Skip Navigation LinksHome > September 1, 1997 - Volume 16 - Issue 1 > Updates of Cost of Illness and Quality of Life Estimates for...
Journal of Acquired Immune Deficiency Syndromes & Human Retrovirology:
Epidemiology

Updates of Cost of Illness and Quality of Life Estimates for Use in Economic Evaluations of HIV Prevention Programs

Holtgrave, David R.; Pinkerton, Steven D.

Free Access
Article Outline
Collapse Box

Author Information

Center for AIDS Intervention Research, Medical College of Wisconsin, Milwaukee, Wisconsin, U.S.A.

Address correspondence and reprint requests to David R. Holtgrave, Ph.D., Associate Professor and Director of Cost-Effectiveness Studies Core Center for AIDS Intervention Research, Medical College of Wisconsin, 1249 North Franklin Place, Milwaukee, WI 53202, U.S.A.;holtgrav@post.its.mcw.edu.

Manuscript received October 8, 1996; accepted February 28, 1997.

Collapse Box

Abstract

To allocate limited economic and other resources for HIV prevention and treatment for maximum benefit, health policy planners and decision makers require accurate, current estimates of the lifetime costs of HIV-related illness and the impact of therapy on the quality of life of HIV-infected persons. These data are central input parameters to the economic evaluation methodology known as cost-utility analysis. The estimates available in the literature are already outdated, and this paper presents updated estimates of the projected lifetime health care costs associated with HIV disease in the United States and the number of quality-adjusted life years(QALYs) lost to HIV in light of recent advancements in HIV diagnostics and therapeutics. Results indicate that the lifetime cost of HIV medical care has grown from about$55,000 U.S. to more than $155,000 U.S., while the number of QALYs lost per case of HIV infection has decreased from 9.26 to 7.10, when discounted at a 5% annual rate. When these figures are discounted instead at the newly recommended 3% rate, lifetime costs rise to more than $195,000 U.S. and lost QALYs increase to 11.23. The net effect of these increases in the medical costs of care and treatment saved by averting an HIV infection and in QALYs makes HIV prevention a relatively more cost-effective strategy than other, non-HIV health-related programs.

The development of new generations of antiretroviral drugs, including protease inhibitors and nonnucleoside reverse transcriptase inhibitors; combination drug therapies; and improved techniques for monitoring disease progression and therapeutic effectiveness have radically altered how HIV disease is perceived. Despite their promise to prolong survival and improve the quality of life of persons infected with HIV, the new therapeutic regimens are also much more costly than their predecessors. Moreover, as persons live longer, they consume greater health care resources, driving the overall health care costs associated with HIV infection even higher. Resources to fund HIV prevention programs are limited and must be used wisely to maximize their prevention potential(1). Cost-effectiveness analysis (CEA) and related methods of economic evaluation can be used to help decision makers allocate available HIV prevention funds prudently(2,3).

The U.S. Public Health Service Panel on Cost-Effectiveness in Health and Medicine issued guidelines for the conduct of economic evaluation studies(4) in which they recommended cost-utility analysis (CUA) as the standard methodology for economic evaluation. The U.S. Centers for Disease Control and Prevention (CDC) defines CUA as "A type of cost-effectiveness analysis in which benefits are expressed as the number of life years saved adjusted to account for loss of quality from morbidity of the health outcome or side effects from the intervention"(2).

The most common outcome measure employed in CUA is the number of quality-adjusted life years(QALYs) saved by the health services program under consideration. One QALY equals 1 year at full health(e.g., 2 years of survival at a half-diminished quality of life would equal 1 QALY). In general, the number of QALYs remaining in a person's life from any given year of life (i = 1) until death (i = I) is obtained by summing yi * qi, for which yi equals unity for each year that the person is alive, and qi is the person's perceived health-related quality of her or his life for that year. One of the principal advantages to the use of QALYs saved by an intervention is that they provide a disease-independent measure of programmatic effectiveness, unlike, for example, the number of HIV infections averted or cases of tuberculosis prevented.

