HomePracticeIs the commuted overtime model the best option for SA hospitals?

Is the commuted overtime model the best option for SA hospitals?

Could South Africa’s over-burdened public health system and its exhausted doctors change to an alternative, shift-based staffing model that would make economic sense? J Mthethwa, H Hlela, T Moji, S Mtshali, and O Mahomed weigh up the fiscal implications in the SA Medical Journal.

They write:

South Africa’s public health system relies on commuted overtime (COT) to provide 24-hour medical cover in hospitals, but there have long been concerns about excessive working hours, fatigue and patient safety, renewing interest in shift-based staffing models.

Local evidence on the workforce and fiscal implications of such a transition remains limited, however.

Our recent study undertook to determine the staffing complement required to deliver continuous 24-hour surgical coverage under a shift-based system, and to model the personnel cost implications across three staffing configurations compared with the current COT arrangement

Our findings suggest that replacing COT with a 12-hour shift system would require a substantially greater number of medical officers (MOs), and increases total personnel costs by 48%-79% when estimated overtime is included in the baseline. Hybrid models reduced but do not eliminate this impact, suggesting that any transition should be phased, locally costed using verified payroll data, and aligned with national workforce planning priorities.

Clinical outcomes and optimal staffing levels were not assessed in this retrospective cost analysis at a regional hospital in eThekwini, KwaZulu-Natal, using the Department of General Surgery as the case study.

Comparisons

For the study, three staffing models were compared: the current COT model; a hybrid model with medical officers (MOs) on 12-hour shifts and specialists retained on COT; and an expanded hybrid model with additional specialist posts. Annual personnel costs were calculated using 2024/25 occupation-specific dispensation salary notches under grade 1 and mixed-grade assumptions. Sensitivity analyses examined variations in specialist salaries, workload estimates and absence allowances.

The baseline complement comprised eight MOs and four specialists. Delivering continuous shift-based coverage required scaling the MO complement to 22, a 175% increase. On a base-salary-only comparison, annual costs rose from R13.54m (all grade 1) to R27.56m under hybrid model A (+103.7%) and to R31.71m under hybrid model B (+134.3%).

When estimated COT payments are incorporated into the current model baseline – using group 3 parameters and a 1.3 × ordinary hourly rate multiplier applied to 16 contracted overtime hours per week – total estimated annual remuneration rises to R20.58m (all grade 1) and R23.21m (mixed grades).

Under this COT-inclusive comparison, hybrid model A represents a cost increase of ~48% and hybrid model B ~79%.

Shift-based models increase total personnel expenditure under both costing approaches; the COT-inclusive figures are the primary policy-relevant comparator. Sensitivity analyses confirmed robustness across all scenarios tested.

Cost and labour intensive

Providing continuous medical cover is a core requirement of hospital care, yet it remains one of the most resource-intensive aspects of service delivery. Both routine and emergency services, especially those operating outside regular hours, rely heavily on doctors, and the structure of their work has a direct impact on patient safety, workforce sustainability and costs.

International experience shows that systems relying on extended duty hours differ markedly from those using regulated shift models, with important trade-offs among staffing levels, fatigue and expenditure. Decisions about after-hours cover are therefore not only operational, but fundamentally economic.

In South Africa’s public sector, after-hours medical services are predominantly delivered through commuted overtime (COT): doctors work a standard 40-hour week with additional contracted overtime hours. COT has enabled hospitals to maintain 24-hour services despite longstanding staff shortages.

Our evidence from a large tertiary hospital showed that interns frequently worked substantially more overtime than they were rostered for, often without formal recognition or compensation.

At the same time, national analyses have identified COT as a major and growing component of provincial compensation budgets, amounting to several billion rands annually. These trends have intensified policy interest in alternative staffing models, including shift-based systems.

Despite this policy focus, there is little local evidence on the fiscal implications of replacing COT with shift-based arrangements in our public hospitals.

International studies on shift systems largely originate from high-income settings and are not directly applicable to the SA context, where staffing constraints, remuneration structures and budget ceilings differ substantially.

Practicality

Our study was conducted at a regional 500-bed public hospital in eThekwini. The Department of General Surgery, which served as the unit of analysis, offers 24-hour services and is responsible for 112 inpatient beds across multiple wards, including high care and ICUs. It operates a surgical outpatient clinic five days per week and runs two operating theatres daily for elective and emergency cases.

