Gabapentinoids and Risk for Severe Exacerbation in Chronic Obstructive Pulmonary Disease : A Population-Based Cohort Study.

BACKGROUND: North American and European health agencies recently warned of severe breathing problems associated with gabapentinoids, including in patients with chronic obstructive pulmonary disease (COPD), although supporting evidence is limited. OBJECTIVE: To assess whether gabapentinoid use is associated with severe exacerbation in patients with COPD. DESIGN: Time-conditional propensity score-matched, new-user cohort study. SETTING: Health insurance databases from the Régie de l'assurance maladie du Québec in Canada. PATIENTS: Within a base cohort of patients with COPD between 1994 and 2015, patients initiating gabapentinoid therapy with an indication (epilepsy, neuropathic pain, or other chronic pain) were matched 1:1 with nonusers on COPD duration, indication for gabapentinoids, age, sex, calendar year, and time-conditional propensity score. MEASUREMENTS: The primary outcome was severe COPD exacerbation requiring hospitalization. Hazard ratios (HRs) associated with gabapentinoid use were estimated in subcohorts according to gabapentinoid indication and in the overall cohort. RESULTS: The cohort included 356 gabapentinoid users with epilepsy, 9411 with neuropathic pain, and 3737 with other chronic pain, matched 1:1 to nonusers. Compared with nonuse, gabapentinoid use was associated with increased risk for severe COPD exacerbation across the indications of epilepsy (HR, 1.58 [95% CI, 1.08 to 2.30]), neuropathic pain (HR, 1.35 [CI, 1.24 to 1.48]), and other chronic pain (HR, 1.49 [CI, 1.27 to 1.73]) and overall (HR, 1.39 [CI, 1.29 to 1.50]). LIMITATION: Residual confounding, including from lack of smoking information. CONCLUSION: In patients with COPD, gabapentinoid use was associated with increased risk for severe exacerbation. This study supports the warnings from regulatory agencies and highlights the importance of considering this potential risk when prescribing gabapentin and pregabalin to patients with COPD. PRIMARY FUNDING SOURCE: Canadian Institutes of Health Research and Canadian Lung Association.

Variable selection in high dimensions for discrete-outcome individualized treatment rules: Reducing severity of depression symptoms.

Despite growing interest in estimating individualized treatment rules, little attention has been given the binary outcome setting. Estimation is challenging with nonlinear link functions, especially when variable selection is needed. We use a new computational approach to solve a recently proposed doubly robust regularized estimating equation to accomplish this difficult task in a case study of depression treatment. We demonstrate an application of this new approach in combination with a weighted and penalized estimating equation to this challenging binary outcome setting. We demonstrate the double robustness of the method and its effectiveness for variable selection. The work is motivated by and applied to an analysis of treatment for unipolar depression using a population of patients treated at Kaiser Permanente Washington.

Intraoperative phlebotomies and bleeding in liver transplantation: a historical cohort study and causal analysis.

BACKGROUND: Liver transplantation is associated with major bleeding and red blood cell (RBC) transfusions. No well-designed causal analysis on interventions used to reduce transfusions, such as an intraoperative phlebotomy, has been conducted in this population. METHODS: We conducted a historical cohort study among liver transplantations performed from July 2008 to January 2021 in a Canadian centre. The exposure was intraoperative phlebotomy. The outcomes were blood loss, perioperative RBC transfusions (intraoperative and up to 48 hr after surgery), intraoperative RBC transfusions, and one-year survival. We estimated marginal multiplicative factors (MFs), risk differences (RDs), and hazard ratios by inverse probability of treatment weighting both among treated patients and the whole population. Estimates are reported with 95% confidence intervals (CIs). RESULTS: We included 679 patients undergoing liver transplantations of which 365 (54%) received an intraoperative phlebotomy. A phlebotomy did not reduce bleeding, transfusion risks, or mortality when estimated among the treated but reduced bleeding and transfusion risks when estimated among the whole population (MF, 0.85; 95% CI, 0.72 to 0.99; perioperative RD, -15.2%; 95% CI, -26.1 to -0.8; intraoperative RD, -14.7%; 95% CI, -23.2 to -2.8). In a subgroup analysis on 584 patients with end-stage liver disease, slightly larger effects were observed on both transfusion risks when estimated among the whole population while beneficial effects were observed on the intraoperative transfusion risk when estimated among the treated population. CONCLUSION: The use of intraoperative phlebotomy was not consistently associated with better outcomes in all targets of inference but may improve outcomes among the whole population. STUDY REGISTRATION: www. CLINICALTRIALS: gov (NCT04826666); registered 1 April 2021.

