Initiation of guideline-preferred first-line antiretroviral therapy (ART) was associated with better ART adherence and persistence and similar total healthcare expenditures among Medicaid-insured HIV patients.
To compare antiretroviral therapy (ART) adherence and persistence and total healthcare expenditures in Medicaid-insured patients with human immunodeficiency virus (HIV) initiating preferred or nonpreferred first-line ART based on March 2012 HHS HIV treatment guidelines.
Retrospective observational study using Medicaid administrative healthcare claims from 15 states.
Subjects were HIV patients 18 to 64 years who initiated first-line HIV-related ART between January 1, 2007, and September 30, 2011, with continuous enrollment for 6 months prior to and at least 3 months following ART initiation. Patients were classified as having initiated preferred or nonpreferred ART based on March 2012 HHS HIV treatment guidelines. Outcomes were: ART adherence (proportion of days covered dichotomized at ≥80% and ≥95%), time to ART nonpersistence, and per patient per month (PPPM) total healthcare expenditures. Outcomes were evaluated using multivariable regressions.
Sample included 1979 patients initiating preferred ART regimens and 1614 patients initiating nonpreferred ART; overall mean age was 41 years; 48% of subjects were female. In the multivariable analyses, patients initiating preferred ART regimens had significantly greater odds of adherence ≥80% (odds ratio [OR], 1.38; 95% CI, 1.07-1.77) and adherence ≥95% (OR, 1.26; 95% CI, 1.05-1.51), and a significantly lower hazard of nonpersistence (HR, 0.48; 95% CI, 0.44-0.52). PPPM total healthcare expenditures were numerically lower for patients initiating preferred ART regimens (—$341; 95% CI, –$888 to $255) but the difference was not deemed significant.
This study reinforces the value of HHS recommendations for first-line ART. The potential impact of these findings will grow as more HIV patients become Medicaid-eligible under the Patient Protection and Affordable Care Act.
Am J Manag Care. 2014;20(6):448-455
More than 1 million people in the United States are currently infected with human immunodeficiency virus (HIV).1,2 While there is currently no cure for HIV, antiretroviral therapy (ART) can reduce HIV-associated morbidity and mortality.3 Based on accumulated clinical evidence, the March 2012 HHS guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents recommend 4 specific “preferred” ART regimens for ART-naïve patients.3 These are: efavirenz/tenofovir/ emtricitabine (EFV/TDF/FTC), ritonavir-boosted atazanavir + tenofovir/emtricitabine (ATV/r + TDF/ FTC), ritonavir-boosted darunavir + tenofovir/emtricitabine (DRV/r + TDF/FTC), and raltegravir + tenofovir/ emtricitabine (RAL + TDF/FTC).
These guideline-preferred ART regimens have been studied in randomized controlled trials and have been shown to confer the benefits of “optimal and durable virologic efficacy, favorable tolerability and toxicity profiles, and ease of use.”3 Comparatively, nonpreferred ART regimens (defined by guidelines as alternative, acceptable, or not recommended) may have “reduced virologic activity, lack of efficacy data from large clinical trials, or other factors (such as greater toxicities, pill burden, drug interaction potential, or need for additional testing).”3 Though the recommendations for preferred ART regimens are unambiguous, real-world evidence suggests that some healthcare providers nonetheless prescribe nonpreferred ART regimens.4,5 While the beneficial clinical aspects of guideline- preferred ART regimens are well described, the extent to which such benefits translate into improved ART adherence and persistence or lower healthcare expenditures in the ‘real world’ setting of routine clinical practice is poorly understood.
Among the various clinical settings in which HIV patients are commonly treated, the Medicaid setting is one of the most important. Medicaid is the largest source of health insurance for HIV patients living in the United States.6 Furthermore, as of January 2014, Medicaid coverage was expanded to cover all Americans with income below 133% of the federal poverty level, a provision of the Patient Protection and Affordable Care Act (PPACA). This change is predicted to significantly expand and improve access to, and quality of care for, HIV patients living in the United States.7-9 Thus, the objective of this study was to compare ART adherence and persistence and total healthcare expenditures in Medicaid- insured HIV patients initiating current guideline-preferred or nonpreferred first-line ART.
