Technical efficiency analysis of the Basic Health Care Provision Fund for the funded primary health care facilities in Nigeria

Abstract

Introduction

The National Health Act of 2014 established the Basic Health Care Provision Fund (BHCPF) to strengthen Nigeria’s primary healthcare (PHC) delivery system. With over $190 million disbursed, we conducted this study to assess the technical efficiency of the BHCPF-funded PHC centers in Nigeria.

Methods

We obtained data from a national PHC assessment conducted in 2022 led by the National Primary Healthcare Development Agency (NPHCDA). A two-stage data envelopment analysis was conducted among 3531 BHCPF-funded PHC facilities in Nigeria. In the first stage, the constant return to scale (CRS) and variable returns to scale (VRS) models were used to evaluate the technical efficiency of the PHCs using an input orientation. We used two inputs (number of clinical staff and number of facility inpatient beds) and four outputs (number of fully immunized children, outpatient attendance, antenatal care attendance, and facility deliveries). In the second stage, a Tobit regression model was used to assess the effect of the per-capita subnational allocation to health (PSAH), catchment area’s population (CAP), patient-reported quality of care (PRQoC), and patient-reported post-partum and newborn care (PPNC) on inefficiency among the PHCs.

Results

The mean CRS and VRS efficiency scores of the 3,531 facilities were 11% (± 12%) and 30% (± 21%), respectively, with efficiency scores ranging from 1.2% to 100% for CRS and from 5.1% to 100% for VRS. The Tobit regression model showed that the exploratory variables assessed explain only 0.17% and 0.31% of the inefficiencies recorded by the CRS and VRS models among the PHC facilities. The Tobit regression model also showed that PPNC is significantly positively associated with VRS inefficiency, while the CAP and PSAH are significantly negatively associated with inefficiencies of health facilities. PRQoC was, however, not a significant determinant of the inefficiencies of health facilities.

Conclusion

Our study concluded that the level of efficiency of BHCPF-funded PHCs in Nigeria is very low. To improve the impact of BHCPF on Nigeria’s PHC system, there needs to be increased emphasis on efficiency. Efforts to improve efficiency should be consolidated by institutionalizing measurement of health facilities’ efficiency and tracking progress towards attaining health systems goals and universal health coverage.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12875-026-03284-8.

Background

Primary healthcare service delivery is the bedrock of health systems serving as the entry point to the health system in many countries [1]. Due to this critical role, a well-functioning national primary healthcare (PHC) system is desirable, especially for many developing countries, including Nigeria [2]. PHC in Nigeria is delivered through a mix of publicly and privately owned health facilities, though public health facilities are dominant [3, 4].

Despite the critical role it plays, the primary healthcare system in Nigeria is faced with complex challenges across all the building blocks of the health system – leadership and governance, health financing, human resources for health, health informatics, essential medicines and technologies, and service delivery [5]. Over the past decade, concerted efforts have been made by the government and other stakeholders to strengthen the national primary health care system through policies like; primary health care under one roof, minimum standards for PHC services, and the Basic health care provision fund(BHCPF) [6]. The BHCPF was created by the National Health Act of 2014 as “an earmarked fund for health to be financed with not less than 1% of the Federal Government’s Consolidated Revenue Fund (CRF), in addition to contributions from donors, the private sector, and other sources” [7]. The fund seeks to strengthen the demand for, and supply of PHC services, especially those related to reproductive, maternal, newborn, child, and adolescent and nutrition services. It aims to improve the demand for PHC services through a basic minimum package of health services and the supply of PHC services, through funds disbursement for the provision of essential drugs, vaccines, and consumables, maintenance of facilities, laboratories, equipment, and transport. It was also designed to strengthen the development of Human Resources for Primary Health Care for eligible health facilities [8].

