, and Memory Limitations Using Medicare Claims and Post-Acute Care Assessments.

Functional impairment predicts mortality and health care utilization. However, validated measures of functional impairment are not routinely collected during clinical encounters and are impractical to use for large-scale risk-adjustment or targeting interventions. This study's purpose was to develop and validate claims-based algorithms to predict functional impairment using Medicare Fee-for-Service (FFS) 2014-2017 claims data linked with post-acute care (PAC) assessment data and weighted to better represent the overall Medicare FFS population. Using supervised machine learning, predictors were identified that best predicted two functional impairment outcomes measured in PAC data-any memory limitation and a count of 0-6 activity/mobility limitations. The memory limitation algorithm had moderately high sensitivity and specificity. The activity/mobility limitations algorithm performed well in identifying beneficiaries with five or more limitations, but overall accuracy was poor. This dataset shows promise for use in PAC populations, though generalizability to broader older adult populations remains a challenge.

Doubly robust evaluation of high-dimensional surrogate markers.

When evaluating the effectiveness of a treatment, policy, or intervention, the desired measure of efficacy may be expensive to collect, not routinely available, or may take a long time to occur. In these cases, it is sometimes possible to identify a surrogate outcome that can more easily, quickly, or cheaply capture the effect of interest. Theory and methods for evaluating the strength of surrogate markers have been well studied in the context of a single surrogate marker measured in the course of a randomized clinical study. However, methods are lacking for quantifying the utility of surrogate markers when the dimension of the surrogate grows. We propose a robust and efficient method for evaluating a set of surrogate markers that may be high-dimensional. Our method does not require treatment to be randomized and may be used in observational studies. Our approach draws on a connection between quantifying the utility of a surrogate marker and the most fundamental tools of causal inference-namely, methods for robust estimation of the average treatment effect. This connection facilitates the use of modern methods for estimating treatment effects, using machine learning to estimate nuisance functions and relaxing the dependence on model specification. We demonstrate that our proposed approach performs well, demonstrate connections between our approach and certain mediation effects, and illustrate it by evaluating whether gene expression can be used as a surrogate for immune activation in an Ebola study.

Measuring Equity in the Hospital Setting: An HCAHPS Application of the Health Equity Summary Score.

BACKGROUND: Health care quality varies by patient factors, including race-and-ethnicity and preferred language. Addressing inequities requires identifying them and incentivizing equity. OBJECTIVES: We apply an approach first implemented in the Medicare Advantage setting to measure equity in patient experiences by race-and-ethnicity [Asian American and Native Hawaiian or Pacific Islander (AA and NHPI), Black, Hispanic, vs. White] and language preference (English-preferring vs. another-language-preferring). We identify characteristics of hospitals providing high-quality equitable care. RESEARCH DESIGN: We estimated, standardized, and combined performance measures into a Health Equity Summary Score (HESS) using 2016-2019 Hospital Consumer Assessment of Healthcare Providers and Systems (HCAHPS) survey data. The HCAHPS HESS considered current cross-sectional performance, within-hospital improvement, and overall improvement by race-and-ethnicity and language preference. SUBJECTS: A total of 3333 US hospitals with 2019 HCAHPS Star Ratings. RESULTS: The HCAHPS HESS was calculable for 44% of hospitals. High-scoring (4-5 diamonds on a 1-diamond to 5-diamond scale) hospitals tended to be smaller than intermediate-scoring [3 diamonds (14% of high-scoring hospitals had <100 beds vs. 7% of intermediate-scoring hospitals, P <0.001) and were less often for-profit (20% vs. 31%, P <0.001)]. While a significant percentage (29%) of patients served by high-scoring hospitals were AA and NHPI, Black, or Hispanic, and 9% were another-language-preferring, there were smaller proportions of Black and Hispanic patients in high-scoring versus other hospitals. HESS performance was negatively associated with the percentage of patients preferring another language to English. HESS scores were moderately correlated with overall Star Ratings ( r =0.70). CONCLUSIONS: The HCAHPS HESS and practices of high-scoring hospitals could promote more equitable patient experiences.

Predicting Race And Ethnicity To Ensure Equitable Algorithms For Health Care Decision Making.

Algorithms are currently used to assist in a wide array of health care decisions. Despite the general utility of these health care algorithms, there is growing recognition that they may lead to unintended racially discriminatory practices, raising concerns about the potential for algorithmic bias. An intuitive precaution against such bias is to remove race and ethnicity information as an input to health care algorithms, mimicking the idea of "race-blind" decisions. However, we argue that this approach is misguided. Knowledge, not ignorance, of race and ethnicity is necessary to combat algorithmic bias. When race and ethnicity are observed, many methodological approaches can be used to enforce equitable algorithmic performance. When race and ethnicity information is unavailable, which is often the case, imputing them can expand opportunities to not only identify and assess algorithmic bias but also combat it in both clinical and nonclinical settings. A valid imputation method, such as Bayesian Improved Surname Geocoding, can be applied to standard data collected by public and private payers and provider entities. We describe two applications in which imputation of race and ethnicity can help mitigate potential algorithmic biases: equitable disease screening algorithms using machine learning and equitable pay-for-performance incentives.

Association of Postsurgical Opioid Refills for Patients With Risk of Opioid Misuse and Chronic Opioid Use Among Family Members.

