Paul G Farnham

Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States

Are you Paul G Farnham?

Claim your profile

Publications (29)106.4 Total impact

  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: An alternative HIV testing algorithm, designed to improve the detection of acute and early infections and differentiate between HIV-1 and HIV-2 antibodies, has been developed by the Centers for Disease Control and Prevention and the Association of Public Health Laboratories. While it promises greater sensitivity, it also raises concerns about costs. We sought to compare the most commonly used algorithm which was developed in 1989, a third-generation (3G) immunoassay (IA) and Western blot confirmatory test, to a newer algorithm. The new algorithm includes either a 3G or a fourth-generation (4G) initial IA, followed by confirmatory testing with a HIV-1/HIV-2 differentiation IA and, if needed, a nucleic acid amplification test (NAT). We conducted an analysis of HIV testing costs from the perspective of the laboratory, and classified costs according to IA testing volume. We developed a decision analytic model, populated with cost data from 17 laboratories and published assay performance data, to compare the cost-effectiveness of the testing algorithms for a cohort of 30,000 specimens with a 1% HIV prevalence and 0.1% acute HIV infection prevalence. Costs were lower in high-volume laboratories regardless of testing algorithm. For specimens confirmed positive for HIV antibody, the alternative algorithm (IA, Multispot) was less costly than the current algorithm (IA, WB); however, there was wide variation in reported testing costs. For our cohort, the alternative algorithm initiated with a 3G IA and 4G IA identified 15 and 25 more HIV infections, respectively, than the 1989 algorithm. In medium-volume laboratories, the 1989 algorithm was more costly and less effective than the alternative algorithm with a 3G IA; in high-volume laboratories, the alternative algorithm with 3G IA costs $162 more per infection detected. The alternative algorithm with 4G instead of 3G incurred an additional cost of $14,400 and $4865 in medium- and high-volume labs, respectively. HIV testing costs varied with IA testing volumes. The additional cost of 4G over 3G IA might be justified by the additional cases of HIV detected and transmissions averted due to earlier detection. The alternative HIV testing algorithm compares favorably to the 1989 algorithm in terms of cost and effectiveness.
    Journal of clinical virology: the official publication of the Pan American Society for Clinical Virology 12/2013; 58 Suppl 1:e2-7. · 3.12 Impact Factor
  • Source
    JAIDS Journal of Acquired Immune Deficiency Syndromes 10/2013; 64(2):e15-8. · 4.65 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: BACKGROUND:: Lifetime costs of care and quality of life estimates for human immunodeficiency virus (HIV)-infected persons depend upon the disease stage at which these persons are diagnosed, enter care, and start antiretroviral therapy (ART). Using updated estimates, we analyzed the effect of late versus early diagnosis/entry on U.S. lifetime care costs, quality of life estimates, and HIV transmissions. METHODS:: We used the Progression and Transmission of HIV/AIDS (PATH) model to estimate discounted (3%) lifetime treatment costs ($US 2011) and quality of life variables from time of infection for cohorts of 10,000 HIV-infected index patients in four categories of CD4 count (cells/µL) at diagnosis: (I) ≤ 200; (II) 201 - 350; (III) 351 - 500 and (IV) 501 - 900. We assumed index patient diagnoses were uniformly distributed across the CD4 count range in each category and that patients entered care at the time of diagnosis, remained in care, and were eligible to initiate ART at a CD4 count of 500 cells/µL. We also estimated lifetime transmissions of the index patients. RESULTS:: Discounted average lifetime costs varied from $253,000 for category (I) index patients to $402,000 for category (IV) patients. Discounted quality-adjusted life years lost decreased from 7.95 to 4.45 across these categories, additional years of life expectancy increased from 30.8 to 38.1, and lifetime transmissions decreased from 1.40 to 0.72. CONCLUSIONS:: Early diagnosis and treatment of HIV infection increases lifetime costs but improves length and quality of life and reduces by nearly 50% the number of new infections transmitted.
