Readmission to a hospital within 30 days is common in the general US population. For adults under 65 years of age, the 30-day readmission rate for all causes has been estimated at 13.3% (extrapolated from Wier et al. 2008) , and for Medicare beneficiaries 65 and older, at 19.5% . The rate of readmission among persons living with HIV (PLWH) may be relatively high given the multiple complications of HIV, potentially toxic drug regimens, and social marginalization of many PLWH. Data on 30-day readmissions in HIV are limited to a single-site study, which reported a rate during 2006–2008 of 25.2% .
Thirty-day readmission rates are increasingly becoming a benchmark for hospital quality of care and reimbursement [4–7]. A basis of comparison specific to HIV will be important to clinicians, hospital administrators, and policymakers attempting to create best-practice readmission targets for populations which include substantial numbers of PLWH.
Understanding factors associated with readmission among PLWH may indicate ways to improve care and reduce readmissions. Of particular interest are the reason for the initial (index) hospitalization, CD4+ cell count, and use of antiretroviral therapy (ART) at discharge. Also of interest is having a postdischarge clinic follow-up visit. Such visits have been associated with decreased risk of readmission among (HIV-uninfected) patients with chronic obstructive pulmonary disease (COPD) and congestive heart failure (CHF) [8,9].
This study has two aims. The first is to describe the 30-day readmission rate in a multicenter HIV cohort. The second is to evaluate associations between readmission and clinical factors of interest.
Patient population and data collection
The HIV Research Network (HIVRN) is a research collaboration that includes 12 sites providing longitudinal HIV care to adult patients in 11 US cities . Sites collect demographic, laboratory, inpatient, and outpatient utilization data, strip these data of identifying characteristics, and submit them to a coordinating center, where they are reviewed and combined. Ethical review boards at each site and at the coordinating center have approved the collection and use of these data.
Nine HIVRN sites for which no readmission data has previously been published were included in this study. These sites collect complete hospitalization data, including International Classification of Diseases, Ninth Revision (ICD-9) discharge codes. Six of these sites had data for 2005–2010, and three had only 2010 data. After preliminary analyses identified no univariate or multivariate associations between readmission and calendar year, data from these three sites were combined with the initial six sites in all final analyses. The nine sites are located in the northeast (3), west (3), south (2), and midwest (1) regions of the United States.
Index hospitalizations and 30-day readmissions
Following the Centers for Medicare and Medicaid Services’ methodology, index hospitalizations were defined by the following criteria: being a first-ever hospitalization or being any hospitalization which occurs more than 30 days after the most recent previous hospitalization, and having a live discharge . In other words, most hospitalizations are index hospitalizations; the exceptions are those which are, themselves, 30-day readmissions and those which end in death.
Because of the need to observe 30 days before a potential index hospitalization, hospital admissions prior to 1 February of the first year of hospitalization data were excluded as potential index hospitalizations. Because of the need to observe for 30-day readmissions, index hospitalizations with discharges after 30 November 2010 were excluded.
When an index hospitalization was followed by a ‘chain’ of multiple readmissions, each with less than 30-day intervals (the overall span of the chain may have exceeded 30 days), all of the readmissions in the chain were excluded from being index hospitalizations.
Same day readmissions were considered readmissions.
Patients were in active care if they had a visit to the HIV clinic and a CD4+ cell test recorded in a calendar year. Because patients not in active care may have received care at other centers, any available hospitalization records for such patient years (n = 1042) were excluded. Patients could resume active care after a period out of care. Hospitalizations for clinical trials (ICD9 V70.7; n = 87), for rehabilitation (V57.8-V57.9; n = 32), and with illogical dates (e.g. beginning within the span of another hospitalization; n = 9) were excluded.