To become eligible to provide BHCPF services, healthcare providers were accredited based on the minimum standards for primary healthcare centers, as defined by the National Primary Healthcare Development Agency (NPHCDA) and the National Health Insurance Authority (NHIA) [8]. The fund’s objective is to ensure that at least one functional primary health care supported by the fund exists across all the political wards in the country [8]. A political or electoral ward in Nigeria is an officially recognized sub-division of a Local Government Area (LGA), and it represents the smallest unit of electoral administration in the country. Each LGA is divided into multiple wards, and each ward contains several settlements or communities. Overall, improved financing through BHCPF is expected to translate into better health outcomes, especially for maternal and child health in the country [9]. However, without a focus on efficiency, these increased investments in primary health care through BHCPF will not translate into the best health outcomes in the country, which can further lead to a waste of scarce resources [10]. Hence, maximizing the efficiency of these facilities will improve care delivery with existing resources, and ensure that the objectives of BHCPF are fully realized.

Efficiency measurement in healthcare is a performance measurement tool that appraises the translation of healthcare inputs into ensuring health sector goals are met in the form of outputs through the production process by healthcare facilities or systems [11]. Three types of efficiency measures are commonly used in healthcare – technical efficiency, allocative efficiency, and cost (economic) efficiency [12]. Allocative efficiency in healthcare is concerned with the best combination of inputs to produce the health outputs in a healthcare setting. Cost efficiency on the other hand is concerned with getting benefits at the lowest possible cost, while technical efficiency deals with using a given set of healthcare inputs is used to produce the maximum set of outputs [13, 14]. The focus of most efficiency measurements in healthcare is on technical efficiency, because of its ease of measurement and ability to evaluate different units or systems and compare them [15, 16].

Different methods have been used in the technical efficiency measurement of health facilities and health systems. They include ratio analysis, econometric analysis, stochastic frontier analysis (SFA), and data envelopment analysis (DEA). Two commonly used techniques for measuring technical efficiency in healthcare are Stochastic Frontier Analysis (SFA) and Data Envelopment Analysis (DEA). SFA is a parametric method that estimates a statistical production frontier and separates inefficiency from random error. DEA, on the other hand, is a non-parametric method that constructs an efficiency frontier based on observed input and output data, allowing comparison across multiple facilities without requiring a specific functional form [17, 18]. In a systematic review of efficiency measures of health facilities in low and middle income countries (LMICs), DEA accounted for 59% of efficiency studies, especially for studies published after 2000 [19]. Since its development in the 1970s, DEA has been widely utilized in healthcare. DEA assesses the efficiency of each health facility, referred to as a Decision-Making Unit (DMU) by comparing how well it converts inputs (e.g., staff, beds) into outputs (e.g., immunizations, deliveries). DEA constructs a theoretical maximum performance line called the Production Possibility Frontier (PPF). Facilities on the frontier are considered efficient; those below it are inefficient [20]. It is widely used in healthcare because of its relative ease of use as no production function is specified and the model can consider multiple inputs and outputs at the same time [21]. SFA on the other hand uses a stochastic frontier whose shape is determined by data with some deviations attributed to random factors [22, 23].

There have been reports from several studies that applied efficiency measurements in healthcare at the health systems and health facilities levels. In sub-Saharan Africa, some studies have reported the use of efficiency measurements to analyze healthcare facilities across all levels of care delivery in the public and private sectors [24–27]. In Nigeria however, there is a dearth of information on efficiency studies. Although, Ichoku et al., 2011 measured the technical efficiencies of secondary health facilities in two Southeastern states in Nigeria [28], however, Sede and Ohemeng, 2012 only evaluated the technical efficiency of a mix of selected tertiary and secondary health facilities across the six geopolitical zones of the country [29]. Adejoh et al., 2022 on the other hand conducted a cross-efficiency analysis of selected PHCs in Lagos state [30].

In addition to measuring technical efficiency, understanding the contextual and organisational factors that influence it is critical for identifying targeted strategies to improve facility performance and resource use. Prior studies have shown that efficiency at the facility level can be affected by factors such as catchment population size, patient satisfaction, and investments in service delivery. Building on this evidence, our study incorporates four contextual variables drawn from the PHC dataset: catchment population size, patient-reported quality of care, patient-reported maternal and newborn care, and state-level per-capita health allocations. These variables were selected for their theoretical relevance and data availability and are used in a second-stage Tobit regression analysis to explore drivers of inefficiency.