The results of a CUA are typically expressed as a cost-utility ratio that compares the cost per QALY saved of two health service programs(2,4,5). As a simplified example, suppose a decision maker wishes to compare a policy of funding a new HIV testing and counseling program with a "do nothing" option in which the program is left unfunded. The simplified cost-utility ratio would take this form: Equation 1 for which C is the cost of the counseling and testing program relative to the do-nothing option, A is the number of HIV infections averted by the program, T is the current value of the medical costs of care and treatment saved by averting an HIV infection, and Q is the number of quality-adjusted life years (QALYs) saved by preventing an infection.

Equation 1
Equation 1
Image Tools

The Panel on Cost-Effectiveness in Health and Medicine recommends that common methods be used across CUA studies to ensure maximum comparability(4). When applied to a specific illness area, such as HIV prevention, this generic recommendation can be taken a step further. Comparability would be enhanced if all economic evaluation studies of HIV prevention programs used common values of T and Q, at least in referent cases. However, a review of the economic evaluation literature on HIV prevention found little evidence of cross-study standardization of analytic methods (1). This is unfortunate, because methodologic choices made by the various research teams may limit the comparability of studies across types of HIV prevention programs. If this comparability is limited, so too is the potential for economic evaluation studies to inform public health decision makers who are responsible for allocating fiscal resources. If methodologic choices and certain common parameter values (such as those for T and Q) can be employed across studies, the entire literature on HIV-related economic evaluations can become more useful.

Although carefully estimated T and Q values for the equation have been reported in the literature previously(6,7), they should be updated to reflect important scientific developments in antiretroviral therapy and the greatly increased costs associated with these break-through therapies. A panel of clinical investigators convened by the International AIDS Society-U.S.A. (IAS) recommended combination drug therapy with two or more antiretroviral agents for most HIV-infected patients(8). The recommended treatment regimens, especially those that incorporate one or more protease inhibitors, have dramatically changed the costs of treating HIV disease and improved the projected survival of persons living with HIV infection(8-15). As a result, previous characterizations of the course of HIV disease may no longer be valid. As assumptions about disease progression change, so must calculations of the QALYs saved when HIV infections are prevented, because these calculations depend on the course of disease progression and its impact on quality of life. Moreover, the empiric literature on measurements of HIV-infected persons' perceptions of their quality of life also has expanded.

In this paper, we describe previously published estimates of T and Q values and review the literatures relevant to the task of updating these estimates. We also compare the available information for calculating Q with the ideal information to meet all recommendations of the Panel on Cost-Effectiveness in Health and Medicine. Several scenarios of HIV disease progression and treatment-related costs are constructed that incorporate explicit assumptions about the impact of protease inhibitors and other promising therapies. Estimates of T and Q values for persons undergoing treatment in the United States are calculated for each of these scenarios. This range of estimates provides updated T and Q values that are appropriate for use in economic evaluations of HIV prevention programs. We also propose a single paired estimate of T and Q values for use as a reference case in all economic evaluation studies of HIV prevention programs. Such standardization in referent case analyses can substantially increase the comparability of important, policy-relevant economic evaluation study results. However, we also emphasize the limitations of the available information to calculate T and Q values and the need for further research in these areas.

Back to Top | Article Outline

METHODS

Previous Estimates of Cost of Illness and Quality of Life

Guinan et al.(5) previously estimated the discounted value of the cost of medical care and treatment services saved each time an HIV infection is prevented. They assumed that the length of time from HIV infection to death was 12 years. This 12-year survival period was divided into four phases: (A) 6 years of being unaware of one's HIV seropositive status; (B) 3 years of being aware of one's HIV seropositivity with a CD4 cell count of between 200 and 499/mm3; (C) 1 year of AIDS as defined by a CD4 cell count of less than 200/mm3; and (D) 2 years of AIDS as defined by clinical conditions such as opportunistic infections. In calculating overall costs, Guinan et al. used Hellinger's empirically derived cost of illness estimates(16) -which were based on the AIDS Cost and Service Utilization Survey-for each year that a person is aware that he or she is infected. The total, undiscounted cost of HIV-related illness was $93,696 (in 1992 U.S. dollars). Discounting the cost of illness for each year into current values using a 5% discount rate reduced this estimate to $55,640 (in 1992 U.S. dollars).

