Example 161 Natalizumab plus interferon beta1a for relapsing multiple sclerosis the Sentinel study

Dr Garys MS Treatment System

Multiple Sclerosis Homeopathic Cure

Get Instant Access

Figure 16.1 provides the participant flow in relation to this placebo-controlled trial of natalizumab. (Rudick et al. (2006)).

Figure 16.1 Patient disposition in the SENTINEL trial. (Rudick RA Stuart WH, Calabresi PA, Confavreux C et ai. for the SENTINEL Investigators (2006) 'Natalizumab plus interferon beta-1a for relapsing multiple sclerosis' New England Journal of Medicine, 354, 911-923. © (2006) Massachusetts Medical Society.)

Figure 16.1 Patient disposition in the SENTINEL trial. (Rudick RA Stuart WH, Calabresi PA, Confavreux C et ai. for the SENTINEL Investigators (2006) 'Natalizumab plus interferon beta-1a for relapsing multiple sclerosis' New England Journal of Medicine, 354, 911-923. © (2006) Massachusetts Medical Society.)

The quality of publications is certainly increasing and in part this is due to guidance of the type already described in this section. It is unfortunately the case, however, that many mistakes are still made, even in leading journals, despite apparently rigorous refereeing procedures. Particular areas of statistics seem to cause consistent difficulty:

• The correct design, analysis and interpretation of non-inferiority trials -remember conventional p-values have no role

• Conforming to the principle of intention-to-treat to avoid bias - ITT means all randomised subjects or something very close to that

• Incorrect analysis of time-to-event data in terms of the definition for the origin of the measurement - the point of randomisation is the only origin that can be used in a randomised trial

• Adjusting the analysis for baseline factors - with change from baseline as the outcome variable, include baseline as a covariate to avoid regression towards the mean and note that it is incorrect to use covariates that are measured after randomisation

• The correct statistical methods for combining data in meta-analysis -the summary statistics that are combined must come from independent data sets

It is important to enlist the help of statistical colleagues when putting publications together, not only in terms of the actual analysis, but in terms of the interpretation and reporting in the publication itself. Further, do cast a critical eye over the statistical methodology in the papers you review in order to spot these problem areas and again request the help of your statistical colleagues.

Presentations at conferences do not, of course, go through the same critical review process as publications. Even though abstracts are often submitted and reviewed in advance, mistakes and bad practice will still slip through. It is important to have statistical input when putting these presentations together. Errors in the statistics will invariably get picked up by some members of the audience and the resulting bad press could well be damaging. From the opposite perspective look critically at what is being presented in terms of the statistics and challenge if you feel that inappropriate methods are being used.

Was this article helpful?

0 0

Post a comment