Linda E. Lévesque

Problem of immortal time bias in cohort studies: example using statins for preventing progression of diabetes

Immortal time in observational studies can bias the results in favour of the treatment group, but it is not difficult to identify and avoid

Interrupted time series analysis in drug utilization research is increasing: systematic review and recommendations

OBJECTIVES To describe the use and reporting of interrupted time series methods in drug utilization research. STUDY DESIGN AND SETTING We completed a systematic search of MEDLINE, Web of Science, and reference lists to identify English language articles through to December 2013 that used interrupted time series methods in drug utilization research. We tabulated the number of studies by publication year and summarized methodological detail. RESULTS We identified 220 eligible empirical applications since 1984. Only 17 (8%) were published before 2000, and 90 (41%) were published since 2010. Segmented regression was the most commonly applied interrupted time series method (67%). Most studies assessed drug policy changes (51%, n = 112); 22% (n = 48) examined the impact of new evidence, 18% (n = 39) examined safety advisories, and 16% (n = 35) examined quality improvement interventions. Autocorrelation was considered in 66% of studies, 31% reported adjusting for seasonality, and 15% accounted for nonstationarity. CONCLUSION Use of interrupted time series methods in drug utilization research has increased, particularly in recent years. Despite methodological recommendations, there is large variation in reporting of analytic methods. Developing methodological and reporting standards for interrupted time series analysis is important to improve its application in drug utilization research, and we provide recommendations for consideration.

Time variations in the risk of myocardial infarction among elderly users of COX-2 inhibitors.

Background: The timing of cardiovascular risks associated with the use of cyclooxygenase-2 (COX-2) inhibitors is unclear. Using data collected in a previous population-based cohort study of elderly people starting nonsteroidal anti-inflammatory drug (NSAID) therapy, we evaluated the temporal nature of the risk of a first myocardial infarction (MI) associated with the use of rofecoxib and celecoxib. Methods: We identified people 66 years of age or older without previous MI who were currently taking rofecoxib and celecoxib using Quebecs computerized health databases (January 1999 to June 2002). Data on use and MI outcome were analyzed using a time-matched, nested case–control approach with rate ratios, comparing current users and non-users of rofecoxib and celecoxib in the year preceding the index date, estimated using conditional logistic regression. Results: The risk of MI was highest following first-time use of rofecoxib (adjusted rate ratio [RR] 1.67, 95% confidence interval [CI] 1.21–2.30), with events occurring within a median of 9 (6–13) days after therapy was started. The risk increase for first-time use of celecoxib was not statistically significant (RR 1.29, 95% CI 0.90–1.83). Repeated exposure to rofecoxib was associated with a small but statistically nonsignificant delayed risk (RR 1.17, 95% CI 0.98–1.40), but no risk was seen with celecoxib (RR 0.97, 95% CI 0.82–1.14). Treatment duration was not associated with increasing risk for either agent. The risk remained elevated for the first 7 days after rofecoxib was discontinued (RR 1.23, 95% CI 1.05–1.44) but appeared to return to baseline between day 8 and 30 (RR 0.82, 95% CI 0.61–1.09). Interpretation: A small proportion of patients using rofecoxib for the first time had their first MI shortly after starting the drug. This risk did not increase with the length of treatment and returned to baseline shortly after treatment was discontinued. More research is needed to identify those most susceptible to cardiotoxicity mediated by COX-2 inhibitor therapy.