    JAIDS Journal of Acquired Immune Deficiency Syndromes 04/2013; · 4.65 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: One of the goals of the National HIV/AIDS Strategy (NHAS) is to increase the proportion of HIV-infected individuals linked to care within 3 months of diagnosis (early linkage) from 65% to 85%. Earlier access to care, and eventually, to treatment, increases life expectancy and quality of life for HIV-infected persons. However, longer treatment is also associated with higher costs, especially for antiretroviral drugs. We evaluated the cost effectiveness of achieving the NHAS goal and estimated the maximum cost that HIV programs could spend on linkage to care and remain cost effective. We used the Progression and Transmission of HIV/AIDS model to estimate the effects on life measures and costs associated with increasing early linkage to care from 65% to 85%. We estimated an incremental cost-effectiveness ratio as the additional cost required to reach the target divided by the quality-adjusted life years (QALYs) gained and assumed that programs costing $100,000 or less per QALY gained are cost effective. Achieving the NHAS linkage-to-care goal increased life expectancy by 0.4 years and delayed the onset of AIDS by 1.2 years on average for every HIV-diagnosed person. Increasing early linkage to care cost an extra $62,200 per QALY gained, considering only benefits to index persons. The maximum that could be cost effectively spent on early linkage-to-care interventions was approximately $5100 per HIV-diagnosed person. Considerable investment can be cost effectively made to achieve the NHAS goal on early linkage to care.
    JAIDS Journal of Acquired Immune Deficiency Syndromes 05/2012; 61(1):99-105. · 4.65 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: The Centers for Disease Control and Prevention (CDC), Division of HIV/AIDS Prevention, spends approximately 50% of its $325 million annual human immunodeficiency virus (HIV) prevention funds for HIV-testing services. An accurate estimate of the costs of HIV testing in various settings is essential for efficient allocation of HIV prevention resources. To assess the costs of HIV-testing interventions using different costing methods. We used the microcosting-direct measurement method to assess the costs of HIV-testing interventions in nonclinical settings, and we compared these results with those from 3 other costing methods: microcosting-staff allocation, where the labor cost was derived from the proportion of each staff person's time allocated to HIV testing interventions; gross costing, where the New York State Medicaid payment for HIV testing was used to estimate program costs, and program budget, where the program cost was assumed to be the total funding provided by Centers for Disease Control and Prevention. Total program cost, cost per person tested, and cost per person notified of new HIV diagnosis. The median costs per person notified of a new HIV diagnosis were $12 475, $15 018, $2697, and $20 144 based on microcosting-direct measurement, microcosting-staff allocation, gross costing, and program budget methods, respectively. Compared with the microcosting-direct measurement method, the cost was 78% lower with gross costing, and 20% and 61% higher using the microcosting-staff allocation and program budget methods, respectively. Our analysis showed that HIV-testing program cost estimates vary widely by costing methods. However, the choice of a particular costing method may depend on the research question being addressed. Although program budget and gross-costing methods may be attractive because of their simplicity, only the microcosting-direct measurement method can identify important determinants of the program costs and provide guidance to improve efficiency.
    Journal of public health management and practice: JPHMP 05/2012; 18(3):259-67. · 1.47 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To develop a model to assist clinical setting decision makers in determining how much they can spend on human immunodeficiency virus (HIV) screening and still be cost-effective. The authors developed a simple mathematical model relating the program cost per new HIV diagnosis to the cost per HIV infection averted and the cost per quality-adjusted life year (QALY) saved by screening. They estimated outcomes based on behavioral changes associated with awareness of HIV infection and applied the model to US sexually transmitted disease clinics. The authors based the cost per new HIV diagnosis (2009 US dollars) on the costs of testing and the proportion of persons who tested positive. Infections averted were calculated from the reduction in annual transmission rates between persons aware and unaware of their infections. The authors defined program costs from the sexually transmitted disease clinic perspective and treatment costs and QALYs saved from the societal perspective. They undertook numerous sensitivity analyses to determine the robustness of the base case results. In the base case, the cost per new HIV diagnosis was $2528, the cost per infection averted was $40,516, and the cost per QALY saved was less than zero, or cost-saving. Given the model inputs, the cost per new diagnosis could increase to $22,909 to reach the cost-saving threshold and to $63,053 for the cost-effectiveness threshold. All sensitivity analyses showed that the cost-effectiveness results were consistent for extensive variation in the values of model inputs. HIV screening in a clinical setting is cost-effective for a wide range of testing costs, variations in positivity rates, reductions in HIV transmissions, and variation in the receipt of test results.
