Financial incentives and initiatives to improve the quality of care in South Africa

http://resyst.lshtm.ac.uk@RESYSTresearch

Financial Incentives and Initiatives to Improve the Quality of Primary Care in South Africa

Duane Blaauw & Mylene Lagarde

University of the WitwatersrandLondon School of Hygiene & Tropical Medicine


Background South Africa

Upper middle-income country Poor health outcomes & persistent health inequalities Entrenched dual health care system

Sector Source Popn

Health Expenditu

re

1º Care Provider

s

1º Care

Doctors

Payment

Private

Private health insurance 20% 60% GPs 70% FFS

Public Tax-funded 80% 40% Nurses 30% Salary Recent reforms to improve the quality of primary care1. Specialist MCH support teams2. CHW outreach teams3. Improved school health services4. GP contracting initiative

General Practitioner (GP) Contracting

Contract private sector GPs to work number of hours each month in public sector clinics

Sessional payments Maximum public sector rate much lower than

remuneration in private practice Appeal to the altruism of private GPs National campaign by Minister to mobilise GP support

Related Literature

Strong evidence of pro-social behaviour from experimental economics (Fehr & Schmidt, 1999)

Public sector employees have stronger pro-social preferences than private sector employees (Perry & Wise, 2010)

Doctors expected to be motivated by patient benefit and ethical practice rather than profit (Arrow, 1963; McGuire, 2000)

Utility functions of doctors include benefit to patients (Ellis & McGuire, 1986; Farley, 1986)

Demonstrated heterogeneity in altruism of medical students (Godager & Wiesen, 2013)

More altruistic medical students choose to work in public sector (Serra et al, 2010; Kolstad & Lindkvist, 2012)

Study Objectives

Likely uptake of sessional contracts by private GPs Quantify relative importance of different contract

elements Contrast personal income and benefit to patients

Using a stated preference discrete choice experiment (DCE)

Methods

DCE Design [Ngene] Generic design: 2 alternatives + opt-out 8 attributes derived from preliminary research Incremental design strategy

Orthogonal design Priors from first 25 responses Bayesian D-Efficient design (Rose & Bliemer, 2009)

DCE Analysis [Limdep/Nlogit] Multinomial logit (MNL) Heterogeneity of preferences

MNL interactions with GP and practice characteristics Mixed (Random parameters) logit (MXL) Latent class logit (LCM)

DCE DesignATTRIBUTE LEVELS

1. Distance to nearest public sector doctor

20km 40km

2. Basic contract rate R 265 per hour (11ZAR=1USD) R 350 per hour (33% increase) R 435 per hour (66% increase) R 520 per hour (100% increase)

3. Deprivation allowance

None An additional R 85 per hour (33% increase)

4. Performance bonus

None An additional R 85 per hour if meet specified quality

targets for consultation records, referrals & adherence to treatment protocols

5. Travel reimbursement

R 100 per trip (35% increase) R 130 per trip (50% increase)

6. Free CPD points for induction and training

None 15 points

7. Type and location Fully-functional container clinic in an informal settlement

Fixed clinic in the township 8. Distance from your

current practice 10km 30km

DCE Choice Task

Which of these two contracts would you choose?

CONTRACT A CONTRACT B NEITHER⃝� ⃝� ⃝�

Respondents and Sampling

National database of ~8000 active GPs Random sample of 493 Email invitation

15.1% response rate Online survey

DCE + socio-demographic questionnaire

Respondents

Male 64.4%

Age 48.8 ± 10.4

Time working as private GP 18.5 ± 10.5 yearCharge per consultation for insured patients

R 297.70 ± 49.78

Charge per consultation for cash patients R 254.71 ± 62.23

Estimated turnover per hour R 1119.95 ± 600.39

Have done sessional work in public sector 22.5%Said were likely to do sessional work in public clinic 50.0%

Coeff (SE)Contract characteristicsNearest public sector doctor 20km further away 0.000 (0.093) Increase in basic contract rate of R85 per hour 0.569 (0.175) ***

R170 per hour 1.447 (0.174) ***R255 per hour 2.013 (0.171) ***

Additional deprivation allowance of R85 per hour 0.571 (0.110) ***Additional performance bonus of R85 per hour 0.486 (0.095) ***Additional transport allowance of R100 per trip 0.580 (0.153) ***

