-
Data on the CDC system workforce were available from two sources - China Health Statistical Yearbooks (CHSY)[1], which are published annually by the National Health Commission (NHC, formerly the Chinese Ministry of Health) and the Chinese Center for Disease Control and Prevention basic information system. We analyzed discrepancies across the two data sources, which use different data collection instruments and methodologies. Because the basic information system was newly established in 2004, we used province level data for the CDC system workforce and resident population data from CHSY during the period 2008–2017, covering 31 provinces, municipalities, and autonomous regions of China, with the exception of Hong Kong, Macao, and Taiwan.
-
According to the document, ‘Opinions of the CPC Central Committee and the State Council on Deepening the Health Care System Reform’ in 2009[18], CDCs and professional public health agencies involved in health education, maternal and child health, mental health, emergency response, blood, health supervision, and family planning, based on primary level medical and health care networks, must work together to develop a functional public health system. As the most integral part of a highly efficient and sustainable public health system, CDCs are pyramidal in structure and have a centralized system of administration that runs from their center (the Chinese Center for Disease Control and Prevention [China CDC]), through an intermediary level (provincial CDCs), culminating at the primary level (municipal CDCs and district/county CDCs). The different levels of health commission/government have authority to enforce collaboration between corresponding health-related agencies/departments and CDCs (Figure 1). In 2017, in additon to China CDC, there were 3456 CDCs covering all counties/municipalities/provinces throughout China [1] (Table 1).
Level of CDCs Number of CDCs Public health workforce size CDCs health workers Provincial 31 11,129 7,801 Municipal/city 411 42,654 32,237 County-level city 1,220 60,429 44,982 County 1,553 69,501 51,795 Othersb 241 7,017 5,299 Total 3,456 190,730 142,114 Note. aThe latest data are from the 2018 CHSY, providing information from the previous year; bThe category of ‘others’ includes CDCs owned by railway, agriculture, and reclamation, or other systems. Table 1. Numbers of CDCs and size of workforce, 2017a
-
By type of work performed, there were four public health workforce categories[1]: (i) health workers; (ii) other technicians; (iii) management staff; and (iv) logistics staff. They are defined as follows:
Health workers: doctor or assistant doctor, nurse, pharmacist, laboratory technician, and other technical staff in CDCs (of which 81.8% are licensed staff who pass a licensing examination or are registered at a county or higher level health authority) — representing 74.5% of the total workforce.
Other technicians: professionals who engage in engineering, economics, information technology, editors, and other non health-related workers in the CDCs — representing 7.7% of the total.
Management staff: individuals responsible for managing or administering staff and systems to enhance the quality and efficiency in CDCs — representing 7.3% of the total.
Logistics staff: individuals who provide logistics support and services — representing 10.5% of the total.
Employees of immunization clinics, specialized preventive institutions, and women and child care agencies were not included in the study.
-
The density of the public health workforce of CDCs was measured by the combined number of public health workers per 10,000 population, defined as
$$ {\text{D}} ={\text{h} _{\text{i}}}{\rm{/}}{{\text{p} }_{\text{i} }} $$ Where hi refers to the number of public health workers in province i, and pi refers to the total population of province i.
-
According to the traditional administrative district definition by the National Bureau of Statistics, China is stratified into three distinct regions (East, Central and West), excluding Hong Kong, Macao, and Taiwan[1]. The East includes 11 provinces or directly-controlled cities: Beijing, Tianjin, Hebei, Liaoning, Shanghai, Jiangsu, Zhejiang, Fujian, Shandong, Guangdong, and Hainan; Central China includes eight provinces: Shanxi, Jilin, Heilongjiang, Anhui, Jiangxi, Henan, Hubei, and Hunan; and the West covers 12 provinces, autonomous regions, or directly-controlled cities: Inner Mongolia, Chongqing, Guangxi, Sichuan, Guizhou, Yunnan, Tibet, Shaanxi, Gansu, Qinghai, Ningxia, and Xinjiang.
