how to report drop outs while reporting the results of clinical research trials?
statistically and as well as ethically?
what are the impacts of drop outs on the results of clinical research trials?
We have a simple standard practice. Do two analyses. One with all observations. Second excluding drop outs. Of course this will work only for observations taken before people drop out. If both are in agreement, there is no problem.
If endpoint is survival, then drop outs give you censored data and you should use methods suitable for censored data.
In some cases, it is possible to predict final outcome based on information available up to the time of dropping out. Then you can use predicted outcome.
As far as the ethical angle is concerned, I do not know what the regulatory expectation is; however I feel that anything is ok as long as you spell it out clearly up front. In other words, users of your results can judge whether the approach you have taken is acceptable or not.
What is clearly unethical (and dangerous) is to change the rules of the game midstream.
Read the story of Paxil by GSK. They changed the endpoint. Got approval, sold the drug and earned huge profits and were fined $3 Billion in 2012.
There is one simple golden rule with regard to dropouts in clinical studies. All dropouts have to be accounted for and reported. The reasons for dropouts have to be mentioned. If the reason is adverse event, it has to be included in safety analysis.
The previous experts have already outlined the implications which dropouts have on data management and statistical analysis.
For estimating intervention effects in a randomized controlled trial (RCT) there are three main ways: Per-Protocol Analysis restricted to the participants who fulfil the protocol in the terms of the eligibility, interventions, and outcome assessment, As-Treated Analysis comparing the subjects with the treatment regimen that they received not considering which treatment they were assigned and Intention-to-Treat analysis (ITT), evaluation of the treatment groups that includes all patients as originally allocated after randomization. The Consolidated Standards of Reporting Trials, the Cochrane Collaboration, the US Food and Drug Administration, the Nordic Council on Medicine in Europe, and the American Statistical Associations Group have recommended ITT as the way to analyze RCT data.To improve the applicability of study results to individual patients, investigators should improve study design to ensure protocol adherence with minimal loss to follow-up, because this loss at the final recall can result in exactly the same sort of bias as a Per- Protocol Analysis.
As mentioned in previous answers, including both Intention to Treat and Per Protocol analyses is the standard way of accounting for participant attrition (for whatever reason) in clinical trial results. This should be explicitly stated and planned for in the analysis plan that you draw up in the planning phase of your clinical trial, and it would be a good idea to consult with a statistician experienced in clinical trial analyses before you start recruiting, as this will also influence the numbers of patients that you will need to recruit to have sufficient numbers to get reliable and high quality data.
- Badges
- Science topic
- Similar topics
- Correction