Visible minority status and occupation were associated with increased COVID-19 infection in Greater Vancouver British Columbia between June and November 2020: an ecological study.

Background: The COVID-19 pandemic has highlighted health disparities, especially among specific population groups. This study examines the spatial relationship between the proportion of visible minorities (VM), occupation types and COVID-19 infection in the Greater Vancouver region of British Columbia, Canada. Methods: Provincial COVID-19 case data between June 24, 2020, and November 7, 2020, were aggregated by census dissemination area and linked with sociodemographic data from the Canadian 2016 census. Bayesian spatial Poisson regression models were used to examine the association between proportion of visible minorities, occupation types and COVID-19 infection. Models were adjusted for COVID-19 testing rates and other sociodemographic factors. Relative risk (RR) and 95% Credible Intervals (95% CrI) were calculated. Results: We found an inverse relationship between the proportion of the Chinese population and risk of COVID-19 infection (RR = 0.98 95% CrI = 0.96, 0.99), whereas an increased risk was observed for the proportions of the South Asian group (RR = 1.10, 95% CrI = 1.08, 1.12), and Other Visible Minority group (RR = 1.06, 95% CrI = 1.04, 1.08). Similarly, a higher proportion of frontline workers (RR = 1.05, 95% CrI = 1.04, 1.07) was associated with higher infection risk compared to non-frontline. Conclusion: Despite adjustments for testing, housing, occupation, and other social economic status variables, there is still a substantial association between the proportion of visible minorities, occupation types, and the risk of acquiring COVID-19 infection in British Columbia. This ecological analysis highlights the existing disparities in the burden of diseases among different visible minority populations and occupation types.

Cohort profile: the British Columbia COVID-19 Cohort (BCC19C)-a dynamic, linked population-based cohort.

Purpose: The British Columbia COVID-19 Cohort (BCC19C) was developed from an innovative, dynamic surveillance platform and is accessed/analyzed through a cloud-based environment. The platform integrates recently developed provincial COVID-19 datasets (refreshed daily) with existing administrative holdings and provincial registries (refreshed weekly/monthly). The platform/cohort were established to inform the COVID-19 response in near "real-time" and to answer more in-depth epidemiologic questions. Participants: The surveillance platform facilitates the creation of large, up-to-date analytic cohorts of people accessing COVID-19 related services and their linked medical histories. The program of work focused on creating/analyzing these cohorts is referred to as the BCC19C. The administrative/registry datasets integrated within the platform are not specific to COVID-19 and allow for selection of "control" individuals who have not accessed COVID-19 services. Findings to date: The platform has vastly broadened the range of COVID-19 analyses possible, and outputs from BCC19C analyses have been used to create dashboards, support routine reporting and contribute to the peer-reviewed literature. Published manuscripts (total of 15 as of July, 2023) have appeared in high-profile publications, generated significant media attention and informed policy and programming. In this paper, we conducted an analysis to identify sociodemographic and health characteristics associated with receiving SARS-CoV-2 laboratory testing, testing positive, and being fully vaccinated. Other published analyses have compared the relative clinical severity of different variants of concern; quantified the high "real-world" effectiveness of vaccines in addition to the higher risk of myocarditis among younger males following a 2nd dose of an mRNA vaccine; developed and validated an algorithm for identifying long-COVID patients in administrative data; identified a higher rate of diabetes and healthcare utilization among people with long-COVID; and measured the impact of the pandemic on mental health, among other analyses. Future plans: While the global COVID-19 health emergency has ended, our program of work remains robust. We plan to integrate additional datasets into the surveillance platform to further improve and expand covariate measurement and scope of analyses. Our analyses continue to focus on retrospective studies of various aspects of the COVID-19 pandemic, as well as prospective assessment of post-acute COVID-19 conditions and other impacts of the pandemic.

Clinical Severity of Severe Acute Respiratory Syndrome Coronavirus 2 Omicron Variant Relative to Delta in British Columbia, Canada: A Retrospective Analysis of Whole-Genome Sequenced Cases.

