Researchers from the University Medical Center Groningen developed multimorbidity patterns for adults aged 60 years and older using a large number of chronic conditions. The derived variables are avaialble for the first assessment (1A), second assessment (2A), and third assessment (3A). (sections: Diseases & symptoms (TOBEADDED ) and secondary & linked variables).

The score can be requested in the Lifelines catalogue in the future.When this data has been used in your research, you will have to include a reference to the the following paper:

• Ogaz-González, R., Zou, Q., Du, Y., Gutiérrez-Robledo, L. M., Escamilla-Santiago, R., López-Cervantes, M., & Corpeleijn, E. (2025). Multimorbidity patterns in older adults from Northern Netherlands: Comparing factor analysis and latent class analysis solutions. BMC Geriatrics, 25, 381. DOI: https://doi.org/10.1186/s12877-025-06029-x
  

Multimorbidity is commonly defined as the presence of two or more chronic non-communicable diseases (NCDs) in the same individual. However, a simple disease count does not describe which diseases occur together or distinguish between people with substantially different health profiles. The multimorbidity configurations in this secondary dataset provide a more informative classification of chronic disease combinations among older adults participating in Lifelines.

The patterns were derived among adults aged 60 years or older. Forty-five chronic conditions were grouped into 14 non-communicable disease domains: cancer, major cardiovascular events, other heart and peripheral vascular conditions, hypertension, arthritis, osteoporosis, digestive diseases, diabetes, kidney diseases, respiratory diseases, thyroid diseases, neurodegenerative disorders, depression, and obesity. A domain was considered present when the participant met at least one applicable criterion based on self-reported diagnosis, treatment, medication use, clinical measurements, laboratory results, or medical procedures.

Latent class analysis was used to identify groups of participants with similar combinations of disease domains. Models containing one to seven classes were compared, and the final solution was selected using model fit, interpretability, and consistency between the first and second assessment waves. Five configurations were identified: Vascular, Metabolic Risk, Heart & Vascular Conditions, Major Cardiovascular Disease (CVD) & Vascular Conditions, and Complex Treatment Spectrum (Ogaz-González et al., 2025).

These variables allow researchers to study multimorbidity beyond conventional disease counts. They can be used as exposures, outcomes, stratification variables, or covariates in studies of healthy ageing, frailty, disability, healthcare use, mortality, lifestyle, and social inequalities. The class names summarize the dominant characteristics of each group; they are not clinical diagnoses and do not imply that every participant has all conditions represented by a class name.

For each general assessment, the number of affected disease domains was calculated by summing the 14 binary indicators. Participants with fewer than two affected domains were classified as having no multimorbidity, while those with two or more were classified as having multimorbidity.

Latent class membership was estimated separately for the first and second assessment waves. The resulting variable describes membership in the latent disease configurations and includes a separate No NCDs category. The publication variable mcs combines the baseline latent class membership with the conventional multimorbidity definition. Participants with fewer than two disease domains are classified as No multimorbidity, regardless of their initial latent class. The mcs variable is therefore recommended for reproducing the baseline analyses reported in the publication.