Accordingly, the study aims to:

• i.Assess the technical efficiency of BHCPF-funded PHC centers in Nigeria.
• ii.Evaluate how contextual factors (catchment population size, patient-reported quality of care, patient-reported maternal and newborn care, and state-level per-capita health allocations) influence inefficiency among these facilities.

Methodology

Setting

Nigeria is Africa’s most populous country with an estimated population of 200 million people, and a gross domestic products (GDP) per capita of $5,408 [31]. Nigeria is classified as a lower middle-income country and it is administered by three levels of government – national, state, and local governments [32]. The health system in Nigeria has a tri-partite structure comprising primary, secondary, and tertiary healthcare facilities. Local governments are constitutionally responsible for the management and financing of PHCs in Nigeria; though they have been historically supported by the Federal and State governments, as well as local and international donors and partners [32, 33].

Nigeria is comprised of 774 Local Government Areas (LGAs), each with an average of 10 to 15 political wards. PHC facilities are distributed across all LGAs, with at least one BHCPF-funded facility targeted per ward under the fund’s design. As defined by the minimum standards for PHCs in Nigeria, there are three levels of PHC facilities: health post, level 1 PHC (PHC clinic), and level 2 PHC (PHC centers) [34]. Health posts are facilities that serve about 500 people and are often located in a single dispersed village, PHC clinics (level 1 PHC) serve a group of villages with an average catchment size of 2,000–5,000 persons while PHC centers (level 2 PHCs) also referred to as comprehensive health centers serve a catchment area of about 8,000 people. The BHCPF-funded facilities are level 1 and level 2 PHCs.

Data

This study was based on secondary analysis of data collected through the 2022 Comprehensive National PHC Assessment. The assessment was conducted by the Federal Ministry of Health (FMoH) and the National Primary Health Care Development Agency (NPHCDA), with support from partners led by Sydani Group. The dataset includes facility-level information on service delivery, human resources, infrastructure, and patient-reported experience measures, collected from PHC facilities across all 36 states and the Federal Capital Territory. A total of 3,959 BHCPF-funded PHC facilities with complete data for the second quarter of 2022 were initially included in the analysis. As the dataset comprised all eligible and reporting facilities nationwide, no sampling was applied. Following the exclusion of facilities with incomplete data and outlier values, variables of interest, including inputs, outputs, and contextual factors were extracted for 3,939 facilities. The distribution of health facilities by level is presented in (Table 1).

Table 1.

Number of BHCPF facilities by level

Level of PHC	Number of BHCPF Facilities
Level 1 PHC	408
Level 2 PHC	3531

Selection of variables

A disaggregated selection of input variables was employed in this analysis because of the difficulty in aggregating and costing for all inputs. Inputs were selected based on relevance, use in DEA studies, and availability across facilities. The number of beds in a healthcare facility shows the size of the PHC facility and serves as a good proxy for capital inputs [35]. All BHCPF facilities are expected to have inpatient wards for prenatal, postnatal, and delivery services. The number of staff is a good proxy for labour inputs in a DEA [36], especially where it is demanding to obtain the monetary value for the labour cost. It also ensures uniformity when there are differences in workers’ wages and remuneration across different facilities, as in our study.

Output variables were selected to reflect priority services rendered at the primary healthcare centers that serve as a proxy for health outcomes improvement. Figure 1 shows the flowchart of the relationship between inputs and output variables. The output variables were:

Fig. 1.

Framework of PHC technical efficiency

• i.Attendance of Antenatal Care.
• ii.Number of Skilled Birth Deliveries.
• iii.Number of Outpatient visits.
• iv.Number of fully vaccinated children.

Data analysis

A two-stage data analysis was conducted for this study. In the first stage analysis, the efficiencies of PHCs were estimated using Charnes, Cooper, and Rhodes (CCR) and Banker, Charnes, and Cooper (BCC) models on the assumption of constant returns to scale (CRS) and variable returns to scale (VRS) respectively. Then, a Tobit regression model was used to assess the effect of contextual variables on the efficiency of PHCs.

