Economic Issues
An economic discussion of health policies designed to prevent, treat, and manage musculoskeletal conditions in developing countries is inherently difficult for a variety of reasons, but primarily because of the lack of both epidemiological and cost-effectiveness data for most developing countries. Some progress has been made by Symmons, Mathers, and Pfleger (2004a, 2004b),-who-provide incidence estimates for OA and RA from epidemiological data on prevalence and relative mortality risks, although data from many areas are scant.
Perhaps a more important constraint on economic evaluations in this field is the surprising number of interventions for which trial data on efficacy are inadequate. Another issue, currently the target of a concerted effort to improve practice in the field, is the lack of cross-study comparability of the results of economic evaluations of interventions for OA, RA, and OP. One of the most important variables is the choice of comparator used to assess the cost-effectiveness of interventions.
The Outcome Measures in Rheumatology Clinical Trials Economics Working Group, which was established in 1996, has made some progress toward redressing this problem. In principle, the relevant comparator is generally the next-best alternative or alternatives to the intervention of interest. The choice of comparator is especially important for cost-effectiveness analysis, because cost-effectiveness is a relative, not an absolute, concept; whether a particular intervention is considered efficient depends on the efficiency of other interventions and on budget constraints. This issue is a fundamental one, because a great many health sector innovations involve new ways of producing desirable effects with existing technology. The relevant consideration in such cases is the additional benefits that the innovation is expected to confer and the relative cost of achieving those benefits. In such circumstances, the computation of incremental cost-effectiveness ratios (ICERs) on the basis of a no-treatment alternative is of limited use, unless that scenario is genuinely under consideration. Unfortunately, the no-treatment (or, more accurately, the placebo treatment) option is precisely the comparator that much of the literature has used.
Another characteristic of economic evaluations in this field is that they have been performed almost entirely for developed countries. In the sections that follow, we discuss the steps we have taken in an attempt to minimize the adverse consequences of reliance on the literature for developed countries. Nevertheless, the pragmatic approach that we have adopted is subject to some important limitations and caveats.
Cost-effectiveness is a relative concept, in the sense that cost-effectiveness ratios (CERs) are useful only for comparing alternative ways of achieving a desired outcome—for instance, improving the quality and length of life. Assertions that an intervention is, in its own right, cost-effective are usually based on the notion that a particular CER represents a cutoff between those interventions that are efficient and those that are not. Thresholds of this kind involve an assumption about the value of life—for example, that a quality-adjusted life year (QALY) is worth US$30,000. Nevertheless, the literature routinely uses cost-effectiveness rules that are based on thresholds without the theoretically necessary explicit consideration of implicit budget constraints.
We have tried to avoid using a threshold type of approach in relation to the discussion of cost-effectiveness. Instead, we critically reviewed the cost-effectiveness literature in rheumatology to provide an indication of the relative costs and consequences of available interventions. In some cases, an intervention appears to be inefficient because it costs more and produces fewer benefits than a competing alternative or because two interventions produce identical effects but one costs less than the other. Nevertheless, we have provided a summary of our views—for ease of reference—as table 51.5. This table summarizes our thoughts on the weight of the current effectiveness and cost-effectiveness evidence and the likelihood that developing countries might realistically consider each intervention. For the reasons given above, though, we have articulated the evidence in more detail in the text.
[Table .]
Cross-country differences in the epidemiology of conditions of interest, the age structure of populations, and the access to health care, along with differences in relative prices, are liable to affect the cost-effectiveness of any given intervention. Some of the substantive gaps between the developed and developing worlds may compound the problem. For example, if the price of labor relative to that of capital is consistently lower in the developing countries, capital-intensive interventions may be relatively less attractive than they are in the developed countries, especially if labor-intensive alternatives exist.
To improve comparability across the literature, we adjusted reported CERs by converting them to 2001 U.S. dollar prices (therefore, those we report generally differ from those the original authors cite). Generally, we adjusted outcomes to U.S. dollars for studies outside the United States that reported CERs in local currencies using the official exchange rate in effect at that time, but wherever such studies reported only U.S. dollar equivalents, we took these as given. Finally, we used the U.S. Bureau of Labor Statistics consumer price index data for 2004 to inflate (deflate) the U.S. dollar CERs to 2001 prices. Thus, unless otherwise stated, all price data are expressed in 2001 U.S. dollars.
For clarity, we have classified cost-effectiveness results by condition and also according to whether the intervention constitutes a primary, secondary, or tertiary intervention. The exception is RA, for which the management protocols are less amenable to this type of abstraction. For RA, we found that categorizing the evidence according to a taxonomy that is problem or intervention based was more useful.
