Incentives to change: effects of performance-based financing on health workers in Zambia

Clinical health worker encompasses 12 categories: clinical officers, dental surgeons, doctors, nutritionists, lab scientists, pharmacists, physical therapists, radiographists, midwives, nurses, environmental health personnel, and oral health. Clinical health workers are not the same as administrators, who also work in health facilities.

Poor staff morale and weak incentives lead to emigration or migration, which in turn result in a health worker shortage [54, 55]. An initial wave of Zambian health workers migrated to countries in sub-Saharan Africa, such as South Africa, Botswana, and Namibia [56]. Subsequent waves migrated to Europe, North America, Australia, and New Zealand. An exodus of health professionals has also been observed within Zambia from rural to urban areas, from the public to the private sector, and from curative to preventive care [57, 58].

PBF is a strategy to address inadequate performance-reward linkages and, more generally, health system reform [40, 59]. In this paper we adopt Soeter et al.’s definition of PBF as “fee-for-service conditional on quality of care” [60].

Monetary incentives (i.e., bonus payment) can either be awarded for achievement of predefined performance targets or withheld if targets are not reached. Individuals and organizations may also respond to non-monetary incentives of PBF schemes such as enhanced supervision, data system development, and institutional rankings [61, 62].

A wide array of moderators on the “meso-levels,” or intermediary levels, that could weigh on PBF’s influence on worker outcomes, which we acknowledge them though not account for them explicitly in this study. Beyond concerns over PBF implementation is fungibility, i.e., whether health managers allocate revenue to health worker bonuses or to other assets of equivalent value, say, equipment, supplies, and capital improvements to the health care facility. Toonen et al. considered autonomy, management capacity, and an understanding of PBF concepts to be important for implementing PBF programs in sub-Saharan Africa’s health sector [63]. Mohammed et al. urged a closer look at the “know–do” gap, or the gap between what health workers know how to do—and actually do (Mohammed RL, Herbst C, Leonard K, Goldberg J.: Crossing the Three-Gap Divide with PBF, unpublished). PBF interventions can reduce this gap through improved accountability and supervision structures, and more generally via an improved working environment.

Eisenberger et al. argued that rewards could have a positive, negative or null effect on motivation depending on the nature of the task assigned in controlled laboratory settings [64]. Woolhandler et al. further questioned the causal direction posed by rewards on motivation in health care settings in high-income countries [65].

We make this assertion following Bhatnagar and George’s observations in Nigeria [66] and Kalk et al.’s observations in Rwanda [67].

Job satisfaction is workers’ personal satisfaction relative to their work situation [68]. The correlation between income and subjective well-being has been observed both within and across countries [69]. Yet job satisfaction is a multi-faceted concept. Heneman et al.’s results indicated a positive relationship between pay-for-performance perceptions and pay satisfaction [70]. Judge et al.’s meta analysis results suggested that pay level is only marginally related to job satisfaction [71]. Job satisfaction is not just about job conditions but also personality. In an earlier study, Judge et al. found that traits of core self-evaluations, or positive self-concept, are positive dispositional predictors of job satisfaction [72]. The PBF scheme in Zambia primarily responds workers’ satisfaction with their compensation, but it might also respond to the other six aspects of job satisfaction we measure. Therefore, like motivation, we hypothesize that PBF has a positive effect on job satisfaction. We further hypothesize that enhanced financing would also have a positive effect on job satisfaction, but with a lower magnitude than PBF because enhancing financing primarily addresses one aspect of job satisfaction, or satisfaction with work conditions.

Porter et al.’s longitudinal study results further suggested that satisfaction with pay is most acute at time points closest to when psychiatric technicians intend to leave their organization [73].

We expect to see stronger effects on HRH results for the intervention (PBF) group than for the control 1 (C1; enhanced financing) group, but the direction is expected to remain the same for the following reason: RBF can be interpreted by health workers as a reward for their individual efforts, whereas enhanced financing can be interpreted by them as a signal of recognition for collective efforts through improvements on working conditions. Both incentive schemes are expected have positive, knock-on effects on HRH outcomes compared with the control 2 (C2; pure control) group.

The amount of bonus received by each staff member was dependent on a number of factors: individual performance scores taken during a performance appraisal, actual PBF income made, investment priorities, the number and composition of staff at the health center, and individual government salary levels. The ratio was higher at the start of the PBF project but dropped after about 6 months of the PBF project due to an increase in government salaries ranging from 100 to 200% [27]. Nonetheless, PBF staff bonuses led to an absolute increase in the personal income for staff but by different margins/percentages across staff and health centers.

It should also be noted that, unlike the PBF districts, health centers in C1 districts did not have devolved autonomy, enhanced supervision, training, access to PBF reference materials, and data monitoring and verification that were also part of the PBF intervention.

The surveys collected information on human resources and physical capacity, facility governance, practitioner knowledge, outreach activities and other initiatives, and quality of care and practitioner behavior through patient exit interviews.

At baseline, health centers and DMO offices in Gwembe (PBF group), Itezhi-Tezhi (C1 group) and Mazabuka (C2 group) in the Southern Province of Zambia were included in the sample. Southern Province reflects the median levels of socio-economic status and health indicators across all the ten Zambian rural provinces, so focusing data collection efforts on this province enabled easier identification of cross-cutting issues for PBF and human resources. Five additional interviews with conducted with Provincial Medical Officers in Livingstone. At endline, health centers in four districts (Isoka, Gwembe, Itezhi-Tezhi, Mazabuka) were purposely sampled based on a multitude of characteristics such as remoteness, size of catchment area, urban/rural split, performance, and staff-mix.

Cognitive evaluation theory (CET) predicts that extrinsic rewards would diminish, if not displace, intrinsic interest [36]. Deci et al.’s meta analysis concluded that extrinsic awards decreased intrinsic motivation regardless of the form of incentive: engagement-, completion-, or performance-contingent awards [39]. Decreased intrinsic motivation negatively affects worker autonomy, purpose, altruism and competence [74]. Huillery and Seban found lower staff attendance and on-the-job effort after the PBF pilot ended in the Democratic Republic of Congo, which they attribute to not only lowered motivation, but a switch from intrinsic to extrinsic motivation on the workers’ behalf [75]. However we do not find evidence in support of CET.

Likewise, Brock et al. have found that peer scrutiny and encouragement alone can determine quality of care provided by clinicians in their sample [76].

During administrative audit and quality assessment visits, the verifiers would look at all aspects of PBF program implementation and see how health centers were performing against standards and benchmarks. After their visit, the verifiers would debrief health center staff on their strengths and weaknesses, and with the health center cadres devise potential solutions.

PBF funds were determined from performance and then disbursed from the MoH directly into each health center’s bank account. PBF funds did not pass through the DMO. The health cadres of Isoka and Gwembe districts, both PBF groups, had high involvement in prioritizing needs and use of PBF money without having to wait for the DMO to tell them what to do.

The amount should be 40%.

Health staff who were engaged in income-generating activities in order to augment their salaries has also been observed in Democratic Republic of Congo [77], in Sierra Leone [78], and in Tanzania [79].

For example, Zambia raised salaries for all government workers in 2011 and in 2013 by between 100 to 200% which had an effect on the size of the staff bonus.

For example, health workers from either of the control groups may have sought employment in one of the health centers in the PBF group. Health workers across the three study groups could also have been sharing information about how to generate additional income from means other than the PBF program [80].