Purpose:Selection of osteoporosis (OP) treatment is affected by patients' disease severity and fracture risk, potentially confounding real-world comparative effectiveness and safety studies of antiresorptive medications. To inform the choice of valid treatment contrasts for subsequent real-world comparative studies, we assessed comparability of antiresorptive OP treatment groups using negative control outcomes (NCOs). Patients and Methods:Women aged ≥55 years in Optum© Clinformatics® Data Mart from October 2010 through June 2019 who received denosumab, zoledronic acid (ZA), or oral bisphosphonates (BPs) were included. We estimated the 1-year cumulative risks for 12 NCOs by treatment group among treatment-naïve and treatment-experienced women using augmented inverse-probability of treatment and censoring weighted (AIPW) estimation. A Bayesian sensitivity analysis was conducted to aggregate estimates and associated variances into a form characterized by magnitude and probability. Results:Women in both treatment-naïve (n = 199,335) and treatment-experienced (n = 33,296) cohorts initiated treatment at a mean age of 71.8 years. Treatment-naïve women initiating denosumab had similar 1-year risks of most NCOs compared with initiators of ZA (maximum observed RD = 2.47% for colon cancer screening). However, significant risk differences were observed for seven NCOs when comparing ZA or denosumab with oral BPs. Among treatment-experienced women, all NCOs indicated similar risks when comparing denosumab to alendronate alone. Only one NCO (dementia: RD = 0.42%) was associated with treatment when comparing denosumab to oral BPs, and one (influenza vaccine: RD = 3.57%) was associated with treatment when comparing ZA to oral BPs. Results of the Bayesian analysis aligned with our qualitative interpretations. Conclusion:Comparative studies including denosumab or ZA versus oral BPs among treatment-experienced, commercially insured women aged ≥55 years in the United States are likely valid with respect to comparability of treatment groups. Our results do not support conducting observational studies examining these treatment contrasts in the overall treatment-naïve population. However, comparison of ZA and oral BPs among treatment-naïve women with a prior fracture may be undertaken with minimal expected residual bias.

BACKGROUND:Studies comparing treatment strategies based on initiation timing-such as starting PCSK9 inhibitor (PCSK9i) therapy sooner versus later after a myocardial infarction (MI)-are prone to immortal time bias. Clone-censor-weight methods can address these issues and allow the researcher to emulate a trial in which patients are assigned to protocols dictating when PCSK9i is initiated. This study aimed to evaluate the comparability of patients in a clone-censor-weight setup who initiated a PCSK9i within 12 months post-MI versus non-initiators. METHODS:We included adult patients hospitalized for MI in Sweden (2015-2021) and followed them for 3 years. We considered two treatment strategies: initiating PCSK9i within 12 months versus not initiating PCSK9i during the same period. We applied the clone-censor-weight method to address immortal time bias and assessed remaining bias using covariate balance metrics and negative control outcomes. RESULTS:The primary study sample included 38  627 episodes of MI, with 561 (1.5%) initiating PCSK9i treatment within 12 months. These patients were younger, had higher baseline LDL-C levels, and were more frequently treated with ezetimibe during their post-MI follow-up compared to non-initiators. Although clone-censor-weight estimation was free of immortal time bias, it faced challenges in achieving adequate balance of covariates due to the high rates of censoring (relatively small number of people initiating a PCSK9i in the first year) and strong association between covariates and censoring. Truncation of weights provided more stable estimates but at the expense of some covariate imbalances. CONCLUSIONS:The clone-censor-weight method is a promising approach that allows researchers to answer questions about the effect of treatment policies. But practical guidance is needed to address problems that arise from small, highly imbalanced groups, which is common with most newly introduced treatments.

