Current trends in HIV-related hospital admissions and their impact on hospital resource utilization in Ontario
Meaghen J. Hyland,* BSc
Geri Bailey*
Mustafa Rawji,* BSc
Leslie Lee Peck†
Robert H. Hyland,* MD
Charles K. Chan,*† MD
Clin Invest Med 1997;20(2):95-101.
[résumé]
From the Departments of Medicine of *the Wellesley Hospital and †the Toronto Hospital, University of Toronto, Toronto, Ont.
(Original manuscript submitted Oct. 6 1995; received in revised form Jan. 2, 1997; accepted Jan. 8, 1997)
Reprint requests to: Dr. C.K. Chan, The Toronto Hospital, Eaton Building 10-N220, 200 Elizabeth St., Toronto ON M5G 2C4; fax 416 971-6427
Contents
Abstract
Objective: To assess changes in patterns of hospital admissions, in frequency of admissions and in average length of stay (ALOS) at a tertiary HIV referral centre, and to investigate the overall impact of care for patients with HIV infection or AIDS on peer hospitals in Ontario.
Design: Descriptive study.
Participants: Data were obtained on patients with HIV infection or AIDS treated at the Wellesley Hospital in Toronto for the fiscal years (May 1 to Apr. 31) 199091, 199192 and 199293, and on admissions for HIV or AIDS in 9 peer hospitals in Ontario during the same period.
Interventions: For the Wellesley Hospital, review of medical records of HIV-related admissions to determine the reasons for admission and to examine concurrent illnesses. For the Wellesley Hospital and peer hospitals, analysis of changes in ALOS and Resource Intensity Weights (RIWs).
Results: Between May 1, 1990, and Apr. 31, 1993, the number of admissions for treatment of Pneumocystis carinii pneumonia (PCP) fell, but admissions for respiratory infections other than PCP remained very common, although they decreased slightly. Overall, infection remained the main reason for admission. The frequency of gastrointestinal complications necessitating admission increased. The frequency of admissions remained high, although the ALOS decreased significantly. In the period between Apr. 1, 1991, and Mar. 31, 1994, the proportion of HIV-related discharges and total hospital discharges among the 9 peer hospitals remained stable. The HIV-related ALOS decreased substantially. Although the HIV-related average RIW decreased slightly, this measure and the mortality rate are still much higher for HIV-related admissions than for overall admissions.
Conclusions: This contemporary survey suggests that nonrespiratory infection complications have become the main reason for admission of patients with HIV infection or AIDS, but that the HIV tertiary hospitals are coping with the load of HIV-related admissions and the high average RIW associated with these patients by reducing the ALOS.
Résumé
Objectif : Évaluer l'évolution des tendances des hospitalisations, de la fréquence des admissions et de la durée moyenne du séjour (DMS) à un centre tertiaire de référence de personnes injectée par le VIH et étudier l'impact global du soin des patients infectés par le VIH ou atteints du SIDA sur les hôpitaux pairs en Ontario.
Conception : Étude descriptive.
Participants : On a réuni des données sur des patients infectés par le VIH ou atteints du SIDA traités à l'Hôpital Wellesley de Toronto au cours des exercices (1er mai au 30 avr.) 19901991, 19911992 et 19921993, ainsi que sur les admissions de personnes infectées par le VIH ou atteintes du SIDA dans 9 hôpitaux pairs de l'Ontario au cours de la même période.
Interventions : Dans le cas de l'Hôpital Wellesley, étude des dossiers médicaux d'admissions liées au VIH pour déterminer les motifs de l'admission et étudier la présence de maladies simultanées. Dans le cas de l'Hôpital Wellesley et des hôpitaux pairs, analyse des modifications de la DMS et facteurs de pondération de la teneur en ressources (PTR).
Résultats : Entre le 1er mai 1990 et le 30 avr. 1993, le nombre d'admissions pour traitement d'une pneumonie à Pneumocystis carinii (PCP) a diminué, mais les admissions à cause d'infections respiratoires autres que la PCP sont demeurées très fréquentes, même si leur nombre a diminué légèrement. Dans l'ensemble, l'infection est demeurée le principal motif d'admission. La fréquence des complications gastro-intestinales qui ont nécessité une admission a augmenté. La fréquence des admissions est demeurée élevée, même si la DMS a diminué sensiblement. Au cours de la période du 1er avr. 1991 au 31 mars 1994, la proportion des congés liés au VIH et des congés totaux de l'hôpital entre les 9 hôpitaux pairs est demeurée stable. La DMS pour causes liées au VIH a diminué sensiblement. Même si les facteurs PTR moyens liés au VIH ont diminué légèrement, cette mesure et le taux de mortalité sont encore beaucoup plus élevés dans le cas des admissions liées au VIH que dans celui des admissions générales.
