Cost-Effectiveness of Syringe Exchange as an HIV Prevention Strategy

Objective: To analyze the cost-effectiveness of New York State-approved syringe exchange programs (SEPs) and estimate the cost-saving potential of these programs.

Design and Methods: The cost-effectiveness analysis used cost and process data provided by seven SEPs for the calendar year 1996 or the most recent 12-month period available at the time of the study. Alternative estimates of the number of HIV infections prevented were calculated using published data and a simplified circulation model. HIV treatment costs were taken from the literature.

Cited Here...: A cost-effectiveness ratio of $20,947 per HIV infection averted was calculated based on an estimated 87 HIV infections averted across the seven programs and total program costs of $1.82 million (all amounts given in US dollars). Sensitivity analyses were also performed. Using imputed costs for donated services and estimates of lifetime costs to treat an HIV infection, syringe exchange was demonstrated to be a cost-saving strategy from a societal perspective.

Conclusions: This research demonstrates that syringe exchange is a cost-effective and cost-saving strategy for reducing HIV transmission.

The Centers for Disease Control and Prevention of the U.S. Department of Health and Human Services (CDC) report that 753,907 persons with AIDS have been reported in the United States through June 2000. Of these, 36% of the adult AIDS cases and 52% of the pediatric AIDS cases (children less than 13 years old) are reported in conjunction with an exposure category involving injecting drug use ⁽¹⁾. Recent surveillance by the New York State (NYS) Department of Health's Bureau of HIV/AIDS Epidemiology reveals a similar, although more disconcerting, picture of HIV transmission in NYS resulting from injecting drug use. As of June 2000, the Bureau of HIV/AIDS Epidemiology reported 142,376 AIDS cases in NYS, with injecting drug use reported as either the primary or secondary risk factor for 47% of the adult AIDS cases and 64% of the pediatric AIDS cases ⁽²⁾. These data as well as those reported by the CDC at the national level demonstrate the significant role that injection drug use continues to play in the HIV/AIDS epidemic.

Syringe exchange is one intervention that can reduce HIV transmission among intravenous drug users ^(3-6). The objective of syringe exchange is to reduce the transmission of HIV and other blood-borne pathogens such as hepatitis B virus (HBV) and hepatitis C virus (HCV) by decreasing the average circulation time of syringes, thereby reducing the potential for the reuse of contaminated syringes. Data from surveys of participants in four syringe exchange programs (SEPs) reveal that in the absence of SEPs, injection drug users (IDUs) reused their syringes more than seven times on average ⁽⁵⁾. Kaplan ⁽⁷⁾ and Kaplan and Heimer ⁽⁸⁾ demonstrated that SEPs could reduce the average circulation time of a syringe from 23.5 days to less than 3 days. Recent experiments have revealed that HIV-1 survives for up to 30 days in blood within syringes ⁽⁹⁾. Given this extended ability of syringes to transmit viable HIV-1, the effectiveness of syringe exchange rests in its removing a large proportion of contaminated syringes from circulation by reducing their circulation time.

Where applicable, cost-effectiveness analysis provides important information to inform policy debates. The question of whether to establish or to continue to promote syringe exchange as a component of a strategy to decrease HIV transmission is such a case. To better inform policy within NYS, the NYS Department of Health's AIDS Institute undertook a cost-effectiveness analysis of SEPs operating in the state in 1996. It was the intent of this analysis to report cost-effectiveness in terms of cost per HIV infection averted, consistent with other economic analyses found in the HIV prevention literature [the reader is referred to the article by Holtgrave et al. ⁽¹⁰⁾ for a review], so as to provide some measure of comparability with non-SEP prevention activities.

SYRINGE EXCHANGE PROGRAMS IN NEW YORK STATE

Through the enactment of state regulation authorizing legal syringe exchange in NYS by approved programs, the NYS Department of Health permitted the establishment of five SEPs in New York City (NYC) in 1992 ⁽⁴⁾. At the time of this study (April 1997), 12 SEPs were legally operating in NYS. Nine programs were located in NYC, with programs also located in Buffalo, Rochester, and White Plains. State regulations authorizing the legal operation of SEPs require that such programs provide syringe exchange in the context of comprehensive harm reduction services such as outreach, distributing condoms and bleach kits, making referrals for HIV counseling and testing, and providing literature and instruction on HIV prevention and safer injection techniques. Although not required, a few programs also provide the opportunity for clients to participate in individual or group counseling sessions or to obtain HIV counseling and testing services on-site.

