Comments on: The effect of an educational program on quality of life in patients undergoing hemodialysis in Western Saudi Arabia

To the Editor

I have read with interest the study of Bakarman et al,1 on the effect of an educational program on quality of life in patients undergoing hemodialysis in Western Saudi Arabia. I found the subject very important, however, I want to share some of my opinions especially on the methodology of the study. They said that they used the Short Form-36 (SF-36) but they did not give any reference about the manuscript on the validation of Arabic version of SF-36. Similarly, they described the necessary parameters for sample size calculation but they did not give any reference for the program if they did not calculate it by using the formulas by themselves. I realize that the calculation was made by using G*Power, if so, the reference should be written as the researchers, Faul F et al² requested from all who use G*Power. For the sample size calculation, I believe that there is also one more issue. Might the calculation be made post hoc? Because it is not common to target to achieve a 84% of power. I believe this is a post hoc power for the determined parameters, for which I calculated “0.8421493” by G*Power.^3,4 If I had been the reviewer of the study I had offer to use the CONSORT checklist which might be very useful for the presentation of the study. For example, I cannot understand the total of patients included in the study, how many were excluded for which reason, etc. Without giving this information, the reader can think that the researcher determined the criteria and included the first 50 to the intervention group and 50 to the control group. The design may be so, but it should be written as what it is performed. There are few issues as well, lack of information mostly, also on other presentations: They said that “Paired t test was used and the results are presented in Table 2.”, but on the bottom of the Table 2 it is written as “* Independent t-test, † Significance level at p<0.05”. There was no information on the standardization of the trainings of the trainers-an important bias source, routine information given to control group, etc.

Because of all of these methodological problems, I presume that these results should be cautiously interpreted.

Reply from the Author

We really appreciate valuable feedback from the reviewer. Reviewer highlighted some queries.

Sample size

We have not used G.Power for sample size calculations, we use PASS for calculation of sample size. Details can be verified form screenshot of the program (Figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Power and Sample Size Estimation details.