Emulating a trial of joint dynamic strategies: An application to monitoring and treatment of HIV-positive individuals.

Decisions about when to start or switch a therapy often depend on the frequency with which individuals are monitored or tested. For example, the optimal time to switch antiretroviral therapy depends on the frequency with which HIV-positive individuals have HIV RNA measured. This paper describes an approach to use observational data for the comparison of joint monitoring and treatment strategies and applies the method to a clinically relevant question in HIV research: when can monitoring frequency be decreased and when should individuals switch from a first-line treatment regimen to a new regimen? We outline the target trial that would compare the dynamic strategies of interest and then describe how to emulate it using data from HIV-positive individuals included in the HIV-CAUSAL Collaboration and the Centers for AIDS Research Network of Integrated Clinical Systems. When, as in our example, few individuals follow the dynamic strategies of interest over long periods of follow-up, we describe how to leverage an additional assumption: no direct effect of monitoring on the outcome of interest. We compare our results with and without the "no direct effect" assumption. We found little differences on survival and AIDS-free survival between strategies where monitoring frequency was decreased at a CD4 threshold of 350 cells/µl compared with 500 cells/µl and where treatment was switched at an HIV-RNA threshold of 1000 copies/ml compared with 200 copies/ml. The "no direct effect" assumption resulted in efficiency improvements for the risk difference estimates ranging from an 7- to 53-fold increase in the effective sample size.

Who Will Show? Predicting Missed Visits Among Patients in Routine HIV Primary Care in the United States.

Missed HIV medical visits predict poor clinical outcomes. We sought to identify patients at high risk of missing visits. We analyzed 2002-2014 data from six large US HIV clinics. At each visit, we predicted the likelihood of missing the next scheduled visit using demographic, clinical, and patient-reported psychosocial variables. Overall, 10,374 participants contributed 105,628 HIV visits. For 17% of visits, the next scheduled appointment was missed. The strongest predictor of a future missed visit was past-year missed visits. A model with only this predictor had area under the receiver operator curve = 0.65; defining "high risk" as those with any past-year missed visits had 73% sensitivity and 51% specificity in correctly identifying a future missed visit. Inclusion of other clinical and psychosocial predictors only slightly improved performance. Past visit attendance can identify those at increased risk for future missed visits, allowing for proactive allocation of resources to those at greatest risk.

Physical Activity Intensity is Associated with Symptom Distress in the CNICS Cohort.

Symptom distress remains a challenging aspect of living with HIV. Physical activity is a promising symptom management strategy, but its effect on symptom distress has not been examined in a large, longitudinal HIV-infected cohort. We hypothesized that higher physical activity intensity would be associated with reduced symptom distress. We included 5370 people living with HIV (PLHIV) who completed patient-reported assessments of symptom distress, physical activity, alcohol and substance use, and HIV medication adherence between 2005 and 2016. The most frequent and burdensome symptoms were fatigue (reported by 56%), insomnia (50%), pain (46%), sadness (45%), and anxiety (45%), with women experiencing more symptoms and more burdensome symptoms than men. After adjusting for age, sex, race, time, HIV medication adherence, alcohol and substance use, site, and HIV RNA, greater physical activity intensity was associated with lower symptom intensity. Although individual symptoms may be a barrier to physical activity (e.g. pain), the consistent association between symptoms with physical activity suggests that more intense physical activity could mitigate symptoms experienced by PLHIV.

HIV Preexposure Prophylaxis as a Gateway to Primary Care.

OBJECTIVES: To determine whether HIV preexposure prophylaxis (PrEP) use is associated with use of non-HIV-related health care. METHODS: We conducted a cross-sectional study of potential PrEP candidates at a Boston, Massachusetts, community health clinic during 2012 to 2016, comparing the proportion of PrEP users and non-PrEP users receiving primary care. RESULTS: Of 5857 PrEP candidates, 2047 (35%) were prescribed PrEP. After adjustment for demographics and number of visits, more PrEP users received influenza vaccination (prevalence ratio [PR] = 1.28; 95% confidence interval [CI] = 1.20, 1.37), tobacco screening (PR = 1.06; 95% CI = 1.02, 1.09), and depression screening (PR = 1.07; 95% CI = 1.04, 1.11) compared with non-PrEP users. After additional adjustment for diabetes, hypertension, and overweight or obesity, more PrEP users received glucose testing (PR = 1.64; 95% CI = 1.56, 1.72) but fewer received hemoglobin A1c testing (PR = 0.81; 95% CI = 0.71, 0.93) compared with non-PrEP users. CONCLUSIONS: PrEP use was associated with receipt of influenza vaccination, tobacco and depression screening, and glucose but not hemoglobin A1c testing. Among PrEP users receiving routine care, the benefits of PrEP may extend to behavioral health, mental health, and prevention and treatment of other infectious and chronic diseases.

