Introduction and Objective: COVID-19 infection has been associated with an increased risk of diabetes mellitus (DM). Limited data is available regarding whether patients with COVID-19 infection preceding DM have different disease severity than those without a prior infection. The goal of this study was to assess diabetes-related outcomes among patients with positive vs. negative COVID-19 tests prior to a new diagnosis. Methods: We conducted a retrospective cohort study using electronic health record data from 31 healthcare systems participating in the PCORnet national research network and the NIH RECOVER program. Eligible patients were 20-80 years old with a COVID-19 test (positive or negative) prior to a new DM diagnosis from March 2020 to December 2022. Outcomes were HbA1c at DM diagnosis, healthcare utilization, and long-acting insulin use in the 365 days after diagnosis. Adjusted multivariable linear and logistic regression analyses evaluated the association between COVID-19 infection (i.e., positive test) and these outcomes. Results: A total of 26,936 patients were diagnosed with new DM after a COVID-19 test (19% positive, 81% negative). We found little association between test result and mean HbA1c (estimated mean HbA1c difference: 0.09, 95% CI: 0.01, 0.16), though HbA1c was higher among patients diagnosed 50-200 days after testing positive. A positive test was associated with a lower risk of an inpatient encounter (OR: 0.84, 95% CI: 0.78 - 0.90) and a higher risk of an emergency department encounter (OR: 1.15, 95% CI: 1.08, 1.23). No difference was evident in long-acting insulin use (OR: 1.02, 95% CI: 0.88 - 1.18). Conclusion: Shorter latency between COVID-19 infection and DM diagnosis was associated with slightly higher HbA1c. COVID-19 infection was associated with modest differences in health care utilization, but no difference in long-acting insulin use. More research is needed to further evaluate the nature of the association between COVID-19 and DM progression. Disclosure J.G. Lyons: None. S. Rifas-Shiman: None. J.G. Young: None. T. Eggerman: None. B. French: None. M. Hivert: None. S. McGrath: None. A.C. Powers: None. B. Rasouli: None. R.L. Rothman: Research Support; Current; Dexcom, Inc. J.S. Schildcrout: None. J.P. Block: None. Funding National Institutes of Health, NIDDK (U01DK137533)

We estimated sex-specific population effects of hypothetical interventions to limit sugar sweetened-beverages (SSBs) and 100% fruit juice throughout childhood on central adiposity, insulin resistance, and glycemic outcomes in adolescence in Project Viva prebirth cohort. Among 481 females and 491 males, mothers reported beverage intake from 3 to 10 years from a food frequency questionnaire. The primary outcome was the homeostatic model assessment for insulin resistance (HOMA-IR), and secondary outcomes were waist circumference, truncal fat mass, fasting glucose, and glycated hemoglobin in late adolescence. We applied inverse probability weighting of longitudinal marginal structural models to account for baseline and time-varying confounding, and censoring. We estimated that limiting SSBs to 1 serving weekly across childhood would reduce HOMA-IR by 0.28 units (95% confidence interval [CI], -0.61 to 0.02), waist circumference by 1.91 cm (95% CI, -3.79 to -0.05), truncal fat mass by 0.64 kg (95% CI, -1.33 to 0.05), and fasting glucose by 1.02 mg/dL (95% CI, -2.40 to 0.35) in males compared to no intervention. In females, effect estimates were near 0 and less precise than males. Effect estimates for 100% fruit juice were small, with imprecise CI in both sexes. Overall, limiting SSBs in childhood may have small effects on insulin resistance, central adiposity, and glycemia in males in this population of low consumers. Trial registration: NCT02820402; https://clinicaltrials.gov/study/NCT02820402.

