| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
RESEARCH AND PRACTICE |
Benjamin P. Linas, Hui Zheng, Rochelle P. Walensky, and Kenneth A. Freedberg are with the Division of General Medicine, Massachusetts General Hospital, Boston, Mass. Benjamin P. Linas, Rochelle P. Walensky, and Kenneth A. Freedberg are also with the Division of Infectious Diseases, Massachusetts General Hospital. Benjamin P. Linas, Hui Zheng, Rochelle P. Walensky, and Kenneth A. Freedberg are also with the Harvard Center for AIDS Research, Harvard Medical School, Boston, Mass. Kenneth A. Freedberg and Elena Losina are with the departments of Biostatistics and Epidemiology, Boston University School of Public Health, Boston, Mass.
Correspondence: Requests for reprints should be sent to Benjamin P. Linas, MD, Massachusetts General Hospital, 50 Staniford St, Ninth floor, Boston, MA, 02114 (e-mail: blinas{at}partners.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
Objectives. The United States allocates more than $900 million annually for the prevention of HIV infection. We assessed the impact of this funding on HIV testing and knowledge.
Methods. We linked data from the Behavioral Risk Factor Surveillance System with tracking of Centers for Disease Control and Prevention (CDC) HIV prevention funding. We developed and validated regression models of the relation between HIV prevention funding to a respondent’s state and the odds that the respondent (1) had been tested for HIV, and (2) was aware of methods to prevent mother-to-child HIV transmission (MTCT).
Results. The odds of having been tested for HIV increased with increased CDC funding to states (P=.009), as did awareness of prevention of MTCT (P=.002). We estimate that CDC HIV prevention funds led to 12.8 million more people being tested for HIV between 1998 and 2003 than would have been tested had all states received funds equal to the lowest quintile of funding.
Conclusions. Federal HIV prevention funds independently correlate with increased HIV testing and knowledge of prevention of MTCT. Proposed reductions in HIV prevention spending would likely have adverse public health consequences.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
As part of this prevention effort, the CDC seeks to reduce barriers to early diagnosis of HIV infection.5 One of the major goals of the CDC HIV prevention initiative is increasing the proportion of HIV-infected individuals in the United States who are aware of their HIV infection. A second objective is increasing the proportion of HIV-infected pregnant women who choose to take antiretroviral medications to prevent mother-to-child HIV transmission (MTCT).3
In fiscal year 2005, the US government spent approximately $900 million on HIV prevention across 9 federal agencies.6 The CDC is the nation’s lead federal agency for HIV prevention, receiving nearly 90% of total HIV prevention funds.6,7 To accomplish its prevention goals, the CDC funds state and local departments of health, as well as community-based organizations, to provide education and prevention services at the local level.3
The CDC places great emphasis on program evaluation, and several publications demonstrate the effectiveness of individual CDC-funded programs.8,9 There have been no investigations, however, of the relation between HIV prevention funds and success at achieving stated HIV prevention goals at the national level. We therefore sought to assess the impact of CDC HIV prevention funds on HIV test utilization and knowledge about prevention of MTCT.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
The study cohort was all BRFSS respondents aged 18 to 54 years living in the 50 United States.
Data
We used state of residence to link data from 3 sources: the CDC BRFSS, NASTAD, and CDC HIV/AIDS surveillance reports.
CDC BRFSS. The BRFSS is a scientifically sampled annual telephone survey of the US population that monitors state-level prevalence of the major behavioral risks associated with premature morbidity and mortality among adults.10
The BRFSS asks respondents, "As far as you know, have you ever been tested for HIV? Do not count tests that you may have had as part of blood donation." Respondents who answer yes provide the month and year of their most recent HIV test. BRFSS also asks, "True or false? A pregnant woman can get treatment to help reduce the chances that she will pass the virus on to her baby."11 BRFSS also provides demographic data including the respondent’s age, gender, race, marital status, employment status, educational level, and income. The 2003 BRFSS included an additional question about HIV risk behaviors.12
NASTAD tracking of CDC HIV prevention funds. NASTAD documents CDC HIV prevention funding to state and local departments of health in a publicly available data set (data available from author). These NASTAD data provided the total amount of CDC HIV prevention funds granted to each state department of health in each year from 1996 through 2003.