Setting and Data
The setting of this retrospective observational cohort study was routine clinical practice in a sample of Medicaid programs in 15 states of varying sizes and sociodemographic compositions across the United States. The Medicaid data were drawn from the 2004-2011 Truven Health Analytics MarketScan Multi-State Medicaid Database, which comprises covered inpatient medical, outpatient medical (including lab claims but not lab results), and outpatient pharmaceutical claims and encounter records, along with patient enrollment records and demographic information such as age, sex, and race. During the study period, over 25 million patients were represented in the database. Further public disclosure of identifying information about the state Medicaid programs covered by the database is restricted by confidentiality agreements between Truven Health Analytics and the data-contributing states.
The data contained in the Medicaid database are statistically deidentified and compliant with the Health Insurance Portability and Accountability Act privacy regulations. Therefore, Institutional Review Board approval and written informed consent were not required for this study.
Patients were identified and study variables were constructed from the data using International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes, Current Procedural Terminology codes, Healthcare Common Procedure Coding System codes, and National Drug Codes (NDCs), as appropriate.
HIV patients meeting all of the following inclusion criteria were selected into the study sample: initiated HIVrelated ART between January 1, 2007, and September 30, 2011 (the date of ART initiation was designated the index date); aged 18 to 64 years on the index date; had continuous enrollment with pharmacy benefits for 6 months prior to, and for at least 3 months following, the index date. Patients meeting any of the following exclusion criteria were excluded from the study sample: eligible for Medicare; showed evidence of HIV-related ART prescriptions any time prior to the index date, with data extending back to January 1, 2004, being used to check for prior ART use; showed any evidence of pregnancy or hepatitis B infection.
The study period comprised an index date, a baseline period, and a follow-up period. As noted above, the date of ART initiation was designated as the index date. This could have been the date of initiation on a single or combination of ART agents. Outpatient prescription claims for HIV-related ART were evaluated during the first 14 days after the index date (index date included) to establish ART regimens. The fixed 6-month period ending on the day before the index date was designated the baseline period. A variable-length period beginning on the index date was designated as the follow-up period. The duration of the follow-up period equaled the number of days from the index date until the first occurrence of (1) a prescription claim for an ART that was not part of the initiated ART regimen; (2) at least a 30-day continuous gap in any agent from the ART regimen (in which case, the end of the follow- up period was set to the last day on which the patient had possession of ART); (3) disenrollment from Medicaid benefits; (4) or the study end date of December 31, 2011.
Classification of Patients’ Initiated ART
Based upon the March 2012 HHS HIV treatment guidelines,3 patients were classified as initiating either guideline-preferred or nonpreferred first-line ART. This classification was applied for all data years and did not vary according to the preferred guidelines during the timeframe in which the patient actually initiated ART. As noted above, the March 2012 guideline-preferred ART regimens were: EFV/TDF/FTC, ATV/r + TDF/FTC, DRV/r + TDF/FTC, and RAL + TDF/FTC.
Study outcomes were ART adherence, ART persistence (ie, time to ART nonpersistence), and total healthcare expenditures. ART adherence and persistence were measured using ART prescription fill dates and days supplied. ART adherence was measured using the proportion of days covered (PDC), which was calculated as the proportion of days that patients had all components of their ART regimen “on hand” during the follow-up period. Accordingly, any days on which the patient did not possess all components of their ART regimen were counted as nonadherent days. ART adherence was dichotomized at 80% or greater (vs <80%) and 95% or greater (vs <95%). The 95% threshold was chosen on the basis of World Health Organization (WHO) HIV/AIDS program guidelines for antiretroviral therapy for HIV-infected adults and adolescents, which recommend maintenance of adherence levels at or above 95%.10 The 80% threshold was chosen on the basis of evidence showing that contemporary regimens may be slightly more “forgiving” to lapses in adherence.11 Time to ART nonpersistence was equal to the duration of the follow-up period. Patients were flagged as being nonpersistent if their follow-up period terminated with a filled prescription claim for an ART that was not part of the initiated first-line ART regimen or at least a 30-day continuous gap in therapy for any agent within the initiated first-line ART regimen. Patients were flagged as being censored if they were not flagged as being nonpersistent.
