In this assessment of PBI for VHWs, we found overall satisfaction with the PBI system was high, though both VHWs and supervisors also reported specific concerns about the PBI system – most notably dissatisfaction with payment amounts. Attrition from the program was fairly low both before and after implementation of the PBI system. Our quantitative analysis of VHW earnings suggests that the PBI system operates fairly with respect to education level, gender, and proximity to the district hospital.
Previous research suggests that dissatisfaction with payment amounts represents one drawback of monetary incentives [5, 16, 17]. In this evaluation, both supervisors and VHWs felt that the PBI system distributed payments according to the amount of work accomplished. However, both VHWs and supervisors agreed that VHWs should be paid more. VHWs expressed dissatisfaction with both the overall level of payment as well as with the uncertainty of reward for case-finding (i.e., that screening did not always lead to case-finding). These concerns may reflect decreased intrinsic motivation in the setting of PBI. However, VHWs also expressed dissatisfaction with their pay under the prior incentive system (in which they were paid per household visited). Most VHWs in fact believed they were earning more under the PBI system, suggesting that this dissatisfaction with payment amounts is not specifically linked to a PBI system. We were unable to assess whether VHWs’ dissatisfaction with payment amounts adversely impacted the quality of their work, or merely represented an expression of the near-universal human desire to be paid more. However, the low level of attrition perhaps suggests that the VHWs’ dissatisfaction with payment amounts has not significantly impacted the success of the program.
In response to these results, DGH and KDH have added additional incentives for routine data collection and training performance, increasing the average payment level to $20 per month ($240 per year). Future program monitoring and evaluation will reveal whether this increase resolves VHWs’ concerns about payment amounts or merely establishes a new set point from which to bargain.
VHWs expressed mixed views about volunteerism in the program. VHWs felt they should be compensated well enough to ensure that the time they spent on their work as VHWs did not have a negative effect on their families’ wellbeing. However, they framed their request as a desire to paid at the same level as an unskilled farm laborer, suggesting some level of volunteerism, since they were not seeking payment at the level of the ‘market value’ for their skills. In some VHW programs, the desire for future, better-paid employment is a key motivating factor [13, 16]; this did not appear to be the case in the program evaluated here. All VHWs interviewed mentioned intrinsic motivations, such as knowledge gained and the satisfaction of seeing health improvements in their village, as reasons for their participation. However, the fact that VHWs are paid emerged as a potential area of friction between VHWs and their communities due to jealousy among some community members that the VHWs had secured part-time paid employment.
The PBI system requires close supervision of VHWs, which has the benefit of bringing clinicians into the communities. Close supervision also enables the continuous collection of well-validated data via the PBI system, creating the potential for rigorous program monitoring. The benefits of the PBI system described by supervisors and VHWs may have resulted in part from the closer supervision built into the PBI system. In one sense, this represents an obstacle to assessing the PBI system, since it is difficult to disentangle the effect of the PBI system from the effect of closer supervision. In another sense, as noted in the interviews with supervisors, closer supervision is facilitated by the PBI system, in that the PBI system provides greater structure for the supervisory role. Previous research suggests that close supervision is critical to VHW program functioning . Regardless of the incentive model selected, we suggest that VHW programs are unlikely to be successful without providing frequent supervision. In this light, it is encouraging that the actual incentive payments comprised less than 30% of total program costs.
The quantitative analysis of VHW earnings reveals several notable findings. VHWs with greater education earned no more than those with less education, suggesting that VHWs with limited formal education can master a PBI system. Contrary to our expectations based on gender inequality , female VHWs out-earned male VHWs under the PBI system. The success of female VHWs under the PBI system may stem from greater effort expended, or from the greater trust that other women place in them (since many of the incentives pertain to maternal and child health). We hypothesized two possible effects of proximity to the hospital: that VHWs in more distant villages would find it more difficult to refer patients for services mainly available at the hospital (such as family planning and cervical cancer screening), or that people in more remote villages would have less access to other health services and would rely more on their VHW. However, we did not find evidence of an association between proximity and VHW earnings under the PBI system.
