Economic Consequences of Epidemiological Changes in Diabetes in Middle-Income Countries: The Mexican case

Direct costs of health care services were obtained from the management of standardized cases, adjusted by type of institution. The cost-evaluation method was designed according to an instrumentation technique that identified production and supply functions for each case management, a diagnostic-related group variant adjusted for Mexico (12). Eight tools were used to establish costs per production function, which were concentrated in cost-evaluation matrices, according to the institution.

Case management was defined for an average case with the corresponding adjustments for each institution. For each disease and event to be evaluated, management of the average case was defined on the basis of the disease’s natural history and the results of shadow study reviews. The point of view of a group of expert clinicians and administrators was considered in order to obtain a homogeneous opinion of how to manage each case. These definitions refer to the demand of hospital or ambulatory services, according to each disease.

As explained above, standard management was defined per disease/event to be cost evaluated, considering the natural history of the disease, results found in the clinical files review and the shadow study, and the opinions of a group of expert clinicians and administrators. Results obtained depended on hospital or ambulatory management, the case, and the institution’s regulations.

Direct costs refer to all monetary resources required to satisfy health care services demanded by users; these included consultation/diagnosis, drugs, hospitalization, and treatment of complications. Following the human capital approach to calculate indirect costs in Latin America (3,13), the model included three categories of monetary costs attributable to type 2 diabetes in three public institutions: mortality costs, cost of permanently disabled patients, and cost of temporarily disabled patients. The annual demand for health care services for diabetes was calculated from the number of cases being treated, adjusted by type of institution. This information was obtained from the statistics bulletin on health impairment of the National Health System, 1989–2002 (14,15).

To determine the financial requirements, a time series was performed from 1989 to 2002, with a probabilistic design according to the Box-Jenkins technique (16), using a 95% CI with P < 0.05. Analysis variables included changes in morbidity for the study period and changes in health care policies and programs for diabetes and chronic diseases. The following methodological phases were considered in the development of the probabilistic models: identification, estimation, diagnostic check, and prediction.

The resulting models for estimating the expected demand of diabetes have an average movement operator of order 1. As in other studies with similar methods (17), the basis of each model was to include people who received a diagnosis of type 2 diabetes as well as costs of annual hospital or ambulatory management at one of the three institutions under study. We should emphasize that the projection period was limited to 3 years because previous studies have advised against using periods >4 years because they may cause budgeting uncertainty. To calculate costs and financial consequences caused by changes in both the epidemiological profile and the demand by type of institution, an inflationary index projected to 2005 was developed and applied, based on the Banco de Mexico’s price index for consumers. Results are in Mexican pesos and U.S. dollars, at the exchange rate of 1 U.S. dollar = 9.35 Mexican pesos, corresponding to January 2003 (18).

Figure 1 shows the estimated direct, indirect, and total costs of type 2 diabetes. This included all people and was only for health care costs that will be demanded of the three main public health care providers in Mexico, up to 2005; it does not include costs of health care demand for diabetes in the private health care sector. The overall contribution of direct costs was 45% of total costs, and indirect costs were 55% of total costs. Regarding the four categories of estimated direct costs, results were: consultations/diagnosis 11.64%, drugs 38.77%, hospitalization 11.64%, and complications 32.18%. With respect to the three estimated categories of indirect costs, the results were: mortality costs 4.52%, costs of permanently disabled patients 94.06%, and costs of temporarily disabled patients 1.42%.

With respect to the costs of complications, the five estimated categories were: costs of retinopathy 11%, costs of cardiovascular disease 10%, costs of nephropathy 73%, costs of neuropathy 3.6%, and costs of peripheral vascular disease 2.4% (Fig. 1). Going into detail about costs, from these results it was possible to establish minimum and maximum ranges of financing requirements at the three most important health care institutions. The cost of a hospital case ranged from $613.71 to $887.14, with the lowest at the SSA and the highest at the IMSS. The same trends were observed for costs of ambulatory cases.