Healthcare utilization and expenditures were measured for all causes throughout the follow-up period and were summarized as per patient per month (PPPM) to account for variability in the duration of follow-up. Healthcare utilization and expenditures were categorized by place of service and included inpatient admissions, emergency department (ED) visits, outpatient office visits, outpatient laboratory services, other outpatient services, and pharmacy. Healthcare expenditures were expressed in 2011 constant dollars, adjusted using the Medical Care component of the Consumer Price Index.12 Healthcare expenditures included the gross covered payments for all healthcare services or products (ie, the amount eligible for payment after applying pricing guidelines such as fee schedules and discounts, but including deductibles, copayments, and coordination of benefits).
Covariates were measured to describe the study sample and to adjust for across-group differences in the multivariable models. These included demographics such as age, sex, and race; clinical characteristics such as proxies of overall health status and specific comorbidities; and ART regimen characteristics such as the number of ART agents in the initiated regimen and the components of the initiated regimen. The full listing of covariates is provided within the results tables.
Multivariable generalized linear models (GLMs) with logit link functions and binomial error distributions were used to conduct adjusted analyses of ART adherence of 80% or greater and 95% or greater, separately. These analyses excluded patients who were persistent for 30 days or less. The reason for this exclusion is that most ART prescriptions had 30 days of supply, which results in a 100% adherence value when the period over which persistence is measured is 30 days or less. Multivariable Cox proportional hazards models with time-invariant covariates were used to conduct an adjusted analysis of the time to ART nonpersistence. Multivariable log-ordinary least squares (OLS) regressions were used to conduct an adjusted analysis of PPPM total healthcare expenditures. Multivariable log-OLS models were chosen because kurtosis, with respect to log scale residuals, was greater than 3, which is indicative of a heavy-tailed distribution that can result in precision losses for alternative healthcare expenditure models such as log-gamma GLMs.13 PPPM total healthcare expenditures were retransformed to the dollar scale using 2 “smearing” factors, and expenditure-related inference was based on the “recycled prediction” method, with 500 bootstrap repetitions. 14,15 The list of covariates included in the multivariable regressions are provided in the results section below.
RESULTSStudy SampleTable 1 presents the sample attrition associated with the study inclusion and exclusion criteria. A total of 3593 patients met all of the study selection criteria. Among 1979 patients initiating guideline-preferred ART regimens, 1259 (63.6%) initiated EFV/TDF/FTC, 498 (25.2%) initiated ATV/r + TDF/FTC, 143 (7.2%) initiated DRV/r + TDF/ FTC, and 79 (4.0%) initiated RAL + TDF/FTC. Among patients initiating nonpreferred ART, the most commonly used ART agents that are not recommended as part of guideline-preferred ART regimens included 3 nucleoside reverse transcriptase inhibitors (NRTIs) (lamivudine, zidovudine, and abacavir) and 1 protease inhibitor (PI) (lopinavir).
Table 2 presents patient baseline demographic information, and Table 3 presents patient baseline clinical characteristics. Compared with patients initiating nonpreferred ART regimens, a slightly greater proportion of patients initiating guideline-preferred ART regimens was black, a lower proportion of patients was enrolled in health plans with capitated payment arrangements, and this group also showed a higher prevalence of renal diseases.
Results of Unadjusted AnalysesTable 4 presents the results of the unadjusted analyses. Compared with patients initiating nonpreferred ART, those initiating guideline-preferred ART regimens had greater unadjusted proportions of patients achieving ART adherence of 80% or greater and 95% or greater; lower unadjusted incidence rates of nonpersistence per 10 person years; lower unadjusted mean PPPM healthcare expenditures on inpatient admissions (by $1345), ED visits (by $15), outpatient office visits (by $2), “other” outpatient services (by $386), outpatient pharmacy (by $6), and total healthcare expenditures (by $1746); and had slightly higher unadjusted mean PPPM healthcare expenditures on outpatient laboratory (by $8).
