Brief Report: Clinical Science
The HIV epidemic arrived in South Africa some years after other African areas, but, once established, seroprevalence levels rose rapidly. By the end of 2003, more than 5 million South Africans were HIV infected out of a population of 40 million, placing a severe strain on the provision of health care.1 Studies from urban tertiary referral hospitals in Africa have documented the effect of HIV on medical admissions: in Kenya admissions to Kenyatta National hospital rose greatly between 1988/1989 and 1997 as did bed occupancy; and in the University Hospital Blantyre, Malawi, 70% of medical patients were found to be HIV infected and 8% of patients were classified as having AIDS, whereas over two-thirds of deaths were related to bacterial infections, tuberculosis, and AIDS.2,3 In South Africa a number of studies have addressed the impact of HIV on admissions to tertiary institutions: in Pietermaritzburg, KwaZulu/Natal, 28% of medical admissions were diagnosed with tuberculosis, many were also HIV infected, and in a large tertiary hospital in Durban 54% of medical inpatients were found to be HIV infected and 56% of these patients had tuberculosis.4,5 Few data exist on the effect of HIV on adult patients in rural areas. These studies are important as they allow informed decisions to be made about the distribution of limited public resources. Here we present the results of a 12-year survey of admission trends among adult medical patients to a hospital in rural KwaZulu/Natal, South Africa.
Study Population and Setting
This study was conducted in Hlabisa Hospital, a government hospital situated in the rural district of Hlabisa, KwaZulu/Natal, approximately 250 km north of Durban, on the east coast of South Africa. The district has an area of 3729 km2 (1492 sq miles) and a population density of 53 people per km2 (21 per sq mile). The population speaks mainly isiZulu and in 2000 numbered approximately 220,000 (46% male and 54% female). Unemployment is high, and a high proportion of the male population works away from home for long periods. The main sources of income in the area are farming, tourism, and social grants.6 Few people have private health insurance and most of the population seeks care at 1 of the 14 government-run primary health care clinics and are referred as necessary to Hlabisa Hospital, the only hospital in the district. HIV prevalence measured in antenatal clinics rose from 4% in 1992 to 35% in 2002 in the Hlabisa district.7,8
In this study, hospital medical admissions were examined for the same 2-month period in 1991, 1995, 1998, and 2002. The period chosen was the months of March and April. These months were chosen as they are usually outside the periods of sporadic seasonal epidemics of cholera and typhoid that affect the district; also, they are outside the time when migrant workers return home. A Shigella epidemic and the end of the seasonal malaria period occurred during the period studied.
Case notes of patients admitted during the study period were examined by 1 of 2 physicians (AR, MD) who worked in the hospital and were familiar with hospital protocols. Admissions were identified from the admission register in the outpatient department and the ward admission and discharge registers. Of the notes identified from the admission registers, 626 of 626 (100%) were available from 2002, 650 of 653 (99.5%) from 1998, 407 of 421 (96.7%) from 1995, and 228 of 237 (96.2%) from 1991. All adult patients admitted to the male and female medical wards during the study period were included. All patients had been seen by a physician on admission and on discharge or prior to transfer. Data were extracted from the case notes, including age, sex, presenting syndrome, discharge diagnosis, length of stay, and discharge destination. Standardized coding and diagnostic criteria were used by both clinicians. Patients were followed up from admission until death, discharge home, or transfer to another ward or hospital. Outcomes were coded as improved, which was judged clinically and by asking the patient; died; and “other.” The “other” category consisted of patients who had either absconded or who had been transferred to other hospitals. The presenting syndrome was recorded based on the patient's predominant presenting clinical signs and symptoms. Discharge diagnosis was based on the best available evidence; a laboratory test, imaging, histology, or clinical impression. Patients transferred to the tuberculosis ward were coded as discharged for the purposes of analysis as transfer to this ward from the medical ward occurred prior to discharge home and the performance status of these patients was equivalent to patients being sent home from the medical ward. HIV status was not consistently available during all of the study periods and is therefore not given here. Data were entered and analyzed using EpiInfo version 6.04d (Centers for Disease Control, Atlanta, GA). Categorical data were analyzed using a χ2 for trend and continuous data were analyzed using parametric and nonparametric tests as appropriate.
Overall 1911 of 1936 of case notes (98.7%) for the period under study were available for examination, 228 of 237 (96.2%) from 1991, 407 of 420 (96.9%) from 1995, 650 of 653 (99.5%) from 1998, and 626 of 626 (100%) from 2002. Only patients whose case records were available were included in the analysis. The number of patients admitted in the 2-month period under study rose from 228 in 1991 to 407 in 1995 (a 179% rise compared with 1991), then to 646 in 1998 (a 159% rise compared with 1995) with a slight fall to 626 in 2002 (a 3% fall compared with 1998). The proportion of women compared with men increased over the study period (Table 1), but the trend was not significant (χ2 1 = 1.6, P = 0.2). The distribution of ages of the patients at several of the time points was not normal; therefore comparisons are presented between medians using the Mann-Whitney U test. The median age of women in the study fell from 50 years (interquartile range [IQR] 24) in 1991 to 34 years (IQR 33); in 2002 this was statistically significant (Z = −2.91, P < 0.01). For men the median age fell from 45 years (IQR 33) in 1991 to 39 years (IQR 29) in 2002, which was also significant (Z = −3.2, P < 0.01). For all 3 study periods after 1991, female patients were younger than the male patients (Table 2).