The average monthly activity consists of ~180 general surgical cases and 200 gastrointestinal procedures. Continuous senior and junior medical cover is required.

Salary data were obtained from official provincial occupation-specific dispensation (OSD) salary notches for the 2024/25 financial year, as published by the Department of Public Service and Administration (DPSA).

Three staffing models were evaluated. The current model (COT) reflects existing practice. The department is staffed by eight medical officers (MOs) and four specialists. All doctors are contracted for a 40-hour work week with COT to provide after-hours cover.

Two hybrid shift-based models were developed to reflect a transition to 12-hour shifts for MOs while retaining specialists on COT.

In hybrid model A, the MO complement was increased to 22 and employed on 12-hour shifts. Four specialists were retained on COT for call duties.

In hybrid model B, 22 MOs were employed on shifts, and the specialist complement was increased to seven. Specialists remained on COT. The configuration of seven specialists was derived from a clinical supervision ratio analysis: with 22 MOs rostered across a 24-hour cycle, ~3 – 4 MOs would be concurrently on duty per shift. Applying a conservative 1:3 specialist-to-MO supervision ratio, consistent with clinical training and governance standards in Commonwealth surgical contexts, to concurrent shift coverage yields a minimum requirement of six or seven specialists to ensure immediate specialist availability on each shift.

So seven was selected as the upper bound under conservative supervision standards.

Hybrid model A, with four specialists, represents the lower bound, preserving the current specialist complement unchanged as minimum feasible configuration.

A full shift system for specialists was not modelled, and considered impractical given current specialist shortages in the public sector and existing contractual arrangements.

Staffing requirements

Under the COT model, the Department of General Surgery was staffed by eight MOs and four specialists. Workload modelling indicated a requirement of ~616 doctor-hours per week for continuous service coverage. After applying a 25% allowance for leave, training and other absences, this translated to a requirement of 22 MOs to sustain a 12-hour shift system – a 175% increase in the MO complement compared with the current establishment.

Personnel costs: All grade 1 assumptions

These figures reflect base salary costs only, and are retained as a methodologically conservative: they don’t include overtime, employer contributions, or allowances.

Assuming all staff were appointed at grade 1 entry-level salaries, the current COT model costs R13.54m per year. Hybrid model A, with 22 MOs on shifts and four specialists retained on COT, resulted in annual personnel costs of R27.56m, representing a 103.7% increase relative to the baseline. Hybrid model B, which increased the specialist complement to seven, further increased costs to R31.71m, a 134.3% increase.

Personnel costs: Mixed-grade assumption

Under a mixed-grade distribution (50% grade 1, 25% grade 2 and 25% grade 3), the current COT model incurs an annual cost of R15.27m. Hybrid model A increased costs to R30.91m, a 102.4% increase, while hybrid model B increased costs to R35.66m, a 133.5% increase.

Employer contributions were excluded from the primary analysis to maintain comparability across models. For planning purposes applying a conservative 20% loading to account for Government Employees Pension Fund contributions (16% of pensionable salary), medical scheme contributions and statutory levies yields the following indicative employment costs under all grade 1 assumptions: R24.70m for the current COT-inclusive model, R36.52m for hybrid model A and R44.09m for hybrid model B.

Discussion

Replacing COT with a 12-hour shift system for continuous surgical cover would require a large increase in staff and more than double the annual personnel costs. Across all scenarios, costs increased by >100% and, in some cases, by >130%. However, when estimated COT payments are incorporated into the current model baseline, the fiscal differential narrows substantially.

Under COT-inclusive assumptions using group 3 parameters and a 1.3 × ordinary hourly rate multiplier, hybrid model A increases total estimated remuneration by ~48% and hybrid model B by ~79%. The findings are consistent across both costing approaches: shift-based models increase total personnel expenditure in every scenario tested. However, the magnitude of the fiscal gap is substantially more moderate in the COT-inclusive analysis.

This reflects a structural feature of the shift model that is obscured in base-salary comparisons: replacing contracted overtime with a larger headcount eliminates the 52% overtime premium for the MO cohort, thereby partially offsetting the cost of employing 14 additional doctors.

This distinction matters considerably for policy. A finding that shift models double personnel costs risks being dismissed as fiscally prohibitive without further deliberation. A finding that shift models increase total remuneration by 48%-79%, while simultaneously eliminating the overtime premium currently paid to the MO cohort and replacing informal excess labour with formally rostered coverage, is a meaningfully different policy signal.