The DEA CCR model which Charnes et al., 1978 developed assumes that an increase in input results in a corresponding increase in output [37]. Using this model, the efficiency (e_j) of a given PHC j with m inputs (v = 1,2….m), generating n outputs (u = 1,2,….m) is given by (Eq. 1).

1

The DEA BCC model developed by Banker et al. assumes that not all increases in input result in a corresponding increase in outputs, therefore a variable return to scale [38].

Using this model, the efficiency (e_j) of a given PHC j with m inputs (v = 1,2….m), generating n outputs (u = 1,2,….m) is given by (Eq. 2).

2

Before conducting the Tobit regression in the second stage, we performed a Spearman rank correlation analysis between the contextual variables and the inefficiency scores (CRS, VRS, and scale inefficiency). This step served as an exploratory diagnostic to identify potential associations and to assess the strength and direction of bivariate relationships. It also helped to check for any signs of multicollinearity or variable redundancy that could influence the robustness of the Tobit regression estimates. The results of these preliminary correlation tests are presented in Appendix Tables A1 (Level 1 PHCs) and A2 (Level 2 PHCs).

For the second stage analysis, we first determined the correlation between the VRS, CRS and scale inefficiency and the contextual variables using Spearman rank correlations. We then conducted a Tobit stage regression of selected contextual variables with technical and scale inefficiencies of PHCs to assess the effect of these variables on PHCs technical efficiency. The selection of contextual variables in the second-stage Tobit regression was informed by theoretical reasoning and the structure of Nigeria’s PHC system. Four variables were included: catchment area population, patient-reported quality of care (PR-QoC), patient-reported postpartum and newborn care (PR-QMNC), and per-capita subnational allocation to health. Catchment population was used as a proxy for the scale of potential demand. Facilities in larger catchment areas are likely to serve more people, potentially generating more outputs relative to inputs, and thereby appearing more efficient. Conversely, low catchment populations may lead to underutilization despite comparable resource inputs, resulting in inefficiency. PR-QoC was included as an indicator of user satisfaction and perceived service acceptability. Prior studies have shown that perceived quality of care influences healthcare utilization, which in turn affects the volume of services delivered. While most previous efficiency studies have relied on objective quality measures, our use of PR-QoC reflects an adaptation based on available data and aligns with Nigeria’s health system emphasis on people-centered care. We also included PR-QMNC as a domain-specific quality measure focused on maternal and newborn health. While higher reported quality in this area may reflect better service provision, it may also involve higher input intensity per client (e.g., longer consultations, more staff time), which may not always translate into higher output volumes. Finally, per-capita subnational allocation to health was included to reflect the fiscal environment in which PHC facilities operate. In Nigeria’s decentralized system, state-level investment in health influences facility resourcing and oversight. Higher per-capita allocations may support improved service readiness, enabling better conversion of inputs into outputs, and potentially reducing inefficiency.

We assumed a zero-censoring point in the Tobit model, such that efficient PHCs have a score of zero, while inefficient PHCs have scores greater than zero. Like Alatawi et al. (2021) [39], we calculated VRS, CRS, and scale inefficiency scores of health facilities using the formula:

Inefficiency score = (1/Technical efficiency score) – 1)

Orientation

We used an input orientation for the technical efficiency analysis because the objective of the BHCPF is to directly finance PHCs and give managers control over resources for PHC services, thereby using these funds discretionally in improving PHC services.

Results

In the second quarter of 2022, there were 40,929 workers and 30,483 beds across the 3,939 PHCs. These workers fully vaccinated 852,754 children, attended to 2,831,031 outpatients, and 706,383 pregnant women, and delivered 135,836 children. The summary statistics of input and output variables by facility level are presented in (Table 2).

Table 2.