RéSUMé: CONTEXTE: La transplantation hépatique est associée à des saignements importants et à de multiples transfusions de globules rouges (GR). Aucune analyse causale bien conçue sur l'effet d'interventions servant à réduire les transfusions, comme une phlébotomie peropératoire, n'a été menée dans cette population. MéTHODE: Nous avons mené une étude de cohorte historique incluant toutes les transplantations hépatiques réalisées dans un centre canadien de juillet 2008 à janvier 2021. L'exposition d'intérêt était une phlébotomie peropératoire. Les critères d'évaluation étaient le saignement peropératoire, les transfusions de GR périopératoires (peropératoires et jusqu'à 48 heures après la chirurgie), les transfusions de globules rouges peropératoires et la survie à un an. Des facteurs multiplicatifs (FM), des différences de risque (DR) et des rapports de risques instantanés marginaux ont été estimés en utilisant une pondération par l'inverse de la probabilité de traitement parmi les patients traités et parmi l'ensemble de la population. Les effets estimés ont été rapportés avec des intervalles de confiance (IC) à 95 %. RéSULTATS: Nous avons inclus 679 transplantations hépatiques dont 365 (54 %) ont bénéficié d'une phlébotomie peropératoire. La phlébotomie n'a pas réduit les saignements, le risque de transfusion ou la mortalité lorsque ses effets ont été estimés parmi les patients traités, mais a réduit les risques de saignement et de transfusion lorsque ses effets ont été estimés parmi l'ensemble de la population (FM = 0,85 (IC 95 %, 0,72 à 0,99); DR périopératoire = −15,2 % (IC 95 %, −26,1 % à −0,8 %); DR peropératoire = −14,7 % (IC 95 %, −23,2 % à −2,8 %)). Dans une analyse de sous-groupe portant sur 584 patients atteints d'une hépatopathie terminale, des effets légèrement plus importants ont été observés sur les deux risques transfusionnels lorsqu'estimés dans l'ensemble de la population, tandis que des effets bénéfiques ont été observés sur le risque transfusionnel peropératoire lorsqu'estimés parmi les patients traités. CONCLUSION: L'utilisation de la phlébotomie peropératoire n'a pas été systématiquement associée à de meilleurs résultats dans toutes les populations cibles, mais semble améliorer les résultats lorsque les effets sont estimés dans l'ensemble de la population. ENREGISTREMENT DE L'éTUDE: www.ClinicalTrials.gov (NCT04826666); enregistrée le 1er avril 2021.

Privacy-preserving estimation of an optimal individualized treatment rule: a case study in maximizing time to severe depression-related outcomes.

Estimating individualized treatment rules-particularly in the context of right-censored outcomes-is challenging because the treatment effect heterogeneity of interest is often small, thus difficult to detect. While this motivates the use of very large datasets such as those from multiple health systems or centres, data privacy may be of concern with participating data centres reluctant to share individual-level data. In this case study on the treatment of depression, we demonstrate an application of distributed regression for privacy protection used in combination with dynamic weighted survival modelling (DWSurv) to estimate an optimal individualized treatment rule whilst obscuring individual-level data. In simulations, we demonstrate the flexibility of this approach to address local treatment practices that may affect confounding, and show that DWSurv retains its double robustness even when performed through a (weighted) distributed regression approach. The work is motivated by, and illustrated with, an analysis of treatment for unipolar depression using the United Kingdom's Clinical Practice Research Datalink.