It suggests that the transition, while still costly, involves a partial reallocation of existing overtime expenditure into headcount rather than a straightforward doubling of the wage bill.

The increase was driven by headcount rather than salary levels.

To provide 24-hour cover without extended working hours, the number of MOs had to increase from 8 to 22. This was necessary once a standard working week and leave allowances were applied. The cost increase, therefore, reflects the need to employ more doctors to deliver the same volume of service.

Under COT-inclusive assumptions, each additional MO adds ~R1.52m per year in total remuneration at grade 1, compared with ~R1.0m on a base salary-only basis. The difference reflects the overtime premium that would have been paid had those posts been COT-contracted rather than shift-based.

The cost increase therefore reflects the need to employ more doctors to deliver the same volume of service, partially offset by the elimination of overtime payments for the expanded MO cohort.

These findings are consistent with existing evidence on public sector staffing pressures in SA. National expenditure reviews have shown that COT has grown steadily and now accounts for a substantial share of provincial compensation budgets, reaching almost R15bn
by the 2023/24 financial year.

Much of this growth has been linked to increases in the number of doctors rather than changes in overtime rates.

Our findings reflect the same dynamic. When systems replace extended hours with shifts, staffing numbers rise, and costs follow.

It is equally important to acknowledge that the current COT model, as costed here, reflects only base salaries.

The true cost of the current model is higher than presented, indicating that the fiscal gap between the COT and shift models is smaller than the figures in this study suggest. The reported cost increase of ~100% – 134% should be interpreted as an upper bound. It is also worth noting that the current COT system has been subject to documented governance concerns beyond its fiscal scope.

The Government Technical Advisory Centre’s technical review and successive reports from the Auditor-General of SA have identified structural weaknesses in COT monitoring and verification, including irregular payments for unverified overtime claims and cases where
COT allowances were paid without corresponding documented service delivery.

If a proportion of current COT cost represents payment for work not performed or inadequately verified, then the governance-adjusted cost of the current model may be lower than its nominal cost, and the relative cost advantage of a formally rostered shift system, which includes built-in attendance monitoring, may be understated in purely salary-based comparisons.

Future analyses should incorporate an estimate of governance leakage in COT cost to further refine the comparison.

At the same time, concerns about the current COT system are well described. Studies from large academic hospitals have shown that doctors often work far beyond their rostered overtime hours, sometimes exceeding 150% of their scheduled time. These hours are not always formally recorded or compensated.

This suggests the current model relies partly on excess labour. While this limits immediate cost, it raises concerns about fatigue, patient safety and the long-term sustainability of the workforce. Removing this informal buffer has significant financial implications.

A critical nuance in interpreting the cost differential between models is the role of this uncompensated excess labour in the cost comparison. If the current eight-MO complement delivers its coverage in part because individual MOs work substantially more than their contracted hours without formal compensation, then the current model achieves its apparent cost efficiency in part through a systematic underpayment of medical labour.

Were this labour to be properly costed, either by paying at contractual overtime rates for actual hours worked, or by employing additional staff to cover excess uncontracted hours, the true cost of the current model would be higher than presented here. This means the true fiscal gap between COT and a formally rostered shift model is narrower than the base salary comparison this study suggests.

This observation does not alter the finding that shift models are more expensive on a base-salary basis, but it is an important policy consideration: the apparent cost efficiency of COT may be partly illusory and is neither equitable nor sustainable over the long term.

International experience points in the same direction. Countries that introduced stricter controls on working hours required additional staff to maintain services. These reforms reduced fatigue and improved compliance, but increased costs and disrupted established service arrangements. Importantly, these changes occurred in health systems with greater staffing depth and fiscal flexibility than those found in SA.

The local environment is therefore considerably more constrained.

From a practical perspective, the funding implications are substantial. Each additional MO adds ~R1m in base salary per year. Increasing the complement from 8 to 22, therefore, adds ~R15m annually for a single department. Provincial health finances would be severely strained if this strategy were implemented in all regional hospitals.

Beyond the direct fiscal implications modelled here, the operational feasibility of expanding the MO complement by 175% must be considered within SA’s current public-sector hiring environment. Most provincial health departments are subject to personnel cost ceilings imposed by provincial treasuries.

