HIV/AIDS has progressed to a complex, chronic manageable disease in many nations.1,2 However, estimates are that only 19% of all HIV-positive patients in the United States have maximal virological control (HIV RNA level below limits of quantification of the assay, “BLQ”).3 Although the exact level of antiretroviral therapy (ART) adherence necessary to achieve BLQ is debated, it is widely accepted that ART adherence can be influenced by factors at the patient, provider, and system levels.4 In the United States, HIV care is often provided by an HIV multidisciplinary care team (MDCT), comprising varying professionals including nurses, pharmacists, case mangers, and social workers. The composition of these teams can vary greatly by clinical site; yet, the impact of this varied composition on ART adherence has not been determined.5,6
Improved ecological or system factors including access to optimized multidisciplinary care should improve the individual's health care, thus improving the individual's willingness to adhere to the treatment program.7,8 The Department of Health and Human Services guidelines for use of antiretroviral agents recommend a health care team approach with nurses, pharmacists, and peer counselors to improve adherence to ART but provide no further specific recommendation.9 Recommendations for HIV MDCT have stemmed from studies that show that MDCTs are associated with increased receipt of ancillary services results and more patients with HIV seeking and remaining in care.10,11 In addition, US health care reform has given renewed emphasis on the “patient-centered medical home,” of which HIV MDCT (as in Ryan White Part C clinics) can be a paradigm.12,13 Typically, HIV MDCT comprises at least a physician, nurse and/or case manager, and often a clinical pharmacist.13,14 Other personnel cited as beneficial are nurse practitioners,15 mental health workers,16 social workers,17,18 dieticians,19 health educators,14 and transportation service consultants.13,20 Although some studies have shown benefit to MDCT,13,14,21,22 studies did not measure ART adherence, nor account for the contribution of MDCT components in their analyses.
Kaiser Permanente (KP) is one of the largest providers of integrated HIV care in the United States. Although KP endorses an HIV specialty and multidisciplinary model of HIV care, our clinic compositions are quite diverse and include some clinics where an MDCT is not present. The primary goal of our study was to explore which combinations of MDCT components are most associated with maximal ART adherence, when compared with the HIV/infectious disease (ID) specialist without an MDCT. We also sought to examine which MDCT combinations were associated with improvements in other HIV-related outcomes.
We performed a retrospective cohort analysis of all new ART regimen starts in KP California from 1996 to 2006 and followed through 12 months after initiation. We determined which components of the MDCT (clinician, clinical pharmacist, nurse manager, non-nurse care coordinator, dietician, social worker/benefits counselor, mental health worker, clinical health educator) were most influential on maximal ART adherence. We determined the optimized MDCT associated with maximal ART adherence. We also determined if the optimized MDCTs for adherence were associated with improvements in other HIV-related outcomes, including achieving HIV RNA BLQ, changes in CD4+ counts, and odds of new AIDS-defining events.
KP California is an integrated health care system serving 18.3% of the California population and 12.3% of its HIV-infected population. Patients in our system receive multidisciplinary health care, including HIV specialty care. The HIV-positive population in KP is demographically representative of the state23; data indicate that members overall are very similar to the general population with regard to age, gender, and race/ethnicity, with only slight underrepresentation of those in lower and higher income and education categories.24
We identified all HIV-positive patients at least 18 years old initiating a first or new ART regimen dispensed from a KP pharmacy in the years 1996–2006, where HIV RNA before starting the new regimen was above lower limits of quantification. We classified the patients associated with the ART regimen as antiretroviral naive if they had no record of any prior antiretroviral regimen; otherwise, new regimens were classified as among ART-experienced patients. We defined an ART regimen among ART-naive patients as ≥3 antiretroviral drugs used in combination as defined by the Department of Health and Human Services guidelines,9 with ritonavir at doses ≤400 mg/d not considered an active drug in the regimen. New regimens among ART-experienced patients were defined as ≥2 antiretroviral drugs not previously used. Eligible patients had ≥6 months KP membership before initiation of the first or new regimen for baseline values to be determined and presence of antiretroviral history to be established.