The number of discounted QALYs saved each time an HIV infection is prevented has also been estimated previously. Holtgrave and Qualls(7) used the survival and disease progression framework employed by Guinan et al.(5), with the additional assumption that the average age at the time of HIV infection was 26 years. Their review of the then available literature on HIV-infected persons' quality of life led to the following estimates for quality of life in each disease stage: (A) full health for persons unaware of their HIV infection status; (b) 0.9 of full health for persons aware of their HIV infection with CD4 cells counts of between 200 and 499/mm3 (based on one prior, expert judgment study); (C) 0.65 of full health for persons with AIDS as defined by a CD4 cell count of less than 200/mm3 (based on three prior, empiric studies with AIDS patients); and (D) 0.40 of full health for persons with AIDS as defined by clinical conditions (based on four prior, empiric studies with AIDS patients, including one international study). Assuming that non-HIV-infected individuals enjoy full health, the average number of undiscounted QALYs saved (before age 65) each time an infection is averted is 28.85. When discounted at a 5% rate, this figure is reduced to 9.26 QALYs saved; it is standard practice in economic evaluation to discount health benefits at the same rate as monetary costs.

Holtgrave and Qualls(7) cited three reasons for truncating their analysis at age 65: the number of years of potential life lost before age 65 is a common measure used in public health research; given that HIV infection occurs largely among young persons (i.e., infection on average at age 26 in the United States), discounted QALYs saved after age 65 have a very small effect on estimates of QALYs saved by preventing HIV infection; and based on 1991 life tables, life expectancy at birth would be approximately 69.9 to 71.1 years if adjusted for gender and race or ethnicity characteristics matched to demographic characteristics of cumulative AIDS case statistics, and this figure would be lower if appropriate levels of adjustment were known to account for the detrimental health effects of HIV-related risk behaviors such as injection drug use.

Back to Top | Article Outline
Reasons for Updating T and Q Parameters

The previously published estimates of T and Q values need to be updated for several reasons. The primary reason is that new drug therapies have dramatically changed assumptions about the course of illness for persons living with HIV infection(8-15). These promising drugs, however, are also expensive(14). Combination therapies with two or more antiretroviral agents, especially with regimens that include one or more protease inhibitors, can reduce plasma viral load by up to several log units(10,11). Several studies have demonstrated that viral load reductions of one or more log units are associated with significant clinical benefit. Investigators also believe that early treatment with effective antiretroviral regimens can lower the "set point" at which viral levels stabilize after the brief period of primary infection. Additional research has established that this steady-state viral load is predictive of long-term clinical outcomes, including time of progression to AIDS(8,13).

Several different combinations of antiretroviral drugs have been shown to decrease plasma viral titers to below the level of detection. At these low levels of infection, the rate of HIV genetic mutation is slowed, which delays possible development of drug-resistant viral subtypes. Moreover, combination therapy multiplies the number of genetic sites at which HIV must mutate to remain viable, further slowing viral evolution toward resistance(10,11). Although long-term effects of combination antiretroviral therapy remain uncertain, there is substantial basis for optimism. Besides possible side effects, the most significant downside to these promising therapies is the high cost of the new drugs, which limits their availability to some persons in developed countries, at least for the near future(11).

Another reason for updating T and Q estimates is to comply with the recent recommendation of the U.S. Public Health Service's Panel on Cost-Effectiveness in Health and Medicine that all economic evaluation studies employ a 3% discount rate in the base case analysis(4).(They also recommended performing analyses at 0% and 5% rates as primary comparison cases.) Previous estimates of T and Q values were based on a 5% rate, which results in smaller estimated values. Because HIV disease progresses over a time horizon of more than a decade, the selection of a discount rate value can substantially affect derived estimates of T and Q.