BACKGROUND: In late 2021, the Omicron severe acute respiratory syndrome coronavirus 2 variant emerged and rapidly replaced Delta as the dominant variant. The increased transmissibility of Omicron led to surges in case rates and hospitalizations; however, the true severity of the variant remained unclear. We aimed to provide robust estimates of Omicron severity relative to Delta. METHODS: This retrospective cohort study was conducted with data from the British Columbia COVID-19 Cohort, a large provincial surveillance platform with linkage to administrative datasets. To capture the time of cocirculation with Omicron and Delta, December 2021 was chosen as the study period. Whole-genome sequencing was used to determine Omicron and Delta variants. To assess the severity (hospitalization, intensive care unit [ICU] admission, length of stay), we conducted adjusted Cox proportional hazard models, weighted by inverse probability of treatment weights (IPTW). RESULTS: The cohort was composed of 13 128 individuals (7729 Omicron and 5399 Delta). There were 419 coronavirus disease 2019 hospitalizations, with 118 (22%) among people diagnosed with Omicron (crude rate = 1.5% Omicron, 5.6% Delta). In multivariable IPTW analysis, Omicron was associated with a 50% lower risk of hospitalization compared with Delta (adjusted hazard ratio [aHR] = 0.50, 95% confidence interval [CI] = 0.43 to 0.59), a 73% lower risk of ICU admission (aHR = 0.27, 95% CI = 0.19 to 0.38), and a 5-day shorter hospital stay (aß = -5.03, 95% CI = -8.01 to -2.05). CONCLUSIONS: Our analysis supports findings from other studies that have demonstrated lower risk of severe outcomes in Omicron-infected individuals relative to Delta.

Association Between Prescription Opioid Therapy for Noncancer Pain and Hepatitis C Virus Seroconversion.

Importance: Initiation of injection drug use may be more frequent among people dispensed prescription opioid therapy for noncancer pain, potentially increasing the risk of hepatitis C virus (HCV) acquisition. Objective: To assess the association between medically dispensed long-term prescription opioid therapy for noncancer pain and HCV seroconversion among individuals who were initially injection drug use-naive. Design, Setting, and Participants: A population-based, retrospective cohort study of individuals tested for HCV in British Columbia, Canada, with linkage to outpatient pharmacy dispensations, was conducted. Individuals with an initial HCV-negative test result followed by 1 additional test between January 1, 2000, and December 31, 2017, and who had no history of substance use at baseline (first HCV-negative test), were included. Participants were followed up from baseline to the last HCV-negative test or estimated date of seroconversion (midpoint between HCV-positive and the preceding HCV-negative test). Exposures: Episodes of prescription opioid use for noncancer pain were defined as acute (<90 days) or long-term (≥90 days). Prescription opioid exposure status (long-term vs prescription opioid-naive/acute) was treated as time-varying in survival analyses. In secondary analyses, long-term exposure was stratified by intensity of use (chronic vs. episodic) and by average daily dose in morphine equivalents (MEQ). Main Outcomes and Measures: Multivariable Cox regression models were used to assess the association between time-varying prescription opioid status and HCV seroconversion. Results: A total of 382 478 individuals who had more than 1 HCV test were included, of whom more than half were female (224 373 [58.7%]), born before 1974 (201 944 [52.8%]), and younger than 35 years at baseline (196 298 [53.9%]). Participants were followed up for 2 057 668 person-years and 1947 HCV seroconversions occurred. Of the participants, 41 755 people (10.9%) were exposed to long-term prescription opioid therapy at baseline or during follow-up. The HCV seroconversion rate per 1000 person-years was 0.8 among the individuals who were prescription opioid-naive/acute (1489 of 1947 [76.5%] seroconversions; 0.4% seroconverted within 5 years) and 2.1 with long-term prescription opioid therapy (458 of 1947 [23.5%] seroconversions; 1.1% seroconverted within 5 years). In multivariable analysis, exposure to long-term prescription opioid therapy was associated with a 3.2-fold (95% CI, 2.9-3.6) higher risk of HCV seroconversion (vs prescription opioid-naive/acute). In separate Cox models, long-term chronic use was associated with a 4.7-fold higher risk of HCV seroconversion (vs naive/acute use 95% CI, 3.9-5.8), and long-term higher-dose use (≥90 MEQ) was associated with a 5.1-fold higher risk (vs naive/acute use 95% CI, 3.7-7.1). Conclusions and Relevance: In this cohort study of people with more than 1 HCV test, long-term prescription opioid therapy for noncancer pain was associated with a higher risk of HCV seroconversion among individuals who were injection drug use-naive at baseline or at prescription opioid initiation. These results suggest injection drug use initiation risk is higher among people dispensed long-term therapy and may be useful for informing approaches to identify and prevent HCV infection. These findings should not be used to justify abrupt discontinuation of long-term therapy, which could increase risk of harms.

