Annals of African Medicine Annals of African Medicine Society ISSN: 1596-3519 Vol. 5, Num. 3, 2006, pp. 142-148

Annals of African Medicine, Vol. 5, No. 3, 2006, pp. 142-148

Impact of HIV/AIDS on Trends in Major Causes of Death at a Rural Mission Hospital in Kenya: Review of 4858 Records

Peter A. Leblanc

4349 Weather Stone crossing, Zionsville, Indiana 46077, U. S. A.
Reprint request to: Peter Leblanc. E-mail: leblanc_peter@hotmail.com

Code Number: am06033

Abstract

Background:  Acquired Immune Deficiency Syndrome (AIDS), caused by the human immunodeficiency virus (HIV), is a worldwide public health issue.  Hospital death records can be used to study the impact of HIV in Africa.  The explanation of mortality figures through hospital records identifies the evolution of the pandemic at that point.  
Methods: This study was framed with the objective, to describe trends in the leading causes of death from 1980 to 2000 at Kijabe Hospital; determining the proportion of deaths attributed to HIV/AIDS. Data were examined from death records stored in an ACCESS database at Kijabe Hospital.  The numbers of deaths in categories of causes of death were used to determine trends in the most frequent causes of death over the time period.  In the case of HIV/AIDS the frequency of this diagnosis as recorded on the death certificates was tracked.  The study design was a retrospective review of the death records.  
Results:  Larger proportions of young people died in at Kijabe Hospital over the study period.  HIV/AIDS became the leading cause of death for every year after 1991. 
Conclusion: These trends may help rural hospitals plan and allocate resources.  The data in this study may influence local resource distribution and future programs in similar settings.

Key words:HIV/AIDS, rural, hospital mortality, East Africa

Résumé

Introduction :Syndrôme immunodéficitaire acquis (SIDA) causé par le virus immunodéficitaire humaine (VIH), est un problème de la santé publique mondial.  On peut utiliser les dossiers des morts dans l’hôpital pour étudier l’impact du VIH en Afrique.  L’explication des chiffres de la mortalité à travers les dossiers dans l’hôpital a identifié l’évolution de cette pandémique à cet endroit.
Méthodes : L’objet de cette étude est de décrire les tendences dans les causes principales des morts de 1980 au 2000 dans l’hôpital de Kijabe ; décider la proportion des morts attribuable au VIH/SIDA.  Des données ont été étudiées à travers les dossiers des morts mis en reserve dans un ACCESS base de données à l’Hôpital de Kijabe.  Les nombres des morts selon des catégories des causes de morts ont été utilisés pour décider les tendances dans les causes des morts les plus fréquentes au cours de la période d’étude.  Dans le cas de VIH/SIDA, la fréquence de cet diagnostic comme noté sur les certificats des morts était  retrouvé.  Plan d’étude ;  Il s’agit d’un bilan rétrospectif dans les dossiers des morts.
Résultats : Une proportion nombreuses des jeunes morts à l’hôpital de Kijabe au cours de la période d’étude.  VIH/SIDA est devenu la cause principale de la mort pour chaque année après 1991.
Conclusion : Ces tendances pourront aider des hôpitaux ruraux pour mieux organiser et affecter des ressources   Les données dans cette étude pourront influcer la répartition des ressources locales et des programmes futur dans des milieux semblabes.

Mot clés : VIH/SIDE, rural, mortalité dans l’hôpital , Afrique de l’est

The Human Immunodeficiency Virus (HIV) epidemic continues to grow internationally; particularly affecting Africa. ¹ The HIV pandemic is still continuing and African hospitals treat a growing percentage of patients who experience HIV related conditions.  Unlike larger urban hospitals, rural hospitals are often run by mission organizations, churches or non-governmental organizations (NGOs).  These hospitals have varied sources of funding and often rely on outside donor funds.  This study wil report trends in the leading causes of death at Kijabe hospital, a smaller rural mission hospital about an hour’s drive from Nairobi, Kenya.  The results show the evolution of the HIV pandemic as experienced at the rural hospital level.