Antipsychotic Drugs and Hyperglycemia in Older Patients With Diabetes

BACKGROUND Evidence suggests that there is an association between antipsychotic drugs and new-onset diabetes, but little is known about the risk of hyperglycemia among persons with preexisting diabetes. METHODS Using a nested case-control design and population-based health databases in Ontario, Canada, persons aged 66 years or older with diabetes who started treatment with an antipsychotic drug from April 1, 2002, to March 31, 2006, were followed up from treatment start until March 31, 2007. The cohort was subdivided into 3 groups: insulin-treated, oral hypoglycemic agent only-treated, and no diabetes treatment. We defined cases as patients hospitalized (emergency department visit or hospital admissions) for hyperglycemia. Each case was matched with up to 10 controls. We compared the likelihood of hyperglycemia among current users of atypical and typical antipsychotic agents with that among remote antipsychotic users (discontinued >180 days), based on prescriptions closest to event date. RESULTS Of 13 817 patients studied, 1515 (11.0%) were hospitalized for hyperglycemia. Current antipsychotic treatment was associated with a higher risk of hyperglycemia compared with remote antipsychotic use in all diabetes treatment groups (overall adjusted rate ratio, 1.50; 95% confidence interval, 1.29-1.74). The risk was increased among patients who were treated with atypical and typical antipsychotic agents and was extremely high among patients who were just starting treatment (only 1 prescription before event). CONCLUSIONS Among older patients with diabetes, the initiation of treatment with antipsychotic drugs was associated with a significantly increased risk of hospitalization for hyperglycemia (P < .001). The risk was particularly high during the initial course of treatment and was increased with the use of all antipsychotic agents.

The coronary risk of cyclo-oxygenase-2 inhibitors in patients with a previous myocardial infarction

Background: Cyclo-oxygenase-2 selective inhibitors have been associated with cardiovascular side effects, but previous studies have generally excluded people with previous myocardial infarction, thereby limiting our knowledge of their cardiotoxicity in this population. Objectives: To determine whether a history of myocardial infarction modified the risk of acute myocardial infarction associated with the use of various non-steroidal anti-inflammatory drugs (NSAIDs). Methods: A population-based cohort of 122 079 elderly people with and without previous myocardial infarction newly treated with an NSAID between 1 January 1999 and 30 June 2002 were identified using the computerised health databases of Québec, Canada. A nested-case–control approach was used for the analysis, with controls matched by cohort entry and age. Current users of NSAIDs, those whose last prescription overlapped with the index date, were compared with those who were not exposed to NSAIDs in the year preceding the event. Rate ratios of acute myocardial infarction were estimated using conditional logistic regression and adjusted for potential confounders. Results: Users of rofecoxib, both with and without previous myocardial infarction, were at increased risk of myocardial infarction, with a trend for greater risk among those with a previous event (rate ratio (RR) 1.59, 95% confidence interval (CI) 1.15 to 2.18 v RR 1.23, 95% CI 1.05 to 1.45; p = 0.14 for interaction). By contrast, celecoxib was only associated with an increased risk in people with previous myocardial infarction (RR 1.40, 95% CI 1.06 to 1.84 v RR 1.03, 95% CI 0.88 to 1.20; p = 0.04 for interaction). The available power was insufficient to reliably assess risks among patients with previous myocardial infarction treated with other NSAIDs, dose–response relationships or interaction with aspirin. Conclusions: Although only rofecoxib use was associated with an increased risk of myocardial infarction in those without a previous event, both rofecoxib and celecoxib were associated with an excess risk of acute myocardial infarction for current users with a history of myocardial infarction. A large randomised trial is required to more completely and reliably assess the cardiovascular safety of celecoxib and traditional NSAIDs in this population of high-risk patients.

Factors associated with initiation and completion of the quadrivalent human papillomavirus vaccine series in an ontario cohort of grade 8 girls

