Modelling the supply and need for health professionals for primary health care in Ghana: Implications for health professions education and employment planning

Background The health workforce (HWF) is critical in developing responsive health systems to address population health needs and respond to health emergencies, but defective planning have arguably resulted in underinvestment in health professions education and decent employment. Primary Health Care (PHC) has been the anchor of Ghana’s health system. As Ghana’s population increases and the disease burden doubles, it is imperative to estimate the potential supply and need for health professionals; and the level of investment in health professions education and employment that will be necessary to avert any mismatches. Methods Using a need-based health workforce planning framework, we triangulated data from multiple sources and systematically applied a previously published Microsoft® Excel-based model to conduct a fifteen-year projection of the HWF supply, needs, gaps and training requirements in the context of primary health care in Ghana. Results The projections show that based on the population (size and demographics), disease burden, the package of health services and the professional standards for delivering those services, Ghana needed about 221,593 health professionals across eleven categories in primary health care in 2020. At a rate of change between 3.2% and 10.7% (average: 5.5%) per annum, the aggregate need for health professionals is likely to reach 495,273 by 2035. By comparison, the current (2020) stock is estimated to grow from 148,390 to about 333,770 by 2035 at an average growth rate of 5.6%. The health professional’s stock is projected to meet 67% of the need but with huge supply imbalances. Specifically, the supply of six out of the 11 health professionals (~54.5%) cannot meet even 50% of the needs by 2035, but Midwives could potentially be overproduced by 32% in 2030. Conclusion Future health workforce strategy should endeavour to increase the intake of Pharmacy Technicians by more than seven-fold; General Practitioners by 110%; Registered general Nurses by 55% whilst Midwives scaled down by 15%. About US$ 480.39 million investment is required in health professions education to correct the need versus supply mismatches. By 2035, US$ 2.374 billion must be planned for the employment of those that would have to be trained to fill the need-based shortages and for sustaining the employment of those currently available. Based on a scoping review of various analytical applications of needs-based health workforce planning approach [21], and building upon previous works [4, 22–26], we developed a needs-based analytical health workforce planning model, built in Microsoft® Excel and suitable for health sector-wide application in any country [27]. The paper reports on how the model was applied to conduct health workforce projection in PHC in Ghana. As illustrated in Fig 2, the underpinning conceptual framework of the model consists of need analysis, supply analysis, gap analysis, and resource implications. Source: Adopted from Asamani et al. [27]. The need analysis assumes that the required calibre and quantity of health professionals are derived from the population’s need for health services [2, 5, 28, 29]; as the population in any jurisdiction at any given time have some need for health services, regardless of whether or not they have demanded such service or if it can even be afforded [2]. Therefore, forecasting the optimal requirement for health professionals in a health system must directly flow from an estimate of the population’s need for health services which can be modelled as a function of (a) the size of the population and its demographic characteristics (b) the state of health or level of health of the population and (c) the level of services (type and frequency) necessary to attain and maintain optimal health of the population. The aggregate need for health services can then be translated into the health workforce requirements if the category of health professional competent to deliver the service was identified [24] with clear work division [26], and an established measure of their standard workload (or productivity) [4, 25]. Box 1 summarises the empirical formulae for estimating the need for health services and translating the same into health workforce requirements. See S1 File for the fully functional excel-based model. Where: Where: Where: Sources: Adopted from Asamani et al. [27]. Building on existing stock and distribution of the health workforce, the model uses the stock and flow approach to estimate health professionals’ future supply. It comprises determining the inflow or entry in the current workforce and the outflow or attrition from the current workforce. While the inflow depends on the training capacity and immigration, the outflow/attrition is influenced by natural and unnatural disengagements such as retirements, emigration, deaths, resignations and dismissals. Box 2 presents the supply projection formula, explained in detail in a prior publication [27]. Where: Sources: Adopted from Asamani et al. [27]. The gap analysis compares the need analysis and the supply analysis to determine if there were a potential need-based shortage or oversupply of the health professionals [30]. Absolute gaps are presented as the actual number of deficit or surplus of health professionals, whilst relative gaps were presented as the proportion of the need-based requirements that the anticipated supply levels could potentially meet. While absolute gaps are essential for planning the number of additional health professionals to train, relative gaps are significant in prioritising the health professional group that required immediate attention. It also has implications for quality of care as it could be interpreted to mean the degree to which professionals standards will be met [31]. Finally, the resource implications are computed in terms of the investment required to sustain jobs for the anticipated supply can be compared with the investment needed to fill the need-based requirement assuming there were no supply-side barriers (see Box 3 for formulae). Where there are need-based shortages, the required investment in health professions education to fill the gaps are estimated. All these cost implications are then compared with budgetary allocations or potential financial space to examine their feasibility. Where: Source: Adopted from Asamani et al. [27]. Input data for the model application (Eqs 1–8) were triangulated from multiple sources. This section highlights the nature and process of data acquisition and data validation procedure and limitations. The population size, gender, age composition (age cohorts) and their distribution by geographical regions and rural and urban distribution were taken from the Ghana Statistical Service projections [32]. The list of diseases and risk factors that accounted for 98% of the burden of mortalities, outpatient attendance and hospital admissions were profiled from the Global Burden of Disease Study [33] and the Facts and Figures report of the Health Sector in Ghana [34, 35]. We then conducted a desk review of survey reports, analytical