METHODS

Programs were asked to provide cost and other information either for the calendar year 1996 or for the most recent 12-month period available. The analysis was undertaken from the perspective of the programs in the aggregate as well as from that of society, where all costs would be considered regardless of source.

Ten of the 12 programs in operation at the time of the study (8 in NYC and the Buffalo and Rochester programs) participated in the study. Because the data reported by three programs were incomplete, our analysis is limited to 7 programs, 6 of which were operating in NYC and 1 in Rochester. Programs were requested to report the number of unduplicated individuals participating in their program as well as the number of syringes distributed and returned. The reported number of unduplicated participants is an important factor in the analysis, as it is this number on which HIV incidence and the number of HIV infections averted, the effectiveness measure for the analysis, are based as described below. (The terms participant and client are used interchangeably to refer to individuals participating in SEPs.)

Effectiveness Measures

HIV Infections Averted

Two methods were used to estimate the number of HIV infections averted during the time of the study, the primary measure of effectiveness for this analysis. The first method is based on the reduction in the probability of HIV transmission to an HIV-negative SEP participant. The following equation uses this effect to estimate the number of HIV infections averted (A):EQUATION

Equation U1

for which all parameters are as defined in Table 1. In addition to the estimated HIV seroprevalence of the clients participating in the SEP (p), this equation recognizes that the number of HIV infections averted among HIV-negative participants is a function of these clients' SEP participation rate (a) (NYS Department of Health, AIDS Institute, unpublished data) and the reduction in risk from SEP participation (r) as estimated by Des Jarlais et al. ⁽⁴⁾.

Table 1

The number of HIV infections averted can also be estimated using a simplified circulation model ⁽¹¹⁾. This model uses the number of needles exchanged per client year (E) and the number of shared injections per IDU per year (s) to estimate the decrease in HIV incidence through SEP participation. The decrease in HIV incidence can be estimated as E/ (E +s), which can be applied to the projected number of SEP clients who would contract HIV in the absence of the SEP to estimate the number of HIV infections averted.

Costs

Costs of Syringe Exchange Programs

The SEPs participating in this study were requested to provide cost information for either the 1996 calendar year or the latest 12-month period available. Programs reported costs for personal services (including fringe benefits) for the syringe exchange activities as well as for other required (e.g., condom and bleach kit distribution, HIV prevention education) and ancillary (e.g., counseling and support groups) activities. Expenses relating to supplies, materials, travel, subcontracts, and other nonpersonal services were also collected. The seven programs incurred costs of $1,822,426 for personal and other than personal service costs of operation, including syringe exchange and other required services as well as ancillary services.

To attempt a societal perspective rather than to perform the analysis from the perspective of any one funding source (e.g., government or other funding agency), the programs were instructed to report costs regardless of the source of funding support. As a result, in-kind or donated services were also reported and valued by each program either at the level at which similar services were being purchased or at estimated market value. Across all programs, these services were valued at $320,914, the majority of which was for personal services ($228,361 or 71.15% of all donated services costs). Costs incurred by participants (e.g., costs for travel or lost wages) were not included in the analysis, however, thereby limiting the extent to which the analysis can be said to reflect a societal perspective.

Costs of Treatment

In addition to the costs incurred in operating a program, costs saved by not having to treat individuals for infection are also considered. Estimates of treatment costs have been recently calculated to incorporate the costs of antiretroviral therapy (e.g., reverse transcriptase and protease inhibitors) and other sources or components of care. Pinkerton and Holtgrave ^(12,13) provide estimates of HIV-related medical care costs for three level-of-care scenarios (low, intermediate, and high), each of which is dependent on access and treatment regimen. In addition, the projected costs of each scenario are discounted at rates of 0%, 3%, and 5%. The intermediate scenario, which reflects initiation of combination therapy with a pair of nucleoside analog reverse transcriptase inhibitors on detection of infection and adding a protease inhibitor when a patient's CD4 count falls below 500 mm ³ , is recommended by Pinkerton and Holtgrave ^(12,13) for use in all base case analyses and is used for that purpose here.