R130 per trip 0.686 (0.171) ***10 CPD points for induction and training 0.061 (0.089) Fixed clinic 0.429 (0.096) ***Facility 20km nearer to current practice 1.024 (0.156) ***

Opt-out constant 3.083 (0.237) ***

Pseudo R-squared 0.123*** p<0.01, ** p<0.05, * p<0.10

MNL

Facility FixedClinic

FixedClinic

FixedClinic Container Fixed

ClinicFixedClinic

FixedClinic

FixedClinic

FixedClinic

Basic rate R265 R265 R265 R265 R265 R350 R265 R265 R265

Distance 30km 30km 30km 30km 10km 30km 30km 30km 30km

Transport allowance 0 0 0 0 0 0 0 0 R100

Deprivation allowance 0 0 0 0 0 0 R85 0 0

Performance –related bonus

0 0 0 0 0 0 0 R85 0

Nearest doctor 20km 40km 20km 40km 20km 20km 20km 20km 20km

CPD points 0 0 10 0 0 0 0 0 0

Total/Hour R265 R265 R265 R265 R265 R350 R350 R350 R365

UPTAKE 7.5% 7.9% 8.0% 5.2% 16.9% 12.2% 13.3% 12.6% 12.9%

Model Simulations

Facility FixedClinic

FixedClinic

FixedClinic

FixedClinic

FixedClinic

FixedClinic

Basic rate R350 R435 R520 R350 R520 R520

Distance 30km 30km 30km 30km 30km 30km

Transport allowance 0 0 0 R100 R130 R130

Deprivation allowance 0 0 0 R85 R85 R85

Performance –related bonus

0 0 0 R85 R85 R85

Nearest doctor 20km 20km 20km 20km 40km 40km

CPD points 0 0 0 0 10 10

Total/Hour R350 R435 R520 R620 R820 R820

UPTAKE 12.2% 21.8% 34.6% 43.6% 77.2% 78.6%

Model Simulations

Coef (SE)

Contract characteristicsNearest public sector doctor 20km further away -0.051 (0.109) Increase in basic contract rate of R85 per hour 0.723 (0.213) ***

R170 per hour 1.721 (0.211) *** R255 per hour 2.315 (0.210) ***

Additional deprivation allowance of R85 per hour 0.630 (0.131) ***Additional performance bonus of R85 per hour 0.565 (0.113) ***Additional transport allowance of R100 per trip 0.689 (0.181) ***

R130 per trip 0.870 (0.204) ***10 CPD points for induction and training 0.017 (0.105) Fixed clinic 0.333 (0.114) ***Facility 20km nearer to current practice 1.128 (0.190) ***

Opt-out constant 3.784 (0.315) ***Demographic characteristics Interaction with opt-outUnder 50 years old -0.252 (0.162) Upper tertile of turnover per hour 0.980 (0.181) ***Currently doing sessional work -0.584 (0.196) ***Likely to accept sessional work in public clinic -1.089 (0.156) ***Pseudo R-squared 0.212*** p<0.01, ** p<0.05, * p<0.10

MNL Interactions

Mixed Logit Mean (se) SD (se)Contract characteristicsNearest public sector doctor 20km further away 0.028 (0.180) 0.241 (0.255) Increase in basic contract rate of R85 per hour 1.955 (0.436) *** 1.496 (0.444) ***

R170 per hour 4.579 (0.496) *** 0.772 (0.297) ***R255 per hour 5.869 (0.543) *** 1.933 (0.302) ***

Additional deprivation allowance of R85 per hour 1.446 (0.247) *** 0.930 (0.188) ***Additional performance bonus of R85 per hour 1.014 (0.241) *** 1.441 (0.272) ***Additional transport allowance of R100 per trip 1.267 (0.318) *** 0.793 (0.371) **

R130 per trip 1.964 (0.349) *** 0.465 (0.416) 10 CPD points for induction and training 0.412 (0.184) ** 0.340 (0.242) Fixed clinic 1.794 (0.435) *** 2.986 (0.416) ***Facility 20km nearer to current practice 2.399 (0.680) *** 4.333 (0.520) ***

