Introduction
Overview of the public health sector in Ghana
Type of health facility | Number | Percentage (%) |
---|---|---|
CHPS | 3,463 | 76.8 |
Health Centre | 871 | 19.3 |
Polyclinic | 36 | 0.8 |
District (Primary) Hospital | 127 | 2.8 |
Regional Hospital | 10 | 0.2 |
Grand total | 4507 |
Tools and methods
An overview of the WISN implementation and staffing norms development in Ghana
No. | Generic WISN steps | How it was applied in Ghana |
---|---|---|
1 | Governance and technical processes | Following a capacity building workshop facilitated by WHO, a National Steering Committee (NSC) was established to provide political and technical leadership for the application of WISN for the purpose of developing staffing norms in Ghana. The NSC also led in mobilising funding for the process. A 17-member Technical Working Group (TWG), drawn from various agencies was also constituted to undertake the WISN application. The TWG routinely reported the progress of work to the NSC and received guidance as and needed. In each health facility that was visited, an Expert Group was formed by occupational category to assist in the setting of activity standards |
2 | Determining the priorities for WISN application | Based on the policy direction of the Ministry of Health, it was prioritised to apply WISN for all health workers in the country (across 141 categories of clinical and non-clinical staff). In a first phase from 2013 to 2014, 70% of the categories prioritised were covered while the rest were covered in a second phase in 2017–2018 |
3 | Estimating available working time (AWT) for health professionals | In determining the AWT, national leave policy which stipulates the number of days each category of health workers was entitled to was used, alongside, the average number of sick leave taken by health workers which was obtained from each of the health facilities visited, national public holidays, average maternity leave and training days per year were deducted from the total annual working days |
4 | Defining the workload components | The workload components were defined as the tasks (duties) performed by staff on a typical day. These workload components were classified into three: Health Service Activities (or Administrative Activities in case of administration staff), Support activities and Additional activities Health Service Activities refer to tasks performed by all members of a staff category for which regular statistics are collected. Example, number of deliveries, OPD, surgeries etc Support activities are tasks performed by all members of a staff category, but statistics are not collected regularly. Example, documentation of patient care, meetings, etc. Additional activities are tasks performed by some (not all) members of a staff category, but statistics are not collected regularly. Example, administrative duties Data was collected from 54 randomly selected health facilities and institutions to develop the workload components and activity standards which was validated and applied in a nationally representative sample of 138 health facilities countrywide across all levels of the public health system. Expert Groups were formed at the health facilities visited who provided technical insights into their work determine the workload components using a purposely deigned job components tool |
5 | Setting activity standards | Activity Standard (or Service Standard) is the time it takes a trained and well-motivated member of a particular staff category to perform his/her duties to acceptable professional standards in the circumstances of the country/facility. Setting of the activity standards was undertaken concurrently with that of the workload components. Aimed to achieve a technical consensus, the Expert Group in the first health facility provided a list of health service activities they perform, and the corresponding time spent on each. These were then collected and sent to the next health facility, where the completed tool was given to another batch of health professionals (in the same category) to indicate if they agreed with the previous batch of health professionals' proposal. The process continued until a near consensus was achieved where no new workload components were added and the standard time acceptable to all. Where there were still divergent views, non-obtrusive direct observation to determine the standard time was carried out. In all the 192 health facilities used (54 pilot sites for workload components and activity standards development and 136 scale-up application), the institutional staffing requirement was calculated and discussed with the health workers and their management whose comments were used to refine the analysis |