CDC HIV/AIDS surveillance reports. We used HIV/AIDS surveillance data to control for correlation between AIDS prevalence and HIV testing. The quintile of AIDS prevalence was included as a categorical variable in logistic regression models.
Analysis
We conducted analyses with SAS version 9.113 and used weights provided by the BRFSS to adjust for clustering biases introduced by the BRFSS sampling methodology.
Development phase.
We chose mean annual per capita CDC HIV prevention funds granted during the preceding 5 years as the measure of CDC HIV prevention funding to states. We adjusted funding to year 2000 dollars and controlled for regional differences in inflation with the region-specific Consumer Price Index for Urban Consumers.14 We ranked states by the amount of CDC HIV prevention funding that they received during the preceding 5 years (1996–2001 in the 2001 data set, 1998–2003 in the 2003 data set), and grouped BRFSS responses into quintiles on the basis of CDC prevention funding to the respondent’s state. We observed the percentage of respondents tested for HIV during the preceding 5 years and the percentage aware of the prevention of MTCT by quintile of CDC prevention funding, using the Mantel-Haenszel 
2 test for linear trends.
To adjust for possible confounders, we then constructed logistic regression models of the likelihood that an individual had been tested for HIV during the preceding 5 years and the likelihood that an individual was aware of the prevention of MTCT. We first included the quintile of CDC funding as a categorical variable and then tested for linear trends among quintiles by including the CDC funding quintile as an ordinal variable and observing the effect estimate. With demographic information from BRFSS and linked AIDS prevalence data, the regression models controlled for the effects of age, race, gender, income, educational level, marital status, employment status, and state AIDS prevalence.
Validation phase. To validate the relation, we repeated the analyses with the 2003 BRFSS data set. Because the 2003 BRFSS asked respondents if they had ever engaged in unprotected anal sex, injection drug use, or sex work, the 2003 models controlled for high-risk HIV behavior in addition to the demographic variables included in the development models. We report the impact of CDC funding on HIV testing and knowledge as measured by the validated logistic regression models.
Additional validity testing. To further test the validity of the relation between CDC funding and HIV testing and knowledge, we performed additional analyses examining logical extensions of the argument that CDC HIV prevention funds increase the likelihood that an individual is tested for HIV and is aware of prevention of MTCT. We first examined the temporal nature of the relation. Logically, HIV prevention funds in fiscal year 2003 should not predict testing behaviors between 1996 and 2001 or awareness of prevention of MTCT in 2001. We therefore performed logistic regression of the likelihood of having been tested for HIV between 1996 and 2001 and the likelihood of being aware of the prevention of MTCT in 2001 with fiscal year 2003 HIV prevention funds as the primary predictor.
Similarly, we reasoned that if CDC funds impact HIV testing rates and knowledge, states with large increases in CDC funds should enjoy large increases in testing and knowledge. We therefore changed the unit of analysis to the state, using BRFSS data to calculate the change between 2001 and 2003 in the percentage of people in each state who had been tested for HIV in the preceding 5 years and who were aware of the prevention of MTCT. We then calculated the change in CDC prevention funding to each state between fiscal year 1998 and fiscal year 2003 and used linear regression to assess the relation between changes in CDC funding and changes in HIV testing rates and awareness of prevention of MTCT at the state level.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
. Between 1996 and 2001, median total CDC prevention funding to each state (adjusted to year 2000 dollars) was $18104683 (range $2969695 to $268384235). Median annual CDC prevention funding per capita was $1.59 (range $0.69 to $4.41).
