The Mayer Hashi Large-Scale Program to Increase Use of Long-Acting Reversible Contraceptives and Permanent Methods in Bangladesh: Explaining the Disappointing Results. An Outcome and Process Evaluation

BACKGROUND Bangladesh has achieved a low total fertility rate of 2.3. Two-thirds of currently married women of reproductive age (CMWRA) want to limit fertility, and many women achieve their desired fertility before age 30. The incidence of unintended pregnancy and pregnancy termination is high, however. Long-acting reversible contraceptives (LARCs), consisting of the intrauterine device and implant, and permanent methods (PM), including female sterilization and vasectomy, offer several advantages in this situation, but only 8% of CMWRA or 13% of method users use these methods. PROGRAM The Mayer Hashi (MH) program (2009-2013) aimed to improve access to and the quality of LARC/PM services in 21 of the 64 districts in Bangladesh. It was grounded in the SEED (supply-enabling environment-demand) Programming Model. Supply improvements addressed provider knowledge and skills, system strengthening, and logistics. Creating an enabling environment involved holding workshops with local and community leaders, including religious leaders, to encourage them to help promote demand for LARCs and PMs and overcome cultural barriers. Demand promotion encompassed training of providers in counseling, distribution of behavior change communication materials in the community and in facilities, and community mobilization. METHODS We selected 6 MH program districts and 3 nonprogram districts to evaluate the program. We used a before-after and intervention-comparison design to measure the changes in key contraceptive behavior outcomes, and we used a difference-in-differences (DID) specification with comparison to the nonprogram districts to capture the impact of the program. In addition to the outcome evaluation, we considered intermediate indicators that measured the processes through which the interventions were expected to affect the use of LARCs and PMs. RESULTS The use of LARCs/PMs among CMWRA increased between 2010 and 2013 in both program (from 5.3% to 7.5%) and nonprogram (from 5.0% to 8.9%) districts, but the rate of change was higher in the nonprogram districts. Client-provider interaction and exposure to LARCs/PMs were lower in the program than nonprogram districts, and the MH program districts had higher vacancies of key providers than the nonprogram areas, both indications of a more difficult health system environment. CONCLUSION The weaknesses in the health system in the MH districts apparently undermined the effectiveness of the program. More attention to system weaknesses, such as additional supportive supervision for providers, might have improved the outcome.