[The confounding factor and interaction in studies of pathogenesis--an analysis of the secondary infection of viral hepatitis with corticoid therapy].

A total of 762 cases of viral hepatitis admitted to the two teaching hospitals of the Third Military Medical College were used for study. They were divided into two groups depending upon whether corticoid treatment was given. As the result of cRR = 6.06, chi 2 = 174.16, P < 0.01, it is obvious that the corticoid therapy is a risk factor for the secondary infection in patients suffered from viral hepatitis. For getting rid of the interference of confounding factor and interaction a stratified analysis was performed. These patients were again divided into severe type and moderate type according to the severity of the disease. The results of stratified analysis were aRR (F) = 3.05 and aRR (F) = 1.48, cRR > aRR. These findings demonstrated that the degree of severity of patients played a primary role in the secondary infection of hepatitis and showed that the confounding factor was present. On the other hand, the result of aRR (F) not equal to aRR (F) indicated that the interaction also existed simultaneously. This study suggests that special attention should be paid to maintain and increase the level of immunological defence function in patients with hepatitis. And, it is necessary to restrict the indiscriminate usage of corticoids, particularly in severe patients, for preventing nosocomial infection.