Results of Multivariable-Adjusted AnalysesTable 5 presents the results of the multivariableadjusted analyses. Compared with patients initiating nonpreferred ART, those initiating guideline-preferred ART regimens had: 37.4% greater adjusted odds (P = .014) of ART adherence of 80% or greater and 25.9% greater adjusted odds (P = .014) of ART adherence of 95% or greater; 51.9% lower adjusted hazards (P <.001) of nonpersistence; lower adjusted mean PPPM total expenditures (by $341), though this difference did not reach statistical significance.
This is the first study to compare ART adherence and persistence and total healthcare expenditures in Medicaid- insured HIV patients initiating guideline-preferred or nonpreferred first-line ART. The overall findings of this study suggest that guideline-preferred ART regimens may be associated with several important beneficial outcomes in the “real-world” setting of routine clinical practice.
It has been shown that suboptimal ART adherence and persistence can lead to reduced treatment response and sometimes irreversible loss of treatment options caused by resistance mutations.3,16,17 Accordingly, HHS guidelines recommend that physicians encourage patients to “adhere as closely as possible” to their ART,3 and HHS and WHO guidelines emphasize the importance of interventions to improve ART adherence.3,9 The present study found that when compared with patients initiating nonpreferred ART, those initiating guideline-preferred ART regimens had significantly greater adjusted odds of ART adherence of at least 80% and at least 95%, and had significantly lower adjusted hazards of ART nonpersistence.
The current findings are consistent with previous studies. Purdum and colleagues (2004) conducted a retrospective, administrative claims—based study of 1791 HIVinfected patients treated with ART between February 1998 and July 1999.18 They found that patients treated with “preferred” or “alternate” ART (based on 1998 HHS guidelines) had higher ART adherence, longer time-to-inpatient admission, and lower medical expenditures, but higher prescription and total healthcare expenditures than patients treated with “not generally recommended” or “not recommended” ART. Cocohoba and colleagues (2008) examined outcomes according to type of initial ART self-reported by 217 Women’s Interagency HIV Study participants between April 1998 and October 2004. They found that patients initiating “preferred” or “alternate” ART (based on HHS guidelines at the time) had a lower incidence of ART switching and more favorable short-term immunologic and virologic outcomes than patients initiating “not generally recommended” or “not recommended” ART.19
Finally, there are a number of other observational studies that were not explicitly designed to compare guidelinepreferred and nonpreferred first-line ARTs, but included comparisons of adherence and persistence between EFV/ TDF/FTC and other regimens, including those that are nonpreferred within the current HHS guidelines.20-22 The findings from these studies are also consistent with the findings from the present study.
The present study found that compared with patients initiating nonpreferred ART, those initiating guidelinepreferred ART regimens had $1746 lower unadjusted mean PPPM total healthcare expenditures. After multivariable adjustment, PPPM total healthcare expenditures were $341 lower for patients initiating guideline-preferred ART regimens, but the difference was not statistically significant. The improved ART adherence and persistence observed among patients initiating guideline-preferred ART regimens may be reflective of fewer adverse events, fewer drug switches, and lower related healthcare expenditures. Thus, these findings suggest that the potential economic benefits associated with guideline-preferred ART regimens may offset potentially higher drug acquisition costs.
The present study has limitations. Among patients initiating guideline-preferred ART regimens, DRV/r + TDF/FTC and RAL + TDF/FTC represented only 11.2% of guideline-preferred ART regimens, primarily because DRV/r and RAL were initially included as components of preferred first-line ART in the November 3, 2008, and December 1, 2009, HHS guidelines, respectively, which is later than the initial inclusion of EFV and ATV.23 Consequently, the present study’s findings are primarily driven by EFV/TDF/FTC and ATV/r + TDF/FTC. The data used for this study were not originally collected for research purposes, and the diagnostic coding on administrative claims is recorded by physicians to support reimbursement. Noncoded or miscoded administrative claims can result in measurement error when constructing variables that rely on diagnostic or procedure coding. Calculations of ART adherence and persistence were based on the service date and days of supply from outpatient prescription drug claims and implicitly assumed that patients took their ART as directed by their physicians. Consequently, it is unknown whether patients skipped doses or discontinued the ART prior to the end of the last observed day’s supply. The impact of this limitation, however, is tempered by prior studies demonstrating that prescription refill records correlate well with clinical outcomes in HIV.24-26 One other limitation is that the Multistate Medicaid database may not be generalizable to the entire US Medicaid population. While multivariable adjustment was performed when comparing the study outcomes between the various study groups, there remains the possibility of residual confounding from unmeasured variables that affect outcomes and the choice to treat with a guideline-preferred or nonpreferred first-line ART regimen.
