Table 3.

Adjusted IRR of multimorbidity-related variables by primary care consultation type across three model specifications^a

	IRR (95% CI)
	Total	GP	Nurse	Other	Face to face	Telephone	Home
Model 1: multimorbidity clusters, count LTCs, interaction, and polypharmacy
Mental health+	1.24 (1.23 to 1.24)	1.24 (1.23 to 1.25)	1.12 (1.11 to 1.14)	1.26 (1.24 to 1.28)	1.22 (1.21 to 1.23)	1.28 (1.26 to 1.30)	1.33 (1.29 to 1.38)
Cardiovascular+	1.10 (1.09 to 1.10)	1.10 (1.10 to 1.11)	1.05 (1.04 to 1.07)	1.10 (1.08 to 1.11)	1.06 (1.06 to 1.07)	1.18 (1.17 to 1.19)	1.23 (1.21 to 1.24)
Pain+	1.11 (1.10 to 1.11)	1.11 (1.11 to 1.11)	1.08 (1.07 to 1.09)	1.12 (1.11 to 1.13)	1.09 (1.09 to 1.09)	1.17 (1.16 to 1.18)	1.24 (1.23 to 1.25)
Liver+	1.22 (1.14 to 1.29)	1.20 (1.12 to 1.28)	1.27 (1.13 to 1.43)	1.23 (1.06 to 1.43)	1.20 (1.13 to 1.27)	1.37 (1.21 to 1.55)	1.20 (0.96 to 1.49)
Dependence+	1.33 (1.29 to 1.37)	1.35 (1.32 to 1.40)	1.14 (1.03 to 1.26)	1.41 (1.33 to 1.49)	1.30 (1.26 to 1.35)	1.48 (1.40 to 1.55)	1.36 (1.25 to 1.48)
Unclustered LTCs	1.11 (1.10 to 1.11)	1.11 (1.10 to 1.11)	1.07 (1.06 to 1.08)	1.13 (1.12 to 1.14)	1.09 (1.08 to 1.09)	1.16 (1.15 to 1.16)	1.17 (1.16 to 1.18)
Model 2: multimorbidity only
Multimorbidity (yes)	2.64 (2.63 to 2.66)	2.74 (2.72 to 2.75)	2.28 (2.27 to 2.32)	2.68 (2.64 to 2.69)	2.56 (2.53 to 2.59)	3.16 (3.10 to 3.19)	5.47 (5.37 to 5.58)
Model 3: multimorbidity and polypharmacy
Multimorbidity (yes)	2.30 (2.29 to 2.32)	2.36 (2.36 to 2.49)	1.92 (1.90 to 1.93)	2.36 (2.34 to 2.39)	2.24 (2.20 to 2.25)	2.61 (2.59 to 2.64)	3.83 (3.74 to 3.91)
Polypharmacy (yes)	2.20 (2.18 to 2.21)	2.29 (2.27 to 2.30)	2.36 (2.32 to 2.39)	1.95 (1.93 to 1.97)	2.16 (2.14 to 2.16)	2.53 (2.51 to 2.56)	4.04 (3.97 to 4.10)

^a n = 5 243 478 person–years, corresponding to 826 166 individuals. Data from April 2005 to March 2020 are used. All models also adjust for age, sex, ethnic group, Index of Multiple Deprivation quintiles, and language. Multimorbidity clusters: Mental health+ includes anxiety and depression; Cardiovascular+ includes heart failure, Peripheral Arterial Disease (PAD), osteoporosis, atrial fibrillation, coronary heart disease, chronic kidney disease, stroke/transient ischaemic attack, and dementia; Pain+ includes osteoarthritis, cancer, chronic pain, hypertension, and diabetes; Liver+ includes chronic liver disease and viral hepatitis; Dependence+ includes alcohol dependence, substance dependence, and HIV; Unclustered LTCs include: Parkinson’s disease, chronic obstructive pulmonary disease, asthma, inflammatory bowel disease, lupus, multiple sclerosis, rheumatoid arthritis, morbid obesity, cognitive and learning disabilities, sickle-cell anaemia, serious mental illness, and epilepsy. The reference category for both multimorbidity and multimorbidity clusters is not having multimorbidity. Categories of self-ascribed ethnic group include White, Black (Black/African/Caribbean/Black British), Asian (Asian/Asian British), mixed ethnicity, other, or unknown. In model 3, the count of LTCs and multimorbidity clusters are included as main effects, along with an interaction between the two variables. Parameter estimates of the main effects cannot be interpreted by themselves anymore because of the interaction. Simple slopes (marginal effect of the continuous variable — number of LTCs — across the different levels of the categorical variable–clusters) are computed instead, and incidence rates generated by exponentiating simple slopes. For example, in model 3, the IRR for each cluster indicates the effect of developing one more LTC for individuals in that specific cluster. For the Dependence+ cluster, IRR 1.33, so for a one unit increase in the number of LTCs, the incidence rate of primary care consultations increases by 33%, while in the Cardiovascular+ cluster it increases by 10%. IRR = incidence rate ratio. LTC = long-term condition.