The distribution of presenting complaints changed over the study period (Table 3.) Increases were seen in the proportion of patients presenting with respiratory, neurologic, and febrile illnesses. Significant decreases were seen in the proportion of patients presenting with cardiac and psychiatric conditions. From 1995, tuberculosis became the most frequent discharge diagnosis and made up an increased proportion of all discharges, compromising 37% of all adult patients discharged in 2002. Lower respiratory tract infection was the second most frequent discharge diagnosis from 1995 onwards, although the proportion of patients with this diagnosis fell from 21% in 1991 to 15% in 2002 (Table 4). The length of time patients spent on the medical wards decreased from 10.9 days in 1991 to 7.9 days for the rest of the study period; this was significant (P < 0.01). The fall in length of stay was greater for women than men.
There were significant changes in patient outcomes over the study period. The proportion of patients considered to have improved fell from 90% in 1991 to 77% in 2002 (P < 0.001). The proportion of patients dying as inpatients increased from 8% in 1991 to 20% in 2002 (P < 0.001). There was no change in the proportion of patients who absconded or took their own discharge during the study period. After not figuring in the top 3 causes of death in years 1991, 1995, and 1998, tuberculosis was the most common cause of death in 2002, accounting for 30% of all deaths (Table 5).
Few studies have examined the impact of HIV/AIDS on rural district hospitals in Africa. The data presented here demonstrate changes occurring in the demographics of adult medical hospital admissions in a rural part of South Africa over 12 years corresponding to a rise in HIV prevalence from 4% to 35%. This dramatic rise in HIV prevalence has been shown to affect health services at many levels. An 88% rise in attendances at a primary health care clinic in the district, with no increase in staffing levels, has already been documented.9 A large proportion of this increase was due to respiratory presentations attributed to HIV infection. An 81% rise in general hospital admissions has also been documented at Hlabisa Hospital between 1991 and 1998.10 A verbal autopsy study conducted in the Hlabisa district attributed 48% of deaths in adults aged between 15 and 60 years as being due to AIDS with or without tuberculosis, in broad agreement with our findings.11 Studies from tertiary hospitals in South Africa have also shown that a large proportion of admissions and deaths on medical wards are due to HIV/AIDS and tuberculosis.4,5 A large study of admissions in a tertiary hospital in Bulawayo, Zimbabwe, has also documented a rise in HIV/AIDS-and tuberculosis-related admissions as well as an increase in lower respiratory tract infections.12
Data presented in this study show a 275% increase in adult medical admissions over the 12-year study period in the absence of changes in hospital policy, health service provision in the district, or the demographics of the district population. In partial response, average length of stay decreased by 28% over the study period without any increase in number of medical beds or medical and nursing staffing levels. The decreased length of stay was not due to an increase in early deaths, as the mean length of stay of a patient dying on the ward was 9 days, compared with 7.7 days for a patient discharged improved, and 7 days for patients transferred to the tuberculosis ward. The stable number of admissions and length of stay in 1998 and 2002 suggest that despite rising HIV prevalence the medical wards are functioning at capacity within the current organization of the district health service. The fall in length of stay in is contrast to findings in Kenya at a tertiary institution, where the length of stay remained constant over a 10-year period.2 More admissions attributed to HIV and increased incidence of tuberculosis have been documented previously in a rural Kenyan hospital.13 Median age of admissions fell significantly over the study period, with the fall being greater in women than in men, reflecting the impact of HIV on young adults, who normally place little demand on health care services.
The increase in HIV-related infectious disease may result in patients with other conditions receiving inadequate care. Most notably, the proportion of patients presenting with cardiac conditions and psychiatric conditions decreased over the study period. Psychiatric conditions were the 4th most common discharge diagnosis in 1991; in 2002 combined psychiatric conditions made up only 2% of all discharges. During the 10-year study period there were no community-based interventions to reduce psychiatric admissions. This may indicate that ambulatory patients with chronic conditions are less likely to be admitted due to the increasing number of critically ill patients with HIV/AIDS-related conditions. This could reverse the increased survival gained through inpatient treatment of people with uncontrolled chronic conditions such as cardiac failure, diabetes, and psychiatric illness. In support of this argument, a large study in a tertiary hospital in Nairobi, Kenya, found that the increasing numbers of HIV-related admissions was at the expense of HIV-uninfected patients who were in effect crowded out of the hospital.14 Community-based management programs for patients with chronic conditions are currently being implemented in Hlabisa to address these problems.