Summary of inputs and outputs

	Inputs		Outputs
	Number of beds	Number of workers	Number of outpatients	Number of vaccinations	Antenatal care attendance	Number of deliveries
Level 1 facilities
Min	1	1	7	1	2	1
Max	20	57	9342	2584	1214	339
Mean	6.68	9.00	627.708	193.19	137.2	28.8
SD	5.09	7.83	895.108	270.27	181.2	44.8
Level 2 facilities
Min	1	1	3	1	2	1
Max	20	65	9919	6080	3800	818
Mean	7.96	10.55	729.23	219.17	184.20	35.14
SD	4.99	9.12	1061.6	374.74	274.01	51.95

Efficiency of level 1 PHCs

The average efficiency score for PHC clinics under the assumption of CRS was 27.6 ± 23.3%, while the efficiency scores ranged from 1.1% to 100%. Thirteen of the facilities had CRS technical efficiency of 100% and 59% of the facilities had CRS technical efficiency scores of less than 25%.

Under the assumption of VRS, the average efficiency score of PHC clinics was 44 ± 27%, with the efficiency scores ranging between 6.28% and 100%. There were 49 VRS technically efficient PHC clinics and 62.5% of health clinics had efficiency scores of less than 50%.

The average scale efficiency score was 62% with scores ranging from 3.4% to 100%. Fourteen PHC clinics had scale efficiency scores of 100%, which shows they produced optimum amounts of outputs for their size. The scale efficiency scores for 96.6% (394) were less than 100, which made them inefficient at their current scale of operation (Fig. 2).

Fig. 2.

Efficiency of BHCPF-funded level 1 facilities

Efficiency of level 2 PHCs

The average efficiency score for level 2 PHCs under the assumption of CRS was 11 ± 12%, with scores ranging between 1.2% and 100%. Ten facilities had a CRS technical efficiency score of 100%, and 91% of facilities had constant return-to-scale technical efficiency scores of less than 25%.

Under the assumption of VRS, the average efficiency score of level 2 PHCs was 30% ±21%, with scores ranging from 5.1% to 100%. There were 176 VRS technically efficient PHC centers and 82% of health centers had efficiency scores of less than 50%.

The average scale efficiency score was 39.1% ranging from 1.3% − 100%. Ten PHC centers had scale efficiency scores of 100%, which shows they produced optimum amounts of outputs for their size. The scale efficiency scores for 3521 (99.7%) was less than 100, which makes them inefficient at their current scale of operation (Fig. 3).

Fig. 3.

Efficiency of BHCPF-funded level 2 PHCs

There are wide subnational variations in the average efficiency scores of level 1 and 2 PHCs as shown in (Figs. 4 and 5).

Fig. 4.

Technical efficiency of level 1 PHCs by state

Fig. 5.

Technical efficiency analysis of level 2 PHCs by states

Tobit regression of contextual variables and inefficiencies of level 1 PHC facilities

Tobit regression of contextual variables and technical inefficiency scores showed that 0.49%, 0.04% and 1.33% of the CRS, VRS, and scale inefficiencies can be explained by the contextual variables. Details of significance of variables and nature of relationships is presented in (Table 3).

Table 3.

Tobit regression of contextual variables and level 1 PHC facilities

PR-QOC	CRS Inefficiency	VRS Inefficiency	Scale Inefficiency
	0.019	0.0115	− 0.049
PR-QMNC	-0.268	0.0352	− 0.172**
Catchment Area Population	− 0.0001**	0.00000140	− 0.00004***
State per-capital allocation to healthcare	− 0.0011***	0.0000169	− 0.00036***
LR Chi2 [4]	14.54	0.63	26.65
Prob> Chi2	0.0058	0.9598	0.0000
Pseudo R2	0.0049	0.0004	0.0133
Log-likelihood function	-1476.56.04	-891.53	-991.76

Abbreviations: *,**,*** - significance at p<0.1, p<0.05 and p<0.01 respectively, PR-QoC – the patient-related quality of care, PR-QMNC – the patient-reported quality of maternal and newborn care

Results showed that VRS inefficiency scores does not have any statistically significant relationship with any of the contextual variables. However, CRS inefficiency has a negative and statistically significant relationship with catchment area population (p = 0.025) and per-capital allocation to healthcare (p = 0.007). The negative relationship implies an inverse relationship between inefficiency, catchment area population and per-capita allocation to healthcare. However, the degree of influence is weak due to very low co-efficients.