Low-density lipoprotein cholesterol response after statin initiation among persons living with human immunodeficiency virus.

BACKGROUND: Meta-analyses of general population studies report mean low-density lipoprotein cholesterol (LDL-C) reductions of 30% to <50% with moderate-intensity and ≥50% with high-intensity statins. Persons living with human immunodeficiency virus (PLWH) are at high risk for atherosclerotic cardiovascular disease (ASCVD), yet many have elevated LDL-C. OBJECTIVE: To evaluate LDL-C response after statin initiation among PLWH. METHODS: We conducted a retrospective cohort study of PLWH initiating statins between 2009 and 2013 (N = 706). Patients were categorized into mutually exclusive groups in the following hierarchy: history of coronary heart disease (CHD), diabetes, prestatin LDL-C ≥190 mg/dL, 10-year predicted ASCVD risk ≥7.5%, and none of the above (ie, unknown statin indication). The primary outcome was a ≥30% reduction in LDL-C after statin initiation. RESULTS: Among patients initiating statins, 5.8% had a history of CHD, 13.6% had diabetes, 6.2% had LDL-C ≥190 mg/dL, 35.4% had 10-year ASCVD risk ≥7.5%, and 39.0% had an unknown statin indication. Among patients with a history of CHD, 31.7% achieved a ≥30% LDL-C reduction compared with 25.0%, 59.1%, and 33.9% among those with diabetes, LDL-C ≥190 mg/dL, and 10-year ASCVD risk ≥7.5%, respectively. In multivariable adjusted analyses and compared to patients with an unknown statin indication, LDL-C ≥ 190 mg/dL was associated with a prevalence ratio for an LDL-C reduction ≥30% of 1.81 (95% confidence interval, 1.34-2.45), whereas no statistically significant association was present for history of CHD, diabetes, and 10-year ASCVD risk ≥7.5%. CONCLUSION: A low percentage of PLWH achieved the expected reductions in LDL-C after statin initiation, highlighting an unmet need for ASCVD risk reduction.

Factors Associated With Gaps in Medicaid Enrollment Among People With HIV and the Effect of Gaps on Viral Suppression.

INTRODUCTION: Gaps in Medicaid enrollment may affect HIV outcomes. We evaluated factors associated with Medicaid enrollment gaps and their effect on viral suppression (VS) within the HIV Research Network. METHODS: We used a combined data set with Medicaid enrollment files from 2006 to 2010 and HIV Research Network demographic and clinical data. A gap was defined as ≥1 month without Medicaid and gap length was determined. We used multivariable logistic regression to determine factors associated with a gap and multivariable logistic regression with generalized estimated equations to evaluate factors associated with VS after gap. RESULTS: Of 5836 participants, the majority were male, of black race, and aged 25-50 years. More than half had a gap in Medicaid. Factors associated with a gap included male sex [adjusted odds ratio (aOR) 1.79, (1.53, 2.08)] and younger age (aORs ranging from 1.50 to 4.13 comparing younger age groups to age >50, P < 0.05 for all). About a quarter of gaps had VS information before and after gap. Of those, 53.7% had VS both before and after gap and 25.8% were unsuppressed both before and after gap. The strongest association with VS after gap was VS before gap [aOR 15.76 (10.48, 23.69)]. Transition into Ryan White HIV/AIDS Program coverage during Medicaid gaps was common (28% of all transitions). CONCLUSIONS: Gaps in Medicaid enrollment were common and many individuals with pre-gap VS maintained VS after gap, possibly due to accessing other sources of antiretroviral therapy coverage. Implementing initiatives to maintain Medicaid enrollment and to expedite Medicaid reenrollment and having alternate resources available in gaps are important to ensure continuous antiretroviral therapy to optimize HIV outcomes.

Evaluating the Population Impact on Racial/Ethnic Disparities in HIV in Adulthood of Intervening on Specific Targets: A Conceptual and Methodological Framework.