Can the Risk of Severe Depression-Related Outcomes Be Reduced by Tailoring the Antidepressant Therapy to Patient Characteristics?

Selective serotonin reuptake inhibitors (SSRIs) are the first-line treatment for patients with unipolar depression, yet there is little guidance on which SSRI provides the most benefit to a patient, based on personal characteristics. In this work, we explore whether an individualized treatment strategy can be used by health-care providers to adapt their prescription pattern to reduce the risk of a severe depression-related outcome (SDO) when choosing between citalopram and fluoxetine, 2 commonly prescribed SSRIs. Our population-based cohort study used data from the Clinical Practice Research Datalink, the Hospital Episode Statistics repository, and the Office for National Statistics database in the United Kingdom to create a cohort of individuals diagnosed with depression who were prescribed citalopram or fluoxetine between April 1998 and December 2017. Patients were followed from treatment initiation until occurrence of the SDO outcome, treatment discontinuation, or end of study. To find an optimal treatment strategy, we used dynamic weighted survival modeling, considering patient features such as age, sex, body mass index, previous psychiatric diagnoses, and medications. Our findings suggest that using patient characteristics to tailor the antidepressant drug therapy is associated with an increase of 4 days in the median time to SDO (95% confidence interval: 2, 10 days).

Weighted regression analysis to correct for informative monitoring times and confounders in longitudinal studies.

We address estimation of the marginal effect of a time-varying binary treatment on a continuous longitudinal outcome in the context of observational studies using electronic health records, when the relationship of interest is confounded, mediated, and further distorted by an informative visit process. We allow the longitudinal outcome to be recorded only sporadically and assume that its monitoring timing is informed by patients' characteristics. We propose two novel estimators based on linear models for the mean outcome that incorporate an adjustment for confounding and informative monitoring process through generalized inverse probability of treatment weights and a proportional intensity model, respectively. We allow for a flexible modeling of the intercept function as a function of time. Our estimators have closed-form solutions, and their asymptotic distributions can be derived. Extensive simulation studies show that both estimators outperform standard methods such as the ordinary least squares estimator or estimators that only account for informative monitoring or confounders. We illustrate our methods using data from the Add Health study, assessing the effect of depressive mood on weight in adolescents.

Estimating the marginal effect of a continuous exposure on an ordinal outcome using data subject to covariate-driven treatment and visit processes.

In the statistical literature, a number of methods have been proposed to ensure valid inference about marginal effects of variables on a longitudinal outcome in settings with irregular monitoring times. However, the potential biases due to covariate-driven monitoring times and confounding have rarely been considered simultaneously, and never in a setting with an ordinal outcome and a continuous exposure. In this work, we propose and demonstrate a methodology for causal inference in such a setting, relying on a proportional odds model to study the effect of the exposure on the outcome. Irregular observation times are considered via a proportional rate model, and a generalization of inverse probability of treatment weights is used to account for the continuous exposure. We motivate our methodology by the estimation of the marginal (causal) effect of the time spent on video or computer games on suicide attempts in the Add Health study, a longitudinal study in the United States. Although in the Add Health data, observation times are prespecified, our proposed approach is applicable even in more general settings such as when analyzing data from electronic health records where observations are highly irregular. In simulation studies, we let observation times vary across individuals and demonstrate that not accounting for biasing imbalances due to the monitoring and the exposure schemes can bias the estimate for the marginal odds ratio of exposure.

Revisiting sex differences in outcomes in non-valvular atrial fibrillation: a population-based cohort study.

