AIM: To design a medical cost calculator and show that diabetes care is beyond reach of the majority particularly patients with complications.

METHODS: Out-of-pocket expenditures of patients for medical treatment of type-2 diabetes were estimated based on price data collected in Benin, Burkina Faso, Guinea and Mali. A detailed protocol for realistic medical care of diabetes and its complications in the African context was defined. Care components were based on existing guidelines, published data and clinical experience. Prices were obtained in public and private health facilities. The cost calculator used Excel. The cost for basic management of uncomplicated diabetes was calculated per person and per year. Incremental costs were also computed per annum for chronic complications and per episode for acute complications.

RESULTS: Wide variations of estimated care costs were observed among countries and between the public and private healthcare system. The minimum estimated cost for the treatment of uncomplicated diabetes (in the public sector) would amount to 21%-34% of the country’s gross national income per capita, 26%-47% in the presence of retinopathy, and above 70% for nephropathy, the most expensive complication.

CONCLUSION: The study provided objective evidence for the exorbitant medical cost of diabetes considering that no medical insurance is available in the study countries. Although the calculator only estimates the cost of inaction, it is innovative and of interest for several stakeholders.

Non-communicable diseases (NCDs) including diabetes pose a serious challenge to health systems already overstretched with acute and infectious diseases in Africa[1]. Following the High Level Meeting of the United Nations on NCDs in New-York in 2011[2], heads of governments made commitments towards prevention and control. However, the misconception that NCDs are diseases of the affluent is still widespread in low-income countries in spite of accumulating evidence against this myth. Strong advocacy is therefore required and the International Diabetes Federation (IDF), for instance, has developed a series of tools and guidelines to influence decision makers[3]. The following definition of advocacy is relevant for all health professionals[4]: “Blending science, ethics and politics, advocacy is self-initiated, evidence-based, strategic action that health professionals can take to help transform systems and improve the environments and policies which shape their patients’ behaviours and choices, and ultimately their health”. Simple and culture-sensitive advocacy tools are direly needed to foster the prevention and control of chronic diseases such as diabetes.

Several methods for identifying the economic consequences of diseases have been defined. Standard costs of diabetes in Africa, including the direct and indirect costs of the illness, have been estimated based on secondary data[5]. Such cost-of-illness (COI) studies, in addition to direct and indirect costs, may also include the intangible costs although these are seldom measured owing to the complexity of such measurements[6]. COI methods are simpler than, and very different from, the macro- and micro-economic models that have been developed to estimate the economic burden of the disease (cost of inaction) and the cost-effectiveness of action, such as CORE[7] and other Markov-based models, and WHO CHOICE[8]. Markov models are computer simulations of probabilistic progression of a disease in a hypothetical cohort which have been adapted to forecast mortality and complications of diabetes, as well as medical costs[9]. WHO CHOICE has been used to estimate the cost-effectiveness of interventions to combat chronic diseases, including diabetes, at the regional level, thereby assisting decision makers[10]. Comprehensive computer models of diabetes economic burden are many and they are in constant evolution[11,12]. However, such models are highly complex and ill-suited to the field.

As part of a university partnership project on the double burden of malnutrition in French-speaking West-Africa[1], diabetes advocacy instruments that are more specific to sub-Saharan African countries were developed. The principal tool was a simple medical cost calculator which is described and the results discussed in the present paper.

Individual medical costs for uncomplicated diabetes and additional costs associated with complications are shown in Table 2 for each study country (in United States dollars). Uncomplicated diabetes costs and chronic complication costs are given on an annual basis. Costs for acute complications include uncomplicated diabetes annual costs plus additional costs per episode of a given complication. Medical costs of uncomplicated diabetes ranged from 126 United States dollars to 1093 United States dollars in Guinea, 137 United States dollars to 869 United States dollars in Mali, 212 United States dollars to 828 United States dollars in Benin, and 224 United States dollars to 859 United States dollars in Burkina-Faso. Wide ranges were also observed within countries, with at least a twofold increase from the minimum to the maximum cost in the public healthcare system as well as in the private sector. The minimum cost in the public sector was lower by a factor of 4 to 8 than the maximum cost in the private sector in all countries. In the treatment of uncomplicated diabetes, drugs and medical supplies represented the highest cost share, ranging from 52% to 75%. Figure 1 illustrates the range of medical costs per person per year for uncomplicated diabetes, in the public and private healthcare systems and for each study country. The medical costs increased steadily from the public to the private sector, except in Benin and Burkina-Faso where the maximum medical cost in the public sector was higher than the minimum cost in the private sector. Figure 2 provides illustrated examples of the software outputs for incremental medical costs per person per year in the presence of one chronic complication, in this instance for Mali. It shows that except for retinopathy, the additional cost for treating complications represents at least twice the medical cost of basic diabetes care. Of all chronic complications considered, the most costly was nephropathy in all countries.

Open in New Tab   Full Size Figure   Download Figure

Figure 1 Computed medical costs per individual per year, basic care, no complications.