The majority of the world’s current HIV-positive individuals are found in Africa. ² Hospital records, including death certificates, have been used to describe the impact of diseases on varied populations. ^{3 - 10} Hospital records were used to approximate population trends of infectious disease and healthcare deficits in a transient war-torn population in Uganda. ⁴ In a setting where the population was uprooted by conflicts, hospital records provided an estimate of trends in the community. Accorsi et al, with the Italian-Ugandan AIDS cooperation program, found malaria, respiratory TB, pneumonia, measles, diarrhea, meningitis, septicemia and AIDS to be the leading infectious causes of admission to Lacor Hospital, in Uganda.  At this hospital HIV/AIDS was the 11^th leading cause of admission, but it was the fourth leading cause of death. ³

The HIV epidemic was the subject of two studies at the Kenyatta national hospital in Nairobi, Kenya. ^{6, 7} It was found that fewer AIDS patients were treated, while larger numbers of HIV-positive patients with less acute illnesses were treated at the hospital.  These patients presented in larger numbers but were not as ill, and the hospital had lower morbidity and mortality among all patients. ⁷

There are many unknowns in the overall HIV/AIDS picture in Africa.  There are limited tracking tools in place for information collection. ¹¹ Information from hospital records and readily available data should be analyzed to lay the ground work for interventions in these settings.  A retrospective analysis of hospital records helps hospitals improve their function and future planning.

 Materials and Methods

The study period spanned the years 1980 to 2000, the deaths at Kijabe Hospital, over this time period, were used for this study.  An ACCESS database was created from the death records at the hospital for these years in the study period.  Each death over this time period was stored as a record in the database.  The information analyzed for this study was the number of deaths per year, the characteristics and patterns of these deaths. 

At the hospital, physicians complete the governmental death certificates.  Staff members from the medical records department of the hospital transfer the information from each death certificate to a handwritten log kept for the hospital’s records.  The original death certificates are then submitted to the regional government office.  The investigator transferred the information, from this hospital log and available death certificates, into the database. 

The governmental requirements varied over the time period and there were multiple different physicians who filled out the death certificates.  The investigator looked at the causes of death and picked the corresponding immediate cause of death for each record.  In cases of an underlying cause of death, that directly led to the death, such as HIV/AIDS, TB or another identified disease; this was recorded as the immediate cause of death.  The classifications and groupings did not follow any international coding or classification system.  Admission data was not studied over the study period at all, for this research project only death records were studied.  Deaths due to the same cause were grouped and tallied by year that the death occurred at the hospital, in this way the leading cause of death by year was determined.  Two other variables in addition to cause of death were grouped and tallied for this study, these were gender and age.  Age, gender and cause of death were the three main pieces of information taken from each death record and analyzed for trends.

Tabulation was performed using Minitab, SPSS (a statistical software package) and a simple calculator.  Graphs were made with Excel. Some of the results were grouped, by time period, and by age groups chosen to represent infancy (<1), young childhood (1-5), older childhood (6-15), youth (16-25), young adult (26-45), older adult (46-65), and geriatric (65+).  SPSS was used to test for significance in the differences found from year to year in the proportion of deaths due to HIV/AIDS.  This was done by comparing the percentages of deaths due to HIV/AIDS for each year during the study period.  The two-sided chi-square test was applied, to test for significance in the differences seen in the numbers of deaths due to HIV/AIDS.  The proportion of deaths in each year from 1991 to 2000 was tested against the baseline year of 1990.

Kijabe Hospital confirmed permission, in writing, to the Director of Research at the University of Texas, School of Public Health.  A copy of the database, as released by the owner (Kijabe Medical center), is kept by the investigator.  The database is kept on a secured disk.  Only the investigator has access to the disk as part of records safeguarding.

There were 4858 death records reviewed over the study period from the year 1980 to the year 2000.  23 records were eliminated because the date of death could not be confirmed.  There was an increase in the total numbers of deaths at Kijabe Hospital over the study period (Figure 1); the majority of the deaths occurred in the second half of the study period (1991-2000). 