Scale inefficiency scores of level 1 PHCs also showed a negative and statistically significant relationship with patient reported quality of maternal and newborn care (p = 0.017), catchment area population (p = 0.003) and per-capital allocation to healthcare (p = 0.004).

Tobit regression of contextual variables and inefficiencies of level 2 PHC facilities

Tobit regression analysis was done to relate technical inefficiency scores to the contextual variables. The results reveal that contextual variables explain only 0.17%, 0.31%, and 0.50 of the inefficiencies found among PHCs. The results are presented in (Table 4).

Table 4.

Tobit regression of contextual variables and level 2 PHC facilities

	CRS Inefficiency	VRS Inefficiency	Scale Inefficiency
PR-QOC	0.3413	0.0340	- 0.703
PR-QMNC	0.2459	0.0601**	- 0.0108
Catchment Area Population	- 0.0001***	0.00000751*	- 0.00005***
State per-capital allocation to healthcare	- 0.0001	0.00027***	- 0.00034***
LR Chi2 [4]	54.94	54.14	110.15
Prob>Chi2	0.0000	0.0000	0.0000
Pseudo R2	0.0017	0.0031	0.0050
Log-likelihood function	15712.04	-8620	- 10911.15

Abbreviations: PR-QOC patient related quality of care, PR-QMNC Patient reported quality of maternal and newborn care,  *,**,*** significance at p<0.1, p<0.05 and p<0.01% respectively

Regarding the contextual factors, results indicated that the patient-reported quality of care is not statistically significantly related to technical inefficiency, VRS inefficiency, or scale inefficiency. The Patient-reported quality of maternal and newborn care had a significantly positive relationship with VRS inefficiency (p = 0.041), indicating that the health centers with higher patient-reported quality of maternal had higher levels of inefficiency. However, no significant relationship was observed between CRS inefficiency and scale inefficiency.

The catchment area population of PHC facilities has a significant relationship with CRS inefficiency, VRS inefficiency, and scale inefficiency. A significantly positive relationship exists between VRS efficiency and the per-capital allocation to health (p = 0.0000). However, there is a very significant but negative relationship between CRS inefficiency and catchment area population, a unit increase in catchment area population would lead to 0.00000751 and 0.00005 reductions in CRS inefficiency and scale inefficiency respectively. Similarly, a significant negative relationship exists between scale inefficiency and per-capita budgetary allocation to healthcare (p = 0.000). The relationship between CRS inefficiency and per capita allocation to healthcare is statistically insignificant, while there is a positively significant relationship between VRS inefficiency and per capita expenditure.

Discussion

Our study analyzed the efficiency of BHCPF-funded facilities in Nigeria using data envelopment analysis; a non-parametric technique that uses a deterministic frontier - determining the DMU operating at the PPF and ranks efficient decision-making units DMUs based on this frontier [20]. We also assessed the effect of contextual environmental and institutional variables on the efficiency of these facilities. In this study, the DEA efficiency scores of both PHC centers and PHC are very low, yielding lower than 50% for pure technical efficiencies and Scale efficiencies for both PHC clinics and PHC centers. There are subnational and regional differences in the average efficiency scores of the health centers and health clinics, these subnational differences are reflective of the decentralized governance of PHC facilities in Nigeria, as subnational governments are responsible for the management of PHC facilities within their domains [40] Similarly, the large standard deviations observed for both CRS and VRS efficiency scores nearly equal to their respective mean values also reflect the substantial variation in facility performance across the country. This widespread may be indicative of heterogeneity in service readiness, population catchment, or managerial capacity among PHC facilities, and is consistent with the decentralized nature of PHC governance in Nigeria. The average efficiency score from our analysis is lower than that reported by Adejoh et al., 2022 [30] who examined the cross-efficiency of primary health care in the Lagos metropolis and found an average efficiency score of 75.3%. This difference might have resulted from the focus on Lagos state. The state also performed comparatively better than other states in our analysis [30].

