Background With increasing life expectancy and aging populations, the prevalence of multimorbidity (two or more conditions in a person) is rising. Multimorbidity is progressively more common in older age, socioeconomically disadvantaged groups and associated with increased mortality and morbidity. Most studies on multimorbidity to date are cross-sectional in design or with limited follow-up.

We studied the development of multimorbidity across adulthood and early old age in a nationally representative birth cohort study.

Methods This study analysed data from the 1946 National Survey of Health and Development (NSHD). The analysis sample included participants who attended the age 36 assessment in 1982 and any one of the follow-up assessments (ages 43, 53, 63 & 69; N=3,723, 51% males). Information on 18 common conditions was based on a combination of self-report, prescribed medications and health records.

Conditions included diabetes, dyslipidaemia, hypertension, obesity, coronary heart disease, stroke, cancer, anaemia, respiratory-, kidney-, gastro intestinal-, skin-disorders, arthritis, Parkinson’s disease, epilepsy, depression & psychosis.

For all participants, we calculated a multimorbidity score at each age indicating the number of conditions accumulated over time and was the outcome of interest. Linear-spline mixed-effects modelling was used to study the population-average accumulation of conditions over time in different periods. We also assessed sex and socioeconomic differences in longitudinal trajectories of multimorbidity across the five ages (18,615 data points, mean: 5 data points/participant).

Childhood social class and adulthood educational level were used as socioeconomic indicators. Missing data was addressed using multiple imputation.

Results Proportion of participants with no conditions decreased progressively from 52% at age 36 to 7% at age 69. Multimorbidity (the number of conditions) increased progressively across all 4 periods (0.55, 95% CI [0.5, 0.6] for 1982–89, 0.63 [0.58, 0.7] for 1989–99, 0.70 [0.63, 0.78] for 1999–09 and 1.15 [1.04, 1.25] for 2009–15). Disadvantaged social class in childhood was associated with marginally increased multimorbidity in adulthood (0.08 [0.01, 0.15] for skilled/unskilled and 0.07 [-0.01, 0.15] for manual groups compared to professional/intermediate group). Higher educational attainment was associated with decreased risk for multimorbidity (-0.09 [-0.2, 0.01] for university degree and -0.13 [-0.2, -0.06] for General Certificate of Education (GCE) compared to those without education). Estimates for childhood social class were attenuated and no longer significant when adjusted for educational level. Women had marginally higher risk for multimorbidity compared to men (0.15 [0.09–0.2]).

Conclusion Multimorbidity increases progressively with age, with the socioeconomically more disadvantaged having greater multimorbidity.