Measuring Technical Efficiency and Productivity of Secondary Maternal and Child Health Hospitals in Hubei,China: Some Evidences from Hubei Province of China 2019 to 2021

Abstract

Environmental factors like COVID-19 can have significant impact on technical efficiency (TE) and total factor productivity (TFP) of health services provided. In this study, focusing on Maternal and Child Health (MCH) hospitals in Hubei Province of China in 2019 to 2021, we aimed to measure their TE and TFP, identify some influential environmental factors, and propose some policy recommendations. Altogether 62 secondary MCH hospitals were selected as the study sample. Four input indicators, 3 output indicators, and 4 environmental indicators were selected to analyze the panel data from 2019 to 2021. Three-stage Data Envelopment Analysis (DEA) and Malmquist Productivity Index (MPI) model were employed to estimate the TE and TFP of these hospitals. During 2019 to 2021, the inputs of the sample hospitals had increased, while the outputs had decreased. The inputs redundancy was negatively associated with birth rate, number of residents, and GDP per capita (P < .05). It was positively associated with number of COVID-19 infections (P < .05). The adjusted TE scores in 2019 to 2021 were 0.822, 0.784, and 0.803, respectively. The TFP declined in 2020 and 2021 compared to 2019, with scores being 0.845 and 0.762. The technical efficiency change (TEC) scores from 2019 to 2021 were 0.926 and 1.063. The technological change (TC) scores from 2019 to 2021 were 0.912 and 0.716. During 2019 to 2021, the operation of sample hospitals had been significantly influenced by environmental factors like COVID-19 pandemic, low birth rate, number of residents, and GDP per capita. The inputs had increased but outputs had decreased, leading to an increase in inputs redundancy and a decline in TE. The TFP showed a downward trend, with TC and SEC being the priority directions for improvement. Some recommendations are made for both hospitals and government to continuously improve the TE and TFP.

Keywords

maternal and child health hospitals the 3-stage DEA model technical efficiency total factor productivity technological change

What do we already know about this topic?

Most MCH studies applied 1-stage and 2-stage DEA model to analyze the TE and TFP of MCH hospitals in China. Many of these studies did not generate the adjusted efficiency scores causing by environmental factors and random errors, potentially causing them to deviate from authentic scores. It is necessary to employ SFA to adjust the impact of environmental factors and random errors on TE and TFP.

How does your research contribute to the field?

In this study, we employed the 3-stage DEA and MPI models to measure the TE and TFP of the secondary MCH hospitals in Hubei Province of China, and identified some environmental indicators. Our study found that COVID-19, low birth rate, number of residents, and GDP per capita were found to have impact on the TE and TFP of the MCH Hospitals. After eliminating the impact of environmental factors and random errors, the TFP showed a downward trend, with TC and SEC being the priority direction for improvement.

What are your research’s implications toward theory, practice, or policy?

This study suggests that MCH hospitals in Hubei need to optimize inputs resource allocation and enhance medical technological innovation and services, as well as strengthen their information systems for introducing regional healthcare performance evaluation. Policy support, internal management and technological innovation in this direction are key to improving the overall efficiency and productivity of these hospitals.

Introduction

Promoting Maternal and Child Health (MCH) has long been a priority for developing countries to achieve Millenium Development Goals (MDGs) and Sustainable Development Goals (SDGs).¹ However, the MCH care system in China still faces challenges such as a scarcity of institutions (only 25.80% in number compared with general hospitals²); inequitable allocation of resources across institutions; and low level technical efficiency; and so forth.³ Due to the influence of external factors such as the COVID-19 and birth rate, the service efficiency of the MCH hospitals has been significantly influenced.⁴ Some studies have demonstrated the impact of COVID-19 pandemic on MCH services in Africa,⁵ tertiary MCH hospitals in China, etc. Compared to 2019, in 2020, the number of services provided by MCH hospitals decreased by 23.6% in China.⁶ Hubei Province is one of the largest economic regions in Central China, with a population of 58.3 million in 2021, which is close to the population of Italy (59.03 million) and Kenya (53.01 million).⁷ Moreover, Hubei was the first province in China hit by COVID-19. In 2021, the gross domestic product (GDP) of Hubei still reached $775 214 million, ranking top number 7 in China.⁸ Moreover, from 2019 to 2021, the maternal mortality rate and infant mortality rate in Hubei Province were 8.98/100 000 and 2.11‰, respectively, significantly lower than the national average (16.1/100 000 and 5.0‰).⁹ This indicates that Hubei Province has effectively fought against the COVID-19 pandemic, and safeguarded the health of women and children. As a cornerstone for meeting the increasing demand for MCH services at the grassroots level, the technical expertise, guidance, and service capacity of the secondary MCH hospitals directly impact the development of local MCH care programs.¹⁰ Jiang et al¹¹ found an resource redundant phenomenon of tertiary MCH hospitals in Hubei Province in 2020, while secondary MCH hospitals faced issues such as inadequate input of resources and shortage of talents. Given the limited health resources available, it is crucial to optimize input resources based on scientific evidences to improve the efficiency and productivity of secondary MCH hospitals.