    Medical Decision Making 01/2012; 32(3):459-69. · 2.89 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Over a 3-year period, the Centers for Disease Control and Prevention invested $102.3 million in a large-scale HIV testing program, the Expanded HIV Testing Initiative for populations disproportionally affected by HIV. Policy makers, who must optimize public health given a set budget, are interested in the financial return on investment (ROI) of large-scale HIV testing. We conducted an ROI analysis using expenditure and outcome data from the program. A health system perspective was used that included all program expenditures including medical costs of treating newly diagnosed patients. We incorporated benefits of HIV transmissions averted from persons diagnosed of their infection through the Initiative compared with when, on average, those persons would have been diagnosed without the Initiative (3 years later in the base case). HIV transmissions were derived from a published mathematical model of HIV transmission. In sensitivity analysis, we tested the effect of 1-year to 5-year alternate testing intervals and differences in the prevalence of undiagnosed HIV infection. Under the Initiative, 2.7 million persons were tested for HIV, there was a newly diagnosed HIV positivity rate of 0.7%, and an estimated 3381 HIV infections were averted. It achieved a return of $1.95 for every dollar invested. ROI ranged from $1.46 to $2.01 for alternative testing intervals of 1-5 years and remained above $1 (positive return on investment) with a prevalence of undiagnosed HIV infection as low as 0.12%. The expanded testing Initiative yielded ROI values of >$1 under a broad range of sensitivity analyses and provides further support for large-scale HIV testing programs.
    JAIDS Journal of Acquired Immune Deficiency Syndromes 11/2011; 59(3):281-6. · 4.65 Impact Factor
  • JAIDS Journal of Acquired Immune Deficiency Syndromes 10/2011; 58(2):e51-2. · 4.65 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Although previous studies have shown that HIV screening in emergency departments (EDs) is feasible, the costs and outcomes of alternative methods of implementing ED screening have not been examined. We compared the costs and outcomes of a model that used the hospital's ED staff to conduct screening, a supplemental staff model that used non-ED staff hired to conduct screening and a hypothetical hybrid model that combined aspects of both approaches. We developed a decision analytic model to estimate the cost per HIV-infected patient identified using alternative ED testing models. The cost per new HIV infection identified was $3,319, $2,084 and $1,850 under the supplemental, existing staff and hybrid models, respectively. Assuming an annual ED census of 50,000 patients, the existing staff model identified 29 more HIV infections than the supplemental model and the hybrid model identified 76 more infections than the existing staff model. Our findings suggest that a hybrid model should be favored over either a supplemental staff or existing staff model in terms of cost per outcome achieved.
    AIDS education and prevention: official publication of the International Society for AIDS Education 06/2011; 23(3 Suppl):58-69. · 1.51 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Identifying and treating persons with human immunodeficiency virus (HIV) infection early in their disease stage is considered an effective means of reducing the impact of the disease. We compared the cost-effectiveness of HIV screening in three settings, sexually transmitted disease (STD) clinics serving men who have sex with men, hospital emergency departments (EDs), settings where patients are likely to be diagnosed early, and inpatient diagnosis based on clinical manifestations. We developed the Progression and Transmission of HIV/AIDS model, a health state transition model that tracks index patients and their infected partners from HIV infection to death. We used program characteristics for each setting to compare the incremental cost per quality-adjusted life year gained from early versus late diagnosis and treatment. We ran the model for 10,000 index patients for each setting, examining alternative scenarios, excluding and including transmission to partners, and assuming HAART was initiated at a CD4 count of either 350 or 500 cells/µL. Screening in STD clinics and EDs was cost-effective compared with diagnosing inpatients, even when including only the benefits to the index patients. Screening patients in STD clinics, who have less-advanced disease, was cost-effective compared with ED screening when treatment with HAART was initiated at a CD4 count of 500 cells/µL. When the benefits of reduced transmission to partners from early diagnosis were included, screening in settings with less-advanced disease stages was cost-saving compared with screening later in the course of infection. The study was limited by a small number of observations on CD4 count at diagnosis and by including transmission only to first generation partners of the index patients. HIV prevention efforts can be advanced by screening in settings where patients present with less-advanced stages of HIV infection and by initiating treatment with HAART earlier in the course of infection.
