Table 2.

Included Cohort Studies

Author Publication Year Country	Population and setting	Definition and prevalence of multimorbidity (MM)	Functional decline outcome measure/s	Follow-up period Losses to follow-up (%)	Results
Abizanda 2014 Spain	General population (FRADEA Study) 842 adults aged >70 yrs	MM ≥2 chronic diseases in a specific period of time. 14 pre-specified conditions selected for prevalence and impact on disability/mortality Chronic diseases identified from medical records and coded via ICD-10 580 ≥ 2 conditions (69.0 %)	Barthel index (disability) Fried’s criteria (frailty)	2 years 7.5 % loss to follow-up	Disability and frailty was not associated with MM over two years.
Aarts 2012 Netherlands	Primary care (Maastricht Aging Study) 1184 adults aged 21–84 years	MM ≥2 chronic diseases co-occurring within one person Morbidities sourced from GP database including all current and past health problems by clinician 96 included conditions based on medical literature and clinical experience 35.5 % ≥ 2 chronic diseases	SF-36	3 and 6 years 16.4 % loss to follow-up	MM significantly associated with poorer physical functioning at all 3 follow-up points (p < 0.001) Significant steep decline in physical function between 3 and 6 year follow up in those with MM (p < 0.001) Participants whose morbidity status changed from baseline to 3 year follow up (either to single or MM) associated with significantly lower physical function (p < 0.001)
Bayliss 2004 USA	Primary care (Medical Outcomes Study) 2708 adults, mean age 57.6 years	No definition of MM reported Self-report of 7 pre-specified chronic conditions. Condition presence also sourced from records Conditions chosen as of high prevalence in practice and in literature 686 ≥ 2 chronic diseases (25.3 %)	SF-36 (PCS scores)	4 years 41.9 % loss to follow up	≥4 chronic diseases associated with significant decline in physical function (p < 0.05) Reduction in PCS by 6.5 used as criteria for clinically significant <4 chronic diseases no association with physical decline Congestive Heart Failure, diabetes and/or respiratory disease predictive of clinically significant decline in PCS (p <0.05)
Byles 2005 Australia	Primary care (Veteran’s Affairs Preventative Care Trial) 1417 adults ≥70 years	Co-existence of multiple diseases in the same individual Self-reported MM questionnaire consisting of 25 conditions Severity measure incorporated and included mild cognitive decline 1107 > 3 conditions (78.1 %)	SF-36	2 years 7.2 % lost to follow up	Quality of Life (QoL) decreases as number of conditions increases The presence of each condition associated with significantly lower SF-36 scores (except heart bypass, stroke and diabetes) Data not shown
Drewes 2011 Netherlands	General population (Leiden 85–plus study) 594 adults aged 85 years	MM ≥2 chronic diseases at age 85 years Chart confirmed presence of 9 common conditions pre-specified 234 ≥ 2 chronic diseases (39.4 %)	Groningen Activity Restriction Scale	5 years 53.9 % loss to follow up	Participants with MM had an accelerated progression of ADL (activities of daily living) disability over time compared to those without MM (95 % CI 0.21 -0.63, p < 0.001) MM demonstrated accelerated increase in ADL disability in older people with optimal cognitive function (95 % CI 0.39-0.95, p < 0.001) This was not observed in participants with lower MMSE scores.
Kiely 1997 USA	Community based (Sample first drawn 1982: Massachusetts state-supported home care programme) 1060 adults aged ≥65 years	No definition of MM reported Self-report of 5 pre-specified medical conditions MM numbers not reported	Functional Dependency Index (FDI)	3 years 22.5 % loss to follow up	Each additional medical condition resulted in a significant increase in the FDI score (p < 0.001) Rate of decline did not differ by total number of medical conditions (p =0.67)
Nikolova 2011 Canada	Community based (Research Program on Integrated Services for the Elderly) 1164 disabled adults ≥65 years Disability status estimated using the Functional Autonomy Measurement System (SMAF) Score ≥10 excluded	Comorbidity : number of chronic diseases Self-report of comorbidities using 16 item questionnaire Diseases not specified but grouped into four categories: 0-1 disease 2-3 diseases 4-5 diseases ≥6 diseases 1084 ≥ 2 diseases (93.1 %)	Functional status measured using 7 item IADL subscale of the OAR and Katz ADL index	3 years High rate of attrition discussed but loss to follow up number NR	Comorbidity burden is a strong predictor in developing IADL and ADL disability 6 diseases vs 0–1 disease OR (95 % CI) IADL 6.42 (1.52; 27.18) ADL 16.73 (3.08; 91.06) 4 –5 diseases vs 0–1 disease OR(95 % CI) IADL 1.20 (0.52; 2.80) ADL 0.89 (0.26; 2.98) 2–3 diseases vs 0 –1 disease OR(95 % CI) IADL 1.00 (0.46; 2.20) ADL 1.44 (0.49; 4.15) ≥6 morbidities-6 times more likely to develop ADL disability and 17 times more likely to develop IADL disability
Prior 2011 UK	Primary care 4672 adults aged ≥50 years	Comorbidity –number of chronic diseases Record confirmed condition counts over previous 2 years In addition to number of GP consultations for morbidity in 2 year period Specific cardiovascular and musculoskeletal conditions (n = 15) chosen as most prevalent in developed countries Stage of disease as proxy for severity 561 ≥ 1 CVD & MSK condition MM in overall group not reported	SF-12 (PCS)	3 years 46 % loss to follow up	Cardiovascular cohort: higher comorbidity and increasing severity in disease associated with greater deterioration in PCS. Significant deterioration shown for HTN (p < 0.001) with PCS score deteriorating by -0.86 over three years Musculoskeletal cohort: no association
Rigler 2002 USA	Community based (Veteran’s Affairs Medical Centre) 492 adults aged ≥65 years	Comorbidity scores: based on sum of the domains affected, and the sum of the domains which patients reported affected function. Self-report of 18 prevalent conditions from 8 organ domains via self-report 335 ≥ 2 diagnoses (68.1 %)	MOS-36 Physical Function Index Self-report ADL and IADL	1 year 7.2 % loss to follow up	Increasing comorbidity significantly associated with increased risk of future functional decline (p <0.001) OR 1.09: 2 conditions OR 2.41: ≥ 3 conditions Presence of ADL and IADL problems at baseline demonstrated to have a significant impact on new ADL problems developing at one year (p <0.001) OR 4.77: 1 IADL problem at baseline OR 15.6: 1 ADL problem at baseline