-
We used three indices to measure inequality of density of the CDCs public health workforce: the Gini coefficient, Theil L, and Theil T[19]. The Gini index is widely used to measure aggregate level of inequality and ranges between 0 and 1, with higher values indicating higher levels of inequality. A Gini coefficient < 0.2 is considered absolute equality, 0.2–0.3 is considered proper inequality, 0.4–0.5 is considered large inequality, and > 0.5 is considered severe inequality. The Theil index is a relative indicator, with no universal assessment standard, that shows contributions within subgroups and between subgroups on the basis of a calculated contribution rate. Values range from 0 to 1, with higher values indicating lower levels of inequality.
-
Spatial autocorrelation uses variable values and their spatial locations, and reflects the degree of spatial dependence between random variable values in geographic terms. The global Moran’s I method has been widely used to reflect the degree of spatial autocorrelation of variables and to estimate spatial agglomeration and divergence distributions[20]. It is defined by the equation:
$$ I=\frac{N{\sum }_{i}{\sum }_{j}{W}_{ij}({X}_{i}-\stackrel{-}{X})({X}_{j}-\stackrel{-}{X})}{{\left({\sum }_{i}{\sum }_{j}{W}_{ij}\right){\sum }_{i}({X}_{i}-\stackrel{-}{X})}^{\text{2}}} $$ where N is the total number of provinces in the study area; Xi and Xj are the public health workforce indices of the ith and jth provincial units;
$ \stackrel{-}{X} $ is the mean of the variable and Wij is the spatial weight matrix. The global Moran’s I value range is [–1,1]. Positive values indicate spatial agglomeration; negative values indicate spatial divergence.We used local spatial autocorrelation to explore distributions of provinces. A Moran scatterplot can be divided into four quadrants: the first quadrant is high value and high value (H–H), which indicates a province that has a high workforce density that is adjacent to a province with a high workforce density; the second quadrant is low value and high value (L–H), indicating a province with a low workforce density adjacent to a province with a high workforce density; the third quadrant is low value and low value (L–L), indicating low density in two adjacent provinces; and the fourth quadrant is high value and low value (H–L), indicating a province with a high density neighboring a province with low workforce density. H–H and L–L quadrants show spatial clusters, while H–L and L–H quadrants are regarded as spatial outliers.
-
The inequality indices were computed using Excel 2007. GeoDa 1.14.0 was used to produce a LISA map for identifying spatial clusters and outliers and for calculating the global Moran’s I and local Moran’s I. Spatial weights were analyzed by the neighboring weights method. The first order queen contiguity was selected as the rule for spatial weights, and Hainan province was considered to be connected with Guangdong province. The density distribution was mapped by ArcGIS 10 (Esri Inc., Redlands, CA, USA)
-
The public health workforce density decreased year by year. The aggregate ratio of workforce to population decreased from 1.47 to 1.42 per 10,000 from 2008 to 2017, consistently lower than the NHC’s recommended critical shortage threshold of 1.75 per 10,000. The standard deviation (SD) of the ratio varied between 0.56 and 0.67, while the coefficient of variation (CV) ranged from 34% to 40% during the ten years (Figure 2A). When the population refers to the population of the corresponding year, the aggregate density decreased more sharply than when calculated using 2010 as the year of reference (Figure 2B).
Figure 2. Mean, SD, and coefficient for density of public health workforce to population per 10,000, 2008–2017. (A) The population refers specifically to 2010. ‘The guidelines for the establishment standards for CDCs’ issued by the Chinese government in 2014, require that the population of the 6th national population census in 2010 is taken as the reference year. (B) The population refers to the corresponding year.
-
Table 2 summarizes the three inequality measures of the public health workforce distribution in each of the three regions — East, Central, and West. Based on regional population sizes and available public health personnel, inequality in each category was higher in the West than the East and Central regions. Gini coefficients ranged from 0.147 to 0.165 nationwide, and ranged from 0.087 to 0.135 in East China, 0.127 to 0.144 in Central China, and 0.113 to 0.186 in West China during 2008 to 2017. Theil L and Theil T inequalities were almost twice as high in the West than in the East. The density of the public health workforce in West China was 1.688 per 10,000 population compared with 1.146 in East China, and 1.404 in Central China in 2017. The density range in the East was 1.146 to 1.372 per 10,000 population compared with 1.602 to 1.688 in the West.