Discounting

Discounting accounts for the differential timing of costs and benefits. Costs incurred today may result in effects and other costs realized sometime in the future. This adjustment, not to be confused with adjustments made to reflect the effect of inflation on a dollar's purchasing power, is intended to express costs in current dollars. Dollars spent in the future have less immediate value. Similarly, benefits may also be subject to discounting ^(14-16). In this analysis, however, a discount rate was not applied to the SEPs' reported operating costs, because such costs were incurred in the present (i.e., within a 1-year period). The recommended intermediate base case cost of HIV treatment used here, as calculated originally by Pinkerton and Holtgrave ^(12,13), was discounted at a rate of 3%.

Sensitivity Analysis

Sensitivity analyses were conducted to examine the effect of varying two key parameters, the estimated number of shared injections per IDU per year (s) in the simple circulation model and the HIV incidence among non-SEP users as estimated by Des Jarlais et al. ⁽⁴⁾, on the estimated number of HIV infections averted. Sensitivity analysis was not performed on program costs, as these costs were retrospectively obtained and reflect actual rather than projected costs of program operation.

RESULTS

The number of HIV infections averted was estimated using the formula and the simplified circulation model. Separate calculations were made using the formula for the programs in NYC and Rochester so as to recognize the difference in estimated HIV prevalence in each city's IDU population. Combining the results of these calculations results in an estimated 92 infections averted as a result of SEP participation. (Estimated HIV infections averted are rounded to the nearest whole number.)

To use the simplified circulation model, the parameters of the model need to be calculated. The needle exchange rate (E) is obtained by dividing the reported 1,667,682 needles distributed by the programs by an estimated 4507.84 client-years of participation (9333 clients × 0.483 = 4507.84 SEP client-years) and is estimated to be 369.99 per client-year. The number of shared injections per year (s) is calculated by multiplying IDUs' injection frequency by the proportion of injections that are shared. Kaplan and O'Keefe ⁽¹⁷⁾ reported a needle-sharing rate of 31.5% of all injections based on the proportion of needles distributed to one person that were returned by another person in their evaluation of the New Haven, Connecticut needle exchange program. Researchers have reported an injection frequency of approximately 780 injections per year among the IDU populations studied ^(4,17-19), and applying the 31.5% estimated needle-sharing rate to this injection frequency results in an estimated 245.70 (0.315 × 780) shared injections per IDU per year. Inserting these values in the formula results in an estimated decrease in HIV incidence of 60.09% (E/ [E +s] = 369.99 / [369.99 + 245.70]). The number of SEP clients who would contract HIV in the absence of the SEP is calculated by multiplying the number of full-time HIV-negative clients (as estimated in the first method above) by 0.0526, which is the estimated HIV incidence among non-SEP users as estimated by Des Jarlais et al. ⁽⁴⁾. Applying the 60.09% decrease in HIV incidence to the estimated 144.15 SEP clients who would contract HIV in the absence of the SEP (2740.43 × 0.0526) results in about 87 HIV infections averted.

The number of HIV infections averted as estimated by the two methods just described was not expected to be exactly the same but provides reasonably similar results. For the purposes of the base case analysis, we use the more conservative estimate of 87 HIV infections averted calculated using the simplified circulation model. It should be noted, however, that these calculations estimate the number of primary infections averted (i.e., infections among participants in the SEPs) and that no attempt was made to estimate or project the number of secondary infections that might arise among the sexual or needle-sharing partners of SEP participants who might already be infected.

Program Perspective

From the SEP perspective, an estimated 87 HIV infections would be averted across all programs because of the reduction in risk resulting from the use of syringe exchange, and total reported costs from the seven programs were $1,822,426. The resulting cost-effectiveness ratio for this base case scenario is $20,947 per HIV infection averted across all reporting programs. On a program-specific basis, the median cost per HIV infection averted was $41,011 for the program averting two infections, and the cost per HIV infection averted across all programs ranged from $11,648 to $129,008, with these programs averting seven and three infections, respectively.