Conclusions : Cette enquête récente indique que les complications d'infections non respiratoires sont devenues la principale cause d'admission de patients infectés par le VIH ou atteints du SIDA, mais que les hôpitaux tertiaires de traitement du VIH réduisent la DMS pour faire face à la charge de travail constituée par les admissions liées au VIH et les facteurs PTR moyens élevés liés à ces patients.
[ Top of document ]
Introduction
Now that we are well into the second decade of the HIV epidemic, it is pertinent to review the trends in the use of health care resources by patients with HIV infection to aid in the planning of hospital programs for future patient needs. A large amount of the work on the impact of the HIV epidemic on health care services originates from the United States,[14] whereas published information from the United Kingdom[58] and Canada[9] is more limited.
Historically, the most common pulmonary manifestation of HIV infection has been Pneumocystis carinii pneumonia (PCP). We have previously reported on changes in the patterns of cogent diseases necessitating the hospital admission of patients with HIV infection in 3 major HIV referral centres in Toronto just before and after the introduction of widespread use of PCP prophylaxis in May 1989.[9] This study was designed to compare contemporary results with those of our previous study.[9] In addition, it was designed to determine whether subsequent advances in antiretroviral drugs and popular use of systemic agents such as trimethoprimsulfamethoxazole and dapsone as well as aerosol pentamidine for PCP prophylaxis in the subsequent 3 years (from May 1, 1990, to Apr. 31, 1993) have further changed the frequency, average length of stay (ALOS) and primary reason for admission at our centre.
To provide a global perspective, we also examined the overall impact of patients with HIV infection or AIDS on the resources of comparable tertiary HIV hospitals in Ontario during the same period, to determine how these hospitals are coping.
[ Top of document ] Methods
Local perspective: Wellesley Hospital
Medical records of all HIV-related admissions for the 3 fiscal years from May 1 to Apr. 31, 199091, 199192 and 199293 were systematically reviewed. Charts were obtained from the Health Records Department at the Wellesley Hospital, a community-based and tertiary HIV care institution affiliated with the University of Toronto. The number of charts reviewed and the number missing were recorded.
Data were collected by 2 trained research assistants who were not blinded, since this was a descriptive study with no initial hypothesis. The principal cause of hospital admission was recorded as the presenting illness leading to admission. Thus, for a patient afflicted with a number of conditions, only the primary condition most responsible for the patient's admission to hospital was recorded as the admitting diagnosis. All other conditions were recorded as concurrent illnesses. The results of any investigations or procedures, such as radiographs, operations or cultures, were also recorded and analysed to help determine the primary illness necessitating admission. The principal cause of hospital admission was always cross-referenced with the discharge diagnosis to reveal any discrepancy.
For patients with multiple hospital admissions, each was treated as a separate data-entry point. The annual readmission rate was calculated by dividing the total discharges by the number of patients treated per year.
Provincial perspective: peer hospitals in Ontario
To determine trends in demands on resources by patients with HIV infection or AIDS in comparable hospitals in Ontario, data from all of the Wellesley Hospital's peer hospitals in Ontario were combined. Information was obtained from the Hospital Medical Records Institute (HMRI). (HMRI became known as the Canadian Institute for Health Information in February 1994, having been formed by the merger of Management Information System and HMRI.[10])
The total HIV-related average Resource Intensity Weight (RIW) was examined as an indicator of the amount of resources required by this group of patients. The RIW is calculated by the Canadian Institute for Health Information for each Case Mix Group (CMG). Each CMG is defined by the diagnostic-related grouping. The atypical average RIW is an average of all of the RIWs for different types of patients, including those who have signed out, died, transferred or who are statistical outliers.
Nine hospitals in Ontario were classified as peer hospitals to the Wellesley Hospital, according to HMRI. The other centres were: Hamilton General Division of Hamilton Civic Hospitals, Kingston General Hospital, University Hospital (London), Victoria Hospital Corporation (London), Ottawa Civic Hospital, the Ottawa General Hospital, St. Michael's Hospital (Toronto), the Toronto Hospital and Sunnybrook Health Science Centre. Information from the peer hospitals was combined with information from the Wellesley Hospital for the 3-year period between Apr. 1, 1991, and Mar. 31, 1994.