Objective A lack of transparent reporting of race and ethnicity in clinical research limits the ability to identify health inequities and evaluate to what extent clinical research includes diverse populations. Our objectives are: (1) to identify study characteristics associated with reporting race and ethnicity of clinical study participants and (2) to document temporal trends in race and ethnicity reporting on ClinicalTrials.gov.Design Cross-sectional analysis of interventional trials and observational studies from 2009 to 2024; multivariable logistic regression assessed study-level factors associated with reporting race and ethnicity.Setting Global registry of clinical studies (ClinicalTrials.gov).Participants 58 163 studies with posted results and without early termination.Exposures Study characteristics: sponsor trial phase, study type and country.Main outcomes and measures Reporting of race, reporting of ethnicity, reporting of both.Results Among 58163 studies (mean enrolment=1215 participants), 44.8% did not report race or ethnicity to the repository (mean enrolment=1481 participants). The proportion of studies reporting both race and ethnicity rose from 7.4% in 2013 to 54.6% in 2024. In multivariable models, observational studies had lower odds of reporting race and ethnicity (OR 0.55, 95% CI 0.49 to 0.61) compared with interventional trials. Phase 4 trials were least likely phase to report race and ethnicity (OR=0.32; 95% CI 0.29 to 0.35), and studies with only National Institute of Health funding were more likely to report race and ethnicity compared with studies with any industry funding or sponsorship (OR=1.70, 95% CI 1.61 to 1.79). For studies that reported race, White participants comprised ≥50% each year based on study-level percentages; proportions of Asian participants declined, and Black participants fluctuated. ‘Not Hispanic or Latino’ remained ≥80% of reported ethnicity annually.Conclusions Race and ethnicity reporting on ClinicalTrials.gov has improved markedly yet remains incomplete, with shortfalls in late-phase and observational studies.

BACKGROUND:Discrimination in health care can influence patient behavior and potentially lead to poor quality care. When patients perceive discrimination, they may disengage from health care, have heightened stress, and identify biased treatment practices. Atherosclerotic cardiovascular disease (ASCVD) is a condition that requires adequate disease prevention. Our objective is to assess the association between discrimination in health care and ASCVD risk. METHODS:We examined data from adults aged 50 to 90 years enrolled in the 2008 to 2020 waves of the Health and Retirement Study who were followed for up to 12 years. Participants reported how frequently they perceived receiving poorer treatment than other people from doctors or hospitals; we characterized this as discrimination. Nonfatal ASCVD events were ascertained from dates of a doctor-diagnosed myocardial infarction or stroke during follow-up. Cox models were used to estimate hazards of ASCVD outcomes with experiencing discrimination using propensity score weights to adjust for confounding. Models included covariate adjustment, and differences by sex, race, and ethnicity were examined. RESULTS:Of the 17 632 study participants (mean age: 65.86, 41.97% male), 3347 (18.9%) reported discrimination in health care at baseline, and 1785 (10.1%) had an ASCVD event over 10 years of follow-up. Among those who did not have an event, 2983 (18.82%) reported discrimination, and among those who had an event, 364 (20.39%) reported discrimination. Discrimination was associated with higher risks of nonfatal ASCVD within the first 2-years of follow-up (hazard ratio [HR], 1.54 [95% confidence interval (CI), 1.24-1.91]) and was partially attenuated after 5 years (HR, 1.29 [CI, 1.11-1.50]) and 10 years (HR, 1.27 [CI, 1.10-1.46]). The associations remained largely unchanged after covariate adjustments (2-year HR, 1.44 [CI, 1.16-1.80]; 5-year HR, 1.24 [CI, 1.06-1.44]; 10-year HR, 1.23 [CI, 1.10-1.46]). CONCLUSIONS:Discrimination in health care was associated with increased risks of nonfatal ASCVD in middle-aged and older adults. The risks persisted over time and suggest that discrimination is an important factor to consider for the prevention of ASCVD.

A common approach used to estimate per-protocol effects, or the effect of perfect adherence to a defined protocol on a specified outcome, is to artificially censor individuals when they deviate from the protocol. As adherence is not under control of the investigator, selection bias is a threat to the validity of per-protocol analyses. Here, we describe nonparametric bounds for the per-protocol effect. Importantly, these bounds describe the range of all possible per-protocol effects concordant with the observed data under the sole additional assumption—beyond those of a standard intent-to-treat analysis—that adherence was accurately measured. Further, these bounds naturally follow from an intuitive understanding of the “best” and “worst” cases among trial participants. Finally, the per-protocol bounds are straightforward to compute with standard statistical analysis software. We describe and illustrate these bounds with two open-source data sets, one in the context of a binary outcome and the other with a survival outcome. These examples are accompanied by open-source code in several different software programs to aid implementation. Per-protocol bounds offer a simple and intuitive assessment of the per-protocol effect under relatively weak assumptions and offer insight into the strength of the assumptions like no selection bias.

