Effects of non-medical health coaching on multimorbid patients in primary care: a difference-in-differences analysis

Table 7 Difference-in-differences estimates of effect of Enhanced Primary Care on the population (primary analysis)

Primary Analysis
Outcomes	Unadjusted means				Unadjusted difference-in-differences			N	Adjusted^a difference-in-differences (95% CI)
Outcomes	Treatment Pre/Post		Control Pre/Post		Simplified	Wave 1	Wave 2	N	Adjusted^a difference-in-differences (95% CI)
Health status
EQ-5D-5L	0.8192	0.8139	0.8060	0.8036	−0.0030	−0.0054	0.0054^‡‡	3,147,503	0.0004	(−0.0103 to 0.0111)
Physical functioning	0.8557	0.8571	0.8438	0.8455	−0.0004	−0.0029	0.0048^‡‡	3,242,123	0.0044	(−0.0116 to 0.0204)
Psychological wellbeing	0.9013	0.8892	0.8888	0.8824	−0.0056	−0.0143	0.0073^‡‡	3,267,482	−0.0034	(−0.0078 to 0.0011)
Resilience	0.8927	0.8929	0.8829	0.8817	0.0015	0.0022	0.0046^‡‡	3,238,682	0.0038	(−0.0139 to 0.0214)
Experience of care
Person-centeredness	0.8759	0.8723	0.8503	0.8523	−0.0056	−0.0055^‡‡	0.0010^‡‡	2,919,957	−0.0155	(−0.0494 to 0.0184)
Continuity of care	0.7465	0.7083	0.7037	0.6787	−0.0132	−0.0173^‡‡	−0.0187^‡‡	1,774,615	−0.0794	(−0.2058 to 0.0470)
Health behaviour
Smoking habit	0.2113	0.2034	0.2204	0.2128	−0.0002	0.0030	0.0046^‡‡	3,368,788	0.0088	(0.0009 to 0.0167)^‡
Health care utilisation
Primary care utilisation^b	0.7347	0.7299	0.7366	0.7294	0.0023	−0.0165^‡	0.0091	3,384,804	−0.0331	(−0.0448 to -0.0214)^‡‡

All unadjusted and adjusted values are weighted. Simplified unadjusted difference-in-differences assumes that all treated practices join the treatment at a single point in time. Wave 1 and 2 unadjusted difference-in-differences represent the isolated effects on treated practices that joined in each wave. Adjusted intervention difference-in-differences incorporates the gradual implementation of the treatment, and controls for covariates and practice and time fixed effects
^aAdjusted for gender, age, ethnicity, employment status, number of chronic conditions, time since last GP appointment, and practice and time fixed effects
^bAdjusted for gender, age, ethnicity, employment status, number of chronic conditions, and practice and time fixed effects
^‡p < 0.05, ^‡‡p < 0.01

ISSN: 1472-6963