Cost-Effectiveness of Interventions and Priorities
Most of the interventions to prevent and treat diabetes and its complications significantly affect the use of health services. The limitations of clinical trials include their failure in most cases to capture the entire intervention effect over a lifetime and to include all segments of a population to whom the intervention may apply. Evaluating the cost-effectiveness of interventions often requires the use of computer simulation models, but data availability, technical complexity, and resource needs present a significant barrier to constructing such models for developing countries. Furthermore, data on interventions are often available only from developed countries, and these data are often extrapolated to developing countries.
Estimating the Cost-Effectiveness of Interventions in Developing Countries
To assess the cost-effectiveness of interventions in developing countries, we updated the results from Klonoff and Schwartz's (2000) comprehensive review by including studies that were published up to 2003. Table 30.2 summarizes the cost-effectiveness of interventions for the developed countries, mainly in the United States. The results show that the cost-effectiveness of interventions varies greatly—from cost saving (an intervention is both more effective and less expensive than the comparator) to US$73,500 per quality-adjusted life year (QALY) gained.
We estimated the cost-effectiveness ratio of diabetes interventions for the six developing regions shown in table 30.3. We assumed that the effectiveness of these interventions, as measured in QALYs, was the same as in developed countries but that the cost of interventions and other diabetes care differed between developed and developing countries and also among the six developing regions. Using this assumption, we estimated the cost-effectiveness ratio for a developing region as the cost-effectiveness ratio in the developed country, mainly represented by the United States, multiplied by the ratio of costs in the developing region to the cost in the developed countries, which we calculated as follows. These cost-effectiveness ratios are based on costs and benefits over a lifetime, except for preconception care for women of reproductive age.
[Table .]
We estimated that the cost of intervention and other diabetes care in the United States was 8.6 times the cost in Latin America and the Caribbean. This cost ratio was an average of four cost ratios—each weighted by its share (Barcelo and others 2003)—for outpatient care, inpatient care, drugs and laboratory tests, and treatment for diabetic complications. The cost ratio for each cost component was calculated as the cost of medical services or drugs in the United States divided by the cost of the same services or drugs in Latin America and the Caribbean. U.S. data for medical services and drugs for routine diabetes care, plus treatment cost for diabetes complications, were obtained from a 1998 cost-effectiveness Markov model of the U.S. Centers for Disease Control and Prevention (CDC). Data for laboratory service were obtained from the 2001 Clinical Diagnostic Laboratory Fee Schedule from the U.S. Centers for Medicare Services (available from http://www.cms.gov). Data for Latin America and the Caribbean were obtained from three countries—Argentina (Gagliardino and others 1993), Brazil (Health Policy Division of the Brazilian Ministry of Health), and Mexico (Villarreal-Rios and others 2000).
We applied Mulligan and others' framework (2003) to estimate the costs of intervention and diabetes care in each developing region. Assuming that cost estimates are available for one of the regions, this framework allows the development of a relative cost index for health care services that can then be used to obtain cost estimates for the other five regions. Using costs estimated by Mulligan and others (2003), we first estimated three health service indexes, including hospital bed days, outpatient and inpatient services, and laboratory tests and procedures. We then combined the three indexes into one overall index for diabetes care in accordance with the share of each component in developing countries (Barcelo and others 2003). Finally, we estimated the costs of intervention and diabetes care in the other five developing regions by multiplying the cost of care in the Latin America region by the overall regional relative cost index.
Ranking Implementation Priorities
We assessed the implementation priority and feasibility of interventions, as explained in table 30.3.
Level 1 Interventions
All three interventions in this category are cost saving and are also feasible in terms of all four aspects considered. The barrier to implementing these interventions may be a short-term hike in intervention costs.
Glycemic control in a population with poor control (hemoglobin A1c greater than 9 percent or another measure of glucose control in situations where HbA1c tests may be unaffordable) is cost saving because the reduction in medical care costs associated with both short-term and long-term complications is greater than is the cost of intervention. Glycemic control for people with type 1 diabetes involves insulin use and, for people with type 2 diabetes, depending on the stage and severity of the disease, consists of diet and physical activity, oral glucose-lowering agents, and insulin. Patient education is an essential component of these interventions to encourage patients to comply with medication regimes and to change to and maintain healthy lifestyles.
Glucose is generally poorly controlled in people with both type 1 and type 2 diabetes, mostly because of lack of access to insulin and other diabetes supplies in developing countries. For example, the mean HbA1c level for people with diabetes in India was 8.9 percent in 1998 (Raheja and others 2001). A survey conducted by the International Diabetes Federation in 1997 (2003b) showed that no country in Africa had 100 percent accessibility to insulin. Ensuring adequate access to insulin should be an important priority for developing countries.