Prescription opioid treatment for non-cancer pain and initiation of injection drug use: large retrospective cohort study.

OBJECTIVE: To assess the association between long term prescription opioid treatment medically dispensed for non-cancer pain and the initiation of injection drug use (IDU) among individuals without a history of substance use. DESIGN: Retrospective cohort study. SETTING: Large administrative data source (containing information for about 1.7 million individuals tested for hepatitis C virus or HIV in British Columbia, Canada) with linkage to administrative health databases, including dispensations from community pharmacies. PARTICIPANTS: Individuals age 11-65 years and without a history of substance use (except alcohol) at baseline. MAIN OUTCOME MEASURES: Episodes of prescription opioid use for non-cancer pain were identified based on drugs dispensed between 2000 and 2015. Episodes were classified by the increasing length and intensity of opioid use (acute (lasting <90 episode days), episodic (lasting ≥90 episode days; with <90 days' drug supply and/or <50% episode intensity), and chronic (lasting ≥90 episode days; with ≥90 days' drug supply and ≥50% episode intensity)). People with a chronic episode were matched 1:1:1:1 on socioeconomic variables to those with episodic or acute episodes and to those who were opioid naive. IDU initiation was identified by a validated administrative algorithm with high specificity. Cox models weighted by inverse probability of treatment weights assessed the association between opioid use category (chronic, episodic, acute, opioid naive) and IDU initiation. RESULTS: 59 804 participants (14 951 people from each opioid use category) were included in the matched cohort, and followed for a median of 5.8 years. 1149 participants initiated IDU. Cumulative probability of IDU initiation at five years was highest for participants with chronic opioid use (4.0%), followed by those with episodic use (1.3%) and acute use (0.7%), and those who were opioid naive (0.4%). In the inverse probability of treatment weighted Cox model, risk of IDU initiation was 8.4 times higher for those with chronic opioid use versus those who were opioid naive (95% confidence interval 6.4 to 10.9). In a sensitivity analysis limited to individuals with a history of chronic pain, cumulative risk for those with chronic use (3.4% within five years) was lower than the primary results, but the relative risk was not (hazard ratio 9.7 (95% confidence interval 6.5 to 14.5)). IDU initiation was more frequent at higher opioid doses and younger ages. CONCLUSIONS: The rate of IDU initiation among individuals who received chronic prescription opioid treatment for non-cancer pain was infrequent overall (3-4% within five years) but about eight times higher than among opioid naive individuals. These findings could have implications for strategies to prevent IDU initiation, but should not be used as a reason to support involuntary tapering or discontinuation of long term prescription opioid treatment.

Impact of direct-acting antivirals for HCV on mortality in a large population-based cohort study.

