This study quantifies for the first time the health impact of health system strengthening, showing that integrating HSS with disease-specific program scale-up delivers substantially greater health benefits and value for money than standalone vertical interventions. The diagonal approach averted almost four times as many DALYs as vertical programs alone, primarily due to accelerated expansion of the health workforce and improved availability of essential consumables. It also proved more cost-effective, with an ICER of $109 per DALY averted against the baseline, compared to vertical HTM expansion, which had an ICER of $132. This aligns with earlier studies showing the importance of addressing system constraints to maximise the effectiveness of disease-specific interventions ^1,9,16.

Numerous previous studies emphasize that disease-specific programs often face diminishing returns in constrained health systems^1,14–17,32. These results support the growing body of evidence advocating for diagonal approaches in health sector investments. By improving workforce capacity and consumable availability, integrated strategies not only enhance the effectiveness of HTM programs but also strengthen the overall sustainability and effectiveness of the health system^33,34.

While our individual-based simulation model accounts for real-world constraints such as workforce shortages and stockouts, uncertainties in epidemiological parameters and cost estimates may affect projection precision. We incorporate stochastic variation in key processes, reporting uncertainty across aggregated runs and consider a number of alternative assumptions through the sensitivity analysis. Although the Historical HRH Scale-up increased the health workforce more rapidly than the Pessimistic HRH Scale- up, it averted fewer DALYs, suggesting that other health system bottlenecks may have constrained the translation of workforce expansion into health gains; the overlap in uncertainty intervals indicates that these differences should be interpreted with caution. Similarly, despite positive health gains for HIV-specific outcomes, overall DALYs and deaths increased under the HIV Program scale-up without accompanying health system strengthening (-0.79% DALYs and -0.47% deaths averted). This counterintuitive result likely reflects the high resource demands of intensified HIV services (such as repeated appointments and follow-up care) outstripping available health system capacity in the absence of expanded human resources and consumables. Consequently, care for other conditions was constrained, leading to a net increase in DALYs and deaths from non-HIV causes.

Our ICER and ROI estimates reflect both direct costs of expanding health system inputs and attributable indirect costs under vertical and horizontal expansion scenarios. However, they do not account for potential additional administrative, regulatory, and implementation costs (above service-level costs³¹) that may arise when operationalizing these scenarios and so we present ROI figures across a range of hypothetical above service level costs over which the ROI remains positive. This is particularly important in the case of modelling improvements in consumable availability, for which our incremental costs include the cost of additional consumables stocked (along with proportional supply chain costs) and not the broader operational interventions that may be needed to enable these improvements.²³. Additionally, our costing method does not systematically account for incremental equipment costs relative to the baseline³⁶. This is because the quantity of medical equipment is assumed to match the levels prioritized in the national strategic plan across all scenarios, due to a lack of data on the throughput of each equipment type.

In the absence of robust estimates of such expenditures tied to specific vertical and horizontal system expansion options, we instead identify threshold levels of above service-level costs over which the diagonal approach would still yield a higher ROI than the vertical approach. The two illustrative thresholds, assuming either no incremental above service level cost for the vertical scenario or a 58% overhead based on estimates from Opuni et al. (2023), serve as lower and upper bounds for interpretation. The fact that the diagonal approach continues to outperform the vertical strategy across such a wide range of above service cost assumptions lends confidence to the robustness of our conclusions.

While our model accounts for several important constraints within the health system, there remain additional factors such as weak information systems, inadequate infrastructure, and governance challenges that could influence both the effectiveness and wider applicability of HSS efforts. In these scenarios, we have assumed an immediate rise in healthcare-seeking behaviour, modelled as almost perfect healthcare utilisation following improvements in appointment availability and workforce numbers. However, increases in demand alone are unlikely to translate into substantial health gains without parallel investments to expand the health system’s capacity³⁷. Although we model the expansion of the health workforce and improvements in supply chains as mechanisms to increase service coverage, it is important to recognise that the success of these measures also depends heavily on factors such as the quality of training, staff retention, and the equitable distribution of personnel across different regions and levels of care ³⁸. Strengthening health systems requires consistent, long-term investment, and the benefits of such efforts are rarely immediate. While our projections are based on relatively rapid improvements, particularly in the expansion of the health workforce, it is likely that real-world progress would unfold more gradually, shaped by local contexts and operational constraints. Although integrating vertical disease programmes with broader system-strengthening strategies can offer important synergies, the reality of limited resources means that tensions and trade-offs inevitably arise. In the simulated scenarios, services are ordered by priority with emergency care addressed first, followed by HIV, tuberculosis, malaria, and subsequently other services. Yet, even with efforts to increase HRH in key programmes, there is a risk that staff may be diverted from other essential areas of care. Careful, context-specific analysis of how resources are distributed, and the consequences of prioritisation decisions, is therefore critical to maximise impact and ensure sustainability.

The evidence from this work points to the need for a significant shift in global health financing ^1,39. With growing financial pressures and changing donor interests, it is no longer viable to rely solely on narrowly focused, vertical programmes if meaningful improvements in health outcomes are to be achieved. What is needed are approaches that combine disease-specific priorities with efforts to tackle broader weaknesses in health systems. Such integrated strategies offer a more practical, resilient, and ultimately cost-effective route to improving health at the population level. They also help to safeguard against the uncertainties of fluctuating donor support, placing countries in a stronger position to deliver a wider range of essential services through their own health systems, with the long-term aim of building sustainability.