For the period between Apr. 1, 1991, and Mar. 31, 1992, the Toronto Western Hospital and the Toronto General Hospital had not yet amalgamated to form the Toronto Hospital. Thus, to illustrate the trends for the Toronto Hospital, the results for the Toronto Western Hospital and the Toronto General Hospital for the fiscal year 199192 were combined in our analysis. As well, for the period between Apr. 1, 1991, and Mar. 31, 1992, there were no data for the Hamilton General Division of Hamilton Civic Hospitals; therefore, the summary for this year excludes this hospital.
Data analysis
The HIV-related admission diagnoses at the Wellesley Hospital were organized to allow us to examine the primary reason for admission according to specific diagnosis and organ system. The demographic aspects of the population, including patient sex and HIV-associated risk behaviour, were also examined.
In addition, the data were also organized to compare several categories for all cases and all HIV cases in peer hospitals in each of the 3 fiscal years examined. These categories included: (1) number of discharges, (2) ratio of number of HIV cases to total number of cases, (3) ALOS, (4) average RIW and (5) mortality rate.
The primary objective was a qualitative analysis of the trends in hospital admissions and resource utilization. Statistical comparisons were conducted with the chi2 test for categorical variables and Student's t-test for continuous variables. A p value of less than 0.05 was considered statistically significant.
[ Top of document ] Results
Local perspective: the Wellesley Hospital
There were a total of 587 HIV-related hospital admissions between May 1, 1990, and Apr. 31, 1993, at the Wellesley Hospital. We reviewed 98% of these admissions (Table 1). Eleven charts could not be located despite intensive searching. The annual number of HIV-related admissions at the Wellesley Hospital almost doubled from 199091 to 199293 (Table 1); however, the readmission rate during this period went up only slightly, from 1.21 to 1.43 discharges per patient per year. Over the 3 years, the number of deaths during admission increased slightly, and the ALOS was progressively shorter (Table 1).
Although the major risk behaviour associated with HIV infection was homosexual or bisexual behaviour, there was a significant decline in this risk behaviour among patients in 199293, which corresponded to an increase in the proportion of other risk behaviour (Table 1).
Although only 3 years were reviewed in this study, we were able to extract data specific to the Wellesley Hospital from our previous study.[9] Consequently, Table 2 demonstrates the primary diagnoses driving HIV-related admissions at the Wellesley Hospital over a 5-year period.
We noted a significant increase in the number of annual admissions in the most recent 2 years (Table 2). The overall number of admissions due to infections was stable and remained the number one cause of HIV-related admissions, accounting for about two-thirds of the total number of admissions over the 5 years studied (1988 to 1993). The proportion of total admissions due to PCP fell (from 32% to 18%) during this period and was accompanied by a corresponding increase in the proportion of total admissions due to infectious causes other than PCP over the 5 years (from 39% to 51%). The decline in PCP admissions also corresponded to a decrease in the importance of the respiratory tract as the major system affected among patients during this period (from 49% to 35% of admissions). The proportion of admissions due to neoplasms fluctuated over the 5 years.
The annual HIV-related hospital admissions were also classified according to the major organ system involved (Table 3). An overall decrease in the proportion of respiratory causes for admission was observed in our previous study,[9] (from 50% in 198889 to 41% in 198990), and this pattern has continued (Table 3).
Gastrointestinal causes of admission have gradually become more prevalent in the recent era, accounting for 20% to 30% of admissions (Table 3). The frequency of admissions associated with the nervous system and the "other" category (which included admissions not associated with a specific organ system) appeared to be stable over the 3 years (Table 3).
Provincial perspective: peer hospitals in Ontario
HMRI data was assembled to examine province-wide statistics for the care of patients with HIV infection and AIDS among peer hospitals. The results, shown in Tables 4 and 5, illustrate that, during the 3 fiscal years (Apr. 1, 1991, to Mar. 31, 1994), the number of total hospital discharges and the number of HIV cases remained fairly stable.
The ALOS of the overall hospital admissions and of the HIV-related admissions both decreased over the 3 years. However, the HIV-specific ALOS decreased more dramatically than the overall ALOS, from 15.15 to 13.70 days (p < 0.0005). The overall hospital average RIW increased slightly over the 3-year period (p < 0.0001) (Table 4), whereas the HIV-specific average RIW decreased slightly (Table 5). Nevertheless, the HIV-specific average RIW is still about 50% higher than the overall average RIW. Both the HIV-specific and the overall hospital mortality rates remained stable over the 3 years. However, the HIV-specific mortality rate remained 5 times greater than the total hospital mortality rate (Tables 4 and 5).
