Study cannot be generalised

EDITOR—In their study on the effects of postnatal health education on mothers in Nepal Bolam et al conclude that “individual health education for postnatal mothers in poor communities has no impact on infant feeding, care, or immunisation….”1 This finding cannot be generalised to mothers who do not enjoy the health advantages of the mothers studied by Bolam et al. The effectiveness of health education programmes should be tested among those they are intended to benefit. The authors admit that their group is select, having opted to pay to have their babies in hospital. This reflects a high awareness of health issues and also economic security in a country in which only 7% of births are attended by trained staff and gross national product is $200 (£125) per capita.

Ninety per cent of the deaths that occurred among infants in the study were not preventable, being associated with prematurity and congenital abnormality. Only one death was attributed to respiratory infection and none to diarrhoea; this is remarkable when, as reported by the authors, infant mortality is 98/1000 live births in Nepal. The highly select nature of the women studied is further reflected by the 95% rate of infant immunisation among the group. The authors acknowledge this; they had anticipated rates of 40% and thus miscalculated their sample size. Running a pilot study would have avoided such problems and indicated the need to select a more vulnerable group.

Similarly the statistics on outcome at 6 months reflect a well nourished cohort of Nepali infants; it would be difficult to improve this outcome even if the intervention had improved breast feeding practice. That the rate of breast feeding was below average probably is not a …