Regarding the effect of the expected epidemiological changes in hospital and ambulatory service demand for 2003–2005, the results are shown in Fig. 2. An increase is expected, although the results show a greater increase for hospital cases. As shown in Fig. 2, the increase tends to be higher among the insured when compared with the uninsured population. Figure 3 shows the trends in the economic resources needed to finance the minimum service demanded by the population. To cover service demand by diabetes patients, IMSS requirements will increase from $78,813,110 to $83,689,444; for the ISSSTE they will increase from $17,670,392 to $19,582,796; and for the SSA they will increase from $33,038,568 to $37,079,578. Increases in the resources required for hospital or ambulatory case management followed along similar lines.

When comparing direct costs for different health care models, depending on the disease and the institution, results were as follows: for hospital cases of diabetes, the IMSS will pay the highest cost, followed by the SSA and ISSTE, with a lower direct cost (Fig. 4). The same trend is observed in relation to direct costs for ambulatory care. The highest average direct costs per managed case, as well as the economic impact of the global management of diabetes for 2003–2005, were found at the IMSS. Indirect costs showed a similar trend for the three institutions, representing 30% above the total direct costs. Financial requirements for ambulatory and hospital care represent ∼4.7% of the total budget assigned to the uninsured population and 6.5% of that allocated to the insured population.

The economic consequences for insured and uninsured populations also have a different impact with respect to the total budget for both population groups. For the low-income uninsured population (40% of all Mexicans), in recent years health care expenditure was on average 15% of the total, whereas for the insured population (50% of all Mexicans), health care expenditure was 43% of total expenditures (19). According to the results of our study, including resources required for hypertension, of the total financial burden for annual management of cases of diabetes and hypertension, 25% will be for the uninsured and 75% for the insured population.

In this study we estimated the costs of type 2 diabetes for the three main public institutions that provide health care in Mexico. We also estimated the financial requirements for the next 3 years, taking into account the expected changes in the epidemiological profile of diabetes in Mexico.

The estimated costs of our study reported a total of $317 million, $140 million in direct costs, and $177 million in indirect costs. With respect to the cost results, several studies in Mexico on the cost of diabetes reported $100 million in direct costs and $330 million in indirect costs (20). Another recently published study reported $1.9 billion in direct costs and $13.1 billion in indirect costs. There are differences between our results and those of previous studies because of differences in the estimates of the prevalence of diabetes used in the developed models and because our study only reported costs of type 2 diabetes for the three main public institutions.

With regard to the out-of-pocket expenses for diabetes, we have to highlight this situation in Mexico. In terms of other costs for the control and follow-up of ambulatory management of diabetes, the use of financial resources attributable to direct user expenditures is a considerable contribution to the annual cost of case management. In Mexico, users must make direct out-of-pocket payments, which are relatively considerable, for control studies, monitoring, group therapy, diabetes education, etc. Although this affects all users of the health care system, this cost has a greater impact on uninsured users. Because the infrastructure needed to carry out studies is not always available, these users have to take on the cost of these studies, without being reimbursed by the health care system.

This has important repercussions in terms of quality of care. The health care system proposes integral management of the diabetic patient, including monitoring, control and follow-up with laboratory studies, health education, group therapy, etc. However, in view of the lack of infrastructure in the system, the patient is the one who finally determines his/her ambulatory care, making integral management more difficult. For this reason, the quality of care with which diabetic patients are managed requires adjustments and corrections to improve this quality and decrease the financial impact on the user’s pocket.

With respect to the epidemiological changes expected for diabetes and their effects on financial requirements, we should point out that financial pressure will not only depend on the changing disease epidemiology but also on the type of population to be served. For example, in the case of insured patients at the IMSS, the demand for services for diabetes is much lower than the demand for other chronic diseases such as hypertension. Health care needs for patients with diabetes only at the IMSS are almost the same as those at the SSA; however, they are much greater than the corresponding service needs for insured patients with diabetes at the ISSSTE. On the other hand, if risk factors and the different ambulatory and hospital care models remain more or less as they are in the three studied institutions, the financial burden will be higher for the IMSS, followed by the SSA and then the ISSSTE.

