Presentation is loading. Please wait.

Presentation is loading. Please wait.

Performance based supply-side incentives in the context of health system reform: A tale of two experiments in rural China Winnie Yip Professor of Health.

Similar presentations


Presentation on theme: "Performance based supply-side incentives in the context of health system reform: A tale of two experiments in rural China Winnie Yip Professor of Health."— Presentation transcript:

1 Performance based supply-side incentives in the context of health system reform: A tale of two experiments in rural China Winnie Yip Professor of Health Policy and Economics

2 Motivations Contributing to Chinese health care reform: – Rapid health expenditure growth – “Over” use of drugs and hi-tech diagnostics – A major cause: fee-for-service payment method with a distorted price schedule

3 THE growth rate: 16.51% p.a. GDP growth rate: 13.86% p.a. Rapid Health Expenditure Growth

4 Motivations Contributing to Chinese health care reform: – Rapid health expenditure growth – “Over” use of drugs and hi-tech diagnostics – A major cause: fee-for-service payment method with a distorted price schedule Filling gaps in the p4p literature: – Very limited experimental evidence – Still know very little about various design issues of p4p

5 Design Issues What to incentivize? – Activities/volume – Quality: structure, process, outcome How to incentivize? – Reward versus penalty – Targets – Relative comparisons – Improvements from last period – % of income at risk Who to incentivize? – Organization or individual

6 Background Two pilot experiments – Ningxia – Shandong New Cooperative Medical Scheme (NCMS) – rural health insurance scheme, has traditionally paid providers using FFS Primary health care in rural areas: township health centres (THCs) and village clinics (VCs)

7 Drug prescriptions for common cold in Ningxia, 2009

8 Baseline condition: drug prescriptions, Shandong

9 Incentive Design Ningxia Capitated budget w/ p4p Capitation rate includes outpatient services at both THCs and VCs Capitated budget: capitate rate x pop in the catchment area Capitation rate: – Based on visits per enrollee and expenditure per visit for THCs and VCs in the last two years Shandong Capitated budget w and w/o p4p Capitation rate includes outpatient services at both THCs and VCs Capitated budget: capitate rate x pop in the catchment area Capitation rate: – Take 30% of total premium for NCMS, divide by the number of enrolled population

10 Ningxia—Phase I ItemItem DescriptionWeightScore Antibiotic prescription Percentage of visits prescribed with antibiotics150Weight x (1 - % visits with antibiotics prescribed) IV-antibiotic prescription Among visits with antibiotics prescribed, percentage given via IV injection 150Weight x (1 - % with IV injection among those prescribed antibiotics) Falsifying visits Percentage of visits with identical patient name and health problems repeated within one day 50Weight x (1 - % of visits with identical patient name and health problems repeated within one day) Completeness of visit record Percentage of visit record that has at least one of the following incomplete: symptoms; diagnosis; drugs prescribed (types, dosage); exam/test prescribed (for township health centres); expenditure 50Weight x (1 - % visit record deemed incomplete) Patient satisfaction Are you satisfied with the cleanliness of the clinic; are providers patient and careful in explaining your health problems; explain to you re medication; explain to you how to take medication; satisfied with the providers technical quality; waiting time 20 eachWeight x (% response “1”) TOTAL500

11 Shandong Quality indicatorTarget and penalty Qualified prescriptions T: 100%. P: For each pp < T, half a point was deducted. Prescriptions containing 6 or more drugs T: 10%. P: For each pp > T, half a point was deducted. Divided prescriptions T: zero. P: Each occurrence results in a deduction of 1/10 of a point. Prescriptions combined multiple antibiotics For THCs, T is 10-15% For village clinics, T is 15-20% Prescriptions using intravenous (IV) transfusion For THCs, T is 25% for THC, 35% for village clinics For each pp > T, half a point was deducted. Prescriptions using steroids For THCs, T is 10%; for village clinics, T is 20- 22% For each pp>T, half a point was deducted. Score on test of skill in diagnosing and treating chronic illnesses T is 80 marks. For each mark<80, 1/5 or 1/10 of a point was deducted. Patient satisfaction T: 80% P: each pp below T, 1/20 deducted Environment and hygiene of the village clinic Subjective evaluation if not satisfactory, 1/10 of a point was deducted. Environment and hygiene of the THC Same as above

12 P4P Incentives on THCs Ningxia 30% with-held Tournament – Each THC and VC receives a score – THC score includes scores of its VCs – Within each county, an avg score is calculated for all THCs. THCs that score at the avg take 30% of with-hold; those that score above avg receives >30% of with-hold; those that score < avg receives < 30% of with-held Shandong 20% with-held Targets and penalty – Each THC and VC receives a score – THC scores include scores of its VCs