We surveyed each medical center in KP California to determine the presence of MDCT components of interest at that medical center in each year from 1996 to 2006. Components of interest were as follows: HIV or ID specialist (MD/DO, HIV nurse practitioner, HIV physician assistant), HIV nurse case manager, non-nurse care coordinator, clinical pharmacist, social worker (or benefits coordinator), dietician, health educator, or mental health worker. Components were counted as present if they had a formal assignment to the HIV clinic. We determined whether providers were an HIV specialist according to either HIV Medicine Association or American Academy of HIV Medicine definitions.25,26 Using the data collected in the survey, we identified the components of the MDCT available to a particular patient at the time of his or her ART regimen initiation.
KP pharmacy records provide details of each antiretroviral dispensed to a patient, which allowed us to assess date of first antiretroviral prescription fill and regimen composition, and also identify refills for patients during study follow-up. Adherence to the ART regimen over the 12-month observation period was calculated using established methods developed for administrative pharmacy databases. These methods account for all the component medicines of an individual patient's ART regimen.27–29 This measure of adherence is computed across all antiretroviral medications as the number of doses in an interval (bounded by a first and last fill date of drug, requiring at least 2 fills per drug) for which the patient has drug in possession as a percentage of total intended doses in the span between the first and last filling. The computation takes quantity supplied and frequency of dosing into account. Employing pharmacy records to ascertain ART adherence has been validated and widely used in previous studies at other institutions and KP.30–33 We classified ART regimens as nonnucleoside reverse transcriptase inhibitor (NNRTI) based, nucleoside reverse transcriptase inhibitor only, protease inhibitor (PI) based, PI + NNRTI, or new class.
We measured HIV RNA closest to and before regimen initiation and closest to 12 months after initiation. We defined BLQ as <500 copies per milliliter through 2000 and <75 copies per milliliter subsequently. We also recorded all clinical AIDS-defining events (1993 Centers for Disease Control and Prevention classification except CD4 count <200/μL)34 and all CD4+ counts at regimen initiation and through 12 months after initiation. In addition, we recorded gender, race/ethnicity, HIV risk behavior, coinfection with hepatitis C virus, and the year the first/new ART regimen was initiated.
We used recursive partitioning to determine potential MDCT combinations that had maximal ART adherence through 12 months, the a priori primary outcome for the study. We constructed a regression tree using standard classification and regression tree (CART) methodology, a commonly used tool for identifying interactions (in this case, the different components of MDCT and their combined effects on ART adherence), providing clues to data structure that might not be apparent from linear regression analysis.35,36 A regression tree is built through a process known as binary recursive partitioning, an iterative process of splitting groups of patients (parent node) into 2 groups (child nodes) on the basis of an independent variable (eg, MDCT component × present vs absent and its effect on percent 12-month adherence), and then further splitting each new node into 2 groups. For a given node, the tree algorithm considers every possible binary split on every variable under consideration, choosing the split that partitions the data into 2 parts such that it minimizes the sum of the squared deviations from the mean adherence in the separate parts. The partitioning is then applied to each of the new nodes, continuing until each node reaches a user-specified minimum node size (or if the sum of squared deviations from the mean within a node is 0), specified here as n = 5, thus becoming a terminal node. Each terminal node will have a unique combination of MDCT components (Fig. 1). Generally, this process results in very large trees suffering from overfitting (ie, it is “explaining” random elements of the data that are not likely to be features of the larger population). Using 10-fold cross-validation, the tree is “pruned” to obtain a set of mutually exclusive and exhaustive groups of patients characterized by different combinations of MDCT components, selected to minimize unexplained variance (sum of within-node variance across terminal nodes) and complexity (number of terminal nodes). In this analysis, for parsimony (ie, fewest number of terminal nodes), we chose the smallest tree within 1 standard error of the best-performing tree.