Finally, it is important to note that patients were classified as initiating either guideline-preferred or nonpreferred first-line ART based upon the March 2012 HHS HIV treatment guidelines.3 This classification was applied for all data years and did not vary according to the preferred guidelines during the time frame in which the patient actually initiated ART. Thus, a number of regimens considered nonpreferred based on the March 2012 HHS HIV treatment guidelines were actually preferred in earlier parts of our study period, which spanned January 1, 2007, through September 30, 2011. For example, fosamprenavir + ritonavir + 2 NRTIs and lopinavir/ritonavir + 2 NRTIs were both preferred PIbased regimens in the December 1, 2007, HHS HIV treatment guidelines.27 Because of this aspect of our preferred/ nonpreferred classification, data on the proportion of patients initiating nonpreferred regimens should not be interpreted as evidence of physicians failing to follow guidelines.
Compared with patients initiating nonpreferred ART, those initiating guideline-preferred ART regimens had significantly longer durations of time to ART nonpersistence, were significantly more likely to adhere to their ART, and had similar PPPM total healthcare expenditures. This study reinforces the value of HHS recommendations for first-line ART. The potential impact of these findings will increase as more HIV patients become Medicaideligible under the PPACA.Author Affiliations: Truven Health Analytics, Bethesda, MD (SSJ, AMF, B-CC); Bristol-Myers Squibb (BMS), Plainsboro, NJ (TJ, TH).
Source of Funding: This study was supported by BMS.
Author Disclosures: Mr Johnston, Ms Farr, and Dr Chu report employment with Truven Health Analytics, which received payment from BMS to complete this study. Drs Juday and Hebden report employment with BMS at the time of study completion.
Authorship Information: Concept and design (SSJ, TJ, B-CC, TH); acquisition of data (SSJ, AMF, B-CC); analysis and interpretation of data (SSJ, TJ, AMF, B-CC, TH); drafting of the manuscript (SSJ, TJ, B-CC, TH); critical revision of the manuscript for important intellectual content (SSJ, TJ, AMF, B-CC, TH); statistical analysis (SSJ, AMF, B-CC); obtaining funding (SSJ); supervision (TJ, TH).
Address correspondence to: Stephen S. Johnston, MA, Truven Health Analytics, 7700 Old Georgetown Rd, Suite 650, Bethesda, MD 20814. Email: firstname.lastname@example.org. Hall HI, Song R, Rhodes P, et al. Estimation of HIV incidence in the United States. JAMA. 2008;300(5):520-529.
2. CDC. Twenty-five years of HIV/AIDS—United States, 1981-2006. MMWR. 2006;55(21):585-589.
3. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. HHS website. http://www.aidsinfo.nih.gov/Content- Files/AdultandAdolescentGL.pdf. Published March 27, 2012. Accessed April 10, 2012.
4. Hirsch JD, Gonzales M, Rosenquist A, Miller TA, Gilmer TP, Best BM. Antiretroviral therapy adherence, medication use, and health care costs during 3 years of a community pharmacy medication therapy management program for Medi-Cal beneficiaries with HIV/AIDS. J Manag Care Pharm. 2011;17(3):213-223.
5. Stone VE, Mansourati FF, Poses RM, Mayer KH. Relation of physician specialty and HIV/AIDS experience to choice of guideline-recommended antiretroviral therapy. J Gen Intern Med. 2001;16(6):360-368.
6. Bozzette SA, Berry SH, Duan N, et al. The care of HIV-infected adults in the United States. N Engl J Med. 1998;339(26):1897-1904.
7. Pub L No. 111Y148, 124 Stat 119, to be codified as amended at scattered sections of the Internal Revenue Code and in 42 USC.