The quality of life literature relevant to HIV disease has expanded with the implementation of new therapies. Because the empiric basis for qualify of life estimates (on which calculations of Q rely) has grown, that figure too should be updated. The Panel on Cost-Effectiveness in Health and Medicine(4) has recommended that quality weights should be empirically measured by means of a representative sample of the national population; hence, most respondents would be expressing preferences for health states that by and large they have not experienced. As subsequently discussed, no such studies have been accomplished. Our review of the HIV-related quality of life literature therefore was focused on finding the closest approximations to the Panel's ideal standard for measuring quality weights.

Back to Top | Article Outline
Scenario Analysis

The reasons elucidated previously for updating T and Q values are interrelated. To take this into account, we have constructed three different scenarios, each of which is a plausible interpretation of the state of the field regarding HIV treatment and quality of life. The three scenarios are labeled low cost, intermediate cost, and high cost. Throughout, all costs are expressed in June 1996 dollars(17).

Table 1 displays the definitions of the disease phases employed in the analyses. Phases 2 through 4 are self-explanatory, but phase 1 is slightly more complicated. For phase 1, there are four different subphases(A through D), which range from being unaware of one's HIV seropositivity to being aware and receiving viral load monitoring and various levels of drug treatment. In the three scenarios described, some of the subphases of phase 1 may be skipped.

Table 1
Table 1
Image Tools

Table 2 displays quality of life estimates from several empiric studies(18-23). Because quality of life adjustments should be preference based(4), we included only empiric studies that purported to measure the perceived value of HIV-related health states, as opposed to studies measuring the health states as such. We focused on studies done in the United States so as not to mix results from patients in different national systems of health care. We searched for studies whose results could be matched to one of the disease phases listed in Table 1; in many cases, the investigators had labeled disease stages somewhat differently from the scheme outlined in Table 1, but CD4 cell count levels were used as a matching criterion whenever possible. If a study presented baseline and follow-up assessments of patients' quality of life perceptions, only the follow-up assessment was used. If a study had more than one estimate of the quality of life of a disease phase (using the same assessment methodology), we averaged the two figures.

Table 2
Table 2
Image Tools

HIV-infected patients formed the study population of almost all the identified studies; the one exception surveyed physicians as proxies for their patients(23). We included this physician proxy survey because it could be considered the one study that comes closest to the Panel on Cost-Effectiveness in Health and Medicine's recommendation that quality adjustments should be made on the basis of community-wide surveys of persons not necessarily living with the particular disease in question(4). None of the studies fully met the Panel's recommendation.

Because the studies used various methodologies to assess quality of life and included different numbers of subjects, we chose the median of the estimates within each disease phase as a summary statistic to minimize the impact of outliers; in this case, an outlier figure may reflect particular methodologic choices of the researchers. The resulting quality of life estimates for phases 2, 3, and 4 are 0.76, 0.65, and 0.62, respectively. For an overview of the HIV-related health status and quality of life measurement studies that include but are not limited to those meeting our inclusion criteria, see the work of Holzemer and Wilson(24); this review and the original articles cited in Table 2 are important sources for psychometric information on the various quality of life measures.

Persons living with HIV and AIDS perceive disease phases 3 and 4 to be approximately equal (0.65 versus 0.62, respectively). This may result from methodologic issues of assessments or from patients learning to cope with HIV disease and using these coping strategies to improve or maintain their quality of life even as health problems mount. This finding from the growing literature on quality of life of persons living with HIV and AIDS is counter to the a priori assumption that quality of life must markedly diminish as disease progresses substantially; if a community-wide survey were done (according to the recommendations of the Panel on Cost-Effectiveness in Health and Medicine), the results might be quite different.

