Missing Data Imputation in Two Phase III Trials Treating HIV1 Infection

Virologic and immunologic effectiveness of darunavir-based salvage therapy in HIV-1-infected adults in a Brazilian clinical practice setting: results of a multicenter and retrospective cohort study

Background

Darunavir has been proven efficacious for antiretroviral-experienced HIV-1-infected patients in randomized trials. However, effectiveness of darunavir-based salvage therapy is understudied in routine care in Brazil.

Methods

Retrospective cohort study of HIV-1-infected patients from three public referral centers in Belo Horizonte, who received a darunavir-based therapy between 2008 and 2010, after virologic failure. Primary endpoint was the proportion of patients with viral load < 50 copies/mL at week 48. Change in CD4 cell count was also evaluated. Outcome measures were analyzed on an intent-to-treat basis applied to observational studies. Sensitivity analysis was conducted to evaluate the impact of missing data at week 48. Predictors of virologic failure were examined using rare-event, finite sample, bias-corrected logistic regression.

Results

Among 108 patients, the median age was 44.2 years, and 72.2% were male. They had long-standing HIV-1 infection (median 11.6 years) and advanced disease (76.9% had an AIDS-defining event). All patients had previously received protease inhibitors and nucleoside reverse transcriptase inhibitors, 75% nonnucleoside reverse transcriptase inhibitors, and 4.6% enfuvirtide. The median length of protease inhibitor use was 8.9 years, and 90.8% of patients had prior exposure to unboosted protease inhibitor. Genotypic resistance profile showed a median of three primary protease inhibitor mutations and 10.2% had three or more darunavir resistance-associated mutations. Virologic success at week 48 was achieved by 78.7% (95% CI = 69.7–86%) of patients and mean CD4 cell count increase from baseline was 131.5 cells/μL (95% CI = 103.4–159.6). In multiple logistic regression analysis, higher baseline viral load (RR = 1.04 per 10,000 copies/mL increase; 95% CI = 1.01–1.09) and higher number of darunavir resistance-associated mutations (RR = 1.23 per each; 95% CI = 0.95–1.48) were independently associated with virologic failure.

Conclusion

Virologic suppression is a realistic endpoint for most treatment-experienced patients who begin a darunavir-based therapy outside the controlled conditions of a randomized trial, at routine care settings.

Virologic and Immunologic Effectiveness at 48 Weeks of Darunavir–Ritonavir-Based Regimens in Treatment-Experienced Persons Living with HIV-1 Infection in Clinical Practice

Introduction:

Published data addressing the effectiveness of darunavir–ritonavir (DRV/r)-based therapy for multiexperienced patients in developing countries are scarce. This study evaluated the 48-week virologic and immunologic effectiveness of salvage therapy based on DRV/r for the treatment of multidrug-experienced HIV-1-infected adults in Brazil.

Materials and Methods:

A multicenter retrospective cohort study was carried out with multidrug-experienced adults who were on a failing antiretroviral therapy and started a DRV/r-based salvage therapy between 2008 and 2010. The primary effectiveness end point was the proportion of patients with virologic success (plasma HIV-1 RNA <50 copies/mL at week 48).

Results:

At 48 weeks, 73% of the patients had HIV-RNA <50 copies/mL and a mean increase of 108 CD4 cells/mm3. Higher baseline viral load, lower baseline CD4 count, younger age, and 3 or more DRV/r-associated resistance mutations were significantly predictive of virologic failure. Concomitant use of raltegravir was strongly associated with virologic success.

Conclusion:

The use of DRV/r-based regimens for salvage therapy is an effective strategy in the clinical care setting of a developing country.

A web-based intervention to improve and prevent low back pain among office workers: a randomized controlled trial

STUDY DESIGN

Randomized controlled trial.

OBJECTIVES

To test the feasibility, safety, and efficacy of a web-based multidisciplinary intervention for office workers with subacute, nonspecific low back pain.

BACKGROUND

Low back pain is one of the most frequent ailments seen in primary care consultations.

