Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Socioeconomic status and multimorbidity: a systematic review and meta‐analysis

Socioeconomic status and multimorbidity: a systematic review and meta‐analysis The increasing prevalence of chronic conditions1,2 and the growth of the ageing population has led to an increase in multimorbidity worldwide.3 In this article, we distinguish between multimorbidity (the co‐existence of two or more chronic conditions) and co‐morbidity (co‐occurrence of disease/s with a specific index disease). The impact of multimorbidity on the health and wellbeing of individuals, the burden on healthcare systems and the effect on economies has created a major global public health problem. A recent systematic review reported that the prevalence of multimorbidity among the general adult population in high‐income countries ranges from 12.9% in participants aged 18 years and older to 95.1% in a community‐dwelling elderly population aged 85 years.4 However, while the prevalence of multimorbidity is highest among the oldest (above 85 years of age), the growing burden of multimorbidity among older adults still of working age and among lower socioeconomic groups in some countries is of particular public health concern.5,6This expansion of morbidity is leading to individuals living longer but with more co‐existing chronic disease from a younger age, placing an even greater burden on healthcare systems.While multimorbidity incidence and prevalence is known to vary by measures of socioeconomic status (SEP), with an excess burden in lower socioeconomic groups, there are some gaps in our understanding of this relationship. It is unclear whether the association is true for all SEP measures and whether the magnitude of association varies by age, gender and country. Since different SEP variables measure different aspects of individual circumstances and environmental characteristics, a better understanding of which socioeconomic factors are more strongly associated with multimorbidity may help us to better understand the underlying mechanisms. In turn, this will help inform the design of intervention approaches aimed at preventing or reducing the development of multimorbidity.ObjectivesThe objective of this review was to systematically identify, critically appraise and synthesise the existing literature on the association between SEP and multimorbidity occurrence.MethodsThis manuscript was prepared in accordance with the PRISMA guidelines.7 The protocol for this review was not registered.Search strategyWe sought studies published in English in Medline and Embase between 1946 and 1974, respectively, and December 2014 reporting on the association between SEP and multimorbidity using a comprehensive electronic search strategy (Supplementary file 1: Appendix A) and perusal of reference lists of all relevant identified articles. We included terms for multimorbidity and comorbidity, since these are often used interchangeably. One author (TP) screened all references by title and abstract and, where necessary, the full text of the article. All potentially relevant articles were reviewed by the co‐author, and any disagreements on inclusion were resolved through discussion between the two authors and consensus was reached.Inclusion and exclusion criteriaWe included studies of any quantitative design that reported on the occurrence of multimorbidity (defined by the co‐occurrence of multiple conditions) with respect to any measure of SEP, in adult populations only. We excluded: qualitative studies; studies that included selected populations (e.g. patients with psychiatric conditions only, substance abuse problems or those who had undergone a specific medical procedure); and studies that reported on co‐morbidity rather than multimorbidity. Where multiple articles on overlapping study populations were identified, we included the study with the largest population.Data extractionBoth authors independently extracted information on: study design; study population; demographics; sample size; exclusion criteria; definition and ascertainment of multimorbidity; number of diseases ascertained; inclusion of mental health among chronic diseases; ascertainment and measure/s of SEP; and results, including numbers with and without multimorbidity, for the purpose of meta‐analysis, where appropriate.Data synthesisWe aimed to narratively summarise study findings or, where possible, combine study‐specific estimates of effect using meta‐analysis to obtain a pooled summary estimate. Due to the substantial variation in how SEP measures were defined, and/or the different methods of analysis used, meta‐analysis was only possible for the association between education and multimorbidity. This SEP measure was the most consistently defined and a sufficient number of studies adopted the same statistical analysis approach (i.e. logistic regression) to allow us to formally pool the results.Meta‐analysisWe performed meta‐analysis using Stata version 13 and adhered to the MOOSE guidelines.8 We combined studies that either reported unadjusted odds ratios (ORs) with 95% confidence intervals (CIs) for the association between education and multimorbidity or presented raw numbers that allowed us to calculate ORs. Where we extracted raw numbers, we defined the non‐multimorbidity comparison group as participants with fewer than two conditions, in line with the most common comparison group used across studies. Our rationale for pooling together minimally or unadjusted ORs was firstly to harmonise findings from studies, in order to facilitate inclusion of as many studies as possible. Studies varied in terms of: the type of effect estimates presented; whether they presented unadjusted and/or adjusted estimates; and specific factors adjusted for, and reference category used for effect estimates for education. Secondly, our interest lay in determining the association between SEP and multimorbidity without seeking to identify the mechanisms underlying the association. Extraction of only effect estimates adjusted for additional factors, some of which may lie on the causal pathway between SEP and multimorbidity, would therefore have potentially obscured the true nature of the association between SEP and multimorbidity. We obtained a pooled summary OR for the odds of multimorbidity comparing low versus high education using the Mantel Haenszel random effects model, assessing heterogeneity between studies using the chi2 (Cochrane Q) and I2 statistics. We sought to explore potential explanations for any observed heterogeneity using sub‐group analysis. We aimed to assess the influence of three a priori determined study characteristics: age of study population; sex; and method of ascertainment of chronic disease (i.e. objective versus subjective, as described below). Where relevant studies on education and multimorbidity did not provide sufficient information to be included in the meta‐analysis, we contacted the authors to obtain results in the necessary format. However, none of the authors who replied were able to provide this information, because they no longer had access to the datasets.ResultsWe identified 2,496 articles, 63 of which were potentially relevant and underwent full‐text review. Of 28 relevant studies, two were excluded9,10 because the study populations overlapped with a third included study.11 Two studies were excluded because they defined multimorbidity using a score that incorporated severity of disease (Figure ).12,13 The remaining 24 included studies were cross‐sectional in design or entailed cross‐sectional analysis of SEP factors related to multimorbidity (Table ).5,6,11,14–34 In general, participants were recruited through population‐based primary care databases or national or regional surveys, with more than half (N=15) conducted in high‐income countries. The number of chronic diseases included in each study ranged from 5 to 335 (Supplementary file 2: Appendix B), with just 12 studies reporting mental health diseases were included (Table ).5,6,11,14,19,20,23,24,28,29,31,33 Ten studies5,11,22–25,28,29,31,33 ascertained multimorbidity through objective sources such as health records (which capture doctor‐diagnosed conditions), while the remainder relied on self‐report of doctor‐diagnosed conditions by participants, which may be subject to recall error (Table ). Multimorbidity was defined in 18 studies as the co‐occurrence of two or more conditions, and in one study as the co‐occurrence of three or more conditions, with the comparison group being fewer than two (or three) conditions (Table ). The exceptions to this were the studies by Jerlui et al. and Marengoni et al. in which multimorbidity was compared to single morbidity.21,24 In the remaining five studies, multimorbidity was analysed as a continuous count of conditions, without a cut‐off being employed.Flow diagram of literature search and included studies.Characteristics of identified studies.Author, yearCountryStudy populationTotal NMale (%)Age (years)aDefinition of multimorbidity [comparison group]Ascertainment of morbiditiesNumber of conditions includedMental health includedAgborsangaya, 201214CanadaA representative sample of the adult (aged ≥18 years) population of the province of Alberta, as identified in the 2010 Patient Experience Survey by Health Quality Council of Alberta5,01047.746.7≥2 concurrent conditions [<2 conditions]Self‐report via telephone‐administered survey questionnaire16YAhluwalia, 200515USNon‐institutionalised, non‐pregnant women aged 18–44 years, identified from the Behavioural Risk Factor Surveillance System, an ongoing random digit dialled telephone survey of 50 States;163,5660Mean NR Range 18–44≥2 concurrent conditions [no formal statistical comparison made]Self‐report via telephone‐administered survey questionnaire5NAlaba, 201316South AfricaStratified two‐stage cluster sampling of 53 district councils of South Africa, including adults aged ≥15 years; for this study adults aged ≥18 were included11,6383940≥2 concurrent conditions [<2 conditions]Self‐report of doctor‐diagnosed condition6NAndrade, 201017BrazilNon‐institutionalised adults aged ≥18 years included in a household survey of two boroughs of Sao Paulo1,46457.4Mean NR (82% aged <60)Morbidity included in modelling as a continuous variableFace‐to‐face interviews, with information on morbidities obtained via self‐reported doctor diagnosis of various conditions8NbBarnett, 20125ScotlandPatients of any age registered with one of 314 GP practices, covering one third of the Scottish population1,751,84149.5Mean NR (83% aged <65)≥2 concurrent conditions [<2 conditions]GP medical records and prescription data40YDroomers, 200418NetherlandsRespondents aged 25 or over included in the Netherlands Health Interview Surveys from 1990–1998 which included non‐institutionalised persons53,339NRMean NR (all ≥25)≥2 concurrent conditions [<2 conditions]Self‐report in questionnaire14NEnroth, 201319FinlandSurvey of adults aged 90+ living in Tampere, irrespective of dwelling place (80% of eligible population responded)1,28319Mean NR (All ≥90)≥2 concurrent conditions [<2 conditions]Self‐report of