There is very little empiric guidance in the literature on quality of life values for phases 1A through 1D. We therefore assumed that the quality of life in phase 1A is 0.94, in accordance with published estimates of the quality of life of the U.S. population at large, as determined through empiric study(25). We then imposed the condition that the quality of life difference between phases 1C and 1B should be larger than the difference between phases 1B and 1A and then proceeded similarly for phases 2, 1D, 1D, and 1B. As disease progresses and additional treatments are imposed (with possibly toxic side effects), quality of life diminishes at an accelerating rate. Although it could be argued that monitoring or treatment should prolong survival or enhance quality of life, we have instead assumed that the initiation of treatment marks a decline in overall health and quality of life. The quality of life figures of 0.91, 0.87, and 0.82 for phases 1B, 1C, and 1D, respectively, satisfy the assumption described previously. However, we emphasize that these estimates are largely arbitrary.

Back to Top | Article Outline
Low-Cost Scenario

The low-cost scenario is displayed in Table 3. This scenario may be thought of as describing the cost of illness for someone with a very low level of access to current HIV disease treatments. In this scenario, the disease progression assumptions are exactly those made by Guinan et al.(6) and Holtgrave and Qualls (7) in deriving previous estimates of T and Q values. In particular, it assumes that the person is unaware of HIV seropositivity until phase 2 commences. Thereafter, only zidovudine(ZDV) monotherapy is accessed. No viral load monitoring is available.

Table 3
Table 3
Image Tools

The costs for HIV disease other than ZDV monotherapy are taken from two sources. The first source is the work of Hellinger(16), which was based on patients' actual experiences incurring HIV-related costs, and the second source is the work of Gable et al.(26), which was based on expert judgments about the implementation of various protocols for treating HIV disease. The quality of life estimates for each disease phase are also displayed in Table 3. This scenario describes a person receiving care and treatment below currently accepted standards and perhaps reflects the experience of someone with very low access to state-of-the-art treatment.

Back to Top | Article Outline
Intermediate-Cost Scenario

The intermediate-cost scenario (Table 4), which we consider the base case analysis, reflects as closely as possible current recommendations for the treatment of HIV disease(8). There is still uncertainty in a number of the assumptions made, and we recognize that individual patients' experiences will differ from our illustration of an average patient. In this scenario, the time that a patient is unaware of his or her seropositivity is substantially decreased relative to the low-cost scenario. As more and more persons learn of the availability of promising new treatments, we believe that there will be a greater impetus for individuals to be tested earlier in the course of illness.

Table 4
Table 4
Image Tools

Consistent with IAS recommendations(8), we assume that the typical patient progresses through periods of viral load monitoring only, two-drug combination therapy, and three-drug combination therapy. The cost of a viral load test has been estimated to be between $71 and$100 U.S.(14,27); assuming approximately two tests per year, we estimate the annual cost of viral load monitoring to be $175 U.S. in this scenario.

For two-drug combination therapy, we assume that the patient receives ZDV and lamivudine (2'-deoxy-3'-thiacytidine [3TC]), in accordance with IAS recommendations; three-drug therapy adds a protease inhibitor to this regimen. Standard dosing schedules and drug costs to pharmacists were obtained from a published source(12). Saquinavir was selected to represent the class of protease inhibitors because of its intermediate cost (i.e., greater than indinavir and less than ritonavir)(12). We assumed that after treatment with a drug was initiated, the drug would not be withdrawn, or if withdrawn, that another drug of equal cost would be substituted. This yielded a $12,900 U.s. estimate of the annual cost of three-drug therapy. A review of combination therapies(14) estimated that the monthly cost of three-drug therapy ranges from $904 to $1201 U.S.; taking the midpoint of this range and multiplying by 12 months yields an annual estimate of$12,630 U.S., a figure quite close to our estimate.

The impact of these expanded treatments on survival and disease progression can only be estimated at this point. However, there is substantial reason for optimism that combination therapy can extend average survival and prolong the early, relatively disease-free phases of HIV infection. The key to this optimism is the pronounced reduction in viral load evident in many patients receiving these therapies(8). Because increased viral load is associated with immune system deterioration and more rapid progression to AIDS, it is believed that durable suppression of viral levels is crucial for enhanced clinical outcomes. After immune function has significantly deteriorated, the benefit of the new therapeutics is likely to be attenuated. In our determinations, we assumed that phase 1B would last for 1 year, phase 1C for 3 years, and phase 1D for 3 years. The total of all phase 1 subphases is 9 years in this scenario, as opposed to the 6-year duration assumed by Guinan et al.(6) (compare also the low-cost scenario). The length of time spent in phase 2 is increased from 3 years (as in the low-cost scenario) to 4 years. The intermediate-cost scenario expands overall survival from 12 years to 16 years. We are aware of the high level of uncertainty in the survival estimate in this scenario.