In the presence of competing events, many investigators are interested in a direct treatment effect on the event of interest that does not capture treatment effects on competing events. Classical survival analysis methods that treat competing events like censoring events, at best, target a controlled direct effect: the effect of the treatment under a difficult to imagine and typically clinically irrelevant scenario where competing events are somehow eliminated. A separable direct effect, quantifying the effect of a future modified version of the treatment, is an alternative direct effect notion that may better align with an investigator's underlying causal question. In this paper, we provide insights into the implications of naively applying an estimator constructed for a controlled direct effect (i.e., "censoring by competing events") when the actual causal effect of interest is a separable direct effect. We illustrate the degree to which controlled and separable direct effects may take different values, possibly even different signs, and the degree to which these two different effects may be differentially impacted by violation and/or near violation of their respective identifying conditions under a range of data generating scenarios. Finally, we provide an empirical comparison of inverse probability of censoring weighting to an alternative weighted estimator specifically structured for a separable effect using data from a randomized trial of estrogen therapy and prostate cancer mortality.

BACKGROUND:Antihypertensive class medications are prescribed long-term for adolescents, including for conditions other than hypertension. Evidence on weight effects is limited. OBJECTIVE:To assess the weight effects of commonly prescribed antihypertensive class medications among adolescents, regardless of clinical indication for use. METHODS:This retrospective study was conducted using electronic health record (EHR) data from a multi-site US research network. The study cohort included adolescents aged 13.0-19.5 years with ≥ 1 antihypertensive class medication order during 2010-2019. Weight and height documented in the EHR were used to calculate BMI z-scores. Inverse probability weighting with marginal structural models was used to account for selection factors and covariates. RESULTS:Overall, 23 853 adolescents were prescribed an antihypertensive class medication, including clonidine (prescribed to n = 7491), guanfacine (n = 6411), lisinopril (n = 3584), propranolol (n = 3015), spironolactone (n = 1926), and atenolol (n = 1426). At baseline, 13% had a hypertension diagnosis. At 12 months after medication initiation, the estimated population-level BMI z-score was significantly increased for propranolol (BMI z-score 0.19 [95% CI 0.09, 0.29]) and significantly decreased for lisinopril (-0.06 [95% CI -0.10, -0.01]); clonidine, guanfacine, spironolactone, and atenolol were not associated with a significant change in BMI z-score. CONCLUSIONS:Most antihypertensive class medications were not associated with statistically significant weight gain when prescribed to adolescents.

BACKGROUND:Long-acting cabotegravir (CAB-LA) was approved as HIV preexposure prophylaxis (PrEP) in the United States in December 2021, but data are limited on uptake, adherence, and persistence in clinical practice. METHODS:We extracted electronic health records of adults receiving oral or injectable PrEP during December 2021 to June 2024 at Kaiser Permanente (KP) Northern California and Mid-Atlantic States, 2 large integrated healthcare systems. We used χ 2 tests to compare characteristics of CAB-LA users and oral-PrEP-only users. Among CAB-LA users, we assessed adherence to bimonthly injections after lead-in doses (weeks 0 and 4) and used Kaplan-Meier methods to estimate persistence. RESULTS:Among 23,311 individuals accessing oral or injectable PrEP, 180 (0.8%) received CAB-LA, with 23.9% having no documentation of previous PrEP use at KP. Compared with oral-PrEP-only users, a lower proportion of CAB-LA users were commercially insured (82.2% vs 89.2%; P = 0.014) and a higher proportion were Black (18.9% vs 10.2%) or Hispanic (34.4% vs 23.6%; P < 0.001 across race/ethnicity categories). Of 688 non-lead-in CAB-LA injections, 90.4% were administered within 8 weeks +7 days after the previous injection. Persistence on CAB-LA was 87.9% and 74.9% at 6 and 12 months, respectively. There were no incident HIV infections during CAB-LA use. CONCLUSIONS:CAB-LA is engaging new users, including populations traditionally underrepresented in PrEP uptake, and adherence and persistence are high in clinical practice. However, uptake of CAB-LA is extremely low, suggesting population affect will be limited without efforts to expand implementation and use.