|
Table 2 presents logistic regression models of the likelihood that an individual had been tested for HIV during the preceding 5 years. After control for all potential confounders, the odds ratio of having been tested for HIV for individuals living in states in the second through fifth quintiles of CDC funding, compared with those living in first quintile states, ranged from 1.02 to 1.21 (Table 2, model 3).
|
. The median annual per capita CDC funds granted to states between 1998 and 2003 were $1.56 (range $0.74 to $4.54) (Figure 1).
|
Multivariate analysis confirmed the relation established in the development data set. The odds ratio that an individual had been tested for HIV between 1998 and 2003 increased as a function CDC funds granted to the respondent’s state. Compared with respondents living in first-quintile states, those living in second- through fifth-quintile states were 8% to 14% more likely to have been tested for HIV (Table 2, model 3). When we entered the quintile of CDC funding into the model as an ordinal variable, we found a statistically significant upward, linear trend in the likelihood that an individual had been tested for HIV (P= .009 for trend).
Using the odds ratios generated by the model to approximate the relative risk of having been tested for HIV, we estimated that CDC HIV prevention funds led to approximately 12.8 million more people being tested for HIV between 1998 and 2003 than would have been tested had every state received funds equivalent to those received by states in the lowest quintile of funding (95% confidence interval [CI]=4.5 million, 21.8 million).
Additional Validity Testing
We found no association between CDC HIV prevention funds in fiscal year 2003 and the likelihood that an individual had been tested for HIV between 1996 and 2001. The odds of being tested for HIV between 1996 and 2001 for those living in states in the second through fifth quintiles of fiscal year 2003 CDC HIV prevention funding compared with those living in first-quintile states were 0.98 (95% CI = 0.92, 1.03), 0.97 (95% CI = 0.91, 1.03), 1.07 (95% CI = 0.99, 1.14), and 1.03 (95% CI = 0.96, 1.11), respectively.
Changes in CDC HIV prevention funds to a state did correlate with changes in the percentage of people in that state who had been tested for HIV in the preceding 5 years. In linear regression analysis, an increase of $1 per capita in CDC HIV prevention funds granted to a state between fiscal year 1998 and fiscal year 2003 correlated with an 8% increase in HIV testing in that state between 2001 and 2003 (95% CI = 0.3%, 16%).
Awareness of the Prevention of MTCT
Development phase.
Fifty-four percent of those in the 2001 data set reported that they were aware of the prevention of MTCT (range by state 41% to 64%). In univariate analysis, 50% of those living in states in the lowest quintile of CDC HIV prevention funds was aware of the prevention of MTCT, compared with 50%, 53%, 57%, and 58% in the second through fifth quintiles, respectively (P< .0001 for trend).
In multivariate analysis, the percentage aware of the prevention of MTCT increased as a function of mean annual per capita CDC prevention funds. The odds ratio of being aware of the prevention of MTCT for those living in second- through fifth- quintile states, compared with those in first-quintile states, ranged from 1.01 to 1.19 (Table 3, model 3).
|
After adjustment for possible confounders, the odds ratio of being aware of the prevention of MTCT for those living in states in the second through fifth quintiles of CDC funding compared with those in first-quintile states ranged from 0.95 to 1.14 (Table 3, model 3). When we entered the quintile of CDC funding into the model as an ordinal variable, we found a significant upward linear trend in the likelihood that an individual was aware of the prevention of MTCT (P= .002 for trend).
Additional validity testing. We found no association between CDC HIV prevention funds in fiscal year 2003 and awareness of prevention of MTCT in 2001. The odds of being aware of the prevention of MTCT in 2001 for those living in states in the second through fifth quintiles of fiscal year 2003 CDC HIV prevention funding compared with those living in first-quintile states were 0.87 (95% CI = 0.83, 0.92), 0.87 (95% CI = 0.82, 0.92), 0.86 (95% CI = 0.81, 0.92), and 0.95 (95% CI = 0.89, 1.02), respectively.