Back to Top | Article Outline
High-Cost Scenario

The high-cost scenario describes the possible experiences of an HIV-infected person who has high levels of access to health care services(Table 5). This scenario also makes the most optimistic assumptions about the effects of new drug therapies on survival.

Table 5
Table 5
Image Tools

This scenario assumes almost instant awareness of HIV seropositivity and continuous access to viral load monitoring services. It also assumes that three-drug therapy is initiated in phase 1D and extends through phase 4, with a consequent doubling of the average duration of phases 1 and 2 to 12 and 6 years, respectively. The total duration of survival after HIV infection in this scenario is 21 years.

Back to Top | Article Outline

RESULTS

The results of the scenario analysis are displayed inTable 6. Results are shown for each of the three scenarios, for each of the two different sources of nondrug costs(16,26), and for each of three different discount rates (0%, 3%, and 5%). Cost of illness figures (estimates of T) are displayed in pairs with QALYs saved values(estimates of Q); these figures should always be used in pairs, as reflected in the cost-utility ratio (see Eq. 1).

Table 6
Table 6
Image Tools

The Hellinger-based (Gable-based) estimates reported in Table 6 range from a low of T = $71,143 ($46,236) and Q = 8.57 to a high of T = $424,763 ($351,053) and Q = 20.37 (all in U.S. dollars). The intermediate cost, base case estimate of T is $195,188 (using Hellinger's nondrug costs), while the base case value of Q is 11.23. The base case results reflect a discount rate of 3%; at a 5% rate, the corresponding estimates are T = $157,348 and Q = 7.10.

Back to Top | Article Outline

DISCUSSION

The base case results for T = $195,188 and Q = 11.23 are both higher than previous estimates for T (estimated at $55,640 [6]) and Q (estimated at 9.26 [7]). The increase in T occurs for several reasons. First, even within a particular disease stage, the costs of treatment and disease monitoring are rising. Second, we assume increases in survival from new therapies (i.e., the costs are accumulated for a longer period). Third, a lower discount rate is now the standard in the field, and this drives up the discounted cost figure. Fourth, costs are now expressed in June 1996 dollars. As T increases, HIV prevention programs generally appear to be more cost-effective because the numerator of the cost-utility ratio is reduced.

The parameter Q is higher than previous estimates. This occurs primarily because of the lower discount rate now employed. This is true even though two factors tend to decrease the size of Q. First, the new quality of life weights employed in our calculations of Q are higher than previous estimates of quality of life weights; this tends to decrease Q. Second, our assumption of increased survival from new treatments also tends to decrease the parameter Q. Both of these downward tendencies, however, are more than offset by the change in the discount rate. The resultant higher estimate of Q also makes HIV prevention programs appear to be more cost-effective, because as Q increases, the denominator of the cost-utility ratio increases, and the overall cost-utility ratio decreases. In the future, Q may decrease over time if the discount rate and quality weights stabilize in the literature and therapies become more effective at increasing survival.

Increases in the size of T and Q individually make HIV prevention programs appear more cost-effective. The joint effects of increased values for T and Q are even more pronounced in reducing the size of the cost-utility ratio for any given pairs of values for C and A. To see this, consider the cost-utility ratio introduced above (see Eq. 1). Although there is no single, universally accepted cost-utility ratio for determining whether a program is cost-effective, health service programs with cost-utility ratios less than about $30,000 U.S. per QALY saved are generally considered cost-effective, whereas those with ratios more than $140,000 U.S. per QALY saved are difficult to defend as cost-effective(28-30). Because the cost-utility ratio associated with a particular C and A pair is always smaller when the new estimates of T and Q are used, the ratio using the new T and Q values is more likely to be in the range of cost per QALY saved that is usually considered cost-effective. Some HIV prevention programs that might not have been considered cost-effective with the previous estimates of T and Q may now be considered to be so.