doctor‐diagnosed conditions6YHosseinpoor, 20122041 low and middle‐income countriesData from the 2002–04 World Health Survey of adults aged ≥18 years in 41 low and middle‐income countries; surveys are nationally representative except in 6 countries where WHS is conducted in geographically limited regions170,29845.5Mean NR (all >18)≥2 concurrent conditions [<2 conditions]Self‐report in questionnaire; symptom‐based classification based on symptoms in previous 12 months used for all conditions except diabetes, which was reported directly; diabetes by self‐report5YHudon, 201213CanadaRegular patients aged ≥18 years attending a family medicine clinic (primary care) of a regional health center in Saguenay, Quebec1033549.9Used a multimorbidity scoreSelf‐reported questionnaire using simplified version of the Disease Burden Morbidity Assessment, including 11 of 21 diseases included in the original instrument, based on the high prevalence of these diseases in this setting11NJerliu, 201321KosovoA nationwide, population‐based representative sample of adults aged ≥65 years1,89050.273.4≥2 concurrent conditions [1 condition]Self‐report via an interviewer‐administered questionnaireUnclear*Unclear*Khanam, 201122BangladeshRandomly selected individuals age >=60 from two areas of a rural area of Bangladesh, based on the sampling frame of the Health and Demographic Surveillance System45245.169.5≥2 concurrent conditions [<2 conditions]Clinical examination by physicians, with independent evaluation by senior physicians/geriatricians9NMacleod, 200423ScotlandPatients aged 18 or over registered at one practice in an area of high deprivation in Glasgow City7,286NRMean NR (76.7% aged <65)≥2 concurrent conditions [<2 conditions]GP records16YMarengoni, 200824SwedenParticipants in the first follow up of the Kungsholmen Project, a study on elderly people (aged ≥75 years in October 1987) living in Stockholm.1,099?NR84.6≥2 concurrent conditions [1 condition]Clinical assessment by a physician, medical history (from Stockholm inpatient register that records discharge diagnoses from Stockholm hospitals), laboratory data and current drug use.30YMercer, 200725ScotlandConsecutive unselected patients of one GP nominated from medium‐sized practices in the upper or lower quartiles of deprivation in 4 healthboard regions in West of Scotland; 26 GPs included3,044NR44.6**Morbidity assessed as a count, not a cut‐offGP recordsNR, but based on GP records so presumably all conditions includedNRNagel, 200826GermanyHeidelberg cohort of European Prospective Investigation in to Cancer and nutrition (EPIC) recruited from general population. Cross‐sectional analysis of multimorbidity using data on prevalent chronic conditions collected at baseline and during follow‐up13,78152.9Median 55–58 across male and female subgroups≥2 concurrent conditions [<2 conditions]Prevalent diseases identified at recruitment by face to face interviews; new diseases identified by active follow up through questionnaire on physician diagnosed chronic conditions. Cancer cases verified with medical records and validation studies conducted for incident cases of diabetes, MI, asthma and stroke24NNeeleman, 200127NetherlandsParticipants of the 1996 wave of NEMESIS, a study of the incidence and prevalence of mental disorders in the general population; random sampling of 90 municipalities of adults aged 18–64 years7,076NRNRMorbidity assessed as a count, not a cut‐offSelf‐report based on interviewer‐administered questionnaire30NbOrueta, 201328SpainAdults aged ≥65 years covered by public health insurance in the Basque country on 31st August 2011 for at least 6 months in the previous year, regardless of whether or not they had made any contact with or use of Basque Health Service452,69842.5Mean NR≥2 concurrent conditions [<2 conditions]Primary care electronic medical records, hospital admissions, outpatient care databases and prescription data47YSalisbury, 201129EnglandRandom sample of patients aged 18 or over registered with one of 18 General Practitioners on index date of 1 April 200599,997NRNR≥2 concurrent conditions [<2 conditions]GP records19YSantos Machado, 201330BrazilSimple random sampling of women aged ≥50 living in 68 census districts of the city of Campinos, Sao Paulo6220Mean NR (61.3% aged >60)≥2 concurrent conditions [<2 conditions]Self‐report based on interviewer‐administered questionnaire12NSchafer, 201231GermanyPatients born between 1st July 1923 and 30th June 1943 who consulted GP at one of 158 General Practices in 8 study centres were randomly selected3,18959.374.4Morbidity included in modelling as a continuous variableGP interviews using standardized instrument covering 46 chronic conditions46YTaylor, 20106AustraliaParticipants aged ≥18 years from the North West Adelaide Study, a population‐based cohort study; included those from stage 2 of the cohort study, which involved all or a combination of: a computer assisted phone interview; self‐complete questionnaire; and a biomedical examination at a clinic3,206NRMean NR (75% aged <60)≥2 concurrent conditions [<2 conditions]Self‐report of doctor diagnosed condition7YTucker‐Seeley, 201132USRespondents aged ≥50 years from 2004 wave of the Health and Retirement study7,30546.465≥2 concurrent conditions (for this review) [<2 conditions]Self‐report of doctor diagnosed condition6NUijen, 200833NetherlandsPatients of all ages enlisted in the Continuous Morbidity Registration, Nijmegen, which includes 4 General Practices, including 10 GPs with approximately 13,500 enlisted patients;13,584NRMean NR Range 0 – >75Examined distribution of 0,1,2,3 and ≥4 conditions by SEPGP recordsNRcYVan den Akker, 199811NetherlandsParticipants identified from Registration Network Family Practices, a continuous computerised primary care database including 42 GPs in 15 different practices60,85748.7Mean NR (80% aged <60)≥2 concurrent conditions [<2 conditions]GP records335YWalker, 200734AustraliaParticipants aged ≥ 20 years in the Australian 2001 National Health Survey or the 2003 Survey of Disability Ageing and Carers, both of which are nationally representative17,45045Mean NR≥3 concurrent conditions [<3 conditions]Self‐report in household surveysNRNR* Survey included questions on which groups of diseases participants had and also included an open‐ended question on other chronic conditions** Weighted mean calculated from data presented in papera: Mean unless specified otherwiseb: Psychiatric conditions analysed separatelyc: Total number of conditions not reported, but included all obligatory and conditionally registered chronic diseases recorded in the Continuous Morbidity RegistrationNR = not reported; GP = general practitioner; MI = myocardial infarction; SEP = socioeconomic positionMost studies (N=17; 359,507 participants) measured SEP using education.6,11,14–21,24,26,27,30–32,34 Eight studies (209,186 participants) reported on income6,14,16,17,20,27,31,32 and six (2,332,316 participants) reported on deprivation.5,23,25,28,29,34 In addition, one study reported on literacy,22 three on occupational social class,24,31,33 one on non‐defined social class,17 two on employment status16,17 and one on self‐perceived poverty.21EducationSix studies used comparable methods of analysis and reported ORs with 95% CI for the association between education and multimorbidity. An additional 4 studies16,30,32,34 reported sufficient raw numbers to allow the calculation of unadjusted ORs. Thus 10 studies presenting information on 13 study populations (N=122,858 participants) were included in the meta‐analysis. There was substantial heterogeneity between studies (I2=89.3%; p‐value<0.001), therefore we must be cautious when interpreting the pooled effect estimate. For what it's worth, low education was associated with a 64% increased odds of multimorbidity (summary OR: 1.64, 95%CI 1.41 to 1.91; Figure ). ORs were adjusted for age in only three11,21,24 of the 12 study populations. However, age‐adjusted odds ratios were consistent with the overall finding, with low education associated with a 60% increased odds of multimorbidity.21,24 Sub‐group analyses suggested that the effect of low education on multimorbidity varied according to the method of disease ascertainment, with the effect stronger among studies relying on self‐report of chronic conditions than in studies using healthcare records to ascertain disease history (summary ORs 1.79, 95%CI 1.45 to 2.21 and 1.40, 95%CI 1.28 to 1.53, respectively; Figure ). Unfortunately, the majority of studies reported findings for both genders combined, limiting scope for investigation of consistency across men and women. When we grouped studies according to age (using a cut‐off of 65 years, which was the most common age restriction applied across studies), the association appeared stronger in older than younger populations (Supplementary Table ). However, very few studies actually investigated age and sex within the same study population.6,14,19,26 Eleven study populations among six studies reported odds ratios adjusted for other sociodemographic factors and (less commonly) lifestyle behaviours.6,11,14,16,21,26 The pooled summary estimate indicated an attenuation of the association between education level and multimorbidity (pooled summary OR 1.27, 95%CI 1.21 to 1.33; Supplementary Figure ), with no heterogeneity between studies (I2=0%; p‐value=0.52). A funnel plot for the association between education and multimorbidity revealed no suggestion of publication bias (p‐value for small study effects=0.95; Figure ).Meta‐analysis of the association between education and multimorbidity, showing study‐specific and summary odds ratios with 95% confidence intervals, and sub‐group analysis according to whether multimorbidity was ascertained using objective or subjective methods.Funnel plot for the association between education and multimorbidity.Seven studies (N=236,649 participants) reporting on education were not included in the meta‐analysis due to: incomparable methods of analysis and insufficient data to calculate ORs;17,19,20,27,31 lack of CIs for effect estimates;18 and insufficient data to calculate ORs.15 Findings from almost all of these studies were consistent with those included in the meta‐analysis.DeprivationThe association between deprivation and multimorbidity was generally investigated using primary care datasets. As such, analyses were unadjusted for health behaviours, apart from one study that used survey data.34 In all studies, higher deprivation level was associated with a statistically significant greater risk of multimorbidity (Table ). Differing methods of analysis and data presentation precluded formal pooling of these studies. Just two studies stratified by age and/or gender. In one study, the association between deprivation and multimorbidity was most striking in those aged 40–70 years, with the gap narrowing in those aged over 70 years. Young and middle‐aged adults living in the most deprived areas had rates of multimorbidity equivalent to those aged 10–15 years older in the most affluent areas.5 Orueta et al. stratified by both age and sex.28 The study population was aged 65 years or over and so, in