These results suggest that the PBI system operates in a fair way, at least with respect to the demographic characteristics of VHWs and their villages. Payment amounts were higher for VHWs with more households. While this could suggest that it is easier for these VHWs to succeed under the PBI system, it seems more likely that VHWs serving more households must expend more time and effort in order to care for a larger population. Furthermore, if the PBI system acts as an accurate measure of performance, then VHWs in larger villages are simply being paid for greater performance (e.g., helping a greater number of people gain access to services such as improved sanitation, family planning, malnutrition treatment, etc.).
This work has a number of limitations. The most significant limitation is that we evaluated a relatively small VHW program over a relatively brief period of time, within a specific programmatic and cultural context. Thus, this process evaluation serves as an initial overview of a new system rather than any sort of definitive assessment. In addition, while this evaluation allowed us to assess issues such as the feasibility and acceptability of PBI for VHWs, it does not provide information on outcomes. Due to the data available, we are unable to assess the impact of the PBI system on utilization of priority health services in Kisoro relative to other possible payment systems. As the productivity data presented in the results demonstrate, the PBI system generates metrics (such as malnutrition case-finding and treatment) that could be used as outcome metrics. However, we did not have data for any of these metrics prior to the implementation of the PBI system, preventing a before-after comparison. A larger evaluation with the resources to prospectively track case-finding and utilization of care would be needed for an impact evaluation.
The specific components of this research also have their own limitations. Some VHWs have only basic literacy, and may have had difficulty in fully comprehending the questions on the survey. One potential limitation of the interviews is that they reflect the views of only 6 VHWs. Although there was substantial thematic overlap across the 6 interviews, it is possible they do not represent the full range of views. The complex issue of reflexivity also poses potential problems. The first author previously spent a year as a DGH volunteer in Kisoro, Uganda, working on a variety of community health programs including the VHW program. Thus, it is possible that VHWs and supervisors were reluctant to offer criticism to someone they viewed as a colleague. Alternatively, given the prevailing culture of politeness and group solidarity in Kisoro, it is also possible that VHWs and supervisors would be more willing to offer criticism to someone they knew than to an outsider. It is also possible that the VHWs viewed the surveys and interviews as an opportunity to lobby for greater pay, and exaggerated that concern. The small number of VHWs in this program also limited the power of statistical tests used in the quantitative analysis. This process evaluation also did not include any evaluation of community perspectives on the program and the PBI system.
Further research is also needed to assess the long-term impact of PBI incentives, and particularly whether they diminish intrinsic motivation over time and, if so, whether diminished intrinsic motivation harms program functioning. Considerable evidence in the psychological and economic literature suggests that extrinsic rewards undermine intrinsic motivation [26, 27]. However, it is unclear whether this research applies to long-term employment, or whether different payment schemes for VHWs (e.g., salary vs. PBI) would have differing effects on intrinsic motivation. An additional concern presented in the literature suggests that PBI may lead to distortions, in which workers focus on incentivized activities and neglect non-incentivized activities [28, 29]. While this evaluation did not specifically address this issue, we suggest that distortions are less likely in PBI for VHWs than for other healthcare workers, since the more limited scope of practice of VHWs makes it easier to create a comprehensive PBI system that incentivizes many if not all of the desired activities of VHWs. In addition, if incentive amounts for different accomplishments are accurately calibrated with the health benefits, then VHWs will have an incentive to allocate their effort appropriately. Incentive amounts can also be adjusted at different times of the year based on seasonal variation in health needs, e.g., increasing incentives for malaria prevention and diagnosis in the rainy seasons. Incentive amounts can easily be adjusted upward over time to keep pace with inflation or to reward increasingly long-serving and experienced VHWs, just as salaries might increase over time. However, assessment of these issues will require long-term follow-up.