Journal of Public Health Advance Access originally published online on June 28, 2005
Journal of Public Health 2005 27(3):292-297; doi:10.1093/pubmed/fdi038
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Unsung heroes who put their lives at risk? Informal caring, health and neighbourhood attachment
Stephen Barrow
Stephen Barrow, Senior Research Fellow, Institute for Public Health Research & Policy, University of Salford, The Crescent, Salford M5 4QA
Roger A. Harrison
Roger A. Harrison, Research Scientist in Public Health, Bolton Primary Care Trust, St Peter’s House, Silverwell St, Bolton BL1 1PP
Evidence for Population Health Unit, School of Epidemiology and Health Sciences, University of Manchester, Oxford Rd, Manchester M13 9PT
Address correspondence to Stephen Barrow. E-mail: s.barrow{at}salford.ac.uk
Background Previous studies have highlighted the effects of informal caring on mental health, but evidence for its wider impact remains scant. We explored associations between informal care and a range of health (and health-related) measures. We also considered previously neglected differences between informal care ‘at home’ and care elsewhere, along with neighbourhood attachment as a possible modifier of the associations we found.
Methods The study involved a large population survey in two Primary Care Trusts. Data were collected by postal survey of 15465 adults; subjects were selected from the local General Practice register. 12.7 per cent of respondents identified themselves as carers. Health measures included psychiatric morbidity, bodily pain, self-assessed health, health-related behaviours, obesity, prescribed drugs and high levels of GP consultation.
Results Of nine measures considered, care at home was associated with psychiatric morbidity (OR 1.46, 95 per cent CI 1.25–1.70), bodily pain (OR 1.19, 95 per cent CI 1.02–1.39), and obesity (OR 1.59, 95 per cent CI 1.34–1.89). Care ‘away’ was associated with smoking (OR 1.26, 95 per cent CI 1.03–1.54), and inversely with both sedentary living (OR 0.70, 95 per cent CI 0.58–0.85) and poor self-assessed health (OR 0.78, 95 per cent CI 0.62–0.99). Health tended to be poorer when carers lacked a sense of neighbourhood attachment.
Conclusions Informal carers are likely to face serious health challenges besides anxiety and depression. Caring is associated with several aspects of poor health, which are themselves predictors of premature mortality. Proactive and wide-ranging support is required, the more so in neighbourhoods where carers feel alienated. Research and policy should distinguish carers ‘at home’ from carers ‘away’.
Keywords: informal carers, place of delivery, neighbourhood attachment, premature mortality
 |
Introduction |
|---|
 TOP Introduction MethodsResultsDiscussionAcknowledgementsReferences |
|---|
In the UK, seven million adults provide unpaid help to relatives or neighbours with chronic illness or disability.1 The extent of this informal care is increasing,2 along with a commensurate recognition of its social value.3 Informal caring enhances the ‘social capital’ of an area, and high levels of social capital are considered (although not by all) to have beneficial effects on population health.4–6
1 Informal carers, however, are not without their own needs for support. A growing literature suggests that carers are more likely than non-carers to suffer poor health.7 This has been recognized in recent government initiatives, including the right of carers to seek a formal assessment of service needs.8 And yet our understanding of informal care remains scant in many respects. Aspects of health other than mental state are seldom reported.7 ‘Place of delivery’ is similarly neglected, although carer health is likely to differ when care is provided within, rather than outside, the carer’s own home.9
We investigated the association between informal caregiving and several measures of health and health-related behaviour. We distinguished care ‘at home’ from care ‘away’, hypothesizing that carers in the former setting would have the greater risk of poor health. A further novelty was our consideration of neighbourhood attachment, a dimension of social cohesion with implications for population health.10 We hypothesized that any association between caring and poor health would be attenuated by a sense of belonging to the community in which the care was provided.
|
Methods |
|---|
|
TOPIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Sample
The present study is cross-sectional in design. Data were collected in 2001 through a population-based health and lifestyle survey in the metropolitan boroughs of Bolton and Wigan, north-west England. The collection of data on informal care was only one aim of the survey, the exercise being typical of general community health surveys in the NHS. Postal questionnaires were sent to a sample of 21923 adults (or 5 per cent of the total population aged 18 or over). Potential respondents were selected by systematic sampling from a random start-point in the general practice register. We had to rely on health authority sampling procedures, which precluded access to a completely random sample. Non-responders were sent two postal reminders. The survey was approved by the local ethics committees. Further details of the procedures are given elsewhere.11–13
Measures
Respondents identified themselves as carers by confirming that they cared ‘for someone with a long-term illness or disability, other than as part of a job’. Details were recorded of the number of people supported, along with the locus of that support (‘caring in own home’ or ‘caring away from home’).