Mixed effects linear regression was used to examine the association between ART adherence and the terminal node MDCT combinations identified in the CART analyses, with clustering by medical center and provider and adjusting for ART experience, age, gender, race/ethnicity, HIV risk behavior, hepatitis C status, ART regimen class, and year of regimen initiation, compared with the HIV specialist only. We further analyzed only the MDCT teams found most associated with improved ART adherence. We employed mixed effects logistic regression to assess odds of achieving BLQ and odds of new AIDS-defining events (compared with the HIV specialist only) among these more limited MDCT combinations. Change in CD4+ count in the 12-month follow-up period among these more limited MDCT combinations was examined employing mixed effects linear regression models, using all available repeated measurements per patient during follow-up, again compared with specialist only. These analyses were adjusted for the same potential confounders as in the adherence analysis, with the addition of further clustering by patient in the longitudinal CD4+ count analyses. The CD4+ count analyses used linear splines, allowing for different slopes in the 2 periods of 0 to 90 days and the latter 270 days, in methodology previously employed by the authors.32 Analyses were conducted using CART Pro 6.0 (Salford Systems, San Diego, CA) and Stata10 (Stata Corp, College Station, TX).
We obtained approval from KP Institutional Review Boards, which waived the requirement for informed patient consent.
We analyzed data on 10,801 regimen starts among 9669 unique patients. Of those regimen starts, 7071 were first regimen starts among ART-naive patients and 3730 were new regimen starts among ART-experienced patients. A total of 1742 naive patients also contributed data as experienced patients when they started a subsequent new regimen. Characteristics of the patients at regimen start are shown in Table 1. Patients contributing data were predominantly male, men having sex with men, and a majority was Caucasian. The vast majority of patients were started on NNRTI-based or PI-based regimens. Consistent with KP practice, a large majority of patients were exposed to an HIV or ID specialist (Table 1). However, the HIV clinics were quite heterogeneous, with a wide variation of exposure to individual components of the MDCT.
The recursive partitioning results are shown in Figure 1. The first splitting variable was the clinical pharmacist. The eventual potential combinations (12 terminal nodes including HIV/ID specialist only) identified in the CART analysis are quite varied. For example, 2 of the potential MDCTs were clinical pharmacist or nurse case manager with primary care provider, and other combinations were a clinical pharmacist or nurse case manager along with a social worker or a non-nurse care coordinator and primary care provider.
Five MDCTs, when compared with the HIV specialist only, were found to have statistically significantly greater ART adherence with linear regression modeling (Table 2). The clinical pharmacist with primary care provider was associated with a mean 3.3% improved adherence (P = 0.01). In the modeling, the clinical pharmacist was the only single allied health professional with a physician found to have a statistically significantly improved adherence. The highest mean improved adherence was with the pharmacist plus non-nurse care coordinator plus primary care provider (8.1%, P = 0.003). Three other MDCT combinations with significantly improved adherence were nurse plus social worker/benefits coordinator plus primary care provider (7.5%, P < 0.001), pharmacist plus social worker/benefits coordinator plus primary care provider (5.7%, P < 0.001), and HIV/ID specialist plus mental health worker (6.5%, P = 0.001). Of note, these 5 combinations were not found to be statistically significantly different from each other (Wald test, P = 0.29).
We analyzed these 5 MDCT combinations for increased odds of achieving BLQ or new AIDS events in the first 12 months of follow-up (Table 3). Although 3 of the MDCTs in unadjusted analyses had significantly increased odds of BLQ compared with the specialist only, these results attenuated in adjusted analysis (Wald test: unadjusted P < 0.001, adjusted P = 0.08). We did not find any statistically significant association between these MDCTs and odds of new AIDS events in either unadjusted or adjusted analyses.
We performed a similar analysis for change in CD4+ count through 12 months (Table 4). There was no association between MDCTs and change in CD4+ count for either the 90-day period or the subsequent 270-day period (P = 0.19 for the 90-day period and P = 0.07 for the 270-day period). The MDCT with clinical pharmacist plus primary care provider had statistically significant increases compared with the specialist only in both time periods (P < 0.001 and P = 0.02, respectively). In adjusted analysis, statistically significant covariates associated with change in CD4+ count were younger age, later year, NNRTI class, and injection drug use (but injection drug user negatively associated).