contrast to the latter study, there was a less obvious narrowing of the deprivation gap in multimorbidity risk by age. Disparities were, however, larger in women compared to men.Summary of findings from studies on the association between each of area‐based deprivation/disadvantage and income, and multimorbidity.Study, yearCountryMeasure of deprivation/incomeEffect on multimorbidity*Area‐based deprivationaBarnett, 20125ScotlandCarstairs Deprivation Index; based on census information for postcode sectors↑Macleod, 200423ScotlandCarstairs Deprivation Index; based on census information for postcode sectors↑Mercer, 200725ScotlandMultiple index of deprivation score based on geographical areas↑Orueta, 201328SpainGeographical deprivation index based on census information for small geographical units↑Salisbury, 201129EnglandTownsend Deprivation Index; based on census information for postcode sectors↑Walker, 200734AustraliaSocioeconomic indexes for Areas (SEIFA) Index of relative disadvantage; based on census information for geographical areas↑IncomeaAgborsangaya, 201214CanadaAnnual household income; 4 categories ranging from <$30,000 to ≥$100,000↑bAlaba, 201316South AfricaQuintiles of annual household income↓cAndrade, 201017BrazilFamily income; categorised into top 25%, middle 50% and lowest 24%↔Hosseinpoor, 201220Low‐ and middle‐income countriesAn index of the long‐running economic status of households based on owning selected assets and/or using certain services was created to give a household wealth index, split into quintiles↑bHudon, 201213CanadaAnnual household income; 4 categories ranging from <$10,000 to ≥$50,000↑bNeeleman, 200127NetherlandsAnnual income; 3 categories – lowest to highest↑ (unadjusted) ↔ (adjustedb)Schafer, 201231GermanyHousehold‐size adjusted net income per month↑ (unadjusted)Taylor, 20106AustraliaAnnual household income; 3 categories; ranging from <$20,000 to >$80,001↑ for ages 45–59 (unadjusted) ↔ for ages 45–59 (adjustedd) ↔ for ages ≥60Tucker‐Seeley, 201132USLifetime earnings, based on average annual lifetime earnings during young and middle adulthood↑ (unadjusted) ↔ (adjustedb)*↑ = increasing deprivation is statistically significantly associated with an increased risk of multimorbidity and decreasing income is statistically significantly associated with an increased risk of multimorbidity; ↓ = decreasing income is associated with a decreased risk of multimorbidity; ↔ = no statistically significant association between income and multimorbiditya: For deprivation, all results are unadjusted for other factors; for income, unless specified, arrows represent the results for both unadjusted and adjusted analyses;b: Adjusted for demographic factors and educationc: Adjusted for sociodemographic factors, smoking, obesity, health facility visits and civic participationd: Adjusted for demographic, risk factor and health‐related variables (health service use and medicines)IncomeThe findings for income in relation to multimorbidity risk were inconsistent across studies (Table ). Four studies reported an increasing risk of multimorbidity with decreasing income,13,14,20,31 three of which had adjusted for demographic factors and education level.13,14,20 In contrast to their findings on education and multimorbidity, the South African study reported that multimorbidity risk increased with increasing income. This suggests that some SEP measures, such as income, might actually be positively associated with risk of chronic disease and multimorbidity in some low‐income countries.16 A Brazilian study reported no significant association between income and multimorbidity,17 while a US study reported an association between income and multimorbidity in unadjusted analyses only.32 An Australian study found the risk of multimorbidity increased with decreasing income level among those aged 45–59 years, but not 60 years or over,6 whereas no age differences were observed in a Canadian study.14 Just one study reported findings stratified by sex,20 with low income associated with increased risk of multimorbidity in men in both low‐ and middle‐income countries, but no association observed in women in low‐income countries after adjusting for education, marital status and rural/urban area.Other SEP measuresEvidence for the association between other SEP measures and multimorbidity is limited. Occupational social class was associated with multimorbidity in one study33 (although statistical significance was not tested), but not in two other studies.24,31 Social class (not defined) was not associated with multimorbidity in a Brazilian study.17 Employment status was associated with a decreased risk of multimorbidity in one study16 but not in another.17 A study in Bangladesh found an association between low literacy level and increased multimorbidity risk, which did not persist after adjustment for other factors.22 Finally, one study reported that self‐perceived poverty was associated with increased multimorbidity risk.21DiscussionWe identified a high number of studies examining the occurrence of multimorbidity according to SEP, primarily in high‐income settings. Education was most commonly studied, with meta‐analysis showing that low‐ versus high‐education level was significantly associated with an increased odds of multimorbidity, albeit with substantial heterogeneity between studies. Higher area‐based deprivation was consistently associated with greater multimorbidity, but the picture was less clear for income. There is little evidence on how the associations between SEP and multimorbidity varies by sex and age.The association between each of education and deprivation and prevalence of multimorbidity is unsurprising, given the well‐established evidence base for the association between these SEP measures and the risk of various individual chronic diseases.35–37 The less consistent association between income and multimorbidity may reflect differences in setting and population and the fact that income is less of a robust SEP measure particularly among retired people. Few studies stratified associations between SEP and multimorbidity by age and/or sex, with conflicting results from the studies that did stratify. Findings from one of the largest studies indicated that multimorbidity onset may occur at a younger age in the most deprived versus affluent areas.5 While further studies are needed to fully understand the reasons for this disparity, these findings have implications for intervention approaches aimed at reducing multimorbidity,38 which need to be targeted at much younger age groups, particularly among those living in adverse circumstances.While the underlying reasons for the observed association between education, deprivation and risk of multimorbidity are likely complex and multifactorial, intermediary factors such as lifestyle, access to and use of health services, and neighbourhood context will be important.39 Studies on deprivation and multimorbidity did not tend to adjust for any of these factors, while the few studies on education and multimorbidity that did adjust for lifestyle behaviours found that the association persisted. Developing constructs that capture more refined elements of socioeconomic circumstances, including for example social capital, might also yield a richer understanding of why inequalities in multimorbidity exist.40 Fresh perspective on this may come from the field of syndemics, which refers to the synergistic clustering of health conditions that results from and contributes to complex social and economic inequalities. This theory highlights the importance of the wider context of multimorbidity and reinforces the importance of understanding how macro‐ level factors interact with and promote the clustering of chronic diseases at the population level.41 We found relatively little data on SEP and multimorbidity occurrence in low‐ and middle‐income countries.16,17,20–22,30 While some measures of SEP may actually be associated with an increased risk of various chronic conditions in some low‐income settings, once these countries undergo epidemiological transition we can expect to see greater burden of multimorbidity among lower socioeconomic groups.Methodological shortcomings of some of the identified studies limit the robustness of the results. In particular, given the cross‐sectional nature of the existing studies on this topic, we must exert caution when drawing conclusions about SEP and the association with multimorbidity incidence. Education is perhaps an exception, given that it is a marker of young adult socioeconomic status. However, ability to work, nature of occupation, level of income and to some extent area‐based deprivation could themselves be influenced by a person's level of morbidity. The quality of the evidence from existing studies is mixed, with aspects of the study design in many instances potentially introducing bias and contributing to significant heterogeneity between studies. Specifically, some studies included a limited number of morbidities, which may have underestimated the prevalence of multimorbidity and affected the association with SEP in an unpredictable manner.15,16,19,20,32 Some studies ascertained disease occurrence through self‐report,6,14,16,21,30,32,34 which we demonstrated in sub‐group analyses to lead to an overestimation of the association between education and multimorbidity.Not all studies included mental health conditions in their definition of multimorbidity, and so conclusions on the association between SEP and multimorbidity in these studies relate specifically to physical disease multimorbidity. There were no differences in study findings between those that did include mental disorders versus those that didn't. Mental health disorders are likely to be under‐ascertained, particularly in low‐ and middle‐income settings where substantial treatment gaps for mental health exist.42 However, studies on multimorbidity should endeavour to capture both physical and mental health disease occurrence, which are known to be strongly linked.43Finally, although multimorbidity was consistently defined as two or more conditions in the majority of identified studies, there is no universally accepted definition of multimorbidity.44 A simple count of conditions may be too crude and may not necessarily reflect ‘burden’ of disease in terms of morbidity that impacts on quality of life for example. Different conditions or combinations of conditions may also relate to SEP to differing degrees, associations that would be masked by the use of a single multimorbidity construct. In some scenarios, co‐morbidity or frailty measures might be more applicable or appropriate than a measure of multimorbidity. Also, there is no consensus as to what constitutes a single disease when studying multimorbidity. There is some support for a definition of multimorbidity that reflects the existence of disease in multiple body systems as opposed to a count of conditions, irrespective of whether they reflect the same ‘bodily’ disease.45To our knowledge, this is the first systematic review of studies investigating the association between SEP and multimorbidity. Our review