6 | Establishing standard workloads | Standard Workload is the amount of work (within one activity) that one person could do in a year. Standard Workload is the Available working time divided by the activity standard (Service Standard) of a particular task (Standard Workload = AWT ÷ Activity Standard) Standard workload for all activities and categories of staff were determined with the aid of the WISN software |
7 | Calculating allowance factors | Allowance factor (AF) is the estimation of the number of health workers required to cover support activities and additional activities. There are two types of allowance factors—category and individual The category allowance factor (CAF) is a multiplier that is used to calculate the total number of health workers, required for both health service and support activities The individual allowance factor (IAF) is the staff required to cover additional activities of certain cadre members These calculations have been automated in the WISN software |
8 | Determination of staff requirements | In determining the staff requirements, in each health facility, the annual workload statistics was obtained from the annual report, health information system and admission and discharge books in the wards, as appropriate. For each workload component, the annual service statistics was used to divide by its respective standard workload. A sum of all workload components was then put together to get the total staff requirement for all health service activities. The allowance factors are then applied to get the true staffing requirement using the formula below Total required number of staff = (A×B) + C, where A = required staff for health service activities B = category allowance factor C = individual allowance factor The staffing requirements for the individual staff categories and health facilities/institution was computed using the stated formula using the WISN software |
9 | Development of national staffing norms from the WISN analyses | The facility-level WISN results (staffing requirements) were validated and meta-analysed to establish a national staffing norm for the various categories of health facilities. This process included data preparation and validation, statistical analysis for setting staffing norms and validation (a) Data preparation facility-based WISN results were compiled in an excel template for inspection and comparison by facility and staff category. Each facility WISN output (staffing requirement) was assessed for internal and external validity. For internal validation, the facility WISN output was checked to see if the results generally made sense in the light of expert knowledge about the general staffing situation in that facility; and for relativities among cadres in the facility—for example the ratio of doctors to nurses from the WISN results. For external validation, each facility WISN output was assessed to find out if there is any significant difference between that facility and others of similar status and service utilisation. Health facilities were then grouped into workload categories. Whenever unexplained discrepancies were detected, a verification of the inputted data vis-à-vis expert consultation and a re-run of the WISN study was made to correct the errors (if any) (b) Determining the national staffing norms from WISN results: The facility based WISN results (grouped by type of health facility and similarity of workload) was meta-analysed using random effect model of meta-analysis (the random effect model assumes that when pooling results, there could have been variations within and across studies). The pooled average staffing requirements for each cadre based on the meta-analysis and its boundaries of uncertainties (95% confidence limits) were considered the ‘statistical limits’ for setting the staffing norms: The lower limit of the 95% confidence interval of the pooled mean requirement of each category of the staff was considered the minimum staffing limit on the staffing norm for that cadre The upper limit of the 95% confidence interval of the pooled mean requirement of each category of the staff was considered the maximum staffing limit on the staffing norm (c) Validation and adoption The draft staffing norm was then reviewed and validated in a series of consultation and validation workshops across the country with stakeholders across all levels of the health system including health professions regulators, labour unions, and frontline health managers. Feedback from the series of validation workshop was used to finalise the staffing norms document before it was adopted as a national policy for health workforce planning, distribution and management in the public health sector |
Analytical framework
Establishing absolute HRH gaps
Establishing relative HRH gaps (Staff Availability Ratio, SAR)
Crude equity index
Costing of the staffing norms and gaps
-
Gross annual salaries
-
Gross market premium (this is an allowance paid to health workers supposedly in short supply)
-
Other allowances (such as housing, on-call duty facilitation, fuel, utility).