Data Envelopment Analysis (DEA), a method for evaluating the relative efficiency of Decision-Making Units (DMUs) including healthcare institutions,¹² has gained widespread recognition and application worldwide. As one of the DEA models, the Malmquist Productivity Index (MPI) model is competitive to address both horizontal and longitudinal panel data, enabling in-depth analysis of efficiency changes. Manavgat and Audibert¹³ conducted a literature review on the efficiency of the global healthcare system (including MCH system) during the COVID-19 pandemic. They found that the efficiency of global healthcare systems has significantly been affected by the COVID-19 pandemic. Some empirical measurements and analysis on the efficiency of healthcare systems during the COVID-19 period have been explored as well in different countries. Kamel and Mousa¹⁴ used the DEA-tobit model to measure and found that during the COVID-19 pandemic, the operational efficiency of isolation hospitals in Egypt was low, indicating the need to optimize the allocation of healthcare resources. Liu et al¹⁵ evaluated that the productivity of public tertiary traditional Chinese medicine hospitals in Hubei Province was affected by the outbreak of the COVID-19 pandemic. Sülkü et al¹⁶ concluded that the COVID-19 pandemic significantly impacted the operation of public hospitals in Türkiye. Many studies have used the DEA/MPI models to measure the technical efficiency (TE) and total factor productivity (TFP) of MCH hospitals.^13,17 However, most of these studies have applied 1-stage DEA model (BCC and MPI), which did not consider the impact of environmental factors and random errors; some used the 2-stage DEA model, which did not generate the adjusted efficiency scores causing by environmental factors and random errors.^18
-20 Aigner, Lovell, and Schmidt proposed Stochastic Frontier Analysis (SFA) model in 1997, Fired et al²¹ further constructed a 3-stage DEA model by combining the SFA model to introduce environmental factors and random errors into the DEA model. At Stages 1 and 3, traditional DEA models were applied, and at Stage 2 SFA was applied to eliminate the influence of environmental factors and random errors, and at Stage 3 the authentic efficiency score of medical institutions were estimated.²² Therefore, in this study we constructed a 3-stage DEA-Malmquist combination model, with the aim to measure the technical efficiency (TE) and total factor productivity (TFP) of secondary MCH hospitals in Hubei Province, identify the environmental factors, and make policy recommendations.

Methods

Study Setting

MCH hospitals in China are specialized healthcare facilities dedicated to providing MCH services. These institutions are categorized into 3 levels, which vary in building areas for hospital operation, functional orientations, departmental configuration standards, number of health professionals and other aspects. For instance, a secondary MCH hospital is required to have between 20 and 49 inpatient beds, with a minimum of 40 health professionals. In contrast, a tertiary MCH hospital should have a minimum of 50 inpatient beds and at least 60 health professionals; a primary MCH hospital is required to have between 5 and 19 inpatient beds, with a minimum of 20 health professionals. In 2019, altogether there were 106 MCH hospitals in Hubei Province.

Study Design

Figure 1 illustrates the approach we employed to measure the TE and TFP scores of secondary MCH hospitals in Hubei Province. The analysis was conducted in both horizontal and longitudinal directions. The DEA-BCC model was applied to analyze each of the 3 year cross-sectional data from 2019 to 2021. TE scores were derived from the 3-stage DEA model for cross-sectional comparison, while total factor productivity change (TFPC) was calculated using the MPI model for longitudinal analyses.

Figure 1.

Diagram measuring the technical efficiency and productivity of secondary MCH hospitals.

Selection of Indicators

Input and output indicators

A systematic review of hospital efficiency evaluation indicators in China by Dong et al, found that there was often an incorrect mixture of monetary and ratio indicators in input and output indicators, which resulted in potential issues, such as confusion between technical and allocative efficiency.^23,24 Jiang et al¹¹ established an efficiency evaluation index system for MCH hospitals that only selects direct quantitative indicators. This study is based on the research insights of Li and Dong, combined with relevant literature on DEA efficiency analysis. Considering the representativeness and availability of data, the number of health professionals, the number of actual beds, the number of equipment above 10 000 RMB yuan, and the building area for hospital operation were selected as the most straightforward resource input indicators.²⁵ There is a lack of evaluation indicators for MCH hospitals that consider the specific characteristics of MCH.¹¹ Therefore, it is necessary to add indicators for the number of birth delivers. As a result, the number of total diagnoses, discharges, and birth delivers at the institution were selected as output indicators.

Environmental indicators

Based on previous studies using 3-stage DEA, the external influences of healthcare institutions include general environmental factors such as economic conditions in the jurisdiction, government policies and regulations, and the population of the jurisdiction.²⁶ Based on the availability of data and the indicator system used in the research by Dai and Chen,^27,28 our study selected the birth rate and the number of residents in the jurisdiction to reflect the demographic environment, and GDP per capita to reflect the economic status. In addition, policies related to COVID-19 may impact the actual efficiency of hospitals.²⁹ Therefore, we selected the cumulative number of COVID-19 infections in the jurisdiction to assess the impact of the pandemic and policy on institutions. The final selection included the cumulative number of COVID-19 infections in the jurisdiction, birth rate, the resident population, and GDP per capita.

Three-Stage DEA Model

The 3-stage DEA model can simultaneously adjust the impact of external environment and random errors, in order to calculate more accurately TE scores that reflect the internal management level of DMUs.