This study uses a novel approach to analyzing the many components of the HIV MDCT on ART adherence. We determined that several MDCT combinations had a positive influence on adherence, compared with the HIV/ID specialist alone. Although we did not see a consistent improvement in odds of BLQ or new AIDS events among the groups found significantly associated with improved adherence compared with the HIV specialist only, we did see with the clinical pharmacist plus non–HIV-specialized primary care provider a significant increase in CD4+ counts through the 360 days of follow-up.
Our finding that team members other than the HIV specialist would have a significant impact on ART adherence is supported by prior studies indicating that doctors do not necessarily emphasize adherence in their patient interactions, including HIV care providers.37 Clinical pharmacists are trained to help patients manage and adhere to complex medication regimens.38,39 In an earlier study, we showed that clinical pharmacists improved adherence and decreased outpatient office visits among patients both antiretroviral naive and experienced patients.32,40 Our earlier research with HIV clinical pharmacists also indicated that there was an interaction with provider experience, prompting us to initiate this study.
The potential MDCT combinations shown here associated with improved adherence are consistent with the literature. Earlier studies indicate that nurses improve patient adherence knowledge.41 The nurse case manager and social worker have different scopes of practice, but each help address patient unmet needs, which should improve the likelihood of being adherent to medications. The same reasoning would apply to the clinical pharmacist plus the social worker or plus the non-nurse care coordinator. The HIV specialist and mental health worker could address different significant impacts on the patient's ability to adhere to treatment plans and regimens. Of course, another primary outcome (like care retention) may derive different results.
We did not specifically designate a case manager among our MDCT components of interest because many different personnel in KP function in that role, including the nurse, non-nurse care coordinator, social worker, clinical pharmacist, and others, depending on the clinic's structure. The ability of many different personnel to have the case manager role has made prior research in this area problematic.13,42–44 However, the activities of case management in a medical system, including improving access to and retention in care, treatment plan adherence, and meeting unmet patient needs,13,45 are likely represented by the personnel found to be associated with improved adherence in this study—obviating the need for a specifically designated case manager in all clinics.
There is a lack of prior research on health educators, dieticians, and social workers in the area of HIV MDCT and improved outcomes.13,14,18,19 We found improved adherence with the social worker when combined with both the nurse and the clinical pharmacist, demonstrating a likely synergy of roles. Given the many tasks that social workers provide (including arranging housing, transportation, insurance benefits, and medication availability), our results are not surprising. We were disappointed to find that neither dieticians nor health educators were specifically associated with improved adherence, but it is possible that their services could be associated with improving other HIV-related outcomes.
We acknowledge a few limitations to our study. A small percent of patients (<5%) do not receive their medications through our pharmacies. It is possible that their adherence results could be quite different, but this is unlikely. It is also possible that some patients use outside case management services and the impact of those services are not considered adequately, but this likelihood is quite low with our total integrated care system. As with all observational studies, there could be residual confounding, but we have accounted for the most significant factors in ART adherence and even account for provider-level and system-level factors, the contributions and interactions of which are rarely considered. Also, some patients could have self-selected which clinic to attend based on the MDCT composition, but most patients in KP select their clinic based on geography.
We likely see some confounding by indication in our results. This is especially true for mental health because patients with mental health issues were likely referred to the mental health specialist (and medical centers where they were located); patients with mental health issues have previously been shown to have lower adherence.46 There are parts of the CART analysis where mental health was associated with reduced adherence, although it is unlikely that mental health providers would have such a negative impact on adherence. In fact, our prior research has shown that HIV-positive patients with depression but on antidepression (selective serotonin reuptake inhibitor) medication had ART adherence similar to HIV-positive patients without a depression diagnosis.46 This will require further study but should not take away from the key findings of the study.
