Over the past two decades, GIZ supported a number of complex costing studies in Africa and Asia. Looking back, the experts involved in four of the studies share what they learned in the process.

Table 1: GIZ-supported costing studies

Country Years (data collection)	Study overview
Indonesia 2008 2010 – 2011	Two-pronged study: a normative or ‘best practice’ costing model (what should services cost); and, an empirical study (what do they cost in reality), drawing on a nationally representative sample across 15 provinces, 30 districts and 464 public and private health facilities.
Tanzania 2011 – 12	Three-pronged study: ‘best practice’ normative costs; ‘expected practice’ normative costs; and ‘actual’ empirical costs. A nationally representative sample including 155 facilities from dispensary to specialised national hospital (public, faith-based and private sector).
Cambodia 2017 – 2018, 2019; 2020 – 2021	Study comprising two main rounds of data collection. A nationally representative sample across 8 provinces, 25 operational districts and 117 health facilities (excluding national hospitals and private hospitals).
Pakistan 2015 – 2016; 2020	Study comprising two rounds of data collection: 4 District Hospitals (round 1); and 4 private hospitals in the social health insurance programme (round 2) in Khyber Pakhtunkhwa province.

Unpacking the cost and utility of health service costing studies

As we move towards 2030, designing health financing systems that enable access to health services based on need without incurring financial hardship is a policy priority of nearly every country that is working towards Universal Health Coverage. At the same time, global economic pressures – fuelled by interconnected risks of inflation, climate change, pandemics and conflict, to name a few – mean that, governments are increasingly required to pay attention to the value of public spending.

Understanding the value of something requires knowledge of its costs. However, the literature on costing of health services in low- and middle-income countries provides little indication of how the cost data have been incorporated into decision making and price setting (Guiness et al., 2022). This is also true of the studies supported by GIZ, where a lack of documentation has made it challenging to understand how the results have been used, or to gauge their real impact.

In many cases, the primary actors involved in these studies have moved to new posts and different countries. And yet, all the studies generated useful knowledge and have influenced developments in the sector. Below, we attempt to unpack some of the lessons related to the challenges, cost and utility of conducting costing studies, illustrated with examples drawn from the four GIZ countries.

Recognise that costing is highly political

There is no such thing as “real cost” or “true cost.” The delivery of health services does not exist in a vacuum and is affected by ongoing, real-world factors and decisions.

So state the editors in their introduction to the Joint Learning Network manual Costing of health services for provider payment. Not only are health service costs not static, they vary widely in accordance with the approach(es) taken to measuring them. How, and to what extent information on costs is used also depends on many factors, including the legal and policy frameworks in place, power dynamics both in and outside the health sector, and of course the resources available. In other words, costing is highly political.

An in-depth understanding of the political economy is therefore a prerequisite for undertaking costing studies; who is interested in health care costs and why, who bears these costs, and who has the power to use or impede the use of the cost data, are just some of the questions that must be answered before launching a costing study. The extent to which the roles and responsibilities of institutions concerned with financing healthcare are clear – as well as clearly understood – is also important for determining how and to what extent the study results will be used.

Sokunthea Koy, Healthcare Financing Advisor at GIZ in Cambodia, points out that power dynamics play an important role not only between financing institutions, but also within them. A government department assigned to work on a costing study may provide the needed technical and contextual knowledge, but might lack the authority to take decisions based on the results. Reflecting on this, Kai Straehler-Pohl, Programme Manager for the Improving Healthcare Provision project in Tanzania and a Health Adviser for P4H at the time of the study, says,

regardless of what stakeholders might like to do, if those involved in the study do not have the power and autonomy to effect change based on the results, then the study will not achieve its objectives.

Stakeholder analysis can help answer these questions and identify the institutions, and groups within them, that need to be involved from the outset. In addition to being consulted on the costing study scope and design, the different actors need to understand and be prepared for the changes that may follow from the analysis, such as changes in incentives and workloads, and reforms to their systems. Looking back from her vantage point as a Senior Economist at the World Bank, Mariam Ally, who led the study in Tanzania on behalf of the Ministry of Health, recognises that they did not sufficiently appreciate the highly political nature of costing.

Our objective in doing the costing study was to inform pricing, but pricing is not only technical, it is also political. Technically, we worked well, but once we started engaging with politicians, we found it was difficult for them to accept the results – we had not fully understood their role in the study.