Stage 1: DEA-BCC model

The traditional DEA model is divided into 2 main approaches: input-oriented DEA model and output-oriented DEA model. The input-oriented DEA model aims to minimize inputs to enhance efficiency while maintaining a certain level of outputs. The production technology of the healthcare system is variable returns to scale (VRS), so the input-oriented DEA-BCC model was chosen to conduct a cross-sectional analysis of the input-output data of the secondary MCH hospital in Hubei Province, obtaining its input slack variables and initial efficiency scores. The formula for DEA-BCC is as follows:

\begin{array}{l} \min θ - ε ({\hat{e}}^{T} S^{-} + e^{T} S^{+}) \\ s . t . {\begin{cases} \sum_{j = 1}^{n} X_{j} λ_{j} + S^{-} = θ X_{0} \\ \sum_{j = 1}^{n} Y_{j} λ_{j} - S^{+} = Y_{0} \\ λ_{j} \geq 0, S^{-}, S^{+} \geq 0 \end{cases} \end{array}

Where: $X_{j}$ is the input, $Y_{j}$ is the output, $λ_{j}$ is the weight coefficient. $θ$ is the efficiency scores of the DMU. S⁺, S⁻ represents corresponding input and output slack variables.

Stage 2: Stochastic Frontier Analysis (SFA)

The slack variables calculated in the first stage are not only due to ineffective management of the organization, but also influenced by external environment and random errors. To improve the accuracy of efficiency scores, it is necessary to eliminate the impact of external environment and random errors. Therefore, the SFA model proposed by Aigner et al²² is employed in the second stage to achieve this objective. The previous stage’s slack variables serve as dependent variables in the analysis, with environmental factors acting as independent variables. Based on the type of orientation in Stage 1, refer to the formulas provided by Liu et al³⁰ and Fired et al,²¹ the SFA formulation is as follows:

S_{n i} = f (Z;^{i} β_{n}) + ν_{n i} + μ_{n i}; i = 1, 2, \dots, I; n = 1, 2, \dots, N

Where $S_{ni}$ represents the slack scores of the $n$ input in the i DMU. $Z_{i}$ represents the environmental variable, while $β_{n}$ represents the coefficient of the environmental variable. $V_{ni}$ + $μ_{ni}$ represents the mixed error term, $V_{ni}$ is the random disturbance term, representing the impact of random interference factors on the input slack variable, assuming that it follows an independent distribution $ν \sim N (0, σ_{ν}^{2})$ ; $μ_{n i}$ is the non-negative random variable, representing management inefficiency, it is assumed to follow a normal distribution truncated at the null point, that is, $μ \sim N$ ⁺(0, $σ_{μ}^{2}$ ).

Estimates of the parameters $β_{n}$ , σ², and γ are calculated by maximum likelihood estimation. The individual inputs are then calculated, Refer to the formulas provided by Jondrow et al³¹ and Chen et al,²⁷ the formula for the separation adjustment is as follows:

E (μ | ε) = σ_{*} [\frac{ϕ (λ \frac{ε}{σ})}{Φ (\frac{λ ε}{σ})} + \frac{λ ε}{σ}]

Of these, $σ_{*} = \frac{σ_{μ} σ_{ν}}{σ}, σ = \sqrt{σ_{μ} 2 + σ_{ν} 2}$ , $λ = σ_{μ} / σ_{ν}$ .

Stage 3: Recalculation of the DEA-BCC model

In this stage, we incorporated the second-stage SFA-adjusted input indicators data and the original output indicators data into the DEA-BCC model to calculate the efficiency score after excluding the influence of environmental factors and random errors. The scores obtained at this point can more accurately reflect the efficiency level of the secondary MCH hospital in Hubei province.

Malmquist Productivity Index Model

Due to the fact that the TE scores calculated by the 3-stage DEA are at a static level, in order to further analyze the dynamic changes in the efficiency of secondary MCH hospitals in Hubei Province, we chose the MPI model.³² The MPI was originally proposed by Malmquist³³ in 1953. Fried et al²¹ combined this model with the DEA method, allowing for the decomposition of TFP. By inputting the revised input values and original output values back into the MPI model, efficiency can be longitudinally compared and analyzed over time. TFP consists of 2 components: technical efficiency change (TEC) and technical progress (TC). TEC can be further divided into pure technical efficiency change (PTEC) and scale efficiency change (SEC). Each index greater than 1 indicates that its corresponding efficiency has improved compared to the previous year.³⁴

Data Collection and Processing

The input and output data were collected from the Health Commission of Hubei Province. These data were obtained by the study members during their research on the topic of Primary Public Health Services Performance Evaluation in Hubei Province (2022). The environmental variable data were obtained from various local statistical yearbooks published by the Hubei Provincial Bureau of Statistics and the China Health Statistics Yearbook. Eighty-four secondary MCH hospitals in Hubei Province were selected for this study, spanning from 2019 to 2021. All 3-year data from the 84 hospitals were imported from an Excel spreadsheet into a dataset file, and then double-checked for missing values and outliers. Of these, 22 hospitals had vacancies and incomplete data on the key indices. After excluding the number of absent samples, 62 secondary MCH hospitals met the requirements for data analysis.