Reducing racial/ethnic disparities in human immunodeficiency virus (HIV) disease is a high priority. Reductions in HIV racial/ethnic disparities can potentially be achieved by intervening on important intermediate factors. The potential population impact of intervening on intermediates can be evaluated using observational data when certain conditions are met. However, using standard stratification-based approaches commonly employed in the observational HIV literature to estimate the potential population impact in this setting may yield results that do not accurately estimate quantities of interest. Here we describe a useful conceptual and methodological framework for using observational data to appropriately evaluate the impact on HIV racial/ethnic disparities of interventions. This framework reframes relevant scientific questions in terms of a controlled direct effect and estimates a corresponding proportion eliminated. We review methods and conditions sufficient for accurate estimation within the proposed framework. We use the framework to analyze data on 2,329 participants in the CFAR [Centers for AIDS Research] Network of Integrated Clinical Systems (2008-2014) to evaluate the potential impact of universal prescription of and ≥95% adherence to antiretroviral therapy on racial disparities in HIV virological suppression. We encourage the use of the described framework to appropriately evaluate the potential impact of targeted interventions in addressing HIV racial/ethnic disparities using observational data.

Sociodemographic and Clinical Factors Associated With Increasing Bacterial Sexually Transmitted Infection Diagnoses in Men Who Have Sex With Men Accessing Care at a Boston Community Health Center (2005-2015).

BACKGROUND: The reasons why bacterial sexually transmitted infections (BSTIs) are increasing in US men who have sex with men (MSM) have not been fully characterized. METHODS: An open cohort of MSM accessing medical care at a Boston community health center was used to assess secular trends in BSTI diagnoses. Frequency of infection and the estimated population size were used to calculate diagnosis rates. Poisson models were fit for multivariable analyses. RESULTS: Between 2005 and 2015, 19 232 men had at least 1 clinic visit. Most (72.4%) were white; 6.0% were black, and 6.1% were Latino. Almost half had documented self-report of identifying as gay (42.6%) or bisexual (3.2%). Most had private health insurance (61.7%); 5.4% had Medicare, 4.6% had Medicaid, and 8.4% reported no insurance. Between 2005 and 2015, BSTI diagnoses increased more than 8-fold. In 2015, of 1319 men who were diagnosed with at least 1 BSTI; 291 were diagnosed with syphilis, 554 with gonorrhea (51.4% rectal, 31.0% urogenital), and 679 with chlamydia (69.1% rectal, 34.3% urogenital). In 2015, 22.7% of BSTIs were diagnosed among HIV-infected patients (15.4% of the clinic population), and 32.8% of BSTIs were diagnosed among HIV-uninfected patients using pre-exposure prophylaxis (PrEP; 10.1% of all men in care). In multivariable analyses, age 18 to 24 years, being HIV-infected, using PrEP, being nonwhite, or reporting Medicaid or not reporting having private insurance or Medicare were independently associated with being diagnosed with a new BSTI. CONCLUSIONS: Over the past decade, BSTI diagnosis rates increased in HIV-infected and uninfected MSM, with disproportionate increases in PrEP users, racial and ethnic minority MSM, those aged 25 to 34 years, and those without stable health insurance, warranting focused education, screening, and accessible services for these key subpopulations.

Comparison of dynamic monitoring strategies based on CD4 cell counts in virally suppressed, HIV-positive individuals on combination antiretroviral therapy in high-income countries: a prospective, observational study.