identifies important methodological issues of studies on multimorbidity, which have implications for future primary studies. Our review also identifies important gaps in our understanding of how SEP relates to multimorbidity, which should inform the design of future research.Our review does have some limitations. As discussed above, some of these relate to the limitations of the studies themselves. Although we could not include all identified studies on education in our meta‐analysis, it is reassuring that the findings from studies not included were in keeping with the meta‐analysis findings. However, we were unable to identify all underlying explanations for the observed heterogeneity. Due to limited resources, we did not search grey literature or include non‐English published articles. Finally, while we did carefully consider and critique the methodological quality of included studies, we did not formally assess methodological quality using a quality assessment tool.ConclusionsExisting evidence demonstrates that low education level and living in a deprived area are associated with an increased risk of multimorbidity. Much of this evidence stems from studies based in high‐income settings, some of which are limited by methodological shortcomings. Future studies should: minimise the risk of reverse causation through prospective study of the temporal association between socioeconomic factors and multimorbidity risk; and use objective ascertainment of a comprehensive list of chronic conditions, including mental health conditions. More broadly, further investigation into how multimorbidity should be defined is needed, with a view to obtaining a universally accepted definition or suite of definitions that can be used for research. There is an urgent need for more studies in low‐ and middle‐income countries, where multimorbidity is already a significant public health challenge. A deeper understanding of the mechanisms underlying these associations should help to identify pathways amenable to intervention aimed at reducing multimorbidity in the most vulnerable groups.AcknowledgementsCAJ was supported by the Australian National Health and Medical Research Council (grant number: APP1000986). TP was funded by a University of Queensland Summer Research Scholarship to complete this project. During the period of the project, she was supported by the Fogarty International Centre, National Institutes of Health, under award number D43TW008332 (ASCEND Research Network). The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the ASCEND Research Network. The funders had no role in the design or conduct of this review.ReferencesLozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2095–128.Murray CJ, Vos T, Lozano R, Naghavi M, Flaxman AD, Michaud C, et al. Disability‐adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2197–223.United Nations. World Population Ageing: 1950–2050. Report No.: ST/ESA/SER.A/207. New York (NY): UN Department of Economics and Social Affair Population Division; 2001.Violan C, Foguet‐Boreu Q, Flores‐Mateo G, Salisbury C, Blom J, Freitag M, et al. Prevalence, determinants and patterns of multimorbidity in primary care: A systematic review of observational studies. PLoS One. 2014;9(7):e102149.Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: A cross‐sectional study. Lancet. 2012;380(9836):37–43.Taylor AW, Price K, Gill TK, Adams R, Pilkington R, Carrangis N, et al. Multimorbidity – not just an older person's issue. Results from an Australian biomedical study. BMC Public Health. 2010;10:718.Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta‐analyses: The PRISMA statement. Ann Intern Med. 2009;151(4):264–9.Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta‐analysis of observational studies in epidemiology: A proposal for reporting. Meta‐analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283(15):2008–12.van den Akker M, Buntinx F, Metsemakers JF, Knottnerus JA. Marginal impact of psychosocial factors on multimorbidity: Results of an explorative nested case‐control study. Soc Sci Med. 2000;50(11):1679–93.van den Akker M, Buntinx F, Metsemakers JF, van der Aa M, Knottnerus JA. Psychosocial patient characteristics and GP‐registered chronic morbidity: A prospective study. J Psychosom Res. 2001;50(2):95–102.van den Akker M, Buntinx F, Metsemakers JF, Roos S, Knottnerus JA. Multimorbidity in general practice: Prevalence, incidence, and determinants of co‐occurring chronic and recurrent diseases. J Clin Epidemiol. 1998;51(5):367–75.Gold R, Michael YL, Whitlock EP, Hubbell FA, Mason ED, Rodriguez BL, et al. Race/ethnicity, socioeconomic status, and lifetime morbidity burden in the women's health initiative: A cross‐sectional analysis. J Womens Health. 2006;15(10):1161–73.Hudon C, Fortin M, Poitras ME, Almirall J. The relationship between literacy and multimorbidity in a primary care setting. BMC Fam Pract. 2012;13:33.Agborsangaya CB, Lau D, Lahtinen M, Cooke T, Johnson JA. Multimorbidity prevalence and patterns across socioeconomic determinants: A cross‐sectional survey. BMC Public Health. 2012;12:201.Ahluwalia IB, Mack KA, Mokdad A. Report from the CDC. Changes in selected chronic disease‐related risks and health conditions for nonpregnant women 18–44 years old BRFSS. J Womens Health (Larchmt). 2005;14(5):382–6.Alaba O, Chola L. The social determinants of multimorbidity in South Africa. Int J Equity Health. 2013;12:63.Andrade LH, Bensenor IM, Viana MC, Andreoni S, Wang YP. Clustering of psychiatric and somatic illnesses in the general population: Multimorbidity and socioeconomic correlates. Braz J Med Biol Res. 2010;43(5):483–91.Droomers M, Westert GP. Do lower socioeconomic groups use more health services, because they suffer from more illnesses? Eur J Public Health. 2004;14(3):311–3.Enroth L, Raitanen J, Hervonen A, Jylha M. Do socioeconomic health differences persist in nonagenarians? J Gerontol B Psychol Sci Soc Sci. 2013;68(5):837–47.Hosseinpoor AR, Bergen N, Mendis S, Harper S, Verdes E, Kunst A, et al. Socioeconomic inequality in the prevalence of noncommunicable diseases in low‐ and middle‐income countries: Results from the World Health Survey. BMC Public Health. 2012;12:474.Jerliu N, Toci E, Burazeri G, Ramadani N, Brand H. Prevalence and socioeconomic correlates of chronic morbidity among elderly people in Kosovo: A population‐based survey. BMC Geriatr. 2013;13:22.Khanam MA, Streatfield PK, Kabir ZN, Qiu C, Cornelius C, Wahlin A. Prevalence and patterns of multimorbidity among elderly people in rural Bangladesh: A cross‐sectional study. J Health Popul Nutr. 2011;29(4):406–14.Macleod U, Mitchell E, Black M, Spence G. Comorbidity and socioeconomic deprivation: An observational study of the prevalence of comorbidity in general practice. Eur J Gen Pract. 2004;10(1):24–6.Marengoni A, Winblad B, Karp A, Fratiglioni L. Prevalence of chronic diseases and multimorbidity among the elderly population in Sweden. Am J Public Health. 2008;98(7):1198–200.Mercer SW, Watt GC. The inverse care law: Clinical primary care encounters in deprived and affluent areas of Scotland. Ann Fam Med. 2007;5(6):503–10.Nagel G, Peter R, Braig S, Hermann S, Rohrmann S, Linseisen J. The impact of education on risk factors and the occurrence of multimorbidity in the EPIC‐Heidelberg cohort. BMC Public Health. 2008;8:384.Neeleman J, Ormel J, Bijl R. The distribution of psychiatric and somatic ill health: Associations with personality and socioeconomic status. Psychosom Med. 2001;63:239–47.Orueta JF, Nuno‐Solinis R, Garcia‐Alvarez A, Alonso‐Moran E. Prevalence of multimorbidity according to the deprivation level among the elderly in the Basque Country. BMC Public Health. 2013;13:918.Salisbury C, Johnson L, Purdy S, Valderas JM, Montgomery AA. Epidemiology and impact of multimorbidity in primary care: A retrospective cohort study. Br J Gen Pract. 2011;61(582):e12–21.Santos Machado V, de S, Valadares AL, Costa‐Paiva LH, Osis MJ, Sousa MH, Pinto‐Neto AM. Aging, obesity, and multimorbidity in women 50 years or older: A population‐based study. Menopause. 2013;20(8):818–24.Schafer I, Hansen H, Schon G, Hofels S, Altiner A, Dahlhaus A, et al. The influence of age, gender and socio‐economic status on multimorbidity patterns in primary care. First results from the multicare cohort study. BMC Health Serv Res. 2012;12:89.Tucker‐Seeley RD, Li Y, Sorensen G, Subramanian SV. Lifecourse socioeconomic circumstances and multimorbidity among older adults. BMC Public Health. 2011;11:313.Uijen AA, van de Lisdonk EH. Multimorbidity in primary care: Prevalence and trend over the last 20 years. Eur J Gen Pract. 2008;14 Suppl 1:28–32.Walker AE. Multiple chronic diseases and quality of life: Patterns emerging from a large national sample, Australia. Chronic Illn. 2007;3(3):202–18.Dalstra JA, Kunst AE, Borrell C, Breeze E, Cambois E, Costa G, et al. Socioeconomic differences in the prevalence of common chronic diseases: An overview of eight European countries. Int J Epidemiol. 2005;34(2):316–26.Hippisley‐Cox J, Coupland C, Vinogradova Y, Robson J, May M, Brindle P. Derivation and validation of QRISK, a new cardiovascular disease risk score for the United Kingdom: Prospective open cohort study. BMJ. 2007;335(7611):136.Woodward M, Brindle P, Tunstall‐Pedoe H. Adding social deprivation and family history to cardiovascular risk assessment: The ASSIGN score from the Scottish Heart Health Extended Cohort (SHHEC). Heart. 2007;93(2):172–6.Smith SM, Ferede A, O'Dowd T. Multimorbidity in younger deprived patients: An exploratory study of research and service implications in general practice. BMC Fam Pract. 2008;9(1):6.Wang HHX, Wang JJ, Wong SYS, Wong MCS, Li FJ, Wang PX, et al. Epidemiology of multimorbidity in China and implications for the healthcare system: Cross‐sectional survey among 162,464 community household residents in southern China. BMC Med. 2014;12:188.Ong BN, Richardson JC, Porter T, Grime J. Exploring the relationship between multi‐morbidity, resilience and social connectedness across the lifecourse. Health (London). 2014;18(3):302–18.Mendenhall E, Kohrt BA, Norris SA, Ndetei D, Prabhakaran D. Non‐communicable disease syndemics: poverty, depression, and diabetes among low‐income populations. Lancet. 2017;389(10072):951–63.Semrau M, Evans‐Lacko S, Alem A, Ayuso‐Mateos JL, Chisholm D, Gureje O, et al. Strengthening mental health systems in low‐ and middle‐income countries: The Emerald programme. BMC Med. 2015;13:79.Royal College of Psychiatrists. No Health Without Public Mental Health: The Case for Action. London (UK): RCP; 2010.Diederichs C, Berger K, Bartels DB. The measurement of multiple chronic diseases – A systematic review on existing multimorbidity indices. J Gerontol A Biol Sci Med Sci. 2011;66(3):301–11.Fortin M, Stewart M, Poitras ME, Almirall J, Maddocks H. A systematic review of prevalence studies on multimorbidity: Toward a more uniform methodology. Ann Fam Med. 2012;10(2):142–51. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australian and New Zealand Journal of Public Health Wiley

Socioeconomic status and multimorbidity: a systematic review and meta‐analysis

Loading next page...
