Advertisement

Comparative Cost Analysis of Clinical Reminder for HIV Testing at the Veterans Affairs Healthcare System

  • Kee Chan
    Correspondence
    Address correspondence to: Kee Chan, College of Health and Rehabilitation Sciences: Sargent College, Boston University, 635 Commonwealth Avenue, Boston, MA 02215.
    Affiliations
    Department of Health Sciences, College of Health and Rehabilitation Sciences: Sargent College, Boston University, Boston, MA, USA

    Center of Health, Quality, and Economics Research, Veterans Affairs, Bedford, MA, USA
    Search for articles by this author
  • Leilani Hernandez
    Affiliations
    Department of Health Sciences, College of Health and Rehabilitation Sciences: Sargent College, Boston University, Boston, MA, USA
    Search for articles by this author
  • Heidi Yang
    Affiliations
    Department of Health Sciences, College of Health and Rehabilitation Sciences: Sargent College, Boston University, Boston, MA, USA
    Search for articles by this author
  • Matthew Bidwell Goetz
    Affiliations
    Infectious Diseases Section, Department of Medicine, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA

    Infectious Diseases Section, Department of Medicine, VA Greater Los Angeles Healthcare System, Los Angeles, CA, USA
    Search for articles by this author
Open ArchivePublished:May 17, 2014DOI:https://doi.org/10.1016/j.jval.2014.03.001

      Abstract

      Objective

      To estimate the cost and health outcomes associated with a new HIV testing strategy that utilizes routine-based clinical reminders.

      Methods

      We conducted an economic analysis of 1) traditional pretest/post-test counseling; 2) counseling and a new clinical reminders system; and 3) only clinical reminder in the veterans’ health care system. A payer-perspective decision model was conducted to calculate the 1-year budget impact of three HIV testing strategies. Parameter values were obtained from the literature, including patients’ probability of accepting test, and costs associated with HIV testing procedures. Deidentified patient data, including total population screened and number of new HIV cases, were collected from one clinic in Los Angeles, California, from August 2004 to December 2011. Annual total costs and costs per new case were calculated on the basis of parameter values and patient data. Sensitivity analyses were conducted to evaluate the robustness of the critical variable on costs.

      Results

      The total cost of the clinical reminder system with pretest counseling was $81,726 over 1 year compared with $109,208 for traditional HIV testing. Under a clinical reminder system with no pretest counseling, the number of HIV tests performed and the number of new diagnoses increased for that year. In addition, cost per new diagnoses was the lowest.

      Conclusions

      The clinical reminder system can reduce the cost per cases identified and promote better performance of HIV testing compared with traditional HIV testing. The fundamental decision model can be used for hospital facilities outside the Veteran Affairs adopting a similar program for improving the HIV testing rate.

      Keywords

      Introduction

      Early identification of HIV infection provides clinical benefit to the infected individuals and reduces the risk of disease transmission. The Centers for Disease Control and Prevention estimates that approximately 19% of the US population is unaware of its HIV status. This gap has led the Centers for Disease Control and Prevention, the American College of Physicians, and US Preventive Services Task Force to recommend that routine, voluntary HIV testing be offered to adults [
      • Branson B.M.
      • Handsfield H.H.
      • Lampe M.A.
      • et al.
      Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings.
      ,
      • Qaseem A.
      • Snow V.
      • Shekelle P.
      • et al.
      Screening for HIV in health care settings: a guidance statement from the American College of Physicians and HIV Medicine Association.
      ,
      • Moyer V.A.
      U.S. Preventive Services Task Force*. Screening for HIV: U.S. Preventive Services Task Force Recommendation Statement.
      ]. In response, as the single largest health care system in the United States, the Veteran Health Administration (VHA) has implemented a series of programs to increase early identification by promoting HIV testing [

      Testing for human immunodeficiency virus in Veterans Health Administration facilities. Available from: http://www1.va.gov/vhapublications/ViewPublication.asp?pub_ID=2056. [Accessed August 20, 2009].