8. National HIV/AIDS Strategy for the United States. The White House Office of National AIDS Policy. www.whitehouse.gov/sites/default/files/ uploads/NHAS.pdf. Published July 13, 2010. Accessed February 16, 2012.
9. Kates J. How the ACA changes pathways to insurance coverage for people with HIV. The Henry J. Kaiser Family Foundation website. http:// policyinsights.kff.org/2012/september/how-the-aca-changes-pathwaysto- insurance-coverage-for-people-with-hiv.aspx. Accessed January 31, 2013.
10. HIV/AIDS Programme: antiretroviral therapy for HIV infection in adults and adolescents: recommendations for a public health approach: revised 2006. World Health Organization website. www.who.int/hiv/ pub/guidelines/artadultguidelines.pdf. Accessed October 1, 2012.
11. Maggiolo F, Ravasio L, Ripamonti D, et al. Similar adherence rates favor different virologic outcomes for patients treated with nonnucleoside analogues or protease inhibitors. Clin Infect Dis. 2005;40(1): 158-163.
12. Consumer Price Index Detailed Report Tables Annual Average 2008. US Bureau of Labor Statistics website. www.bls.gov/cpi/cpid10av.pdf. Accessed January 31, 2013.
13. Manning WG, Mullahy J. Estimating log models: to transform or not to transform? J Health Econ. 2001;20(4):461-94.
14. Kleinman LC, Norton EC. What’s the risk? a simple approach for estimating adjusted risk measures from nonlinear models including logistic regression. Health Serv Res. 2009;44(1):288-302.
15. Duan N. Smearing estimate: a nonparametric retransformation method. J Am Stat Assoc. 1983:78:605-610.
16. Kantor R, Shafer RW, Follansbee S, et al. Evolution of resistance to drugs in HIV-1-infected patients failing antiretroviral therapy. AIDS. 2004;18(11):1503-1511.
17. Napravnik S, Edwards D, Stewart P, Stalzer B, Matteson E, Eron JJ Jr. HIV-1 drug resistance evolution among patients on potent combination antiretroviral therapy with detectable viremia. J Acquir Immune Defic Syndr. 2005;40(1):34-40.
18. Purdum AG, Johnson KA, Globe DR. Comparing total health care costs and treatment patterns of HIV patients in a managed care setting. AIDS Care. 2004;16(6):767-780.
19. Cocohoba J, Wang QJ, Cox C, et al. Consistency of initial antiretroviral therapy with HIV treatment guidelines in a US cohort of HIVinfected women. J Acquir Immune Defic Syndr. 2008;47(3):377-383.
20. Juday T, Gupta S, Grimm K, Wagner S, Kim E. Factors associated with complete adherence to HIV combination antiretroviral therapy. HIV Clin Trials. 2011;12(2):71-78.
21. Juday T, Grimm K, Zoe-Powers A, Willig J, Kim E. A retrospective study of HIV antiretroviral treatment persistence in a commercially insured population in the United States. AIDS Care. 2011;23(9):1154-1162.
22. Taneja C, Juday T, Gertzog L, et al. Adherence and persistence with non-nucleoside reverse transcriptase inhibitor-based antiretroviral regimens. Expert Opin Pharmacother. 2012;13(15):2111-2118.
23. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents.HHS website, archived guidelines. http://aidsinfo.nih .gov/guidelines/archive/adult-and-adolescent-guidelines/. Accessed January 31, 2013.
24. Berg KM, Arnsten JH. Practical and conceptual challenges in measuring antiretroviral adherence. J Acquir Immune Defic Syndr. 2006;43:S79-S87.
25. Grossberg R, Bross R. Use of pharmacy refill data as a measure of antiretroviral adherence. Curr HIV/AIDS Rep. 2007;4(4):187-191.
26. Kitahata MM, Reed SD, Dillingham PW, et al. Pharmacy-based assessment of adherence to HAART predicts virologic and immunologic treatment response and clinical progression to AIDS and death. Int J STD AIDS. 2004;15:803-810.
27. Panel on Antiretroviral Guidelines for Adults and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. HHS. December 1, 2007:1-144.