AIMS: In patients with non-valvular atrial fibrillation (NVAF), it is uncertain whether the higher risk of ischaemic stroke in women reported in some studies is due to residual confounding. We assessed this association using standard time-fixed and more accurate time-dependent adjustment for confounders. METHODS AND RESULTS: Using the computerized databases of the Régie de l'assurance maladie du Québec (RAMQ), we identified a cohort of patients with NVAF during 2000-2009 and RAMQ medication coverage. Cox proportional hazards models were used to estimate the hazard ratio (HR) of ischaemic stroke, death, and bleeding, associated with sex, adjusting for time-fixed covariates at cohort entry. This was compared with adjustment for time-dependent covariates using an age and time-matched nested case-control analysis. The cohort included 147 622 patients. During a mean follow-up of 2.9 years 11 326 patients had a stroke (incidence rate 2.6 per 100 per year). Using time-fixed adjustment for confounders, women had a moderately higher risk of ischaemic stroke than men (HR 1.16 (Confidence interval (CI) 95% 1.11-1.21). Matching on age and using time-dependent adjustment for confounders, women were not at higher risk of stroke than men (Rate Ratio 1.01; 95% CI 0.97-1.05). Mortality and bleeding rates were lower in women compared with men in both analyses. CONCLUSION: In NVAF, women were not at higher risk of thromboembolic events than men in our study. The small increased risk reported in previous studies may be related to residual confounding, in particular from insufficient control for age.

Confounding by Pre-Morbid Functional Status in Studies of Apparent Sex Differences in Severity and Outcome of Stroke.

BACKGROUND AND PURPOSE: Several studies have reported unexplained worse outcomes after stroke in women but none included the full spectrum of symptomatic ischemic cerebrovascular events while adjusting for prior handicap. METHODS: Using a prospective population-based incident cohort of all transient ischemic attack/stroke (OXVASC [Oxford Vascular Study]) recruited between April 2002 and March 2014, we compared pre-morbid and post-event modified Rankin Scale score (mRS) in women and men and change in mRS score 1 month, 6 months, 1 year, and 5 years after stroke. Baseline stroke-related neurological impairment was measured with the National Institutes of Health Stroke Scale. RESULTS: Among 2553 patients (50.6% women) with a first transient ischemic attack/ischemic stroke, women had a worse handicap 1 month after ischemic stroke (age-adjusted odds ratio for mRS score, 1.35; 95% confidence interval, 1.12-1.63). However, women also had a higher pre-morbid mRS score compared with men (age-adjusted odds ratio, 1.58; 95% confidence interval, 1.36-1.84). There was no difference in stroke severity when adjusting for age and pre-morbid mRS (odds ratio, 1.10; 95% confidence interval, 0.90-1.35) and no difference in the pre-/poststroke change in mRS at 1 month (age-adjusted odds ratio, 1.00; 95% confidence interval, 0.82-1.21), 6 months, 1 year, and 5 years. Women had a lower mortality rate, and there was no sex difference in risk of recurrent stroke. CONCLUSIONS: We found no evidence of a worse outcome of stroke in women when adjusting for age and pre-morbid mRS. Failure to account for sex differences in pre-morbid handicap could explain contradictory findings in previous studies. Properties of the mRS may also contribute to these inconsistencies.

Trends in Epidemiology of Pediatric Inflammatory Bowel Disease in Canada: Distributed Network Analysis of Multiple Population-Based Provincial Health Administrative Databases.

OBJECTIVES: The incidence of pediatric-onset inflammatory bowel disease (IBD) is increasing worldwide. We used population-based health administrative data to determine national Canadian IBD incidence, prevalence, and trends over time of childhood-onset IBD. METHODS: We identified children <16 years (y) diagnosed with IBD 1999-2010 from health administrative data in five provinces (Alberta, Manitoba, Nova Scotia, Ontario, Quebec), comprising 79.2% of the Canadian population. Standardized incidence and prevalence were calculated per 100,000 children. Annual percentage change (APC) in incidence and prevalence were determined using Poisson regression analysis. Provincial estimates were meta-analyzed using random-effects models to produce national estimates. RESULTS: 5,214 incident cases were diagnosed during the study period (3,462 Crohn's disease, 1,382 ulcerative colitis, 279 type unclassifiable). The incidence in Canada was 9.68 (95% CI 9.11 to 10.25) per 100,000 children. Incidence was similar amongst most provinces, but higher in Nova Scotia. APC in incidence did not significantly change over the study period in the overall cohort (+2.06%, 95% CI -0.64% to +4.76%). However, incidence significantly increased in children aged 0-5y (+7.19%, 95% +2.82% to +11.56%). Prevalence at the end of the study period in Canada was 38.25 (95% CI 35.78 to 40.73) per 100,000 children. Prevalence increased significantly over time, APC +4.56% (95% CI +3.71% to +5.42%). CONCLUSIONS: Canada has amongst the highest incidence of childhood-onset IBD in the world. Prevalence significantly increased over time. Incidence was not statistically changed with the exception of a rapid increase in incidence in the youngest group of children.