Opt-out constant 7.626 (0.771) *** 6.630 (0.708) *** Pseudo R-squared 0.504*** p<0.01, ** p<0.05, * p<0.10

Latent Class Analysis Class 1 Class 2 Class 3

Mean (se) Mean (se) Mean (se)Contract characteristicsNearest public sector doctor 20km further away -1.372 (0.895) -0.022 (0.134) -0.038 (0.156) Increase in basic contract rate per R85 / hour 2.648 (0.959) *** 1.172 (0.094) *** 0.876 (0.134) ***Additional deprivation allowance of R85 / hour 2.829 (1.174) ** 0.888 (0.148) *** 0.846 (0.175) ***Additional Performance bonus of R85 / hour 1.847 (0.963) * 0.823 (0.148) *** 0.719 (0.180) ***Additional transport allowance of R100 / trip 3.238 (1.423) ** 1.149 (0.239) *** 0.943 (0.293) ***

R130 / trip 3.323 (1.441) ** 1.521 (0.253) *** 1.015 (0.331) ***10 CPD points for induction and training 1.117 (0.798) 0.054 (0.129) 0.080 (0.150) Fixed clinic 0.348 (0.642) 0.883 (0.148) *** 0.251 (0.173) Facility 20km nearer to current practice 4.514 (1.471) *** 1.784 (0.254) *** 1.586 (0.342) ***

Opt-out constant 15.639 (4.315) *** 4.841 (0.431) *** -0.644 (0.631)

Class probabilities 0.296 0.462 0.242 Pseudo R-squared 0.394*** p<0.01, ** p<0.05, * p<0.10

Baseline rate

↑ patient benefit

CPD points

25% ↑ P4P

25% ↑ deprivation allowance

25% ↑ basic rate

25% ↑ transport allowance

Closer to own practice

50% ↑ basic rate

75% ↑ basic rate

0.0% 25.0% 50.0% 75.0% 100.0%

7.5%

7.9%

8.0%

11.2%

12.6%

12.9%

13.3%

16.9%

21.8%

34.6%

Uptake of Sessional ContractGroup Average (MNL)

Baseline rate

↑ patient benefit

CPD points

25% ↑ P4P


25% ↑ basic rate



50% ↑ basic rate

75% ↑ basic rate

0.0% 25.0% 50.0% 75.0% 100.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

0.0%

Uptake of Sessional ContractLatent Class 1 (29.6%)

Baseline rate

↑ patient benefit

CPD points

25% ↑ P4P


25% ↑ basic rate



50% ↑ basic rate

75% ↑ basic rate

0.0% 25.0% 50.0% 75.0% 100.0%

4.8%

4.9%

5.1%

10.4%

11.0%

14.1%

13.8%

23.2%

34.6%

63.1%


Baseline rate

↑ patient benefit

CPD points

25% ↑ P4P


25% ↑ basic rate



50% ↑ basic rate

75% ↑ basic rate

0.0% 25.0% 50.0% 75.0% 100.0%

44.3%

45.3%

46.3%

62.1%

65.0%

65.7%

67.2%

79.6%

82.1%

91.7%


GP Concerns

Sessional rates not market-related Security risks Likelihood and timeliness of payment by the

Department of Health Availability of medicines and equipment Lack of consultation in the formulation and design of

the policy initiative

Limitations

Sampling bias Non-response bias Hypothetical bias Specification of patient benefit Decisions influenced by considerations outside of

contract design

Main Findings

Low uptake of proposed public sector contracts by private GPs

Private GPs more motivated by own financial welfare than potential benefit to public sector patients

But significant heterogeneity in the pro-social preferences of private GPs Some GPs completely opposed to public sector work Largest proportion mainly motivated by payment rates Small group with more pro-social orientation

Policy Implications

Inform improvements in contract design Would require significant financial resources to

increase contract uptake Framing of financial incentives makes little difference No significant opposition to performance monitoring Target policy initiative to more pro-social GPs


RESYST is funded by UK aid from the UK Department for International Development (DFID). However, the views expressed do not necessarily reflect the Department’s official policies.


Financial incentives and initiatives to improve the quality of care in South Africa

Health & Medicine

Transcript of Financial incentives and initiatives to improve the quality of care in South Africa