Estimating the causal effects of treatment

Collaborative care for depression in older adults: How much is enough?

Collaborative care in primary care has been shown to be effective for subthreshold depression in older adults in the 'CASPER' trial. However, to understand the impact of adherence, and to explore the minimum effective dose of collaborative care, we reanalysed the trial data using a complier average causal effect (CACE) analysis. Data were available for 705 participants, 519 with 12-month PHQ-9 scores. 'Compliance' could be observed for participants in the intervention group. Latent complier status in the control group was estimated. Completion of five or more sessions of care was defined as 'compliance'. Sensitivity analyses, using alternative cut-offs of two to eight sessions, assessed the impact of changing the definition of 'compliance'. Compliers in the intervention group had lower PHQ-9 scores at 12-month follow up than assumed compliers in the control group (1.75 lower, 95% confidence interval 0.29 to 3.21, p = 0.02), a greater effect than originally reported. Sensitivity analyses confirmed statistically significant differences between the intervention and control groups in those attending five or more sessions. We conclude that collaborative care is causally effective in reducing subthreshold depressive symptoms in older people who adhere to treatment. Our findings suggest the minimum effective dose is five sessions.

Estimating treatment effects from randomized clinical trials with noncompliance and loss to follow-up: the role of instrumental variable methods

Perfectly implemented randomized clinical trials, particularly of complex interventions, are extremely rare. Almost always they are characterized by imperfect adherence to the randomly allocated treatment and variable amounts of missing outcome data. Here we start by describing a wide variety of examples and then introduce instrumental variable methods for the analysis of such trials. We concentrate mainly on situations in which compliance is all or nothing (either the patient receives the allocated treatment or they do not - in the latter case they may receive no treatment or a treatment other than the one allocated). The main purpose of the review is to illustrate the use of latent class (finite mixture) models, using maximum likelihood, for complier-average causal effect estimation under varying assumptions concerning the mechanism of the missing outcome data.

An IV for the RCT: using instrumental variables to adjust for treatment contamination in randomised controlled trials

Although the randomised controlled trial is the "gold standard" for studying the efficacy and safety of medical treatments, it is not necessarily free from bias. When patients do not follow the protocol for their assigned treatment, the resultant "treatment contamination" can produce misleading findings. The methods used historically to deal with this problem, the "as treated" and "per protocol" analysis techniques, are flawed and inaccurate. Intention to treat analysis is the solution most often used to analyse randomised controlled trials, but this approach ignores this issue of treatment contamination. Intention to treat analysis estimates the effect of recommending a treatment to study participants, not the effect of the treatment on those study participants who actually received it. In this article, we describe a simple yet rarely used analytical technique, the "contamination adjusted intention to treat analysis," which complements the intention to treat approach by producing a better estimate of the benefits and harms of receiving a treatment. This method uses the statistical technique of instrumental variable analysis to address contamination. We discuss the strengths and limitations of the current methods of addressing treatment contamination and the contamination adjusted intention to treat technique, provide examples of effective uses, and discuss how using estimates generated by contamination adjusted intention to treat analysis can improve clinical decision making and patient care.

The impact of the CONSORT statement on reporting of randomized clinical trials in psychiatry

To determine whether the CONSORT recommendations influenced the quality of reporting of randomized controlled trials (RCTs) in the field of psychiatry, we evaluated the quality of clinical trial reports before and after the introduction of CONSORT statement. We selected seven high impact journals and retrieved the randomized, clinical trials in the field of psychiatry during the period of 1992-1996 (pre-CONSORT) and 2002-2007 (post-CONSORT). Among the total 5201 articles screened, 736 were identified and entered in our database. After critical review of the publications, 442 articles met the inclusion and exclusion criteria. The CONSORT Index (sum of 22 items of the checklist) during the post-CONSORT period was significantly higher than that during the pre-CONSORT period. However, over 40% of post-CONSORT studies did not adhere to CONSORT statement for reporting the process of randomization, and details of the process for obtaining informed consent were still insufficient. Furthermore, adherence to the CONSORT guidelines of reporting how blinding was accomplished and evaluated actually decreased after publication of the CONSORT statement. Although the overall quality of reporting on psychiatric RCTs generally improved after publication of the CONSORT statement, reporting the details of randomization, blinding, and obtaining informed consent remain insufficient.

Improvement and decline of cognitive function in schizophrenia over one year: a longitudinal investigation using latent growth modelling

BACKGROUND

Long-term follow-up studies of people with schizophrenia report stability of cognitive performance; less is known about any shorter-term changes in cognitive function.

