Tables 3 and 4 present the results for the length of stay and charge data for the six clinical conditions. Taking into account disease severity, the results reflect the pervasive influence of hospital bed size. Larger-size hospitals have both longer length of stay and higher charges for each condition. Whether or not a hospital is a formal CQl/l'QM site, however, is significantly associated with lower length of stay and/or charges in 4 of the 12 equations and is consistently negative in all other equations. Thus, some support exists for the first hypothesis, suggesting that CQl/f(pd sites would experience shorter length of stay and lower charges for selected clinical conditions. There is little support, however, for hypothesis 4, suggesting that the degree of QI implementation would be associated with shorter length of stay and lower charges. Only two predicted associations are observed (i.e., charges for stroke and length of stay for congestive heart failure). While the other relationships involving implementation are consistently in the predicted direction (i.e., greater QI implementation associated with shorter length of stay and lower charges), they do not approach statistical significance. The major reason for this is that employees in larger-size hospitals report lower QJ implementation than employees in smaller-size hospitals
( r = - .41; p .05). In fact, when bed size is deleted from the equation, QI implementation is significant in the predicted direction in 11 of the 12 equations. Additional analysis of the effects of region, location, and market competition did not change these relationships.