The repercussions of all of the above will have a greater relevance for health care services for the uninsured because the pressure of inequality in resource allocation will be added to that caused by the relatively large increase in service demand for chronic or degenerative diseases. These services will be part of the popular health care insurance that the present national health care program (2001–2006) intends to implement for the low-income population with no social security coverage (21,22). In terms of the financial pressure generated from both, the epidemiological changes and changes required in the new health care programs will present different challenges for health care planning at all three institutions. For example, expected financial consequences suggest the need to redesign and invest more in programs for health promotion and prevention in order to reduce demand for hospital services and promote treatment of diabetes in an ambulatory setting. The greatest challenge will be for the social security services because, since their creation, these services have received the greatest amount of funding for treatment and rehabilitation programs.

Estimations of financial requirements are basic information for strategic planning. The financial consequences of the expected epidemiological changes are a basis and justification not only for more investment in diabetes or management of chronic diseases but also for the allocation of more resources for their prevention, thus minimizing and controlling the disease and lessening the economic burden on the health care systems of middle-income countries. This means that despite our evidence for the need to increase health care expenditures, the present economic situation does not seem to favor this and instead points to a decrease in health care expenditures, not only for 2003 but at least until 2005.

In terms of the validity of our results, it is important to highlight that, in relation to epidemiology, the expected results were validated by reference to trends established during the previous 10 years. The models used in our projections were similar to those in the statistics bulletin of the National Health System for 1990–2002. The number of expected cases follows trends that are similar to those projected by the International Diabetes Federation for Latin American countries (23). Case management costs are within the management range of diabetes found in previous studies at the national and international level.

We show that the models are internally valid and predict direct and indirect costs and financial requirements for the next 3 years. These results and the relative weight of each category of costs (complications, hospitalization, mortality, disability, etc.) are consistent with those observed in the reference population and institutions under study. We extrapolate the results only to the costs of diabetes in public institutions because we recognize that it is much more difficult to prove external validity without data on the entire health care sector, particularly when we do not have access to data in the private sector. The models developed in other studies (24–25) are intentionally conservative and are unlikely to overestimate the economic impact of diabetes.

On the other hand, it is important to emphasize that studies in the U.S. (26) and Canada (27) reported that direct costs are higher than indirect costs. In our results, as we showed before, they were similar: direct costs accounted for 45% and indirect costs for 55% of the total costs. As we have seen in other studies on the economic impact of diabetes, our study has limitations basically because of the methodological difficulties found in developing probabilistic models in countries where there is no access to (or an absence of) data, particularly from private institutions.

Despite the above-mentioned limitation, results reported several relevant aspects of the diabetes problem: in the first place, if risk factors and health care programs for the main public institutions continue as they are in 2003, the economic burden of direct costs to satisfy future demands of health care services will go beyond the resources these institutions have allocated for diabetes. In the second place, the economic burden represented by indirect costs is a challenge that should be faced by individuals, society, and the health care system as a whole. In other words, the indirect cost of the diabetes problem establishes the need to involve different sectors of society with a greater commitment to action.

Finally, total reported direct and indirect costs set forth, as challenges to be solved, the prevention of diabetes and its complications within the health care sector as well as in other sectors, particularly that of education. As actions to be taken to face these challenges, we propose the following strategies: a greater investment of financial resources allocated to prevention in order to decrease the burden attributable to complications and new cases of diabetes; a greater rationality by type of institution in the allocation and use of financial resources for diabetes; and, just as importantly, a greater promotion of educational programs directed toward the general population, and particularly for the population with diabetes. These programs should be promoted and coordinated by the health care sector, but with strong and ample participation of various social actors (nongovernmental organizations; researchers; other sectors from federal, state, and municipal government levels; community leaders; and users).

The authors express special thanks to the National Council for Science and Technology of Mexico (CONACYT), and the International Development Research Center from Canada (IDRC) for providing funding for the first phase of this study.