BACKGROUND & AIMS: We evaluated the effect of direct-acting antiviral (DAA)-induced sustained virologic response (SVR) on all-cause, liver- and drug-related mortality in a population-based cohort in British Columbia, Canada. METHODS: We used data from the British Columbia Hepatitis Testers Cohort, which includes people tested for HCV since 1990, linked with data on medical visits, hospitalizations, prescription drugs and mortality. We followed people who received DAAs and people who did not receive any HCV treatment to death or December 31, 2019. We used inverse probability of treatment weighting to balance the baseline profile of treated and untreated individuals and performed multivariable proportional hazard modelling to assess the effect of DAAs on mortality. RESULTS: Our cohort comprised 10,851 people treated with DAAs (SVR 10,426 [96%], no-SVR: 425) and 10,851 matched untreated individuals. Median follow-up time was 2.2 years (IQR 1.3-3.6; maximum 6.2). The all-cause mortality rate was 19.5/1,000 person-years (PY) among the SVR group (deaths = 552), 86.5/1,000 PY among the no-SVR group (deaths = 96), and 99.2/1,000 PY among the untreated group (deaths = 2,133). In the multivariable model, SVR was associated with significant reduction in all-cause (adjusted hazard ratio [aHR] 0.19; 95% CI 0.17-0.21), liver- (adjusted subdistribution HR [asHR] 0.22, 95% CI 0.18-0.27) and drug-related mortality (asHR 0.26, 95% CI 0.21-0.32) compared to no-treatment. Older age and cirrhosis were associated with higher risk of liver-related mortality while younger age, injection drug use (IDU), problematic alcohol use and HIV/HBV co-infections were associated with a higher risk of drug-related mortality. CONCLUSIONS: DAA treatment is associated with a substantial reduction in all-cause, liver- and drug-related mortality. The association of IDU and related syndemic factors with a higher risk of drug-related mortality calls for an integrated social support, addiction, and HCV care approach among people who inject drugs. LAY SUMMARY: We assessed the effect of treatment of hepatitis C virus infection with direct-acting antiviral drugs on deaths from all causes, liver disease and drug use. We found that treatment with direct-acting antiviral drugs is associated with substantial lowering in risk of death from all causes, liver disease and drug use among people with hepatitis C virus infection.

Impact of HCV infection and ethnicity on incident type 2 diabetes: findings from a large population-based cohort in British Columbia.

INTRODUCTION: Increasing evidence indicates that chronic hepatitis C virus (HCV) infection is associated with higher risk of diabetes. Previous studies showed ethnic disparities in the disease burden of diabetes, with increased risk in Asian population. We described the incidence of type 2 diabetes related to HCV infection and assessed the concurrent impact of HCV infection and ethnicity on the risk of diabetes. RESEARCH DESIGN AND METHODS: In British Columbia Hepatitis Testers Cohort, individuals were followed from HCV diagnosis to the earliest of (1) incident type 2 diabetes, (2) death or (3) end of the study (December 31, 2015). Study population included 847 021 people. Diabetes incidence rates in people with and without HCV were computed. Propensity scores (PS) analysis was used to assess the impact of HCV infection on newly acquired diabetes. PS-matched dataset included 117 184 people. We used Fine and Gray multivariable subdistributional hazards models to assess the effect of HCV and ethnicity on diabetes while adjusting for confounders and competing risks. RESULTS: Diabetes incidence rates were higher among people with HCV infection than those without. The highest diabetes incidence rate was in South Asians with HCV (14.7/1000 person-years, 95% CI 12.87 to 16.78). Compared with Others, South Asians with and without HCV and East Asians with HCV had a greater risk of diabetes. In the multivariable stratified analysis, HCV infection was associated with increased diabetes risk in all subgroups: East Asians, adjusted HR (aHR) 3.07 (95% CI 2.43 to 3.88); South Asians, aHR 2.62 (95% CI 2.10 to 3.26); and Others, aHR 2.28 (95% CI 2.15 to 2.42). CONCLUSIONS: In a large population-based linked administrative health data, HCV infection was associated with higher diabetes risk, with a greater relative impact in East Asians. South Asians had the highest risk of diabetes. These findings highlight the need for care and screening for HCV-related chronic diseases such as type 2 diabetes among people affected by HCV.

Cohort profile: development and characteristics of a retrospective cohort of individuals dispensed prescription opioids for non-cancer pain in British Columbia, Canada.