Factors evaluated for association with readmission included year of index hospitalization, geographic region, age, gender, race/ethnicity, injection drug use (IDU, either alone or in combination with other HIV risk factors), CD4+ cell count, HIV RNA less than 400 copies/ml, ART (defined as one or more antiretroviral medicines) at discharge, length of stay (LOS), primary insurance, diagnostic category for index hospitalization, and outpatient follow-up. For CD4+ cell count, HIV RNA, and insurance category, the first values within the year of hospitalization were used. We hypothesized that ART use in the 30 days after discharge may increase readmission risk either through direct adverse effects, through drug interactions, or because of an immune reconstitution inflammatory syndrome (IRIS). ART at discharge was categorized as none, discharge on ART after being admitted on ART, and discharge on ART when not admitted on ART (felt to be the highest risk scenario for toxicity or IRIS). In the second case, the ART regimen need not have been the same on admission and discharge; in the third case, patients may have previously used ART but were not using it on the admission date.
Assigning each hospitalization to a single diagnostic category was done using previously published methods [12–14]. In brief, the first listed ICD-9 code referring to neither HIV nor chronic hepatitis C infection was defined as the primary code. This code was assigned to one of 18 categories using multilevel Clinical Classification Software from the Agency for Healthcare Research and Quality . The classification was modified in two ways. First, infections were reassigned from organ system categories to the infection category (for example, pneumonia was reassigned from pulmonary to infection). Second, an AIDS-defining illness (ADI) category was constructed from ICD-9 codes according to the 1993 Centers for Disease Control and Prevention criteria . Recurrent bacterial pneumonia was defined as any bacterial pneumonia hospitalization occurring within more than 30 days, but 365 or less days of a previous such hospitalization.
For each index hospitalization, a dichotomous dependent variable indicated whether any readmission occurred within 30 days or less. To estimate associations with independent variables, we used repeated measures logistic models adjusted for clustering of multiple index hospitalizations within patient using generalized estimating equations with robust variance estimators.
Supplementary analysis examining the effect of outpatient visits
Including an indicator of any outpatient visit prior to readmission was problematic for the logistic regression analysis as it could introduce time-dependent bias (patients who were readmitted within a few days would have had little chance for a visit). Cox proportional hazard models, with time to outpatient visit treated as a time-dependent covariate, can estimate an association without introducing this bias [2,8]. Time to first outpatient visit (primary HIV or specialty care) that occurred within 30 days, but prior to any readmission, was determined for each index hospitalization. Outpatient visits that occurred on the same day as a readmission were excluded on the premise that the patient appeared for the visit in need of hospitalization. For analysis, outpatient visits that occurred on the day of discharge from an index hospitalization were assigned a follow-up time of 0.1 day; same day readmissions were assigned 0.25 day. We fit Cox models with variance adjusted for clustering of multiple index hospitalizations within patient. However, the proportional hazard assumption was violated (Schoenfeld residuals P < 0.05) for four variables: year of hospitalization, CD4+ cell count, LOS, and diagnostic category. In order to obtain a valid, fully adjusted estimate, the multivariate Cox model was stratified by each of these four variables. The full, stratified multivariate model results are presented in a Supplementary Table, http://links.lww.com/QAD/A350.
In all analyses, a two-sided type 1 error of 5% was considered statistically significant. Analyses were performed using Stata 12.1 (StataCorp LP, College Station, Texas, USA) .
A total of 19 943 adult patients were in active care for at least 1 year during 2005–2010 at the nine study sites. A total of 5536 patients had at least one index hospitalization, of whom 3926 (71%) were never readmitted within 30 days and 1610 (29%) were readmitted. Patients were more likely to be readmitted within 30 days if they had more index hospitalizations, were female (vs. male), black (vs. white), or had a history of IDU (vs. MSM; Table 1).
Among the full patient sample, 15 951 total hospitalizations occurred in 2005–2010. Of these hospitalizations, 303 (1.9%) were not counted as index hospitalizations because they occurred in January of the first analytic year, 224 (1.4%) because discharge occurred after 30 November 2010, and 142 (0.9%) because of death during hospitalization. Of the remaining 15 282 hospitalizations, 11 651 (75.5%) were index hospitalizations; 2252 (14.6%) were first readmissions within 30 days; and 1379 (8.9%) were subsequent readmissions in a chain.