Psychological morbidity was assessed with the 12-item General Health Questionnaire (GHQ), a measure validated for the screening of anxiety and depression in community samples.14 ‘Caseness’ was defined by a score of three symptoms or more. General health was assessed by asking respondents to rate their health over the last 12 months in terms of five categories, which ranged from ‘excellent’ to ‘poor’. This measure of ‘Self-Assessed Health’ (SAH) is a commonly recognized predictor of mortality.15 Here it was recoded as ‘good SAH’ (excellent/very good/good) and ‘poor SAH’ (fair/poor). Respondents were also classified by the level of bodily pain they had experienced in the last 4 weeks. We distinguished ‘pain’ (i.e. moderate/severe/very severe) and ‘no pain’ (none/mild/very mild).
Respondents were considered to have ‘unhealthy diets’ if they reported fewer than four from six established dietary indicators of health.16 ‘Sedentary lifestyles’ were said to characterize those who took part in fewer than three 15-minute sessions per week of moderate or strenuous physical activity, based on the Godin scale.17 In addition, respondents were classified by smoking status as either ‘smokers’ or ‘non-smokers’.
Several other health measures were investigated. These were obesity (defined as a Body Mass Index score of 30 or more), ‘high’ levels of general practitioner visiting (taken to be six or more consultations per year), and whether or not the respondent was taking any prescribed drugs or medicines (excluding the contraceptive pill).
Neighbourhood attachment was assessed by asking respondents for their level of agreement with the statement ‘I feel like I belong to this neighbourhood’ (adapted from Lochner etal.).18
Socio-demographic factors included sex, age, employment status, tenure and ethnic group (recoded as ‘White’ or ‘Non-white’). Additional measures of socio-economic status were perceived financial position (‘living comfortably’, ‘doing alright’, ‘just about getting by’, ‘finding it difficult’ and ‘finding it very difficult’), and residential level of deprivation (based on quintiles of the Townsend Index for enumeration districts).
Analysis
The data were analysed using SPSS 12.0. The characteristics of caring were first analysed in terms of basic statistics. Logistic regression was then used to examine associations between caregiving and each of our nine health measures. Caregiving was analysed as a single, three-category variable to include the locus of care (none, away, at home). Associations were described in terms of odds ratios (ORs) along with their 95 per cent confidence intervals (CIs). Caregiving was initially considered on a univariate basis with ORs adjusted only for age and gender. The reference group was composed of people with no involvement in the provision of informal care. Relations with health measures were then explored in multivariate analyses with caregiving adjusted for all seven of our social factors (age, gender, employment, tenure, ethnic group, financial position and residential deprivation). All variables were categoric and entered simultaneously in every regression. The analyses at each stage included all cases for which the relevant data was available.
Further regressions were undertaken to investigate neighbourhood attachment as a potential modifier of associations between caregiving and health. A variable was created to represent combinations of caregiving (none, away, home) and sense of neighbourhood belonging (belongs, doesn’t belong). Following Cannuscio etal.,19 this approach was adopted as an alternative to the use of multiplicative interaction. Odds ratios and (95 per cent CIs) were calculated to describe associations between each combination and our several health measures. For this purpose we retained only those measures of health that were significantly related to caregiving in our previous analysis. Estimates were adjusted for the same social factors as before. In this analysis the reference group consisted of respondents with no caregiving responsibilities and a sense of belonging to their local communities.
|
Results |
|---|
|
TOPIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Of the 21923 adults sent a questionnaire, 15465 (70.5 per cent) returned a useable reply. Men accounted for 45.2 per cent of responders, the mean age was 49.8 years (SD 17.6, range 18–102 years), and the majority described themselves as ‘white’ (95.5 per cent). Responders and non-responders were similar in terms of age and sex, but no other information for comparison was available.
