Archived Comments for:
Knowledge, attitudes and practices with regard to the presence, transmission, impact, and control of cystic echinococcosis in Sidi Kacem Province, Morocco
Justification of sample size adequecy and representiveness is needed
Siddharudha Shivalli, Yenepoya Medical College, Yenepoya University, Mangaluru-575018, Karnataka, India
27 November 2015
I read the article by El Berbri I et al, with a great interest. Authors’ efforts are commendable. Based on a large scale community based survey, authors highlight the burden and socio-demographic disproportional distribution of Cystic echinococcosis (CE) in Sidi Kacem Province, Morocco. In addition, authors also emphasize the higher infection rate in slaughter animals and the main drivers of CE transmission in the study area.
Authors state that 543 community members in 39 douars (villages) across the 27 communes studied. And they have followed multistage random sampling. Authors should have justified the adequacy and representativeness of sample size, although, sample size appears good enough. This is essential to ensure the internal validity of the study findings. In my opinion, sample size should have been calculated based on the total population, anticipated community knowledge of CE, design effect (owing to multistage random sampling), power of the study, confidence levels and anticipated non-response. Same should have been done for the Abattoir study component also. In addition, authors should have explicitly stated the inclusion and exclusion criteria for the study participants.
It is recommended to follow the STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) checklist while reporting cross sectional studies. It is endorsed by many biomedical journals. In Table 4 of the article, authors have mentioned the p value as ‘0.000’. Statistical software, by default setting, displays p value as zero if it extends beyond 3 decimal points (i.e. p=0.0000001 would be displayed as p=0.00). Practically, the value of p cannot be zero and hence, I would suggest to report it as p<0.0001. In addition, median values for various continuous variables in Table 4 should have been reported with inter quartile range.
Nonetheless, I must congratulate the authors for exploring an important public health problem in the study area.
Competing interests
The author declares that there is no conflict of interest about this publication.
Justification of sample size adequecy and representiveness is needed
27 November 2015
I read the article by El Berbri I et al, with a great interest. Authors’ efforts are commendable. Based on a large scale community based survey, authors highlight the burden and socio-demographic disproportional distribution of Cystic echinococcosis (CE) in Sidi Kacem Province, Morocco. In addition, authors also emphasize the higher infection rate in slaughter animals and the main drivers of CE transmission in the study area.
Authors state that 543 community members in 39 douars (villages) across the 27 communes studied. And they have followed multistage random sampling. Authors should have justified the adequacy and representativeness of sample size, although, sample size appears good enough. This is essential to ensure the internal validity of the study findings. In my opinion, sample size should have been calculated based on the total population, anticipated community knowledge of CE, design effect (owing to multistage random sampling), power of the study, confidence levels and anticipated non-response. Same should have been done for the Abattoir study component also. In addition, authors should have explicitly stated the inclusion and exclusion criteria for the study participants.
It is recommended to follow the STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) checklist while reporting cross sectional studies. It is endorsed by many biomedical journals. In Table 4 of the article, authors have mentioned the p value as ‘0.000’. Statistical software, by default setting, displays p value as zero if it extends beyond 3 decimal points (i.e. p=0.0000001 would be displayed as p=0.00). Practically, the value of p cannot be zero and hence, I would suggest to report it as p<0.0001. In addition, median values for various continuous variables in Table 4 should have been reported with inter quartile range.
Nonetheless, I must congratulate the authors for exploring an important public health problem in the study area.
Competing interests
The author declares that there is no conflict of interest about this publication.