Discussion
The GSHRH highlighted a projected health workforce shortage of 18 million health workers by 2030.3 It identified gaps in data reporting and triggered dedicated policies and efforts to monitor health workforce through the progressive implementation of the NHWA.17 Five years on, data availability and quality significantly improved allowing the present reassessment to take place. The latest NHWA data for year 2020 measures a total global stock of 65.1 million health workers (dentists, medical doctors, midwives, nurses, pharmacists and other health workers) which indicates an increase of 14 million net jobs (a 27% increase) as compared with 2013. This growth in the size of the health workforce is >3 times faster than the population growth rate over the same period. Although some data improvement could have had a role in the change in reported number, it is important to note that the figure used for 2013 is not the number that was reported in the Global Strategy on Human Resources for Health: workforce 2030, but on a revised estimate that includes the retrospective reporting, hence could be now considered as a solid reference stock.
The analysis to 2030 indicates a projected growth rate of 2.7% annually (as compared with an 0.9% growth of global population), which represents an increase in 19 million net jobs that could reach 84 million health workers in 2030. This increasing stock of employed health workers in the health economy is consistent with the predictions stated in the GSHRH3 and the UN High Level Commission on Health Employment and Economic Growth.18 Both identified strong economic demand—mostly in upper middle-income and high-income countries—that could create up to 40 million additional jobs by 2030. In spite of this consistency, the 2030 estimate of 84 million health workers should be viewed with a degree of caution as it is based on the prevailing graduation rates and cannot as yet factor in the multidimensional impact of the COVID-19 pandemic neither on healthcare workers19 20 nor on the education market.
Unequivocally, the pace of progress has significantly reduced the previously estimated global shortage to 15.4 million in 2020 and is projected to further decline to 10.2 million in 2030: lower than that estimated in the GSHRH3 or through other methods.21 Part of the this can be explained by the availability of better and more recent data allowing for a better assessment of the shortage. For instance, the GSHRH estimates predicted a 17% decrease in shortage between 2013 and 2030, while this reassessment estimates a 49% decline between 2013 and 2030, largely explained by better data on additional health occupations and their collective contributions in expanding access to UHC.
This global progress is however uneven. WHO African and the Eastern Mediterranean regions did not show improvements of their health workforce shortage between 2013 and 2020, which is projected to remain stagnant by 2030. One possible explanation is their notable population growth rate of 28% and 18%, respectively, between 2020 and 2030, compared with about 5% in other WHO regions. Our findings further suggest that at this pace, the projected 2030 shortage will be clustered in 47 countries (mostly in WHO African region), included in the WHO HWF-SSL and between them encompass two-third of the global shortage by 2030. This is an alarming situation for the African region’s heath systems that are facing an increasing demand for health services by a growing population size and the rising demand to accelerate progress towards the SDG health-related targets.22 Despite the clear focus on SIDS articulated in SDG3.c, these countries show less progress in reducing their shortages in comparison with other countries. SIDS countries are facing several challenges, such service delivery as in hard-to-reach areas and limited national education capacity, which may explain the modest reduction of the shortage. This finding underscores the urgency for dedicated actions and support to SIDS countries.
Limitations
While this paper presents an estimate for year 2020 with latest data for almost all countries, it must be noted that in the case of few countries’ data the latest year available was sometimes >3 years earlier than 2020 (online supplemental appendix 5). For most countries, the density used for year 2020 and applied to the population of 2020 came either from 2018, 2019 or 2020 (78% of countries for dentists, 76% for medical doctors, 71% for midwifes, 83% for nurses and 64% for pharmacists). For these countries, the potential change in density between 2018 and 2020 can be considered as negligible. However, for the remaining countries, when latest data dated before 2018, there is a slight possibility of underestimating the stock of 2020 given that density expected to have increased over time. Alternative approaches could be used, such as multiple imputation, to fill empirical data gaps, however those would require specific development to include both the profiling of countries with missing data and the interpolation of missing values in trends by occupation group.
The method used for projecting the global health workforce to 2030 relied on a simple stock and flow model. It has however a few limitations. First, the model assumes a fixed production capacity throughout the period. The model used is therefore conservative and may slightly underestimate or overestimate the production capacity of individual countries over time. Second, no data on health worker migration were used in the modelling. It can be expected that the attrition factor due to outmigration varies by country and might affect the shortage estimate. In particular, the COVID-19 pandemic could have disrupted international mobility and countries most dependent on international migration, such as high-income countries, may experience a slower progression of their overall stock to 2030. On the whole, the impact of attrition due to migration is difficult to estimate, however it can be speculated that the degree of potential variation on our estimation would be equivalent to the average share of foreign-trained nurses from OECD Stat of about of 6.3%.15 It can be expected that countries in African region and South-East Asia region would be affected also by having higher retention of health workers during this period, but this might not necessarily have translated into new jobs. Third, the age distribution used as an attrition factor was considered as fixed. The age at retirement was set at 65 years, which is in line with data from OECD.14 A faster attrition rate might occur because the majority of health workers are women,23 who may retire on average at a slightly younger age than men. Fourth, our reassessment assumed that 70% of new graduates would be constantly employed over the coming years. It implies that 30% of graduates would be trained as health workers but being unemployed or working in other sectors than health such as in public administration, in education, etc. As a consequence, the overall stock of health workers in each country by 2030 would be higher than the estimation of practising health workers considered in the present paper. This fixed 70% share of graduates becoming employed is an uncertain assumption in light of potential fiscal space constraints linked to the social and economic impact of the COVID-19 pandemic.24 A sensitivity analysis was also conducted and considered ‘what if’ only 50% of new graduates were absorbed in the health sector annually in all countries, the projected workforce would be around 75.9 million health workers by 2030 as compared with 84 million with a 70% absorption capacity. If all countries were to experience a major reduction of 30% absorption capacity, the projected health workforce would be 67.5 million health workers by 2030. These sensitivity analyses stipulate the necessity to sustain a high absorption capacity in the years leading to 2030. Beyond the impact of COVID-19, the absorption capacity of new graduates could also vary by income group and by occupation. A sensitivity analysis using a varying absorption capacity relative to income groups showed a projected stock of 76 million health workers and a shortage of 12.9 million by 2030. Fifth, the attraction to and retention of jobs in the health sector (postpandemic) may become a relative unknown given the negative impact of attrition due to concerns over higher risks of infections, risks to mental health and well-being, stigmatisation19 among others. Put together however, these limitations are least problematic for the short-term projection (over 10 years) presented in this analysis.
The shortage estimation also relies on a global comparison of densities of occupation groups and therefore provide a broad indication of the progress that would be required for countries with lowest density to scale-up their workforce relative to the threshold median density. It also suggests that health systems would become better equipped to ensure essential health services provision without necessarily reaching the UHC service coverage index (SCI) of at least 80.
Therefore, the shortage of health workers can be estimated using alternative methods and scenarios (online supplemental appendix 4). For instance, between 2020 and 2030, our analysis showed that 10 million health workers would retire globally, half of those are in high-income countries. Factoring in the population growth, an alternative scenario on the number of health workers required to compensate attrition due to retirements estimates that 13.7 million health workers will be needed by 2030 primarily in high-income countries.
Alternative scenarios can also consider target densities to reach a specific level of essential service coverage under the assumption of a perfectly causal relationship between UHC SCI and health workforce density. Such scenario would produce an estimated shortage as high as 78 million if all countries of the world were to reach a UHC SCI level of 80 in 2030 (online supplemental appendix 4).