Assessing technical efficiency of HIV prevention interventions in three sub-Saharan countries

Assessing technical efficiency of HIV prevention interventions in three sub-Saharan

countries

S Bautista-Arredondo, SG Sosa-Rubí, D Contreras-Loya, M Opuni, A Kwan, C Chaumont, J Condo, N Martinson, J Coetzee, F Masiye, S Nsanzimana, J Wang'ombe,

K Dzekedzeke, O Galarraga, and R Wamai on behalf of the ORPHEA study team

July · 2014

Motivation

- Need for implementing HIV programs with higher efficiency- Maximizing value for money

- Lack of data on updated performance in the region- Previously published evidence suggested enormous

heterogeneity in HIV prevention costs and potential waste (PANCEA, 2002)

- Need to understand:- Current levels of efficiency- Determinants of more efficient performance

Information needs for optimizing HIV programs

- Allocation among interventions- Effectiveness data

- Allocation among populations/groups- Epidemiological and behavioral data

- Allocation among health inputs- Performance data (M&E)- Determinants of efficiency- Interventions to improve efficiency

- Which incentives work better and are more cost-effective?- How can M&E systems and changing in management practices can facilitate

more efficient results

THE ORPHEAPROJECT

Aims

• Research question– Which characteristics predict the most efficient performance in

the delivery of HIV services?

• Objectives– Measure and explain efficiency:

- To estimate the total costs and the average cost per output, at the facility level

- To estimate levels and determinants of efficiency

– Provide recommendations

HTCHIV testing and counseling

PMTCTPrevention of Mother-to-child

Transmission

Key hypotheses

- Heterogeneity of unit costs- High variability on average cost per service across facilities

- Possible to identify the role of determinants and constraints- Modifiable characteristics that predict higher efficiency - Environment in which facilities operate and make decisions - Not possible to

modify through interventions

- Overlap between economics and management  - Looking at performance at the facility level: potential for improving efficiency

METHODS

Measuring Efficiency

• Four HIV prevention interventions: HTC, PMTC, MC, FSW

• Four African Countries: Kenya, Zambia, South Africa, Rwanda

• Outputs: all services produced in the previous fiscal year

• Inputs: staff, essential recurrent inputs and services, capital, training and

supervision

• Managerial and environmental characteristics: describing the environment

and constraints in which production decisions are made

- Identify constraints and determinants

8

Constraints from the firm’s perspective in the short term

- Country/Location- Urban vs. rural setting- Funding sources- Facility type / Ownership - HIV/AIDS prevalence - Size of demand- Supply of services (utilities)

Determinants, can be adjusted at the facility-level

- Structure and governance- Training and staff composition- Management - Accountability- Incentives- Sanctions

Determinants of efficiency and constraints to more efficient performance

Microeconomic approach

- Micro-costing- One-year retrospective data collection - Effort to measure staff’s time allocation (Time-motion)- Measurement of quality using exit interviews, clinical vignettes

and the cascade approach- Data collection at different levels:Facility-level information

- Staff roster- Drugs and supplies- Utilities- Equipment and buildings

District-level information

- Training - Supervision

National-level information

- Salaries- Prices of supplies (HIV test

kits, ART)

Measuring quality

- Process quality using clinical vignettes and exit interviews

- Try to capture quality of the program through the outcome measures using a

“cascade” approach

- Reflect definition of “comprehensive” prevention packages

- Reflect hierarchy or sense of “effective” coverage

- Assumption: higher quality of services can be captured by higher success of programs in

achieving effective coverage

- Example: PMTCTPregnant women

tested for HIV

Pregnant women tested and positive for

HIV

Pregnant, HIV-positive women linked to ART

Estimating efficiency

- Estimation of total annual input costs – at the facility level

- Estimation of unit average cost per services along the HIV prevention

services cascade

- Correlation of unit average cost vs. scale of production, controlling for

quality

- Estimation of cost functions using a translog specification

- Include determinants and constraints of efficiency in a joint equations system

- Technical efficiency analysis using DEA or other methodology

RESULTS

Kenya, Rwanda and Zambia

Unit cost breakdown

HTC PMTCT

HTC Staff Composition PMTCT Staff Composition

Cost per client across the service cascadeHTC PMTCT

Average cost vs scale for two stages in the cascade

ORPHEA: Policy Implications

- Assessing the determinants of efficiency- Weak evidence of economies of scale in the first stage, much stronger in

the second stage- Supervision seems to have an important role increasing efficiency- Incentives and complex governing structures increase costs- Our results suggest that quality of services is not the most important

predictor of efficiency

- Three promising approaches- Measuring performance at the clinical level and revealing disparities- Fairly simple management training and interventions- Looking into the production function of services: staff compositions

Acknowledgements

- The ORPHEA study is supported by the Bill and Melinda Gates Foundation.

- We gratefully acknowledge the collaboration of our academic partners in Kenya, Rwanda, South Africa, Zambia and the United States.