BACKGROUND: Clinical guidelines vary with respect to the optimal monitoring frequency of HIV-positive individuals. We compared dynamic monitoring strategies based on time-varying CD4 cell counts in virologically suppressed HIV-positive individuals. METHODS: In this observational study, we used data from prospective studies of HIV-positive individuals in Europe (France, Greece, the Netherlands, Spain, Switzerland, and the UK) and North and South America (Brazil, Canada, and the USA) in The HIV-CAUSAL Collaboration and The Centers for AIDS Research Network of Integrated Clinical Systems. We compared three monitoring strategies that differ in the threshold used to measure CD4 cell count and HIV RNA viral load every 3-6 months (when below the threshold) or every 9-12 months (when above the threshold). The strategies were defined by the threshold CD4 counts of 200 cells per µL, 350 cells per µL, and 500 cells per µL. Using inverse probability weighting to adjust for baseline and time-varying confounders, we estimated hazard ratios (HRs) of death and of AIDS-defining illness or death, risk ratios of virological failure, and mean differences in CD4 cell count. FINDINGS: 47 635 individuals initiated an antiretroviral therapy regimen between Jan 1, 2000, and Jan 9, 2015, and met the eligibility criteria for inclusion in our study. During follow-up, CD4 cell count was measured on average every 4·0 months and viral load every 3·8 months. 464 individuals died (107 in threshold 200 strategy, 157 in threshold 350, and 200 in threshold 500) and 1091 had AIDS-defining illnesses or died (267 in threshold 200 strategy, 365 in threshold 350, and 459 in threshold 500). Compared with threshold 500, the mortality HR was 1·05 (95% CI 0·86-1·29) for threshold 200 and 1·02 (0·91·1·14) for threshold 350. Corresponding estimates for death or AIDS-defining illness were 1·08 (0·95-1·22) for threshold 200 and 1·03 (0·96-1·12) for threshold 350. Compared with threshold 500, the 24 month risk ratios of virological failure (viral load more than 200 copies per mL) were 2·01 (1·17-3·43) for threshold 200 and 1·24 (0·89-1·73) for threshold 350, and 24 month mean CD4 cell count differences were 0·4 (-25·5 to 26·3) cells per µL for threshold 200 and -3·5 (-16·0 to 8·9) cells per µL for threshold 350. INTERPRETATION: Decreasing monitoring to annually when CD4 count is higher than 200 cells per µL compared with higher than 500 cells per µL does not worsen the short-term clinical and immunological outcomes of virally suppressed HIV-positive individuals. However, more frequent virological monitoring might be necessary to reduce the risk of virological failure. Further follow-up studies are needed to establish the long-term safety of these strategies. FUNDING: National Institutes of Health.

Comparative effectiveness of single versus multiple tablet antiretroviral therapy regimens in clinical HIV practice.

We determined risk of virologic failure (VF) in individuals initiating tenofovir/emtricitabine/efavirenz as single versus multiple tablet regimens (MTR). We found no significant difference in the risk of VF, though did observe a trend toward more VF and M184 V mutations among persons initiating MTR. Temporal trends in care may have confounded results.

Emerging models of clinical services for men who have sex with men: focused versus comprehensive approaches.

By the early 1970s, it was increasingly recognised that men who have sex with men (MSM) were at risk for specific sexually transmissible infections (STI), and clinician awareness regarding MSM STI grew significantly after the AIDS epidemic was first recognised in 1981. In many urban centres in the USA and other resource-rich countries, the development of clinical infrastructure to address the AIDS epidemic led to the creation of clinics that provided services for large numbers of MSM. During the same time period, other health centres were created that were community-focused, providing comprehensive behavioural health and medical services for all sexual and gender minority patients. Over the next few years, multiple models for MSM sexual health will evolve, ranging from centres that embed STI care in primary care, to more focused centres that can use new technology to provide an efficient assessment for at-risk MSM desiring quick screening services.

Poorly Controlled HIV Infection: An Independent Risk Factor for Liver Fibrosis.

BACKGROUND: Liver disease is a major cause of mortality among HIV-infected persons. There is limited information about the extent to which HIV disease severity impacts liver disease progression. METHODS: We determined the incidence and predictors of advanced hepatic fibrosis measured by the Fibrosis-4 index (≥3.25) in a large diverse population of HIV-infected patients without significant liver disease at baseline (Fibrosis-4 score <1.45) in care between January 2000 and March 2014. We used Cox proportional hazards analysis to examine factors associated with progression to Fibrosis-4 score ≥3.25. RESULTS: Among 14,198 HIV-infected patients, hepatitis C virus (HCV) coinfection [adjusted hazard ratio (aHR) 1.9, 95% confidence interval (CI): 1.6 to 2.1], hepatitis B virus coinfection (aHR 1.5, 95% CI: 1.2 to 1.8), alcohol-use disorder (aHR 1.4, 95% CI: 1.2 to 1.6), and diabetes (aHR 1.9, 95% CI: 1.6 to 2.3) were associated with progression to advanced fibrosis in multivariable analysis. In addition, patients at each lower level of time-varying CD4 cell count had a significantly greater risk of progression, with ∼7-fold higher risk in those with CD4 <100 cells per cubic millimeter (aHR 6.9, 95% CI: 5.8 to 8.3) compared with CD4 ≥500 cells per cubic millimeter. An increasing gradient of risk was also observed among patients with higher time-varying HIV viral load (VL), with the greatest risk noted with VL ≥100,000 copies per milliliter (aHR 2.6, 95% CI: 2.2 to 3.1) compared with VL <500 copies per milliliter. CONCLUSIONS: Lower CD4 cell count and higher HIV VL were significantly associated with progression to advanced hepatic fibrosis in a dose-dependent manner, independent of the risk associated with traditional factors: hepatitis C virus or hepatitis B virus coinfection, alcohol, and diabetes. Our findings suggest that early treatment of HIV infection could mitigate liver disease.

