A Bird’s Eye View on Harms
Dr. Despina Contopoulos-Ioannidis Clinical Associate Research Professor, Pediatrics, Infectious Diseases, Stanford University.
The pediatric use of drugs often depends on adult efficacy and safety data due to the limited availability of RCTs in children. Evidence suggests that several clinically significant differences may exist between adults and children. Large-scale evaluations are needed to understand differences in comparative effectiveness and safety between adults and children.
Exploring methodological caveats in pediatric clinical research through large-scale empirical evaluations is the primary focus of research for Dr. Despina Contopoulos-Ioannidis of Stanford University and her research group. In 2010 this group published Comparative effectiveness of medical interventions in adults versus children in the Journal of Pediatrics. Through the review of 128 meta-analyses (MAs) they found that although treatment effects were similar on average between adults and children, the relative differences in the point estimates of those treatment effects in adults vs. children exceeded 50% in two thirds of the MAs. Could a similar trend of uncertainty exist in comparative safety?
In a similar method of investigation, Dr. Contopoulos-Ioannidis recently published in Pediatrics another large scale empirical evaluation of evidence from the Cochrane Database of Systematic Reviews entitled Safety of medical interventions in children versus adults. The aim was to take a bird’s eye view of the comparative safety of diverse pharmacologic interventions between children and adults. Within the 176 MAs that contained evidence from both adult and pediatric randomized clinical trials, harms/harms-related outcomes were grouped into 7 categories according to severity and specificity. Severe harms and withdrawals due to harms were the primary endpoints.
This study found that in ~40% of MAs the relative difference in harms/harm-related endpoints in children vs. adults was two-fold of greater. Furthermore, when the uncertainty around the point estimates of those relative differences was also considered, a two-fold or greater difference in harms between children and adults could not be excluded in ~80% of the MAs. Collectively these results suggest that the available evidence on the relative differences in harms from pharmacologic interventions between adults and children has large uncertainty. Therefore extrapolation of safety data from adults to children may be tenuous.
Another observation from this research was the paucity of MAs with comparative safety data from both adult and pediatric trials. Clinical trials should target harms more systematically. In order to further understand the relative differences in harms between adults and children, harms need to be better defined across diverse interventions and diseases. One way to facilitate this, would be the standardization in the collection and reporting of harms in clinical research.
When asked about the contributions StaR could make in this line of research Despina emphasized the need for StaR to participate in more empirical research: “Our priority should be to generate more empirical data, publish empirical evaluations in all areas of pediatric clinical research we are targeting with our “Standards” groups, and identify remaining research gaps in this field.”