On the trail of necrotizing fasciitis in children.

T Hsieh and colleagues are to be commended for their attempt to define the pediatric presentation and outcome of necrotizing fasciitis versus cellulitis, using a case–control study design. One major limitation of their study is the paucity of cases of necrotizing fasciitis (8 cases in total), despite a 16-year period for the retrospective analysis. This raises a question about the accuracy of the ICD-9 coding system for identifying cases of necrotizing fasciitis or similar entities. The answer is that it is not particularly accurate. For example, the sensitivity of ICD-9-CM codes appears to be at most 58.3% for laboratory-confirmed pneumococcal pneumoniae and tends to be much lower than that. Identification of common occurrences such as adult stroke, myocardial infarction, childhood accidents and reportable communicable diseases via ICD-9 and ICD-9-CM codes is often equally poor, especially among pediatricians. Furthermore, coding discrepancies are greater with more complex medical cases, such as necrotizing fasciitis. I do not criticize the authors for this limitation, but feel that it may help explain why so few cases were identified over such a long study period. I am concerned that the authors did not estimate a study sample size that would have enabled them to address their question(s) with greater study power and precision. Sample size estimation is an important part of any study design, especially in a case– control study that attempts to examine a rare occurrence like pediatric necrotizing fasciitis. Accordingly, a colleague and I developed a practical paper to assist clinician-researchers in the difficult task of estimating sample size for such studies. As Hsieh and colleagues pointed out, they may have identified a greater number of cases by embarking on a multicentre study, which is often required when rare diseases are studied. It was for this purpose that the Ontario Group A Streptococcal Study Group was formed. During 1992 and 1993 alone, this group identified 323 cases of invasive group A streptococcal disease in Ontario; the highest rates were among children and elderly people. However, necrotizing fasciitis occurred in only 6% of all patients, highlighting the rarity of this disease and its high rate of associated morbidity and mortality.