Blood pressure control for people with diabetes and hypertension reduces the incidence of both microvascular and macrovascular diseases. Major medication interventions include an ACE inhibitor, thiazide diuretics, or a beta blocker. Blood pressure control is cost saving mainly because of its large health benefits and relatively low intervention costs. Even in the United States, moderate blood pressure control costs less than US$250 per patient per year. Because many blood pressure medications are generic drugs, the costs are much lower in developing countries. In addition, the prevalence of people with poor control of blood pressure may be high in developing countries. For example, in Latin America and the Caribbean, 60 percent of people with type 2 diabetes in 2000 had blood pressure higher than 140/90 mmHg (Gagliardino, de la Hera and Siri 2001).
Complications related to foot problems are common among diabetics in developing countries. For example, in India, 43 percent of diabetes patients had foot-related complications (Raheja and others 2001). Interventions for foot care are low tech and require little capital. Interventions for foot care in developing countries should include educational programs for patients and professionals (for example, on foot hygiene, treatment of calluses, awareness of functional infections, and care for skin injuries); access to appropriate footwear; and multidisciplinary clinics. All three interventions could be cost saving, mainly because the cost of the interventions is low and the interventions can reduce the risk of foot ulceration and amputation, which are costly. Applying these interventions for high-risk patients, such as those with at least one previous foot ulcer or amputation, would yield even larger savings (Klonoff and Schwartz 2000).
Level 2 Interventions
The six interventions in this category are either cost saving and not feasible in one or more aspects or cost less than US$1,500 per QALY and are at least moderately feasible. Thus, interventions in this category represent good value for money but may present some difficulties in terms of feasibility.
Preconception care among women of reproductive age includes patient education and intensive glucose management. This intervention reduces short-term hospital costs for both mothers and infants and improves birth outcomes. However, the intervention may not be feasible in some developing countries because of the resources needed for the intervention and the difficulty of reaching the target population.
The lifestyle intervention for preventing type 2 diabetes costs US$60 to US$130 per QALY over a lifetime, depending on the region. The potential population eligible for a lifestyle intervention (those with impaired glucose tolerance or impaired fasting glucose) is large in developing countries. The International Diabetes Federation (2003b) estimates that the prevalence of impaired glucose tolerance was at least as high as the prevalence of diabetes in all regions. The expertise required for the intervention, such as dietitians and exercise physiologists, and the capacity of health care systems to handle the large populations eligible for the intervention may present a barrier to implementing the intervention in many developing countries.
People with diabetes are at higher risk of complications from influenza and pneumococcal infections than those without diabetes. Influenza vaccinations are a relatively cost-effective intervention, mainly because of the low intervention cost. However, the level of adoption for the intervention would depend on a country's ability to deliver the intervention to the targeted population.
The detection of proliferative diabetic retinopathy and macular edema by dilated eye examination followed by appropriate laser photocoagulation therapy prevents blindness. Annual screening and treatment programs for diabetic retinopathy cost US$700 or less per QALY gained in developing countries. The intervention is more cost-effective among older people, those who require insulin (Klonoff and Schwartz 2000), or those with poor glucose control (Vijan, Hofer, and Hayward 2000). In addition, screening less frequently, such as every two years, may be more cost-effective than screening every year (Vijan, Hofer, and Hayward 2000). Eye complications among people with diabetes are common in developing countries; for example, 39 percent of people with diabetes in India had eye-related complications (Rajala and others 1998). Although laser treatment is an effective intervention, such treatment may not be available in many developing countries or may be extremely costly.
ACE inhibitors can lower the blood pressure of those with hypertension and delay the onset or prevent further progression of renal disease for those with diabetes. Compared with screening for microalbuminuria and treating only those who have the condition, offering ACE inhibitors to all people with diabetes was more cost-effective at less than US$1,020 per QALY gained. This intervention was more cost-effective among younger people and was sensitive to the cost of drug. Thus, lowering the cost of the medication is a key factor for the success of this intervention in developing countries.
Smoking cessation includes both counseling and using medication such as a nicotine patch. Smoking cessation appears to be the least cost-effective among the level 2 interventions. However, the benefits of smoking cessation may be underestimated because our calculations only took the reduced risk of CVD into account (Earnshaw and others 2002). Adding the health benefits derived from preventing cancer and pulmonary diseases would improve the cost-effectiveness of smoking cessation. Considering the high prevalence of smoking in developing countries, smoking cessation should be a high-priority intervention, but the availability of the nicotine patch may be a barrier to implementing this intervention in developing countries.