Data sources
Results
Aggregate human resources for health (HRH) gaps by region
Region | Total staff required (a) | Total at post (b) | Absolute HR gaps (c = a−b) | Staff availability ratio (SAR = b/a) (%) |
---|---|---|---|---|
Ashanti | 13,730 | 7854 | 6209 | 57 |
Brong Ahafo | 10,510 | 5009 | 5777 | 48 |
Central | 8283 | 5366 | 3245 | 65 |
Eastern | 14,627 | 6390 | 8579 | 44 |
Greater Accra | 9317 | 8497 | 1041 | 91 |
Northern | 12,716 | 9335 | 4233 | 73 |
Upper East | 6643 | 4011 | 3038 | 60 |
Upper West | 7606 | 3169 | 4757 | 42 |
Volta | 10,800 | 5738 | 5490 | 53 |
Western | 11,208 | 6387 | 5389 | 57 |
National | 105,440 | 61,756 | 47,758 | 59 |
Crude equity index (highest/lowest) | 2.17 |
Descriptive analysis of staff availability across levels of health facilities
Type of health facility | Total staff required (a) | Total at post (b) | Total HR gaps (c = a–b) | Staff availability ratio (SAR = b/a) |
---|---|---|---|---|
CHPS | 14,670 | 10,082 | 7141 | 68.7% |
Health Centre | 29,521 | 15,357 | 14,419 | 52.0% |
Polyclinic | 4211 | 3333 | 879 | 79.1% |
Primary Hospital | 45,068 | 24,817 | 21,094 | 55.1% |
Regional Hospital | 050 | 5505 | 1679 | 78.1% |
District Health Directoratea | 3390 | 1456 | 2048 | 42.9% |
Municipal Health Directoratea | 1371 | 701 | 744 | 51.1% |
Metropolitan Health Directoratea | 162 | 70 | 93 | 43.2% |
National | 105,443 | 61,321 | 48,097 | 58.2% |
Aggregate HRH cost estimates: requirements, deficits and distributional inefficiencies
Region | Total expected cost | Total current cost | Cost of inefficient staff distribution | Total cost of shortage | Proportion of inefficiency to current cost (%) | ||||
---|---|---|---|---|---|---|---|---|---|
(GH¢) | US$ | (GH¢) | US$ | (GH¢) | US$ | (GH¢) | US$ | ||
Ashanti | 307,673,077 | 68,069,265 | 184,695,059 | 40,861,739 | 53,274,031 | 11,786,290 | 176,252,049 | 38,993,816 | 29 |
Brong Ahafo | 232,105,773 | 51,350,835 | 114,796,453 | 25,397,445 | 23,396,508 | 5,176,219 | 140,705,828 | 31,129,608 | 20 |
Central | 182,776,572 | 40,437,295 | 116,987,137 | 25,882,110 | 31,458,889 | 6,959,931 | 97,248,323 | 21,515,116 | 27 |
Eastern | 322,420,490 | 71,331,967 | 146,622,828 | 32,438,679 | 23,017,244 | 5,092,311 | 198,814,906 | 43,985,599 | 16 |
Greater Accra | 216,834,126 | 47,972,152 | 211,487,566 | 46,789,285 | 79,539,377 | 17,597,207 | 84,885,937 | 18,780,075 | 38 |
Northern | 290,656,714 | 64,304,583 | 217,152,425 | 48,042,572 | 78,038,383 | 17,265,129 | 151,542,672 | 33,527,140 | 36 |
Upper East | 148,811,404 | 32,922,877 | 92,074,647 | 20,370,497 | 24,167,133 | 5,346,711 | 80,903,890 | 17,899,091 | 26 |
Upper West | 171,496,126 | 37,941,621 | 69,139,832 | 15,296,423 | 18,871,455 | 4,175,101 | 121,227,749 | 26,820,298 | 27 |
Volta | 235,851,965 | 52,179,638 | 127,988,154 | 28,315,963 | 29,751,003 | 6,582,080 | 137,614,814 | 30,445,755 | 23 |
Western | 249,720,225 | 55,247,837 | 143,387,300 | 31,722,854 | 39,540,309 | 8,747,856 | 145,873,235 | 32,272,840 | 28 |
National | 2,358,346,472 | 521,758,069 | 1,424,331,400 | 315,117,566 | 401,054,332 | 88,728,835 | 1,335,069,404 | 295,369,337 | 28 |
Cost estimates of staffing requirements, gaps and inefficient distribution at various levels of service delivery
Type of facility | Total expected cost | Total current cost | Total cost of inefficient distribution | Total cost of shortage | % of inefficiency to current cost | ||||
---|---|---|---|---|---|---|---|---|---|
GHc | US$ | GHc | US$ | GHc | US$ | GHc | US$ | ||
CHPS | 231,823,176 | 51,288,313 | 191,693,352 | 42,410,034 | 95,147,366 | 21,050,302 | 135,277,190 | 29,928,582 | 50% |
Health Centres | 570,040,243 | 126,115,098 | 320,121,782 | 70,823,403 | 100,744,215 | 22,288,543 | 350,662,676 | 77,580,238 | 31% |
Polyclinics | 98,443,018 | 21,779,429 | 80,600,579 | 17,831,987 | 30,218,681 | 6,685,549 | 48,061,119 | 10,632,991 | 37% |
Primary Hospitals | 1,163,063,130 | 257,314,852 | 634,088,004 | 140,284,957 | 120,239,721 | 26,601,708 | 649,214,847 | 143,631,603 | 19% |
Regional Hospitals | 197,989,928 | 43,803,081 | 149,837,220 | 33,149,827 | 40,162,967 | 8,885,612 | 88,315,674 | 19,538,866 | 27% |
District Health Directorates | 71,349,138 | 15,785,208 | 27,744,496 | 6,138,163 | 3,901,849 | 863,241 | 47,506,490 | 10,510,285 | 14% |