    PLoS ONE 01/2011; 6(5):e19936. · 3.53 Impact Factor
  • JAIDS Journal of Acquired Immune Deficiency Syndromes 08/2010; 54(5):565-7. · 4.65 Impact Factor
  • Paul G Farnham
    [Show abstract] [Hide abstract]
    ABSTRACT: In this article we analyse how the costs of treating patients with HIV infection in the US have changed over time, with an emphasis on the relationship between inpatient hospitalization costs and the costs of highly active antiretroviral therapy (HAART). We examine how HIV treatment modes have evolved by comparing the pre-HAART treatment period before 1996-7 with the subsequent use of HAART. We describe the sources of data on HIV healthcare service utilization, the costs of those services, and the differences between the annual costs of treating all patients with different stages of HIV and the lifetime costs of treating a person with HIV from the time of infection. The major question in estimating HIV treatment costs and their components is how to incorporate a complete set of services utilized from all providers of HIV treatment for a representative sample of patients with HIV. The literature reviewed varies significantly on both of these factors. Although the hospitalization of patients with HIV has been declining over the past 2 decades, this rate of decrease accelerated after the introduction of HAART. Initially, the declines in hospitalization and its associated costs were greater than the increases in drug therapy costs, so the annual total costs of treating patients with HIV decreased. However, subsequent studies failed to show decreases in overall annual treatment costs, given rising drug costs and increases in hospitalizations due to complications from, or resistance to, HAART and due to other diseases impacting HIV-infected patients. Although the lifetime costs of treating a person with HIV have also increased, this treatment has resulted in substantial gains in the length and quality of life for those living with HIV.
    Applied Health Economics and Health Policy 01/2010; 8(2):75-88.
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Detection of acute HIV infection (AHI) with pooled nucleic acid amplification testing (NAAT) following HIV testing is feasible. However, cost-effectiveness analyses to guide policy around AHI screening are lacking; particularly after more sensitive third-generation antibody screening and rapid testing. We conducted a cost-effectiveness analysis of pooled NAAT screening that assessed the prevention benefits of identification and notification of persons with AHI and cases averted compared with repeat antibody testing at different intervals. Effectiveness data were derived from a Centers for Disease Control and Prevention AHI study conducted in three settings: municipal sexually transmitted disease (STD) clinics, a community clinic serving a population of men who have sex with men, and HIV counseling and testing sites. Our analysis included a micro-costing study of NAAT and a mathematical model of HIV transmission. Cost-effectiveness ratios are reported as costs per quality-adjusted life year (QALY) gained in US dollars from the societal perspective. Sensitivity analyses were conducted on key variables, including AHI positivity rates, antibody testing frequency, symptomatic detection of AHI, and costs. Pooled NAAT for AHI screening following annual antibody testing had cost-effectiveness ratios exceeding US$200,000 per QALY gained for the municipal STD clinics and HIV counseling and testing sites and was cost saving for the community clinic. Cost-effectiveness ratios increased substantially if the antibody testing interval decreased to every 6 months and decreased to cost-saving if the testing interval increased to every 5 years. NAAT was cost saving in the community clinic in all situations. Results were particularly sensitive to AHI screening yield. Pooled NAAT screening for AHI following negative third-generation antibody or rapid tests is not cost-effective at recommended antibody testing intervals for high-risk persons except in very high-incidence settings.