Furthermore, over 90% of the health facilities in this study were found to be inefficient, which is below the findings of efficiency studies of PHC facilities in sub-Saharan Africa. Akazali et al. (2008) found that 78% of public health centers in Ghana are inefficient [24], while Kirirgia et al. (2000) found that 70% of PHCs in South Africa were technically inefficient [41]. The he low technical efficiency observed across many PHC facilities in Nigeria may reflect systemic issues in service readiness, uneven demand generation, limited managerial autonomy, and resource misalignment. Despite the introduction of the BHCPF, many facilities continue to face bottlenecks such as inconsistent human resource deployment, poor infrastructure maintenance, and limited community engagement, all of which can hinder optimal use of inputs to generate outputs [42]. This aligns with the findings of Kress et al., 2016 that found that PHC worker productivity in Nigeria is low in an assessment of the Nigerian PHC system [42].

The results also established that the delivery of PHC services is more efficient in PHC clinics than in PHC centers in Nigeria. This could be because PHC centers are larger, run 24-hour services, and have more room for efficiency improvement than PHC clinics. This however contrasts the findings of Farewar et al. (2022) [40], which established that larger facilities that deliver more comprehensive care are more efficient than smaller facilities [43].

Patient-reported quality of maternal and newborn care, facility catchment area population, and state per capital allocation to health had a significant relationship with the efficiency of health facilities, albeit the overall effect is small. This indicates the need for further research to understand the main drivers of inefficiencies in these PHC facilities. The patient-reported quality of maternal and childcare is found to be negatively associated with pure technical efficiency of PHC but did not affect scale efficiency, corroborating the findings of Gorgemans et al., 2018 [44] which highlighted that improvements in quality do not necessarily translate into improvements in the efficiency of health facilities [44]. Notwithstanding, quality of care is an important lever of primary health care which guarantees healthy outcomes within the population.

The facility catchment population was positively associated with the scale efficiency of the health facilities, which is like the finding of Kontodimopoulos et al. (2007) [45]. However, it was negatively associated with VRS efficiency. Such similar differences were observed in the study of Alatawi et al. (2021) [39] to be influenced by the sociodemographic characteristics of the residents from the catchment area as an underlying driver of efficiencies [39], but this was not explored in this study, hence further research will be required to validate it. State per capita allocation to health was a significant predictor of the scale efficiency of PHCs but there was a reverse relationship with VRS efficiency. The per-capita allocation to health care which is an indicator of the priority of healthcare within the state shows that subnational prioritization of health significantly affects the efficiency of health facilities.

The inverse relationships observed in our analysis suggest that certain contextual factors may help buffer against inefficiency. Facilities with larger catchment populations tend to serve more clients, which may enable them to convert fixed inputs into higher service outputs, thus reducing inefficiency through scale advantages. Similarly, higher per-capita health allocations at the state level are likely to improve resource availability, supervision, and supply chain, all of which may enhance operational efficiency at the facility level. In contrast, the positive association between patient-reported postpartum and newborn care (PR-QMNC) and inefficiency may reflect the time- and labor-intensive nature of high-quality maternal and newborn services. These services may require more staff effort per patient but do not necessarily result in higher throughput, thus increasing measured inefficiency in input-oriented DEA models.

While improving the efficiency of healthcare facilities is essential, equity remains a cardinal principle of primary health care. Managing PHC facilities must therefore strike a balance between maximizing output (efficiency) and ensuring fair access, particularly for underserved or hard-to-reach populations [46]. In some cases, facilities serving marginalized communities may appear less efficient due to lower utilization, despite fulfilling critical equity functions. As such, performance management should consider both efficiency metrics and the broader equity objectives of PHC systems.