Importance: The US health care system is experiencing a sharp increase in opioid-related adverse events and spending, and opioid overprescription may be a key factor in this crisis. Ambient opioid exposure within households is one of the known major dangers of overprescription. Objective: To quantify the association between the postsurgical initiation of prescription opioid use in opioid-naive patients and the subsequent prescription opioid misuse and chronic opioid use among opioid-naive family members. Design, Setting, and Participants: This cohort study was conducted using administrative data from the database of a US commercial insurance provider with more than 35 million covered individuals. Participants included pairs of patients who underwent surgery from January 1, 2008, to December 31, 2016, and their family members within the same household. Data were analyzed from January 1 to November 30, 2018. Exposures: Duration of opioid exposure and refills of opioid prescriptions received by patients after surgery. Main Outcomes and Measures: Risk of opioid misuse and chronic opioid use in family members were calculated using inverse probability weighted Cox proportional hazards regression models. Results: The final cohort included 843 531 pairs of patients and family members. Most pairs included female patients (445 456 [52.8%]) and male family members (442 992 [52.5%]), and a plurality of pairs included patients in the 45 to 54 years age group (249 369 [29.6%]) and family members in the 15 to 24 years age group (313 707 [37.2%]). A total of 3894 opioid misuse events (0.5%) and 7485 chronic opioid use events (0.9%) occurred in family members. In adjusted models, each additional opioid prescription refill for the patient was associated with a 19.2% (95% CI, 14.5%-24.0%) increase in hazard of opioid misuse in family members. The risk of opioid misuse appeared to increase only in households in which the patient obtained refills. Family members in households with any refill had a 32.9% (95% CI, 22.7%-43.8%) increased adjusted hazard of opioid misuse. When patients became chronic opioid users, the hazard ratio for opioid misuse among family members was 2.52 (95% CI, 1.68-3.80), and similar patterns were found for chronic opioid use. Conclusions and Relevance: This cohort study found that opioid exposure was a household risk. Family members of a patient who received opioid prescription refills after surgery had an increased risk of opioid misuse and chronic opioid use.

Is Bigger Better?: A Closer Look at Small Health Systems in the United States.

BACKGROUND: Research on US health systems has focused on large systems with at least 50 physicians. Little is known about small systems. OBJECTIVES: Compare the characteristics, quality, and costs of care between small and large health systems. RESEARCH DESIGN: Retrospective, repeated cross-sectional analysis. SUBJECTS: Between 468 and 479 large health systems, and between 608 and 641 small systems serving fee-for-service Medicare beneficiaries, yearly between 2013 and 2017. MEASURES: We compared organizational, provider and beneficiary characteristics of large and small systems, and their geographic distribution, using multiple Medicare and Internal Revenue Service administrative data sources. We used mixed-effects regression models to estimate differences between small and large systems in claims-based Healthcare Effectiveness Data and Information Set (HEDIS) quality measures and HealthPartners' Total Cost of Care measure using a 100% sample of Medicare fee-for-service claims. We fit linear spline models to examine the relationship between the number of a system's affiliated physicians and its quality and costs. RESULTS: The number of both small and large systems increased from 2013 to 2017. Small systems had a larger share of practice sites (43.1% vs. 11.7% for large systems in 2017) and beneficiaries (51.4% vs. 15.5% for large systems in 2017) in rural areas or small towns. Quality performance was lower among small systems than large systems (-0.52 SDs of a composite quality measure) and increased with system size up to ∼75 physicians. There was no difference in total costs of care. CONCLUSIONS: Small systems are a growing source of care for rural Medicare populations, but their quality performance lags behind large systems. Future studies should examine the mechanisms responsible for quality differences.

Measuring Inconsistency in Quality Across Patient Groups to Target Quality Improvement.

BACKGROUND: Quality improvement (QI) may be aimed at improving care for all patients, or it may be targeted at only certain patient groups. Health care providers have little guidance when determining when targeted QI may be preferred. OBJECTIVES: The aim was to develop a method for quantifying performance inconsistency and guidelines for when inconsistency indicates targeted QI, which we apply to the performance of health plans for different patient groups. RESEARCH DESIGN AND MEASURES: Retrospective analysis of 7 Health Care Effectiveness Data and Information Set (HEDIS) measures of clinical care quality. SUBJECTS: All Medicare Advantage (MA) beneficiaries eligible for any of 7 HEDIS measures 2015-2018. RESULTS: MA plans with higher overall performance tended to be less inconsistent in their performance (r=-0.2) across groups defined by race-and-ethnicity and low-income status (ie, dual eligibility for Medicaid or receipt of Low-Income Subsidy). Plan characteristics were usually associated with only small differences in inconsistency. The characteristics associated with differences in consistency [eg, size, Health Maintenance Organization (HMO) status] were also associated with differences in overall performance. We identified 9 (of 363) plans that had large inconsistency in performance across groups (>0.8 SD) and investigated the reasons for inconsistency for 2 example plans. CONCLUSIONS: This newly developed inconsistency metric may help those designing and evaluating QI efforts to appropriately determine when targeted QI is preferred. It can be used in settings where performance varies across groups, which can be defined by patient characteristics, geographic areas, hospital wards, etc. Effectively targeting QI efforts is essential in today's resource-constrained health care environment.