Changes in CDC HIV prevention funds to a state do not correlate with changes in awareness of the prevention of MTCT. In linear regression analysis, an increase of $1 per capita in CDC HIV prevention funds granted to a state during the preceding 5 years correlated with a 2% increase in awareness of the prevention of MTCT in that state between 2001 and 2003 (95% CI = –4.2%, 8.3%).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
That federal funding has an effect on individual behavior and knowledge likely reflects planning by CDC and effective resource allocation by states. It remains possible that CDC funds could be still more efficiently allocated. Indeed, recent reports suggest that the CDC’s emphasis on HIV testing and counseling for the general population may not be the most cost-effective approach to HIV prevention.15 Our results, however, suggest that current planning and implementation have an impact. The finding is also consistent with the CDC’s emphasis on evaluating individual program effectiveness.8,9 Although the CDC pays close attention to the effectiveness of grantee programs, there have been few attempts to quantify the impact of CDC HIV prevention efforts at the macro level. Holtgrave16 has used mathematical modeling to estimate that HIV prevention activities in the United States between 1978 and 2000 averted between 204 000 and 1.6 million cases of HIV. Chesson et al.17 found that CDC funding for HIV and STD prevention reduced state gonorrhea rates. Our findings add to this growing body of evidence that CDC prevention funding has a measurable impact in reaching prevention goals.
Although there is a trend for increased prevention funding to correlate with increased HIV testing, the effect of funding appears to flatten or perhaps decrease when states in the fourth and fifth quintiles of CDC funding are compared. The highest quintile of CDC prevention funding in the 2003 data set included Connecticut, New Jersey, Maryland, Delaware, New York, and Rhode Island, the 6 states with the highest proportion of cumulative AIDS cases attributed to illegal injection drug use.18 This finding is consistent with the CDC’s priority of providing services to vulnerable communities.3 Injection drug use itself does not make an individual less likely to be tested for HIV. In fact, in our regression models, injection drug users were more than twice as likely as others to have been tested for HIV in the preceding 5 years. Having a high prevalence of HIV related to injection drug use, however, may be a proxy measure for a state having an HIV epidemic located in poor, urban communities, locations where underdeveloped social infrastructure make HIV prevention work difficult.19,20 These are states that might have had low rates of HIV testing in the absence of CDC-funded prevention campaigns. If the CDC targets funds specifically to states with difficult HIV prevention campaigns, and these states are also those that would have had the lowest testing rates in the absence of prevention funds, the effect of CDC prevention dollars will appear muted compared with the effect in states with less funding, but less complicated prevention efforts. Were it possible to measure the difference between current HIV testing rates and what the rates would have been in the absence of CDC HIV prevention funds, the change in testing rates might be higher in fifth-quintile states than in fourth-quintile states, even though the absolute percentage tested was the same.
Alternately, it is possible that testing rates in the fourth and fifth quintiles are similar because of decreasing marginal returns of HIV prevention funding. If so, it is possible that redistributing funds from the highest-quintile states to first- through third-quintile states might allow more states to receive the optimal level of funding. Distinguishing between these competing explanations of the observed trend has resource allocation implications, especially given recent reports of the importance of cost-effectiveness analyses in directing the distribution of limited government resources.15 Future research efforts should concentrate on the relative efficacy of CDC HIV prevention dollars at the highest levels of funding.
There are limitations to this analysis. Because there is no mandatory reporting of HIV incidence, we were not able to measure the outcome of ultimate interest for an HIV prevention program—decreased incidence of HIV infection. HIV testing itself, however, is a meaningful endpoint, as early detection has been shown to lead to better HIV outcomes,21 and some research suggests that knowledge of positive HIV serostatus correlates with decreased high-risk behavior.22–24 Above all, increased HIV testing is 1 of the expressed goals of the CDC HIV prevention strategy.5 What is more, because 44 of the 50 states currently have voluntary (as opposed to "opt-out") HIV testing for pregnant women,25 improving awareness is an important intermediate step toward improving the prevention of MTCT.