Year Nation East region Central region West region Mean
densityGini Theil L Theil T Mean
densityGini Theil L Theil T Mean
densityGini Theil L Theil T Mean
densityGini Theil L Theil T 2008 1.507 0.165 0.019 0.019 1.372 0.135 0.013 0.014 1.582 0.143 0.016 0.015 1.613 0.186 0.025 0.027 2009 1.494 0.163 0.018 0.019 1.354 0.125 0.012 0.012 1.575 0.144 0.016 0.015 1.602 0.185 0.025 0.027 2010 1.467 0.161 0.018 0.019 1.277 0.118 0.010 0.011 1.565 0.140 0.015 0.014 1.641 0.177 0.023 0.024 2011 1.452 0.158 0.017 0.018 1.257 0.112 0.009 0.010 1.546 0.135 0.014 0.013 1.640 0.171 0.022 0.023 2012 1.433 0.155 0.017 0.017 1.230 0.109 0.009 0.009 1.514 0.135 0.015 0.013 1.650 0.153 0.018 0.018 2013 1.434 0.155 0.017 0.017 1.232 0.113 0.010 0.010 1.497 0.132 0.014 0.013 1.671 0.149 0.017 0.018 2014 1.412 0.151 0.016 0.016 1.208 0.103 0.008 0.009 1.464 0.127 0.014 0.012 1.665 0.141 0.016 0.016 2015 1.393 0.149 0.015 0.016 1.188 0.096 0.008 0.008 1.434 0.128 0.014 0.013 1.660 0.133 0.014 0.014 2016 1.389 0.147 0.015 0.015 1.180 0.092 0.007 0.008 1.421 0.129 0.014 0.013 1.671 0.117 0.012 0.012 2017 1.374 0.150 0.016 0.016 1.146 0.087 0.007 0.007 1.404 0.128 0.014 0.013 1.688 0.113 0.011 0.011 Table 2. Density and inequality measures by regional strata and year, 2008–2017
-
Figure 3 shows that the public health workforce density varied significantly across the 31 provinces. Density decreased from north-west to south-east in 2008 and 2017. In 2008, Tibet, Qinghai, Xinjiang, Inner Mongolia, and Jilin had the highest densities, while Chongqing, Anhui, Zhejiang, Guangdong, and Fujian had the lowest. In 2017, the western provinces of Tibet, Qinghai, Xinjiang, and Inner Mongolia had the top four densities, whereas Jilin had decreased its density from 2.39 to 1.75 per 10,000 population in the ten years. Changes in public health workforce density led to a continuous decrease in 24 provinces ranging between 1.48% and 28.26%; there were five provinces — Guangxi, Guizhou, Sichuan, Chongqing, and Shaanxi — with more than a 12% increase in public health workforce density per 10,000 population from 2008 to 2017 (Figure 4).
-
The yearly global Moran’s I indices of the public health workforce were 0.460, 0.454, 0.485, 0.470, 0.464, 0.442, 0.422, 0.408, 0.405, and 0.415 over the period 2008–2017, showing a significant spatial autocorrelation during the study period that indicates provincial clustering of public health workforce across provinces.
The local spatial autocorrelation distribution of the public health workforce is shown in Figure 5. From 2008 to 2017, the H–H and L–L cluster provinces experienced little change. The H–H clusters were located in the western provinces, while the L–L clusters included Guangdong and Fujian. Most provinces had similar public health workforce distributions in 2008 and 2017, with the exceptions of Sichuan and Yunnan, which did not have H–H clusters in 2008 but did in 2017. This indicates that Sichuan and Yunnan had a relatively higher level of improvement in workforce density compared to neighboring provinces during the study period.
Examining Inequality in the Public Health Workforce Distribution in the Centers for Disease Control and Prevention (CDCs) System in China, 2008–2017
doi: 10.3967/bes2020.051
- Received Date: 2019-12-27
- Accepted Date: 2020-05-09
-
Key words:
- Inequality /
- Public health workforce /
- CDCs /
- Gini coefficient /
- Geographical distribution
Abstract:
Citation: | LI Yuan Qiu, CHEN Hao, GUO Hao Yan. Examining Inequality in the Public Health Workforce Distribution in the Centers for Disease Control and Prevention (CDCs) System in China, 2008–2017[J]. Biomedical and Environmental Sciences, 2020, 33(5): 374-383. doi: 10.3967/bes2020.051 |