Does patient health behaviour respond to doctor effort?

Incentive pay systems have been introduced in public sectors such as education and health care. In these sectors the output (education or health respectively) depends on the actions of different agents and it is unclear what the effects of such incentive systems are on the behaviour of untargeted agents. In this study we focus on patient health, modelled as a joint product of patient effort (through lifestyle and behaviour) and doctor effort (through diagnosis and treatment). Patient response to doctor effort is shown to be a priori ambiguous and depends on the degree of complementarity or substitution between doctor and patient effort. We build an empirical model to estimate the effect of doctors’ treatment effort on patient behaviour. To address the endogeneity of doctor effort we exploit a change in payments to doctors in the U.K. that led to incentive changes that varied by practice, depending on their prior performance levels. We use panel data on the physical activity, drinking and smoking behaviours of over 2000 cardiovascular disease patients aged over 50 in England and link these data to their primary care practice performance data. Our results indicate that primary care practices increased the proportion of patients with controlled disease from 76% to 83% in response to the payment change. Patients responded by reducing the frequency of drinking alcohol and their cigarette consumption, suggesting that patient efforts are complements to doctor effort. Understanding such complementarities has implications for assessing the design and effectiveness of pay-for-performance schemes which encourage higher doctor effort.