We read with great interest the recent article by Balu, et al. [1].
However, in methodology, sample size considerations and calculations have
not been mentioned.
As the sensitivity of gold standard has been mentioned
as 88%, we calculated the sample size for 80% with a deviation of 5% on
either side and the calculations show that sample size in this study is
woefully inadequate [2]. Thus the study was inadequately powered to assess
the validity of the proposed screening tests. As PPV depends upon
prevalence, likelihood ratio is a better measure to overcome this
inadequacy which has not been calculated in this study. Authors also state
that more training (how much) would be required to get a better result,
which raises the question of internal validity. In the diagram describing
association of pulse oximetry with clinical evaluation and
echocardiography it seems that all newborns went through pulse oximetry
first followed by clinical examination but authors initially mentioned the
reverse sequence. The appropriate sequence of clinical examination
followed by oximetry has also been described earlier [3].
References
1. Vaidyanathan B, Sathish G, Mohanan ST, Sundaram KR,
Warrier KKR, Kumar RK. Clinical screening for congenital heart disease at
birth : A prospective study in a community hospital in Kerala. Indian
Pediatr. 2011;48:25-30.
2. Kumar R, Indrayan A. Receiver operating
characteristic (ROC) curve for medical researchers. Indian
Pediatr.2011;48:277-87.
3. Jonathan DR. Clinical screening for congenital heart disease at
birth: A long way to go. Indian Pediatr. 2011;48:17-8.