Journal of Family Medicine and Primary Care

: 2017  |  Volume : 6  |  Issue : 3  |  Page : 697-

Sample size for cluster sampling

Barbara Olenka Sanchez-Palomino1, Andrea Celi-Villacorta1, Laura Cecilia Gómez-Arrambide1, German F Alvarado2,  
1 Escuela de Terapia Física, Universidad Peruana de Ciencias Aplicadas, Lima, Peru
2 Escuela de Medicina, Universidad Peruana de Ciencias Aplicadas, Lima, Peru

Correspondence Address:
Dr. Barbara Olenka Sanchez-Palomino
Escuela de Terapia Fisica, Universidad Peruana de Ciencias Aplicadas, Lima

How to cite this article:
Sanchez-Palomino BO, Celi-Villacorta A, Gómez-Arrambide LC, Alvarado GF. Sample size for cluster sampling.J Family Med Prim Care 2017;6:697-697

How to cite this URL:
Sanchez-Palomino BO, Celi-Villacorta A, Gómez-Arrambide LC, Alvarado GF. Sample size for cluster sampling. J Family Med Prim Care [serial online] 2017 [cited 2020 Jun 5 ];6:697-697
Available from:

Full Text

Dear Editor,

We read with interest the article published by Ganjiwale et al.[1] To study the association of quality of life of caregivers and the degree of functional independence and the quality of life of their patients with a history of stroke, is a very important issue, especially in specific non-Western populations, where there has been little research.[1] It is known that a lot of people with stroke live with certain functional limitations in their activities of daily living,[2] so they need a caretaker, who suffers physical and mental stress.[3] However, we have some concerns regarding the study design.

The parameters considered for sample size calculation are incomplete; also, did the researchers calculate the power? was the “design effect” taken into account given the fact that the observations are not independent, but correlated? The patient-caregiver pair could be considered as a cluster; the “design effect” is directly proportional to the Intraclass Correlation Coefficient (ICC);[4] thus, if ICC and “design effect” were not taken into account, the study may not have adequate power.[5] Furthermore, power calculation has been made for a difference in proportions, however, the analysis was based on the correlation. In the methods section, data analysis plan was not shown; was the normality assumption evaluated in order to apply Pearson correlation? also, it is not known if Cronbach's alpha of the scales was computed, and if a factorial analysis was performed.

Finally, we found certain discrepancies between the main aim and the conclusions mentioned in the abstract. Definitely, this issue deserves further discussion and studies with more explicit designs.

Financial support and sponsorship


Conflicts of interest

There are no conflicts of interest.


1Ganjiwale D, Ganjiwale J, Parikh S. Association of quality of life of carers with quality of life and functional independence of stroke survivors. J Family Med Prim Care 2016;5:129-33.
2Knoflach M, Matosevic B, Rücker M, Furtner M, Mair A, Wille G, et al. Functional recovery after ischemic stroke – A matter of age: Data from the Austrian stroke unit registry. Neurology 2012;78:279-85.
3Jeong YG, Jeong YJ, Kim WC, Kim JS. The mediating effect of caregiver burden on the caregivers' quality of life. J Phys Ther Sci 2015;27:1543-7.
4Killip S, Mahfoud Z, Pearce K. What is an intracluster correlation coefficient? Crucial concepts for primary care researchers. Ann Fam Med 2004;2:204-8.
5Knox SA, Chondros P. Observed intra-cluster correlation coefficients in a cluster survey sample of patient encounters in general practice in Australia. BMC Med Res Methodol 2004;4:30.