Abstract
We explored the role of age, gender, and socioeconomic status in the occurrence of chronic diseases and multimorbidity in 1099 elderly participants in the Kungsholmen Project. Cardiovascular and mental diseases were the most common chronic disorders. Of the participants, 55% had multimorbidity. Advanced age, female gender, and lower education were independently associated with a more than 50% increased risk for multimorbidity. Multimorbidity is the most common clinical picture of the elderly and may be increased by unhealthy behaviors linked to education.
Because of the aging of the population1 and the association of chronic diseases with advanced age,2 “multimorbidity,” defined as coexistence of 2 or more chronic diseases, is expected to become a common problem in elderly populations. Previous studies have reported multimorbidity prevalence rates ranging from 40% to 80%; these rates are higher among women than among men.3–5 Few studies have evaluated the distribution of chronic conditions and multimorbidity by socioeconomic status (SES).6,7 Recently, Chandola et al.8 showed that physical health deteriorated more rapidly with age among persons from the lowest occupational grade. We used clinical data from the Kungsholmen Project in Stockholm, Sweden (1987–2000)9 to explore the role of age, gender, and SES in the occurrence of chronic diseases and multimorbidity in the elderly population in Sweden.
The study population consisted of participants in the first follow-up (n = 1099) of the Kungsholmen Project, during which physicians performed a complete clinical examination on all persons. The diagnoses were based on clinical assessment, medical history (from the Stockholm Inpatient Register that records discharge diagnoses from Stockholm, Sweden, hospitals), laboratory data, and current drug use. A disease was classified as chronic if it was permanent, caused by nonreversible pathological alteration, or required rehabilitation or a long period of care.10
The International Classification of Diseases, Ninth Revision (ICD-9),11 was used for all diagnoses, with some exceptions. Deafness was defined as being unable to hear the interviewer’s voice, and visual impairment was defined as being blind or almost blind. Major depression was diagnosed by a psychiatrist according to Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV ), criteria12; Diagnostic and Statistical Manual of Mental Disorders, Revised Third Edition (DSM-III-R ), criteria were used to diagnose dementia13; and diagnosis of anemia followed the World Health Organization criteria.14 Multimorbidity was defined as any co-occurrence of 2 or more chronic conditions (among the 30 detected) in the same individual, whether coincidental or not.15 SES was evaluated according to education (years of schooling) and the main life occupation in a subgroup of 770 participants.16
The study population were aged 77 to 100 (mean = 84.6 years), and 77% were women. Nearly half of the participants had high education levels and worked in white-collar occupations (data supplement to the online version of this article at http://www.ajph.org).
Hypertension, dementia, and heart failure were the most common disorders, with a prevalence of 38%, 21%, and 18%, respectively, whereas all other diseases were less frequent (15%; data supplement to the online version of this article at http://www.ajph.org). Figure 1 shows prevalence figures, by age, of chronic diseases, grouped according to the ICD-9 classification. Cardiovascular disease prevalence did not differ by age or gender, whereas a higher proportion of mental disorders was found among the oldest-old (i.e., ≥ 85 years) than among the younger-old persons (i.e., 77–84 years; 36.4% vs 17.9%; P < .001) and among women than among men (29.1% vs 17.2%; P < .01). Although persons with low occupation-based SES had a higher proportion of cardiovascular diseases than did those with high SES (57.1% vs 52.3%; not significant), participants with low education levels had more mental diseases than did participants with high levels of education (31.4% vs 20.5%; P < .001). The prevalence of the other diseases did not significantly differ by SES.
Thirty percent of the population had only 1 disease, whereas 55% had multimorbidity. The median number of diseases among persons with multimorbidity was 3, ranging from 2 to 7. Multimorbidity prevalence figures are reported in Table 1. In multivariate logistic regression analyses, age, gender, and education were independently associated with multimorbidity (Table 1).
Occupation-based SES showed a crude association with multimorbidity (odds ratio [OR] = 1.5; 95% confidence interval [CI] = 1.0, 2.3) but not when adjusted for sociodemographic variables (Table 1). Because of the high correlation between education and SES, we stratified education and occupation-based SES into 4 groups. Low education level showed a strong association with multimorbidity independent of high or low occupation-based SES (OR = 1.9; 95% CI = 1.1, 3.3 and OR = 1.7; 95% CI = 1.0, 2.7, respectively).
In line with previous reports,17,18 cardiovascular diseases and mental conditions emerged as the most common chronic disorders. Although the prevalence of cardiovascular diseases did not differ by age or gender, the prevalence of mental conditions increased with age, mostly because of the high number of oldest-old women affected by dementia.
In agreement with previous reports,19,20 the prevalence of multimorbidity was as high as 55%; significantly higher prevalence rates were found among the oldest-old (P < .05) and lower-educated persons (P < .01).5 Female gender was also independently associated with multimorbidity. Because of the limited number of male survivors, the estimates among men could have lacked precision, but it is also possible that the few men who survive to old ages are a selected group of healthier persons. Lower education level was independently associated with multimorbidity, but a high occupation-based SES could not compensate for the increased risk of multimorbidity caused by low education level. Education increases health-related knowledge, affects lifestyle behaviors, and is strongly associated with family SES.21 Although we should be cautious when drawing conclusions because of possible selective survival, our findings suggest that unhealthy behaviors linked to educational level or SES in early life may still play a role in the health status of the very old.
Note. OR = odds ratio; CI = confidence interval. SES was estimated from main work attainment in a subgroup of 770 participants. aORs and 95% CIs for multimorbidity (i.e., 2 or more diseases vs no disease) caused by sociodemographic factors are also reported. bAdjusted for gender and education. cAdjusted for age and education. dData on educational background were missing for 7 participants. eAdjusted for age and gender. fAdjusted for age, gender, and education. FIGURE 1— Prevalence per 100 people of chronic disorders grouped according to International Classification of Diseases, Ninth Revision (ICD-9), classification, by age: Kungsholmen Project, Sweden, 1987–2000. Note. Distribution by age (77–84 and ≥ 85 years). * P < .01.
Morbidity Multimorbiditya No. Prevalence (95% CI) No. Prevalence (95% CI) OR (95% CI) All 335 30.5 (27.8, 33.2) 602 54.8 (50.8, 58.8) . . . Age, y 77–84 (Ref) 177 29.8 (26.1, 33.5) 306 51.6 (47.6, 55.6) 1.0 ≥ 85 158 31.2 (27.2, 35.2) 296 58.5 (54.2, 62.8) 1.9b (1.3, 2.8) Gender Men (Ref) 71 28.4 (22.8, 33.9) 128 51.2 (45.0, 57.5) 1.0 Women 264 31.1 (27.9, 34.2) 474 55.8 (52.5, 59.1) 1.5c (1.0, 2.2) Educationd High education (Ref) 164 30.9 (26.9, 34.8) 271 51.0 (46.7, 55.3) 1.0 Low education 168 29.9 (26.1, 33.7) 327 58.3 (54.2, 62.4) 1.6e (1.1, 2.3) Occupation-based SES High occupation-based SES (Ref) 138 28.3 (24.3, 32.3) 258 52.9 (48.5, 57.3) 1.0 Low occupation-based SES 76 27.0 (21.8, 32.2) 166 58.9 (53.2, 64.6) 1.1f (0.7, 1.9) 
This study has been supported by C. M. Lerici Foundation and the Swedish Council for Working Life and Social Research.
Human Participant Protection The research follows the guidelines of the Swedish Council for Research in the Humanities and Social Sciences, and the ethics committee of the Karolinska Institutet approved the study.