We propose that the base case estimates of T and Q derived should be used by analysts and others who perform economic evaluations of HIV prevention programs. Such a referent case analysis can render these studies more comparable and more useful to policy makers. Because of the uncertainty in T and Q values, economic evaluation studies should include sensitivity analyses in which these parameters are varied over a range of plausible values. Table 6 provides a wide range of values for use in such sensitivity analyses. At a minimum, the range of values reflected in the intermediate case(i.e., middle rows of Table 6) should be employed in sensitivity analyses of T and Q. For purposes of consistency of assumptions about survival and course of illness, only T and Q pairs should be taken from Table 6.

This analysis is subject to limitations. Most important perhaps is the uncertainty about survival and disease progression under the new drug regimens. It will be several years before empiric data are available on survival changes for all disease phases induced by the new therapies. However, our assumptions were varied across a wide range of possibilities, and base case assumptions are cautious rather than overly optimistic. It is important to make explicit these sources of uncertainty so that they can be debated and updated over time. The quality of life weights in the available literature do not correspond with the assessment recommendations of the Panel on Cost-Effectiveness in Health and Medicine. Further empiric research is needed to meet their suggestion for a population-wide survey to assess quality weights. Our estimate of Q should be updated if and when such data become available.

In summary, the results of this analysis suggests that newly developed drug therapies could have a tangible impact on the cost-effectiveness of different HIV prevention programs. Relative to non-HIV health care or health promotion programs, the cost-effectiveness of HIV prevention is enhanced by the increased savings realized each time a case of HIV infection is averted. The estimates reported reflect limited foreknowledge of the ultimate effect of novel therapies on the course of HIV disease progression. Moreover, the rapid pace of antiretroviral development ensures a limited longevity for even the most accurate estimates of HIV-related treatment costs(31) and quality of life. Nevertheless, HIV policy decisions, especially resource allocation decisions, are being made now, and policy makers deserve the best decision aids possible. Cost-effectiveness analyses can assist their difficult task of making these tough choices, and such analyses should be based on the very best available parameter estimates.

Acknowledgments: This research was supported by NIMH grants P30-MH52776 and RO1-MH55440. We thank Drs. Paul Farnham and Mary Guinan for helpful discussions in conceptualizing this project.

Back to Top | Article Outline

REFERENCES

1. Holtgrave DR, Qualls NL, Graham JD. Economic evaluation of HIV primary prevention. Annu Rev Public Health 1996;17:467-88.

2. Haddix AC, Teutsch SM, Shaffer PA, Duñet DO, eds. Prevention effectiveness. New York: Oxford University Press, 1996.

3. Holtgrave DR. Setting priorities and community planning for HIV-prevention programs. AIDS Public Policy J 1994;9:145-51.

4. Gold MR, Siegel JE, Russell LB, Weinstein MC, eds. Cost-effectiveness in health and medicine. New York: Oxford University Press, 1996.

5. Holtgrave DR, Kelly JA. Preventing HIV/AIDS among high-risk urban women: the cost-effectiveness of a behavioral group intervention.Am J Public Health 1996;86:1442-5.

6. Guinan ME, Farnham PG, Holtgrave DR. Estimating the value of preventing a human immunodeficiency virus infection. Am J Prev Med 1994;10:1-4.

7. Holtgrave DR, Qualls NL. Threshold analysis and HIV prevention programs. Med Decis Making 1995;15:311-7.

8. Carpenter CCJ, Fischl MA, Hammer SM, et al. Antiretroviral therapy for HIV infection in 1996: recommendations of an international panel. JAMA 1996;276:146-54.