[ Top of document ] Discussion
Most of the HIV-related admissions at our centre involved patients from the surrounding community. The Wellesley Hospital catchment area is Metro Toronto, which did not change in size during the study period. The number of discharges related to HIV or AIDS at our centre increased dramatically during the 3-year study period. On the other hand, the province-wide pattern of admissions related to HIV and AIDS remained stable during the period. This discrepancy demonstrates the importance of analysing local data in the context of the global picture. The increase in admissions observed at the local centre is most likely a reflection of bias in referrals and in the popularity of our centre, rather than a true increase in the number of cases in the region.
It was possible to examine the average number of discharges per patient per year at the Wellesley Hospital between Apr. 1, 1990, and Mar. 31, 1993. The increase from 1.21 to 1.43 discharges per patient per year may very well reflect the increasing life span of HIV patients as well as the greater number of adverse reactions to treatment.
The population of patients with HIV infection attending our centre during the study period consisted mainly of white, middle-class, homosexual men who have good access to health care. Their infection is generally diagnosed at an early stage, and they are given prophylaxis and are followed carefully from the start. There is a high degree of compliance with treatment. The population described in this report is very different from cohorts seen at inner-city health centres in the United States, where the patients tend to be from a lower socioeconomic class, have poor access to health care and be less compliant with therapy.[1,4,11] Nevertheless, the demographic characteristics of patients with HIV at our centre are changing. From May 1, 1990, to Apr. 31, 1993, there was a 20% drop in the proportion of patients for whom homosexual or bisexual contact was identified as the HIV-associated risk behaviour. Concurrently, there was a substantial increase in the proportion of patients for whom intravenous drug use was the identified risk behaviour as well as for whom the HIV-associated risk behaviour was unidentified. The increase in the proportion of other unidentified risk behaviours is most likely due to an increase in heterosexual contacts or drug use. Hence, it is likely that, a few years from now, the spectrum of disease necessitating hospital admission at our centre will change because of shifting patient demographic characteristics.
Between May 1, 1990, and Apr. 31, 1991, the first fiscal year examined in this study, we observed a higher proportion of admissions due to PCP (28%) than in previous years. It is possible that the aerosol pentamidine prophylaxis treatment was less effective with time. Alternatively, because of the concern about side effects of aerosol pentamidine, patients may have become less enthusiastic about the regular use of aerosol pentamidine and thus may have forgone prophylaxis. Between May 1, 1991, and Apr. 31, 1993, the proportion of PCP admissions again decreased (to 24% in the first year and 18% in the second), which may have been due to the popular use of other systemic prophylactic agents such as trimethoprimsulfamethoxazole and dapsone.
The reduction in the proportion of admissions due to PCP between May 1, 1989, and Apr. 31, 1993, after the introduction of PCP prophylaxis, was accompanied by an increase in the proportion of admissions due to other infectious conditions, primarily those of nonrespiratory origin. In addition, a small decrease in the number of admissions due to respiratory tract infections other than PCP was also observed. These trends imply that HIV-related respiratory complications are either being prevented or managed effectively in outpatient care.
The proportion of HIV-related admissions due to bacterial pneumonia is lower at our centre (9%, 8% and 9% in the 3 study years) than it was earlier.[9] This reduction may reflect concurrent suppression of bacterial pathogens by systemic PCP prophylactic agents such as trimethoprimsulfamethoxazole or better recognition and treatment of bacterial pneumonia in the ambulatory setting.
Chien and associates[9] reported small increases in the number of admissions due to gastrointestinal and central nervous system manifestations at the 3 HIV tertiary hospitals in Toronto between 1988 and 1990. At our centre, in the 3 subsequent years (May 1, 1990, to Apr. 31, 1993), a further increase in the number of gastrointestinal-related hospital admissions, especially those involving gastrointestinal tract infections, was observed. No increase was observed in admissions due to disorders of the central nervous system. Major changes in the process of initiation of treatment for cytomegalovirus retinitis, including home programs to provide treatment through a central venous catheter, may have resulted in avoidance of hospital admissions and may have contributed to the stable number of admissions for diagnoses involving the central nervous system.
As survival of HIV patients increases, it is interesting to examine whether there is an increase in the frequency of chronic diseases or of neoplastic disorders, which are the anticipated result of chronic immunosuppression.[9] In the recent era, the number of admissions due to lymphoma remained low and fairly stable. The number of admissions due to other neoplastic causes also remained low.