Data Analysis

A descriptive analysis was performed on the input and output indicators, focusing on their mean and standard deviation (SD), utilizing SPSS version 23.0 for statistical analysis. In Stage 1, DEAP 2.1 was used to measure the Initial TE in secondary MCH hospitals. In Stage 2, SFA was conducted using Frontier 4.1 to calculate the adjusted input values, with the statistical significance threshold set at P < .05. To mitigate the impact of environmental variables, the data for the cumulative number of COVID-19 infections in the jurisdiction was normalized using negative normalization, while the remaining indicators were normalized using positive normalization. Moreover, the γ scores converged to 1, illustrating that the SFA model was appropriately constructed and the exclusion of the environmental variable was necessary.²⁷ DEAP 2.1 was applied to Stage 3 to calculate adjusted TE. In order to calculate adjusted TFP scores based on the adjusted input values and original output values, DEAP 2.1 was applied as well using the MPI model. The average scores of TE and TFP were calculated in for the form of geometric means. In order to visually represent the change in efficiency of the sample hospitals, different colors were used to indicate different efficiency score categories: gray to represent TEC; yellow, TC; blue, PTEC; red, SEC; and green, TFP. The trend line was used to represent the change in efficiency scores from 2019 to 2021.

Results

Descriptive Statistics of Input and Output Indicators for Secondary MCH Hospitals

As shown in Table 1, there was a significantly wide variation in capacity among different indicators in secondary MCH hospitals. The average annual growth rates of the input data were 3.52%, 2.25%, 25.93%, and 24.60%, respectively. The building areas for hospital operation and the number of equipment above 10 000 RMB yuan grew rapidly. These hospitals experienced capacity and size development during the 2019 to 2021 period. However, the total number of diagnoses, discharges, and birth deliveries in the sample hospitals displayed an overall decreasing trend. Their average annual growth rates were −3.53%, −11.37%, and −14.92%, respectively. In contrast, the number of births deliveries in hospitals declined much faster than other output indicators, dropping by 27.61% year-on-year. More details of the indicators can be found in Additional File 1.

Table 1.

Descriptive Information of Input–Output Indicators of the 62 MCH Hospitals From 2019 to 2021.

	Year	Maximum	Minimum	Max/Min	Mean	SD
Input indicators
Number of health professionals	2019	546	54	10.11	205.45	93.60
	2020	580	52	11.15	207.68	93.18
	2021	580	56	10.36	220.18	94.06
Number of beds	2019	500	20	25	152.47	87.53
	2020	500	20	25	153.15	91.56
	2021	500	25	20	159.42	100.10
Building areas for hospital operation	2019	48 753	1413	34.50	9533.84	8110.92
	2020	36 316	1413	25.70	10 862.77	7996.26
	2021	64 740	1413	45.82	15 119.50	12 758.94
Number of equipment above 10 000 RMB yuan	2019	363	45	8.07	163.19	77.95
	2020	818	45	18.18	207.45	126.80
	2021	1013	56	18.09	253.34	171.20
Output indicators
Number of total diagnostic patients	2019	299 505	11 823	25.33	125 594.52	76 748.74
	2020	378 695	10 125	37.40	114 174.10	78 527.82
	2021	395 369	11 026	35.86	116 872.58	73 009.96
Number of discharged patients	2019	22 852	308	74.19	5670.76	3985.43
	2020	20 788	278	74.78	4655.68	3362.61
	2021	19 537	388	50.35	4454.90	3290.75
Number of birth deliveries	2019	6258	12	521.50	1556.94	1194.33
	2020	6020	6	1003.33	1650.85	1187.49
	2021	3995	10	399.50	1127.13	831.47

Stage 1 (DEA-BCC) Analysis

According to the original data analysis results, the mean TE of sample hospitals from 2019 to 2021 were 0.822, 0.787 and 0.803, respectively. And the number of effective DEA level hospitals accounted for a small proportion, only 25.81 to 32.25%. When decomposed, 30, 27, and 27 hospitals have reached PTE effective levels, and 20, 16, and 18 hospitals have reached SE effective levels. From the perspective of returns to scale, 32, 41, and 35 hospitals showed increasing returns to scale (IRS), 10, 5, and 9 hospitals showed decreasing returns to scale (DRS), respectively. Detailed results were shown in Additional File 2. Since the first stage did not consider the impact of environmental factors and random errors, the preliminary results could not reflect the most real efficiency. Therefore, the second and third stage analyses were carried out on this basis.

Stage 2 (SFA) Analysis

According to Table 2, the γ scores of each input slack variable tend toward 1, indicating that environmental factors are redundantly associated with the inputs of secondary MCH hospitals. The positive regression coefficient indicates that an increase in this environmental variable would have a negative impact, meaning that an increase in the number of cumulative COVID-19 infections in the jurisdiction would reduce the efficiency of the secondary MCH hospital. Conversely, the negative coefficient means that an increase in this environmental factor will have a positive impact. It is suggested that higher birth rate, larger resident populations, and increased GDP per capita consequently enhance the efficiency of secondary MCH hospitals. Overall, the birth rate, the number of resident populations, and the cumulative number of COVID-19 infections are significant factors affecting hospital TE.