Primary Interventions
This section reviews the evidence on the cost-effectiveness of interventions designed to prevent the onset of OP. The works surveyed analyzed interventions in healthy people, primarily perimenopausal and postmenopausal women, with no established history of OP.
Physical Activity
The prophylactic effects of physical activity are generally well appreciated, and a large proportion of preventable disease is sometimes attributed to sedentary lifestyles. Katzmarzyk, Gledhill, and Shephard (2000) estimate the relative risks for those who are inactive compared with those who are physically active for a range of conditions, including OP. Their results for Canada suggest mean OP relative risk factors of 1.56 to 1.90 for sedentary versus active women, depending on race, and indicate that the population-attributable fraction of OP caused by inactivity was approximately 27 percent and accounted for more than 16 percent of the direct economic costs of physical inactivity.
The effectiveness and cost-effectiveness of programs intended to encourage lifestyle changes are generally not well established. Geelhoed, Harris, and Prince (1994) consider the effect of an intervention in Australia involving exercise and calcium supplements for healthy postmenopausal women to prevent osteoporotic fractures. They find that the cost of the intervention was US$96,119 per QALY; however, note that the authors assumed no toxic effect of the lifestyle regimen on diseases other than OP.
Calcium Plus Vitamin D
Willis (2002) analyzes the cost-effectiveness of administering calcium plus vitamin D3 to healthy postmenopausal women in Sweden and demonstrates that this intervention is a cost-saving one for 50-, 60-, and 70-year-old women with a maternal family history of hip fracture and for 60-and 70-year-old women with either a history of fragility fractures or a smoking habit. In developing regions, calcium plus vitamin D therapy may be a cost-effective or cost-saving intervention if targeted at older, asymptomatic women with maternal histories of hip and other fragility fractures—especially those who smoke. A targeted strategy of this kind is likely to be the most cost-effective in regions where environmental uptake of these elements is limited for dietary or other reasons.
Hormone Replacement Therapy
Geelhoed, Harris, and Prince's (1994) cost-effectiveness analysis of interventions in a hypothetical cohort of 100,000 healthy postmenopausal women includes several HRT strategies: (a) estrogen from age 50 for life, (b) estrogen from age 50 for 15 years, and (c) estrogen from age 65 for life. Compared with a no-therapy alternative, the cost per QALY was US$8,609 for strategy c, US$13,268 for strategy a, and US$30,183 for strategy b.
Armstrong and others (2001) compare HRT with a no-therapy scenario in healthy postmenopausal women and examine how the risks of breast cancer and coronary heart disease (CHD) might influence the cost-effectiveness of the interventions over 5- and 10-year periods, as well as a lifetime intervention of approximately 31 years. They report a relatively low cost per QALY of US$2,238 to US$2,850 for women at a 10 to 15 percent risk of breast cancer. The cost-effectiveness of HRT fell as the risk of breast cancer increased.
Both the base cases in these studies assume that HRT reduces hip fracture rates, and Armstrong and others (2001) also assume reductions in CHD. These constitute important assumptions because, as Kanis and others (2002) point out, data from randomized clinical trials (RCTs) support the hypothesis of no effect of HRT on either appendicular fractures or CHD. Thus, the ICERs reported by both the studies may either understate or overstate the true cost per QALY produced by using HRT as a primary prevention.
Raloxifene
Armstrong and others' (2001) study also includes a cost-effectiveness analysis of raloxifene use (compared with HRT and no intervention) in healthy postmenopausal women. Their results indicate that, by comparison with raloxifene, HRT is a dominant long-term therapy for U.S. women at average risk (in this case, a 10 percent lifetime risk) of breast cancer: the 5-and 10-year period ICERs were US$37,620 and US$33, 472 per QALY, respectively. For women at a 30 percent or higher risk of breast cancer, the ICER for raloxifene versus HRT was less than US$4,160, and decreased with risk.
Kanis and others (2002) argue that existing evidence on raloxifene suggests that it has no significant effect on either appendicular fractures or CHD. On the basis of the existing cost-effectiveness evidence, the use of raloxifene as a prophylactic intervention for OP in the developing regions has little to recommend it.