13 P4p Incentives on VCs Ningxia A VC specific capitated budget is specified, and 30% of the budget is with-held for performance assessment Within each town, all VCs are subject to the tournament incentives Shandong Incentives between THC and its VCs not specified

14 Town x First half yearSecond half year Name 30% With- held amount Actual payment based on QL score QL score Relative town average QL score Relative town average VC11250817387 4341.514360.98 VC284007879 2100.734851.09 VC32906420675 1170.414621.04 VC4847612607 4121.434330.97 VC51853014035 1490.524441.00 VC61752720820 3851.344451.00 VC7979412287 3781.323910.88

15 Source of Data for p4p Assessment Ningxia Prescription: – Management information system – Universe of all visits/prescriptions Patient satisfaction: survey of patients Shandong Prescription: – Random sample – THC: 50 per month – VC: 20 per month Patient satisfaction: survey of patients

16 Experimental Design Ningxia Pair-wise randomization: 14 pairs of towns (and all the village clinics within the town): whether the THC provided hospitalisation services; whether the THC was classified as a “centre”,; distance to the county seat; avg expenditure per OP visit; number of OP visits per year; number of VCs under its management; % agricultural population Lost one pair Baseline balance Shandong Blocked randomization by town and all VCs within a town Political interference Baseline unbalanced

17 Estimation Method Ningxia End line comparison of: – Treatment : capitated budget + p4p – Control: FFS – Effects: capitated budget + p4p Shandong Difference-in-difference – Group A: FFS to Capitated budget + p4p – Group B: FFS to Capitated budget – Effects: p4p – Caveat: in practice, group B also subject to 20% of with-held, except not told how they will be assessed

18 Findings

19 Ningxia: THC antibiotic prescription

20 Ningxia: VC antibiotic prescription

21

22

23 Control mean Adjusted Treatment effect (95% CI) p value Number of patient visits per day in township health centres All37.3-5.32 (-19.59, 8.95)0.450 Male19.1-2.37 (-9.03, 4.30)0.471 Female19.4-3.37 (-11.12, 4.39)0.380 Number of patient visits per day in village posts All 9.7 -0.90 (-3.60, 1.80)0.498 Male 5.7 -0.50 (-1.94, 0.94)0.480 Female 5.2 -0.59 (-1.85, 0.67) 0.342 Patient satisfaction score Township health centres 26.4 -0.03 (-0.69, 0.62) 0.913 Village posts 26.0 -0.10 (-0.63, 0.43)0.693 Causal effect of payment intervention on healthcare utilisation and patient satisfaction

24 Estimated impacts of P4P scheme – THCs, Shandong N§N§ Marginal effect § Marginal effect (%)* Std error¶ Bootstrap adjusted p-value‡ Both counties combined Cost 1917-7.001-0.16010.5820.561 No. drugs in prescription 2410-0.469-0.1260.3040.120 2 or more drugs in prescription 2410-0.071-0.0860.0480.130 6 or more drugs in prescription 2410-0.076-0.3160.0480.126 Prescription contained antibiotic 2410-0.050-0.1050.0580.356 2 or more antibiotics 2410-0.076-0.5110.0400.074 Intravenous injection 2410-0.096-0.2690.0750.194 Prescription contained steroids 2410-0.091-0.5310.0500.078

25 Estimated impacts of P4P scheme – VCs, Shandong N§N§ Marginal effect § Marginal effect (%)* Std error¶ Bootstra p adjusted p-value‡ Both counties combined Cost 3273-3.158-0.1074.6940.535 No. drugs in prescription 3876-0.280-0.0880.2070.168 2 or more drugs in prescription 3876-0.063-0.0790.0540.256 6 or more drugs in prescription 3876-0.026-0.1990.0300.396 Prescription contained antibiotic 3873-0.033-0.0560.0740.635 2 or more antibiotics 38730.0550.4850.0490.304 Intravenous injection 3874-0.066-0.1840.0600.284 Prescription contained steroids 3874-0.034-0.3280.0410.400

26 Conclusions Capitation+p4p has some positive effect in reducing inappropriate drug prescription, beyond removing drug profit margin Antibiotic prescription is still high Gaining the government’s understanding of the value of the project, and their respect/trust for the research team to maintain independence is important But, – Cannot isolate the effects of capitation and P4p separately – Requires setting up routine and automatically generated data for performance measurement – P4p needs to be coupled with feedbacks – Effects on other outcome such as hypertension, diabetes remain to be seen


Download ppt "Performance based supply-side incentives in the context of health system reform: A tale of two experiments in rural China Winnie Yip Professor of Health."

Similar presentations


Ads by Google