Effect of labour market conditions on reporting of limiting long-term illness and permanent sickness in England and Wales.

Abstract

Study objective-To identify any bias in the reporting of limiting long term illness and permanent sickness due to labour market conditions, and show the absence of the effect in mortality rates. Design-A geographically based study using data from the 1991 census. Standardised ratios for mortality and long term illness in people aged 0-64 years and permanent sickness in people of working age were compared with Carstairs deprivation scores in multilevel models which separated the effects operating at three geographical scales: census wards, travel to work areas, and standard regions. Holding ward and regional effects constant, variations between travel to work areas were compared with long term unemployment rates.Setting-Altogether 8690 wards and 262 travel to work areas in England and Wales. Main results-Variations in mortality, limiting long term illness, and permanent sickness were related to Carstairs deprivation scores and standard region. With these relationships controlled, limiting long term illness and permanent sickness were significantly related to long term unemployment levels in travel to work areas, but mortality was not affected. Self reported morbidity was more sensitive to variations in long term unemployment rates in conditions of high social deprivation than in affluent populations. Conclusions-Limiting long term illnessand permanent sickness measures may reflect a tendency for higher positive response in difficult labour market conditions. For average social deprivation conditions, standardised limiting long term illness for people aged 0-64 years was 20% higher in travel to work areas where employment prospects were relatively poor compared with areas with relatively good employment prospects. This casts doubt on the use of limiting long term illness as an indicator of objective health care needs for resource allocation purposes at national level. illness is a selfreported measure and may reflect both "real" health status and other effects, such as differences in the way health is perceived and reported. The relative contribution of such effects to limiting long term illness prevalence rates is therefore of interest, and here we examine the potential influence of labour market conditions on the likelihood of reporting long term illness.The existing evidence of labour market influences on self reported morbidity comes largely from studies of permanent sickness or disability. Data on permanent sickness have been collected in successive UK censuses as a reason given for being unable to work. In the 1991 census the relevant question was: "Which of these things was the person doing last week?" with a range of possible responses including "was unable to work because of long term sickness or disability". Although permanent sickness was one of the few morbidity measures available at the local scale prior to the 1991 census, its use as a measure of general health status has been limited. Only people excluded from employment are covered by the census category, so children, most elderly people, a...

Is There a Place for Geography in the Analysis of Health Inequality?

“…Phillimore and Morris (1991) invoke the differences in environmental impact of the industries in Middlesborough and Sunderland to help to explain differences in health experience in the two towns. Haynes et al (1997) found that for wards in travel-to-work areas with high levels of unemployment, selfreported illness ratios were comparatively high, even allowing for ward deprivation. They suggest that the propensity to report long-term illness may be related to the perceived lack of employment opportunity.…”

“…They recommend multi-level modelling, or some alternative, by which the hierarchical structure of the data can be represented. Research which uses multi-level modelling techniques to represent the simultaneous effects of both compositional and contextual influence more effectively includes: Humphreys and Carr-Hill (1991), Duncan et al (1993), Gould and Jones (1994), Duncan and Jones (1995), Congdon (1994), Shouls et al (1996), Haynes et al (1997), Congdon et al (1997).…”

“…At the national scale, Congdon et al (1997) analysed differences in self-reported long term illness and in mortality using multi-level modelling at ward (level 1) and district (level 2) scales. Haynes et al (1997) also used multi-level modelling to examine the (level 1) ward level relationship between social deprivation and health measured by selfreported illness and mortality and how this varied between travel to work areas (at level 2). Both studies found evidence for variation operating at both (level 1 and level 2), and both suggest that the 'level 2' effects were generally stronger for self-reported illness than for mortality.…”

Interrelations between three proxies of health care need at the small area level: an urban/rural comparison

“…Initial analysis of the 1991 census suggested that LLTI data were reasonably accurate and within a short period LLTI was recommended as a key needs indicator for resource allocation purposes, 7 and it has been adopted as such by the Department of Health. 8 Demographic data have been used to explore health care need but marked social and geographical variation in health status remains after controlling for these. Variations in mortality have been shown to be strongly associated with social deprivation.…”

Changes in socioeconomic inequalities in census measures of health in England and Wales, 1991-2001

“…In contrast, permanent sickness rates in the UK may to be particularly vulnerable to fluctuations in the labour market. 11 There is accumulating evidence that those finding it hard to gain employment, particularly the unskilled, …”