HIV primary care providers--Screening, knowledge, attitudes and behaviors related to alcohol interventions.

BACKGROUND: Alcohol has particularly harmful health effects in HIV-infected patients; therefore, HIV clinics are an important setting for integration of brief alcohol intervention and alcohol pharmacotherapy to improve patient outcomes. Current practices of alcohol screening, counseling, and prescription of pharmacotherapy by HIV providers are unknown. METHODS: We conducted a cross-sectional survey of HIV providers from 8 HIV clinical sites across the United States. Surveys queried knowledge and use of alcohol screening, brief advice, counseling and pharmacotherapy, confidence and willingness to prescribe pharmacotherapy and barriers to their use of alcohol pharmacotherapy. We used multivariable logistic regression to examine provider factors associated with confidence and willingness to prescribe pharmacotherapy. RESULTS: Providers (N=158) were predominantly female (58%) and Caucasian (73%); almost half were infectious disease physicians and 31% had been in practice 10-20 years. Most providers (95%) reported always or usually screening for alcohol use, although only 10% reported using a formal screening tool. Over two-thirds never or rarely treated alcohol-dependent patients with pharmacotherapy themselves. Most (71%) referred alcohol-dependent patients for treatment. Knowledge regarding alcohol pharmacotherapy was low. The major barrier to prescribing pharmacotherapy was insufficient training on use of pharmacotherapy. Provider confidence ratings were positively correlated with their practice patterns. CONCLUSIONS: HIV providers reported high rates of screening for alcohol use, though few used a formal screening tool. Most providers referred alcohol dependent patients to outside resources for treatment. Few reported prescribing alcohol pharmacotherapy. Increased training on alcohol pharmacotherapy may increase confidence in prescribing and use of these medications in HIV care settings.

When to Monitor CD4 Cell Count and HIV RNA to Reduce Mortality and AIDS-Defining Illness in Virologically Suppressed HIV-Positive Persons on Antiretroviral Therapy in High-Income Countries: A Prospective Observational Study.

OBJECTIVE: To illustrate an approach to compare CD4 cell count and HIV-RNA monitoring strategies in HIV-positive individuals on antiretroviral therapy (ART). DESIGN: Prospective studies of HIV-positive individuals in Europe and the USA in the HIV-CAUSAL Collaboration and The Center for AIDS Research Network of Integrated Clinical Systems. METHODS: Antiretroviral-naive individuals who initiated ART and became virologically suppressed within 12 months were followed from the date of suppression. We compared 3 CD4 cell count and HIV-RNA monitoring strategies: once every (1) 3 ± 1 months, (2) 6 ± 1 months, and (3) 9-12 ± 1 months. We used inverse-probability weighted models to compare these strategies with respect to clinical, immunologic, and virologic outcomes. RESULTS: In 39,029 eligible individuals, there were 265 deaths and 690 AIDS-defining illnesses or deaths. Compared with the 3-month strategy, the mortality hazard ratios (95% CIs) were 0.86 (0.42 to 1.78) for the 6 months and 0.82 (0.46 to 1.47) for the 9-12 month strategy. The respective 18-month risk ratios (95% CIs) of virologic failure (RNA >200) were 0.74 (0.46 to 1.19) and 2.35 (1.56 to 3.54) and 18-month mean CD4 differences (95% CIs) were -5.3 (-18.6 to 7.9) and -31.7 (-52.0 to -11.3). The estimates for the 2-year risk of AIDS-defining illness or death were similar across strategies. CONCLUSIONS: Our findings suggest that monitoring frequency of virologically suppressed individuals can be decreased from every 3 months to every 6, 9, or 12 months with respect to clinical outcomes. Because effects of different monitoring strategies could take years to materialize, longer follow-up is needed to fully evaluate this question.