Pneumonia risk in asthma patients using inhaled corticosteroids: a quasi-cohort study.

AIM: Studies have linked the use of inhaled corticosteroids (ICSs) to excess pneumonia risk in chronic obstructive pulmonary disease patients. The risk in asthma patients remains unclear. The objective of the present study was to examine the risk of pneumonia with ICSs in asthma patients aged 12-35 years. METHODS: We formed a cohort of asthma patients treated from 1990 to 2007 using Quebec health insurance databases. Subjects were considered currently exposed if they had had an ICS dispensed within the 60 days prior to their pneumonia index event or matched person-moment. Secondary analyses investigated the risk of pneumonia according to ICS dose and type. Rate ratios (RRs) and rate differences (RDs) were both estimated through a quasi-cohort approach. RESULTS: The cohort included 152 412 subjects, of whom 1928 had a pneumonia event during follow-up. There was an increased risk of pneumonia associated with current use of ICSs [RR 1.83; 95% confidence interval (CI) 1.57, 2.14] or an excess risk of 1.44 cases per 1000 person-years (RD 1.44; 95% CI 1.03, 1.85). There was an excess pneumonia risk with low doses (RR 1.60; 95% CI 1.06, 2.45), moderate doses (RR 1.53; 95% CI 1.12, 2.08) and high doses (RR 1.96; 95% CI 1.64, 2.34) of ICSs, and with budesonide (RR 2.67; 95% CI 2.05, 3.49) and fluticasone (RR 1.93; 95% CI 1.58, 2.36), specifically relative to no use. When accounting for potential protopathic bias, the risk with current use of ICSs was attenuated (RR 1.48; 95% CI 1.22, 1.78). CONCLUSION: ICS use in asthma patients appears to be associated with an increased risk of pneumonia and is present for both budesonide and fluticasone.

Patterns of long-term use of non-vitamin K antagonist oral anticoagulants for non-valvular atrial fibrillation: Quebec observational study.

PURPOSE: Studies on long-term utilization of non-vitamin K antagonist oral anticoagulants (NOACs) in non-valvular atrial fibrillation (NVAF) are scarce. We evaluated predictors of use and long-term persistence of NOACs in a real-world setting. METHODS: This population-based cohort study used the computerized databases of the Canadian Province of Quebec's health insurance. Patients with a first NVAF diagnosis from 2011 until 2014 were included. A logistic regression model yielded adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for predictors of treatment initiation with NOACs versus VKAs. Cox proportional hazards models yielded adjusted hazard ratios (HRs) and 95% CIs for predictors of switching from VKAs to NOACs versus remaining on VKAs, and for predictors of discontinuation of anticoagulation treatment. RESULTS: Of the 62 867 newly diagnosed NVAF patients, 14 646 initiated NOACs and 17 685 VKAs. Initiation with NOACs was less likely for patients ≥ 80 years old (OR 0.55, 95% CI 0.41-0.73) or with CHA2 DS2 -VASc ≥ 2 (OR 0.49, 95% CI 0.42-0.57). Switching from VKAs to NOACs was less likely for patients with chronic kidney disease (HR 0.53, 95% CI 0.48-0.59). After 3 years, persistence was 54% with NOACs and 25% with VKAs. Discontinuation of anticoagulation treatment was less likely for patients ≥ 80 years old (HR 0.47, 95% CI 0.40-0.55) or with CHA2 DS2 -VASc ≥ 2 (HR 0.64, 95% CI 0.57-0.70). CONCLUSIONS: Older, high-risk patients are less likely to initiate NOACs than VKAs. NOAC users show a higher long-term persistence than VKA users, and older, high-risk patients are less likely to discontinue anticoagulation treatment.