The question on caring was answered by 15012 (97 per cent) of the survey’s 15465 respondents. Of these, 1913 (12.7 per cent) reported themselves as caregivers, the majority of whom (83.9 per cent) supported just one other person. Thirty-six per cent of carers provided help away from home; almost two-thirds (64 per cent) assisted someone within their own household. Neighbourhood attachment did not differ significantly between carers away (with 61.8 per cent ‘belonging’), carers at home (63.4 per cent), and non-carers (62.0 per cent) (
2=0.71, df=2, p=0.70).
Table 1 shows prevalence data for our nine outcome measures, with details according to care status (none, care away, care home). Table 2 gives odds ratios and confidence intervals for associations between health and caregiving in both the univariate and multivariate situations (rows OR1 and OR2, respectively). In the latter analysis, care at home was independently associated with elevated levels of psychiatric morbidity. Such carers were 46 per cent more likely to score over a GHQ threshold of three symptoms than respondents with no informal care responsibilities (OR 1.46, 95 per cent CI 1.25–1.70). (In a sensitivity analysis using a GHQ threshold of four symptoms, the equivalent figures were OR 1.41, 95 per cent CI 1.19–1.67.) Care at home was also associated with elevated levels of obesity (OR 1.59, 95 per cent CI 1.34–1.89) and bodily pain (OR 1.19, 95 per cent CI 1.02–1.39). Conversely, there was no independent association between care at home and general health, sedentary living, poor diet, smoking, prescribed drugs and high levels of GP consultation.
|
|
The potential health impacts of care away from home were very different. Such caregivers were 26 per cent more likely to smoke than respondents with no caring responsibilities (OR 1.26, 95 per cent CI 1.03–1.54). Conversely, ‘care away’ was inversely associated with poor general health (OR 0.78, 95 per cent CI 0.62–0.99), along with sedentary living (OR 0.70, 95 per cent CI 0.58–0.85). There were no independent associations between care away and any other health measure, including those for which care at home had been a risk factor.
Table 3 gives fully adjusted odds ratios and 95 per cent confidence intervals for combinations of informal care and neighbourhood belonging. A lack of belonging was associated with notably poorer outcomes for home carers in three of the measures considered. Thus elevated levels of psychiatric morbidity were evident for carers at home whatever their sense of neighbourhood attachment. However, the elevation was markedly greater for home caregivers with no sense of belonging within the community. These isolated home carers were more than twice as likely to score above the GHQ threshold than non-carers who reported a sense of belonging (OR 2.11, 95 per cent CI 1.64–2.70, for three symptoms on the GHQ). By contrast the home caregivers who ‘belonged’ were only about 40 per cent more likely to experience mental distress (OR 1.43, 95 per cent CI 1.17–1.75). Similar results were obtained in an analysis using a threshold of four symptoms on the GHQ.
|
Neighbourhood attachment also appeared to modify the association between home caregiving and bodily pain. Elevated levels were only evident for those carers who felt a lack of belonging (OR 1.38, 95 per cent CI 1.08–1.78); there was no association for home carers with a greater sense of attachment. Lack of belonging had a similarly modifying effect on the relation between home caregiving and self-assessments of health in general. Such carers experienced a significant elevation in ‘poor’ health compared with respondents with no care responsibilities and a feeling of neighbourhood attachment (OR 1.36, 95 per cent CI 1.04–1.77). By contrast there was no association between home care and poor health for carers who ‘belonged’ in their local communities.
For two measures, i.e. obesity and smoking, the relation with home caregiving was not modified by level of neighbourhood attachment. However, a sense of belonging was associated with an improved outcome for home carers in our final measure. Thus home care was associated with a reduced level of sedentary living when it was accompanied by a sense of neighbourhood belonging (OR 0.81, 95 per cent CI 0.67–0.98).
The potential impact of care away from home was also modified by neighbourhood attachment. Care away was associated with an elevated level of psychiatric morbidity when it was undertaken without a sense of belonging. On this occasion the relation only emerged for the four-symptom GHQ (OR 1.56, 95 per cent CI 1.10–2.22); there was no association when the caring was undertaken in circumstances of greater attachment. ‘Belonging’ had a similarly modifying effect on the relation between care away and self-assessed health. Providing such care was inversely associated with poor health but only when carers felt integrated within their local neighbourhood (OR 0.72, 95 per cent CI 0.53–0.99)
|
Discussion |
|---|
|
TOPIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
Many studies of informal care have considered only one health outcome, most typically a psychological measure such as depression or ‘caregiver strain’.20 Our study is unique in reporting a range of important health and health-related outcomes. We also based our investigation on a large sample with data collected at the population level. Previous studies have often been based on small numbers and specific groups of patients.
