Multivitamin and CVD Apprasial 2016 (1) (4).docx

***Among men over the age of 50, does daily supplementation with multivitamin reduce the incidence of major CVD (non-fatal MI, stoke, CV death) over 10 years compared to placebo therapy?*** ***Did intervention and control groups start with the same prognosis?*** - **Were patients randomized?**...

***Among men over the age of 50, does daily supplementation with multivitamin reduce the incidence of major CVD (non-fatal MI, stoke, CV death) over 10 years compared to placebo therapy?*** ***Did intervention and control groups start with the same prognosis?*** - **Were patients randomized?** - **Was group allocation concealed?** - **Were patients in the study groups similar with respect to known prognostic variables?** Yes - randomly assigned using blocks of 16; stratified on age, history of cancer and CVD, and beta-carotene use for \~ half of participants No - there is no mention of group allocation concealment. Although it is likely, the authors do not provide any information on his critical step in the randomization process. Yes - Table 1 would suggest the groups were similar especially for key variables like age and history of CVD that would be very impt in this particular study. Overall, I believe they started with the same prognosis. Although there is no mention of allocation concealment, randomization appears to have worked based on table 1 with no major clinical differences noted. Teaching Note: some of you may have noticed a "run in phase' was used whereby 11 128 potential physicians entered a run in phase but only 7000 (63%) were adherent to their pills and were randomized into the PHS II. Run-in phases are very common and are often centered on adherence. These run-in phase generally assist the ability of the study to find a difference (i.e., power) by maximizing your signal (keep only people who are going to take their intervention properly) and reduces heterogeneity (reduces noise by getting rid of non-adherent people but also adherent people also tend to have other lifestyle factors in alignment -- for example, most of the group tended to exercise 1 day a week!). The downside of run-in phases is that it severely reduces the generalizability or the external validity of studies. People who get into the study are not necessarily the same as the general population (i.e., only the people who are adherent). The reduction in generalizability is the major problem with run-in phases. ***Was prognostic balance maintained as the study progressed?*** - **To what extent was the study blinded?** The authors indicate that patients were blinded through the use of calendar packs. However, there is no mention of how this blinding was completed (most likely similar colored vitamin and placebo pills but we are unsure). It is likely they maintained the blinding as the number of people who were adherent at study end were about the same suggesting that the placebo groups did not figure out what they were taking (as why would you take a placebo for 10 years!). Adherence is often a very good proxy for determining whether people maintained blinding in a study. Moreover, all endpoint evaluators were also blinded to treatment assignment. With respect to caregivers, there is no mention if they were blinded or not; however, in this trial, there was minimal interaction between the caregiver and the patients (as intervention provided remotely via the mail) and so the blinding of these people may have been less important....but it is difficult to tell. Similarly, here was no mention of whether the analysts were blinded. Overall, I believe that blinding was likely sufficient. Although there may have been some issues with the caregiver and analyst, I think the effects would be minimal. It would have been more appropriate to indicate exactly how the blinding of the patients was completed but the major outcome assessors were blinded. ***Were the groups prognostically balanced at the study\'s completion?*** - **Was follow-up complete?** - **Were patients analyzed in the groups to which they were first allocated?** - **Was the trial stopped early?** Follow-up in terms of total time was sufficient to see the impact of a multivitamin on an endpoint like CVD in males \~15 years (avg age 63). We would expect to see treatment differences over a 10-15 year span. The authors also did everything they could to ensure all outcomes were collected in the follow-up period through web searches, medical record and autopsy reviews. Moreover, these assessments appear to be identical between the study arms - that is an important aspect. They indicate \