Municipal Health Directorates | 28,256,462 | 6,251,430 | 13,683,188 | 3,027,254 | 3,027,632 | 669,830 | 17,600,906 | 3,894,006 | 22% |
Metropolitan Health Directorates | 3,417,796 | 756,150 | 1,592,694 | 352,366 | 336,300 | 74,403 | 2,161,401 | 478,186 | 22% |
National | 2,364,382,891 | 523,093,560 | 1,419,361,317 | 314,017,990 | 393,778,730 | 87,119,188 | 1,338,800,304 | 296,194,758 | 28% |
Discussion
Conclusions, policy implications and impact
-
Improved budgetary allocation for recruitment of health workers The overall shortfall in health worker availability was estimated at 61,900, but some 14,142 were also attributed to maldistribution, constituting 23% of the national shortfall. Thus, the net shortfall in staffing (if redistribution were to be made) was projected to be 47,758 across all categories of staff. This evidence was presented as part of the 2019 national budget planning process, which contributed to the allocation of additional recruitment for the health sector in 2019 culminating in the recruitment of 13,271 unemployed health workers [40]. However, this evidence-based planning needs to be further strengthened and sustained.
-
Improving health workforce information To ensure sustainability in the monitoring and analysis of the health workforce distribution and equity, a Human Resource Information and Management System, HRIMS (https://www.ghsnewhrims.org/) was developed and deployed within the context of the National Health workforce Account (NHWA) [41] which at the time of writing this paper had a data completion rate of over 90%. The staffing norms and the gap analysis have been integrated into the HRIMS to sustainably repeat the analysis annually for decision-making on a real-time basis.
-
Redistribution strategy Of the 14,142 staff that were deemed to be inequitably distributed, 11,600 (82%) constituted intra-regional distortions which required district- and regional-level redistribution. Only 2542 (18%) of the maldistribution was inter-regional in nature and required headquarters-led or national-level redistribution. Following extensive stakeholder deliberations, a draft redistribution concept was adopted, costed and the possible efficiency gains analysed over 5 years (this section is being reported in a separate paper).
-
Health workforce production (training and education) Although this analysis did not include supply-side analysis of the health workforce, the results, taken alongside previous works [12], paint a picture of education market failure where the current production of HRH seemed not to be matching health workforce need. For instance, there is the above-optimal availability of staff cadres such as auxiliary nurses or Enrolled Nurses (SAR:139%), Dental Prosthesis Technicians (SAR: 200%), IT Managers (SAR: 660%), Public Health Officers-Nutrition (SAR: 220%) and Opticians (SAR: 380%). It would thus, be imperative for a comprehensive health labour market analysis to engender evidence-based policy dialogue at the highest level to correct the current education market failure in favour of the production of staff cadres high in demand but currently short in supply. The current situation(s) of only 24% of hospitals having Specialist Surgeons, 82% of Primary Hospitals having no Paediatricians and all primary hospitals lacking appropriately qualified Emergency Medicine Physician and Psychiatrist could easily be corrected if the norms were to inform the production of the health workforce.
-
Leveraging task-sharing Given that about 90% of the government’s subvention to the public health sector goes into the payment of employee compensation, the fiscal space for additional recruitment of HRH is increasingly constrained. As a result, GHS Leadership could consider deepening its already adopted task-sharing approach, which allows middle- and lower-level health professionals to assume duties and activities hitherto not part of their traditional roles. Allowing these middle- and lower-level health professionals to perform hitherto untraditional roles, of course with additional training, will bring about a rapid expansion of access to essential healthcare services, increase efficiency, and reduce health worker training and wage bill costs [42].