Background Research on sickness absence has typically focused on single diagnoses, despite increasing recognition that long-term health conditions are highly multimorbid and clusters comprising coexisting mental and physical conditions are associated with poorer clinical and functional outcomes. The digitisation of sickness certification in the UK offers an opportunity to address sickness absence in a large primary care population.

Methods Lambeth Datanet is a primary care database which collects individual level data on GP consultations, prescriptions, Quality and Outcomes Framework (QOF) diagnostic data, sickness certification (fit note receipt) and demographic information (including age, gender, self-identified ethnicity, and truncated postcode). We analysed 326,415 people’s records covering a 40-month period from January 2014 to April 2017.

Results We found significant variation in multimorbidity by demographic variables, most notably by self-defined ethnicity. Multimorbid health conditions were associated with increased fit note receipt. Comorbid depression had the largest impact on first fit note receipt, more than any other comorbid diagnoses. Highest rates of first fit note receipt after adjustment for demographics were for comorbid epilepsy and depression (HR 4.0; 95% CI 3.4–4.6), followed by chronic pain and depression (HR 3.9; 95% CI 3.5–4.4) and cardiac condition and depression (HR 3.9; 95% CI 3.2–4.7).

Conclusion Our results show striking variation in multimorbid conditions by gender, deprivation and ethnicity, and highlight the importance of multimorbidity, in particular comorbid depression, as a leading cause of disability in the population. The findings highlight the importance of multi-morbidity, particularly comorbid depression, as the leading cause of disability among working age adults.