Multivitamins in the Prevention of Cancer in Men, Article Review Example

Article Review of the Physicians Health Study II

The Physicians Health Study II, a randomized controlled trial, deals with the role of multivitamins in the prevention of cancer in men.  Its objective was “to determine whether long-term multivitamin supplementation decreases the risk of total and site-specific cancer events among men”(Gaziano, J. M. et al, 2012).

What is the method used to identify and assign subjects to study and control groups?

According to the write-up of the study, it is a randomized study, which means that from any defined population participants to both the treatment and the control groups are chosen by chance alone. Suppose one wanted to do a survey (no treatment or control groups involved in this example) of the satisfaction level of clients of a 500 person capacity bingo hall.  A decision would be made beforehand as to the evening, the time, and the duration during which would be handed out a very brief questionnaire.  That would be completely random since the conductors of the survey would have no idea as to who would walk through the doorway at that time. (To ensure maximum return of questionnaires they might tell the participants that someone will be around within 15 minutes to pick up the completed form.)  When a study has a treatment group and a control group, the randomized subjects in both groups should be matched as equally as possible on important variables.  In this study, the physicians were categorized into five year age groups, beginning at age 50.  Therefore, as equal as possible a number of participants in the 50 – 55 age group should be in the treatment group and in the control group; and, then, the same should hold true for the 55 – 60 group and so forth.

What are the confounding variables?

However, even when the treatment and the control groups are randomized and then   matched as equally as possible  on all variables pertinent to the study, confounding variables  can still be present. A confounding variable is an extraneous one that correlates with both the independent and the dependent variables. If uncontrolled, they can hurt the internal validity of a study. An example of how a confounding variable can affect study results might be that the occurrence of more rape in the summer correlates positively with the consumption of more ice cream in the summer. In this classical example, the confounding variable is summer (warm temperatures).  In this comprehensive, well-planned and well-supervised study, exercise, a healthy diet, sufficient sleep, low stress level, no family deaths, etcetera may have correlated positively with multivitamins—and even placebos– decreasing the incidence of cancer in men.

Making and blinding: are either (or both) the participants or the researchers aware of who is assigned to what group.

This was a double blind study, which means that neither the researchers nor the subjects were aware as to who was in the experimental group (taking multivitamins) or in the control group taking plac ebos.  This is done to avoid anyone having any expectations of either group.  If, for example, participants of the treatment group got the impression that they were the placebo group, some of them might not take their vitamins as prescribed and, consequently, jeopardize the results of the study. Masking/binding was achieved in this study. Blinded treatment and follow-up continued through June 01, 2011, the scheduled end of the multivitamin phase II” (ibid.).

What were the results of the study?

Men taking multivitamins had a modest cancer reduction in epithelial cell cancer, but none in prostrate cancer, unless there was no parental history of cancer. They had an HR level of only 0.86 level with a significance of p=.02.

Age was a factor in that men 70+ had an HR of 82. However, the test for heterogeneity across age groups for men 50 to 69 was not significant.

Where there any secondary outcomes? If yes, what were they and why did the researchers discuss them in the article?

Some secondary outcomes include eye disease, cognitive function and a number of prespecified secondary analyses.

Who funded this study?

This study was funded by grants CA097193, CA34944, CA40360, HL26490 and HL34595 from the NIH and an investigator-initiated grant from the BASE Corporation.

Study agents and packaging were provided by BASE Corporation and Pfizer.  Study packaging was provided by DSM Nutritional Products, Inc.

There seems to be no bias in the design and the carrying out of this study.  It appears to be legit. There is no indication anywhere that any of the researchers were leaning towards certain desired outcomes. They were interested in finding out the facts, nothing more, nothing less.

If the primary hypothesis was proven was it proven in a legit way?

The primary hypothesis, which was to “determine  whether long term vitamin supplementation decreases the risk of total and site-specific cancer events among men”(ibid.).  was proven with a probability value of .04, which is lower than necessary since the probability for the study had been set at p.= .05 or less.  Probability values for studies are usually set at p.=.05 or less; and if absolute faith in the results is essential, they can be set as low as p.=01.

Yes, there was a small loss to follow-up: To treatment group—48; placebo group:  57. The larger the loss to follow-up, the greater the negative impact on the study. Loss to follow-up occurs for various reasons—no longer interested, no forwarding address, lost in the computer.

All cancer and mortality end points were checked and validated by medical review by the PHS End Points Committee composed of physicians blinded by treatment assignment.  The data on the end points was statistically significant using as the analysis the SAS 9-2 and S-Plus and  the statistical significance set at p<.05 using two-sided tests.

References

Gaziano, J. M., Sesso, H.D.,  Christen, W. G., Bubes, V., Smith, J. P., MacFaydan, J., Schvartz, M., Manson, J. E., Glynn, R. J., and Buring, J. F. (2012). Multivitamins in The Prevention of Cancer in Men.  The Physsicians’ Health Study II Randomized Controlled trial.  American Medical Association.