The first death in the study period from 1980to 2000 recorded as a death related to HIV, occurred in 1989.  In 1990 there were three male deaths and five female deaths associated with HIV/AIDS.  These nine deaths were the only deaths related to HIV/AIDS recorded in the Hospital records prior to 1991.  In the study period from 1980 to 2000, there were also two male deaths due to Kaposi’s sarcoma, one in 1980 and another in 1983.  Other causes of death recorded in the dataset from 1980 to 2000 included TB, pneumonia, CHF, other non-infectious diseases, cancers including cancers of the gastrointestinal tract, and complications of childbirth.  These were common diagnoses seen throughout the study period; however HIV/AIDS became the most important cause of death in the 1990s (Table 3).  The leading causes of death over the study period are shown in Table 3; these are shown in grouped form with a male and female column.

HIV/AIDS was not recorded as one of the five leading causes of death in any year before 1991.  From 1991 to 2000, the leading cause of death for each year at Kijabe hospital was HIV/AIDS.  The percentage of deaths due to HIV/AIDS peaked in 1996, when more than 1 in 5 of the deaths was due to HIV/AIDS (Figure 2).  HIV/AIDS accounted for a higher percentage of deaths as compared to other leading causes of death in the years prior to 1991.  HIV/AIDS accounted for 17.9% of the deaths at Kijabe Hospital from 1996-2000. 

There was an increase in deaths among the 26-45 age groups from 10% to between 20-30% of all deaths (Tables 1 and 2).  Two tables (Tables 1 and 2) show the deaths among each age group and the percentage.  The results are grouped by gender and also into four shorter time periods across the full study period.  Among males in this age group (26-45), the percentage rose from 10.2% to 22.7% over the study period.  Among females (26-45), the increase was from 10.5% to 28.3%.  Although these trends were identified the grouping by age and gender resulted in smaller sample size.  More detail was not possible with the size and quality of records in the study period.  The information was retained in Tables 1 and 2, but no statistical significance was identified.   

The main findings from this study are the upward trend in percentage of deaths occurring among the 26-45 age group as depicted in Tables 1 and 2, and the overall increase of the HIV/AIDS associated deaths.  The evolution of the HIV/AIDS epidemic at this hospital resulted in a very small number of cases in the 1980s.  1990 was the first year where HIV/AIDS deaths were consistently diagnosed.  There was a sharp increase in the recorded numbers of deaths due to this disease in every year after 1991.  In each year from 1991 to 2000, the percentage of deaths due to HIV/AIDS was compared to 1990.  The percentage of deaths due to HIV/AIDS in 1990 was 4.73%.  The 2-sided chi-square statistical test showed significance (p<.05) for every year from 1992 to 2000, except in 1995 (p=.095).  There were a large number of death records with missing data in 1995.  1995 was the only year after 1991 for which the chi-squared test didn’t reveal a statistically significant difference in the percentage of deaths due to HIV/AIDS when compared to 1990.  These percentages are reported in Figure 2.  The primary killer of HIV positive patients was AIDS (Table 4).

Table 1:  Distribution of deaths by age and gender, 1980-1990

*2 deaths in the 1980-1985 period with missing/unknown gender and age
#1 death recorded in the 1986-1990 period of missing/unknown gender and age

Table 2:  Distribution of deaths by age and gender, 1991-2000

++43 deaths recorded in the 1996-2000 period with missing gender field values

Table 3:  Five leading causes of death for each gender, comparing time periods

CHF: Congestive heart failure; CVA: Cerebrovascular accident

Table 4:  Leading causes of death among HIV-positive patients from 1991 to 2000.