9. Voelker R. Can researchers use new drugs to push HIV envelope to extinction? JAMA 1996;276:435-7.

10. Ho, DD. Time to hit HIV, early and hard. N Engl J Med 1995;333:450-1.

11. Richman DD. HIV therapeutics. Science 1996;272:1886-7.

12. Anonymous. New drugs for HIV infection. Medical Letter on Drugs and Therapeutics. 1996;38:35-7.

13. Mellors JW, Kingsley JA, Rinaldo CR Jr, et al. Quantitation of HIV-1 RNA in plasma predicts outcome after seroconversion. Ann Intern Med 1995;122:573-9.

14. Deeks G, Smith M, Holodnly M, Kahn JO. HIV-1 protease inhibitors: a review for clinicians. JAMA 1997;277:145-53.

15. American Health Consultants. Panels to develop standards for antiviral therapy, viral-load testing. AIDS Alert 1997;12:13-5.

16. Hellinger FJ. The lifetime cost of treating a person with HIV.JAMA 1993;270:474-8.

17. U.S. Bureau of Labor Statistics. Consumer price index tables. World Wide Web site, http://stats.bls.gov:80/cgi_bin/surveymost, August 13 1996 (data also available from first author).

18. Tsevat J, Solzan JG, Kuntz KM, et al. Health values of patients with human immunodeficiency virus: relationship to mental health and physical functioning. Med Care 1996;34:44-57.

19. Wu AW, Mathews WC, Brysk LT, et al. Quality of life in a placebo-controlled trial of zidovudine in patients with AIDS and AIDS-related complex. J Acquir Immune Defic Syndr 1990;3:683-90.

20. Revicki DA, Wu AW, Murray MI. Change in clinical status, health status, and health utility outcomes in HIV-infected patients. Med Care 1995;33(Suppl):AS173-AS182.

21. Wu AW, Rubin HR, Mathews WC, et al. A health status questionnaire using 30 items from the Medical Outcomes Study. Med Care 1991;29:786-98.

22. Schag CAC, Ganz PA, Kahn B, Petersen L. Assessing the needs and quality of life of patients with HIV infection: development of the HIV Overview of Problems Evaluation System (HOPES). Qual Life Res 1992;1:397-413.

23. Owens DK, Nease RF Jr. Transmission of HIV infection between provider and patient: a quantitative analysis of risk. In: Kaplan EH, Brandeau ML, eds. Modeling the AIDS epidemic: planning, policy, and prediction. New York: Raven Press, 1994:153-77.

24. Holzemer WL, Wilson HS. Quality of life and the spectrum of HIV infection. Annu Rev Nurs 1995;13:3-29.

25. Patrick DL, Erikson P. Health status and health policy: allocating resources to health care. New York: Oxford University Press, 1993.

26. Gable CB, Tierce JC, Simison D, Ward D, Motte K. Costs of HIV+/AIDS at CD4+ counts disease stages based on treatment protocols. J Acquir Immune Defic Syndr Hum Retrovirol 1996;12:413-20.

27. American Health Consultants. Viral-load testing may lower high cost of therapy. AIDS Alert 1996;11:100-1.

28. Laupacis A, Feeny D, Detsky AS, Tugwell PX. Tentative guidelines for using clinical and economic evaluations revisited. Can Med Assoc J 1993;148:927-9.

29. Owens DK, Nease RF, Harris R. Use of cost-effectiveness and value of information analyses to customize guidelines for specific clinical practice settings [abstract]. Med Decis Making 1993;13:395.

30. Tolley GL, Kenkel D, Fabian R, eds. Valuing health for policy: an economic approach. Chicago: University of Chicago Press, 1994.

31. Pinkerton SD, Holtgrave DR. Lifetime costs of HIV/AIDS medical care [letter]. J Acquir Immune Defic Syndr Hum Retrovirol 1997;14:380-2.

Keywords:

Economc evaluation of HIV/AIDS; Cost and cost-benefit analysis; Quality of life

© Lippincott-Raven Publishers.

Login

Search for Similar Articles
You may search for similar articles that contain these same keywords or you may modify the keyword list to augment your search.