 
/lp/wiley/socioeconomic-status-and-multimorbidity-a-systematic-review-and-meta-ypwDtrknW3

References (48)

Publisher
Wiley
Copyright
© 2018 Public Health Association of Australia
ISSN
1326-0200
eISSN
1753-6405
DOI
10.1111/1753-6405.12762
pmid
29442409
Publisher site
See Article on Publisher Site

Abstract

The increasing prevalence of chronic conditions1,2 and the growth of the ageing population has led to an increase in multimorbidity worldwide.3 In this article, we distinguish between multimorbidity (the co‐existence of two or more chronic conditions) and co‐morbidity (co‐occurrence of disease/s with a specific index disease). The impact of multimorbidity on the health and wellbeing of individuals, the burden on healthcare systems and the effect on economies has created a major global public health problem. A recent systematic review reported that the prevalence of multimorbidity among the general adult population in high‐income countries ranges from 12.9% in participants aged 18 years and older to 95.1% in a community‐dwelling elderly population aged 85 years.4 However, while the prevalence of multimorbidity is highest among the oldest (above 85 years of age), the growing burden of multimorbidity among older adults still of working age and among lower socioeconomic groups in some countries is of particular public health concern.5,6This expansion of morbidity is leading to individuals living longer but with more co‐existing chronic disease from a younger age, placing an even greater burden on healthcare systems.While multimorbidity incidence and prevalence is known to vary by measures of socioeconomic status (SEP), with an excess burden in lower socioeconomic groups, there are some gaps in our understanding of this relationship. It is unclear whether the association is true for all SEP measures and whether the magnitude of association varies by age, gender and country. Since different SEP variables measure different aspects of individual circumstances and environmental characteristics, a better understanding of which socioeconomic factors are more strongly associated with multimorbidity may help us to better understand the underlying mechanisms. In turn, this will help inform the design of intervention approaches aimed at preventing or reducing the development of multimorbidity.ObjectivesThe objective of this review was to systematically identify, critically appraise and synthesise the existing literature on the association between SEP and multimorbidity occurrence.MethodsThis manuscript was prepared in accordance with the PRISMA guidelines.7 The protocol for this review was not registered.Search strategyWe sought studies published in English in Medline and Embase between 1946 and 1974, respectively, and December 2014 reporting on the association between SEP and multimorbidity using a comprehensive electronic search strategy (Supplementary file 1: Appendix A) and perusal of reference lists of all relevant identified articles. We included terms for multimorbidity and comorbidity, since these are often used interchangeably. One author (TP) screened all references by title and abstract and, where necessary, the full text of the article. All potentially relevant articles were reviewed by the co‐author, and any disagreements on inclusion were resolved through discussion between the two authors and consensus was reached.Inclusion and exclusion criteriaWe included studies of any quantitative design that reported on the occurrence of multimorbidity (defined by the co‐occurrence of multiple conditions) with respect to any measure of SEP, in adult populations only. We excluded: qualitative studies; studies that included selected populations (e.g. patients with psychiatric conditions only, substance abuse problems or those who had undergone a specific medical procedure); and studies that reported on co‐morbidity rather than multimorbidity. Where multiple articles on overlapping study populations were identified, we included the study with the largest population.Data extractionBoth authors independently extracted information on: study design; study population; demographics; sample size; exclusion criteria; definition and ascertainment of multimorbidity; number of diseases ascertained; inclusion of mental health among chronic diseases; ascertainment and measure/s of SEP; and results, including numbers with and without multimorbidity, for the purpose of meta‐analysis, where appropriate.Data synthesisWe aimed to narratively summarise study findings or, where possible, combine study‐specific estimates of effect using meta‐analysis to obtain a pooled summary estimate. Due to the substantial variation in how SEP measures were defined, and/or the different methods of analysis used, meta‐analysis was only possible for the association between education and multimorbidity. This SEP measure was the most consistently defined and a sufficient number of studies adopted the same statistical analysis approach (i.e. logistic regression) to allow us to formally pool the results.Meta‐analysisWe performed meta‐analysis using Stata version 13 and adhered to the MOOSE guidelines.8 We combined studies that either reported unadjusted odds ratios (ORs) with 95% confidence intervals (CIs) for the association between education and multimorbidity or presented raw numbers that allowed us to calculate ORs. Where we extracted raw numbers, we defined the non‐multimorbidity comparison group as participants with fewer than two conditions, in line with the most common comparison group used across studies. Our rationale for pooling together minimally or unadjusted ORs was firstly to harmonise findings from studies, in order to facilitate inclusion of as many studies as possible. Studies varied in terms of: the type of effect estimates presented; whether they presented unadjusted and/or adjusted estimates; and specific factors adjusted for, and reference category used for effect estimates for education. Secondly, our interest lay in determining the association between SEP and multimorbidity without seeking to identify the mechanisms underlying the association. Extraction of only effect estimates adjusted for additional factors, some of which may lie on the causal pathway between SEP and multimorbidity, would therefore have potentially obscured the true nature of the association between SEP and multimorbidity. We obtained a pooled summary OR for the odds of multimorbidity comparing low versus high education using the Mantel Haenszel random effects model, assessing heterogeneity between studies using the chi2 (Cochrane Q) and I2 statistics. We sought to explore potential explanations for any observed heterogeneity using sub‐group analysis. We aimed to assess the influence of three a priori determined study characteristics: age of study population; sex; and method of ascertainment of chronic disease (i.e. objective versus subjective, as described below). Where relevant studies on education and multimorbidity did not provide sufficient information to be included in the meta‐analysis, we contacted the authors to obtain results in the necessary format. However, none of the authors who replied were able to provide this information, because they no longer had access to the datasets.ResultsWe identified 2,496 articles, 63 of which were potentially relevant and underwent full‐text review. Of 28 relevant studies, two were excluded9,10 because the study populations overlapped with a third included study.11 Two studies were excluded because they defined multimorbidity using a score that incorporated severity of disease (Figure ).12,13 The remaining 24 included studies were cross‐sectional in design or entailed cross‐sectional analysis of SEP factors related to multimorbidity (Table ).5,6,11,14–34 In general, participants were recruited through population‐based primary care databases or national or regional surveys, with more than half (N=15) conducted in high‐income countries. The number of chronic diseases included in each study ranged from 5 to 335 (Supplementary file 2: Appendix B), with just 12 studies reporting mental health diseases were included (Table ).5,6,11,14,19,20,23,24,28,29,31,33 Ten studies5,11,22–25,28,29,31,33 ascertained multimorbidity through objective sources such as health records (which capture doctor‐diagnosed conditions), while the remainder relied on self‐report of doctor‐diagnosed conditions by participants, which may be subject to recall error (Table ). Multimorbidity was defined in 18 studies as the co‐occurrence of two or more conditions, and in one study as the co‐occurrence of three or more conditions, with the comparison group being fewer than two (or three) conditions (Table ). The exceptions to this were the studies by Jerlui et al. and Marengoni et al. in which multimorbidity was compared to single morbidity.21,24 In the remaining five studies, multimorbidity was analysed as a continuous count of conditions, without a cut‐off being employed.Flow diagram of literature search and included studies.Characteristics of identified studies.Author, yearCountryStudy populationTotal NMale (%)Age (years)aDefinition of multimorbidity [comparison group]Ascertainment of morbiditiesNumber of conditions includedMental health includedAgborsangaya, 201214CanadaA representative sample of the adult (aged ≥18 years) population of the province of Alberta, as identified in the 2010 Patient Experience Survey by Health Quality Council of Alberta5,01047.746.7≥2 concurrent conditions [<2 conditions]Self‐report via telephone‐administered survey questionnaire16YAhluwalia, 200515USNon‐institutionalised, non‐pregnant women aged 18–44 years, identified from the Behavioural Risk Factor Surveillance System, an ongoing random digit dialled telephone survey of 50 States;163,5660Mean NR Range 18–44≥2 concurrent conditions [no formal statistical comparison made]Self‐report via telephone‐administered survey questionnaire5NAlaba, 201316South AfricaStratified two‐stage cluster sampling of 53 district councils of South Africa, including adults aged ≥15 years; for this study adults aged ≥18 were included11,6383940≥2 concurrent conditions [<2 conditions]Self‐report of doctor‐diagnosed condition6NAndrade, 201017BrazilNon‐institutionalised adults aged ≥18 years included in a household survey of two boroughs of Sao Paulo1,46457.4Mean NR (82% aged <60)Morbidity included in modelling as a continuous variableFace‐to‐face interviews, with information on morbidities obtained via self‐reported doctor diagnosis of various conditions8NbBarnett, 20125ScotlandPatients of any age registered with one of 314 GP practices, covering one third of the Scottish population1,751,84149.5Mean NR (83% aged <65)≥2 concurrent conditions [<2 conditions]GP medical records and prescription data40YDroomers, 200418NetherlandsRespondents aged 25 or over included in the Netherlands Health Interview Surveys from 1990–1998 which included non‐institutionalised persons53,339NRMean NR (all ≥25)≥2 concurrent conditions [<2 conditions]Self‐report in questionnaire14NEnroth, 201319FinlandSurvey of adults aged 90+ living in Tampere, irrespective of dwelling place (80% of eligible population responded)1,28319Mean NR (All ≥90)≥2 concurrent conditions [<2 conditions]Self‐report