81517
                  enGender Impact: The World Bank’s Gender Impact Evaluation Database



Using Performance Incentives to Improve Medical Care Productivity and Health Outcomes

      Author(s)            Paul Gertler, Christel Vermeerch
      Contact              cvermeersch@worldbank.org

      Country              Rwanda

  Organizing Theme         Health

       Status              Completed

Intervention Category      Pay for Performance

       Sector              Health, Nutrition & Population
                           We nested a large-scale field experiment into the national rollout of the introduction of
                           performance pay for medical care providers in Rwanda to study the effect of incentives for
                           health care providers. In order to identify the effect of incentives separately from higher
                           compensation, we held constant compensation across treatment and comparison groups –
      Abstract
                           a portion of the treatment group’s compensation was based on performance whereas the
                           compensation of the comparison group was fixed. The incentives led to a 20% increase in
                           productivity, and significant improvements in child health. We also find evidence of a
                           strong complementarity between performance incentives and baseline provider skill.
 Gender Connection         Gender Informed Analysis

  Gender Outcomes          Health of children

     IE Design             Clustered Randomized Control Trial (Clustered at District Level)
                           In 2005, the Rwandan Ministry of health scaled up a pay for performance scheme to the
                           national level. The P4P scheme provides bonus payments to primary care facilities based
    Intervention
                           on the provision of services and the quality of services. P4P payments represent an increase
                           of funding of 24.6% above the base budget.
 Intervention Period       Starting in 2006, bonuses were paid monthly.
                           The sample included 166 of Rwanda's 401 primary care facilities, 80 in treatment districts
  Sample population
                           and 86 in comparison districts.
Comparison conditions      The control group implemented the scheme 23 months after the treatment group.

   Unit of analysis        Children

  Evaluation Period        2006-2008, the most recent survey is 23 months after the initial intervention
                           Performance incentives translated into large and significant improvements in child health
                           outcomes. Incentives led to an increase of .53 std in weight-for-age of children aged 0-11
       Results             months and .25 std in height for age of children 24-49 months. Performance pay led to
                           substantial improvement in provider productivity as measured by knowledge of prenatal
                           care procedures. There are also large complementarities between provider skill and

   Last updated: 14 August 2013                                                                      1
                   enGender Impact: The World Bank’s Gender Impact Evaluation Database


                            performance incentives.


Primary study limitations   There is a small number of districts in the assignment.
                            Spanish Impact Evaluation Fund, Bank-Netherlands Partnership Program, the British
    Funding Source          Economic and Social Research Council, the Government of Rwanda, a Japanese PHRD
                            grant
                            Gertler, P., Vermeerch, C. (2012) "Using Performance Incentives to Improve Medical Care
      Reference(s)          Productivity and Health Outcomes," World Bank Policy Research Working Paper Series
                            No. 6100.
    Link to Studies
                            http://www.nber.org/papers/w19046
       Microdata




    Last updated: 14 August 2013                                                                  2