The HIV-specific ALOS at our centre remained fairly constant at around 16 days during the year before and the 2 years after the introduction of aerosol pentamidine for PCP prophylaxis (May 1, 1988, to Apr. 31, 1991). However, the ALOS decreased to 13 days between May 1, 1991, and Apr. 31, 1992, and decreased further to 11 days between May 1, 1992, and Apr. 31, 1993. In 1991, a nurse specialist in HIV care was introduced to assist the physicians working in HIV care at the Wellesley Hospital. In 1992, a specialized multidisciplinary HIV program was organized; it consisted of a nutritionist, a clinical fellow, a physiotherapist and home care personnel. These factors have likely contributed to the decrease in HIV-specific ALOS. The cost of the HIV nurse and the consolidated HIV team can easily be justified by the improved efficiency in handling patients with HIV infection.[4] This efficiency is demonstrated by the decrease in the ALOS despite the high average RIW of these cases.
Between May 1, 1991, and Apr. 31, 1994, the overall resources used in caring for patients with HIV infection and AIDS at Ontario peer hospitals decreased. Since the yearly number of hospital admissions remained stable, the decrease in global resources used reflects the lowering of both the ALOS and the average RIW. One possible explanation for the gradual decline in the HIV-specific average RIW over this 3-year period is the earlier detection of HIV. Thus, patients with HIV infection and AIDS tend to be better managed and are not as desperately ill when admitted to hospital as those who presented at an advanced stage of the disease in the earlier era. Despite the slight decrease in the HIV-specific average RIW over the 3 years, it is still much greater than the total hospital average RIW, which illustrates the severity of illness in patients with HIV infection and AIDS.
The decrease in the HIV-specific ALOS observed at our centre is consistent with that observed at comparable Ontario peer hospitals. Between Apr. 1, 1991, and Mar. 31, 1994, the overall HIV-specific ALOS at Ontario peer hospitals decreased more, proportionally, than the overall hospital ALOS. This is likely a reflection of the increased availability and quality of community and palliative care for patients with HIV infection. The decrease in the HIV-specific ALOS is unlikely to be a result of health care cutbacks. If cutbacks were the cause, one would expect a corresponding increase in the average RIW and perhaps mortality rate, together with a decrease in the number of hospital admissions, since only sicker patients would be admitted.
[ Top of document ] Conclusions
This contemporary survey suggests that the HIV-tertiary hospitals can cope with the load of HIV-related admissions while reducing the ALOS, despite the high average RIW associated with these patients. With the expected improvement in these patients' survival period, we anticipate that patients with multiple-organ illness due to HIV infection will likely impose costs on ambulatory and home care.
[ Top of document ] References
- Rosenblum LS, Buehler JW, Morgan M, Moien M. Increasing impact of HIV infection on hospitalizations in the United States, 19831988. J Acquir Immune Defic Syndr 1992;5:497-504.
- Chen RT, Rutherford GW, Payne SF, Barnhart JL, Lemp GF. Use of hospitals by patients with AIDS in San Francisco. N Engl J Med 1988;319:1671-2.
- Chang R, Wong G, Gold J, Armstrong D. HIV-related emergencies: frequency, diagnoses, and outcome. J Gen Intern Med 1993;8:465-9.
- Fahs MC, Fulop G, Strain J, Sacks HS, Muller C, Cleary PD, et al. The inpatient AIDS unit: a preliminary empirical investigation of access, economic, and outcome issues. Am J Public Health 1992;82:576-8.
- O'Brien SJ, Burch J, Mayon-White RT. Hospital bed usage by people with HIV disease: experience in a provincial setting. Public Health 1993;107:355-62.
- Murphy M, Misquito N, Pomeroy L, Mulcahy F. Admissions for HIV-1 related disease in a Dublin hospital 19871990. Int J STD AIDS 1991;2(6):436-9.
- Willocks L, Cowan F, Brettle RP, Emmanuel FXS, Flegg PJ, Burns S. The spectrum of chest infections in HIV positive patients in Edinburgh. J Infect 1992;24:37-42.
- Pitkin AD, Grant AD, Foley NM, Miller RF. Changing patterns of respiratory disease in HIV positive patients in a referral centre in the United Kingdom between 19867 and 19901. Thorax 1993;48:204-7.
- Chien SM, Rawji M, Mintz S, Rachlis A, Chan CK. Changes in hospital admissions pattern in patients with human immunodeficiency virus infection in the era of Pneumocystis carinii prophylaxis. Chest 1992;102:1035-9.
- Fallon G, editor. Comparative analysis report: hospital utilization in Canada. Ottawa: Canadian Institute for Health Information, 1994: 1.
- Seage GR III, Oddleifson S, Carr E, Shea B, Makarewicz-Robert L, Van Beuzekom M, De Maria A. Survival with AIDS in Massachusetts, 19791989. Am J Public Health 1993;83(1):72-8.