Of 11 651 index hospitalizations, 2252 were followed by a readmission within 30 days, yielding an overall readmission rate of 19.3%. When readmission occurred, 69% of instances consisted of a single readmission, 18% a chain of two readmissions, and 13% a chain of three or more readmissions. The median time to first readmission was 11 days (interquartile range, IQR, 5–19).
Table 2 lists time-varying patient demographic and clinical factors for the 11 651 index hospitalizations. Thirty-day readmission was more likely when patients had lower CD4+ cell counts (occurring 26.1% of the time at CD4+ cell count ≤50 cells/μl), HIV RNA at least 400 copies/ml, longer LOS (26.7% when LOS ≥9 days), Medicaid insurance, and were discharged on ART but not admitted on ART.
Figure 1 shows 30-day readmission rates associated with the 10 most frequent diagnostic categories for index hospitalizations. Oncologic hospitalizations had the highest readmission rate at 27.8%; however, oncologic hospitalizations comprised only 3.6% of index hospitalizations. ADI had a readmission rate of 26.2%, while comprising 9.6% of index hospitalizations. Non-ADI infection had among the lowest readmission rates (16.6%) but was the most frequent category (26.4% of index hospitalizations).
Table 3 shows univariate and multivariate logistic regression results. Increasing readmission risk was consistently seen with progressively lower CD4+ cell counts. In the multivariate model, compared to CD4+ cell count at least 351, adjusted odds ratios (AORs) were 1.19 (1.04–1.37), 1.42 (1.24–1.64), and 1.80 (1.53–2.11) for CD4+ cell count 201–350, 51–200, and 50 cells/μl or less, respectively. Progressively longer LOS exhibited a similar pattern of increasing readmission risk. ADI was significantly associated in both univariate [OR vs. non-AIDS-defining infection 1.77 (1.51–2.09)] and multivariate [1.40 (1.17–1.66)] analyses. Compared to hospitalizations covered by Medicaid, uninsured hospitalizations had lower readmission risk in both models. No independent associations were seen for age, gender, race, IDU, geographic region, HIV RNA level, and antiretroviral use.
Table 4 shows the characteristics of the 2252 index hospitalization/first readmission pairs. Overall, readmissions occurred in the same diagnostic category as the index hospitalization 35.7% of the time. This varied from 21.9% for kidney/genitourinary to 46.2% for oncologic (Table 4, italicized cells). Non-AIDS-defining infection and ADI were the most common readmission categories when all index hospitalizations were combined.
Supplementary analysis examining the effect of outpatient visits
An outpatient visit within 30 days of discharge and prior to any readmission occurred after 7730 (66%) index hospitalizations. The median time to these visits was 7 days (IQR, 4–13). Having a visit within 30 days was not associated with readmission in a univariate [hazard ratio 1.00 (0.90–1.10)] or in a fully adjusted [0.98 (0.88–1.08)] Cox model (complete model results in the Supplementary Table, http://links.lww.com/QAD/A350). The results for other variables evaluated in Cox regressions were similar to the logistic results.
The 30-day hospital readmission rate among persons engaged in outpatient HIV care in our multisite cohort was 19.3%. Some diagnostic categories substantially exceeded the overall rate, with ADI as high as 26.2%. Factors independently associated with readmission included lower CD4+ cell count and longer LOS. ART use and postdischarge outpatient follow-up were not associated with readmission. The majority of readmissions were for different diagnoses than their associated index hospitalizations.
The rate of 19.3% exceeds the rate of 13.3% for adults 18–64 years old extrapolated from the findings of the 2008 Healthcare Cost and Utilization Project (HCUP, which included all non-Federal hospitalizations in 15 US states) and nearly equals the rate (19.5%) among Medicare beneficiaries aged 65 years and older [1,2]. The median age (IQR) in our cohort was 46 (40–53) years.