METHODS

The trial included 100 office workers with subacute low back pain. The intervention group had access to both the study intervention and standard care. The control group had access to standard care only. Standard care was defined as all existing non-web-based interventions offered by the University of Extremadura's Preventive Medicine Service. The web-based program was offered via the Preventive Medicine Service website. The participants in the intervention group were asked to engage in the web-based program at their work site for 11 minutes each day, 5 days a week. Primary outcomes were functional disability, as measured by the Roland-Morris Disability Questionnaire, and health-related quality of life, as measured by the European Quality of Life-5 Dimensions-3 Levels. Secondary outcomes were the number of episodes of low back pain and trunk muscle endurance. Outcomes were measured before and after the 9-month intervention period.

RESULTS

Over the 9-month study, the score on the Roland-Morris Disability Questionnaire for the participants in the web-based intervention group improved by a mean of -7.36 points (95% confidence interval [CI]: -8.41, -6.31) compared to a worsening of 1.89 points (95% CI: 0.71, 2.65) in the control group. The between-group difference in change on the Roland-Morris Disability Questionnaire over the study period was -9.25 points (95% CI: -10.57, -7.89). Similarly, over the 9-month study, the intervention group had a significant improvement in quality of life of 0.24 points (95% CI: 0.20, 0.29) compared to the control group.

CONCLUSION

A 9-month web-based intervention is feasible and effective to improve function and health-related quality of life and to decrease episodes of low back pain among office workers with a history of subacute, nonspecific low back pain.

Predicting missing quality of life data that were later recovered: an empirical comparison of approaches

Background and Purpose The aim was to compare simple imputation, multiple imputation, and modeling approaches to deal with ‘missing’ quality of life data. Data were obtained from five clinical trials, which employed a reminder system for follow-up questionnaires. Previous studies have compared imputation strategies by artificially removing data according to prespecified mechanisms. Our approach differs from previous study as actual collected data are utilized.

Methods Data obtained by reminder were initially treated as missing. These missing values were imputed using a variety of simple and multiple imputation strategies. The trials were analyzed using the imputed datasets, and the resulting treatment effects compared to analyses using the full dataset including responses following reminders. A repeated measures model was also carried out on the available data and the pattern mixture models were employed. The accuracy of the different strategies was assessed by calculating the bias seen in the calculated treatment difference compared to the actual observed treatment difference.

Results Baseline carried forward or last value carried forward were shown to be the best simple imputation methods in this setting. Multiple imputation using a regression model or predictive mean match model tended to provide treatment difference estimates with the least bias when compared to the actual observed data. Pattern mixture models did not perform well. Overall, the multiple imputation procedures were generally the least biased approaches.

Limitations A number of imputation and modeling procedures have been investigated but this list is not exhaustive. All the example datasets come from the same data source and perhaps studies from additional disease areas would have been useful. However, we feel the results are generalizable to other quality of life outcomes and clinical areas.

Conclusions Multiple imputation is recommended for missing quality of life data as it makes the assumption of missing at random which in the quality of life setting is more plausible than the assumption of missing completely at random for which most simple imputation methods are based. Pattern mixture models can be complex and did not perform well in this setting. Clinical Trials 2010; 7: 333—342. http://ctj.sagepub.com

From trial and error to trial simulation. Part 1: the importance of model-based drug development for antidepressant drugs

Clinical trial simulation (CTS) allows the investigation of the influence of design characteristics on important aspects of clinical trials such as power and type I error. Simulation scenarios may be critical to decision making and prevention of study failure. The analysis and simulation of clinical trials in depression have, however, suffered from a lack of disease/dropout models that describe the individual time course of the clinical scale of interest. We propose a new model with dual random effects (DREM) derived from functional data analysis, which provides unbiased estimates of parameters and is suitable for the purposes of clinical trial simulation. A comparison of model performance is presented, along with standard statistical methods using various goodness-of-fit criteria. Our results show that data simulated using the DREM closely match individual patient data, including real-life dropout scenarios. In addition, parameterization in terms of interindividual variability ensures easier explanation of findings to clinical scientists, who ultimately make the relevant decisions.