System and contextual influences on the surgical decision : a test of ten hypotheses and an exploration of sequenced data

Surgical rates for appendectomy, tonsillectomy, hernia repair, prostatectomy, cholecystectomy, and hysterectomy recorded on a monthly basis were modeled by two different procedures. The first part of the study utilized multiple regression analysis to determine the relationship of hospital ownership, hospital size, source of payment, ambiguity of indicators, sex-type of the surgery, and region of the country with surgical rates. The second part used time series analysis to determine the best generic model for each of the six surgeries involved.

The results show that source of payment, whether or not the surgery was paid for by private insurance, hospital ownership, whether or not the surgery was performed in a proprietary hospital, and region were important predictors of surgical rate. The regression analysis also showed that ambiguity of indicators and sex-type of the surgery moderated the relationship of other variables. Finally, the time series analysis showed that the generic models for surgery varied from one region to another if the surgery was performed primarily on one sex and the indicators for surgery were ambiguous.

It appears that time series analysis, because it determines geographical area differences that could not be legitimate medical variables, is a better way to establish those surgeries that evidence idiosyncratic decision making. It also appears that insurance status, hospital status, and region of the country influence the decisions made by surgeons and patients. Finally, it also appears that future studies should take sex-type of the surgery and ambiguity of indicators into consideration when analyzing idiosyncratic medical decision making.