Background Health services, guidelines and epidemiology largely focus on single conditions. However, multimorbidity (typically defined as >1 long-term condition), is an increasingly important, but inconsistently conceptualised, socio-economically patterned phenomenon challenging this approach. This paper explores whether using both short and long-term conditions to classify multimorbidity helps to better understand the health outcomes of working-age adults.

Methods Population: adults aged 16–64 in Scotland. Data: 2008–2011 Scottish Health Survey. Sample design: stratified, geographically unclustered, random probability Postcode Address File (PAF) sample (designed to facilitate combined years’ analysis). Interviewers recorded long-term conditions (>1 year) and “any other health problems” verbatim (without objective verification). Six health states were derived, combining long-term conditions (0, 1, >1) and absence/presence of other health problems. Outcomes: GHQ12 score >4; Warwick-Edinburgh Mental Wellbeing Scale (WEMWBS) score >1 SD below population mean; life-satisfaction below population mode; bad/very bad self-rated health. Computer-assisted personal interviews and self-completion questionnaires (GHQ12/WEMWBS) were used. The associations between health states and outcomes (adjusted for age, sex and area deprivation), were estimated using logistic regression. The reference category was 1 long-term condition and no other problems (to estimate the effect of multiple conditions, and other problems, separately).

Results Differential selection and non-response weights were applied to percentages and odds ratios (ORs). Denominators are unweighted. Multimorbidity increases markedly with age, so analysis was restricted to ages 16–64 (n=21,369). 36% had >1 long-term condition, 13% reported other health problems (16% of those with multiple long-term conditions, n=3019). For brevity, only GHQ12 results are presented; the patterns for all other outcomes were identical. Compared with people with 1 long-term condition and no additional problems (n=4213), the OR of having a GHQ12 score of >4 were (p > 0.001): >1 long-term condition and additional health problems, 3.5 (95% CI: 2.8–4.4) [n=458]; >1 long-term condition and no additional problems, 2.4 (2.1–2.7) [n=2315]; 1 long-term condition and other health problems, 1.8 (1.5–2.2) [n=631]; No long-term conditions but other health problems, 0.8 (0.7–0.9) [n=1711]; No long-term conditions or other health problems, 0.5 (0.4–0.5) [n=10,482].

Discussion Focusing solely on long-term conditions doesn’t capture the extent of people’s condition burden: having long-term conditions and additional health problems is significantly associated with worse physical and mental health, with the effect-size greatest for people with multiple long-term conditions. Better understanding of the interaction between longer and shorter-term conditions (including better case-finding, iatrogenesis and overtreatment) and their impact on outcomes will help to improve interventions for this highly heterogeneous – and socially disadvantaged – group.

Conclusion Long-term conditions are only part of the story.