Cost-Effectiveness of Interventions in Developing Countries
We determined incremental cost-effectiveness ratios (ICERs) for selected interventions for each condition by calculating total DALYs lost by a population because of the condition with and without treatment and then dividing the difference by the treatment cost. The disability weights used are presented in table 32.2. All analyses in this section followed the volume editors' standardized guidelines for economic analysis, region-specific age structures, and underlying mortality rates. We converted nontradable inputs into U.S. dollars at the market exchange rate. We assumed that the costs of tradable inputs were internationally consistent, as were costs associated with surgical treatments. Table 32.3 presents the costs of drugs and medical services. No fixed costs were assumed; therefore, our results are not linked with the extent of treatment coverage.
[Table .]
[Table .]
AD and Other Dementias
We analyzed the use of acetylcholinesterase inhibitors in the treatment of AD on the basis of the following assumptions: first, only patients who were older than 60 at the time of onset were considered; second, the treatment has no long-term benefits—that is, it does not reduce patient disability and has no effect on mortality.
We computed the benefits of reduced caregiver hours on the basis of reports that the improvement in cognitive function in AD patients associated with treatment using acetylcholinesterase inhibitors was a 1.2 point change in the global assessment scale for cognitive function, as measured by the Mini Mental State Examination. A 1 point improvement in the score was associated with a 0.56 hour per day reduction in caregiver hours, or roughly 205 hours per year (Marin and others 2003).
The cost of using acetylcholinesterase inhibitors per hour of caregiver time saved averaged US$13 across low- and middle-income countries (LMICs) and was at least US$11 in specific regions (the regions are the same as those in table 32.1). This amount is substantially more than the wage rate in these regions, which would generally not exceed US$1 to US$1.50 per hour, even for hired caregivers specifically trained to care for AD patients. We, therefore, conclude that the use of acetylcholinesterase inhibitors in developing countries is not efficient from an economic perspective. Calculating the cost per DALY averted for acetylcholinesterase inhibitors would not be meaningful, because we assume no benefit to the patient. Finally, the use of acetylcholinesterase inhibitors is uncommon in developing countries; therefore, reducing its use is not an important concern.
Epilepsy
We analyzed the cost-effectiveness of phenobarbital in the treatment of epilepsy, and the results are shown in table 32.4. We assumed that phenobarbital was provided to all patients. The cost of using phenobarbital per DALY gained in LMICs was US$89. Table 32.4 shows that the benefits of phenobarbital are large relative to its cost.
[Table .]
We did not look at other AEDs, such as phenytoin or carbamazepine, because the costs of those medications are much greater than that of phenobarbital, but their effectiveness is essentially the same (Aldenkamp, De Krom, and Reijs 2003). Although their use may be justified for specific medical reasons, phenobarbital is much more cost-effective.
We analyzed treatment options for patients who are refractory to treatment with phenobarbital. We assumed that such cases were treated either with a combination of phenobarbital and lamotrigine or with a combination of phenobarbital and surgery. We used the cost for epilepsy surgery of US$2,600, in accordance with a study from Colombia, and applied it to all regions (Malmgren and others 1996; Tureczek, Fandino-Franky, and Wieser 2000). We assumed that roughly half of surgery recipients experience no more seizures and that the remaining half continue to take phenobarbital despite undergoing surgery. Our evaluation of the surgical option included the costs of diagnostic services and the costs associated with screening patients who ultimately may not be eligible for surgery. For patients in LMICs who are refractory to phenobarbital, the ICER of the add-on drug lamotrigine was US$3,000, and the ICER of the surgical option plus phenobarbital was US$3,100. The difference between phenobarbital and the other two options was significant in all regions.
Among refractory epilepsy patients eligible for surgery and according to postoperative outcome studies conducted in developed countries, surgery may be of comparable cost-effectiveness to treatment with a combination of phenobarbital and lamotrigine. Because effectiveness data for developing countries are not available, this calculation is based on cost estimates from a study in Colombia and estimates of the effectiveness of surgery from developed countries. If the surgical outcome in developing countries were worse than in developed countries, the cost-effectiveness of surgery would be lower. Furthermore, we note a number of limitations to the use of surgery in refractory epilepsy, particularly in developing countries, along with the lack of long-term follow-up data on the outcome of surgery. We stress that the primary treatment of epilepsy is with phenobarbital, and effective treatment of epilepsy lies in more efficient use of this highly cost-effective medication to close the treatment gap.
Parkinson's Disease
We evaluated three interventions for PD: a combination of l-dopa and carbidopa, traditional medicines such as the ayurvedic treatment used in India, and deep brain stimulation. We assumed that treatment for all three modalities was effective for 10 years from the onset of treatment. The ICERs in LMICs for these three modalities were US$1500, US$750 and US$31,000, respectively (table 32.4). On the basis of the cost of medication and evidence from clinical trials of effectiveness (Parkinson's Disease Group 1995) and from animal studies (Hussain and Manyam 1997), we found that ayurvedic treatment was the most cost-effective option. The relatively favorable ICER for ayurvedic treatment is due to the extremely low medication cost of this intervention. The relatively high ICER for deep brain stimulation was largely attributable to the extremely high cost of surgery. Table 32.4 shows DALYs gained for US$1 million of health expenditure.
Stroke
We evaluated two sets of interventions for stroke: treatment of acute stroke and prevention of secondary stroke. We assumed that stroke sufferers have fully recovered 10 years after their last stroke.
We evaluated aspirin, heparin, and rt-PA for the treatment of acute stroke. The International Stroke Trial Collaborative Group (1997) reports that, within 14 days of the onset of stroke, mortality with heparin treatment is less than with a placebo; however, after six months, mortality is actually greater for patients treated with heparin than with a placebo—that is, there is a negative cost per DALY gained if this effect is incorporated. The estimates presented here are based on the change in the short-term mortality risk. For LMICs, the cost per DALY averted using aspirin was US$150 (table 32.4). The equivalent costs of interventions using rt-PA and heparin were US$1,300 and US$2,700, respectively. The costs of heparin are higher than the costs of rt-PA, despite the expensive equipment required for rt-PA, because of the lower effectiveness of heparin.
Table 32.4 presents DALYs averted for US$1 million of health expenditure for the three treatments. The cost per DALY gained using aspirin is a conservative estimate, because the use of aspirin has additional benefits in terms of preventing a recurrence of stroke.
Table 32.4 shows the costs of preventing a second stroke within two years of the first stroke. For LMICs, aspirin was the least expensive option at US$3.80 per single percentage point decrease in the risk of a second stroke within two years of the first. This rate translates to roughly US$70 per DALY gained (table 32.4). Combining dipyridamole with aspirin, because of higher cost, was slightly more expensive at roughly US$5.20 per single percentage point decrease in recurrent stroke risk for a single individual, or about US$93 per DALY. In contrast, carotid endarterectomy was US$87 for an equivalent decrease in individual recurrence risk or almost US$1,500 per DALY. The aspirin monotherapy option for preventing a recurrence of stroke was the most cost-effective approach only in South Asia and Sub-Saharan Africa, largely because of the relatively low costs of nontradable inputs, such as hospital and doctors' fees, in those regions. Low input costs of nontradables increase the relative importance of drug costs in determining the most cost-effective intervention; therefore, the cheaper drug, aspirin, was most cost-effective. Table 32.4 shows that, though US$1 million would be most effectively spent on aspirin alone in South Asia and Sub-Saharan Africa, investment in aspirin and dipyridamole treatment would result in a greater DALY gain in the other regions.