The Risk of Sepsis with Inhaled and Oral Corticosteroids in Patients with COPD.

The use of oral and inhaled corticosteroids is associated with increases in the risk of infection, especially pneumonia. The risk of sepsis with corticosteroid treatment in patients with chronic obstructive pulmonary disease (COPD) has been little studied, however. We assessed whether the use of inhaled and oral corticosteroids in COPD is associated with an increase in the risk of sepsis. We carried out a retrospective cohort study using the administrative health databases of the province of Quebec, Canada, over the period 1990-2007. The cohort of patients with COPD included patients aged 55 years or older who used respiratory medications. A quasi-cohort analysis was used to estimate the rate ratio (RR) of sepsis in current users of inhaled corticosteroids and oral corticosteroids separately, after adjusting for differences in COPD disease severity and co-morbid conditions. The cohort included 163,514 patients treated for COPD, including 1,704 who were hospitalized for or died with sepsis during follow-up (incidence rate 1.94 per 1000 per year). The RR of sepsis associated with current use of inhaled corticosteroids was 0.98 (95%confidence interval [CI] 0.84-1.14). Current oral corticosteroid use was associated with a 66% increase in the risk of sepsis (RR 1.66; 95% CI: 1.35-2.05). The increase in risk remains for around 5 months after the oral corticosteroid exposure. Among patients treated for COPD, the risk of sepsis is not increased with inhaled corticosteroids, even at high doses, while the risk is increased with oral corticosteroids. This risk should be considered when treating exacerbations of COPD.

Long-term vitamin K antagonists treatment patterns of Non-Valvular Atrial Fibrillation (NVAF): a population-based cohort study.

BACKGROUND: Recent trends in vitamin K antagonists (VKA) use in non-valvular atrial fibrillation (NVAF) are useful to evaluate the potential improvement in management of NVAF since the introduction of new oral anticoagulants. Our objective was therefore to describe the contemporary VKA treatment patterns following NVAF diagnosis. METHODS AND RESULTS: We used the computerized databases of the Régie de l'assurance maladie du Québec (RAMQ), responsible for administering the universal health care services for all its residents, to identify a population-based cohort of 135,241 patients with an incident diagnosis of NVAF during 2000-2009 and RAMQ medication coverage. Following NVAF diagnosis, 47.1 % of the patients were prescribed VKA, 35.5 % received an antiplatelet only, and 17.4 % did not initiate antithrombotic therapy. The proportion of patients initiating VKA within 3 months of diagnosis increased from 33 % to 39 % over the 10-year study period, mainly driven by a higher proportion of treated patients aged 80 or more (from 29 % to 41 %). At the end of the study period, women were prescribed VKA as frequently as men, except in the subgroup of patients with a low risk of ischemic stroke. The median time from VKA initiation to the first discontinuation varied greatly according to the definition of discontinuation, ranging from 11 months to 5.7 years. CONCLUSION: Although VKA remain underused after NVAF diagnosis, there has been an increase in VKA treatment over the last decade, particularly among older patients. Also the gap in treatment between men and women has been closing within the last decade. Once initiated, most VKA interruptions were temporary rather than definitive.

Health Services Utilization and Specialist Care in Pediatric Inflammatory Bowel Disease: A Multiprovince Population-Based Cohort Study.

