original reports
Relationship between the use of hospital services and deprivation score of place of residence among HIV-positive individuals in the north west of England PA Cook, S Hargreaves, K Tocque, MA Bellis
Summary: We used routine surveillance data to investigate whether deprivation relates to hospital admission in a HIV-positive population. HIV-positive individuals living in the poorest areas were more likely to have spent one or more nights in hospital for HIV-related care (adjusted odds ratio = 1.6, p = 0.009, after controlling for infection route, disease stage and demographic variables). This implies that healthcare networks in poorer areas may incur disproportionately greater costs.
Key words: AIDS deprivation HIV hospital services inequality
Commun Dis Public Health 2004; 7(4): 319-21
Introduction Inequalities in socioeconomic status have long been known to create inequalities in health1. For people who have chronic conditions, poverty may exacerbate their illness (for example, diabetes 2 ). Since treatment became available the view of HIV infection as a chronic condition has been amplified. In 2002, there were 30,281 recipients of HIV treatment in the United Kingdon (UK), of whom 2328 (8%) live in the North West3, one of the most deprived regions in England4. Using surveillance data from the North West, we aimed to determine whether hospital admission was related to deprivation in the HIV-positive population.
Methods Firstly, routine surveillance data were collected as part of the regional monitoring system (enhanced Survey of Prevalent HIV Infections Diagnosed: SOPHID 3) from hospital records for all HIV-positive individuals resident in the North West (n = 2331) attending HIV treatment and care centres in the region during 2002. Sufficient postcode data on place of residence were available for 1934 patients (83%) to allocate ward level
PA Cook, S Hargreaves, K Tocque, MA Bellis Centre for Public Health, Liverpool John Moores University Address for correspondence: Penny A Cook Centre for Public Health Liverpool John Moores University Castle House, North Street Liverpool L3 2AY tel: 0151 231 4517 fax: 0151 231 4515 email:
[email protected]
COMMUNICABLE DISEASE AND PUBLIC HEALTH
index of multiple deprivation scores4. The distribution of patients by their place of residence across three equally-sized deprivation categories was compared between the HIV-positive sample and the general North West population (comprising 1007 wards, average population 6843). Secondly, logistic regression was used to identify the relation between a stay of at least one night in hospital for any HIV-related treatment or care during 2002 (an indicator of health status) and deprivation, controlling for route of HIV infection, ethnicity, sex, age, and stage of disease. A total of 1758 (75%) individuals were included in the analysis after removing those with missing data. Because too few HIV-infected individuals lived in the most affluent North West areas (as above), deprivation scores were recategorised into three equallysized groups within the dataset of people with HIV infection. The analysis was repeated including cases with unknown variables to assess the impact of missing data on the results.
Results In general, HIV-positive people were more likely to live in the poorest areas: 61.3% of the HIV population lived in the poorest third of the North West (chisquared goodness of fit = 682, df = 2, P < 0.0001). When considering relative deprivation in the HIV-positive population, having controlled for the other variables, those living in the most deprived areas were more likely to have required a hospital stay than those living in the most affluent areas. People infected through blood or tissue and injecting drug use and people who were symptomatic, had AIDS, or had died were also more likely to have stayed in hospital (table 1). When the analysis was repeated including cases with missing
VOL 7 NO 4 DECEMBER 2004
319
original reports
TABLE 1 Relations between a stay in hospital of one or more nights for HIV-related care during 2002 and demographic and health variables (univariate chi-squared and logistic regression) Variable
Hospital stay % (number) of participants
Univariate analysis Chi-square test
Ethnicity
0.5
White
14.7
(217)
Non-white
16.2
(57)
Infection route
P value
Multivariate analysis(n=1,758) Adjusted odds ratio
p value
(95% CI)
0.469
0.841 1 0.954
48.7
(0.599 to1.517)