of doctor‐diagnosed conditions6YHosseinpoor, 20122041 low and middle‐income countriesData from the 2002–04 World Health Survey of adults aged ≥18 years in 41 low and middle‐income countries; surveys are nationally representative except in 6 countries where WHS is conducted in geographically limited regions170,29845.5Mean NR (all >18)≥2 concurrent conditions [<2 conditions]Self‐report in questionnaire; symptom‐based classification based on symptoms in previous 12 months used for all conditions except diabetes, which was reported directly; diabetes by self‐report5YHudon, 201213CanadaRegular patients aged ≥18 years attending a family medicine clinic (primary care) of a regional health center in Saguenay, Quebec1033549.9Used a multimorbidity scoreSelf‐reported questionnaire using simplified version of the Disease Burden Morbidity Assessment, including 11 of 21 diseases included in the original instrument, based on the high prevalence of these diseases in this setting11NJerliu, 201321KosovoA nationwide, population‐based representative sample of adults aged ≥65 years1,89050.273.4≥2 concurrent conditions [1 condition]Self‐report via an interviewer‐administered questionnaireUnclear*Unclear*Khanam, 201122BangladeshRandomly selected individuals age >=60 from two areas of a rural area of Bangladesh, based on the sampling frame of the Health and Demographic Surveillance System45245.169.5≥2 concurrent conditions [<2 conditions]Clinical examination by physicians, with independent evaluation by senior physicians/geriatricians9NMacleod, 200423ScotlandPatients aged 18 or over registered at one practice in an area of high deprivation in Glasgow City7,286NRMean NR (76.7% aged <65)≥2 concurrent conditions [<2 conditions]GP records16YMarengoni, 200824SwedenParticipants in the first follow up of the Kungsholmen Project, a study on elderly people (aged ≥75 years in October 1987) living in Stockholm.1,099?NR84.6≥2 concurrent conditions [1 condition]Clinical assessment by a physician, medical history (from Stockholm inpatient register that records discharge diagnoses from Stockholm hospitals), laboratory data and current drug use.30YMercer, 200725ScotlandConsecutive unselected patients of one GP nominated from medium‐sized practices in the upper or lower quartiles of deprivation in 4 healthboard regions in West of Scotland; 26 GPs included3,044NR44.6**Morbidity assessed as a count, not a cut‐offGP recordsNR, but based on GP records so presumably all conditions includedNRNagel, 200826GermanyHeidelberg cohort of European Prospective Investigation in to Cancer and nutrition (EPIC) recruited from general population. Cross‐sectional analysis of multimorbidity using data on prevalent chronic conditions collected at baseline and during follow‐up13,78152.9Median 55–58 across male and female subgroups≥2 concurrent conditions [<2 conditions]Prevalent diseases identified at recruitment by face to face interviews; new diseases identified by active follow up through questionnaire on physician diagnosed chronic conditions. Cancer cases verified with medical records and validation studies conducted for incident cases of diabetes, MI, asthma and stroke24NNeeleman, 200127NetherlandsParticipants of the 1996 wave of NEMESIS, a study of the incidence and prevalence of mental disorders in the general population; random sampling of 90 municipalities of adults aged 18–64 years7,076NRNRMorbidity assessed as a count, not a cut‐offSelf‐report based on interviewer‐administered questionnaire30NbOrueta, 201328SpainAdults aged ≥65 years covered by public health insurance in the Basque country on 31st August 2011 for at least 6 months in the previous year, regardless of whether or not they had made any contact with or use of Basque Health Service452,69842.5Mean NR≥2 concurrent conditions [<2 conditions]Primary care electronic medical records, hospital admissions, outpatient care databases and prescription data47YSalisbury, 201129EnglandRandom sample of patients aged 18 or over registered with one of 18 General Practitioners on index date of 1 April 200599,997NRNR≥2 concurrent conditions [<2 conditions]GP records19YSantos Machado, 201330BrazilSimple random sampling of women aged ≥50 living in 68 census districts of the city of Campinos, Sao Paulo6220Mean NR (61.3% aged >60)≥2 concurrent conditions [<2 conditions]Self‐report based on interviewer‐administered questionnaire12NSchafer, 201231GermanyPatients born between 1st July 1923 and 30th June 1943 who consulted GP at one of 158 General Practices in 8 study centres were randomly selected3,18959.374.4Morbidity included in modelling as a continuous variableGP interviews using standardized instrument covering 46 chronic conditions46YTaylor, 20106AustraliaParticipants aged ≥18 years from the North West Adelaide Study, a population‐based cohort study; included those from stage 2 of the cohort study, which involved all or a combination of: a computer assisted phone interview; self‐complete questionnaire; and a biomedical examination at a clinic3,206NRMean NR (75% aged <60)≥2 concurrent conditions [<2 conditions]Self‐report of doctor diagnosed condition7YTucker‐Seeley, 201132USRespondents aged ≥50 years from 2004 wave of the Health and Retirement study7,30546.465≥2 concurrent conditions (for this review) [<2 conditions]Self‐report of doctor diagnosed condition6NUijen, 200833NetherlandsPatients of all ages enlisted in the Continuous Morbidity Registration, Nijmegen, which includes 4 General Practices, including 10 GPs with approximately 13,500 enlisted patients;13,584NRMean NR Range 0 – >75Examined distribution of 0,1,2,3 and ≥4 conditions by SEPGP recordsNRcYVan den Akker, 199811NetherlandsParticipants identified from Registration Network Family Practices, a continuous computerised primary care database including 42 GPs in 15 different practices60,85748.7Mean NR (80% aged <60)≥2 concurrent conditions [<2 conditions]GP records335YWalker, 200734AustraliaParticipants aged ≥ 20 years in the Australian 2001 National Health Survey or the 2003 Survey of Disability Ageing and Carers, both of which are nationally representative17,45045Mean NR≥3 concurrent conditions [<3 conditions]Self‐report in household surveysNRNR* Survey included questions on which groups of diseases participants had and also included an open‐ended question on other chronic conditions** Weighted mean calculated from data presented in papera: Mean unless specified otherwiseb: Psychiatric conditions analysed separatelyc: Total number of conditions not reported, but included all obligatory and conditionally registered chronic diseases recorded in the Continuous Morbidity RegistrationNR = not reported; GP = general practitioner; MI = myocardial infarction; SEP = socioeconomic positionMost studies (N=17; 359,507 participants) measured SEP using education.6,11,14–21,24,26,27,30–32,34 Eight studies (209,186 participants) reported on income6,14,16,17,20,27,31,32 and six (2,332,316 participants) reported on deprivation.5,23,25,28,29,34 In addition, one study reported on literacy,22 three on occupational social class,24,31,33 one on non‐defined social class,17 two on employment status16,17 and one on self‐perceived poverty.21EducationSix studies used comparable methods of analysis and reported ORs with 95% CI for the association between education and multimorbidity. An additional 4 studies16,30,32,34 reported sufficient raw numbers to allow the calculation of unadjusted ORs. Thus 10 studies presenting information on 13 study populations (N=122,858 participants) were included in the meta‐analysis. There was substantial heterogeneity between studies (I2=89.3%; p‐value<0.001), therefore we must be cautious when interpreting the pooled effect estimate. For what it's worth, low education was associated with a 64% increased odds of multimorbidity (summary OR: 1.64, 95%CI 1.41 to 1.91; Figure ). ORs were adjusted for age in only three11,21,24 of the 12 study populations. However, age‐adjusted odds ratios were consistent with the overall finding, with low education associated with a 60% increased odds of multimorbidity.21,24 Sub‐group analyses suggested that the effect of low education on multimorbidity varied according to the method of disease ascertainment, with the effect stronger among studies relying on self‐report of chronic conditions than in studies using healthcare records to ascertain disease history (summary ORs 1.79, 95%CI 1.45 to 2.21 and 1.40, 95%CI 1.28 to 1.53, respectively; Figure ). Unfortunately, the majority of studies reported findings for both genders combined, limiting scope for investigation of consistency across men and women. When we grouped studies according to age (using a cut‐off of 65 years, which was the most common age restriction applied across studies), the association appeared stronger in older than younger populations (Supplementary Table ). However, very few studies actually investigated age and sex within the same study population.6,14,19,26 Eleven study populations among six studies reported odds ratios adjusted for other sociodemographic factors and (less commonly) lifestyle behaviours.6,11,14,16,21,26 The pooled summary estimate indicated an attenuation of the association between education level and multimorbidity (pooled summary OR 1.27, 95%CI 1.21 to 1.33; Supplementary Figure ), with no heterogeneity between studies (I2=0%; p‐value=0.52). A funnel plot for the association between education and multimorbidity revealed no suggestion of publication bias (p‐value for small study effects=0.95; Figure ).Meta‐analysis of the association between education and multimorbidity, showing study‐specific and summary odds ratios with 95% confidence intervals, and sub‐group analysis according to whether multimorbidity was ascertained using objective or subjective methods.Funnel plot for the association between education and multimorbidity.Seven studies (N=236,649 participants) reporting on education were not included in the meta‐analysis due to: incomparable methods of analysis and insufficient data to calculate ORs;17,19,20,27,31 lack of CIs for effect estimates;18 and insufficient data to calculate ORs.15 Findings from almost all of these studies were consistent with those included in the meta‐analysis.DeprivationThe association between deprivation and multimorbidity was generally investigated using primary care datasets. As such, analyses were unadjusted for health behaviours, apart from one study that used survey data.34 In all studies, higher deprivation level was associated with a statistically significant greater risk of multimorbidity (Table ). Differing methods of analysis and data presentation precluded formal pooling of these studies. Just two studies stratified by age and/or gender. In one study, the association between deprivation and multimorbidity was most striking in those aged 40–70 years, with the gap narrowing in those aged over 70 years. Young and middle‐aged adults living in the most deprived areas had rates of multimorbidity equivalent to those aged 10–15 years older in the most affluent areas.5 Orueta et al. stratified by both age and sex.28 The study population was aged 65 years or over and so, in contrast to the latter study, there was a less obvious narrowing of