BACKGROUND: Patterns of health services utilization among children with inflammatory bowel disease (IBD) are important to understand as the number of children with IBD continues to increase. We compared health services utilization and surgery among children diagnosed <10 years of age (Paris classification: A1a) and between 10 and <16 years of age (A1b). METHODS: Incident cases of IBD diagnosed <16 years of age were identified using validated algorithms from deterministically linked health administrative data in 5 Canadian provinces (Alberta, Manitoba, Nova Scotia, Ontario, Quebec) to conduct a retrospective cohort study. We compared the frequency of IBD-specific outpatient visits, emergency department visits, and hospitalizations across age groups (A1a vs A1b [reference]) using negative binomial regression. The risk of surgery was compared across age groups using Cox proportional hazards models. Models were adjusted for sex, rural/urban residence location, and mean neighborhood income quintile. Province-specific estimates were pooled using random-effects meta-analysis. RESULTS: Among the 1165 (65.7% Crohn's) children with IBD included in our study, there were no age differences in the frequency of hospitalizations (rate ratio [RR], 0.88; 95% confidence interval [CI], 0.74-1.06) or outpatient visits (RR, 0.95; 95% CI, 0.78-1.16). A1a children had fewer emergency department visits (RR, 0.70; 95% CI, 0.50-0.97) and were less likely to require a Crohn's-related surgery (hazard ratio, 0.49; 95% CI, 0.26-0.92). The risk of colectomy was similar among children with ulcerative colitis in both age groups (hazard ratio, 0.71; 95% CI, 0.49-1.01). CONCLUSIONS: Patterns of health services utilization are generally similar when comparing children diagnosed across age groups.

Among 1165 children with inflammatory bowel disease, health services utilization was similar for children diagnosed <10 years of age and those diagnosed ≥10 years of age, except younger children had fewer emergency department visits and Crohn's disease­related surgeries.

Estimating individualized treatment rules in longitudinal studies with covariate-driven observation times.

The sequential treatment decisions made by physicians to treat chronic diseases are formalized in the statistical literature as dynamic treatment regimes. To date, methods for dynamic treatment regimes have been developed under the assumption that observation times, that is, treatment and outcome monitoring times, are determined by study investigators. That assumption is often not satisfied in electronic health records data in which the outcome, the observation times, and the treatment mechanism are associated with patients' characteristics. The treatment and observation processes can lead to spurious associations between the treatment of interest and the outcome to be optimized under the dynamic treatment regime if not adequately considered in the analysis. We address these associations by incorporating two inverse weights that are functions of a patient's covariates into dynamic weighted ordinary least squares to develop optimal single stage dynamic treatment regimes, known as individualized treatment rules. We show empirically that our methodology yields consistent, multiply robust estimators. In a cohort of new users of antidepressant drugs from the United Kingdom's Clinical Practice Research Datalink, the proposed method is used to develop an optimal treatment rule that chooses between two antidepressants to optimize a utility function related to the change in body mass index.

Generating synthetic data from administrative health records for drug safety and effectiveness studies.

Introduction: Administrative health records (AHRs) are used to conduct population-based post-market drug safety and comparative effectiveness studies to inform healthcare decision making. However, the cost of data extraction, and the challenges associated with privacy and securing approvals can make it challenging for researchers to conduct methodological research in a timely manner using real data. Generating synthetic AHRs that reasonably represent the real-world data are beneficial for developing analytic methods and training analysts to rapidly implement study protocols. We generated synthetic AHRs using two methods and compared these synthetic AHRs to real-world AHRs. We described the challenges associated with using synthetic AHRs for real-world study. Methods: The real-world AHRs comprised prescription drug records for individuals with healthcare insurance coverage in the Population Research Data Repository (PRDR) from Manitoba, Canada for the 10-year period from 2008 to 2017. Synthetic data were generated using the Observational Medical Dataset Simulator II (OSIM2) and a modification (ModOSIM). Synthetic and real-world data were described using frequencies and percentages. Agreement of prescription drug use measures in PRDR, OSIM2 and ModOSIM was estimated with the concordance coefficient. Results: The PRDR cohort included 169,586,633 drug records and 1,395 drug types for 1,604,734 individuals. Synthetic data for 1,000,000 individuals were generated using OSIM2 and ModOSIM. Sex and age group distributions were similar in the real-world and synthetic AHRs. However, there were significant differences in the number of drug records and number of unique drugs per person for OSIM2 and ModOSIM when compared with PRDR. For the average number of days of drug use, concordance with the PRDR was 16% (95% confidence interval [CI]: 12%-19%) for OSIM2 and 88% (95% CI: 87%-90%) for ModOSIM. Conclusions: ModOSIM data were more similar to PRDR than OSIM2 data on many measures. Synthetic AHRs consistent with those found in real-world settings can be generated using ModOSIM. Synthetic data will benefit rapid implementation of methodological studies and data analyst training.