We employed a retrospective observational cohort design. We believe that this is justified because patient–MDCT component interaction is likely more than just assigned visit times (phone calls, “curbside” interactions as examples of unrecorded “visits”). Furthermore, it is likely (and probably desirable) that there is an ongoing “intercomponent” education, leading to improved practices by all team members, and not just that single component; this would not be captured if only designated visits were used in our analysis. Ideally, a prospective clinic-based trial, comparing the different MDCTs found significantly associated with improved adherence, should be the next step.
We accounted for common patient-level and medication-level factors associated with ART adherence, including age, gender, HIV risk, race/ethnicity, regimen class, and temporal trend.4 We have previously shown that more recent ART regimens may have a greater impact on adherence than provider experience.47,48 However, it is clear that these factors do not account for all the impact on ART adherence. Our work demonstrates that there is a significant and measurable impact of ecological (system of care) factors on patient-level HIV-related outcomes. There is little research in HIV or other chronic conditions in this area. Our results demonstrate that there is a great opportunity for care and health outcome improvement with such exploration. With US health care reform, improving patient outcomes is a necessity; and often, this starts with improved treatment adherence. Furthermore, health care reform places a strong emphasis on the patient-centered medical home.49 Optimized MDCT is a key element of the medical home model. We demonstrate options for such core parts of the medical home.
Our results have implications for the US National HIV/AIDS Strategy (NHAS), released in 2010.50 A goal of NHAS is to increase the number of HIV-positive Americans on ART and increase the percent with maximal viral control. Although clinicians will always be the MDCT member who ultimately determines which patients should be on medications and which ART regimen to use, our results indicate that other team members have integral roles in ensuring the treatment's success, including adherence and improved outcomes. However, in times of constrained revenues and expensive care (total HIV care costs can be >$24,000/yr),51 it is tempting to think that the solo practitioner or specialist is all that is needed. Our results would indicate otherwise. Furthermore, some of the MDCT combinations discovered here may be less costly per year than just the specialist alone who might experience a higher rate of failed regimens because of poor ART adherence.52
CART analysis can be successfully employed to help discover potential optimal care teams for adherence-related outcomes. We believe that the methodology applied here should be investigated for other HIV-related outcomes, especially those found in NHAS and improved health care quality. For example, NHAS calls for increased accessing care at time of HIV diagnosis and increased retention in care. The HIV MDCT could be optimized through this methodology for such desirable outcomes.
The authors thank Courtney Ellis for her assistance with manuscript preparation.
1. Press N, Tyndall MW, Wood E, et al.. Virologic and immunologic response, clinical progression, and highly active antiretroviral therapy adherence. J Acquir Immune Defic Syndr. 2002;31(suppl 3):S112–S117.
2. Simoni JM, Frick PA, Pantalone DW, et al.. Antiretroviral adherence interventions: a review of current literature and ongoing studies. Top HIV Med. 2003;11:185–198.
3. Gardner EM, McLees MP, Steiner JF, et al.. The spectrum of engagement in HIV care and its relevance to test-and-treat strategies for prevention of HIV infection. Clin Infect Dis. 2010;52:793–800.
4. Chesney MA. Factors affecting adherence to antiretroviral therapy. Clin Infect Dis. 2000;30(suppl 2):S171–S176.
5. Grzywacz JG, Fuqua J. The social ecology of health: leverage points and linkages. Behav Med. 2000;26:101–115.
6. Chesney MA, Morin M, Sherr L. Adherence to HIV combination therapy. Soc Sci Med. 2000;50:1599–1605.
7. Green LW, Richard L, Potvin L. Ecological foundations of health promotion. Am J Health Promot. 1996;10:270–281.
8. Richard L, Potvin L, Kishchuk N, et al.. Assessment of the integration of the ecological approach in health promotion programs. Am J Health Promot. 1996;10:318–328.
9. Department of Health and Human Services. Panel on Antiretroviral Guidelines for Adult and Adolescents. Guidelines for the use of antiretroviral agents in HIV-1-infected adults and adolescents. Department of Health and Human Services; 2011:1–128.
10. Ashman JJ, Conviser R, Pounds MB. Associations between HIV-positive individuals' receipt of ancillary services and medical care receipt and retention. AIDS Care. 2002;14(suppl 1):S109–S118.