    PLoS Medicine 01/2010; 7(9):e1000342. · 15.25 Impact Factor
  • AIDS (London, England) 09/2009; 23(13):1792-4. · 4.91 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In 2006, the Centers for Disease Control and Prevention (CDC) recommended routine human immunodeficiency virus (HIV) screening for people aged 13 to 64 years in all U.S. health-care settings. Earlier recommendations focused on those at high risk for HIV and included more extensive pretest counseling. HIV screening may also involve either rapid or conventional testing. The purpose of this research was to estimate the costs of these different testing procedures and the cost per HIV-infected patient correctly receiving test results in three health-care scenarios that illustrated these policy differences. The study estimated the costs of rapid and conventional HIV testing in the following scenarios: (1) sexually transmitted disease (STD) clinic counseling and testing (CT), (2) STD clinic screening, and (3) emergency department (ED) screening. Costs were estimated from the provider perspective in 2006 dollars. A decision analytic model was developed to estimate the cost per HIV-infected patient notified of test results using the two testing procedures in the three scenarios. Although the complete rapid testing procedure was more expensive than conventional testing, the cost per HIV-infected patient receiving test results was lower for the rapid test compared with conventional testing in all scenarios. Per-patient costs of receiving results were lowest in the ED screening scenario and highest in the STD CT scenario. These costs were sensitive to changes in test costs, HIV prevalence, and return rates following conventional tests. HIV screening in general health-care settings is economically feasible, particularly with rapid tests that lower the cost of HIV-infected patients receiving their test results.
    Public Health Reports 01/2008; 123 Suppl 3:51-62. · 1.42 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Assessing the economic burden of HIV/AIDS can help to quantify the effect of the epidemic on a population and assist policy makers in allocating public health resources. To estimate the economic burden of HIV/AIDS in the United States and provide race/ethnicity-specific estimates. We conducted an incidence-based cost-of-illness analysis to estimate the lifetime cost of HIV/AIDS resulting from new infections diagnosed in 2002. Data from the HIV/AIDS Reporting System of the Centers for Disease Control and Prevention were used to determine stage of disease at diagnosis and proportion of cases by race/ethnicity. Lifetime direct medical costs and mortality-related productivity losses were estimated using data on cost, life expectancy, and antiretroviral therapy (ART) use from the literature. The cost of new HIV infections in the United States in 2002 is estimated at $36.4 billion, including $6.7 billion in direct medical costs and $29.7 billion in productivity losses. Direct medical costs per case were highest for whites ($180,900) and lowest for blacks ($160,400). Productivity losses per case were lowest for whites ($661,100) and highest for Hispanics ($838,000). In a sensitivity analysis, universal use of ART and more effective ART regimens decreased the overall cost of illness. Direct medical costs and productivity losses of HIV/AIDS resulting from infections diagnosed in 2002 are substantial. Productivity losses far surpass direct medical costs and are disproportionately borne by minority races/ethnicities. Our analysis underscores economic benefits of more effective ART regimens and universal access to ART.
    JAIDS Journal of Acquired Immune Deficiency Syndromes 01/2007; 43(4):451-7. · 4.65 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Alternatives to conventional HIV counseling and testing (HIV-CT) have been used to improve receipt of HIV test results. To determine the effectiveness of alternative HIV-CT methods on the receipt of HIV test results. Studies were identified by a systematic search of the literature using English-language databases from 1990 to 2005. Studies were included if they used an alternative method for HIV-CT, reported the receipt of HIV test results and had a comparison group. Pooled effect sizes [risk ratios (RR)] were calculated using a random effects model. Seventeen effect sizes (k) were included n = 21 096). Alternative HIV-CT methods included rapid testing (k = 12), oral fluid testing (k = 2), home testing (k = 1), and telephone post-test counseling (k = 2). All alternatives except for oral fluid testing significantly increased receipt of results compared with conventional testing. In stratified analysis, rapid testing was most effective [RR, 1.80; 95% confidence interval (CI), 1.46-2.22] followed by telephone post-test counseling (RR, 1.38. 95% CI, 1.24-1.47). There is strong evidence that clients are substantially more likely to receive their HIV test results with rapid testing than with conventional tests or other alternatives. Therefore, to increase knowledge of HIV status, rapid testing is preferable in settings with low rates of return for test results.
