Social Science & Medicine 53 (2001) 795–799 Short Report Do area-level population change, deprivation and variations in deprivation affect individual-level self-reported limiting long- term illness? Paul J. Boyle a, *, Anthony C. Gatrell b , Oliver Duke-Williams c a School of Geography and Geosciences, University of St Andrews, St Andrews, Fife KY16 9AL, UK b Institute for Health Research, Lancaster University, Lancaster LA1 4YT, UK c School of Geography, University of Leeds, Leeds LS2 9JT, UK Abstract A previous study showed that variations in deprivation within small localities in England and Wales influenced the rates of self-reported limiting long-term illness, controlling for overall levels of deprivation. These results suggest that while morbidity is related to overall levels of material deprivation, the distribution of resources within small areas have a significant effect on health outcomes. However, it is possible that these area effects become redundant once individual- level characteristics are accounted for. This analysis examines whether area-level deprivation and variations in deprivation are significant indicators of individual-level limiting long-term illness, once individual characteristics have been accounted for. # 2001 Elsevier Science Ltd. All rights reserved. Keywords: Limiting long-term illness; Health variations; England and Wales; Logit models; Material deprivation Introduction Previous studies have indicated that mortality is affected by levels of inequality within areas, as well as overall measures of poverty (Kawachi, Kennedy, Loch- ner, & Prothrow-Stith, 1997; Kennedy, Kawachi, & Prothrow-Stith, 1996; Rodgers, 1979; Waldman, 1992; Wilkinson, 1993, 1996). Ben Shlomo, White, & Marmot, (1996) show that this is the case in England and Wales at the scale of the local authority district. We might anticipate that variations in socio-economic conditions would be even more directly associated with morbidity than with mortality (Davey, 1996). Previous work showed that ward-level standardised morbidity rates were positively associated with ward-level deprivation as well as variations in deprivation within the ward and its neighbours (the locality) measured at the enumeration district level (Boyle, Gatrell, & Duke-Williams, 1999). However, area-level results cannot be used to imply that deprived people necessarily have higher rates of self-reported limiting long-term illness, as this would be an ecological fallacy. It has been shown that area-level deprivation measures become redundant once individual characteristics are accounted for in analyses of mortality (Fiscella & Franks, 1997; Sloggett & Joshi, 1994), although this is not the case for morbidity (Gould & Jones, 1996; Gleave, Bartley, & Wiggins, 1998; Shouls, Congdon, & Curtis, 1996). No studies have examined whether variations in deprivation remain significant in explaining morbidity once individual-level characteris- tics are controlled for and this is redressed here. We aimed to determine whether the odds of an individual reporting limiting long-term illness was significantly related to area-level deprivation and variations in deprivation within SAR-areas, once individual-level characteristics were accounted for. In addition, we also *Corresponding author. Fax: +44-1334-463949. E-mail address: p.boyle@st-andrews.ac.uk (P.J. Boyle). 0277-9536/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved. PII:S0277-9536(00)00373-7