11. Messeri PA, Abramson DM, Aidala AA, et al.. The impact of ancillary HIV services on engagement in medical care in New York City. AIDS Care. 2002;14(suppl 1):S15–S29.
12. Hoang T, Goetz MB, Yano EM, et al.. The impact of integrated HIV care on patient health outcomes. Med Care. 2009;47:560–567.
13. Sherer R, Stieglitz K, Narra J, et al.. HIV multidisciplinary teams work: support services improve access to and retention in HIV primary care. AIDS Care. 2002;14(suppl 1):S31–S44.
14. Le CT, Winter TD, Boyd KJ, et al.. Experience with a managed care approach to HIV infection: effectiveness of an interdisciplinary team. Am J Manag Care. 1998;4:647–657.
15. Wilson IB, Landon BE, Hirschhorn LR, et al.. Quality of HIV care provided by nurse practitioners, physician assistants, and physicians. Ann Intern Med. 2005;143:729–736.
16. Berg MB, Mimiaga MJ, Safren SA. Mental health concerns of HIV-infected gay and bisexual men seeking mental health services: an observational study. AIDS Patient Care STDS. 2004;18:635–643.
17. Olivier C, Dykeman M. Challenges to HIV service provision: the commonalities for nurses and social workers. AIDS Care. 2003;15:649–663.
18. Cox LE. Social support, medication compliance and HIV/AIDS. Soc Work Health Care. 2002;35:425–460.
19. Brunner RL, Larson TA, Scott BJ, et al.. Evaluation of the impact and acceptance of a nutrition program in an HIV community clinic. AIDS Patient Care STDS. 2001;15:533–543.
20. Johnson RL, Botwinick G, Sell RL, et al.. The utilization of treatment and case management services by HIV-infected youth. J Adolesc Health. 2003;33(2 suppl):31–38.
21. Levy RW, Rayner CR, Fairley CK, et al.. Multidisciplinary HIV adherence intervention: a randomized study. AIDS Patient Care STDS. 2004;18:728–735.
22. Sorensen JL, Mascovich A, Wall TL, et al.. Medication adherence strategies for drug abusers with HIV/AIDS. AIDS Care. 1998;10:297–312.
23. California Department of Health Services. Office of AIDS. California AIDS Surveillance Report: Cumulative Cases as of December 31, 2005. Sacramento, CA: California Department of Health Services; 2006.
24. Krieger N. Overcoming the absence of socioeconomic data in medical records: validation and application of a census-based methodology. Am J Public Health. 1992;82:703–710.
25. HIV Medicine Association. Qualifications for physicians who care for patients with HIV infection. Available at: http://www.hivma.org/Content.aspx?id=1782
. Accessed October 30, 2008.
26. Grossman HA. Addressing the need for HIV specialists: the AAHIVM perspective. AIDS Read. 2006;16:479–486.
27. Sikka R, Xia F, Aubert RE. Estimating medication persistency using administrative claims data. Am J Manag Care. 2005;11:449–457.
28. Steiner JF, Koepsell TD, Fihn SD, et al.. A general method of compliance assessment using centralized pharmacy records. Description and validation. Med Care. 1988;26:814–823.
29. 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.
30. Fairley CK, Permana A, Read TR. Long-term utility of measuring adherence by self-report compared with pharmacy record in a routine clinic setting. HIV Med. 2005;6:366–369.
31. Seguy N, Diaz T, Campos DP, et al.. Evaluation of the consistency of refills for antiretroviral medications in two hospitals in the state of Rio de Janeiro, Brazil. AIDS Care. 2007;19:617–625.
32. Horberg MA, Hurley LB, Silverberg MJ, et al.. Effect of clinical pharmacists on utilization of and clinical response to antiretroviral therapy. J Acquir Immune Defic Syndr. 2007;44:531–539.
33. Silverberg MJ, Leyden W, Horberg MA, et al.. Older age and the response to and tolerability of antiretroviral therapy. Arch Intern Med. 2007;167:684–691.