    AIDS 09/2006; 20(12):1597-604. · 6.41 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This study examined changes in healthcare use among perinatally HIV-infected children and developed new estimates of expected lifetime treatment costs. The study analyzed longitudinal medical record data from the Pediatric Spectrum of Disease study on perinatally HIV-infected children enrolled in 6 US sites during 1995 and 2001 for enrollee characteristics including healthcare utilization. For the year 2001, costs were assigned to hospitalization, HIV-related drug usage, and laboratory testing. To estimate lifetime treatment costs based on those categories, median survival times of 9, 15, and 25 years were assumed and average annual healthcare utilization costs were applied to each year of survival. From 1995 to 2001, hospitalization rates fell from 0.67 per child-year to 0.23 per child-year (P < 0.05). In 2001, the average cost of healthcare utilization per child was $12,663, including $2164 for hospitalization, $9505 for HIV-related drugs, and $994 for laboratory tests. The discounted lifetime treatment cost, based on those 3 cost categories, was $113,476 for 9 years of survival, $151,849 for 15 years, and $228,155 for 25 years. Hospitalizations among perinatally HIV-infected children decreased significantly from 1995 to 2001. Compared with previously published estimates, lifetime treatment costs for children perinatally infected with HIV have remained relatively stable. However, as years of survival increase for this population, lifetime costs also are likely to increase.
    JAIDS Journal of Acquired Immune Deficiency Syndromes 05/2006; 41(4):521-6. · 4.65 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: To estimate the incremental societal costs and effectiveness of a second human immunodeficiency virus (HIV) antibody test during the third trimester of pregnancy compared with no second test. We used a decision tree in this cost-effectiveness analysis to model outcomes among pregnant women in high-risk communities and nationwide who received an initial, negative HIV antibody test during the first trimester. The main outcome measure was discounted costs per year of infant life saved. In high-risk communities with estimated HIV incidence of 6.2 per 1000 person-years, a second HIV test compared with no second test would detect 192 infections in women, prevent approximately 37 infant infections, and save 655 infant life-years per 100,000 women tested. Net savings would be 5.2 million US dollars. Applied to an estimated national incidence of.17 per 1000 person-years, a second test would detect 5.3 infections in women, prevent 1.3 infant infections, and save 23.3 infant life-years per 100,000 women tested. Net costs would be 1.06 million US dollars, or 45,708 US dollars for each year of infant life saved. A second test would result in net savings in populations with HIV incidence of 1.2 per 1000 person-years or higher. Health care providers serving women in communities with an HIV incidence of 1 per 1000 person-years or higher should strongly consider implementing a second voluntary universal HIV test during the third trimester. Providers serving lower-risk communities should pilot second testing to assess community-specific costs.
    Obstetrics and Gynecology 11/2003; 102(4):782-90. · 4.80 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Because resources to fund HIV prevention are limited, public health decision makers—such as health departments and HIV prevention community planning groups—need to know which prevention strategies are the most cost-effective. In the past several years, a number of studies have appeared in the literature that assess the cost-effectiveness of interventions to prevent the sexual transmission of HIV in the United States. Here, we comprehensively review 16 such studies and then outline an agenda for further research to advance the cost-effectiveness literature and to make the findings of these studies more useful for public health decision makers. The research summarized here provides compelling evidence that interventions to prevent sexual transmission of HIV can be highly cost-effective. Small-group, community-level, and outreach-based sexual risk reduction interventions, in particular, appear to be very efficient strategies for preventing the spread of HIV in moderate- to high-risk populations.
    AIDS and Behavior 01/2002; 6(1). · 3.49 Impact Factor

Publication Stats

448 Citations
106.40 Total Impact Points

Institutions

  • 2013
    • Johns Hopkins Bloomberg School of Public Health
      Baltimore, Maryland, United States
  • 2006–2013
    • Centers for Disease Control and Prevention
      • • Division of HIV/AIDS Prevention, Intervention and Support
      • • Division of Global HIV/AIDS
      Atlanta, Michigan, United States
  • 2002–2003
    • Georgia State University
      • Department of Economics
      Atlanta, Georgia, United States
    • Emory University
      • Centers for Disease Control and Prevention
      Atlanta, GA, United States
  • 2001
    • Medical College of Wisconsin
      • Center for AIDS Intervention Research
      Milwaukee, WI, United States