Profiling hospital utilization in a mixed public–private system

Megan Gu, Meliyanni Johar
2016 Applied Economics  
While there is an extensive body of literature on the demand for hospital services, little is known about the interaction between public and private hospitals in a mixed system. In this paper, we (i) apply latent class analysis to identify distinct subgroups of patients who use the hospital market differently, (ii) characterise each patient type by their personal characteristics and (iii) link the patient type to future hospital admissions. We apply our analysis to individual-level longitudinal
more » ... patient data from Australia, focusing on three popular procedures that are performed in both public and private hospitals. We find 4 to 5 patient types. The most common types use either a public or a private hospital almost exclusively and absorb a moderate level of hospital resources. The severe types represent 13 to 17% of patients. The type which uses both sectors makes up 10 to 20% and tends to have private health insurance coverage. The patient types are predictive of prospective utilisations, as we find that patients tend to be admitted to the sector they have used in the past. By revealing how patients use coexisting public and private hospitals, our results have direct implications on health resource financing and allocations. Notes: The total sample size is 2,877. The average partial effects (APE) are derived from a multinomial logit model with robust standard errors and PB type as the base group. APEs are corrected for multiple category variables. *, ** and *** indicate significance at the 10%, 5% and 1%, respectively.
doi:10.1080/00036846.2016.1197371 fatcat:dwevkv73c5blfpw6nkegpu6zjy