PURPOSE: Prescription opioids (POs) are widely prescribed for chronic non-cancer pain but are associated with several risks and limited long-term benefit. Large, linked data sources are needed to monitor their harmful effects. We developed and characterised a retrospective cohort of people dispensed POs. PARTICIPANTS: We used a large linked administrative database to create the Opioid Prescribing Evaluation and Research Activities cohort of individuals dispensed POs for non-cancer pain in British Columbia (BC), Canada (1996-2015). We created definitions to categorise episodes of PO use based on a review of the literature (acute, episodic, chronic), developed an algorithm for inferring clinical indication and assessed patterns of PO use across a range of characteristics. FINDINGS TO DATE: The current cohort includes 1.1 million individuals and 3.4 million PO episodes (estimated to capture 40%-50% of PO use in BC). The majority of episodes were acute (81%), with most prescribed for dental or surgical pain. Chronic use made up 3% of episodes but 88% of morphine equivalents (MEQ). Across the acute to episodic to chronic episode gradient, there was an increasing prevalence of higher potency POs (hydromorphone, oxycodone, fentanyl, morphine), long-acting formulations and chronic pain related indications (eg, back, neck, joint pain). Average daily dose (MEQ) was similar for acute/episodic but higher for chronic episodes. Approximately 7% of the cohort had a chronic episode and chronic pain was the characteristic most strongly associated with chronic PO use. Individuals initiating a chronic episode were also more likely to have higher social/material deprivation and previous experience with a mental health condition or a problem related to alcohol or opioid use. Overall, these findings suggest our episode definitions have face validity and also provide insight into characteristics of people initiating chronic PO therapy. FUTURE PLANS: The cohort will be refreshed every 2 years. Future analyses will explore the association between POs and adverse outcomes.

Syndemic profiles of people living with hepatitis C virus using population-level latent class analysis to optimize health services.

BACKGROUND: Hepatitis C (HCV) affects diverse populations such as people who inject drugs (PWID), 'baby boomers,' gay/bisexual men who have sex with men (gbMSM), and people from HCV endemic regions. Assessing HCV syndemics (i.e.relationships with mental health/chronic diseases) among subpopulations using Latent Class Analysis (LCA) may facilitate targeted program planning. METHODS: The BC Hepatitis Testers Cohort(BC-HTC) includes all HCV cases identified in BC between 1990 and 2015, integrated with medical administrative data. LCA grouped all BC-HTC HCV diagnosed people(n = 73,665) by socio-demographic/clinical indicators previously determined to be relevant for HCV outcomes. The final model was chosen based on fit statistics, epidemiological meaningfulness, and posterior probability. Classes were named by most defining characteristics. RESULTS: The six-class model was the best fit and had the following names and characteristics: 'Younger PWID'(n =11,563): recent IDU (67%), people born >1974 (48%), mental illness (62%), material deprivation (59%). 'Older PWID'(n =15,266): past IDU (78%), HIV (17%), HBV (17%) coinfections, alcohol misuse(68%). 'Other Middle-Aged People'(n = 9019): gbMSM (26%), material privilege (31%), people born between 1965-1974 (47%). 'People of Asian backgrounds' (n = 4718): East/South Asians (92%), no alcohol misuse (97%) or mental illness (93%), people born <1945 (26%), social privilege (66%). 'Rural baby boomers' (n = 20,401): rural dwellers (32%), baby boomers (79%), heterosexuals (99%), no HIV (100%). 'Urban socially deprived baby boomers' (n = 12,698): urban dwellers (99%), no IDU (100%), liver disease (22%), social deprivation (94%). CONCLUSIONS: Differences between classes suggest variability in patients' service needs. Further analysis of health service utilization patterns may inform optimal service layout.

A Computable Phenotype Model for Classification of Men Who Have Sex With Men Within a Large Linked Database of Laboratory, Surveillance, and Administrative Healthcare Records.

Background: Most public health datasets do not include sexual orientation measures, thereby limiting the availability of data to monitor health disparities, and evaluate tailored interventions. We therefore developed, validated, and applied a novel computable phenotype model to classify men who have sex with men (MSM) using multiple health datasets from British Columbia, Canada, 1990-2015. Methods: Three case surveillance databases, a public health laboratory database, and five administrative health databases were linked and deidentified (BC Hepatitis Testers Cohort), resulting in a retrospective cohort of 727,091 adult men. Known MSM status from the three disease case surveillance databases was used to develop a multivariable model for classifying MSM in the full cohort. Models were selected using "elastic-net" (GLMNet package) in R, and a final model optimized area under the receiver operating characteristics curve. We compared characteristics of known MSM, classified MSM, and classified heterosexual men. Findings: History of gonorrhea and syphilis diagnoses, HIV tests in the past year, history of visit to an identified gay and bisexual men's clinic, and residence in MSM-dense neighborhoods were all positively associated with being MSM. The selected model had sensitivity of 72%, specificity of 94%. Excluding those with known MSM status, a total of 85,521 men (12% of cohort) were classified as MSM. Interpretation: Computable phenotyping is a promising approach for classification of sexual minorities and investigation of health outcomes in the absence of routinely available self-report data.