Table 2.

SFA Results for Input Indicators of Secondary MCH Hospitals From 2019 to 2021.

Variables	Number of health professionals	Number of beds	Building areas for hospital operation	Number of equipment above 10 000 RMB yuan
A constant (math.)	0.98 (0.07)	−11.16 (−0.83)	−1331.19 (−5.12)	16.78 (0.61)
Cumulative number of COVID-19 infections	11.75 (1.27)	12.5 (1.60)	2778.37*** (10.37)	27.08 (1.52)
Birth rate	−21.98* (−1.74)	−8.09 (−0.72)	−1532.85*** (−9.96)	−49.65** (−2.08)
Number of resident population	1.27 (0.13)	−1.44 (−0.16)	−1911.50*** (−17.09)	−9.69 (1.49)
GDP per capita	−0.38 (−0.03)	12.71 (0.99)	649.32*** (10.23)	5.32 (0.20)
σ²	5637.61*** (4.33)	7611.31*** (3.80)	64 870 900.00*** (64 841 914.00)	19 438.44*** (8.34)
γ	0.93*** (47.08)	0.95*** (64.63)	0.55*** (10.52)	0.88*** (43.22)
LR unilateral detection	119.84	107.50	25.97	80.55

***

P < .01. **P < .05. *P < .1.

Stage 3 (DEA-BCC) Analysis

The changes of TE in secondary MCH hospitals before and after the adjustment during 2019 to 2021 were visually shown in Table 3. The adjusted TE of the sample hospitals decreased after eliminating environmental factors and random errors. The year-on-year decrease rates were 1.70%, 3.95%, and 0.37%, respectively. Additionally, the number of hospitals with effective levels of TE decreased relative to the adjusted levels. The PTE of the sample hospitals increased after adjustment from 2019 to 2021, with growth rates of 2.57%, 2.62%, and 4.85% respectively. However, the adjusted SE decreased, with year-on-year decreases of 4.36%, 5.32%, and 5.20%, respectively. The number of hospitals with IRS was 39, 46, and 41, respectively, which has increased compared with that before adjustment. However, the number of hospitals with DRS was 3, 2, and 4, respectively. which has decreased compared with that before adjustment. These indicated that the secondary MCH hospitals TE and SE were overestimated without considering the environment factors and random errors. In particular, the SE were significantly overestimated. Detailed results were shown in Additional File 3.

Table 3.

Efficiency Analysis of 62 Secondary MCH Hospitals From 2019 to 2021.

		Year	Maximum	Minimum	Number of efficiency = 1 institutions	Mean
First stage	TE	2019	1	0.27	20	0.822
		2020	1	0.211	16	0.784
		2021	1	0.161	17	0.803
	PTE	2019	1	0.420	30	0.895
		2020	1	0.523	27	0.877
		2021	1	0.503	27	0.885
	SE	2019	1	0.369	20	0.917
		2020	1	0.250	16	0.884
		2021	1	0.264	18	0.904
Third stage	TE	2019	1	0.258	17	0.808
		2020	1	0.197	14	0.753
		2021	1	0.159	17	0.800
	PTE	2019	1	0.548	30	0.918
		2020	1	0.573	26	0.900
		2021	1	0.571	31	0.928
	SE	2019	1	0.289	20	0.877
		2020	1	0.255	14	0.837
		2021	1	0.213	18	0.857

MPI Analysis

The TFPC during 2019 to 2020 and 2020 to 2021 both showed a downward trend, with an average annual decrease rate of 5.04%, and the 3-year average TFPC score was only 0.802. When TFP was decomposed into TC and TEC, it was found that TC declined significantly from 2019 to 2021, with an average annual decline of 9.60% and a 3-year average score of 0.808.

Different from TC, from 2019 to 2021, TEC showed a trend of first declining and then rising, with a 3-year average score of 0.993. Compared with the level in 2019, 17 hospitals’ TEC increased to varying degrees in 2020, with a growth rate ranging from 0.50% to 33.30%; 10 hospitals’ TEC remained unchanged; 35 hospitals’ TEC declined, with a decline rate ranging from 0.60% to 57.10%. Compared with the level in 2020, 33 hospitals’ TEC increased in 2021, with a growth rate ranging from 0.20% to 64.10%; 11 hospitals’ TEC remained unchanged; 18 hospitals’ TEC declined, with a decline rate ranging from 1.00% to 49.40%.

When TEC was decomposed into PTEC and SEC, it was found that both PTEC and SEC decreased from 2019 to 2020 and increased from 2020 to 2021. From 2019 to 2020, 16 hospitals’ SEC increased, 11 hospitals’ SEC remained unchanged, and 35 hospitals’ SEC decreased; 19 hospitals’ PTEC increased, 21 hospitals’ PTEC remained unchanged, and 23 hospitals’ PTEC decreased. From 2020 to 2021, 31 hospitals’ SEC increased, 12 hospitals’ SEC remained unchanged, and 19 hospitals’ SEC decreased; 25 hospitals’ PTEC increased, 20 hospitals’ PTEC remained unchanged, and 17 hospitals’ PTEC decreased (Figure 2).

Figure 2.