Sequential nested trial (SNT) emulation is a powerful approach for maximizing precision and avoiding time-related biases. However, there exists little discussion about the implied causal estimands in comparison to a real-world single point trial. We used Monte Carlo simulation to compare treatment effect estimates from an SNT emulation that re-indexed patients annually and a SNT emulation with a treatment decision design to the estimates from a single point trial. We generated 5,000 cohorts of 5,000 people with 3 years of follow-up. For the single point trial, patients were randomized to initiate or not initiate treatment at Visit 1. For the SNT emulations, simulated patients could contribute up to two index dates. When disease severity did not modify the treatment effect, both SNT approaches returned treatment effect estimates identical to the single point trial. In the presence of treatment effect modification by disease severity, both SNT approaches returned treatment effect estimates that diverged from the single point trial even after confounding-adjustment. These findings underscore the difficulties of interpreting causal estimands from a SNT emulation: the target population does not correspond to a single time point trial. Such implications are important for communicating study results for evidence-based decision-making.

BACKGROUND:Improvements in breast cancer therapy since the randomized controlled trials of mammography screening might have reduced the screening benefit. Most observational studies of mammography effectiveness would be confounded by these improvements and other factors. Using a design resistant to this confounding, we evaluated whether mammography in asymptomatic women reduces breast cancer mortality during the treatment era succeeding the trials. METHODS:We designed a quasi-experimental cohort study in regions of Denmark without organized screening. We predicted the number of expected mammograms for each general practice based on observed numbers of mammograms and individual risk factors. Regardless of a woman's individual exposure to mammography, we assigned her the ratio of observed to expected mammograms of her general practice as her instrumental variable. We employed this potential instrumental variable as mammography exposure status and followed women from 1 January 2006 until death, emigration, or 31 December 2014, whichever came first. RESULTS:We included 169,197 women aged 50-66 from 738 general practices and without previous breast cancer as of 1 January 2006. Women affiliated with a practice referring more women than expected, compared with less, had a lower hazard of breast cancer death (hazard ratio = 0.80; 95% confidence interval = 0.68, 0.95). Negative control associations were near null, suggesting no confounding bias. CONCLUSIONS:This quasi-experimental study estimated a continued protective effect of mammography in women where most were presumably asymptomatic. In contrast to conventional observational studies, the use of practice referral ratio as an instrumental variable may avoid bias from uncontrolled confounding.

PURPOSE:Linking claims data to electronic health record (EHR) data can improve completeness, often at a cost of decreased sample size. Quantifying information gained and differences in patient characteristics between EHR and EHR-claims linked cohorts may inform study design. METHODS:Using ConcertAI Patient360 EHR linked to multiple closed insurance claims sources, we compared an EHR cohort of patients with incident metastatic breast cancer (mBC) to an EHR-claims subcohort (requiring ≥ 90 days claims coverage). We analyzed diagnosis coverage, patient time during lookback and follow-up, baseline characteristics, and rates of 14 adverse events (AEs). Analyses were age stratified due to insurance coverage changes at age 65. RESULTS:For the EHR cohort (N = 6289), 1438 (23%) were in the EHR-claims subcohort. A greater proportion were aged ≥ 65 years in the EHR cohort (30%) than in the EHR-claims subcohort (17%). EHR-claims patients had longer observation periods and more unique diagnoses across both age groups. For most AEs, incidences were higher in both age groups in the EHR-claims subcohort than in the EHR cohort. CONCLUSIONS:EHR-claims provided more diagnoses and observation time, at the cost of a reduction in sample size and underrepresentation of patients ≥ 65 years. Differing age proportions support age-stratified or standardized analyses for EHR-claims data. Results aid interpretation of differences between EHR and EHR-claims results due to shifts in age, completeness of diagnosis history, and duration of observation.

BACKGROUND:Variable selection is essential for propensity score (PS)-weighted estimators. Recent work shows that including instrumental variables (IVs), associated with only treatment but not with the outcome, can impact both the bias and precision of the PS-weighted estimators. METHODS:The outcome-adaptive lasso (OAL) is an innovative model-based method adapting the popular adaptive lasso variable selection to causal inference. It attempts to identify IVs, so one can exclude them from the PS model. Unlike the model-based approach, stable balancing weighting (SBW) estimates inverse probability weights directly while minimizing the variance of the weights and covariate imbalance simultaneously. Based on its variance optimization algorithm, SBW may provide some protection against the impact of IVs. Lastly, we considered stable confounder selection (SCS), which assesses the stability of model-based effect estimates. RESULTS:The authors present the results of simulation studies to investigate which method performs the best when moderate or strong IVs are used. The simulation studies consider IVs and spurious variables to generate extreme PSs. In simulations, SBW generally outperformed OAL and SCS in terms of reducing mean squared error, notably when the IVs were strong, and many covariates were highly correlated. Our empirical application to the effect of abciximab treatment demonstrates that SBW is a robust method to effectively handle limited overlap. CONCLUSIONS:Our numerical results support the use of SBW in situations where IVs or near-IVs may lead to practical violations of positivity assumptions.