the deprivation gap in multimorbidity risk by age. Disparities were, however, larger in women compared to men.Summary of findings from studies on the association between each of area‐based deprivation/disadvantage and income, and multimorbidity.Study, yearCountryMeasure of deprivation/incomeEffect on multimorbidity*Area‐based deprivationaBarnett, 20125ScotlandCarstairs Deprivation Index; based on census information for postcode sectors↑Macleod, 200423ScotlandCarstairs Deprivation Index; based on census information for postcode sectors↑Mercer, 200725ScotlandMultiple index of deprivation score based on geographical areas↑Orueta, 201328SpainGeographical deprivation index based on census information for small geographical units↑Salisbury, 201129EnglandTownsend Deprivation Index; based on census information for postcode sectors↑Walker, 200734AustraliaSocioeconomic indexes for Areas (SEIFA) Index of relative disadvantage; based on census information for geographical areas↑IncomeaAgborsangaya, 201214CanadaAnnual household income; 4 categories ranging from <$30,000 to ≥$100,000↑bAlaba, 201316South AfricaQuintiles of annual household income↓cAndrade, 201017BrazilFamily income; categorised into top 25%, middle 50% and lowest 24%↔Hosseinpoor, 201220Low‐ and middle‐income countriesAn index of the long‐running economic status of households based on owning selected assets and/or using certain services was created to give a household wealth index, split into quintiles↑bHudon, 201213CanadaAnnual household income; 4 categories ranging from <$10,000 to ≥$50,000↑bNeeleman, 200127NetherlandsAnnual income; 3 categories – lowest to highest↑ (unadjusted) ↔ (adjustedb)Schafer, 201231GermanyHousehold‐size adjusted net income per month↑ (unadjusted)Taylor, 20106AustraliaAnnual household income; 3 categories; ranging from <$20,000 to >$80,001↑ for ages 45–59 (unadjusted) ↔ for ages 45–59 (adjustedd) ↔ for ages ≥60Tucker‐Seeley, 201132USLifetime earnings, based on average annual lifetime earnings during young and middle adulthood↑ (unadjusted) ↔ (adjustedb)*↑ = increasing deprivation is statistically significantly associated with an increased risk of multimorbidity and decreasing income is statistically significantly associated with an increased risk of multimorbidity; ↓ = decreasing income is associated with a decreased risk of multimorbidity; ↔ = no statistically significant association between income and multimorbiditya: For deprivation, all results are unadjusted for other factors; for income, unless specified, arrows represent the results for both unadjusted and adjusted analyses;b: Adjusted for demographic factors and educationc: Adjusted for sociodemographic factors, smoking, obesity, health facility visits and civic participationd: Adjusted for demographic, risk factor and health‐related variables (health service use and medicines)IncomeThe findings for income in relation to multimorbidity risk were inconsistent across studies (Table ). Four studies reported an increasing risk of multimorbidity with decreasing income,13,14,20,31 three of which had adjusted for demographic factors and education level.13,14,20 In contrast to their findings on education and multimorbidity, the South African study reported that multimorbidity risk increased with increasing income. This suggests that some SEP measures, such as income, might actually be positively associated with risk of chronic disease and multimorbidity in some low‐income countries.16 A Brazilian study reported no significant association between income and multimorbidity,17 while a US study reported an association between income and multimorbidity in unadjusted analyses only.32 An Australian study found the risk of multimorbidity increased with decreasing income level among those aged 45–59 years, but not 60 years or over,6 whereas no age differences were observed in a Canadian study.14 Just one study reported findings stratified by sex,20 with low income associated with increased risk of multimorbidity in men in both low‐ and middle‐income countries, but no association observed in women in low‐income countries after adjusting for education, marital status and rural/urban area.Other SEP measuresEvidence for the association between other SEP measures and multimorbidity is limited. Occupational social class was associated with multimorbidity in one study33 (although statistical significance was not tested), but not in two other studies.24,31 Social class (not defined) was not associated with multimorbidity in a Brazilian study.17 Employment status was associated with a decreased risk of multimorbidity in one study16 but not in another.17 A study in Bangladesh found an association between low literacy level and increased multimorbidity risk, which did not persist after adjustment for other factors.22 Finally, one study reported that self‐perceived poverty was associated with increased multimorbidity risk.21DiscussionWe identified a high number of studies examining the occurrence of multimorbidity according to SEP, primarily in high‐income settings. Education was most commonly studied, with meta‐analysis showing that low‐ versus high‐education level was significantly associated with an increased odds of multimorbidity, albeit with substantial heterogeneity between studies. Higher area‐based deprivation was consistently associated with greater multimorbidity, but the picture was less clear for income. There is little evidence on how the associations between SEP and multimorbidity varies by sex and age.The association between each of education and deprivation and prevalence of multimorbidity is unsurprising, given the well‐established evidence base for the association between these SEP measures and the risk of various individual chronic diseases.35–37 The less consistent association between income and multimorbidity may reflect differences in setting and population and the fact that income is less of a robust SEP measure particularly among retired people. Few studies stratified associations between SEP and multimorbidity by age and/or sex, with conflicting results from the studies that did stratify. Findings from one of the largest studies indicated that multimorbidity onset may occur at a younger age in the most deprived versus affluent areas.5 While further studies are needed to fully understand the reasons for this disparity, these findings have implications for intervention approaches aimed at reducing multimorbidity,38 which need to be targeted at much younger age groups, particularly among those living in adverse circumstances.While the underlying reasons for the observed association between education, deprivation and risk of multimorbidity are likely complex and multifactorial, intermediary factors such as lifestyle, access to and use of health services, and neighbourhood context will be important.39 Studies on deprivation and multimorbidity did not tend to adjust for any of these factors, while the few studies on education and multimorbidity that did adjust for lifestyle behaviours found that the association persisted. Developing constructs that capture more refined elements of socioeconomic circumstances, including for example social capital, might also yield a richer understanding of why inequalities in multimorbidity exist.40 Fresh perspective on this may come from the field of syndemics, which refers to the synergistic clustering of health conditions that results from and contributes to complex social and economic inequalities. This theory highlights the importance of the wider context of multimorbidity and reinforces the importance of understanding how macro‐ level factors interact with and promote the clustering of chronic diseases at the population level.41 We found relatively little data on SEP and multimorbidity occurrence in low‐ and middle‐income countries.16,17,20–22,30 While some measures of SEP may actually be associated with an increased risk of various chronic conditions in some low‐income settings, once these countries undergo epidemiological transition we can expect to see greater burden of multimorbidity among lower socioeconomic groups.Methodological shortcomings of some of the identified studies limit the robustness of the results. In particular, given the cross‐sectional nature of the existing studies on this topic, we must exert caution when drawing conclusions about SEP and the association with multimorbidity incidence. Education is perhaps an exception, given that it is a marker of young adult socioeconomic status. However, ability to work, nature of occupation, level of income and to some extent area‐based deprivation could themselves be influenced by a person's level of morbidity. The quality of the evidence from existing studies is mixed, with aspects of the study design in many instances potentially introducing bias and contributing to significant heterogeneity between studies. Specifically, some studies included a limited number of morbidities, which may have underestimated the prevalence of multimorbidity and affected the association with SEP in an unpredictable manner.15,16,19,20,32 Some studies ascertained disease occurrence through self‐report,6,14,16,21,30,32,34 which we demonstrated in sub‐group analyses to lead to an overestimation of the association between education and multimorbidity.Not all studies included mental health conditions in their definition of multimorbidity, and so conclusions on the association between SEP and multimorbidity in these studies relate specifically to physical disease multimorbidity. There were no differences in study findings between those that did include mental disorders versus those that didn't. Mental health disorders are likely to be under‐ascertained, particularly in low‐ and middle‐income settings where substantial treatment gaps for mental health exist.42 However, studies on multimorbidity should endeavour to capture both physical and mental health disease occurrence, which are known to be strongly linked.43Finally, although multimorbidity was consistently defined as two or more conditions in the majority of identified studies, there is no universally accepted definition of multimorbidity.44 A simple count of conditions may be too crude and may not necessarily reflect ‘burden’ of disease in terms of morbidity that impacts on quality of life for example. Different conditions or combinations of conditions may also relate to SEP to differing degrees, associations that would be masked by the use of a single multimorbidity construct. In some scenarios, co‐morbidity or frailty measures might be more applicable or appropriate than a measure of multimorbidity. Also, there is no consensus as to what constitutes a single disease when studying multimorbidity. There is some support for a definition of multimorbidity that reflects the existence of disease in multiple body systems as opposed to a count of conditions, irrespective of whether they reflect the same ‘bodily’ disease.45To our knowledge, this is the first systematic review of studies investigating the association between SEP and multimorbidity. Our review identifies important methodological issues of studies on multimorbidity, which have implications for future primary studies. Our review also identifies