Evaluating the use of bootstrapping in cohort studies conducted with 1:1 propensity score matching-A plasmode simulation study.

PURPOSE: Bootstrapping can account for uncertainty in propensity score (PS) estimation and matching processes in 1:1 PS-matched cohort studies. While theory suggests that the classical bootstrap can fail to produce proper coverage, practical impact of this theoretical limitation in settings typical to pharmacoepidemiology is not well studied. METHODS: In a plasmode-based simulation study, we compared performance of the standard parametric approach, which ignores uncertainty in PS estimation and matching, with two bootstrapping methods. The first method only accounted for uncertainty introduced during the matching process (the observation resampling approach). The second method accounted for uncertainty introduced during both PS estimation and matching processes (the PS reestimation approach). Variance was estimated based on percentile and empirical standard errors, and treatment effect estimation was based on median and mean of the estimated treatment effects across 1000 bootstrap resamples. Two treatment prevalence scenarios (5% and 29%) across two treatment effect scenarios (hazard ratio of 1.0 and 2.0) were evaluated in 500 simulated cohorts of 10 000 patients each. RESULTS: We observed that 95% confidence intervals from the bootstrapping approaches but not the standard approach, resulted in inaccurate coverage rates (98%-100% for the observation resampling approach, 99%-100% for the PS reestimation approach, and 95%-96% for standard approach). Treatment effect estimation based on bootstrapping approaches resulted in lower bias than the standard approach (less than 1.4% vs 4.1%) at 5% treatment prevalence; however, the performance was equivalent at 29% treatment prevalence. CONCLUSION: Use of bootstrapping led to variance overestimation and inconsistent coverage, while coverage remained more consistent with parametric estimation.

Estimation of average treatment effects among multiple treatment groups by using an ensemble approach.

In observational studies, generalized propensity score (GPS)-based statistical methods, such as inverse probability weighting (IPW) and doubly robust (DR) method, have been proposed to estimate the average treatment effect (ATE) among multiple treatment groups. In this article, we investigate the GPS-based statistical methods to estimate treatment effects from two aspects. The first aspect of our investigation is to obtain an optimal GPS estimation method among four competing GPS estimation methods by using a rank aggregation approach. We further examine whether the optimal GPS-based IPW and DR methods would improve the performance for estimating ATE. It is well known that the DR method is consistent if either the GPS or the outcome models are correctly specified. The second aspect of our investigation is to examine whether the DR method could be improved if we ensemble outcome models. To that end, bootstrap method and rank aggregation method are used to obtain the ensemble optimal outcome model from several competing outcome models, and the resulting outcome model is incorporated into the DR method, resulting in an ensemble DR (enDR) method. Extensive simulation results indicate that the enDR method provides the best performance in estimating the ATE regardless of the method used for estimating GPS. We illustrate our methods using the MarketScan healthcare insurance claims database to examine the treatment effects among three different bones and substitutes used for spinal fusion surgeries. We draw conclusions based on the estimates from the enDR method coupled with the optimal GPS estimation method.

Days' Supply of Initial Opioid Analgesic Prescriptions and Additional Fills for Acute Pain Conditions Treated in the Primary Care Setting - United States, 2014.

During 2017, opioids were associated with 47,600 deaths in the United States, approximately one third of which involved a prescription opioid (1). Amid concerns that overprescribing to patients with acute pain remains an essential factor underlying misuse, abuse, diversion, and unintentional overdose, several states have restricted opioid analgesic prescribing (2,3). To characterize patterns of opioid analgesic use for acute pain in primary care settings before the widespread implementation of limits on opioid prescribing (2,3), patients filling an opioid analgesic prescription for acute pain were identified from a 2014 database of commercial claims. Using a logistic generalized additive model, the probability of obtaining a refill was estimated as a function of the initial number of days supplied. Among 176,607 patients with a primary care visit associated with an acute pain complaint, 7.6% filled an opioid analgesic prescription. Among patients who received an initial 7-day supply, the probability of obtaining an opioid analgesic prescription refill for nine of 10 conditions was <25%. These results suggest that a ≤7-day opioid analgesic prescription might be sufficient for most, but not all, patients seen in primary care settings with acute pain who appear to need opioid analgesics. However, treatment strategies should account for patient and condition characteristics, which might alternatively reduce or extend the anticipated duration of benefit from opioid analgesic therapy.