Using observational data to emulate a randomized trial of dynamic treatment-switching strategies: an application to antiretroviral therapy.

Background: When a clinical treatment fails or shows suboptimal results, the question of when to switch to another treatment arises. Treatment switching strategies are often dynamic because the time of switching depends on the evolution of an individual's time-varying covariates. Dynamic strategies can be directly compared in randomized trials. For example, HIV-infected individuals receiving antiretroviral therapy could be randomized to switching therapy within 90 days of HIV-1 RNA crossing above a threshold of either 400 copies/ml (tight-control strategy) or 1000 copies/ml (loose-control strategy). Methods: We review an approach to emulate a randomized trial of dynamic switching strategies using observational data from the Antiretroviral Therapy Cohort Collaboration, the Centers for AIDS Research Network of Integrated Clinical Systems and the HIV-CAUSAL Collaboration. We estimated the comparative effect of tight-control vs. loose-control strategies on death and AIDS or death via inverse-probability weighting. Results: Of 43 803 individuals who initiated an eligible antiretroviral therapy regimen in 2002 or later, 2001 met the baseline inclusion criteria for the mortality analysis and 1641 for the AIDS or death analysis. There were 21 deaths and 33 AIDS or death events in the tight-control group, and 28 deaths and 41 AIDS or death events in the loose-control group. Compared with tight control, the adjusted hazard ratios (95% confidence interval) for loose control were 1.10 (0.73, 1.66) for death, and 1.04 (0.86, 1.27) for AIDS or death. Conclusions: Although our effective sample sizes were small and our estimates imprecise, the described methodological approach can serve as an example for future analyses.

Cigarette Smokers are Less Likely to Have Undetectable Viral Loads: Results From Four HIV Clinics.

BACKGROUND: The prevalence of smoking among HIV-infected individuals is 2-3 times that of the general population, increasing the risk of smoking-related morbidity and mortality. We examined characteristics associated with smoking behavior among a large cohort of HIV-infected individuals in care in the United States. METHODS: A convenience sample of 2952 HIV-infected patients in the Centers for AIDS Research (CFAR) Network of Integrated Clinical Systems (CNICS) was assessed during routine clinic visits and was included. Multinomial logistic regression was used to examine the relationship between smoking status, depression/panic symptoms, alcohol/substance use, and demographic and clinical characteristics. RESULTS: Compared with never-smokers, current smokers were more likely to have moderate to severe depression (odds ratio [OR] 1.37), endorse current substance use (OR 14.09), and less likely to report low-risk alcohol use on the Alcohol Use Disorders Identification Test (AUDIT-C) (OR 0.73). Current smokers were less likely to have an undetectable viral load (OR 0.75), and more likely to have current substance abuse (OR 2.81) and moderate to severe depression (OR 1.50), relative to smokers who had quit smoking. CONCLUSIONS: HIV-infected smokers are less likely to have undetectable viral loads and frequently have psychosocial comorbidities including depression and substance abuse that impact antiretroviral therapy adherence and viral load suppression. To be effective, smoking-cessation interventions need to address the complex underlying concurrent risks in this population.

Multiple Imputation to Account for Measurement Error in Marginal Structural Models.

BACKGROUND: Marginal structural models are an important tool for observational studies. These models typically assume that variables are measured without error. We describe a method to account for differential and nondifferential measurement error in a marginal structural model. METHODS: We illustrate the method estimating the joint effects of antiretroviral therapy initiation and current smoking on all-cause mortality in a United States cohort of 12,290 patients with HIV followed for up to 5 years between 1998 and 2011. Smoking status was likely measured with error, but a subset of 3,686 patients who reported smoking status on separate questionnaires composed an internal validation subgroup. We compared a standard joint marginal structural model fit using inverse probability weights to a model that also accounted for misclassification of smoking status using multiple imputation. RESULTS: In the standard analysis, current smoking was not associated with increased risk of mortality. After accounting for misclassification, current smoking without therapy was associated with increased mortality (hazard ratio [HR]: 1.2 [95% confidence interval [CI] = 0.6, 2.3]). The HR for current smoking and therapy [0.4 (95% CI = 0.2, 0.7)] was similar to the HR for no smoking and therapy (0.4; 95% CI = 0.2, 0.6). CONCLUSIONS: Multiple imputation can be used to account for measurement error in concert with methods for causal inference to strengthen results from observational studies.