The findings of this study align with multiple Sustainable Development Goals (SDGs), particularly SDG 3 (Good Health and Well-Being), by identifying inefficiencies in PHC service delivery and promoting more equitable and effective use of health resources to advance universal health coverage (Target 3.8). The observed regional disparities in efficiency also reflect ongoing challenges in equity, resonating with SDG 10 (Reduced Inequalities). Additionally, the study underscores the importance of accountable governance in resource allocation, consistent with SDG 16 (Peace, Justice, and Strong Institutions). While not the primary focus, the research further supports SDG 17 (Partnerships for the Goals) by highlighting the role of multi-stakeholder engagement and data-driven planning in improving health system performance.

Study Limitations

The limitations of our study should be acknowledged. This study relied on secondary data from a nationally conducted PHC assessment, and as such, we had no control over the original data collection protocols or completeness. While the dataset was comprehensive and nationally representative, the use of pre-existing data limited our ability to influence variable selection, enforce data quality measures directly, or address facility-level missingness beyond what was already cleaned by the primary data collectors, this made us encounter difficulty in obtaining the economic value of inputs of health facilities that would have allowed for a more precise estimation of capital inputs Similarly, it would have been more desirable to disaggregate the total number of PHC workers into clinical workers and support staff, however, the total number of workers still serves as a good proxy for labor inputs. Finally, we could not disaggregate health centers into rural and urban facilities, due to the absence of a standard metric of delineation; however, we used the existing minimum standards of PHC services to delineate PHCs into PHC centers and PHC clinics and ensured the homogeneity of our DMUs.

Conclusion

This study found that a significant proportion of BHCPF-funded PHC facilities in Nigeria operate with low technical efficiency, with considerable variation across facility types and contextual factors. Catchment population and state-level per-capita health allocations were inversely associated with inefficiency, suggesting that demand-side pressure and stronger public investment may enhance service productivity. In contrast, facilities with higher patient-reported scores for postpartum and newborn care showed greater inefficiency, possibly reflecting the time- and resource-intensiveness of comprehensive maternal care. These findings highlight the need for differentiated performance strategies that account for facility context, population dynamics, and service complexity to improve efficiency in PHC delivery.

Abbreviations

BHCPF

Basic health care provision fund

PHC

Primary health care

CRS

Constant return to scale

VRS

Variable returns to scale

PSAH

Per-capita subnational allocation to health

CAP

Catchment area's population

PRQoC

Patient-reported quality of care

PPNC

Patient-reported post-partum and newborn care

NPHCDA

National primary healthcare development agency

NHIA

National health insurance authority

SFA

Stochastic frontier analysis

DEA

Data envelopment analysis

LMICs

Low and middle-income countries

DMU

Decision-making unit

PPF

Production possibilities frontier

GDP

Gross domestic products

FCT

Federal capital territory

CAPI

Computer-assisted personnel interview

ODK

Open data kit

CCR

Charnes, cooper, and rhodes

BCC

Banker, charnes, and cooper

Author contributions

TO and OA conceptualized the article. TO developed the original draft that was initially reviewed by OA. SDI and HIO subsequently reviewed and refined the original manuscript draft. OWS conducted the final review of the manuscript draft, and TO and SDI incorporated the feedback. SS and HE conducted a dataset review of the data utilized for analysis. FO managed the project implementation process. HE and OO led the team of program analysts on the project implementation. SN, OD, DA, OO, LK, HU, and SA were all program analysts who played major roles in data curation from the implementation of the project.

Funding

This study did not receive any external funding or grant from any funding agency in the public, commercial, or non-profit sectors.

Data availability

The data supporting the findings will be available from the corresponding author upon request following a 1-year embargo from the publication date. Requests will be examined and considered on a case-by-case basis. The embargo has been placed due to potential bureaucratic bottlenecks that may affect the approval of sharing the large study dataset with third parties.

Declarations

Ethical Approval and Consent to Participate

This study involved secondary analysis of de-identified facility-level data and did not require additional ethical approval. However, the primary data collection adhered to the principles outlined in the Declaration of Helsinki. Prior to commencement, informed consent was obtained from each participant. Additionally, the research protocols were approved by the National Emergency Maternal and Child Health Intervention Centre of the National Primary Health Care Development Agency (NPHCDA) in Abuja, Nigeria.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.