Delivering Effective Health Care for AllMonday 29th March, 2010

What policies would attract health professionals to rural areas? Evidence from South Africa, Thailand and Kenya

Duane BlaauwCentre for Health Policy, South Africa

Possible interventions to attract health workers to under-served areas?Selection Geographic origin

EthnicityGenderCareer intentionService orientation

Education Location of training collegesCurriculum contentRural exposureFellowships

Coercion Registration requirementSpecialisation requirementInternational recruitment

Incentives Bursaries & scholarshipsDirect financial incentives (rural allowance)

Support Professional supportPersonal support

Grobler et al (2009)

Evidence of policy effectiveness?•Cochrane review (2009) found no rigorous

controlled studies •Need more rigorous evaluations of impact•Methodological challenges▫RCTs may not be possible▫RCTs may not be sufficient

• Interim solutions?▫Longitudinal HR databases▫Modelling of stated preference data

Study methods

•Discrete choice experiment (DCE) with nursing graduates in South Africa, Thailand & Kenya

•DCE Design•DCE Analysis•Used statistical model to investigate:▫Rural uptake for different policy combinations▫Cost-effectiveness of different interventions

Which of these two public sector facilities would you choose to work in?

RURAL Facility URBAN Facility

Type of facility Hospital Clinic

Monthly salary R120,000 per year R120,000 per year

Rural allowance An additional R12,000 per year None

The number of years you would have to work before getting study leave to specialise

2 years 6 years

The housing provided None None

The number of years you would have to work before being eligible for promotion

2 years 2 years

The car allowance offered None None

The workplace culture and style of management This facility is formal and structured. The

managers emphasise stability, following rules, and keeping things running smoothly.

This facility is personal and supportive. The

managers emphasise teamwork, loyalty, and developing the full potential of staff.

Which facility would you choose? Rural Facility Urban Facility

Annual Salary

DCE designLabelled designTwo choices

6 different rural policy interventions

16 choice sets

Different country preferences

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

30%

rura

l allo

wan

ce

Pref

eren

tial t

rain

ing

30%

rura

l allo

wan

ce

0% 20% 40% 60% 80% 100%

16.9%

3.3%

15.2%

39.0%

5.8%

36.4%

8.1%

5.4%

6.5%

35.5%

1.2%

21.9%

8.6%

6.7%

17.0%

52.9%

87.5%

58.6%

75.0%

90.0%

79.8%

44.1%

89.6%

49.9%

71.5%

85.4%

65.3%

44.6%

90.9%

60.4%

% Choosing Rural Job

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

30%

rura

l allo

wan

ce

Pref

eren

tial t

rain

ing

30%

rura

l allo

wan

ce

0% 100%

36.0%

84.2%

43.4%

36.0%

84.2%

43.4%

36.0%

84.2%

43.4%

36.0%

84.2%

43.4%

36.0%

84.2%

43.4%

% Choosing Rural Job

Impact of single interventions

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

30%

rura

l allo

wan

ce

Non

-fina

ncia

l int

erve

ntio

nsAl

l int

erve

ntio

ns10

% ru

ral a

llow

ance

+ tr

aini

ng

0% 20% 40% 60% 80% 100%

39.0%

5.8%

36.4%

50.3%

13.9%

39.1%

59.2%

14.5%

51.7%

46.0%

3.3%

34.0%

75.0%

90.0%

79.8%

86.3%

98.1%

82.5%

95.2%

98.7%

95.1%

82.0%

87.5%

77.4%

% Choosing Rural Job

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

ZA

TH

KN

30%

rura

l allo

wan

ce

Non

-fina

ncia

l int

erve

ntio

nsAl

l int

erve

ntio

ns10

% ru

ral a

llow

ance

+ tr

aini

ng

0% 100%

36.0%

84.2%

43.4%

36.0%

84.2%

43.4%

36.0%

84.2%

43.4%

36.0%

84.2%

43.4%

% Choosing Rural Job

Impact of packages of interventions

Cost-effectiveness of interventions (ZA)

Policy Scenario Additional Effect(Rural Nurse Years)

Additional Cost(ZAR, R1=£11)

Cost-Effectiveness(ZAR / Rural Nurse Year)

Individual Interventions

Base scenario - - -

Promoted faster 64 46 124 497 720 695

Better housing 136 58 736 999 431 890

Car allowance 415 121 034 432 291 649

Earlier study leave 515 128 885 182 250 262

10% rural allowance (RA) 629 308 091 306 489 811

20% rural allowance 985 557 401 459 565 890

30% rural allowance 1 259 801 537 625 636 646

Packages of Interventions

Quick promotion + Car 461 166 454 292 361 072

Better housing + Car 625 201 971 262 323 154

Quick promotion + 20% RA 1 067 629 126 154 589 622

Earlier study leave + Car 1 362 393 988 330 289 272

10% Rural allowance + Car 1 390 546 995 962 393 522

Better housing + Car + 20% RA 2 273

1 036 555 469 456 030

Study leave soon + 10% RA + Car 2 445 954 698 686 390 470

Study leave soon + 20% RA 3 233

1 420 449 751 439 360

0 500 1,000 1,500 2,000 2,500 3,000R

R 200,000,000

R 400,000,000

R 600,000,000

R 800,000,000

R 1,000,000,000

R 1,200,000,000

R 1,400,000,000

Incremental number of rural nurse-year

Incr

e-

me

nta

l co

st

E1

E2

E4

E5

Early study leave + carICER: R 313,096

Early study leave + car + 10% rural allowanceICER: R 517,831

Early study leave + 20% rural allowanceICER: R 590,988

Early study leaveICER: R 250,065

MORE COST-EFFECTIVE

Incremental cost-effectiveness ratios

Base costICER: R 238,324

0 500 1,000 1,500 2,000 2,500 3,000R

R 200,000,000

R 400,000,000

R 600,000,000

R 800,000,000

R 1,000,000,000

R 1,200,000,000

R 1,400,000,000

Incremental number of rural nurse-year

Incr

e-

me

nta

l co

st

E1E1 (#3)

E2

E2 (#3)

E4

E5

E4 (#3)

E5 (#3)

Base (#3)

Original sample

Sample #3 (75% rural)

Base cost: R 213,170

ICER: R 252,608

ICER: R 318,519

ICER: R 558,983

Incremental cost-effectiveness ratios

ICER: R 590,988

Base costICER: R 238,324

Implications for policy and practice•Policymakers in LMICs should make more

use of modelling data to inform the design of HR policies.

•There are no generic HR solutions. HR policies need to be tailored to individual country contexts.

•A combination of financial and non-financial HR strategies is required.

•Non-financial interventions can be as effective as salary increases and are more cost-effective.

•Changing student selection is a very cost-effective strategy to increase health professionals in under-served areas.

Partners

Kenya Medical Research Institute, NairobiKethi Mullei, Sandra Mudhune, Jackie Wafula, Catherine Goodman, Mike English

Centre for Health Policy, JohannesburgDuane Blaauw, Ermin Erasmus

International Health Policy Program, BangkokNonglak Pagaiya, Thinakorn Noree, Viroj Tangcharoensathien

London School of Hygiene and Tropical MedicineMylene Lagarde