Menu
Menu
Menu
Menu
Background: Evidence emerging from qualitative studies suggests the existence of substantial variation in how health workers experience performance-based financing (PBF) within the same setting. To date, however, no study has quantified or systematically explored this within-setting heterogeneity. Considering that differences in health workers’ affective reactions to PBF likely constitute an important element mediating the effectiveness of PBF in improving health service delivery, systematic and tangible information will be highly valuable to policy-makers and program managers who aim to maximize positive impacts of PBF. Our study aimed at contributing to filling this gap in knowledge by quantifying health workers’ knowledge of, satisfaction with, and perceptions of PBF in Burkina Faso, and exploring factors associated with heterogeneity therein. Methods: The study employed a post-intervention cross-sectional explanatory mixed methods study design with a dominant quantitative component – a structured survey to a total of 1314 health workers from 396 intervention health facilities – and a small and focused qualitative component – key informant interviews with 5 program managers – to triangulate and further elucidate the quantitative findings. Quantitative data were analyzed descriptively as well as using 3-level mixed-effects models. Qualitative data were analyzed in a largely deductive process along the quantitative variables and results. Results: Health workers were on average moderately satisfied with PBF overall, with a slight tendency towards the positive and large variation between individuals. Two-thirds of health workers did not have adequate basic knowledge of key PBF elements. Perceived fairness of the performance evaluation process, of the bonus distribution process, and satisfaction with the individual financial bonuses varied dramatically between respondents. Factors associated with heterogeneity in knowledge, satisfaction, and fairness perceptions included responsibility at the facility, general work attitudes, management factors, and training in and length of exposure to PBF. Conclusion: Findings imply that investments into staff training on PBF and manager training on organizational change processes might be beneficial to positive staff attitudes towards PBF, which in turn would likely contribute to improving the effectiveness of PBF.
Despite improvements over the last years, Burkina Faso continues to suffer from a high burden of morbidity and mortality, with a maternal mortality ratio of 371 per 100000 live births and an under-five mortality rate of 88.6 per 1000 live births (2015). 21 Health services are provided primarily by the public sector in a multi-tier district health system. 22 Health facilities upkeep their operations through a mix of government in-kind inputs and revenues from user fees and drug sales. 23 Formal healthcare service utilization rates have improved substantially in recent years, but remain below target. 24 Quality of health services, however, is often substandard 25-27 for reasons including low pay, substandard infrastructure and equipment, poor supervision, shortages in drugs and other supplies, and few incentives for high performance. 22,28-30 Against this background, PBF was first introduced in 2011 as a pilot scheme in 3 health districts to improve access to and quality of care. Given an initially promising evaluation, 31 PBF was scaled up to another 12 districts between 2014 and 2018, implemented by the Ministry of Health (MoH) with financial support by the World Bank’s Health Results Innovation Trust Fund. The intervention and its background and context are described in detail elsewhere. 32,33 Although the primary objective was to improve utilization and quality of maternal and child health services, the intervention effectively included a broad range of primary- and secondary-level services, including also curative care, TB, and HIV services. In brief, health facilities signed contracts with the MoH stipulating the services purchased by PBF, a comprehensive list of quality indicators, and payment modalities. Facilities reported volume of provided services on a monthly basis. Reports were then verified by an external agency and facilities subsequently paid a pre-defined amount (‘subsidies’) for each service provided. Subsidies per provided service ranged from 100 FCFA (≈ 0.15 EUR) for curative outpatient consultations to 8500 FCFA (≈ 13 EUR) for a cured tuberculosis case. Facilities were further categorized into 9 equity categories based on staffing levels and remoteness, and less privileged facilities received proportionally higher subsidies. Quality was verified by the District Health Management Teams on a quarterly basis. If quality scores surpassed 50% (later changed to 60%) of the maximum, facilities were paid a quality bonus proportional to their service volume and quality level. PBF payments came on top of pre-existing financing structures. Initially, facilities were free to spend PBF funds as they wished, for facility-related investments or as staff bonuses. From October 2016 on, to encourage more intensive investments, staff bonuses were limited to 60% of the revenue from PBF, whereas at least 40% had to be invested to improve the infrastructure or equipment of the health facility. Facilities were provided with a financial management tool called ‘outil d’indice.’ This also included a calculator to determine bonus amounts for individual staff members, based on 5 criteria. In some health facilities, following a randomization process in the context of an impact evaluation, 32 the standard PBF was further complemented with measures intended to increase equity in impact. The impact evaluation of the extended PBF trial showed limited overall effects of PBF, with positive impact only on the utilization of facility-based delivery and postnatal care as well as on certain input dimensions of quality of care, but no impact on the utilization of other services or process quality. 34 A process evaluation of the first twelve months of implementation underlined that although the intervention was implemented as planned in most respects, there were a number of important challenges, most notably delays in setting up the verification process and in payment of the subsidies. 