Two sensitivity analyses were conducted to examine the effect of varying certain parameters on the baseline results obtained from the SEP perspective. The first analysis was conducted by varying the estimated number of shared injections per IDU per year (s). In the baseline case, we used a value of 31.5% as reported by Kaplan and O'Keefe ⁽¹⁷⁾, which is similar to an earlier result for this same parameter reported by Siegel et al. ⁽¹⁸⁾ in their evaluation of bleach program effectiveness, and multiplied this by the number of injections per IDU per year. In their analysis, Siegel et al. ⁽¹⁸⁾ acknowledged the important effect on the course of the AIDS epidemic of heterogeneity of risk behavior by assigning IDUs to three separate groups based on frequency of injection: low injectors (337 injections per year or about 1 injection per day), medium injectors (1311 injections per year or about 4 injections per day), and high injectors (3018 injections per year or about 8 injections per day). Based on survey data, these researchers estimated an average proportion of injections that are shared for each injector subgroup, ranging from 35% for the low group to 22% for the high group across various baseline and sensitivity analysis specifications. Substituting these values in the simplified circulation model and using the results to determine the cost per HIV infection averted results in cost-effectiveness ratios of $21,957 and $18,596 per HIV infection averted, respectively.

The second sensitivity analysis was conducted to examine the effects of varying the estimated HIV incidence among non-SEP users across the possible values presented by the confidence interval with which Des Jarlais et al. ⁽⁴⁾ report their estimate of this parameter. Using the lower boundary of the confidence interval, or an estimated HIV incidence rate of 2.41%, we calculate that about 40 HIV infections may be averted across all programs at a cost of $45,561 per HIV infection averted. At the upper boundary of the confidence interval, or an estimated HIV incidence rate of 11.49%, about 189 HIV infections may be averted across all programs at a cost of $9,624 per HIV infection averted.

Societal Perspective

The results presented here examine cost-effectiveness from a program perspective considering only those costs (including donated services) incurred by the programs in the delivery of SEP services, including those mandated or required and those considered ancillary. To attempt a societal perspective (i.e., to consider costs in addition to those incurred by the SEPs themselves), we include the imputed value of donated or in-kind services as well the costs of HIV treatment avoided by averting HIV seroconversion among the SEP users. As noted previously, Pinkerton and Holtgrave ^(12,13) have calculated a set of HIV treatment scenarios and their costs. They recommend using $195,188 for base case scenarios of a cost-effectiveness analysis, which represents what they term an intermediate level of care, with costs discounted at a rate of 3%. Therefore, averting 87 HIV infections through IDU participation in SEPs would save almost $17 million in treatment costs, making operation of SEPs a cost-saving intervention. Recognizing that IDUs may face certain barriers (self-imposed or otherwise) in obtaining HIV treatment ^(20,21), we can apply the costs provided by Pinkerton and Holtgrave ^(12,13) to represent a low cost level of care, resulting in an estimated savings of $7.6 million in HIV treatment costs (87 HIV infections averted × $87,045).

DISCUSSION

This analysis was conducted to examine whether the operation of SEPs in NYS could be considered an efficient use of limited resources. Recognizing that policy discussions concerning syringe exchange may be politically or socially charged, with parties being either supportive or not, we believed that demonstrating the cost-effective or cost-saving aspect of this intervention might make some contribution to the discussion.

Our analysis of the seven NYS SEPs studied here demonstrates that syringe exchange is a cost-effective and cost-saving HIV prevention intervention not only from the perspective of the programs themselves but from a societal perspective. This is consistent with findings from other similar analyses of SEPs [the reader is referred to the article by Kahn ⁽²²⁾ for a review of evaluations of SEPs and other interventions targeting IDUs]. There are some limitations of our study, however. Syringe exchange has been proposed as an effective means of decreasing transmission of other blood-borne pathogens such as HBV and HCV, but we limited our analysis to the effect of SEPs on the transmission of HIV. [A recent study by Hagan et al. ⁽²³⁾ demonstrated that SEP use had no effect on HBV and HCV seroconversion, contradicting earlier research suggesting that SEP use reduces HBV and HCV seroconversion ⁽²⁴⁾.] The estimated number of HIV infections averted was derived by formula rather than biologically measured; it was limited to the number of primary HIV infections averted and did not consider the additional secondary HIV infections that might also be averted. [Based on 1994 CDC data, Lurie and Drucker ⁽²⁵⁾ estimate that every 100 primary HIV infections among IDUs result in 13 secondary infections among these IDUs' sexual partners and children.] Program costs and process measures were self-reported; each SEP participating in the study completed a survey to provide this information, including the amount and valuation of donated or in-kind services provided. The analysis did not consider the costs for drug treatment or other programs to which SEP clients may be referred. Last, any costs incurred by SEP users to participate in programs and any effects on their productivity from a societal perspective have not been considered [the reader is referred to page 46 of the article by Kahn ⁽²²⁾].