important gaps in our understanding of how SEP relates to multimorbidity, which should inform the design of future research.Our review does have some limitations. As discussed above, some of these relate to the limitations of the studies themselves. Although we could not include all identified studies on education in our meta‐analysis, it is reassuring that the findings from studies not included were in keeping with the meta‐analysis findings. However, we were unable to identify all underlying explanations for the observed heterogeneity. Due to limited resources, we did not search grey literature or include non‐English published articles. Finally, while we did carefully consider and critique the methodological quality of included studies, we did not formally assess methodological quality using a quality assessment tool.ConclusionsExisting evidence demonstrates that low education level and living in a deprived area are associated with an increased risk of multimorbidity. Much of this evidence stems from studies based in high‐income settings, some of which are limited by methodological shortcomings. Future studies should: minimise the risk of reverse causation through prospective study of the temporal association between socioeconomic factors and multimorbidity risk; and use objective ascertainment of a comprehensive list of chronic conditions, including mental health conditions. More broadly, further investigation into how multimorbidity should be defined is needed, with a view to obtaining a universally accepted definition or suite of definitions that can be used for research. There is an urgent need for more studies in low‐ and middle‐income countries, where multimorbidity is already a significant public health challenge. A deeper understanding of the mechanisms underlying these associations should help to identify pathways amenable to intervention aimed at reducing multimorbidity in the most vulnerable groups.AcknowledgementsCAJ was supported by the Australian National Health and Medical Research Council (grant number: APP1000986). TP was funded by a University of Queensland Summer Research Scholarship to complete this project. During the period of the project, she was supported by the Fogarty International Centre, National Institutes of Health, under award number D43TW008332 (ASCEND Research Network). The content of this publication is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the ASCEND Research Network. The funders had no role in the design or conduct of this review.ReferencesLozano R, Naghavi M, Foreman K, Lim S, Shibuya K, Aboyans V, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2095–128.Murray CJ, Vos T, Lozano R, Naghavi M, Flaxman AD, Michaud C, et al. Disability‐adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: A systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380(9859):2197–223.United Nations. World Population Ageing: 1950–2050. Report No.: ST/ESA/SER.A/207. New York (NY): UN Department of Economics and Social Affair Population Division; 2001.Violan C, Foguet‐Boreu Q, Flores‐Mateo G, Salisbury C, Blom J, Freitag M, et al. Prevalence, determinants and patterns of multimorbidity in primary care: A systematic review of observational studies. PLoS One. 2014;9(7):e102149.Barnett K, Mercer SW, Norbury M, Watt G, Wyke S, Guthrie B. Epidemiology of multimorbidity and implications for health care, research, and medical education: A cross‐sectional study. Lancet. 2012;380(9836):37–43.Taylor AW, Price K, Gill TK, Adams R, Pilkington R, Carrangis N, et al. Multimorbidity – not just an older person's issue. Results from an Australian biomedical study. BMC Public Health. 2010;10:718.Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta‐analyses: The PRISMA statement. Ann Intern Med. 2009;151(4):264–9.Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD, Rennie D, et al. Meta‐analysis of observational studies in epidemiology: A proposal for reporting. Meta‐analysis Of Observational Studies in Epidemiology (MOOSE) group. JAMA. 2000;283(15):2008–12.van den Akker M, Buntinx F, Metsemakers JF, Knottnerus JA. Marginal impact of psychosocial factors on multimorbidity: Results of an explorative nested case‐control study. Soc Sci Med. 2000;50(11):1679–93.van den Akker M, Buntinx F, Metsemakers JF, van der Aa M, Knottnerus JA. Psychosocial patient characteristics and GP‐registered chronic morbidity: A prospective study. J Psychosom Res. 2001;50(2):95–102.van den Akker M, Buntinx F, Metsemakers JF, Roos S, Knottnerus JA. Multimorbidity in general practice: Prevalence, incidence, and determinants of co‐occurring chronic and recurrent diseases. J Clin Epidemiol. 1998;51(5):367–75.Gold R, Michael YL, Whitlock EP, Hubbell FA, Mason ED, Rodriguez BL, et al. Race/ethnicity, socioeconomic status, and lifetime morbidity burden in the women's health initiative: A cross‐sectional analysis. J Womens Health. 2006;15(10):1161–73.Hudon C, Fortin M, Poitras ME, Almirall J. The relationship between literacy and multimorbidity in a primary care setting. BMC Fam Pract. 2012;13:33.Agborsangaya CB, Lau D, Lahtinen M, Cooke T, Johnson JA. Multimorbidity prevalence and patterns across socioeconomic determinants: A cross‐sectional survey. BMC Public Health. 2012;12:201.Ahluwalia IB, Mack KA, Mokdad A. Report from the CDC. Changes in selected chronic disease‐related risks and health conditions for nonpregnant women 18–44 years old BRFSS. J Womens Health (Larchmt). 2005;14(5):382–6.Alaba O, Chola L. The social determinants of multimorbidity in South Africa. Int J Equity Health. 2013;12:63.Andrade LH, Bensenor IM, Viana MC, Andreoni S, Wang YP. Clustering of psychiatric and somatic illnesses in the general population: Multimorbidity and socioeconomic correlates. Braz J Med Biol Res. 2010;43(5):483–91.Droomers M, Westert GP. Do lower socioeconomic groups use more health services, because they suffer from more illnesses? Eur J Public Health. 2004;14(3):311–3.Enroth L, Raitanen J, Hervonen A, Jylha M. Do socioeconomic health differences persist in nonagenarians? J Gerontol B Psychol Sci Soc Sci. 2013;68(5):837–47.Hosseinpoor AR, Bergen N, Mendis S, Harper S, Verdes E, Kunst A, et al. Socioeconomic inequality in the prevalence of noncommunicable diseases in low‐ and middle‐income countries: Results from the World Health Survey. BMC Public Health. 2012;12:474.Jerliu N, Toci E, Burazeri G, Ramadani N, Brand H. Prevalence and socioeconomic correlates of chronic morbidity among elderly people in Kosovo: A population‐based survey. BMC Geriatr. 2013;13:22.Khanam MA, Streatfield PK, Kabir ZN, Qiu C, Cornelius C, Wahlin A. Prevalence and patterns of multimorbidity among elderly people in rural Bangladesh: A cross‐sectional study. J Health Popul Nutr. 2011;29(4):406–14.Macleod U, Mitchell E, Black M, Spence G. Comorbidity and socioeconomic deprivation: An observational study of the prevalence of comorbidity in general practice. Eur J Gen Pract. 2004;10(1):24–6.Marengoni A, Winblad B, Karp A, Fratiglioni L. Prevalence of chronic diseases and multimorbidity among the elderly population in Sweden. Am J Public Health. 2008;98(7):1198–200.Mercer SW, Watt GC. The inverse care law: Clinical primary care encounters in deprived and affluent areas of Scotland. Ann Fam Med. 2007;5(6):503–10.Nagel G, Peter R, Braig S, Hermann S, Rohrmann S, Linseisen J. The impact of education on risk factors and the occurrence of multimorbidity in the EPIC‐Heidelberg cohort. BMC Public Health. 2008;8:384.Neeleman J, Ormel J, Bijl R. The distribution of psychiatric and somatic ill health: Associations with personality and socioeconomic status. Psychosom Med. 2001;63:239–47.Orueta JF, Nuno‐Solinis R, Garcia‐Alvarez A, Alonso‐Moran E. Prevalence of multimorbidity according to the deprivation level among the elderly in the Basque Country. BMC Public Health. 2013;13:918.Salisbury C, Johnson L, Purdy S, Valderas JM, Montgomery AA. Epidemiology and impact of multimorbidity in primary care: A retrospective cohort study. Br J Gen Pract. 2011;61(582):e12–21.Santos Machado V, de S, Valadares AL, Costa‐Paiva LH, Osis MJ, Sousa MH, Pinto‐Neto AM. Aging, obesity, and multimorbidity in women 50 years or older: A population‐based study. Menopause. 2013;20(8):818–24.Schafer I, Hansen H, Schon G, Hofels S, Altiner A, Dahlhaus A, et al. The influence of age, gender and socio‐economic status on multimorbidity patterns in primary care. First results from the multicare cohort study. BMC Health Serv Res. 2012;12:89.Tucker‐Seeley RD, Li Y, Sorensen G, Subramanian SV. Lifecourse socioeconomic circumstances and multimorbidity among older adults. BMC Public Health. 2011;11:313.Uijen AA, van de Lisdonk EH. Multimorbidity in primary care: Prevalence and trend over the last 20 years. Eur J Gen Pract. 2008;14 Suppl 1:28–32.Walker AE. Multiple chronic diseases and quality of life: Patterns emerging from a large national sample, Australia. Chronic Illn. 2007;3(3):202–18.Dalstra JA, Kunst AE, Borrell C, Breeze E, Cambois E, Costa G, et al. Socioeconomic differences in the prevalence of common chronic diseases: An overview of eight European countries. Int J Epidemiol. 2005;34(2):316–26.Hippisley‐Cox J, Coupland C, Vinogradova Y, Robson J, May M, Brindle P. Derivation and validation of QRISK, a new cardiovascular disease risk score for the United Kingdom: Prospective open cohort study. BMJ. 2007;335(7611):136.Woodward M, Brindle P, Tunstall‐Pedoe H. Adding social deprivation and family history to cardiovascular risk assessment: The ASSIGN score from the Scottish Heart Health Extended Cohort (SHHEC). Heart. 2007;93(2):172–6.Smith SM, Ferede A, O'Dowd T. Multimorbidity in younger deprived patients: An exploratory study of research and service implications in general practice. BMC Fam Pract. 2008;9(1):6.Wang HHX, Wang JJ, Wong SYS, Wong MCS, Li FJ, Wang PX, et al. Epidemiology of multimorbidity in China and implications for the healthcare system: Cross‐sectional survey among 162,464 community household residents in southern China. BMC Med. 2014;12:188.Ong BN, Richardson JC, Porter T, Grime J. Exploring the relationship between multi‐morbidity, resilience and social connectedness across the lifecourse. Health (London). 2014;18(3):302–18.Mendenhall E, Kohrt BA, Norris SA, Ndetei D, Prabhakaran D. Non‐communicable disease syndemics: poverty, depression, and diabetes among low‐income populations. Lancet. 2017;389(10072):951–63.Semrau M, Evans‐Lacko S, Alem A, Ayuso‐Mateos JL, Chisholm D, Gureje O, et al. Strengthening mental health systems in low‐ and middle‐income countries: The Emerald programme. BMC Med. 2015;13:79.Royal College of Psychiatrists. No Health Without Public Mental Health: The Case for Action. London (UK): RCP; 2010.Diederichs C, Berger K, Bartels DB. The measurement of multiple chronic diseases – A systematic review on existing multimorbidity indices. J Gerontol A Biol Sci Med Sci. 2011;66(3):301–11.Fortin M, Stewart M, Poitras ME, Almirall J, Maddocks H. A systematic review of prevalence studies on multimorbidity: Toward a more uniform methodology. Ann Fam Med. 2012;10(2):142–51.

Journal

Australian and New Zealand Journal of Public HealthWiley

Published: Jan 1, 2018

Keywords: ; ; ; ;

There are no references for this article.