The Dynamics of Medical Care Use in the British Household Panel Survey

Abstract

We explore whether medical care use is persistent over a long panel using 18 waves of the British Household Panel Survey. Of particular interest is high medical care use because a few high users account for a disproportionate amount of use while many individuals use no medical care in a given year. If health is a primary driver of medical care demand, and we control for health, then past medical care use should be uninformative for future use. However, we find that conditional on health, other covariates and unobservable heterogeneity, medical care use remains significantly persistent. "No use" and "high use" are more strongly persistent, and persistence is generally stronger for women, those in poor health, and at older ages. We find that unobservable heterogeneity explains between 10% and 25% of the variation in medical care use. This heterogeneity is significantly correlated with both medical care use and health over our long panel. These findings have implications for the econometric modeling of medical care demand and suggest that policies aimed to reduce aggregate medical care spending by improving health, particularly the health of seniors, may be less effective than projected using static models.

Variations in the persistence of health expenditures and the implications for the design of capitation payments in Taiwan

“…What we found was consistent with a longitudinal study in Britain that concluded that, ''the conditional probability of high use after experiencing high use may be fundamentally different for individuals over their remaining lifetime''. 4 Moreover, our finding that chronic conditions provided the greatest improvement in model performance is similar to the literature: one study reported that diagnoses of cancer and diabetes were associated with higher expenditures. 1 Another study using Taiwan's NHI data also concluded that adding morbidity trajectories increased the predictive power of the risk adjustment model.…”

“…Secondly, despite having 5 years of insurance claims, our data included only one baseline health survey, so we could not use a more advanced dynamic model to control for unobservable heterogeneity. 4 Thirdly, there were some differences between the consent and the nonconsent groups with the latter excluded from our sample, though this is unlikely to have affected our main findings since none of the characteristics on which consenters and non-consenters differed were significant predictors of health care expenditures in our regression analyses.…”

“…Many previous studies have found that past health care use is a significant predictor of future use, after controlling for health and sociodemographic factors. 1,3,4,12 Nevertheless, the pitfall of including prior expenditure as a risk adjustment factor is that the absolute number in dollars could be manipulated and promote provision of more and maybe unnecessary services. 13 This study focused on the rank in health expenditure which is considered a relative measure that cannot be determined by a single health care provider.…”

“…[1][2][3][4] For example, a study showed that the top 5% of people who received any hospital care in the United Kingdom during 1991-2008 accounted for 34% of all inpatient days, while more than 90% of the sample did not have any inpatient days annually. 4 A study of Medicare beneficiaries in the United States found that among the most expensive 5% of users in 2002, 23.8% still had high costs in 2003. 2 Previous studies on the concentration of Taiwan's NHI expenditures reported similar findings, for example, that the top 5% of users consumed more than 50% of total medical expenditures in 2003.…”

New Evidence on the Persistence of Health Spending

“…In addition to the studies of U.S. spending summarized in Table 1, an interesting recent study used an 18-year balanced panel of persons age 16 and over from the British Household Panel Survey to examine the persistence of healthcare utilization (Kohn & Liu, 2013). Key findings were that past use predicted future use even after controlling for health and other characteristics, that past utilization predicted future utilization more strongly at older ages and lower health status, and that first year utilization retained some predictive power throughout the follow-up.…”