Longitudinal causal inference is concerned with defining, identifying, and estimating the effect of a time-varying intervention on a time-varying outcome that is indexed by a follow-up time. In an observational study, Robins’s generalized g-formula can identify causal effects induced by a broad class of time-varying interventions. Various methods for estimating the generalized g-formula have been posed for different outcome types, such as a failure event indicator by a specified time (e.g. mortality by 5 year follow-up), as well as continuous or dichotomous/multi-valued outcomes measures at a specified time (e.g. blood pressure in mmHg or an indicator of high blood pressure at 5-year follow-up). Multiply-robust, data-adaptive estimators leverage flexible nonparametric estimation algorithms while allowing for statistical inference. However, extant methods do not accommodate pooling estimation across time points when multiple outcomes are measured over time, which can lead to substantial loss of precision. We propose a novel multiply-robust estimator of the generalized g-formula that accommodates pooled estimation over numerous available outcome measures, which we refer to as time-smoothing. Our approach accommodates any intervention that can be described as a Longitudinal Modified Treatment Policy, a flexible class suitable for binary, multi-valued, and continuous longitudinal treatments. Our method produces an estimate of the effect curve: the causal effect of the intervention on the outcome at each measurement time, taking into account censoring and non-monotonic outcome missingness patterns. In simulations we find that the proposed algorithm outperforms extant multiply-robust approaches for effect curve estimation in scenarios with high degrees of outcome missingness and when there is low covariate overlap. We apply the method to study longitudinal effects of union membership on wages. The proposed estimator is available in the LMTP package at https://github.com/lmtp/tree/curve .

Introduction and Objective: Prior research identified genetic variants for macronutrient preference in adults, but whether these genetic differences affect early life food intake and metabolic phenotypes is unknown. We examined the associations of polygenic scores (PS) for macronutrient preference with dietary behaviors, BMI trajectories from 3-17y and insulin resistance at 17y. Methods: We used genetics, dietary intake and anthropometrics collected at ages ~3, 7, 12, and 17y, and glycemic data measured at 17y among non-Hispanic White children from US prebirth cohort Project Viva. Dietary behaviors from questionnaire included consumption of sugar-sweetened beverages (SSBs) and fast-food. Macronutrient preference PS were calculated from prior large genome-wide study in adults. We estimated odds ratios (OR) for frequent fast-food (>=1 meal/wk) and SSB intake (>1 serving/wk) with generalized estimating equations from 3-17y, and beta coefficients for BMI z-score with linear mixed models from 3-17y. We used linear regressions to examine associations of PS with homeostatic model assessment for insulin resistance (HOMA-IR). We adjusted models for child’s age, sex, and 3 genetic principal components. Results: Among 516 children, each SD increment in PS for carbohydrate preference was associated with a higher likelihood of frequent fast-food (OR 1.13 for >=1 meal/wk, 95% CI: 1.02 to 1.26) and SSB (OR 1.15 for >1 serving/wk, 95% CI: 1.03 to 1.28) intake on average from 3-17y. Each SD increase in the protein preference PS was associated with lower likelihood of frequent SSB intake (OR for >1 serving/wk: 0.89, 95% CI: 0.79 to 0.99). We did not find significant associations between macronutrient PS with BMI z-score or HOMA-IR. Conclusion: Our findings suggest that genetic predisposition for carbohydrate and protein preference could influence SSB and fast-food intake patterns in children, which could heighten their future risk for diabetes. S. Harnois-Leblanc: None. K. Switkowski: None. J. Young: None. I.M. Aris: None. S.L. Rifas-Shiman: None. E. Oken: None. J.E. Gervis: None. K.E. Westerman: None. H.S. Dashti: None. M. Hivert: None. J. Merino: None. American Diabetes Association (7-23-PDFT2DY-03); National Institutes of Health (R01034568, R24ES030894)