34. Center for Disease Control and Prevention. 1993 revised classification system for HIV infection and expanded surveillance case definition of AIDS among adolescents and adults. MMWR Recomm Rep. 1992;41:1–19.
35. Breiman L, Friedman JH, Olshen RA, et al.. Classification and Regression Trees. Monterey, CA: Wadsworth & Brooks/Cole Advanced Books & Software; 1984.
36. Zhang H, Singer B. Recursive Partitioning
in the Health Sciences. New York, NY: Springer Verlag; 1999.
37. Golin CE, Smith SR, Reif S. Adherence counseling practices of generalist and specialist physicians caring for people living with HIV/AIDS in North Carolina. J Gen Intern Med. 2004;19:16–27.
38. Geletko SM, Poulakos MN. Pharmaceutical services in an HIV clinic. Am J Health Syst Pharm. 2002;59:709–713.
39. Rathbun RC, Farmer KC, Stephens JR, et al.. Impact of an adherence clinic on behavioral outcomes and virologic response in treatment of HIV infection: a prospective, randomized, controlled pilot study. Clin Ther. 2005;27:199–209.
40. Horberg M, Silverberg, MJ, Hurley, LB, et al. impact of HIV clinical pharmacists on antiretroviral adherence and clinical outcomes in ARV experienced patients [Abstract TuPe0107]. Paper presented at: XVI International AIDS Conference; 2006; Toronto, Canada.
41. Holzemer WL. HIV and AIDS: the symptom experience. What cell counts and viral loads won't tell you. Am J Nurs. 2002;102:48–52.
42. Conover CJ, Whetten-Goldstein K. The impact of ancillary services on primary care use and outcomes for HIV/AIDS patients with public insurance coverage. AIDS Care. 2002;14(suppl 1):S59–S71.
43. Lo W, MacGovern T, Bradford J. Association of ancillary services with primary care utilization and retention for patients with HIV/AIDS. AIDS Care. 2002;14(suppl 1):S45–S57.
44. London AS, LeBlanc AJ, Aneshensel CS. The integration of informal care, case management and community-based services for persons with HIV/AIDS. AIDS Care. 1998;10:481–503.
45. Katz MH, Cunningham WE, Fleishman JA, et al.. Effect of case management on unmet needs and utilization of medical care and medications among HIV-infected persons. Ann Intern Med. 2001;135(8 pt 1):557–565.
46. Horberg MA, Silverberg MJ, Hurley LB, et al.. Effects of depression and selective serotonin reuptake inhibitor use on adherence to highly active antiretroviral therapy and on clinical outcomes in HIV-infected patients. J Acquir Immune Defic Syndr. 2008;47:384–390.
47. Horberg M, Hurley L, Towner W, et al. Impact of provider experience characteristics on HIV-related outcomes among cART naive [Abstract 1131]. Paper presented at: Infectious Disease Society of America 48th Annual Meeting; 2010; Vancouver, Canada.
48. Horberg M, Hurley L, Towner W, et al. Impact of provider characteristics on HIV-related outcomes among antiretroviral experienced patients [Abstract MOPE464]. Paper presented at: 6th IAS Conference on HIV Pathogenesis, Treatment and Prevention; 2011; Rome, Italy.
49. Arvantes J. Health care reform legislation will drive adoption of medical home projects, officials say. Available at: http://www.aafp.org/online/en/home/publications/news/news-now/professional-issues/20100805pcpccstakeholders.html
. Accessed August 3, 2011.
50. The White House Office of National AIDS Policy. National HIV/AIDS strategy for the United States. The White House Office of National AIDS Policy; 2010.
51. Meenan R, O'Keeffe Rosetti M, Kimes T, et al. Excess prevalence-based costs of multiple HAART switches among HIV+ patients in a US health maintenance organization [Abstract TUPE0236]. Paper presented at: XVII International AIDS Conference; 2006; Mexico City, Mexico.
52. Salary.com. National average salaries by profession. Available at: http://www.salary.com/mysalary.asp
. Accessed July 30, 2011.