Utilization patterns of cohorts of elderly clients: a structural equation model.

OBJECTIVE

To identify a model that takes into account the interrelationship of health services utilization variables, and that allows examination of the utilization patterns of health services for a cohort of elderly clients.

DATA SOURCES AND STUDY SETTING

The data of each client in the study were taken from three computer databases maintained for administrative purposes by the Ministry of Health in British Columbia. Time frame for the utilization variables is one year before and one year after admission to the long-term care program in BC which occurred in 1981-1982.

STUDY DESIGN

A basic model was fitted to the utilization data for the year before admission and patterns of utilization were assessed for each gender-age group for the year before admission and for the two periods, using LISREL: Fifteen utilization variables were included: number of GP and specialist visits in different settings (office, home, etc.) and number of other services such as lab tests, hospital stay, etc.

DATA COLLECTION

The three files were linked to produce one record per client.

PRINCIPAL FINDINGS

A model was identified that fits the data well. The total effect of GP emergency room visits on hospital stay is 0.30 compared to 0.19 direct effect. The additional impact is produced via the effect of specialist consultations on hospital stay. This and similar findings by age, gender, and period are consistent with the joint dependency of utilization variables.

CONCLUSIONS

The analysis shows that males and females have different utilization patterns, while age has no effect on utilization of health services by male clients and only a small effect on utilization patterns by female clients. Admission to LTC causes more specialist contacts resulting from contact with a GP and generally a more intensive use of diagnostic and surgical procedures. However, there is significantly less acute care hospital services utilization.