Letters to the Editor

Indian Pediatrics 2002; 39:402  

Reply

We agree that systematic allocation is no substitute for randomization. We had an unusual problem of trying to separate the test infants from the controls at any one time in the unit so as to prevent control mothers from getting upset and opting out of the study. This would not have been solved by cluster randomization. As already mentioned in our paper under the para heading "General Measures" all mothers were given the same advice regarding neonatal care except that control mothers were repeatedly told not to administer any form of massage till 4 weeks term corrected age. The reason blinding was not done was mainly logistic. Also, it was during the Brazelton test and post-natal follow-up visits that we re-enforced whether to give massage or not and also confirmed compliance.

Regarding choice of primary variables, data variables can be classified by the way they are encoded or by the way they are measured. In this sense the changes in physiological parameters were seen and measured first and were directly and immediately related to the stimulation and have hence been classified as primary data variables. Also, we wanted to make sure that the stimulation would have no adverse effect on the physiological parameters of preterm neonates before looking at other effects. Only after this was established could we look at the more long-term, clinically important secondary variables (changes in which were seen later) like weight gain. However, for the purpose of sample size calculation we used weight gain as a random variable based on a previous study of normal weight gain in preterm neonates of similar birth weight done at our hospital.

Regarding control of temperature at home the average day temperature (massage was done only during day) in Mumbai is 30-34ºC and this hardly varies through the year compared to other parts of the country. Hypothermia is not a major problem in our city.

The weight gain was calculated from Day 1 of the study. The control group had 23 babies and the typographical error in Table 1 is regretted. Elsewhere in the paper (Table 2 and text) the number 23 remains. Regarding compliance, all mothers who followed up (69%) were compliant regarding massaging or not massaging. One baby in each group was started on top milk (Government packet milk which was the only affordable alternative) and since this was the same in both groups it was not mentioned.

Sheila Mathai,
Neonatologist
INHS Asvini, Mumbai,
India.