Level 3 Interventions
The five interventions included in this category cost at least US$1,640 per QALY but could cost as much as US$8,550 per QALY. Compared with the level 1 and 2 interventions, those in this category are also less feasible. In general, depending on cost-effectiveness and feasibility, these interventions may not always be justifiable for all people in developing countries, given the limited health care resources. However, these interventions may be reasonable for selected subpopulation groups, such as those who can afford them.
Metformin therapy for preventing type 2 diabetes among people at high risk, such as those with prediabetes, is feasible because the drug is affordable in many developing countries; however, the intervention may not be good value for money. Cholesterol control intervention for people with diabetes falls into the same category. The cost-effectiveness of both these interventions would improve if the costs of the drug could be lowered.
The aim of intensive glucose control is to lower the glucose level of a person with diabetes to a level close to that of a person without diabetes. Implementing this intervention is a lower priority, mainly because of its relatively low cost-effectiveness in the context of the limited health care resources in developing countries. Although the U.K. Prospective Diabetes Study clearly demonstrates that lowering glucose levels can prevent or delay long-term diabetes complications (UKPDS Group 1998), the marginal return on very intensive glucose control in developing countries was relatively small.
Screening for undiagnosed diabetes is a low-priority intervention mainly because of its relatively high cost per QALY. However, screening for undiagnosed diabetes can be a worthwhile intervention for subpopulation groups, such as those that have a high prevalence of undiagnosed diabetes. In the United States, for example, screening for undiagnosed diabetes among African Americans was estimated to be 10 times more cost-effective than screening among other population groups (CDC Diabetes Cost-Effectiveness Study Group 1998). In addition, screening for undiagnosed diabetes may be a worthwhile intervention for patients with risk factors for other chronic diseases, such as hypertension, high lipid profiles, and prediabetes.
Annual screening for microalbuminuria was a low-priority intervention because screening added costs with no significant benefits. Treating all persons with diabetes with ACE inhibitors was a better treatment option than screening for microalbuminuria and treating only those who have the condition.
Cost-Effectiveness of a Polypill to Prevent CVD
A meta-analysis estimated that a hypothetical polypill could reduce the risk of CVD by 80 percent among all people over 55 or people with diabetes of any age (Wald and Law 2003). This hypothetical pill is a combination of three half-dose antihypertensive medications—aspirin, statin, and folic acid (see also chapter 33). Currently, neither is it available for use, nor have estimates of its benefits and adverse effects been confirmed in a formal, randomized, controlled trial. The idea is thus still theoretical. The cost-effectiveness of this hypothetical pill was, however, simulated using a computer model of people with newly diagnosed diabetes in the United States (Sorensen and others 2004), and the assessment found that a polypill intervention would cost US$11,000 per QALY gained. The intervention would be cost saving if such a pill cost US$1.28 or less per day. We estimated that the cost-effectiveness ratio of the polypill ranged from US$560 to US$1,280 per QALY gained for the six developing regions. This result was sensitive to changes in the cost of the intervention, but the intervention remained cost-effective within the most likely ranges of its cost (Sorensen and others 2004). A barrier to this intervention, in addition to the feasibility of producing such pill, is that its benefits and side effects would still have to be established in a randomized clinical trial.
Cost-Effectiveness of Diabetes Education
People with diabetes play a central role in managing their disease. Thus, diabetes education is an integral part of diabetes care. The goal of diabetes education is to support the efforts of people with diabetes to understand the nature of their illness and its treatment; to identify emergency health problems at early, reversible stages; to adhere to self-care practices; and to make necessary changes to their health habits (International Diabetes Federation 2003b). Health providers can deliver diabetes education programs in various settings. Evaluating the effectiveness of health education is challenging because of the difficulty of separating out its effect from that of other interventions. Nevertheless, a review of literature published in the United States suggests that self-management diabetes education may be cost-effective (Klonoff and Schwartz 2000).
Training in diabetes self-management reduces medical costs for diabetes care in developing countries in the short term. A multicenter intervention study in 10 Latin American countries demonstrated that an education program could reduce the cost of drugs by 62 percent (International Diabetes Federation 2003b), and another program in Argentina found a reduction in diabetes-related costs of 38 percent (Gagliardino and Etchegoyen 2001). Because the costs of education programs are generally low, the intervention may be cost-effective. Training patients to better manage their diabetes is also feasible because of its low technical complexity, low capital requirements, and cultural acceptability. Thus, diabetes education should be a high-priority intervention for all developing regions.