Changes in TFP and relative efficiency in secondary MCH hospitals in Hubei Province from 2019 to 2021.

Discussion

Changes in Inputs Resources

Table 1 shows that the inputs of the secondary MCH hospitals had stably grown, while the outputs had significantly decreased from 2019 to 2020. The gap in the building areas for hospital operations decreased from 34.50 to 25.70 times, while the average building area increased by 13.94% year-on-year. This can be explained that, due to the COVID-19 pandemic, MCH hospitals increased their space (such as establishing or expanding fever clinics) to identify possible infected patients. The gap in the number of equipment increased from 8.07 to 18.18 times. The number of advanced medical equipment had increased so as to treat more patients with severe or critical conditions. The gaps in the number of beds remained relatively unchanged, which can be explained that the number of discharged patients decreased.

From 2020 to 2021, the gap in the building area for hospital operations increased from 25.7 to 45.82 times, the gap in the number of beds dropped from 25 to 20 times, and the gap in the number of health professionals still remained relatively unchanged. The reason is that at the end of 2020, Hubei Provincial Government launched a 3 year program to build resilient systems to recover from the COVID-19 pandemic, in which health system strengthening is 1 of the 10 major directions.³⁵ Due to this policy incentive, larger MCH hospitals have expanded their scale in building their areas for operations and number of equipment, while smaller MCH hospitals expanded their number of beds.

The Impact of Environmental Factors on the Inputs Allocation

The SFA analysis shows that the birth rate and the number of resident population were negatively associated with inputs redundancy (building area for hospital operations and number of equipment above 10 000 RMB yuan). The resident population in Hubei Province decreased in 2021 compared to 2019. Along with the changes in living pressure and fertility preference, which have constrained people’s willingness to have children,³⁶ thereby diminishing birth rate. The cumulative number of COVID-19 infections and per capita GDP in the jurisdiction were positively correlated with the inputs redundancy. The increase of per capita GDP in Hubei Province has raised public expectations for superior medical care. Tian et al³⁷ indicated that there was an elevated demand for healthcare service among residents in more affluent regions. These indicators affect the demand for medical services. When the inputs of hospital resources remain constant, a decrease in demand leads to resource redundancy, which significantly affects SE.

After eliminating the influence of environmental factors and random errors, Table 2 shows that only 17 (27.42%) secondary MCH hospitals reached the TE effective level in 2021, and 45 (72.58%) hospitals had room for improvement to varying degrees. Among the 45 hospitals, the SE of 44 (97.78%) hospitals and the PTE of 31 (68.89%) hospitals did not reach the effective level (score less than 1). This indicates that a large number of hospitals needed to improve SE, of which 36 (81.82%) had IRS and 3 (18.18%) had DRS. The TE is closly subject to hospital internal management.³⁸ Based on the existing research,¹¹ it is recommended that the TE and the inputs redundancy should be carried out each year for supporting resources optimization and allocation.

Changes in TFP and Relative Efficiency of Secondary MCH Hospitals

Our study results show that during 2019 to 2021, the TFP of secondary MCH hospitals in Hubei Province showed a significant downward trend. In fact, Xie et al³⁹ found a TFP growth rate of 3.3% each year during 2015 to 2017 among country-level MCH hospitals in Hubei Province, which indicates that the TFP had deteriorated after the COVID-19 outbreak. Since the 3-year TC average score (0.808) is much lower than the TEC average score (0.993), TC constructs the first priority for improvement. Our study also found that, although the number of equipment above 10 000 RMB yuan in 2021 increased by 24.83% compared to 2019, the number of equipment above 1 000 000 RMB yuan increased much less (16.53%), the number of equipment above 3 000 000 RMB yuan only increased 2%, indicating that most of the new equipment are ordinary instead of high tech ones. Moreover, these high tech equipment were mainly introduced to treat severe COVID-19 patients. After the epidemic came under control in 2020, many of the newly added medical equipment became vacant and contributed negatively to TC. Therefore, in the future, the utilization rate of these high teach equipment can be either improved by developing new types of services, or transferred to other hospitals to reduce redundancy. On the other hand, hospitals shall be able to decide whether to introduce more MCH high tech equipment according to their own needs.

Different from TC, the 3-year average TEC score (0.993) of sample hospitals had remained relatively unchanged from 2019 to 2021. The 3-year average SEC score (0.986) is lower than the average PTE score (1.006), so the priority order for improvement is SEC and PTEC. Moreover, after the COVID-19 epidemic was controlled, patients with COVID-19 infection mainly went to general hospitals and specialized hospitals, resulting in a low utilization rate of a large number of newly introduced medical equipment in MCH hospitals. In contrast, PTEC remained relatively unchanged (only increased by 0.60%), indicating big potentials for improvement. Therefore, hospitals need to optimize the utilization of building areas and medical equipment. On the other hand, hospitals can endeavor to improve internal management as an approaching to enhance PTEC for excellence.

Limitations

First, this study only obtained data for the years 2019 to 2021, which is a relatively short period. The impacts of some environmental variables on TE and TFP were statistically not as significant as what we expected. It is necessary to monitor more years of data in the future for further verification. Second, limited by the data from the information systems, we were unable to explore the impact of more environmental variables and internal management factors on TE and TFP, further empirical studies can be explored after more complete information systems become available.