      ].
      Over time, in responses to changes in the regulatory environment (i.e., removal of requirements for written informed consent) and modifications of recommendations as to who should be offered HIV testing (i.e., transition from risk-based to routine testing), different interventions have been undertaken in VHA to improve HIV testing rates. Although these programs have been successful as indicated in our previous study, the total costs of these programs and the cost per identified case are incompletely described [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ].
      This study performed the cost analysis for three alternative strategies for HIV testing: 1) physician-based traditional HIV testing and counseling in the absence of clinical reminders; 2) clinical reminders and nurse-based streamlined counseling with telephone notifications for negative results; and 3) clinical reminders without pretest counseling and with telephone notifications for negative results. To assist programs that may be interested in adopting a similar strategy but are uncertain of the cost implication, we have evaluated the cost per test and the cost for identifying a previously undiagnosed case of HIV infection. Because this article focused on evaluating the immediate cost implication of these new strategies, we did not include the long-term cost-effectiveness of HIV testing. These analyses will consider the effects of diagnosis and treatment on quality-adjusted life-year, which is beyond the scope of this study.

      Methods

       Background

      When written informed consent was required for HIV testing, nurse-initiated, streamlined counseling was found to be cost-effective in increasing HIV testing rates in primary care settings [
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
      ]. In August 2005, Goetz et al. [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ] used this strategy as part of a multimodal intervention that included a real-time clinical reminder to prompt providers to offer risk-based HIV testing, provider education, and an audit-feedback program to increase HIV testing rates in VHA medical care facilities in Southern California. The clinical reminder was triggered by any previous evidence of hepatitis B or C infection, illicit drug use, a sexually transmitted disease, homelessness, and certain behavioral risk factors in the patient’s medical record. Implementation of this program tripled the screening rate and led to more HIV diagnoses [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ]. In August 2009, the VHA policy for HIV testing was changed. The revised policy removed requirements for formal pretest and post-test counseling. Meanwhile, verbal consent was substituted for written consent for testing and testing was recommended for all persons regardless of known risk of HIV infection [

      Testing for human immunodeficiency virus in Veterans Health Administration facilities. Available from: http://www1.va.gov/vhapublications/ViewPublication.asp?pub_ID=2056. [Accessed August 20, 2009].

      ]. Following this change, many Veterans Affairs (VA) facilities changed their policies to be consistent with new VA requirements and implemented a non–risk-based clinical reminder to promote HIV testing. Both the original risk-based interventions and the subsequent modification to offer HIV testing to all previously untested patients have been shown to improve HIV testing rates [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ,
      • Goetz M.B.
      • Hoang T.
      • Knapp H.
      • et al.
      Central implementation strategies outperform local ones in improving HIV testing in Veterans Healthcare Administration facilities.
      ].

       Hypothesis and Study Design

      This study intended to test the hypothesis that the implementation of a non–risk-based clinical reminder system for promoting HIV testing is more cost-effective than traditional risk-based counseling. The study tested this hypothesis by estimating the comparative costs of HIV testing strategies in three different scenarios (Fig. 1):
      • 1.
        Strategy A: Risk-based testing with reliance on physician’s recognition of at-risk patients and physician’s responsibility for test ordering, and requirements for written informed consent and in-person pretest and post-test counseling.
      • 2.
        Strategy B: Risk-based testing with reliance on clinical reminders to identify at-risk patients, physician’s responsibility for test ordering, nurse-based streamlined in-person pretest counseling, requirements for written informed consent, and telephone notifications for negative results.
      • 3.
        Strategy C: Routine HIV testing with reliance on clinical reminders to identify previously untested patients, elimination of pretest counseling, substitution of verbal for written informed consent, physician’s responsibility for test ordering, and telephone notification for negative results.
      Figure thumbnail gr1
      Fig. 1Key procedural steps in the three strategies are evaluated. The key difference between the strategies are circled. MD, doctor of medicine; RN, registered nurse.
      The analysis included the costs of the screening and the personnel involved in the implementation of HIV testing and excluded the costs of caring for persons found to be HIV-infected [
      • Anaya H.D.
      • Chan K.
      • Karmarkar U.
      • et al.
      Budget impact analysis of HIV testing in the VA healthcare system.
      ]. Figure 1 is a flowchart for each strategy’s procedure, with the major differences among each circled. We calculated costs and benefits from the perspective of a VHA health care facility. First, we identified the costs per patient associated with each strategy’s protocol by identifying relevant cost factors from the literature. Next, we built a decision tree model on TreeAge Pro 2011 using these differing cost factors and probabilities of test acceptance and notification rates under each strategy. Table 1 lists the base-case value and the range of each variable entered into it. Using the computer software for decision tree analysis, Treeage, we performed a sensitivity analysis on the costs and probabilities in the model to account for assumptions and to test their effect on the findings derived from the model. Last, we estimated 1-year’s overall costs of each strategy, which was dependent on the number of people tested and the number of tested people identified as being (cases) or not being (noncases) HIV-infected. Strategies B and C included the one-time cost of installing the clinical reminders and the cost of quarterly feedback reports. The time horizon of our study was 1 year, and the costs calculated were converted to 2011 dollars [