BackgroundAlthough over a hundred million dollars have been invested in offering free quadrivalent human papillomavirus (HPV) vaccination to young girls in Ontario, there continues to be very little information about its usage. In order to successfully guide future HPV vaccine programming, it is important to monitor HPV vaccine use and determine factors associated with use in this population.MethodsLinking administrative health and immunization databases, we conducted a population-based, retrospective cohort study of girls eligible for Ontarios Grade 8 HPV vaccination program in Kingston, Frontenac, Lennox, and Addington. We determined the proportion of girls who initiated (at least one dose) and completed (all three doses) the vaccination series overall and according to socio-demographics, vaccination history, health services utilization, medical history, and program year. Multivariable logistic regression was used to estimate the strength of association between individual factors and initiation and completion, adjusted for all other factors.ResultsWe identified a cohort of 2519 girls, 56.6% of whom received at least one dose of the HPV vaccine. Among vaccinated girls, 85.3% received all three doses. Vaccination history was the strongest predictor of initiation in that girls who received the measles-mumps-rubella, meningococcal C, and hepatitis B vaccines were considerably more likely to also receive the HPV vaccine (odds ratio 4.89; 95% confidence interval 4.04-5.92). Nevertheless, HPV vaccine uptake was more than 20% lower than that of these other vaccines. In addition, while series initiation was not influenced by income, series completion was. In particular, girls of low income were the least likely to receive all three indicated doses of the HPV vaccine (odds ratio 0.45; 95% confidence interval 0.28-0.72).ConclusionsThe current low level of HPV vaccine acceptance in Kingston, Frontenac, Lennox, and Addington will likely have important implications in terms of the health benefits and cost-effectiveness of its publicly funded program. We identified important factors associated with series initiation and completion that should be considered in efforts to improve HPV vaccine use in this population.

Does the addition of functional status indicators to case-mix adjustment indices improve prediction of hospitalization, institutionalization, and death in the elderly?

Background:Case-mix adjustment is widely used in health services research to ensure that groups being compared are equivalent on variables predicting outcome. There has been considerable development and testing of comorbidity indices derived from diagnostic codes recorded in administrative databases, but increasingly, the benefit of clinical information and patient reported ratings of health and functional status is being recognized. One type of information that is highly valued but has so far not been captured by administrative health databases is functional status indicators (FSI). Objective:The purpose of this study was to estimate the extent to which prediction of health outcomes can be improved on by including information on functional status indicators (FSI). Research Design:The data for the current study was obtained from a clustered randomized trial evaluating computerized decision support for managing drug therapy in the elderly, conducted from 1997 to 1998. A total of 107 primary care physicians participated in this trial and 6465 of their patients (51%) completed a generic health status measure—the SF-12—before the intervention. C statistics and R2 were used to compare the predictive value of sociodemographic factors, 2 comorbidity indices, and 11 FSI predictor variables derived from the SF-12 and coded (possible for 8) using the International Classification of Functioning (ICF). Results:Using stepwise logistic regression, FSI, particularly limitation in stair climbing or doing moderate activities like housework, were found to be strong and independent predictors of all outcomes, even after controlling for sociodemographics and comorbidity. Conclusion:This study indicates that FSI provided as robust a prediction of health events as did complex comorbidity indices. Additionally, the ICF coding system provides a mechanism whereby information on FSI could be incorporated into administrative databases through the use of electronic health records that include a health or functional status measure.

Effect of human papillomavirus (HPV) vaccination on clinical indicators of sexual behaviour among adolescent girls: the Ontario Grade 8 HPV Vaccine Cohort Study

Background: Suboptimal human papillomavirus (HPV) vaccine coverage in some jurisdictions is partly attributed to fears that vaccination may increase risky sexual behaviour. We assessed the effect of HPV vaccination on clinical indicators of sexual behaviour among adolescent girls in Ontario. Methods: Using Ontario’s administrative health databases, we identified a population-based cohort of girls in grade 8 in the 2 years before (2005/06 and 2006/07) and after (2007/08 and 2008/09) implementation of Ontario’s grade 8 HPV vaccination program. For each girl, we then obtained data on vaccine receipt in grades 8 and 9 and data on indicators of sexual behaviour (pregnancy and non–HPV-related sexually transmitted infections) in grades 10–12. Using a quasi-experimental method known as regression discontinuity, we estimated, for each outcome, the risk difference (RD) and relative risk (RR) attributable to vaccination and to program eligibility. Results: The cohort comprised 260 493 girls, of whom 131 781 were ineligible for the program and 128 712 were eligible. We identified 15 441 (5.9%) cases of pregnancy and sexually transmitted infection and found no evidence that vaccination increased the risk of this composite outcome: RD per 1000 girls −0.61 (95% confidence interval [CI] −10.71 to 9.49) and RR 0.96 (95% CI 0.81 to 1.14). Similarly, we found no discernible effect of program eligibility: RD per 1000 girls −0.25 (95% CI −4.35 to 3.85) and RR 0.99 (95% CI 0.93 to 1.06). The findings were similar when outcomes were assessed separately. Interpretation: We present strong evidence that HPV vaccination does not have any significant effect on clinical indicators of sexual behaviour among adolescent girls. These results suggest that concerns over increased promiscuity following HPV vaccination are unwarranted and should not deter from vaccinating at a young age.