35,36 We used a post-intervention cross-sectional explanatory mixed methods study design with a dominant quantitative component and a small and focused qualitative component. The quantitative component employed a structured survey to health workers in all intervention health facilities to quantify the elements printed in bold in Figure 1, namely health workers’ satisfaction with PBF overall as well as knowledge and perceptions related to the key issues having emerged repeatedly in previous research, performance evaluation and individual bonus payments. The quantitative survey further served to quantify associations with key individual- and facility-level determinants. The qualitative component employed key informant interviews with program managers to triangulate and further elucidate the quantitative findings. It also served to capture factors and dynamics which we had not included in the quantitative survey, allowing us to place quantified associations into context. Qualitative interviews were performed after a descriptive analysis of the quantitative data, and results then used to further inform quantitative analyses of heterogeneity in knowledge, perceptions, and satisfaction. Specifically, results from the qualitative study component led us to obtain and include in the final models additional quantitative data on facility performance as described in more detail below. Quantitative data were collected in the context of the above-mentioned impact evaluation. The study design and sampling procedures are described in detail in De Allegri et al. 32 In brief, the study included all 396 primary-level healthcare facilities in all 12 purposely selected intervention health districts that newly received PBF in 2014. In line with the specific objectives set for the study presented in this paper, we only used endline data, collected between April and June 2017, approximately 3 years after the introduction of PBF. In each health facility, we included all clinical skilled personnel who had worked at the health facility for at least 3 months and who were present on the day of the study team visit, resulting in a total of 1314 health workers (health workers per facility: mean = 3.3, sd = 1.7, min = 1, max = 11). Table 1 provides an overview over the distribution of basic demographic and PBF-related characteristics in the sample. Abbreviations: PBF, performance-based financing; AIS, Agent Itinérant de Santé (preventive services and outreach); SD, standard deviation. a Nurse: Infirmier Diplômé d’Etat, Infirmier breveté; Midwife: Sage-Femme d’Etat/Maïeuticien d’Etat; Assistant midwife: Accoucheuse Brevetée, Accoucheuse Auxilliaire. Data was collected with a French-language structured survey administered to all sampled health workers by trained interviewers. The survey assessed overall satisfaction with the PBF intervention as well knowledge and perceptions of the performance evaluation process and the individual incentives as outlined above (6 variables in total, referred to as “outcome variables” in the following). The questionnaire also included questions on demographics, working conditions and perceived working environment, motivation, and clinical knowledge. Questionnaire sections pertaining to satisfaction, attitudes, perceptions, and other psychological aspects were administered in the hybrid mode described in Lohmann et al, 37 whereby interviewers read questions, statements, and answer options to the respondents, but respondents entered their answers themselves into the tablet computers used for data collection so as to maximize perceived confidentiality and reduce answer biases. We extracted data on facility catchment population, staffing levels, and patient numbers from a facility assessment also conducted within the context of the impact evaluation. To complement the quantitative analysis, we further obtained program data on facility performance on quality indicators and on facility equity categories. Outcome variables as well as potential determinants of heterogeneity are aligned with the conceptual understanding described earlier and detailed in Table 2. Abbreviations: PBF, performance-based financing; AIS, Agent Itinérant de Santé (preventive services and outreach). a Only health workers who reported to know the last evaluation results were asked to judge on its fairness b 27% of the sample (distributed across all cadres, responsibility levels, genders, etc) reported not to receive any bonus payments. However, since the question might have been misunderstood to exclude PBF bonuses, we included in the results shown in Figure 3a only those respondents who reported to receive bonus payments. We first performed descriptive analyses of each of the 6 outcome variables. For each, we then employed 3-level (individual, health facility, district) mixed-effects linear (for Likert-type variables as per standard psychometric practice 39 ) or logistic (for dichotomous variables) regression to explore determinants of heterogeneity, using the ‘mixed’ and ‘xtmelogit’ commands in Stata 14.2, respectively. Specifically, we modeled associations of the outcome variables with observed individual- and facility-level factors at level 1 as fixed effects, and further accounted for the organizational environment by modeling facility and district random intercepts at levels 2 (health facility) and 3 (district). To triangulate and validate the quantitative findings and to better understand observed heterogeneity in PBF knowledge, perceptions, and satisfaction, we performed key informant interviews with the 5 program managers in the MoH PBF unit who had followed program implementation from the start. We opted to interview program managers rather than health workers as in their supervisory role, they were in constant contact with health workers enrolled in PBF and therefore had the best possible oversight over the spectrum of PBF knowledge, perceptions, and satisfaction among the health workforce. The first and the second author conducted all interviews in French, adopting a strategy previously agreed upon by all authors. Respondents were shown the quantitative results presented in Figures 22–-44 and asked to comment on them, with interviewers probing for more in-depth information where necessary (“Does this surprise you in any way?”; “Does this correspond to what you have experienced on the ground, or did you have different perceptions?”; “From your perceptions on the ground, what were the reasons for these variations?”). Interviews were audio recorded and verbatim transcribed. Written informed consent was obtained prior to each interview. Distribution of Respondents’ Scores Pertaining to Their Overall Satisfaction With PBF. Abbreviation: PBF, performance-based financing. (a) Proportion of respondents having correctly recalled the result of the last performance evaluation result. (b) Distribution of respondents’ perceived fairness scores regarding the performance evaluation process. (a) Proportion of respondents having correctly recalled the bonus distribution modalities. (b) Distribution of respondents’ perceived fairness scores regarding the bonus distribution process. (c) Distribution of respondents’ scores pertaining to their satisfaction with their individual bonuses. The first and second author independently coded the French material in a mostly deductive process along a predefined codebook, with initial codes that mirrored the quantitative variables in Table 2. The 2 authors further integrated a few new codes that emerged in vivo as they proceeded through the transcribed material. The independent analyses advanced by the 2 authors were discussed among all authors and minor discrepancies in emerging interpretations resolved by referring back to the data and/or by relating findings to the context of the intervention. Quotes illustrating main findings were selected and translated from French to English for the purpose of publication.
N/A