      Bureau of Labor Statistics. Table Containing History of CPI-U U.S. All Items Indexes and Annual Percent Changes from 1913 to Present Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

      ].
      Table 1Input variables and sources.
      VariableBase CaseRangeSource
      HIV test characteristics,%
       Sensitivity of screening test99.560-100Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
      Strategies A, B and C in this paper respectively correspond to Strategy 1, 2 and 3 in Sander papers.
       Specificity of entire sequence of testing99.999460-100Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
       Prevalence of undiagnosed HIV, %0.40.24-0.56Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
      Probability of accepting test,%
       Strategy 14024-56Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
       Strategy 28048-100Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
       Strategy 39036-100Farnham, 2008
      • Farnham P.G.
      • Hutchinson A.B.
      • Sansom S.L.
      • et al.
      Comparing the costs of HIV screening strategies and technologies in health-care settings.
      Probability of receiving negative result, %
       Strategy 13325-41Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
      ; Farnham, 2008
      • Farnham P.G.
      • Hutchinson A.B.
      • Sansom S.L.
      • et al.
      Comparing the costs of HIV screening strategies and technologies in health-care settings.
       Strategy 2, 37556-94Spielberg, 2005
      • Spielberg F.
      • Branson B.M.
      • Goldbaum G.M.
      • et al.
      Choosing HIV counseling and testing strategies for outreach settings: a randomized trial.
      Probability of receiving positive result, %
       Strategy 1, 2, 37053-88Farnham, Sanders, 2010,
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
      Costs, $
      General practitioner wage per hour$85.26$64-107Bureau of Labor Statistics

      Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

      Registered nurse wage per hour$33.23$25-42Bureau of Labor Statistics

      Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

       Clinical reminder (Strategy 2)$0.55$0.41-0.69Based on 1 min for nurse to see and resolve
       Clinical reminder (Strategy 3)$1.42$1-2Based on 1 min for physician to see and resolve
      Costs, $
       General practitioner wage per hour$85.26$64-107Bureau of Labor Statistics

      Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

       Registered nurse wage per hour$33.23$25-42Bureau of Labor Statistics

      Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

      Clinical reminder (CR)
       Physician sees and resolves CR (Strategy 3)$0.71$0.5-0.89Based on 30 seconds for physician
       Nurse sees and resolves CR (Strategy 2)$0.28$0.21-0.35Based on 30 seconds for nurse
      Pre-test counseling
       Conventional counseling (Strategy 1)$21.32$21-36Based on 15 min for physician
       Stream-lined counseling (Strategy 2)$3.88$3-5Based on 7 min for nurse (Anaya, 2008)
      • Anaya H.D.
      • Chan K.
      • Karmarkar U.
      • et al.
      Budget impact analysis of HIV testing in the VA healthcare system.
      Electronically ordering HIV test
       Physician orders test (Strategy 1,3)$0.71$0.5-0.89Based on 30 seconds for physician
       Nurse orders test (Strategy 2)$0.28$0.21-0.35Based on 30 seconds for nurse
      Initial negative test$13.46$10-17Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
      Confirm positive test$56.27$42-70Sanders, 2010
      • Sanders G.D.
      • Ananya H.D.
      • Hoang T.
      • et al.
      Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
      Post-test counseling for negative result
       Strategy 1$21.32$16-27Based on 15 min for physician seen in person
       Strategy 2$2.77$2-3.50Based on 5 min for nurse via phone
       Strategy 3$7.11$5-9Based on 5 min for physician via phone
      Post-test counseling for positive result
       Strategy 1$42.63$32-53Based on 30 min for physician
       Strategy 2$16.62$12-21Based on 30 min for nurse
       Strategy 3 (recommended only)$42.63$32-53Based on 30 min for physician
      low asterisk Strategies A, B and C in this paper respectively correspond to Strategy 1, 2 and 3 in Sander papers.