Recently, NYS amended its public health law to permit the sale or furnishing of up to 10 hypodermic needles and syringes without a prescription, subject to certain conditions, to individuals who are18 years of age or older by licensed pharmacies and health care facilities and certain health care practitioners. This program, the Expanded Syringe Access Demonstration Program, is authorized for the period January 1, 2001 through March 31, 2003 and is anticipated to reach a larger proportion of the state's estimated IDU population than is being served by the SEPs currently in operation ⁽²⁶⁾. The benefits realized by SEP clients from the additional services that SEPs are required to provide in the context of a comprehensive harm reduction model provide additional support for syringe exchange as a cost-effective HIV prevention strategy.

Although economic evaluations demonstrate the cost-effective and cost-saving aspects of syringe exchange and the epidemiologic evidence that SEP use reduces the risk of HIV transmission, questions have been raised about other effects of SEP operation. For example, to what extent, if any, does SEP participation encourage drug use among clients? Does the presence of SEPs convey a message that condones or encourages drug use, especially among young people? Does the access to sterile needles and syringes provided by SEPs increase the number of discarded and potentially contaminated needles and syringes found in the street? Such questions are outside the scope of this research. Vlahov and Junge ⁽⁶⁾ have reviewed the recent research addressing these and other questions concerning the positive and negative effects of needle exchange and conclude that the research results support the positive outcomes and do not support the negative outcomes: The data on needle exchange in the United States are consistent with the conclusion that these programs do not encourage drug use and that needle exchanges can be effective in reducing HIV incidence (⁶, p. 79).

A consensus development statement recently issued by the National Institutes of Health ⁽²⁷⁾ not only reaches similar conclusions but issues a call for narrowing the gap between what scientific evidence clearly supports and the actions of policy makers to provide the legislative and financial means necessary to promote interventions such as SEPs, which, although controversial in some respects, can serve as essential and effective components of efforts to contain the still-growing AIDS epidemic. Although questions concerning any potential negative effects of HIV prevention interventions such as SEPs deserve standing in the policy debate, scientific findings that answer these questions should be disseminated and considered.

Cost-effectiveness analysis and other forms of economic evaluation are but one source of information intended to inform policy formulation. The results of economic evaluation and other evidence supporting the beneficial effects of SEPs and other HIV prevention strategies need to be weighed, along with evidence concerning the existence of negative effects, against the subjective beliefs or judgments held by a community or its representatives concerning SEP operation. Some states or localities may perhaps be more committed than others to promoting the use of SEPs as an HIV prevention intervention as evidenced by the relatively low number of existing programs and the disparity between the number and distribution of SEPs required to serve IDUs nationally. On the one hand, policy analysts and advocates on both sides of the issue should work to narrow the gap between what scientific evidence suggests and people's beliefs about the consequences (intended and otherwise) of SEPs. On the other hand, those who consider such evidence need to recognize that community values are no less compelling.

Acknowledgments:

This project was supported by a grant from the CDC. The author thanks the following individuals for their contributions to this study: Guthrie Birkhead, Alma Candelas, Donna Glebatis, Mark Hammer, Susan Klein, Len Magnus, Dan O'Connell, and James Tesoriero of the NYS Department of Health's AIDS Institute; Denise Paone of Beth Israel Medical Center's Chemical Dependency Institute; and the representatives of the authorized SEPs. He also thanks the anonymous reviewers for their comments and suggestions.