Researchers are often interested in estimating effects of generalized time-varying treatment strategies on the mean of an outcome at one or more selected follow-up times of interest. For example, the Medications and Weight Gain in PCORnet (MedWeight) study aimed to estimate effects of adhering to flexible medication regimes on future weight change using electronic health records (EHR) data. This problem presents several methodological challenges that have not been jointly addressed in the prior literature. First, this setting involves treatment strategies that vary over time and depend dynamically and non-deterministically on measured confounder history. Second, the outcome is repeatedly, non-monotonically, informatively, and sparsely measured in the data source. Third, some individuals die during follow-up, rendering the outcome of interest undefined at the follow-up time of interest. In this article, we pose a range of inverse probability weighted (IPW) estimators targeting effects of generalized time-varying treatment strategies in truncation by death settings that allow time-smoothing for precision gain. We conducted simulation studies that confirm precision gains of the time-smoothed IPW approaches over more conventional IPW approaches that do not leverage the repeated outcome measurements. We illustrate an application of the IPW approaches to estimate comparative effects of adhering to flexible antidepressant medication strategies on future weight change. The methods are implemented in the accompanying R package, smoothedIPW.

OBJECTIVE:The advent of biological drugs has revolutionised management of inflammatory bowel disease (IBD). However, the extent to which these novel pharmacological drugs have reduced the need for surgical treatment remains incompletely quantified.We aimed to investigate the risk of first, major surgery in IBD in a population-based, large epidemiological study. METHODS:We empanelled a cohort comprising all 85 974 patients diagnosed with ulcerative colitis (UC) and 42 760 with Crohn's disease (CD) in Norway and Sweden in 1987 through 2017. We used log-rank tests to compare the cumulative probability of surgical treatment for UC and CD. Using multivariable Cox proportional hazards models, we estimated hazard ratios (HR) with 95% CIs by year of diagnosis, age, sex and extent of disease. RESULTS:During a mean follow-up of 9.9 years, surgery was undertaken in 11 187 (13.0%) patients with UC (12.3 per 1000 person-years) and in 11 307 (26.4%) patients with CD (30.0 per 1000 person-years). In UC, the cumulative 5-year probability of surgery decreased from 16.2% in patients diagnosed in 1987-1994 to 5.8% in those diagnosed in 2011-2017 (p<0.001). In CD, the corresponding decline was from 30.1% to 13.9% (p<0.001). In multivariable analyses, the likelihood of surgical treatment decreased during the study period by 61% (HR 0.39, 95% CI 0.36 to 0.42) in UC and by 31% (HR 0.69, 95% CI 0.65 to 0.75) in CD. CONCLUSIONS:Following the introduction of biologic drugs, the need for surgical treatments has been dramatically reduced in patients with UC and moderately reduced in patients with CD.

OBJECTIVE:This study aimed to estimate population-level effects on weight change of initiating/adhering to additional glucose-lowering medications in adults with type 2 diabetes prescribed metformin. METHODS:We conducted a target trial using electronic health record data from 22,601 patients (age 20 to < 80 years) prescribed metformin to determine initiation/adherence to dipeptidyl peptidase IV (DPP4) inhibitors, glucagon-like peptide-1 receptor agonists (GLP-1RAs), sodium-glucose cotransporter 2 (SGLT-2) inhibitors, long-acting insulin, or sulfonylureas. Inverse probability weighting of marginal structural models with standardization by baseline covariates was used to estimate population-level effects of initiating/adhering to different medications on average 24-month weight change. RESULTS:At 24 months, a mean -5.15 kg (95% CI -10.6, -1.36) and -6.71 kg (95% CI -8.38, -4.34) weight loss would be observed for initiation/adherence to GLP-1RAs and SGLT-2s respectively. At 6 months, weight loss for DPP4s would be observed (-0.89 kg, 95% CI -1.41, -0.32) though not at 12 or 24 months. Glimepiride would be associated with weight gain at 6 and 12 months (0.88 kg, 95% CI 0.44, 1.22; 1.01 kg, 95% CI 0.32, 1.51) but not at 24 months. CONCLUSIONS:Initiation/adherence to GLP-1RAs and SGLT-2s over 24 months could result in average weight losses of 5.15 kg and 6.71 kg, respectively.