       HIV Testing and Receipt of Results

      Different procedural requirements for each strategy affected the costs generated for each patient tested. Using publicly available databases from the Bureau of Labor Statistics (BLS) and the Health Economics Resources Center, to estimate direct costs per patient associated with each testing strategy’s procedure, we ascertained baseline conditions (conditions before August 2005 in the VA) and the condition of each intervention’s change (conditions from August 2005 to July 2009 for strategy B, which included both the clinical reminder system and counseling, and unpublished data collected during the year 2011 for strategy C, which included only the clinical reminder system without counseling) (M.B. Goetz, unpublished data, 2012) [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ,
      • Goetz M.B.
      • Hoang T.
      • Henry S.R.
      • et al.
      Evaluation of the sustainability of an intervention to increase HIV testing.
      ]. When published literature was unavailable, we relied on expert opinion from the senior author (M.B.G.) who has more than 25 years of experience in developing programs to provide HIV care within the VHA.

       Patient Costs Per Strategy

      The time for counseling was multiplied by the median hourly wage listed on the US BLS Web site for the occupations of “Family and General Practitioners” and “Registered Nurses” [

      Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

      ]. In strategy A, traditional time for in-person pretest counseling typically took about 15 minutes [
      • Anaya H.D.
      • Hoang T.
      • Golden J.F.
      • et al.
      Improving HIV screening and receipt of results by nurse-initiated streamlined counseling and rapid testing.
      ]. We assumed that the time for in-person post-test counseling for persons with positive results was 30 minutes of physician time whereas the time for in-person post-test counseling for persons with negative results was 15 minutes. In strategy B, streamlined pretest counseling typically took about 7 minutes of nurse time [
      • Anaya H.D.
      • Hoang T.
      • Golden J.F.
      • et al.
      Improving HIV screening and receipt of results by nurse-initiated streamlined counseling and rapid testing.
      ]. We assumed that telephonic test notification and post-test counseling of persons with negative test results took about 5 minutes of nurse time, whereas it also took nurses 30 minutes to conduct post-test counseling for persons with positive results. Pretest counseling was removed in strategy C. Post-test counseling was also no longer required in strategy C for persons with negative test results but is still strongly recommended for persons with positive test results (cases); we estimated that this cost about 30 minutes of the physician time whereas it took the physician 5 minutes to call and report negative test results. Here, Table 1 highlights the base variables with a range of ±40% of their base-case values. All the strategies accounted for an estimated 30 seconds of labor time for the nurse or physician to electronically order an HIV test for the patient. For each strategy, the initial HIV test performed (enzyme-linked immunosorbent assay) cost $13.46. The cost of testing for positive results was $56.27 to account for the performance of a Western blot test analysis.

      Clinical Reminder System

       Installation and Training Cost

      We estimated that it took nurses with technical computer skills, referred to as “Clinical Applications Coordinators,” about 4 hours to install the previously developed clinical reminder system at a VA facility and 20 hours to ensure that all data links were correct. The BLS wage data for registered nurses were used to calculate the labor time for the clinical applications coordinator [

      Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

      ].

       Audit Feedback and Provider Activation System

      The multimodal program also includes an audit feedback and provider activation component. For the provider activation aspect, the provider is given education materials for free; however, we accounted for the cost associated with the time spent by the clinical champion in promoting the clinical reminder program. On the basis of information obtained during previous studies [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ,
      • Goetz M.B.
      • Hoang T.
      • Knapp H.
      • et al.
      Central implementation strategies outperform local ones in improving HIV testing in Veterans Healthcare Administration facilities.
      ], we estimated that 2 hours per month would be spent during the first 6 months of implementation.

      Analytic Methods

      The cost of each strategy is highly dependent on the number of patients who undergo each testing strategy. The costs associated with each strategy testing protocol recur for each patient tested. In addition, costs are higher for cases than for noncases because multiple screening tests are performed to confirm the positive result and more provider time is spent with the patient. Therefore, the probability of patients accepting the test has a huge impact on the short-term costs generated in a year. Thus, a decision tree model was constructed to analyze each HIV testing strategy’s testing procedure by calculating the cost per case and noncase of each strategy.