The rate of 19.3% is lower than the 25.2% rate reported from a large HIV clinic in Dallas . One reason for the different rates could be differences in health status. The Dallas population had more index hospitalizations occurring in the setting of CD4+ cell count less than 50 cells (31 vs. 18%) and more index hospitalizations due to ADI (41 vs. 9.6%).
The independent associations with lower CD4+ cell count and with ADI indicate that advanced HIV disease contributes more to readmission risk than do demographic factors or complications of ART. Having a low CD4+ cell count predisposes to frequent illnesses, both infectious and noninfectious [12,18–23], and an association between lower CD4+ cell count and hospital readmission has been seen in other studies [3,24,25]. From a public perspective, HIV diagnosis, engagement in care, and use of ART at high CD4+ cell counts may reduce future readmission rates. From the perspective of a patient with a low CD4+ cell count who has an index hospitalization, immune reconstitution with ART can be expected to reduce future hospitalizations. However, even with good adherence to ART, 30 days may be too short to see an effect. Decreasing the incidence of illnesses probably requires several months [13,26,27], and the most relevant readmission window for assessing quality of care may, thus, be longer than 30 days.
A long LOS potentially indicates a more complicated index hospitalization. Similar to our findings, longer LOS was associated with readmission in the general Medicare population .
The association between having no insurance and lower readmission risk resembles findings in the 2008 HCUP analysis and may reflect economic barriers to seeking any medical care, including presenting to an emergency room or clinic for potential readmission .
Outpatient follow-up visits provide an opportunity to review discharge medications and response to treatment. The lack of association between outpatient clinic follow-up and lower readmission risk is surprising in light of recent studies that showed such associations among Medicare beneficiaries over 65 with COPD and CHF [8,9]. The younger age of the HIVRN population may explain the difference in findings. Similar to our results, several general population studies have failed to detect a protective effect of early outpatient follow-up, although these studies have generally examined longer intervals (90, 180, and 365 days) after index discharge [28–31]. Confirmatory studies among PLWH are needed. Because outpatient visits may have been preferentially offered to the sickest patients or to those most likely to keep their appointments, prospective trials of increasing outpatient follow-up would provide more reliable evidence.
The majority of readmissions occurred in different diagnostic categories than their associated index hospitalizations. Hence, readmissions may tend to be from issues that present after discharge, or, if signs are present during the index hospitalization, they may not be the focus of attention.
Clinicians may use our findings in several ways. First, the readmission rate in our multicenter cohort may serve as a basis of comparison for assessing quality of inpatient care of PLWH. Second, clinicians may target readmission prevention efforts toward patients with CD4+ cell count less than 200, with LOS more than 5 days, and for ADI and other high-risk reasons for readmission. Third, increasing early outpatient follow-up visits requires further investigation as a means of reducing readmissions. Other interventions, not assessed in the present study, have also shown promise in the general population and should be evaluated among PLWH. These include patient education, patient navigators, medicine reconciliation, and nurse home visits [32–35].
Hospital administrators and healthcare policymakers may use our findings when estimating expected readmission rates at hospital and regional levels. Compared to hospitals that serve few PLWH, hospitals serving more PLWH may have an increased overall readmission rate. Accurate calculation of expected rates is important. The Affordable Care Act has mandated public reporting and payment penalties for hospitals with higher-than-expected readmission rates for common diagnoses among Medicare enrollees over 65 years of age [4–6,11,36]. Readmissions among PLWH will have increasing impact on hospitals’ Medicare readmission rates as PLWH, already with a median age of 50 , grow older. The effect will also be highly relevant if public reporting and payment penalties are expanded to new groups, for example to Medicaid enrollees.