In many provinces, these ceilings have translated into post freezes.

Creating 14 additional funded MO posts in a single department at a single hospital would require additional budget increases, new work studies to determine exact numbers required for each department and each facility, and approvals at multiple levels of government, which could take years and is not guaranteed in the current fiscal climate.

In several provinces, personnel costs already consume 60%-70% or more of total provincial health budgets, leaving minimal flexibility for new headcount. The finding that a 175% increase in MO complement is required should therefore be read not only as a cost finding but as a signal that the structural and administrative feasibility of reform is at least as significant a constraint as the fiscal one.

Limitations

This study has several limitations. First, while COT-inclusive estimates have now been incorporated, these are based on group 3 planning assumptions and a 1.3 × multiplier consistent with the applicable national policy document, rather than PERSAL-verified payroll data. Actual COT costs depend on individual salary notch placement, exact hours claimed and the applicable KZN provincial tariff schedule.

Second, the 616 doctor-hours-per-week workload estimate – driving the entire staffing calculation – was derived from a coverage requirements analysis using actual departmental rosters and operational data, rather than a prospective time-motion study.

Third, the 25% absence allowance, while grounded in statutory leave provisions, may under-estimate actual absenteeism in the public sector.

Fourth, the analysis assumes equivalent service output across both the current COT model and the proposed shift models, but this assumption may not hold in practice if the current model relies partly on excess uncontracted hours.

Fifth, employer contributions (pension, medical aid, UIF) of ~20% of base salary, and Sunday and public holiday allowances applicable to shift workers, were excluded.

Sixth, this is a single-department, single-hospital study. Results may not generalise to other service types

Finally, transition costs, including recruitment, onboarding, training and induction for 14 additional MOs, were excluded from the analysis.

Conclusion

In this regional hospital case study, introducing a 12-hour shift system for 24-hour surgical coverage required a substantial increase in MO numbers. On a base salary-only comparison, this more than than doubled personnel costs. However, when estimated COT payments are incorporated into the current model baseline – using group 3 parameters and a 1.3 × ordinary hourly rate multiplier – the fiscal differential narrows considerably, with hybrid shift arrangements increasing total estimated remuneration by ~48%-79% rather than the 102%-134% suggested by base salary comparisons alone.

Hybrid staffing arrangements, where MOs move to structured shifts while specialists are retained on COT, reduced but did not remove this increase across both costing approaches.

The current COT model remained the least costly option in every scenario tested.

While shift systems may address legitimate concerns about working hours, fatigue and patient safety, this analysis suggests that their base salary cost implications are substantial.

Our findings should be interpreted as indicative estimates of the fiscal order of magnitude involved in a transition, not as precise projections suitable for direct budget submission.

Any policy move away from COT should therefore begin with facility-level costing using actual COT cost data, followed by phased piloting with embedded evaluation.

Alignment with national workforce planning frameworks and the emerging NHI contracting environment will be essential. Further research is needed, particularly on whether the potential benefits of shift systems, including reductions in fatigue-related adverse events, employment opportunities for unemployed doctors, burnout and Medico-legal exposure, offset the additional personnel costs in the SA public health context.

J Mthethwa,1 MPH, PhD; H Hlela,1 MB ChB, PGDip-Health Management; T Moji,2 MB ChB, MPH; S Mtshali,2 MB ChB, PhD; O Mahomed,3 MMed (PHM), PhD

1 Dr Pixley Ka Isaka Seme Memorial Hospital, KwaZulu-Natal Department of Health, Durban, South Africa
2 General Clinical Services, KwaZulu-Natal Department of Health, Pietermaritzburg
3 Discipline of Public Health Medicine, College of Health Sciences, University of KwaZulu-Natal

 

SA Medical Journal article – The fiscal cost of 24-hour clinical coverage: Commuted overtime v. shift-based models in a South African regional hospital (Creative Commons Licence)

 

See more from MedicalBrief archives:

 

Overtime non-payment for doctors sparks outrage in Gauteng

 

Exploited, exhausted ‘ghost doctors’ trapped in ‘new apartheid’

 

Gauteng plan to review doctors’ overtime pay policy a ‘disaster’

 

Gauteng ‘crisis’ warning as specialists resign from hospitals

 

 

MedicalBrief — our free weekly e-newsletter

We'd appreciate as much information as possible, however only an email address is required.