       Overall Annual Estimates of Each Strategy

      We used the costs for each step in Figure 1 to perform basic calculations to estimate each strategy’s overall 1-year costs. Separate calculations were done for persons newly diagnosed with HIV infection to capture the increased post-test counseling time spent with these individuals. Because both strategies B and C utilized the clinical reminder intervention, costs for these strategies included the one-time installation and feedback costs associated.

       Cost-Effectiveness Analysis of HIV Testing Strategies

      The cost-effectiveness of different HIV testing strategies was compared by calculating the incremental annual total cost per case as well as per noncase. If strategy B was less costly than strategy C per case/noncase, then it would be considered cost-effective.

       Sensitivity Analysis

      The Tree Age model analyzing per-patient costs throughout the HIV testing process was also used to assess which variables have the largest effect on the incremental cost-effectiveness ratio results by performing deterministic, one-way sensitivity analyses. Cost variables of interest were the varying costs of pretest and post-test counseling. Probabilistic variables of interest were test acceptance rates and return rates. Calculations were also performed to determine what number of cases identified and what number of noncases identified would make the total cost of the most cost-effective strategy equivalent to the total cost of the next best alternative. All work was approved by the appropriate institutional review boards.

      Results

      All outcomes are from a single site (strategy A) that has been previously described [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ]. Implementing the clinical reminder system and streamlining the counseling procedure increased the number of people tested, the number of cases identified for that year, and the number of asymptomatic patients identified. For strategy A, 1906 patients were tested and 12 new cases of HIV infection were identified [
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      ]. For strategy B, when both the clinical reminder system and counseling were used, 3858 patients were tested and of those 19 new cases were identified [
      • Goetz M.B.
      • Hoang T.
      • Henry S.R.
      • et al.
      Evaluation of the sustainability of an intervention to increase HIV testing.
      ]. And finally, in the absence of pretesting counseling and the use of only the clinical reminder, 16,172 tests were performed, which resulted in 17 new diagnoses. All results are from differing 12-month periods of time.
      In the presence of clinical reminders for routine testing and no pretest counseling (strategy C), patients visiting the hospital were eight times more likely to undergo an HIV test compared with the period when testing was recommended only for patients at known risk of infection, pretest counseling, which took 15 minutes, was required, and providers were not prompted by the clinical reminder to offer HIV testing (strategy A). In addition to the improved HIV testing, it is important to identify the patient with HIV earlier before the immune system is impaired by HIV. When patients were identified by clinical reminders and pretest/post-test counseling (strategy B), almost half the cases identified were found to have a CD4 count of more than than 200 cells/µl compared with a third of the patients identified primarily with pretest/post-test counseling in the absence of a clinical reminder (strategy A). Removing pretest/post-test counseling and encouraging testing by using a clinical reminder (strategy C) resulted in identifying 82% more asymptomatic patients who had more than 200 CD4 cells/µl. The different strategies have a significant effect on cost and clinical outcomes.
      Annual total costs for three HIV testing strategies are presented in Table 2. Overall, annual costs were strongly influenced by the population screened. Strategies B and C included costs related to implementing the clinical reminder and the provider feedback operation, while strategy A did not include the intervention costs but did include the cost of pretest counseling with a physician’s salary. Strategy B had the lowest annual cost of $81,726, with cost per test of $21.18. Strategy C had the highest annual cost of $243,564, with cost per test of $15.06, due to the large population screened over that time period. Strategy A had the highest cost per case, $120.93, and cost per noncase, $56.80. Strategy C, which completely removed the cost of pretest counseling for each patient, had the lowest cost per case, $57.69, and cost per noncase, $14.88. Interestingly, even though the cost per test was about two times more in strategy C than in strategy B, the cost per test was about $6 less per test done in strategy C because there were more individuals tested in a routine-testing setting.
      Table 2Annual costs for each HIV testing strategy.
      Strategy AStrategy BStrategy CStrategy B- Strategy C