No Change in Surgical and Hospitalization Trends Despite Higher Exposure to Anti-Tumor Necrosis Factor in Inflammatory Bowel Disease in the Québec Provincial Database From 1996 to 2015.

BACKGROUND: Crohn disease (CD) and ulcerative colitis (UC) have high health care expenditures because of medications, hospitalizations, and surgeries. We evaluated disease outcomes and treatment algorithms of patients with inflammatory bowel disease (IBD) in Québec, comparing periods before and after 2010. METHODS: The province of Québec's public health administrative database was used to identify newly diagnosed patients with IBD between 1996 and 2015. The primary and secondary outcomes included time to and probability of first and second IBD-related hospitalizations, first and second major surgery, and medication exposures. Medication prescriptions were collected from the public prescription database. RESULTS: We identified 34,644 newly diagnosed patients with IBD (CD = 59.5%). The probability of the first major surgery increased after 2010 in patients with CD (5 years postdiagnosis before and after 2010: 8% [SD = 0.2%] vs 15% [SD = 0.6%]; P < 0.0001) and patients with UC (6% [SD = 0.2%] vs 10% [SD = 0.6%] ;P < 0.0001). The probability of the second major surgery was unchanged in patients with CD. Hospitalization rates remained unchanged. Patients on anti-tumor necrosis factor (anti-TNF) medications had the lowest probability of hospitalizations (overall 5-year probability in patients with IBD stratified by maximal therapeutic step: 5-aminosalicylic acids 37% [SD = 0.6%]; anti-TNFs 31% [SD = 1.8%]; P < 0.0001). Anti-TNFs were more commonly prescribed for patients with CD after 2010 (4% [SD = 0.2%] vs 16% [SD = 0.6%]; P < 0.0001) in the public health insurance plan, especially younger patients. Corticosteroid exposure was unchanged before and after 2010. Immunosuppressant use was low but increased after 2010. The use of 5-ASAs was stable in patients with UC but decreased in patients with CD. CONCLUSIONS: The probability of first and second hospitalizations remained unchanged in Québec and the probability of major surgery was low overall but did increase despite the higher and earlier use of anti-TNFs.

Long-term Use of Inhaled Corticosteroids in COPD and the Risk of Fracture.

BACKGROUND: It is uncertain whether long-term use of inhaled corticosteroids (ICSs), widely used to treat COPD, increases the risk of fracture, particularly in women, in view of the postmenopausal risks. We assessed whether long-term ICS use in patients with COPD increases the risk of hip or upper extremity fractures, and examined sex-related differences. METHODS: The Quebec health-care databases were used to form a cohort of patients with COPD over 1990 to 2005, followed until 2007 for the first hip or upper extremity fracture. In a nested case-control analysis, each case of fracture was matched with 20 control subjects on age, sex, and follow-up time. The adjusted rate ratio (RR) of fracture with ICS use, by duration and dose, was estimated using conditional logistic regression, with an interaction term to compare the risk in men and women. RESULTS: In the cohort of 240,110 subjects, 19,396 sustained a fracture during a mean 5.3 years (rate, 15.2 per 1,000 per year). Any use of ICSs was not associated with an increased rate of fracture (RR, 1.00; 95% CI, 0.97-1.03). The fracture rate was increased with > 4 years of ICS use at daily doses ≥ 1,000 µg in fluticasone equivalents (RR, 1.10; 95% CI, 1.02-1.19). This risk increase did not differ between men and women. CONCLUSIONS: Long-term ICS use at high doses is associated with a modest increase in the risk of hip and upper extremity fractures in patients with COPD. This dose-duration risk increase does not appear to be higher for women.