METHODS

This longitudinal study aimed to establish whether there was stability, improvement or decline in memory and executive functions over four assessments undertaken prospectively in one year. Cognitive performance was assessed during randomized controlled trials of first- and second-generation antipsychotic medication. Analyses used a latent growth modeling approach, so that individuals who missed some testing occasions could be included and trajectories of cognitive change explored despite missing data.

RESULTS

Over the year there was significant decline in spatial recognition but no change in pattern recognition or motor speed. Improvement was seen in planning and spatial working memory tasks; this may reflect improved strategy use with practice. There were significant individual differences in the initial level of performance on all tasks but not in rate of change; the latter may have been due to sample size limitations. Age, sex, premorbid IQ and drug class allocation explained significant variation in level of performance but could not predict change. Patients randomized to first-generation drugs improved more quickly than other groups on the planning task.

CONCLUSION

We conclude that cognitive change is present in schizophrenia but the magnitude of change is small when compared with the large differences in cognitive function that exist between patients. Analyses that retain patients who drop out of longitudinal studies, as well as those who complete testing protocols, are important to our understanding of cognition in schizophrenia.

Reporting in randomized clinical trials improved after adoption of the CONSORT statement

OBJECTIVE

To examine the extent to which the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines improved clinical trials reporting and subject attrition, which may undermine the credibility of published randomized clinical trials (RCTs).

STUDY DESIGN AND SETTING

Published RCTs reported in two major medical journals before and after the CONSORT guidelines were systematically reviewed; one used the CONSORT statement (JAMA) and one did not (NEJM).

RESULTS

The quality of RCT reporting improved for both journals, but JAMA showed more significant and consistent improvements in all aspects of RCT reporting. Subject attrition was better accounted for after the publication of CONSORT, although the attrition rates for various reasons actually increased. Attrition due to unknown reasons, as a percentage of total attrition, declined dramatically, from 68.7% pre-CONSORT to 13.0% post-CONSORT.

CONCLUSIONS

Attrition of study subjects remains a serious problem in RCTs. Bias from selective attrition can undermine the presumptive scientific advantage of RCTs. The CONSORT guidelines improved RCT reporting when they were implemented but did not substantially improve reported attrition rates.

Improving the quality of reporting randomized controlled trials in cardiothoracic surgery: the way forward

OBJECTIVE

To evaluate the quality of reporting of randomized controlled trials in cardiothoracic surgery, to identify factors associated with good reporting quality, and to assess the awareness of the Consolidated Standards for Reporting of Trials statement and ascertain the views of authors reporting randomized controlled trials on the difficulties in conducting randomized controlled trials and the possible ways to further improve the reporting quality of randomized controlled trials in cardiothoracic surgery.

METHODS

Randomized controlled trials of cardiothoracic surgery published in principal cardiothoracic and 4 general medical journals in 2003 were included. The quality of reporting of randomized controlled trials was assessed by using allocation concealment, the Jadad score, and a Consolidated Standards for Reporting of Trials checklist devised for the purpose. A questionnaire survey of authors reporting randomized controlled trials in principal cardiothoracic journals in 2003 was conducted.

RESULTS

The overall reporting quality of the 64 randomized controlled trials included in the analysis was suboptimal as assessed by the 3 methods adopted. Most of the authors (63.5%) were not aware of the Consolidated Standards for Reporting of Trials statement; however, awareness was not associated with reporting quality. More than 65% of the authors responded that conducting randomized controlled trials in surgical specialties was difficult, and the main difficulties were blinding and obtaining a large-enough sample size to detect statistically significant differences. Fifty-four percent of the authors responded that endorsement of the Consolidated Standards for Reporting of Trials statement by the cardiothoracic journals may improve the reporting quality.

CONCLUSIONS

The quality of reporting randomized controlled trials in cardiothoracic surgery is suboptimal. Endorsement of the Consolidated Standards for Reporting of Trials statement by the cardiothoracic journals may improve the quality of reporting.

Estimating psychological treatment effects from a randomised controlled trial with both non-compliance and loss to follow-up

BACKGROUND

The Outcomes of Depression International Network (ODIN) trial evaluated the effect of two psychological interventions for the treatment of depression in primary care. Only about half of the patients in the treatment arm complied with the offer of treatment, prompting the question:'what was the effect of treatment in those patients who actually received it?'

AIMS

To illustrate the estimation of the effect of receipt of treatment in a randomised controlled trial subject to non-compliance and loss to follow-up.

METHOD

We estimated the complier average causal effect (CACE) of treatment.

RESULTS

In the ODIN trial the effect of receipt of psychological intervention (an average of about 4 points on the Beck Depression Inventory) is about twice that of offering it.

CONCLUSIONS

The statistical analysis of the results of a clinical trial subject to non-compliance to allocated treatment is now reasonably straightforward through estimation of a CACE and investigators should be encouraged to present the results of analyses of this type as a routine component of a trial report.