Beware of attempting to do too much

It can be tempting to overload costing studies with multiple objectives and to try and collect too much data, particularly in countries where there is limited cost information available in the health sector. As Guiness and colleagues (2022) point out in their review of healthcare cost accounting in low- and middle-income countries, there is an important trade-off between accuracy and resources needed to generate the required cost information.

Key informants for the costing studies in Indonesia, Cambodia and Tanzania pointed to the highly ambitious objectives set for their costing studies, which then influenced the scope and complexity of the study design, resulting in large samples and multiple, complex survey instruments. 

Professor Tim Ensor of Leeds University in the United Kingdom, who advised on the studies in Indonesia and Tanzania, reflects, ‘we thought at the time, we are doing this study so why not collect as much data as possible because there are clear synergies in doing so. But in retrospect, we were not focused enough and we did not have a perfect match between objectives and methods’.

And Franziska Fuerst sums up the views of many when she says,

costings have a valid place in public health and are of course needed, but you can design and implement them with fewer resources. In Cambodia, for example, we did not need to look at so many facilities to understand that they aren’t operating at an optimum level.

Consider starting small and going step by step

Daunted by the size, duration, and complexity of the studies in Indonesia, Cambodia and Tanzania, the GIZ team in Pakistan started small, with a specific achievable objective – to support negotiations on reimbursement rates for district hospitals within the social health insurance programme in Khyber Pakhtunkhwa. Dr Imran Masood Durrani, Deputy Project Manager at GIZ for the Support to Social Protection-Social Health Protection project, says,

coming later, we really benefited from the lessons learned from other German-supported costing studies. I therefore kept the study very small with a very clear view on what we wanted to achieve.

A first study was conducted in 2015-16 and calculated average costs for providing health services at district hospitals. Drawing on the lessons learned and relationships built with stakeholders, the team was able to convince four private hospitals in the same districts to take part in a second study in 2019-20.

Gulrukh Mehboob, the Technical Adviser who led the costing study on behalf of the Pakistan-German Social Health Protection Project says, ‘even with the COVID-19 pandemic, the 2020 study went much more smoothly because we were much more experienced and we understood the limitations of the hospitals and the challenges they faced’ (see also Healthy DEvelopments article 2020).

Importantly, the study allowed the GIZ team to bring together two parties that, for many years, had been in constant conflict with each other over the issue of health service pricing, namely the service providers (i.e., the public and private hospitals) with the purchaser of services (the social health insurance agency.

And, despite the limited size and scope of the studies in Khyber Pakhtunkhwa, the results were shared widely, resulting in requests for similar support from other provinces, and were also taken up by the federal government. 

Don’t underestimate the challenges involved

Discussions among the GIZ colleagues involved in steering the four costing studies brought out the similarities, as well as the size of the challenges they all faced.

Even when foreseen at the start of the study, the lack of cost data available at the health facilities proved to be a considerable challenge across all four countries. Large amounts of data were missing and had to be calculated based on assumptions. This was a particular challenge for calculating costs of medicines and equipment, and for valuing fixed assets, such as buildings. While public sector facilities frequently lacked the data and paper records were also often stored haphazardly, private sector facilities were often reticent to share their data.

In Pakistan, several years of preparation preceded the first costing study, entailing discussions with the social insurance agency, hospital managers and accountants, and close observation of health service and financial reporting processes at the hospitals. These findings fed into the development of the costing study methodology and tools. Even then, data collection was problematic. Flexibility and a willingness to adapt the methodology over time proved critical. As Gulrukh Mehboob explains,

despite our preparations, really critical cost data was missing in some district hospitals, so we had to modify the methodology and we had to think of how to manage these and other challenges as we went along.

In Indonesia, the study sampling methodology used a stratified sampling approach to circumvent some of the operational challenges of conducting a costing study in a country spanning nearly 2 million square kilometres and more than 17,000 islands, many of them very remote. However, the final study design incorporated 464 health facilities, took four years to complete from conception to publication of the initial results, and cost around 3 million Euros.