The Early Benefits of Human Papillomavirus Vaccination on Cervical Dysplasia and Anogenital Warts

BACKGROUND: Despite widespread promotion of quadrivalent human papillomavirus (qHPV) vaccination for young girls, there is limited information on the vaccine’s real-world effectiveness and none on the effectiveness of qHPV vaccination programs. We assessed the impact of the qHPV vaccine and Ontario’s grade 8 qHPV vaccination program on cervical dysplasia and anogenital warts (AGW). METHODS: By using administrative health databases of Ontario, Canada, we identified a population-based retrospective cohort of girls in grade 8 before (2005/2006–2006/2007) and after (2007/2008–2008/2009) program implementation. Vaccine exposure was ascertained in grades 8 to 9 and outcomes in grades 10 to 12. A quasi-experimental approach known as regression discontinuity was used to estimate absolute risk differences (RDs), relative risks (RRs), and 95% confidence intervals (CIs) attributable to vaccination and program eligibility (intention-to-treat analysis). RESULTS: The cohort comprised 131 781 ineligible and 128 712 eligible girls (n = 260 493). We identified 2436 cases of dysplasia and 400 cases of AGW. Vaccination significantly reduced the incidence of dysplasia by 5.70 per 1000 girls (95% CI −9.91 to −1.50), corresponding to a relative reduction of 44% (RR 0.56; 95% CI 0.36 to 0.87). Program eligibility also had a significant protective effect on dysplasia: RD −2.32/1000 (95% CI −4.02 to −0.61); RR 0.79 (95% CI 0.66 to 0.94). Results suggested decreases in AGW attributable to vaccination (RD −0.83/1000, 95% CI −2.54 to 0.88; RR 0.57, 95% CI 0.20 to 1.58) and program eligibility (RD −0.34/1000, 95% CI −1.03 to 0.36; RR 0.81, 95% CI 0.52 to 1.25). CONCLUSIONS: This study provides strong evidence of the early benefits of qHPV vaccination among girls aged 14 to 17 years, offering additional justification for not delaying vaccination.

Antipsychotic Drugs and the Risk of Hyperglycemia in Older Adults Without Diabetes: A Population-Based Observational Study

OBJECTIVE To determine whether current antipsychotic use among older persons without diabetes is associated with a higher risk of hospital visits for hyperglycemia, as previous studies in this population have yielded conflicting results. DESIGN, SETTING AND PARTICIPANTS A nested case-control study within a population-based cohort of persons aged 66 years or older without diabetes, who initiated antipsychotic therapy between April 1, 2002, and March 31, 2006. Cohort members were identified using health databases from Ontario, Canada, and were followed from treatment start until March 31, 2007. MEASUREMENTS Cases were patients with a hospital visit (emergency department visit or hospital admission) for hyperglycemia. We matched each case with up to 10 controls. We compared the risk of hyperglycemia among current antipsychotic users to that of remote users (discontinued > 180 days). RESULTS The cohort consisted of 44,121 subjects, mean age of 78.3 years, followed for a mean of 2.2 years. Compared to remote antipsychotic use, current treatment with any antipsychotic was associated with a significantly increased risk of hospital visits for hyperglycemia (adjusted odds ratio [aOR]: 1.52; 95% confidence interval [CI]: 1.07-2.17). The risk was elevated for both atypical (aOR: 1.44; 95% CI: 1.01-2.07) and typical (aOR: 2.86; 95% CI: 1.46-5.59) antipsychotic agents. CONCLUSIONS Current use of either atypical or typical antipsychotic agents was associated with a significantly increased risk of hospital visits for hyperglycemia among older persons without diabetes. These findings highlight the need for close glucose monitoring during antipsychotic therapy in older populations.