The main finding in this study is that young Thai children are four times more likely to suffer a heavy burden of acute respiratory conditions such as acute bronchitis, pneumonia and bronchial asthma when they are raised in homes where 3 or 4 adults are smoking indoors and where family members are carrying children around while they are smoking.
After controlling for income, parent’s educational level, and SHS exposure during pregnancy, the multivariate analysis showed that Thai children who were exposed to SHS while being held by primary caregivers who were smoking were significantly more likely to be admitted for an acute lower respiratory condition, then controls. Having been held by caregivers who were smoking increased their likelihood of being admitted for an acute lower respiratory condition four-fold.
The multivariate analysis also showed that being exposed to SHS outside their homes was also a significant predictor of respiratory status. However, the proportions exposed in both groups were low (8% vs. 2%), suggesting that this mode of SHS exposure was of less importance.
Our analysis also shows that some Thai children who develop respiratory conditions are exposed to toxins in utero. This suggests that young Thai children’s exposure to SHS is a continuation of a household pattern of mothers being exposed to SHS during pregnancy. We suspect that for this reason, the multivariate model did not show this variable to be a significant predictor of respiratory status.
In the background of our analysis of the associations between SHS exposure and acute lower respiratory conditions, we found that children admitted are generally from poor or low-income households where money is tight for spending on primary care. Since the parent’s education level is also independently associated with respiratory status, we can infer that income and health awareness are both important factors influencing how children are raised, and when they receive care.
Through selecting our comparison group from a well-child clinic, we learned that parents who seek preventive primary care services for their under-fives tend to have somewhat higher incomes and be better educated. This suggests that disposable income and health awareness are important determinants of Thai children’s health beyond the scope of SHS exposure.
There are some limitations to this study. The samples were generated from one hospital and one well-child clinic in Bangkok. While it might have been preferable to select samples from multiple institutions, the available information indicates that this hospital and clinic are quite typical of such healthcare service providers. The available information also indicates that families who sought care at this hospital and clinic are like families who seek care at similar healthcare service providers throughout Bangkok and Central Thailand. Thus, we believe our findings are generalizable to much of the wider poor and low-income Thai population.
Additionally, the sample size was not as large. Nevertheless, the sample size provided adequate power for conducting robust bivariate and multivariate analyses. To generate an adequate sample size, we had to include children who were admitted with a variety of respiratory conditions rather than a single condition that is sensitive to SHS exposure. The analysis suggests that the fact of having a lower respiratory condition severe enough to require hospitalization was a sufficient criterion for measuring an outcome associated with SHS exposure.
Our original intended matching procedure did not produce samples that were sufficiently comparable to allow us to analyze the data using statistical tests appropriate for a case–control design. We therefore took a more conservative approach to analyzing the data.
We decided to rely on self-reported smoking status and self-reported exposure to SHS without obtaining biochemical verification. Self-reports can be subject to social desirability bias that can produce underestimates of smoking or overestimates of SHS exposure
. The data produced in this study may be subject to this bias, however the self-reported data are consistent with findings in other studies on Thai women’s smoking and SHS exposure levels
. While biochemical verification might have been useful for verifying parents’ current smoking status, this would not have produced reliable estimates of smoking during pregnancy or estimates of long-term exposure to SHS because cotinine has a half-life of about 2 days. For this reason, we felt the potential utility of biochemical verification for our analysis was outweighed by the cost.