Ensuring sufficient time for testing survey instruments in different locations, helps to predict and address some of these challenges ahead of time, and alignment of survey instruments with existing reporting systems is also crucial, helping to streamline data collection and to secure participation of the health facility teams. In Cambodia, the data collection tools – set out in three different manuals – were so complex compared with the level of data available at the health facilities that the GIZ field staff tasked with coordinating a component of the study, had continually to seek clarification on how to collect and record the cost data. ‘This really slowed down data collection and was very unattractive to the health facility staff who were participating in the data collection’, explains Sokunthea Koy.

Make full use of the opportunities for capacity development that costing studies provide

In all four countries, the costing studies provided a useful entry point for developing in-country technical capacities, which are indispensable where the aim of the study is to use results for price setting negotiations, or to inform health financing policy and strategy development.

Drawing on many years of support to health service costing, Professor Tim Ensor believes that the best option is to have one or even two places where the study is firmly anchored in government (e.g., the Ministry of Health and/or national social insurance agency) and at least one non-government organisation (e.g., a think tank or university) and to ensure institutional capacity is built across both government and non-government sectors.

Where countries lacked the necessary capacities to undertake the costing study themselves, consultants were brought in to support the process. In Cambodia, an individual consultant worked with the National Institute for Public Health (NIPH), which itself played a leading role in the study design and implementation (see also Healthy DEvelopments article 2019). Recognising the lack of health financing expertise available in the country, the NIPH went on to develop post-graduate courses in health financing and health economics with the aim of developing a pool of national specialists.

However, building capacity takes time and investment which is not always possible within the constraints of development partner project cycles and budgets. In Tanzania, the costing study was contracted out to a consulting organisation. There was an initial plan to build capacity for costing in Tanzania, but this was not explicitly funded and consequently it did not happen. Kuki Tarimo, who worked on the study with Mariam Ally in the Ministry of Health and is currently the P4H Focal Point for Tanzania, reflects,

what we really needed was to involve national academic and specialist institutions to do this work in-country, and not to use consultants – we still need to establish a platform where these institutions can network and learn from each other, but of course this takes time and money.

What the costing studies achieved

While the findings from the three large national studies may not have been used in line with original expectations, these studies have provided an evidence base on which to build and the results continue to be updated and utilised, if on an ad-hoc basis. As Dr Ir Por, Associate Professor and Deputy Director of Cambodia’s NIPH explains,

even when outdated, these cost data can be adjusted and continue to be useful, particularly at sub-national level for primary health care where the service package does not change much over the years.

The World Bank Public Spending Review for Cambodia published in February 2024, drew heavily on the GIZ-supported costing data. And according to Professor Ali Ghufron, one of the originators of the costing study in Indonesia and currently head of the country’s Social Health Insurance Scheme, the tariff for one of the key public health services – the delivery of a baby by a midwife – was recently up-dated using analysis from the GIZ-supported study, more than 10 years after it was conducted.

Importantly, all four studies succeeded in bringing diverse stakeholders together to discuss health service costing and pricing. In Tanzania, ‘though it took some time’, says Mariam Ally, ‘the study succeeded in sparking discussions on health service pricing and these discussions were evidence-based because we had the costing data that was lacking before’.

Kuki Tarimo adds in relation to Tanzania, ‘the ministry would now like to conduct another comprehensive costing study – and this is because they understand what costing can contribute to policy development in the country’. Echoing this effect, Dr Imran Masood Durrani, says,

There is now – thanks to the study – a strong realisation within the government at federal and provincial levels that this type of costing should become a more routine exercise in the health sector.

Failing forwards – post-mortems make for great learning

The learning session on costing studies, held on the sidelines of the conference of the GIZ Sector Network for Health and Social Protection, proved to be an invaluable first step for GIZ colleagues to share lessons and to explore ‘how to do costing better’. 

This type of productive learning – including the value of ‘failing forwards’ or taking calculated risks and making mistakes in order to progress and innovate – informs improvements in development cooperation. ‘It proved to be unexpectedly cathartic’, says Franziska Fuerst, ‘allowing us not only to scrutinise the setbacks but also to recognise the unintended successes’. Kai Straehler-Pohl puts it plainly, ‘it’s good to look into the rearview mirror after a few years, because some benefits take time to materialise and can only be seen years later’.

This is what the GIZ sector network health and social protection is all about – opportunities for collaborative knowledge management, which go beyond costing studies. There are many other topics which would benefit from a similar approach.

Corinne Grainger, April 2024