We were correct in our hypothesis that outcomes would vary between care that is provided in the caregivers own home, and care that is provided elsewhere. In our data the former tend to have the poorer health. Thus care at home is generally associated with elevated levels of (potential) psychiatric morbidity, bodily pain and obesity. All three factors are important predictors of premature mortality.21–23 The potential impact of care away from home is less clear. Carers in this group are likely to have poorer outcomes than non-carers in only one of our nine measures, i.e. smoking. Moreover, the providers of care away from home are more likely than non-carers to be physically active and to rate their own health status as ‘good’. The extent to which this is causative is impossible to ascertain in a survey of this sort. Indeed, it may arise from a ‘healthy worker effect’ in which people with good health are inclined to support others less able.24
Our study is also the first to consider whether the potential impact of informal care might be modified by carers’ sense of neighbourhood attachment. Here we find that ‘home carers’ face greater health risks when they lack a sense of neighbourhood belonging: such carers have an increased likelihood of psychiatric morbidity, poor general health and bodily pain. A lack of belonging is also associated with elevated levels of psychiatric morbidity for those who care outside the home.
Limitations of our study should be mentioned. These include the possibility of affect bias arising in the analysis of neighbourhoods and health, data for both being provided by the same respondents. Moreover, the high prevalence of some outcomes (Table 1) means that odds ratios are likely to overestimate the true risks involved.25 Finally, our cross-sectional design makes it difficult to be sure about causation and the direction of relationships. For this reason we would hope that the issues we raise might be further explored in longitudinal studies of the health effects of informal care. (Further research would also be useful concerning variations by place of delivery in the types of people who are supported by informal carers; the Bolton and Wigan survey did not collect sufficient detail for this to be explored.)
Despite these limitations our results have some implications for action. First, it seems important to distinguish care ‘at home’ and care ‘away’ in the development of strategies for carer support and in future research and epidemiological studies. This distinction has often been neglected previously, for example in the items on carers in the 2001 UK Census26 and studies based thereon (such as Doran etal.).27 The distinction should also be made in the assessment of individual carers (a process which is often unsystematic if it exists at all).28 Secondly, our results appear to strengthen the case for social policies that promote social cohesion in local neighbourhoods; these are likely to be of benefit to informal carers as well as the community in general. However, we should not expect the act of caring to create its own sense of community; in our data neither carers at home nor carers away have any greater sense of neighbourhood belonging than non-carers.
Attempts to address the needs of carers are somewhat problematic, of course. Many carers do not request the assessments to which they are entitled (or even recognize themselves as ‘carers’).29 This may explain the patchy implementation of many support initiatives.7 It is also consistent with our finding that carers are no more likely than non-carers to have high rates of GP consultation. Unfortunately, any shortfall in support may be greater than anticipated. We find that carers are likely to face other challenges than the anxiety and depression that have been the focus of many studies in this field. High levels of obesity and bodily pain are particularly worrying outcomes. This leads to our third recommendation. Carers need wide-ranging support to offset the problems they face, especially in community environments in which they feel alienated. Moreover, the support should be provided in a more proactive way than at present.
|
Acknowledgements |
|---|
|
TOPIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
We would like to thank Bolton Primary Care Trust and Ashton, Wigan and Leigh Primary Care Trust for access to this data, especially David Holt, the survey manager.
|
References |
|---|
|
TOPIntroductionMethodsResultsDiscussionAcknowledgementsReferences |
|---|
- Maher J, Green H. Carers 2000. London: The Stationery Office, 2002.
- Hirst M. Transitions to informal care in Great Britain during the 1990s. J Epidemiol Commun Hlth 2002; 56: 579–587.
[Abstract/Free Full Text] - Department of Health. Caring about carers: a national strategy for carers. London: Department of Health, 1999.
- Cooper H, Arber S, Fee L, etal. The influence of social support and social capital on health. London: HEA, 1999.