Because this is an ecological analysis, we cannot definitively establish the causality of the observed relation. However, we have taken steps to rigorously test our findings by validating the models in an independent data set, testing the temporal relation between funding and outcomes, and altering the unit of analysis to assess relations between changes in funding and changes in outcomes. Absent a prospective policy experiment in which states are randomly assigned to alternative prevention funding approaches, it is impossible to assess the impact of CDC HIV prevention funds without using an ecological approach.
This analysis focuses on CDC funds that are granted to state and local departments of public health. Although CDC allocates approximately two thirds of its external resources to departments of health, one third go directly to local community-based organizations and are not included in the analysis.26 In addition, many states also apply nonfederal funds to HIV prevention efforts. Tracking reliable estimates of state and local HIV prevention budgets over several years is not possible, however, because state and local HIV prevention spending often overlaps with HIV care provision and social services. Unpublished reports from 40 states to the CDC in 2000 indicated that federal funding accounted for approximately 60% of total HIV prevention spending in those states (range by state 25% to 100%).17 When one considers that some portion of these "prevention" funds was likely applied jointly to AIDS care and service provision, the proportion of total prevention spending attributable to the federal government is likely greater than 60%.
The finding that federal HIV prevention funds have a substantial impact on HIV testing and awareness of prevention of MTCT is important for setting budget priorities. The president’s fiscal year 2006 budget proposal decreases spending for HIV education and prevention by $4 million.6 This analysis provides evidence that there would likely be adverse consequences to reducing federal HIV prevention funding. Federal HIV prevention funds have a measurable impact on prevention goals. Reducing prevention funding may diminish that impact and could lead to lower HIV testing rates and decreased awareness of the prevention of MTCT.
|
Acknowledgments |
|---|
Human Participant Protection
This study was exempted from institutional review board review by the Partners Health Care human research office.
|
Footnotes |
|---|
Contributors
B.P. Linas originated the study question and design, managed the data set, conducted statistical analyses, and led the writing. H. Zheng provided biostatistical and computer programming support. E. Losina contributed to the study design and supervised biostatistical methods. R.P. Walensky assisted with the study design and data interpretation. K.A. Freedberg contributed to developing the study questions and guided data interpretation. All authors reviewed each draft of the article.
Accepted for publication November 14, 2005.
|
References |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONReferences |
|---|
2. A Glance at the HIV Epidemic. Atlanta, Ga: Centers for Disease Control and Prevention; 2001.
3. HIV Prevention Strategic Plan Through 2005. Atlanta, Ga: Centers for Disease Control and Prevention; 2001.
4. Janssen RS, Holtgrave DR, Valdiserri RO, Shepherd M, Gayle HD, De Cock KM. The serostatus approach to fighting the HIV epidemic: prevention strategies for infected individuals. Am J Public Health. 2001;91: 1019–1024.[Abstract]
5. Advancing HIV prevention: new strategies for a changing epidemic—United States, 2003. MMWR Morb Mortal Wkly Rep. 2003;52:329–332.[Medline]
6. Kates J, Leggoe AW. US Federal Funding for HIV/AIDS: The FY2006 Budget Request. Menlo Park, Calif: Henry J. Kaiser Family Foundation; 2005.
7. Summers T, Alagiri P, Kates J. Federal HIV/AIDS Spending: A Budget Chartbook Fiscal Year 2002. Menlo Park, Calif: Henry J. Kaiser Family Foundation; 2005.
8. Evaluation Guidance Handbook: Strategies for Implementing the Evaluation Guidance for CDC-Funded HIV Prevention Programs. Atlanta, Ga: Centers for Disease Control and Prevention; 2001.
9. Prevention Research Synthesis Project. Compendium of HIV Prevention Interventions With Evidence of Effectiveness. Atlanta, Ga: Centers for Disease Control and Prevention; 2001.