The findings from this study, the increase in the proportion of deaths occurring among the young adult age group and the increased number of deaths related to HIV/AIDS, have been described elsewhere.  This is congruous with trends reported in various hospital settings. ^{5-7, 10, 12} The actual impact of HIV in the hospital setting and community may have been underestimated since not all patients at hospitals are tested for HIV. ³ Hospitals’ records help inform investigators about the burdens of infectious disease, war and AIDS, but they are limited in the ability to describe population trends. ³

There are limitations in this study including, the difficulties which prevent HIV-positive persons from seeking hospital care including cost or lack of confidence in hospital services.  Additionally, at-risk groups have died at high rates and there may have been many HIV-positive patients who died at home.  There may also be a distrust of western biomedicine, so traditional healers or remedies were used.  Future research must be done to explain the relationship between health-seeking behaviors and population infection rates.  Even with these limitations the trends are dramatic and the evolution of the HIV epidemic is clearly seen in this setting.  Areas like maternal health and the health of young people are of particular importance for future HIV interventions.  ^{12 - 21} The trends in overall deaths and HIV/AIDS-related deaths in a mission Hospital in Kenya add information on the evolution of the HIV epidemic in Africa.

There was a contrast in the leading causes of death seen in the database before 1991 and those seen later in the study period.  Non-infectious causes of death, in addition to pneumonia and TB were among the leading causes of death in the 1980s.  After 1990, a higher percentage of deaths and HIV-positive hospitalized patients have causes of death associated with immunosuppression; TB, pneumonia and meningitis were found to be important diseases among HIV-positive individuals.  Hospital data are limited by clinician variability in diagnosis but despite this limitation there is a stark contrast between the numbers of deaths due to HIV/AIDS prior to 1991 and thereafter.

Identification of the rising numbers of HIV/AIDS deaths can help health care providers at this hospital.  Knowing the leading causes of death by proportion along with current resource utilization helps to understand where resources are being underutilized, or where new funding is needed.  The HIV epidemic and the relationship to westernization as well as population growth, urbanization, and the effect on hospitals is still not well understood.  Most of the HIV/AIDS patients who die at Kijabe Hospital are adults and it is unclear if children are receiving medical attention for their illnesses in the same patterns as adults. 

The limitations of this study are due to the study population, the retrospective descriptive study design and the lack of uniformity in collection of the data, hospital data are open to selection bias and inconsistency.  Missing data especially in the 1995 records is another limitation especially given the small sample size.  Additionally the death certificate format changed over the time period as the government changed the requirements.  The investigator has been trained as a nurse, but some information may have been misclassified in the final analysis as it was distilled down and then grouped for tallies.  Finally, the patients at Kijabe Hospital are unique to the geography and setting, the community served is primarily rural.  There are few emergency medical services in Kenya, so proximity to a highway or roadway can result in a skewed distribution of causes of death. 

Although retrospective hospital data are limited, the data that are available and reveal trends should not be ignored.  Decisions that are made based on these data sets have to be very selective, with careful scrutiny of the populations served and the similarities with this study population.  The important finding is that HIV/AIDS has increased to the point of being the number one killer at Kijabe Hospital, a rural setting.  In addition the young adult age group was being affected in a disproportional way, consistent with findings from other studies; ²² this has many important future implications.

Acknowledgements

This work was completed to satisfy the requirement for a master’s degree in public health at the University of Texas, Health Science Center at Houston, School of Public Health.  George Kerr MD served as the academic advisor and thesis supervisor for this work and Mark Williams PhD served as the committee member.