The inability to identify dates of death in insurance claims data in the United States is a major limitation to retrospective claims-based research. Although deaths result in disenrollment, disenrollment can also occur due to changes in insurance providers. We created an algorithm to differentiate between disenrollment from health plans due to death and disenrollment for other reasons. We identified 5 259 735 adults who disenrolled from private insurance between 2007 and 2018. Using death dates ascertained from the Social Security Death Index, inpatient discharge status, and death indicators in the administrative data, 7.6% of all disenrollments were classified as resulting from death. We used elastic net regression to build an algorithm using claims data in the year prior to disenrollment; candidate predictors included medical conditions, individual demographic characteristics, treatment utilization, and structural factors related to health insurance eligibility and coding. Using a predicted probability threshold of 0.9 (selected to reflect the corresponding known prevalence of mortality), internal validation found that the algorithm classified death at disenrollment with a positive predictive value of 0.815, sensitivity of 0.721, and specificity of 0.986 (area under the curve = 0.97). Independent data sources were used for external validation and for an applied example. Code for implementation is publicly available.

Importance:Doxycycline postexposure prophylaxis (doxyPEP) has been shown to decrease the incidence of bacterial sexually transmitted infections (STIs) among people assigned male sex at birth in clinical trials, but data from clinical practice are limited. Objective:To describe early uptake of doxyPEP and evaluate changes in STI incidence following doxyPEP initiation. Design, Setting, and Participants:This retrospective cohort study of adults (aged ≥18 years) dispensed HIV preexposure prophylaxis (PrEP) at Kaiser Permanente Northern California during November 1, 2022, to December 31, 2023, examined electronic health record data to compare HIV PrEP users dispensed and not dispensed doxyPEP and rates of bacterial STIs before and after starting doxyPEP. Individuals were followed up from their first recorded STI test on or after November 1, 2020, until December 31, 2023, or discontinuation of health plan membership. Exposure:Pharmacy dispensing data were used to define doxyPEP recipients. Main Outcomes and Measures:Demographic and clinical characteristics were compared between individuals dispensed and not dispensed doxyPEP. Primary outcomes were incident chlamydia, gonorrhea, or infectious syphilis measured as quarterly STI positivity (proportion of individuals testing positive at least once per quarter). Among doxyPEP recipients, rate ratios (RRs) compared mean quarterly STI positivity from 24 months before to 12 months after starting doxyPEP. In an exploratory analysis, STI trends were evaluated for the full cohort, stratified by receipt of doxyPEP. Results:Among 11 551 HIV PrEP users (mean [SD] age, 39.9 [12.1] years; 95.1% male), 2253 (19.5%) were dispensed doxyPEP, of whom 2228 (98.9%) were male and 1096 (48.6%) had an STI in the year before starting doxyPEP. Compared with individuals not dispensed doxyPEP, doxyPEP recipients were older (mean [SD] age, 40.4 [10.8] vs 39.8 [12.4] years; P = .04) and had used HIV PrEP longer (mean [SD], 4.2 [2.8] vs 3.4 [2.6] years; P < .001), and a higher proportion were commercially insured (2091 [92.8%] vs 8270 [88.9%]; P < .001). Among doxyPEP recipients, quarterly chlamydia positivity decreased from 9.6% (95% CI, 9.0%-10.3%) before starting doxyPEP to 2.0% (95% CI, 1.5%-2.6%) after starting doxyPEP (RR, 0.21; 95% CI, 0.16-0.27; P < .001), with significant declines for each anatomic site of infection. Quarterly gonorrhea positivity decreased from 10.2% (95% CI, 9.6%-10.9%) before starting doxyPEP to 9.0% (95% CI, 8.0%-10.1%) after starting doxyPEP (RR, 0.88; 95% CI, 0.77-1.00; P = .048); site-specific declines were significant for rectal (RR, 0.81; 95% CI, 0.67-0.97; P = .02) and urethral (RR, 0.56; 95% CI, 0.40-0.79; P = .001) gonorrhea, but not pharyngeal gonorrhea. Quarterly syphilis positivity decreased from 1.7% (95% CI, 1.4%-1.9%) before starting doxyPEP to 0.3% (95% CI, 0.2%-0.6%) after starting doxyPEP (RR, 0.20; 95% CI, 0.11-0.37; P < .001). Positivity for STIs remained stable in individuals not dispensed doxyPEP. Conclusions and Relevance:This study found that receipt of doxyPEP was associated with substantial declines in chlamydia and syphilis incidence and modest declines in urethral and rectal gonorrhea incidence among individuals using HIV PrEP. These findings suggest that doxyPEP may offer substantial benefits for reducing population-level STI transmission with broader implementation.