Our study is limited by lack of a matched control group of HIV-uninfected individuals. It is possible that our readmission rate exceeds that found for 18–64-year-olds in the HCUP analysis because our findings reflect high local readmission patterns in the cities where our sites are located. However, there are several reasons to doubt this. First are the strong associations with low CD4+ cell count and with ADI. Second, the interquartile ratio for variation nationwide in Medicare readmission rates between hospital referral regions was only 1.10 , much smaller than the ratio of 1.45 (19.3/13.3%), which exists between our rate and the HCUP rate. Third, though not nationally representative, our sample includes sites in eight geographically distributed cities. These reasons notwithstanding, the implications for public reporting and payment penalties make confirmation of our findings by direct comparison to HIV-uninfected controls important.
Another limitation is the lack of hospitalization data on PLWH who are not engaged in outpatient HIV care. Statewide or national databases may be able to account for such persons. A final limitation is that although our sites make attempts to collect hospitalization data from neighboring hospitals, this may be incomplete. However, an analysis of statewide insurance claims at one HIVRN site revealed that 91% of all hospitalizations occurred at the home hospital . In the Dallas study, 78% of readmissions occurred at the home hospital . Hence, although we feel we are capturing the vast majority of readmissions, our rate may slightly underestimate the true readmission rate.
In summary, we have found a high rate of 30-day readmission among PLWH in our multisite cohort. While awaiting data from additional cohorts, HIV providers may use the 19.3% rate as a preliminary basis of comparison for their practices. Policymakers may note this rate when considering adjustment for HIV in estimating expected hospital readmission rates used in public reporting and reimbursement calculations. The associations of readmission with low CD4+ cell count and ADI, the lack of association with outpatient follow-up visits or with discharge on ART, and the finding that patients more-often-than-not are readmitted for a different cause than for the index hospitalization, together suggest that the majority of readmissions among PLWH are due to the high frequency of illnesses inherently associated with advanced HIV disease and may not be the result of failed treatment plans or medication errors. Prevention of advanced HIV disease and/or recovery from it may have the largest impact in reducing readmissions.
S.A.B. and J.A.F. contributed to study design, data analysis, and article writing. B.R.Y. and A.L.A. contributed to study design and article writing. P.T.K. and R.D.M. contributed to data collection and study design. K.A.G. contributed to data collection, study design, and article writing. All authors have read and approved the final article.
The HIV Research Network (HIVRN) sites (participating sites marked with an *) are as follows:
Alameda County Medical Center, Oakland, California (Howard Edelstein, MD)*; Children's Hospital of Philadelphia, Philadelphia, Pennsylvania (Richard Rutstein, MD); Community Health Network, Rochester, New York (Roberto Corales, DO)*; Drexel University, Philadelphia, Pennsylvania (Jeffrey Jacobson, MD, Sara Allen, CRNP); Fenway Health, Boston, Massachusetts (Stephen Boswell, MD); Johns Hopkins University, Baltimore, Maryland (Kelly Gebo, MD, Richard Moore, MD, Allison Agwu MD)*; Montefiore Medical Group, Bronx, New York (Robert Beil, MD)*; Montefiore Medical Center, Bronx, New York (Lawrence Hanau, MD); Oregon Health and Science University, Portland, Oregon (P. Todd Korthuis, MD)*; Parkland Health and Hospital System, Dallas, Texas (Ank Nijhawan, M.D., Muhammad Akbar, MD); St Jude's Children's Hospital and University of Tennessee, Memphis, Tennessee (Aditya Gaur, MD); St Luke's Roosevelt Hospital Center, New York, New York (Victoria Sharp, MD, Stephen Arpadi, MD)*; Tampa General Healthcare, Tampa, Florida (Charurut Somboonwit, MD)*; University of California, San Diego, California (W. Christopher Mathews, MD)*; Wayne State University, Detroit, Michigan (Jonathan Cohn, MD)*.
Sponsoring agencies are as follows:
Agency for Healthcare Research and Quality, Rockville, Maryland (Fred Hellinger, PhD, John Fleishman, PhD, Irene Fraser, PhD); Health Resources and Services Administration, Rockville, Maryland (Robert Mills, PhD, Faye Malitz, MS).