Open in New Tab   Full Size Figure   Download Figure

Figure 2 Computed annual medical costs per individual, basic care plus complications (Mali). A: Annual medical costs per individual, basic diabetes care plus hypertension; B: Annual medical costs per individual, basic diabetes care plus retinopathy; C: Annual medical costs per individual, basic diabetes care plus nephropathy; D: Annual medical costs per individual, basic diabetes care plus cardiopathy.

Table 3 shows for each study country the estimated cost range in current United States dollars and in percentage of gross national income (GNI) per capita (http://wdi.worldbank.org/table/1.1) and of an economic poverty threshold (2$ per day or 730$ a year according to the World Bank[23]) for uncomplicated diabetes, for diabetes with retinopathy, the complication with the lowest incremental cost, and for diabetes with nephropathy which entails the highest additional cost. GNI per capita in current United States dollars 2013 ranged from 460$ in Guinea to 790$ in Benin. Uncomplicated diabetes cost amounted to a minimum of 21%-34% of the GNI, with a maximum reaching 238% in Guinea. With chronic complications, the share of the GNI soared. The minimum cost of diabetes with nephropathy represented 73%-176% of the GNI. The medical costs of diabetes without complications ranged from 17.2% to 150% of the annual income corresponding to the poverty line. With retinopathy, the least costly complication, medical costs varied from 24.2% in Mali to 170% in Guinea. For nephropathy, the most costly complication, medical costs amounted to at least 67% of the annual poverty line.

The aim of this study was to develop and test a simple medical cost calculator for diabetes to be used primarily for advocacy purposes. Research evidence is not consistently used for policy formulation[24], even by international organizations[25]. Demonstration of unbearable economic and human costs of a chronic disease such as diabetes may be regarded as concrete and therefore more promising as a means to influencing decision makers. While considerations other than financial, whether political or other, are at play, economic arguments are important even if not sufficient by themselves to induce changes in public policy and programmes[26]. Furthermore, figures are compelling for decision-makers who have to address competing priorities. Cost indeed plays a critical role in decision making regarding health, in any country.

What the study showed is that the medical cost calculator was quite user-friendly and that collecting country-specific price data was highly relevant, considering the wide differences observed across countries in estimated medical costs of uncomplicated and complicated diabetes. The country-specific cost data collected using this instrument could actually be used to update regional estimates that are needed for macroeconomic models. The diabetes cost calculator is certainly not intended as a substitute for highly technical and validated models such as CORE[7] and WHO-CHOICE[8]. However, their use is constrained by their complexity, the global scale of the figures, and the numerous assumptions underlying the calculations. Expertise is required to manipulate and adjust the models. Furthermore, since data on local effectiveness of interventions are scanty, particularly in Africa, assumptions have to be made. The costing tool was designed so as to remain simple in order to encourage its use at country level. The participants of the advocacy training workshop during which the calculator was presented indeed felt that the instrument was transparent and easy to use, but at the present time, we have no feedback on the effective use of the calculator for advocacy purposes.

The results confirmed that medical costs vary considerably between the private and public healthcare systems and even within each sector. Similarly, in India, direct costs were nearly four times higher in Chennai than in Delhi, illustrating wide variations within a country[6]. Providing ranges of medical costs was considered more representative than means. Medical costs in other hospitals, clinics and pharmacies of the country would most likely fall within the computed range. Additionally, in order to compute means that would be representative of a whole country’s medical costs, the process would be much heavier and expensive, with the need for a large number of randomly selected health facilities to include in the study. The cost calculator also allows to better size up the medical cost discrepancies between the private and public sector. Public subsidies on drugs, hospitalization and other treatment components were not taken into account and likely contribute to these discrepancies among countries and between the private and public sector, for example in Benin, where the government gives subsidies for dialysis and for hospitalization. Additionally, costs for drugs will depend on the type of molecules and on whether or not generic formulations are available. The specific guidelines provided with the cost calculator regarding the type of drugs to consider (available on demand) contribute to standardize as much as possible cost data compilation.

The costing tool that we developed is the first of its kind and there is no equivalent in the published literature, where only out-of-pocket expenditures based on patients’ surveys are to be found. The cost calculator refers to a standardized medical care protocol so that costs can be compared across countries. The tool is flexible in that prices can be updated as required in the Excel file, with automatic adjustment of the outputs (tables and graphs). This allows for the tool to be updated whenever new data become available. However the treatment components cannot be modified in the software in order to allow comparisons.

