Performance Based Payments to Physicians in Turkish Public

Performance Based Payments to Physicians in Turkish Public Hospitals: Issues in Impact Assessment Burcay Erus and Ozan Hatipoglu Bosphorus University, Istanbul

Outline n Information on Performance Based Supplementary Payment System in Turkey n Impact evaluation-various dimensions n Difficulties ¨ Methods ¨ Data collection ¨ Simultaneous reforms

Reforms in healthcare n Reform in Healthcare: Transformation in Health ¨ Starting in 2003 ¨ With the support of WB ¨ Aim: increase access and efficiency n Unification of three distinct public health insurance, first steps towards universal coverage n Inclusion of private providers in the public insurance package n New policies in public hospitals to increase efficiency n A family physician system

Performance Based Supplementary Payment System n n New compensation scheme for physicians in public hospitals starting in 2004 Previously ¨ based on salary and limited bonus payments ¨ dual practice allowed and common ¨ anecdotal evidence on problems in access to public hospitals, presence of informal payments

Performance Based Supplementary Payment System-continued n New scheme ¨ Physicians paid bonuses based on points collected throughout the month ¨ From outpatient physical exams, inpatient procedures, tests and diagnoses ¨ Payments from hospital’s receipts after hospital expenses are made ¨ Quality dimension is largely lacking, only broad hospital measures n Patient satisfaction in general, number of exam rooms, etc.

Points-Examples Table translated from Mo. H presentations on PBSP

Hospital revenues 45 -85% hospital expenses 0 -40% to performance payments 15% to treasury A hospital performance point is calculated which determines the ratio available for physician payments (currently all hospitals score above. 75) Amount that remains after hospital expenditures is shared among the physicians according to the points collected during the month

Further details: n n n Incentives smaller for dual working physicians: points are multiplied with 0. 3 (were 0. 4 in the beginning) Administrative and support personnel (not involved in point earning procedures directly) receive supplementary payments based on their title and average points of the physicians There are caps on maximum supplementary payments (e. g. 8 times base salary for specialists)

Anecdotal evidence on impact n n n More patients treated¨ initial analysis of the data shows that outpatients per specialist (full time+dual) doubled More physicians working full time ¨ Mo. H statistics indicate an increase from 27% to 80% Even no vacation taken by physicians Unnecessary procedures More drugs prescribed ¨ Drug spending increasing despite considerable decrease in prices

Research objectives n Change in productivity ¨ Impact on quantity n Taking into account selection bias in specialists choice of full-time vs. dual ¨ Impact on quality n No sufficient data-left for further research ¨ Induced demand n Work in progress

Impact assessment-ideal case n n n Ideal Randomized Controlled Experiment ¨ Physicians assigned to different groups randomly ¨ Data collected for control and treatment groups before and after Problems ¨ Difficulties in having two different schemes for physicians working together in the same province ¨ Ethical issues-patients treated differentially ¨ Patients’ self selection into physicians Piloting as an alternative ¨ Differences across pilot provinces ¨ Politicians and reform process eager to move quick-not enough time to observe the impact

In the absence of experiment n n Limited publicly available data Survey of hospitals available 2001 -2006 Use variation in the number of full-time and dual working specialists over time and use fixed effects Deal with selection bias by simultaneously estimating demand for specialist at public hospital

Physician caseload n n n Ideally to measure physician effort ¨ Payments as a proxy ¨ Specific standard procedures-numbers We have ¨ Aggregate number of outpatients, inpatients, surgeries (categorized as major, middle, minor) Issues ¨ Reliability ¨ Lack of detail ¨ Aggregation-no info on value

Number of physicians ¨ Ideally Individual level physician data n Specific data per specialty, type n Socio-economic characteristics ¨ We have aggregate at hospital level ¨ Problems: n No info on decomposition into full time vs part time yet! n Should be available somewhere but not provided by the Ministry n No info on specialty n

Quality n n Ideally health outcomes ¨ Hospital mortality ¨ Surveys on health status ¨ Satisfaction ¨ Exam times ¨… We have ¨ No data at all on the quality

Hospital characteristics n n Ideally would have info on hospital chars that are relevant such as number of beds, infrastructure etc. We have a limited number of variables ¨ Number of beds

Simultaneous reforms n Simultaneous reforms ¨ Increased demand-number of public insurees increased and access is eased ¨ The number of full-time/dual physicians endogeneous ¨ Data needs n Information on number of insurees lacking-mostly because number of dependents are estimated (4 people per insuree) n The number of those with public insurance for poor is not known for earlier years n Need instruments!