Weight gain after starting antihypertensive medications is a frequent concern for patients, but there is limited data on expected weight change after initiation of these medications. A comparative effectiveness trial to evaluate this outcome would not be feasible. To estimate and compare average weight change under initiating and adhering to commonly prescribed, first-line antihypertensive medications as monotherapy by emulating a target trial. Retrospective observational cohort study over 24 months of follow-up using electronic health records (EHR). 141,260 patients prescribed one of seven antihypertensives between 2010 and 2019 across 8 US health systems. We examined mean weight change associated with initiation of and adherence to amlodipine, atenolol, hydrochlorothiazide, losartan, metoprolol, or propranolol, relative to lisinopril, at 6, 12, and 24 months after initiation. To adjust for baseline confounding and informative outcome measurement, we used inverse probability weighting with repeated outcome marginal structural models. After baseline and time-varying covariate adjustment, initiation of and adherence to lisinopril were associated with mean weight loss at 6 months (− 0.69 kg, 95

Limited real-world evidence exists on the patterns of symptomatic treatment use for Alzheimer’s disease and related dementias (ADRD). Using data from the Merative® MarketScan Research Databases, we analyzed treatment patterns in ADRD patients aged ≥ 65 from 2011 to 2021. Initial treatment choices, subsequent changes, and factors influencing these changes were evaluated over an 18-month period. Among 74,212 acetylcholinesterase inhibitor (AChEI) and 15,917 memantine initiators, 56

OBJECTIVE:To estimate long-term weight change after initiation and adherence to commonly used antiseizure medications (ASMs) and examine differences in weight change across ASMs compared to topiramate. METHODS:We included 52,309 adult patients who initiated ASMs, applied a target trial emulation approach to control time-varying confounding and selection bias, and examined the long-term comparative effects on weight change after initiating and adhering to different ASMs at 6 and 12 months post initiation. RESULTS:The most commonly initiated ASM was topiramate (41.2%). In comparison to topiramate, we estimated higher 6-month weight change under initiation and adherence to levetiracetam 0.94 kg (95% CI 0.20, 1.64), lamotrigine 1.44 kg (0.74, 1.99), valproate 2.42 kg (1.71, 2.88), carbamazepine 1.32 kg (0.46, 2.16), and oxcarbazepine 1.74 kg (0.85, 2.71), with similar results at 12 months and in sensitivity and subgroup analyses. These results were driven mostly by weight loss with use of topiramate rather than weight gain with use of other ASMs. Results were similar though attenuated when accounting for medication initiation only. CONCLUSIONS:Topiramate was associated with weight loss at 6 and 12 months under either initiation and subsequent adherence or initiation-only effects; other medications were associated with higher weight change. These results provided important information to help with decision-making regarding ASM initiation.