Conclusions and Policy Implications

In conclusion, we found that the inputs of the secondary MCH hospitals in Hubei Province had been gradually increasing from 2019 to 2021. However, the outputs showed a downward trend. TE were significantly influenced by factors such as inputs redundancy, COVID-19, birth rate, and so on. After eliminating the impact of environmental factors and random errors, the TFP still showed a downward trend, with the priority order to improve TC and TEC. Further decomposition of TFP into TC and TEC indicated the priority order to improve SEC and PTEC, mainly targeting at technology innovation and internal management.

Our study has some policy implications to both China and other developing countries. First, in order to generate qualified data to aid decision making, it is necessary to to build regional information platforms based on well established hospital information systems. In this way, more internal management variables and environmental variables can be verified for their impact to TFP. Second, relative efficiency and productivity construct only one dimension of performance, it is necessary to evaluate more dimensions such as quality and safety, patient satisfaction, access and equity of health services, etc. Therefore, it is necessary to conduct multi-dimensional performance evaluation for further exploration and analysis. Third, after empirical evidences are generated, it’s necessary to strengthen results application, such as benchmarking, inter-organizational learning, and continuous improvement.

Footnotes

Acknowledgements

We would like to thank student Yi Zhou, from the School of Public Health of Wuhan University for delivering charts and additional files based on our discussion. We are grateful to Ms Shumin Deng from Department of Clinical Data Center of The Third Affiliated Hospital of Sun Yat-Sen University for her assistance in polishing the language of our manuscript. Our special thanks are given to Dr.Hao Li, professor from the School of Public Health of Wuhan University for his frequent comments and advice, which are very helpful to improve the quality of our article. Last, we would like to give our sincere thanks to the colleagues in the Health Commission of Hubei Province for supporting us with the data. The findings and conclusions in the article are those of the authors and do not represent the views of the Health Commission of Hubei Province.

Author Contributions

QW developed the design and oversaw the whole writing process. JHC processed and analyzed the data, and wrote the manuscript. CLJ contributed in developing the package for statistical analysis and visualization of the Figures. YZ and YHZ contributed in collecting and processing data.

Availability of Data and Materials

The raw data are in the hands of the authors. They cannot be shared due to confidentiality reasons as required by Health Commission of Hubei Province.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research has been funded by the Primary Public Health Services Performance Evaluation in Hubei Province (2022) (grant number: 202308007).

Ethics Approval and Consent to Participate

Data were from the official government report and did not include personal information of humans. Therefore, there is no need to ask for ethics approval and consent to participate.

Consent for Publication

All the authors listed in this paper read the manuscript and agreed to submit it for publication.

Supplemental Material

Supplemental material for this article is available online.

References

Yoshino

Yabashi

Surenkholoo

Participatory blended maternal and child health (MCH) seminar programs to strengthen the competency among nursing professionals in the developing country. Ann Global Health. 2015;81:61-62.

Wang

Shi

. Statistical Bulletin on the Development of China’’s Health and Wellness Program, 2021. In: National Health Commission of the People’s Republic of China 2021:804-813.

Zhu

Zhang

Ding

Evaluation of the efficiency of county hospitals of maternity and child healthcarein chinabased on data envelopment analysis. Chin Rural Heal Serv Adm. 2023;43(05):343-347.

İlgün

Sönmez

Konca

Yetim

Measuring the efficiency of Turkish maternal and child health hospitals: a two-stage data envelopment analysis. Eval Program Plann. 2022;91:102023.

Adu

Stallwood

Adebola

Abah

Okpani

AI.

The direct and indirect impact of COVID-19 pandemic on maternal and child health services in Africa: a scoping review. Glob Health Res Policy. 2022;7(1):20.

Luo

Huang

Study on the influence of COVID-19 pandemic on the operation and management of tertiary maternal and child health hospitals. Chin J Hosp Adm. 2022;38(7):515-518.

Bank

GDP of all the countries in the world. Published 2021. Accessed December 07, 2022. https://databank.worldbank.org/indicator/NY.GNP.PCAP.CD/1ff4a498/Popular-Indicators

Bureau

HPS

. Provincial statistical yearbook. Published November 21, 2022. Updated November 21, 2022. Accessed September 5, 2023. https://tjj.hubei.gov.cn/tjsj/sjkscx/tjnj/qstjnj/

Province HCoH. Notice on the Issuance of the “14th Five-Year Plan” for the Development of Health Care in Hubei Province. Published 2021. Accessed November 19, 2021. http://wjw.hubei.gov.cn/zfxxgk/fdzdgknr/ghxx/202111/t20211123_3878474.shtml

10.

Heng

Wang

Shen

Research on the development planning strategy of the third level maternal and child care institutions based on SWOT analysis. Chin Hosp Manag. 2021;41(7):94-96.

11.

Jiang

Liu

Chen

, et al The technical efficiency of maternal and child health hospitals in China: a case study of Hubei Province. Reprod Health. 2022;19(1):1-8.

12.

Kohl

Schoenfelder

Fügener

Brunner

JO.

The use of data envelopment analysis (DEA) in healthcare with a focus on hospitals. Health Care Manag Sci. 2019;22(2):245-286.