      Population screened
      Data was extracted from Goetz et al. [5].
      1,9063,85816,17212,314
      New diagnoses
      Data was extracted from Goetz et al. [5].
      121917
      Cases identified with CD4>200
      Data was extracted from Goetz, M.B., Unpublished data on HIV baseline conditions (number of HIV tests and diagnoses) in 2011. 2012
      3914
      HIV testing costs (recurring costs per patient)$109,031.25$79,719.01$241,383.30$161,664.29
       Estimated cost per case$120.93$77.32$57.69−$19.62
       Estimated Cost per Non-Case$56.80$20.38$14.88−$5.50
      Implementation of clinical reminder costs (one-time initial costs)
      Installation of CR was estimated to cost 4 hours of RN time. Troubleshooting of CR was estimated to cost 20 hours or RN time. RN hourly wage is $33.23
      $797.52$797.52
       Clinical applications coordinator installs CR$132.92$132.92
       Clinical applications coordinator troubleshooting CR$664.60$664.60
      Social marketing costs (recurring costs per year)
      Requesting reminder reports was estimated to cost 1 staff member 5 minutes. Generating reports was estimated to cost 1 staff member 40 minutes. Disseminating reports was estimated to cost 1 staff member 50 minutes. The average time for 1 manger to read emailed reports was 10 minutes, for 1 physician was 10 minutes, and for one nurse was 10 minutes. Physician’s hourly wage is $85.26. Manager’s hourly wage is $55. Bachelor’s hourly wage is $20 whereas PhD’s hourly wage is $50.
      $1,210.03$1,383.47$173.43
       1 Staff member (PhD) requests reminder reports$16.67$16.67
       1 Staff member (BA) generates reports$53.33$53.33
       1 Staff member (BA) disseminates reports$66.67$66.67
       3 Manager reads e-mailed report$110.00$110.00
       15 Physicians read e-mailed report$852.60
       5 Nurses read e-mailed report$110.77
       20 Physicians read e-mailed report$1,136.80
      Estimated annual total cost$109,208.98$81,726.57$243,564.29$161,837.72
      low asterisk Data was extracted from Goetz et al.
      • Goetz M.B.
      • Hoang T.
      • Bowman C.
      • et al.
      A system-wide intervention to improve HIV testing in the Veterans Health Administration.
      .
      Data was extracted from Goetz, M.B., Unpublished data on HIV baseline conditions (number of HIV tests and diagnoses) in 2011. 2012
      Installation of CR was estimated to cost 4 hours of RN time. Troubleshooting of CR was estimated to cost 20 hours or RN time. RN hourly wage is $33.23
      § Requesting reminder reports was estimated to cost 1 staff member 5 minutes. Generating reports was estimated to cost 1 staff member 40 minutes. Disseminating reports was estimated to cost 1 staff member 50 minutes. The average time for 1 manger to read emailed reports was 10 minutes, for 1 physician was 10 minutes, and for one nurse was 10 minutes. Physician’s hourly wage is $85.26. Manager’s hourly wage is $55. Bachelor’s hourly wage is $20 whereas PhD’s hourly wage is $50.
      In addition, Table 2 suggested that under the assumption in our model, in strategy B the hospital spent $173.43 more per year in social marketing than under strategy C. In our model, we examined the effect of health professionals on the financial outcomes. On the basis of expert opinions on the workflow at the clinic, we made a reasonable assumption that 15 physicians and 5 nurses read reports in strategy B and 20 physicians read reports in strategy C in our model. Table 2 also presented the incremental cost per case identified between strategy B and strategy C. Strategy A was immediately excluded because it was found to be “Strongly Dominated,” meaning that strategy A is more costly and less effective than strategy B. Note that this strategy does not include the implementation of clinical reminders. Thus, only strategies B and C are compared with cost-effectiveness. Given the same effectiveness, identified as one case identified, strategy B spent $19.62 less than did strategy C.
      One-way sensitivity analysis was performed on variables associated with recurring patient costs. Using TreeAge, plots were created using the variable range as the x-axis and the resulting incremental cost-effectiveness ratio as the y-axis. Variables that the model was sensitive to were patient acceptance rates at each strategy and the costs of post-test counseling for noncases during strategies B and C. Strategies A and B had similar incremental cost-effectiveness ratios when the probability of accepting a test during strategy A was 0.33 and when the probability of accepting a test during strategy B was 0.9. The incremental cost-effectiveness ratios for strategies A and B were approximately equal when post-test counseling for negatives cost $7.50. The model was also found to be sensitive to the cost of pretest counseling during strategy A when they were at $15.20 and $20.40 at strategy B and strategy C, respectively. In addition to the cost-effectiveness of strategy C, the advantage of using strategy C is that more patients are tested and thus more diagnoses are made.