Data coordinating center is as follows:
Johns Hopkins University (Richard Moore, MD, Jeanne Keruly, CRNP, Kelly Gebo, MD, Cindy Voss, MA).
The views expressed in this paper are those of the authors. No official endorsement by the Department of Health and Human Services, the National Institutes of Health, or the Agency for Healthcare Research and Quality is intended or should be inferred.
This work was supported by the Agency for Healthcare Research and Quality (HHSA290201100007C to R.D.M. and K.A.G.); the Health Resources and Services Administration (HHSH250201200008C to K.A.G.); and the National Institutes of Health (K23 AI084854 to SAB, K23 MH097647 to B.R.Y., and K23 AI084549 to A.L.A.).
Conflicts of interest
S.A.B. is supported by the National Institutes of Health (NIH, K23 AI084854); B.R.Y. is supported by the NIH (K23 MH097647); A.L.A. is supported by the NIH (K23 AI084549) and has received payment for lecturing at the University of New Mexico; R.D.M. is supported by the Agency for Healthcare Research and Quality (AHRQ, HHSA290201100007C); K.A.G. is supported by AHRQ (HHSA290201100007C) and the Health Resources and Services Administration (HHSH250201200008C), has been a consultant and received research support from Tibotec, has been a consultant for Bristol Myers Squibb, and has received payment for expert testimony to the US government; J.A.F. and P.T.K. have none to declare.
2. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program
. N Engl J Med
3. Nijhawan AE, Clark C, Kaplan R, Moore B, Halm EA, Amarasingham R. An electronic medical record-based model to predict 30-day risk of readmission and death among HIV-infected inpatients
. J Acquir Immune Defic Syndr
5. Axon RN, Williams MV. Hospital readmission as an accountability measure
6. Berenson RA, Paulus RA, Kalman NS. Medicare's readmissions-reduction program: a positive alternative
. N Engl J Med
7. Goodman DC, Fisher ES, Chiang-Hua C. After hospitalization: a Dartmouth atlas report on postacute care for Medicare beneficiaries. The Dartmouth Institute for Health Policy and Clinical Practice 2011. www.dartmouthatlas.org
[Accessed 28 September 2011].
8. Sharma G, Kuo YF, Freeman JL, Zhang DD, Goodwin JS. Outpatient follow-up visit and 30-day emergency department visit and readmission in patients hospitalized for chronic obstructive pulmonary disease
. Arch Intern Med
9. Hernandez AF, Greiner MA, Fonarow GC, Hammill BG, Heidenreich PA, Yancy CW, et al. Relationship between early physician follow-up and 30-day readmission among Medicare beneficiaries hospitalized for heart failure
10. HIV Research Network. Hospital and outpatient health services utilization among HIV-infected patients in care in 1999. J Acquir Immune Defic Syndr
11. Bernheim SM, Grady JN, Spivack S, Bartczak K. 2012 measures maintenance technical report: acute myocardial infarction, heart failure, and pneumonia 30-day risk-standardized readmission
measures – report prepared for the centers for Medicare and Medicaid services 2012. http://www.qualitynet.org
[Accessed 19 November 2012]
12. Berry SA, Fleishman JA, Moore RD, Gebo KA. HIV Research Network. Trends in reasons for hospitalization in a multisite United States cohort of persons living with HIV, 2001–2008
. J Acquir Immune Defic Syndr
13. Berry SA, Manabe YC, Moore RD, Gebo KA. Hospitalization risk following initiation of highly active antiretroviral therapy
. HIV Med
14. Gebo KA, Diener-West M, Moore RD. Hospitalization rates in an urban cohort after the introduction of highly active antiretroviral therapy
. J Acquir Immune Defic Syndr
16. Centers for Disease Control and Prevention1993 Revised Classification System for HIV Infection and Expanded Surveillance Case Definition for AIDS among Adolescents and Adults