ABSTRACT Purpose Increases in adult stimulant prescribing pose a potential risk due to the higher prevalence of contraindicated conditions among this population. We sought to identify patient, provider, and visit characteristics predictive of potentially inappropriate adult stimulant prescriptions. Methods We conducted a repeated cross‐sectional study using the National Ambulatory Medical Care Survey, a nationally representative weighted sample of 5 453 702 723 ambulatory care visits from 2012 to 2019. Potentially inappropriate prescriptions were defined as prescriptions to patients with potentially contraindicated conditions, as determined by US Food and Drug Administration stimulant labels. Results Of the 5 453 702 723 visits, stimulant use was prevalent at 121384694 (2.23%) visits and newly prescribed at 18880152 (0.34%) visits. Of these, 4 620 138 (24.47%) new stimulant prescriptions and 28 055 947 (23.11%) prevalent prescriptions were potentially inappropriate. Potentially inappropriate prescribing increased over time and with age. Visits to primary care providers (relative risk [RR] 1.65, 95% CI 1.05–2.59) were predictive of inappropriate prescribing. Non‐Hispanic Black (RR 0.48, 95% CI 0.33–0.70) and Hispanic race/ethnicity (RR 0.46, 95% CI 0.35–0.60), coronary artery disease (RR 0.54, 95% CI 0.33–0.86), pregnancy (RR 0.05, 95% CI 0.03–0.11), hypertension (RR 0.69, 95% CI 0.56–0.84), and glaucoma (RR 0.07, 95% CI 0.02–0.24) were predictive of decreased prevalent stimulant prescriptions; substance abuse was predictive of new stimulant prescribing (RR 2.14, 95% CI 1.07–4.27). Conclusions The proportion of potentially inappropriate adult stimulant prescriptions increased over time and with patient age. Visits to primary care providers were predictive of potentially inappropriate prescribing, and a history of substance abuse was predictive of new stimulant prescriptions; therefore, quality improvement interventions regarding safe stimulant prescribing practices may be warranted.

Funding Yes — Amgen, Inc. Background/Synopsis Evolocumab significantly reduces major adverse cardiovascular events (MACEs) in patients with atherosclerotic cardiovascular disease (ASCVD) in randomized controlled trials (RCTs). However, evidence on its effectiveness in real-world clinical practice remains limited. Objective/Purpose To evaluate the real-world effectiveness of evolocumab in reducing MACE in patients with ASCVD. Methods Patients (≥18 years) who initiated evolocumab between 2017 and 2023 with a history of ASCVD were identified from Komodo Health's Healthcare Map®. The index date was defined as day 75 following the initial prescription of evolocumab. Patients were classified as a treated cohort (initiated and remained on evolocumab) or a non-treated cohort (initiated but discontinued prior to the index date). The primary outcome was composite MACE of myocardial infarction (MI), stroke, and coronary revascularization. Cumulative incidence and 4-year risk ratio between treatment cohorts were calculated using inverse probability of treatment and censoring weights. Results The final analysis included 87,102 treated and 24,609 non-treated patients. Patients in the treated cohort had a 20% (RR = 0.80; 95% CI: 0.73–0.86) lower risk of composite MACE at 4 years compared to the non-treated cohort. Conclusions In this real-world study, evolocumab was effective in reducing MACE outcomes in patients with clinical ASCVD. The findings are consistent with the Further Cardiovascular Outcomes Research with PCSK9 Inhibition in Subjects with Elevated Risk (FOURIER) trial and extend the evidence of evolocumab effectiveness to a larger and more diverse ASCVD cohort with a longer follow-up period in the real-world setting.Figures/Tables: 2