reports and peer-reviewed publications to map the prevalence or incidence (as appropriate) of the diseases and risk factors identified and the coverage rates of essential public health interventions. The notable sources of the data included the Ghana Demographic and Health Surveys [36, 37], Ghana Maternal Health Surveys [38, 39], the Facts and Figures of the health sector in Ghana [34, 35], Ministry of Health annual holistic assessment report of the health sector [19, 40, 41], and several peer-reviewed publications (see S1 Table for the details of diseases and risk factors with their prevalence or incidence rate as well as the data sources). The main health services that were being provided or were otherwise necessary at the PHC level to address the diseases and risk factors identified above were extracted from various sources, including the Ghana Standard Treatment Guidelines 2017 [42]; CHPS policy [20]; National Reproductive Health Service Policy and Standards [43]; the Non-Communicable Disease Strategy [44]; Ghana National HIV and AIDS Strategic Plan [45]; and clinical management guidelines for TB and HIV infection [46]. The health service activities were matched to the job description of the health professionals being studied. The list of tasks performed by health professionals (matched with the services required to address the disease burden identified above) was extracted from the job descriptions of the respective health professionals and those of a previous Workload Indicators of Staffing Needs (WISN) study [47]. The standard time it would take the individual health professionals to perform these health service tasks were elicited through a nationally representative cross-sectional survey completed by 503 health professionals (the detailed methodology and results of the survey have been reported in a separate piece) [48]. We adopted a workload division established by the Technical Working Group (TWG) on Staffing Norms of the Ministry of Health in 2014 [16, 47]. The TWG adopted a workload division of 25% of medical laboratory workload for the Biomedical Scientist and 75% for Laboratory Technicians, noting that most laboratory examinations, especially at the PHC level, were not complex and hence the role of the Biomedical Scientist was for quality assurance undertaking the 25% of the workload which may require a higher level of technical knowledge or skill. Based on similar logic, pharmacy-related workload division of 20% for Pharmacists and 80% for Pharmacy Technicians was also adopted by the TWG. Based on observed data, the TWG also concluded that Physician Assistants covered 20% of the workload in terms of outpatient consultations in primary/district hospitals and polyclinics, leaving 80% for General Practitioners. However, the Physician Assistant covered 80% of outpatient consultations at health centres, with the remaining 20% taken care of by nurses. We also adopted the workload ratio of 70% for professional nurses (Registered General Nurses) and 30% for auxiliary nurses (Enrolled Nurses) for clinical nursing care. For Midwives, Community Health Nurses, Nutritionists and Dieticians, their functions at the PHC level are usually not shared with other professionals; hence, we made no workload division assumptions. Data from the health sector holistic assessment by the MOH show that between 2014 and 2017, on average, about 20% of outpatient consultations were provided by health professionals at CHPS compounds/zones; 26% at health centres/polyclinics; 49% at primary (and district) hospitals while the rest of 5% was provided in either secondary or tertiary health facilities [19]. We applied the above-mentioned observed trends to the modelled need for health services for the division of service delivery at the various levels of PHC. The existing stock of health professionals, the rate of labour flow (attrition), the education pipeline (number of admissions into health professions education institutions and pass rates) were obtained from the respective professional regulatory bodies of the health professions (as indicated in Table 1). The health professionals’ average income level was taken from the public sector single spine salary scale obtained from the Ghana Health Service. In the absence of comprehensive data on the cost of training of health professionals, we used the average of annual fees paid by fee-paying students as published by two public universities (the University of Ghana’s College of Health Sciences and the University of Health and Allied Sciences) and one private university (the Central University). Since fee-paying students in public and private universities do not benefit from government subsidies and are usually charged at least for full cost recovery for tuition and other costs of training, we assumed that it better reflected the ’true cost’ of training as compared with the regular student fees which the government substantially subsidizes. However, the estimated cost of training per student per year excluded boarding and lodging, which enormously varies depending on the cities and lifestyles of individual students. *a six-year Doctor of Medical Laboratory programme has been introduced by some universities **a six-year Doctor of Pharmacy programme is being introduced by universities, but the 4-year Bachelor of Pharmacy degree is still the basic requirement to be a Pharmacist in the country. Other forms of attrition included resignations, vacation of posts and dismissals. AHPC–Allied Health Professions Council; CU–Central University; GHS–Ghana health Service; MDC–Medical and Dental Council; MOH–Ministry of Health; NMC–Nursing and Midwifery Council; PCG–Pharmacy Council of Ghana; UG–University of Ghana; UHAS–the University of Health and Allied Sciences Data extracted from official reports and websites of MOH, GHS, and the respective health professions regulatory bodies were sent to focal persons in the respective institutions to confirm the validity of the data and explanations was provided for any inconsistencies observed. They also indicated whether or not any subsequent update to the data or report was made and available. Where there was unexplained data inconsistency, a comparison was made with international datasets (if available) such as the World Development Indicators of the World Bank, WHO’s Global Health Observatory (GHO), and the National Health Workforce Account (NHWA) database. Data obtained from peer-reviewed publications were also compared with papers of similar methodology to ensure consistency of estimates. Whenever there was wide variation in estimation between two publications, additional papers were sought for further comparison, and the closest estimates were used. Two of the authors systematically and consistently scrutinised the data before analysis. This study did not involve the use of human subjects as it was based on publicly available data and documents. However, as larger project, the study received ethics approval by the Health Research Ethics Committee (HREC) of North-West University in South Africa with number NWU-00416-20-A1 and the Ghana Health Service Ethics Committee with number GHS-ER17/07/20. Access to administrative datasets was approved by the Director-General of the Ghana Health Service.