Performing multi stage random sampling in community based surveys
Department of Community Medicine, Dr. Rajendra Prasad Government Medical College, Kangra, Himachal Pradesh, India
Dr. S K Raina
Department of Community Medicine, Dr. Rajendra Prasad Government Medical College, Kangra, Himachal Pradesh
|How to cite this article:|
Raina S K. Performing multi stage random sampling in community based surveys.J Postgrad Med 2014;60:221-222
|How to cite this URL:|
Raina S K. Performing multi stage random sampling in community based surveys. J Postgrad Med [serial online] 2014 [cited 2020 Aug 4 ];60:221-222
Available from: http://www.jpgmonline.com/text.asp?2014/60/2/221/132385
This is in reference to the article entitled "Patient-physician trust among adults of rural Tamil Nadu: A community-based survey" published in J Postgrad Med (2014;60:21-6).  The authors rightly point to the fact that the patient's trust has been associated with patient-reported satisfaction with their care, intent to retain or switch physicians, willingness to recommend a physician to others, and treatment adherence.  The authors attempt to identifying the level of trust therefore is laudable. However, I have a concern with the methodology used to arrive at this aim, in the study. The authors mention that a multistage random sampling method was used for the purpose of this study (material and methods).  However a look at the details of the methods reveals that in actual, multistage random sampling method was not used.
For multistage random sampling method, each stage of the multistage process is itself a sampling technique, each stage must be held to the standards as if it were the only stage. Therefore, selection at all stages must be random. However the authors have resorted to convenience sampling at the last of their multistage sample. Using convenience sampling at last stage is expected to introduce a bias in this study, which defeats the very purpose of randomizing at earlier stages. If one stage is done poorly, then this has ramifications for the rest of the data collection as well as for the statistical analyses conducted further along in the research process.
True, as there is no strict definition to multistage random sampling, there is no formulaic way as to how to combine the various sampling options (such as clustering, stratified, and simple random). The multistage sampling procedure should therefore be constructed in such a way to be cost and time effective while retaining both the randomness and sufficient size of the sample. So the key to a good multistage random sampling method will be the random selection of samples at all stages.
Furthermore, there appears some sort of contradiction among the authors in choosing the sample population for this study. The authors state that this cross-sectional household survey included men and women above 40 years of age who may or may not be suffering from any illness during the time of study but have visited a physician at least once during the past 5 years. Then in the second sentence state that since healthy people, who are not in a state of vulnerability in terms of health, are more likely to be capable of thinking about trust in a conscious, calculative fashion they were chosen for the study. So the question arises, as to who were chose, healthy, ill or both.
Also labeling scores as high trust and low trust for the purpose of analysis does not seem justifiable as the authors clearly point out themselves that "high trust" and "low trust" do not reflect the levels of trust in the community.
|1||Baidya M, Gopichandran V, Kosalram K. Patient-physician trust among adults of rural Tamil Nadu: A community-based survey. J Postgrad Med 2014;60:21-6.|