. MMWR Recomm Rep
17. StataCorp. Stata statistical software: Release 12.1; 2012.
18. Hasse B, Ledergerber B, Furrer H, Battegay M, Hirschel B, Cavassini M, et al. Morbidity and aging in HIV-infected persons: the Swiss HIV cohort study
. Clin Infect Dis
19. Crum-Cianflone NF, Grandits G, Echols S, Ganesan A, Landrum M, Weintrob A, et al. Trends and causes of hospitalizations among HIV-infected persons during the late HAART era: what is the impact of CD4 counts and HAART use?
. J Acquir Immune Defic Syndr
20. Lichtenstein KA, Armon C, Buchacz K, Chmiel JS, Buckner K, Tedaldi EM, et al. Low CD4+ T cell count is a risk factor for cardiovascular disease events in the HIV outpatient study
. Clin Infect Dis
21. Kalayjian RC, Lau B, Mechekano RN, Crane HM, Rodriguez B, Salata RA, et al. Risk factors for chronic kidney disease in a large cohort of HIV-1 infected individuals initiating antiretroviral therapy in routine care
22. Petoumenos K, Worm SW, Fontas E, Weber R, De Wit S, Bruyand M, et al. Predicting the short-term risk of diabetes in HIV-positive patients: the data collection on adverse events of anti-HIV drugs (D:A:D) study
. J Int AIDS Soc
23. Yong MK, Elliott JH, Woolley IJ, Hoy JF. Low CD4 count is associated with an increased risk of fragility fracture in HIV-infected patients
. J Acquir Immune Defic Syndr
24. Nosyk B, Sun H, Li X, Palepu A, Anis AH. Highly active antiretroviral therapy and hospital readmission: comparison of a matched cohort
. BMC Infect Dis
25. Palepu A, Sun H, Kuyper L, Schechter MT, O'Shaughnessy MV, Anis AH. Predictors of early hospital readmission in HIV-infected patients with pneumonia
. J Gen Intern Med
26. Bosch RJ, Wang R, Vaida F, Lederman MM, Albrecht MA. Changes in the slope of the CD4 cell count increase after initiation of potent antiretroviral treatment
. J Acquir Immune Defic Syndr
27. Chu C, Selwyn PA. An epidemic in evolution: the need for new models of HIV care in the chronic disease era
. J Urban Health
28. Gill JM, Mainous AG 3rd, Nsereko M. Does having an outpatient visit after hospital discharge reduce the likelihood of readmission?
. Del Med J
29. Weinberger M, Oddone EZ, Henderson WG. Does increased access to primary care reduce hospital readmissions? Veterans Affairs Cooperative Study Group on Primary Care and Hospital Readmission
. N Engl J Med
30. Einstadter D, Cebul RD, Franta PR. Effect of a nurse case manager on postdischarge follow-up
. J Gen Intern Med
31. Piette JD, Moos RH. The influence of distance on ambulatory care use, death, and readmission following a myocardial infarction
. Health Serv Res
32. Hansen LO, Young RS, Hinami K, Leung A, Williams MV. Interventions to reduce 30-day rehospitalization: a systematic review
. Ann Intern Med
33. Naylor MD, Brooten D, Campbell R, Jacobsen BS, Mezey MD, Pauly MV, Schwartz JS. Comprehensive discharge planning and home follow-up of hospitalized elders: a randomized clinical trial
34. Naylor MD, Brooten DA, Campbell RL, Maislin G, McCauley KM, Schwartz JS. Transitional care of older adults hospitalized with heart failure: a randomized, controlled trial
. J Am Geriatr Soc
35. Coleman EA, Parry C, Chalmers S, Min SJ. The care transitions intervention: results of a randomized controlled trial
. Arch Intern Med
36. Report to congress. Medicare Payment Policy
, 2007; Washington, DC: Medicare Payment Advisory Committee.