      Discussion

      We performed the comparative cost analysis of three HIV testing strategies; two strategies included the first-year implementation costs of clinical reminders, an electronic reminder that helps facilitate the HIV testing process during a patient’s visit. We measured short-term costs of these strategies for the first year of implementation of clinical reminders. The implementation of the clinical reminders is a one-time cost that would not be added in additional years of using this strategy. Main health outcomes measured in our model were total annual cost, cost/case identified, and incremental cost per case or cost per test.
      Decreasing the time spent on counseling and substituting nurses for physician significantly cut annual costs for strategy B than for strategy A, even with the addition of clinical reminders and testing costs for the additional patients tested. We found that at the test level, strategy B and strategy C were less expensive than strategy A. These two strategies also relatively resulted in more cases identified and more cases identified with higher CD4 counts than with strategy A.
      The Centers for Disease Control and Prevention, the American College of Physicians, and the US Preventive Services Task Force recommend routine HIV screening in all health care settings [
      • Branson B.M.
      • Handsfield H.H.
      • Lampe M.A.
      • et al.
      Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings.
      ,
      • Qaseem A.
      • Snow V.
      • Shekelle P.
      • et al.
      Screening for HIV in health care settings: a guidance statement from the American College of Physicians and HIV Medicine Association.
      ,
      • Moyer V.A.
      U.S. Preventive Services Task Force*. Screening for HIV: U.S. Preventive Services Task Force Recommendation Statement.
      ]. Although strategies B and C both use clinical reminders, strategy C incorporates routine rather than risk-based HIV testing. Our results suggested that reducing the time spent on counseling reduced cost per case identified. To screen one HIV positive case, strategy C spent $19.62 less in HIV testing than did strategy B. Our finding also suggested that the number of cases identified increased in strategy C than in strategy B. Therefore, the total HIV testing costs for strategy C were higher than for strategy B. Despite the possibility of higher total costs of HIV testing, the results of strategy C will still facilitate hospital manager’s decision to implement clinical reminders as the routine procedure for HIV testing because it will promote the performance of HIV testing by increasing the number of patients tested in the long run.
      Social marketing cost consists of providers generating, requesting, reading, and disseminating the clinical reminder report. Our study found that annual social marketing costs were $173.43 higher in strategy C than that in strategy B. The social marketing cost does not affect the costs per test; on the other hand, it is the changes in policies due to social marketing that change the cost. Therefore, social marketing increases the total cost by increasing the numbers of tests that are performed.
      Our study has several limitations. Costs were also calculated under the assumption that the provider already has an electronic medical records system in place. Data were collected from patients of a veteran hospital who may have different characteristics than do non-VA patient populations. Another limitation is that strategies A, B, and C were introduced sequentially at different times and that as a consequence the population being offered HIV testing differed; that is, the highest risk patients were subject to being offered HIV testing before the implementation of strategy C, which might then partially explain the lower rates of new case finding when strategy C was used. Also, the actual hourly wages of physicians and nurses in VA hospitals may not be the same as extracted from the BLS database, the HIV prevalence rate in VA hospital areas in this study may be quite different from that in other areas, or physician and nurse time for counseling may vary across hospitals. Although this analysis was limited to a single year of results, it is unlikely that per test costs would change substantially in subsequent years because we included the initial costs of installing the clinical reminder software and more than 90% of all costs were directly related to HIV testing per se. In addition to factors affecting cost, health managers may need to be mindful of possible barriers early on so that they can strategize or minimize any delay in adoption, activation, and installation of clinical reminders. For example, acquiring additional resources and reserving time are required to educate coordinators on the expected performance of the clinical reminder so that they can identify errors in adapting the program to the specific data structure of a particular medical facility (e.g., differences in laboratory names require customization of the program to identify persons who have previously undergone HIV testing). Furthermore, identifying key stakeholders who can serve as the “local champion” to conduct social marketing and academic detailing has been proven to be a critical component of the success of programs promoting HIV testing and the use of the HIV testing clinical reminder within the VA [
      • Goetz M.B.
      • Hoang T.
      • Knapp H.
      • et al.
      Central implementation strategies outperform local ones in improving HIV testing in Veterans Healthcare Administration facilities.
      ,

      Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

      ].
      The modeling and economic analysis model provides a useful tool for decision makers considering new policy, recommendations, or programs under different scenarios. As illustrated in our model and assumptions, we projected the cost of 1-year budget of the three scenarios of HIV testing strategies. In 2012, our research group showed the marginal effect of the cost on the basis of the budget impact analysis of improving HIV testing rates, HIV positive rates, and counseling time in hypothetical scenarios [
      • Anaya H.D.
      • Chan K.
      • Karmarkar U.
      • et al.
      Budget impact analysis of HIV testing in the VA healthcare system.
      ]. As a follow-up to our study, using deidentified patient data, here we focused on the budget impact analysis of improving the HIV testing rate in one hospital facility in Greater Los Angeles, California. Most recently, because the Affordable Care Act encourages consumers to use preventive services such as HIV screening, this could further help increase the HIV testing rate by removing an economic barrier. It is likely, however, that HIV testing rates will remain far below the desired rates with the use of programs such as we have developed to promote and facilitate HIV testing. In particular, clinical reminders could be an efficient tool to streamline check-up and screening services; and thus, physicians can use the time to focus on patient-centered issues related to specialty services.
      Source of financial support: The authors have no other financial relationships to disclose.

      References

        • Branson B.M.
        • Handsfield H.H.
        • Lampe M.A.
        • et al.
        Revised recommendations for HIV testing of adults, adolescents, and pregnant women in health-care settings.
        MMWR Recomm Rep. 2006; 55 (quiz CE1–4): 1-17
        • Qaseem A.
        • Snow V.
        • Shekelle P.
        • et al.
        Screening for HIV in health care settings: a guidance statement from the American College of Physicians and HIV Medicine Association.
        Ann Intern Med. 2009; 150: 125-131
        • Moyer V.A.
        U.S. Preventive Services Task Force*. Screening for HIV: U.S. Preventive Services Task Force Recommendation Statement.
        Ann Intern Med. 2013; 159: 51-60
      1. Testing for human immunodeficiency virus in Veterans Health Administration facilities. Available from: http://www1.va.gov/vhapublications/ViewPublication.asp?pub_ID=2056. [Accessed August 20, 2009].

        • Goetz M.B.
        • Hoang T.
        • Bowman C.
        • et al.
        A system-wide intervention to improve HIV testing in the Veterans Health Administration.
        J Gen Intern Med. 2008; 23: 1200-1207
        • Sanders G.D.
        • Ananya H.D.
        • Hoang T.
        • et al.
        Cost-effectiveness of strategies to improve HIV testing and receipt of results: economic analysis of a randomized controlled trial.
        J Gen Intern Med. 2010; 25: 556-563
        • Goetz M.B.
        • Hoang T.
        • Knapp H.
        • et al.
        Central implementation strategies outperform local ones in improving HIV testing in Veterans Healthcare Administration facilities.
        J Gen Intern Med. 2013; 28: 1311-1317
        • Anaya H.D.
        • Chan K.
        • Karmarkar U.
        • et al.
        Budget impact analysis of HIV testing in the VA healthcare system.
        Value Health. 2012; 15: 1022-1028
      2. Bureau of Labor Statistics. Table Containing History of CPI-U U.S. All Items Indexes and Annual Percent Changes from 1913 to Present Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

        • Goetz M.B.
        • Hoang T.
        • Henry S.R.
        • et al.
        Evaluation of the sustainability of an intervention to increase HIV testing.
        J Gen Intern Med. 2009; 24: 1275-1280
      3. Bureau of Labor Statistics. May 2011 National Occupational Employment and Wage Estimates United States 2011. Available from: http://www.bls.gov/cpi/. [Accessed May 1, 2014].

        • Anaya H.D.
        • Hoang T.
        • Golden J.F.
        • et al.
        Improving HIV screening and receipt of results by nurse-initiated streamlined counseling and rapid testing.
        J Gen Intern Med. 2008; 23: 800-807
        • Farnham P.G.
        • Hutchinson A.B.
        • Sansom S.L.
        • et al.
        Comparing the costs of HIV screening strategies and technologies in health-care settings.
        Public Health Rep. 2008; 123: 51-62
        • Spielberg F.
        • Branson B.M.
        • Goldbaum G.M.
        • et al.
        Choosing HIV counseling and testing strategies for outreach settings: a randomized trial.
        J Acquir Immune Defic Syndr. 2005; 38: 348-355