10. Behavioral risk factor surveillance system Web site. Centers for Disease Control and Prevention, 2003. Available at: http://www.cdc.gov/brfss/. Accessed February 27, 2006.
11. 2001 Behavioral Risk Factor Surveillance System questionnaire. Centers For Disease Control and Prevention, 2001. Available at: http://www.cdc.gov/brfss/questionnaires/pdf-ques/2001brfss.pdf. Accessed February 27, 2006.
12. 2003 Behavioral Risk Factor Surveillance System questionnaire. Centers for Disease Control and Prevention, 2003. Available at: http://www.cdc.gov/brfss/questionnaires/pdf-ques/2003brfss.pdf. Accessed February 27, 2006.
13. SAS version 9.1 [computer program]. Cary, NC: SAS Institute Inc; 2002.
14. US Department of Labor Bureau of Labor Statistics. Consumer Price Index detailed statistics. 2005. Available at: http://data.bls.gov. Accessed February 27, 2006.
15. Cohen D, Wu S, Farley T. Cost-effective allocation of government funds to prevent HIV infection. Health Aff (Millwood). 2005;24:915–926.
16. Holtgrave DR. Estimating the effectiveness and efficiency of US HIV prevention efforts using scenario and cost-effectiveness analysis. AIDS. 2002;16:2347–2349.[CrossRef][ISI][Medline]
17. Chesson HW, Harrison P, Scotton CR, Varghese B. Does funding for HIV and sexually transmitted disease prevention matter? Evidence from panel data. Eval Rev. 2005;29:3–23.[Abstract]
18. AIDS Public Information Dataset, 2001. Centers for Disease Control and Prevention, 2001. Available at http://www.statehealthfacts.kff.org/cgi-bin/healthfacts.cgi?action=compare&category=HIV%2fAIDS&subcategory=Cumulative+AIDS+Cases&topic=Adult%2fAdolescent+by+Exposure+Category&link_category=&link_subcategory=&link_topic=&datatype=&printerfriendly=0&viewas=&showregions=0&sortby=Injection+Drug+Use#sorttop. Accessed February 27, 2006.
19. Latkin CA, Williams CT, Wang J, Curry AD. Neighborhood social disorder as a determinant of drug injection behaviors: a structural equation modeling approach. Health Psychol. 2005;24:96–100.[CrossRef][ISI][Medline]
20. Latkin CA, Forman V, Knowlton A, Sherman S. Norms, social networks, and HIV-related risk behaviors among urban disadvantaged drug users. Soc Sci Med. 2003;56:465–476.[CrossRef][ISI][Medline]
21. Enger C, Graham N, Peng Y, et al. Survival from early, intermediate, and late stages of HIV infection. JAMA. 1996;275:1329–1334.[Abstract]
22. Higgins DL, Galavotti C, O’Reilly KR, et al. Evidence for the effects of HIV antibody counseling and testing on risk behaviors. JAMA. 1991;266:2419–2429.[Abstract]
23. Weinhardt LS, Carey MP, Johnson BT, Bickham NL. Effects of HIV counseling and testing on sexual risk behavior: a meta-analytic review of published research, 1985–1997. Am J Public Health. 1999;89:1397–1405.
24. Wolitski RJ, MacGowan RJ, Higgins DL, Jorgensen CM. The effects of HIV counseling and testing on risk-related practices and help-seeking behavior. AIDS Educ Prev. 1997;9(3 suppl):52–67.[ISI][Medline]
25. Henry J. Kaiser Family Foundation. HIV Testing for Mothers and Newborns 2004. Available at: http://www.statehealthfacts.org/cgi-bin/healthfacts.cgi?action=compare&category=HIV%2fAIDS&subcategory=HIV+Testing&topic=HIV+Testing+for+Mothers+and+Newborns. Accessed February 27, 2006.
26. Institute of Medicine. No Time to Lose: Getting More from HIV Prevention. Washington, DC: National Academy Press; 2001.
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
