Severity-Adjusted Differences in Hospital Utilization by Gender

Abstract

Gender-based differences in hospital use may result from biological differences or may suggest problems of access to health services and quality of care. We hypoth esized that there should be no difference in hospital care between men and women, given the same diagnosis. Hospitalizations were characterized by severity of ill ness, as this may indicate the timeliness of hospital care. Hospitalizations may be too late (with higher severity of illness) resulting in long stays and high costs, or too early (with lower severity of illness) resulting in care that could be given in alternative treatment settings. Three abdominal conditions were examined which could be misdiagnosed or confused with other diseases involv ing the female reproductive system: appendicitis, diver ticulitis, and cholecystitis. The National Hospital Dis charge Survey (NHDS) was used for analysis. Disease staging was used to assign a severity of illness indicator, ranging from stage 1 (conditions with no complications) to stage 3 (multiple site involvement, poor prognosis). For each disease, the percentage of discharges and the age-adjusted discharge rate per 1000 population was examined by stage of illness and gender. For appendec tomy, there was a significantly greater percentage of men at stage 1 (lower severity) compared to women (73% versus 67%). For diverticular disease, women had higher proportions of stage 2/3 discharges than men for both medical and surgical hospitalizations. For chole cystitis, women had a greater percentage of hospitaliza tions at stage 1 than men, notably for surgical treatment (63% compared with 38%), although more men were admitted at stage 2 for both medical and surgical treat ment. Gender differences in the proportions of hospital ization by severity of illness suggest that there are prob lems in the timely recognition and treatment of appendicitis and diverticulitis in women. However, for cholecystitis, it appears that men, rather than women, have a problem of timely recognition and treatment. Future research may seek to identify additional patient characteristics, physician attitudes, and treatment pro tocols.