Association of Osteoporosis Medication Use After Hip Fracture With Prevention of Subsequent Nonvertebral Fractures: An Instrumental Variable Analysis.

Importance: Osteoporosis medication treatment is recommended after hip fracture, yet contemporary estimates of rates of initiation and clinical benefit in the patient population receiving routine care are not well documented. Objectives: To report osteoporosis treatment initiation rates between January 1, 2004, and September 30, 2015, and to estimate the risk reduction in subsequent nonvertebral fractures associated with treatment initiation in patients with hip fracture. Design, Setting, and Participants: In this cohort study, data from a commercial insurance claims database from the United States were analyzed. Patients 50 years and older who had a hip fracture and were not receiving treatment with osteoporosis medications before their fracture were included. Exposure: Prescription dispensing of an osteoporosis medication within 180 days of a hip fracture hospitalization. Main Outcomes and Measures: Each initiation episode was matched with 10 nonuse episodes on person-time after the index hip fracture event to preclude immortal time bias and followed up for the outcome of nonvertebral fracture until change in exposure or a censoring event. An instrumental variable analysis using 2-stage residual inclusion method was conducted using calendar year, specialist access, geographical variation in prescribing patterns, and hospital preference. Results: Among 97 169 patients with a hip fracture identified, the mean (SD) age was 80.2 (10.8) years, and 64 164 (66.0%) were women. A continuous decline over the study years was observed in osteoporosis medication initiation rates from 9.8% (95% CI, 9.0%-10.6%) in 2004 to 3.3% (95% CI, 2.9%-3.8%) in 2015. In the effectiveness analyses, the hospital preference instrumental variable had a stronger association with treatment (pseudo R2 = 0.20) than the other 3 instrumental variables (specialist access: pseudo R2 = 0.04; calendar year: pseudo R2 = 0.05; and geographic variation: pseudo R2 = 0.07). Instrumental variable analysis with hospital preference suggested a rate difference of 4.2 events (95% CI, 1.1-7.3) per 100 person-years in subsequent fractures associated with osteoporosis treatment initiation compared with nonuse in an additive hazard model. Conclusions and Relevance: Low rates of osteoporosis treatment initiation after a hip fracture in recent years were observed. Clinically meaningful reduction in subsequent nonvertebral fracture rates associated with treatment suggests that improving prescriber adherence to guidelines and patient adherence to prescribed regimens may result in notable public health benefit.

Propensity scores based methods for estimating average treatment effect and average treatment effect among treated: A comparative study.

Propensity score based statistical methods, such as matching, regression, stratification, inverse probability weighting (IPW), and doubly robust (DR) estimating equations, have become popular in estimating average treatment effect (ATE) and average treatment effect among treated (ATT) in observational studies. Propensity score is the conditional probability receiving a treatment assignment with given covariates, and propensity score is usually estimated by logistic regression. However, a misspecification of the propensity score model may result in biased estimates for ATT and ATE. As an alternative, the generalized boosting method (GBM) has been proposed to estimate the propensity score. GBM uses regression trees as weak predictors and captures nonlinear and interactive effects of the covariate. For GBM-based propensity score, only IPW methods have been investigated in the literature. In this article, we provide a comparative study of the commonly used propensity score based methods for estimating ATT and ATE, and examine their performances when propensity score is estimated by logistic regression and GBM, respectively. Extensive simulation results indicate that the estimators for ATE and ATT may vary greatly due to different methods. We concluded that (i) regression may not be suitable for estimating ATE and ATT regardless of the estimation method of propensity score; (ii) IPW and stratification usually provide reliable estimates of ATT when propensity score model is correctly specified; (iii) the estimators of ATE based on stratification, IPW, and DR are close to the underlying true value of ATE when propensity score is correctly specified by logistic regression or estimated using GBM.