1. De Cock KM, Weiss HA.  The global epidemiology of HIV/AIDS. Trop Med Int Hlth 2000;5:A3-A9
2. Morb Mortal Wkly Rep 2001;50:429-444
3. Accorsi S, Fabiani M, Lukwiya M, Onek PA, di Mattei PD, Declich S.  The increasing burden of infectious Diseases on hospital services at St. Mary’s Hospital Lacor, Gulu, Uganda. Am J Trop Med Hyg 2001;64: 154-158
4. Accorsi S, Fabiani M, Lukwiya M, et al. Impact of insecurity, the AIDS epidemic, and poverty on population health:  disease patterns and trends in Northern Uganda.  Am J Trop Med Hyg 2001;64: 214-221
5. Aiken CG.  HIV-1 infection and perinatal mortality in Zimbabwe.  Arch Dis Child 1992;67:595-599
6. Arthur G, Bhatt SM, Muhindi D, Achiya GA, Kariuki SM, Gilks CF.  The changing impact of HIV/AIDS on Kenyatta national hospital, Nairobi from 1988/89 through 1992 to 1997.  AIDS 2000; 14:1625-1631
7. Gilks CF, Floyd K, Otieno LS, Adam AM, Bhatt SM, Warrell DA.  Some effects of the rising case load of adult HIV related disease on a hospital in Nairobi.  J Acquir Immune Defic Syndr Hum Retrovirol 1998;18:234-240
8. Meyers TM, Pettifor JM, Gray GE, Crewe-Brown H, Galpin JS. Pediatric admissions with Human Immunodeficiency Virus infection at a regional hospital in Soweto, South Africa.  J Trop Pediatr 2000;46:224-230
9. Walraven G, Nicoll A, Njau M, Timaeus I.  The impact of HIV-1 on child health in sub-Saharan Africa:  the burden on the health services.  Trop Med Int Hlth 1996;1:3-14
10. Zwi K, Pettifor J, Soderlund N, Meyers T.  HIV infection and in-hospital mortality at an academic hospital in South Africa.  Arch Dis Child 2000;83:227-230
11. Timaeus IM.  Impact of the HIV epidemic on mortality in sub-Saharan Africa:  evidence from national surveys and censuses.  AIDS 1998;12 (Suppl 1):S15-S27
12. Asiimwe-Okiror G, Opio AA, Musinguzi J, Madraa E, Tembo G, Carael M.  Change in sexual behaviour and decline in HIV infection among young pregnant women in urban Uganda.  AIDS 199;11:1757-1763
13. Batter V, Matela B, Nsuami M, et al. High HIV-1 incidence in young women masked by stable overall seroprevalence among childbearing women in Kinshasa, Zaire:  estimating incidence from serial seroprevalence data.  AIDS 1994; 8:811-817
14. Gichangi P, Fonck K, Sekande-Kigondu C, et al.  Partner notification of pregnant women infected with syphilis in Nairobi, Kenya.  Int J STD AIDS 2000;11:257-261
15. Guest E.  Children of AIDS Africa’s orphan crisis. Pluto Press, London, 2001
16. Hunter DJ, Maggwa BN, Mati JK, Tukei PM, Mbugua S.  Sexual behavior, sexually transmitted diseases, male circumcision and risk of HIV infection among women in Nairobi, Kenya.  AIDS 1994;8:93-99
17. McElroy PD, Ter Kuile FO, Hightower AW, et al.  All-cause mortality among young children in Western Kenya.  VI:  the Asembo bay cohort project.  Am J Trop Med Hyg 2001;64(1-2 Suppl):18-27
18. Nicoll A, Timaeus I, Kigadye R, Walraven G, Killewo J.  The impact of HIV-1 infection on mortality in children under 5 years of age in sub-Saharan Africa:  a demographic and epidemiologic analysis.  AIDS 1994;8:995-1005
19. Nzyuko S, Lurie P, McFarland W, Leyden W, Nyamwaya D, Mandel JS. Adolescent sexual behavior along the Trans-Africa Highway in Kenya.  AIDS 1997;11 (Suppl 1):S21-S26
20. Taha TE, Kumwenda NI, Broadhead RL, et al.  Mortality after the first year of life among human immunodeficiency virus type 1-infected and uninfected children.  Pediatr Infect Dis J 1999;18:689-694
21. Zaba B, Boerma T, White R.  Monitoring the AIDS epidemic using HIV prevalence data among young women attending antenatal clinics: prospects and problems.  AIDS 2000;14:1633-1645
22. Boerma JT, Nunn AJ, Whitworth AG.  Mortality impact of the AIDS epidemic:  evidence from community studies in less developed countries.  AIDS 1998;12 (Suppl 1):S3-S1

Copyright 2006 - Annals of African Medicine

The following images related to this document are available:

Photo images