A compilation of factors over the past decade—including the availability of increasingly large and rich healthcare datasets, advanced technologies to extract unstructured information from health records and digital sources, advancement of principled study design and analytic methods to emulate clinical trials, and frameworks to support transparent study conduct—has ushered in a new era of real-world evidence (RWE). This review article describes the evolution of the RWE era, including pharmacoepidemiologic methods designed to support causal inferences regarding treatment effects, the role of regulators and other health authorities in establishing distributed real-world data networks enabling analytics at scale, and the many global guidance documents on principled methods of producing RWE. This article also highlights the growing opportunity for RWE to support decision making by regulators, health technology assessment groups, clinicians, patients, and other stakeholders and provides examples of influential RWE studies. RWE holds promise to address important questions that clinical trials typically do not answer about treatment benefits and risks, and to ultimately impact public health by helping to guide decision making across the healthcare ecosystem.

OBJECTIVE:This systematic review synthesizes existing evidence to quantify racial and ethnic disparities in low back pain (LBP) incidence and prevalence in the United States across stages of chronicity (acute, subacute, and chronic LBP). METHODS:Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, we systematically searched MEDLINE, Embase, the Cumulative Index to Nursing and Allied Health Literature (CINAHL), and Web of Science (through January 7, 2025) for studies reporting LBP incidence or prevalence by race and ethnicity in US adults. Risk of bias was assessed using the Risk of Bias in Non-randomized Studies - of Exposure (ROBINS-E) tool. RESULTS:Of 8,145 citations, 23 studies met inclusion criteria (10 on incidence, 13 on prevalence). Some incidence studies found higher risk of chronic LBP among Black adults compared to White adults, whereas data on Hispanic and Latino adults remain limited. Prevalence studies showed higher rates in White and American Indian and Alaska Native adults, with lower prevalence in Black, Hispanic and Latino, and Asian adults. Military studies consistently reported that Black service members experienced higher LBP incidence compared to other races. No studies examined the subacute state. CONCLUSION:This review highlights persistent race-based differences in LBP, with critical gaps in research on acute LBP incidence and community-based prevalence. Future studies should prioritize population-based research to better capture racial differences in LBP burden and inform targeted interventions.

Journal Article Accepted manuscript RE: "Invited Commentary: Influence of Incomplete Death Information on Cumulative Risk Estimates in United States Claims Data" Get access Julie Barberio, Julie Barberio Department of Epidemiology, Emory University, Atlanta, GA and Center for Observational Research, Amgen Inc., Thousand Oaks, CA Correspondence Address: Julie Barberio, Department of Epidemiology Rollins School of Public Health, Emory University, 1518 Clifton Rd, Atlanta, GA 30322 ([email protected]) Search for other works by this author on: Oxford Academic PubMed Google Scholar Ashley I Naimi, Ashley I Naimi Department of Epidemiology, Emory University, Atlanta, GA Search for other works by this author on: Oxford Academic PubMed Google Scholar Rachel E Patzer, Rachel E Patzer Department of Epidemiology, Emory University, Atlanta, GA and Regenstrief Institute, Indianapolis, IN Search for other works by this author on: Oxford Academic PubMed Google Scholar Christopher Kim, Christopher Kim Center for Observational Research, Amgen Inc., Thousand Oaks, CA Search for other works by this author on: Oxford Academic PubMed Google Scholar Rohini K Hernandez, Rohini K Hernandez Center for Observational Research, Amgen Inc., Thousand Oaks, CA Search for other works by this author on: Oxford Academic PubMed Google Scholar M Alan Brookhart, M Alan Brookhart Target RWE/NoviSci, Inc, Chapel Hill, NC and Department of Population Health Sciences, Duke University, Durham, NC Search for other works by this author on: Oxford Academic PubMed Google Scholar David Gilbertson, David Gilbertson Chronic Disease Research Group, Minneapolis, MN Search for other works by this author on: Oxford Academic PubMed Google Scholar Brian D Bradbury, Brian D Bradbury Center for Observational Research, Amgen Inc., Thousand Oaks, CADepartment of Epidemiology, University of California, Los Angeles, CA, USA Search for other works by this author on: Oxford Academic PubMed Google Scholar Timothy L Lash Timothy L Lash Department of Epidemiology, Emory University, Atlanta, GA Search for other works by this author on: Oxford Academic PubMed Google Scholar American Journal of Epidemiology, kwae229, https://doi.org/10.1093/aje/kwae229 Published: 03 January 2025