The large increase in the proportion of patients dying on the ward reflects the increase in critically ill patients being admitted as a result of HIV infection and a shortage of home-based care and hospice facilities. Many of the deaths were in patients in whom tuberculosis had been diagnosed, who were also likely to have HIV. Seventy percent of tuberculosis patients in Hlabisa were HIV positive in an unpublished study by Alasdair Reid from 1997. In 1991, despite being the 3rd most common discharge diagnosis, there were no deaths among inpatients with tuberculosis; in 2002, 38 deaths (30%) were in patients with tuberculosis. Tuberculosis patients infected with HIV have been shown to have excess mortality in Africa, compared with those not infected with HIV.15 It has also been shown that there is excess mortality in patients with tuberculosis and HIV if antiretroviral therapy (ART) is delayed.16 Tuberculosis/HIV-coinfected patients who are eligible for ART (ie, those with a CD4 count <200 cells/μL or those with a World Health Organization stage IV condition) will be a priority for treatment in South Africa as access to ART is scaled up.
The rise in medical admissions places an increased burden of care on health professionals who are already suffering from stress and exhaustion.17 It has been estimated that 40% to 50% of South Africa's work force, including health care workers, could die of AIDS over the next 10 years without ART. This combined with the loss of health care professionals to developed countries, the difficulties in recruiting and retaining staff in rural areas, and the increased burden of care will compromise the ability of the public sector to provide health care, unless alternative means of delivering services can be implemented.
The HIV epidemic has increased the number of medical admissions to our hospital, primarily with infectious diseases such as tuberculosis, lower respiratory tract infection, and diarrheal illness. Improved early management of these conditions by community health care workers and in primary health care clinics may reduce the burden of these conditions on secondary facilities. To reduce the burden of opportunistic infections, increased access to voluntary counseling and testing for HIV should be encouraged. Accelerated access to ART as well as tuberculosis case finding and preventive measures such as isoniazid preventive therapy should be prioritized for people living with HIV/AIDS; this may lead to reduced hospital admissions due to infections.18 Improvement of community-based palliative care may help to reduce the burden of terminal care on the hospital.
1. UNAIDS. AIDS Epidemic Update
. 2004. Geneva, Switzerland: UNAIDS; 2004. UNAIDS/04.45E.
2. Arthur G, Bhatt SM, Muhindi D, et al. The changing impact of HIV/AIDS on Kenyatta National Hospital, Nairobi from 1988/89 through 1992 to 1997. AIDS
3. Lewis DK, Callaghan M, Phiri K, et al. Prevalence and indicators of HIV and AIDS among adults admitted to medical and surgical wards in Blantyre, Malawi. Trans R Soc Trop Med Hyg
4. Alvarez GG, Thembela BL, Muller FJ, et al. Tuberculosis at Edendale Hospital in Pietermaritzburg, Kwazulu Natal, South Africa. Int J Tuberc Lung Dis
5. Colvin M, Dawood S, Kleinschmidt I, et al. Prevalence of HIV and HIV-related diseases in the adult medical wards of a tertiary hospital in Durban, South Africa. Int J STD AIDS
6. Curtis B, Bradshaw D, Njilana B. Sociodemographic Profile of the Magisterial District of Hlabisa: 1996 Census
. Capetown, South Africa: South African Medical Research Council; 2001.
7. Floyd K, Reid RA, Wilkinson D, et al. Admission trends in a rural South African hospital during the early years of the HIV epidemic. JAMA
8. Rollins NC, Dedicoat M, Danaviah S, et al, for the Child Health Group. Prevalence, incidence, and mother-to-child transmission of HIV-1 in rural South Africa. Lancet
9. Dedicoat M, Grimwade K, Newton R, et al. Changes in the patient population attending a primary health care clinic in rural South Africa between 1991 and 2001. S Afr Med J
10. Floyd R, Reid RA, Wilkinson D, et al. Admission trends in a rural South African hospital during the early years of the HIV pandemic. JAMA
11. Hosegood V, Vanneste AM, Timaeus IM. Levels and causes of adult mortality in rural South Africa: the impact of AIDS. AIDS
12. Mudiayi TK, Onyanga-Omara A, Gelman ML. Trends of morbidity in general medicine at United Bulawayo Hospitals, Bulawayo, Zimbabwe. Central African Medical Journal
13. Torrens JK. HIV and tuberculosis in a rural hospital in Kenya. East Afr Med J
14. Gilks CF, Floyd C, Otieno L, et al. Some effects of the rising case load of adult HIV-related disease on a hospital in Nairobi. J Acquir Immune Defic Syndr
15. Banergee A, Moyo S, Salaniponi F, et al. HIV testing and tuberculosis treatment outcome in a rural district of Malawi. Trans R Soc Trop Med Hyg
16. Dheda K, Lampe FC, Johnson MA, et al. Outcome of HIV associated tuberculosis in the era of highly active antiretroviral therapy. J Infect Dis
17. Shisana O, Hall E, Maluleke KR, et al. The Impact of HIV/AIDS on the Health Sector
. Pretoria, South Africa: South African Department of Health; 2003.
18. World Health Organization. Interim Policy on Collaborative TB/HIV Activities
. Geneva, Switzerland: World Health Organization; 2004. WHO/HTM/TB/2004.330; WHO/HTM/HIV/2004.1.
Keywords:© 2005 Lippincott Williams & Wilkins, Inc.
Africa; rural; HIV