- McCulloch A. Social environments and health: cross sectional national survey. Br Med J 2001; 323: 208–209.
[Free Full Text] - Adams J, White M. Evidence concerning social capital and health inequalities is still lacking [Letter]. J Publ Hlth Med 2003; 25: 184–185.
[Free Full Text] - Hirst M. Caring-related inequalities in psychological distress in Britain during the 1990s. J Publ Hlth Med 2003; 25: 336–343.
[Abstract/Free Full Text] - The Carers (Recognition and Services) Act 1995. London: HMSO, 1995.
- Horsley S, Barrow S, Gent N, Astbury J. Informal care and psychiatric morbidity. J Publ Hlth Med 1998; 20: 180–185.
[Abstract/Free Full Text] - Pevalin D. Intra-household differences in neighbourhood attachment and their associations with health. In: Morgan A, Swan C, eds. Social capital for health: issues of definition, measurement and links to health. London: Health Development Agency, 2004.
- Harrison RA, Holt D, Elton P. Do postage stamps increase response rates to postal health surveys: a randomised controlled trial. Int J Epidemiol 2002; 31: 872–874.
[Free Full Text] - Harrison RA, Holt D, Pattison D, etal. Who and how many people are taking herbal supplements? A survey of 21,923 adults. Int J Vit Nutr Res 2004; 74: 183–186.[CrossRef]
- Harrison RA, Holt D, Pattison D, etal. Are those in need taking dietary supplements? A survey of 21,923 adults. Br J Nutr 2004; 91: 617–623.[CrossRef][Web of Science][Medline]
- Goldberg D, Williams P. A user’s guide to the General Health Questionnaire. Windsor: NFER-Nelson, 1991.
- Burstrom B, Fredlund P. Self rated health: is it as good a predictor of subsequent mortality among adults in lower as well as in higher social classes? J Epidemiol Commun Hlth 2001; 55: 836–840.
[Abstract/Free Full Text] - Steingrimsdottir L, Ovesen L, Moreiras O, Jacob S. Selection of relevant dietary indicators for health. Eur J Clin Nutr 2002; 56 (Suppl 2): S8–11.
- Godin G, Shephard RJ. A simple method to assess exercise behavior in the community. Can J Appl Sport Sci 1985; 10: 141–146.[Medline]
- Lochner K, Kawachi I, Kennedy BP. Social capital: a guide to its measurement. Hlth Place 1999; 5: 259–270.
- Cannuscio CC, Colditz GA, Rimm EB, etal. Employment status, social ties and caregivers’ mental health. Social Sci Med 2004; 58: 1247–1256.
- van den Heuvel ETP, de Witte LP, Schure LM, etal. Risk factors for burn-out in caregivers of stroke patients, and possibilities for intervention. Clin Rehabil 2001; 15: 669–677.
[Abstract/Free Full Text] - Robinson KL, McBeth J, Macfarlane GJ. Psychological distress and premature mortality in the general population: a prospective study. Ann Epidemiol 2004; 14: 467–472.[CrossRef][Web of Science][Medline]
- Macfarlane GJ, McBeth J, Silman AJ. Widespread body pain and mortality: prospective population-based study. Br Med J 2001; 323: 1–4.
[Free Full Text] - Hu FB, Willett WC, Li T, etal. Adiposity as compared with physical activity in predicting mortality among women. N Engl J Med 2004; 351: 2694–2703.
[Abstract/Free Full Text] - McMichael AJ. Standardised mortality ratios and the ‘healthy worker effect’: scratching beneath the surface. J Occup Med 1976; 18: 165–168.[Web of Science][Medline]
- Davies HTO, Crombie IK, Tavakdi GM. When can odds ratios be misleading? Br Med J 1998; 316: 989–991.
[Free Full Text] - Office for National Statistics. Census 2001. ONS, 2003.
- Doran T, Drever F, Whitehead M. Health of young and elderly informal carers: analysis of UK census data. Br Med J 2003; 327: 1388.
[Free Full Text] - Syron M, Shelley E. The needs of informal carers: a proposed assessment tool for use by public health nurses. J Nursing Mgmt 2001; 9: 31–38.[CrossRef]
- Travers AF. Caring for older people: carers. Br Med J 1996; 313: 482–486.
[Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||