Secondary Interventions
The following studies were concerned with interventions in people with some indication of OP, either from a bone mineral density assessment or a fracture. Some of the general studies include Jonsson and others' (1999) study based on Swedish epidemiological data. The authors consider two different levels of intervention costs, those associated with HRT and those associated with HRT plus bisphosphonates, and find that the higher-cost intervention (HRT plus bisphosphonates therapy) was dominant for the 80-year-old group modeled. In the context of developing countries, note both the relatively higher incidence of osteoporotic fractures among 80-year-olds and the relatively larger size of this demographic group in Sweden.
Screening
Norlund (1996) conducted a cost-benefit analysis of fracture prevention in osteoporotic women age 50 to 54 in Sweden, assuming 70 percent participation in the screening program and an offer of HRT with 30 percent acceptance. The study provides evidence of a negative net benefit, indicating that the costs of a population screening program of this kind exceed its benefits. Thus, population-based bone mineral density screening programs aimed at perimenopausal or post-menopausal women are likely to be a poor use of health resources in the developing world.
Calcium and Calcium Plus Vitamin D
Citing trial evidence, Kanis and others (2002) assume that calcium supplements alone reduce only vertebral fracture risks in women with established OP. Assuming a compliance rate of 70 percent, the authors find that the intervention's cost per QALY for 50-, 60-, 70-, and 80-year-old cohorts were approximately US$64,995, US$31,548, US$10,271, and US$10,527, respectively. They also examine the cost-effectiveness of calcium plus vitamin D on the basis of trial evidence that this combination also reduces appendicular fractures. Assuming a 70 percent compliance rate, they find that calcium plus vitamin D was cost saving in 80-year-olds and either cost saving or a low-cost intervention (mean cost per QALY of US$584) in 70-year-olds. For 50- and 60-year-olds, the mean costs per QALY were US$29,357 and US$13,730, respectively. Thus, in developing regions, calcium plus vitamin D therapy may be an attractive investment for elderly women with established OP.
Hormone Replacement Therapy
Fleurence, Torgerson, and Reid (2002) demonstrate an ICER of US$12,800 to US$19,700 for HRT for their Scottish sample. Kanis and others (2002) show that while HRT was generally a dominant therapy for 80-year-olds, with a cost per QALY of US$4,527, it was an expensive therapy for 50-year-olds at a cost per QALY of US$42,940. These results suggest that HRT is likely to be an attractive intervention for established OP for some age groups in the developing regions. Differences in life expectancy and the underlying incidence of OP will, however, have a considerable bearing on the age at which HRT interventions may be considered desirable in each region.
Raloxifene
Kanis and others (2002) find that the cost per QALY associated with raloxifene was approximately US$835,622 in 50-year-olds, and although this cost generally fell with age, it remains an expensive intervention. Raloxifene therapy is not an attractive investment for the developing regions.
Calcitonin and Bisphosphonates
The cost-effectiveness evidence on nasal calcitonin is unambiguous. It is a particularly expensive intervention and represents an unattractive investment of health care resources even in wealthy developed countries. The most favorable cost-effectiveness results for nasal calcitonin come from a study by Coyle and others (2001), who find that both calcitonin and alendronate reduced wrist, hip, and vertebral fractures in postmenopausal women but that etidronate had no such effect on hip and wrist fractures. The ICERs for nasal calcitonin for 65-year-old women for five years of therapy were US$34,166 per QALY compared with no therapy and US$23,952 per QALY compared with etidronate. The results of this study were sensitive to the underlying fracture rate.
Kanis and others (2002, iv) also conclude that calcitonin is "not cost-effective at any age largely because of its costs." Indeed, their estimates of costs per QALY for 70- and 80-year-old women, the groups for which the intervention is most cost-effective, equate to approximately US$245,373 and US$181,109, respectively. By contrast, both alendronate and etidronate were dominating interventions for 80-year-olds. At current prices, calcitonin therapy is not an attractive investment for the developing regions.
Fluoride
Kanis and others (2002) find that fluoride was generally a dominant intervention in women with established OP, because it appears to decrease the risk of vertebral fracture but to increase the risk of hip fracture, although the latter result is statistically insignificant. When they assume that fluoride has a neutral effect on hip fractures, the authors find that the cost per QALY was in the acceptable range for interventions in the United Kingdom—that is, less than US$46,684. Fluoride is unlikely to be a desirable intervention for preventing OP in developing countries.
Alfacalcidol
Kanis and others (2002) report wide confidence intervals on the cost per QALY of an alfacalcidol intervention. This result is largely due to substantial variation in the apparent vertebral, hip, and humeral fracture risk available from RCTs. Thus, alfacalcidol does not appear to be a good investment for developing economies; however, additional RCTs are required to reduce the uncertainty regarding the cost-effectiveness of this intervention.