The estimated total medical cost for the basic treatment of diabetes at the individual level in a given country using this standard procedure may be of value to health professionals, governments and other potential payers such as insurance companies in many ways. Firstly and of foremost importance, the burden of the expenses that households or individuals would have to incur for basic care of uncomplicated diabetes can be appraised based on income levels and income distribution in the country, where such data are available. In the absence of recent household budget surveys, a basis for comparison could be the country’s minimum wage, the GNI per capita, or else an economic poverty threshold, as was done in the present study. This may allow to clearly show that several patients are not minimally treated because they simply cannot afford the medical follow-ups and even the drugs, in the absence of government subsidies or insurance (or even when these are available). These cost data may therefore represent in themselves powerful advocacy arguments. In Cameroon, a study on out-of-pocket expenditures of more than 350 diabetes patients revealed that monthly medical costs reached 148$ United States in 2009-2011 (89.40$ for medicines, 10.40$ for consults, 35.0$ for tests, and 13.20$ for glycemia monitoring), which amounted to more than twice the minimum wage[27]. According to the present study, basic treatment cost range of uncomplicated diabetes would represent 27%-105% of GNI per capita in Benin, 21%-130% of GNI in Mali, 27%-238% in Guinea, and in Burkina, where the minimum estimated cost was the highest of the four countries, from 34% to 138%. Actual expenditures of patients probably lie somewhere in-between the minimum and the maximum. Analysis of data from a survey of several hundred individuals living with diabetes in Mali showed that annual medical expenditures of subjects free from complications were within the range of estimated costs using the calculator, at least in the public sector (unpublished data). A study on the economic burden of diabetes in Africa[5] showed that average medical costs (including the same components as the present study) represented 36% of GNI for countries with a GNI lower than 2000$ United States, which includes all four countries of the present study. Furthermore, this is the basic cost only. Treatment costs soared when diabetes complications are present, which is the case for a majority of persons living with diabetes -70% according to the survey in Mali[28]. In India, the cost-ratio for those having complications vs those without was around 2.0[6]. In our study, for instance, the treatment of nephropathy, considering only the minimum cost in the public sector, would more than treble the yearly basic cost, ranging from a factor of 3.5 in Burkina Faso to 6 in Guinea. A few studies have shown that the cost of diabetes is indeed beyond reach of a sizeable proportion of the population. In Côte d’Ivoire, for instance, 35% to 55% of the household income would have to be spent for diabetes care[29]. In Mali, it was estimated that for insulin only, households with a diabetic member spent 38% of their total income[30]. This confirms that acceptable diabetes care is likely unaffordable to most. In our study, we used the GNI per capita as a proxy of income, as well as the World Bank poverty threshold of 2$/d (or 730$/year). It is noteworthy that the GNI per capita itself is below this poverty threshold in three of the four study countries. Roughly 75% of the population of these countries lives with less than 2$/d, which betrays rampant poverty. It is therefore likely that a good majority of the people would not afford even the most basic medical treatment of type-2 diabetes. The diabetes cost calculator provides for an estimation of minimum incurred medical expenses as a percentage of income proxy without having to conduct lengthy and expensive surveys for that purpose. Furthermore, demonstrating the cost increment associated with complications related to a late diagnosis may help convince decision-makers to make the screening more efficient at primary health care level. Comparing the medical costs of diabetes in neighbouring countries may also be of interest for policy purposes, even if only the cost of the disease is computed at this stage.

There are obviously several limitations to this costing tool. Only direct medical costs are estimated, on the basis of diabetes treatment component price data collected locally in public and private care institutions. Other direct costs incurred by families, such as transportation, traditional medicine, time of the care-provider and extra-expenditures for the diet are not taken into account and besides, these can hardly be standardized. However, according to other studies on actual expenditures, direct costs tend to be higher than indirect costs[6].

Although the treatment protocol was considered realistic for Africa, all included elements of care may not be absolutely necessarily in spite of their relevance. For instance, some could argue that care may be acceptable even if some costly tests are not performed. Moreover, although consensus was required in the working group to define the treatment components, some arbitrariness was unavoidable, whether in the frequency of medical visits or in medicine posology.

The cost calculator does not include either the indirect costs to the health system (salaries and training of health personnel; health facilities; subsidies, etc.), the families and the society as a whole (loss of productivity, of income, of healthy life years...). The medical costs estimated with the calculator represent only a fraction of the total economic burden of the disease.

A useful addition to the costing tool would be to estimate cost-effectiveness of interventions for primary prevention among high-risk individuals, combined with secondary prevention among diagnosed individuals, in order to demonstrate the savings that may accrue from earlier detection and treatment of diabetes, using the costing tool. However, there are no cost-effectiveness data that are relevant for Africa and the only controlled interventions in low- and middle-income countries were conducted in India[31] and China[32]. Additionally, cost-effectiveness analyses do not take into account non-health benefits, such as income gains, which may be important[33].

In conclusion, the study confirms in an objective and standardized fashion that the basic medical cost of diabetes is likely beyond reach of a majority of people in West-African countries considering that no medical insurance is available in most of them. In spite of its limitations, the medical cost calculator, which can be used in different countries, is deemed important in Africa, considering the paucity of data on diabetes cost in the whole region. It is also flexible enough since cost data can easily be changed. No study had so far designed a simple costing tool which would take into account the various components of medical care of diabetes and its complications in sub-Saharan Africa.

The authors thank Mr Stéphane Proulx for software development, Drs. Adébayo Alassani (Benin), Joseph Samah (Guinea) and Alexandre Zoungrana (Burkina-Faso) for their support in price data collection, and Pr Amédégnato Degno and Dr. Denis Agbétiafa for pretesting the calculator in Togo.