13.

Manavgat

Audibert

Healthcare system efficiency and drivers: re-evaluation of OECD countries for COVID-19. SSM - Health Syst. 2024;2:100003.

14.

Kamel

Mousa

MES

. Measuring operational efficiency of isolation hospitals during COVID-19 pandemic using data envelopment analysis: a case of Egypt. Benchmarking. 2021;28(7):2178-2201.

15.

Liu

Jia

Cheng

, et al Measuring productivity and relative efficiency of public tertiary traditional Chinese Medicine Hospitals in Hubei, China. Inquiry. 2023;60:1-9.

16.

Sülkü

Mortaş

Küçük

Measuring efficiency of public hospitals under the impact of Covid-19: the case of Türkiye. Cost Eff Resour Alloc. 2023;21(1):70.

17.

Amare

Yitayal

Amare

Technical efficiency of maternal health services provision in public hospitals of northwest Ethiopia: a two-stage data envelopment analysis. Risk Manag Healthc Policy. 2020;13:3135-3146.

18.

Emrouznejad

Yang

GL.

A survey and analysis of the first 40 years of scholarly literature in DEA: 1978–2016. Socioecon Plann Sci. 2018;61:4-8.

19.

Nepomuceno

TCC

Piubello Orsini

de Carvalho

VDH

Poleto

Leardini

. The core of healthcare efficiency: a comprehensive bibliometric review on frontier analysis of hospitals. Healthcare. 2022;10(7):1316.

20.

Rosko

Mutter

RL.

Stochastic frontier analysis of hospital inefficiency: a review of empirical issues and an assessment of robustness. Med Care Res Rev. 2008;65(2):131-166.

21.

Fried

Lovell

CAK

Schmidt

Yaisawarng

Accounting for environmental effects and statistical noise in data envelopment analysis. J Prod Anal. 2002;17(1-2):157-174.

22.

Aigner

Lovell

CAK

Schmidt

Formulation and estimation of stochastic frontier production function models. J Econom. 1977;6(1):21-37.

23.

Dong

Liu

Relative efficiency and productivity: a preliminary exploration of public hospitals in Beijing, China. BMC Health Serv Res. 2014;14:158.

24.

Dong

Zuo

Tao

, et al Study on DEA-based Chinese hospital efficiency and applied indicators. Chin J Heal Policy. 2014;7(10):40-45.

25.

Dong

Measuring and benchmarking technical efficiency of public hospitals in Tianjin, China: a bootstrap-data envelopment analysis approach. Inq J Heal Care Organ Provis Financ. 2015;52:5.

26.

Method and Empirical Study of Three-Stage DEA Model [Master]. Peking Union Medical College; 2013.

27.

Chen

Zhang

Liu

Research on three-stage DEA model. Syst Eng. 2014;32(9):144-149.

28.

Dai

Impact of Medicare reform on the efficiency of medical services in China: an empirical analysis based on the three-stage DEA. J Xiamen Univ. 2011;(06):97-103.

29.

Ahmed

Kumar

Mehraj

, et al Perceptions of health care workers on maternal and child health services in Pakistan during COVID-19: a cross-sectional study. Dialogues Heal. 2023;3:100145.

30.

Liu

Yao

A study on the efficiency of pediatric healthcare services and its influencing factors in China ——estimation of a three-stage DEA model based on provincial-level data. Socioecon Plann Sci. 2022;84:101315.

31.

Jondrow

Knox Lovell

Materov

Schmidt

On the estimation of technical inefficiency in the stochastic frontier production function model. J Econom. 1982;19(2-3):233-238.

32.

Fuentes

Lillo-Bañuls

Smoothed bootstrap Malmquist index based on DEA model to compute productivity of tax offices. Expert Syst Appl. 2015;42(5):2442-2450.

33.

Malmquist

Index numbers and indifference surfaces. Trab Estad. 1953;4(2):209-242.

34.

Jing-jian

Bing

Evaluation results of malmquist productivity index-new insight of technical change. Oper Res Manage Sci. 2010;19(1):170-175.

35.

Government HMPs. Notification of the people’s government of the province on issuing the three-year action plan (2020-2022) For the “ten major projects” for post-epidemic revitalization and strengthening of weaknesses in Hubei Province. Accessed August 19, 2023. https://media.hubei.gov.cnzyxmt/index_list/index_weixin/202008/t20200819_2810809.html

36.

Song

Jiang

Chen

China’s low fertility and provincial differences: Fertility Index-based observation. Popul Res. 2023;47(05):74-87.

37.

Tian

Tong

Fiscal expenditure,government competition and medical and healthcare service level:an empirical analysis based on provincial panel entropy tobit model. J Yunnan Univ Financ Econ. 2022;38(08):19-36.

38.

Zhao

Zhou

Tian

The efficiency evaluation of municipal-level traditional Chinese medicine hospitals using data envelopment analysis after the implementation of hierarchical medical treatment policy in Gansu Province, China. Inquiry. 2022;59:00469580221095799.

39.

Xie

Yue

Niu

Liu

Efficiency evaluation of county maternal and child health hospital based on DEA-Tobit model. Matern Child Heal Care China. 2020;35(7):1172-1175.