Table 5.

Results on sensitivity to change – linear additive model predicting the value of the health metric

	Number	Coefficient	SE	p-value
Intercept		18.98	0.65	<0.0001
Health metric in wave 3		0.71	0.01	<0.0001
Gender (female)		−0.86	0.30	0.0047
Education (middle)		1.42	0.38	0.0002
Education (high)		1.22	0.38	0.0012
Income (middle)		0.86	0.37	0.0200
Income (high)		1.79	0.39	<0.0001
Incidence of:
High cholesterol	503	−0.57	0.61	0.3563
Angina	180	−0.73	1.00	0.4655
Heart attack	265	−0.94	0.83	0.2571
Osteoporosis	121	−1.17	1.21	0.3323
High blood pressure	325	−1.46	0.75	0.0502
Other heart disease	126	−1.94	1.19	0.1019
Abnormal heart rhythm	137	−2.27	1.15	0.0481
Asthma	91	−2.72	1,40	0.0514
Diabetes	138	−2.95	1.13	0.0093
Heart murmur	57	−3.06	1.73	0.0771
Arthritis	361	−4.06	0.71	<0.0001
Cancer	138	−4.15	1.12	0.0002
Stroke	91	−4.73	1.41	0.0008
Lung disease	91	−5.90	1.39	<0.0001
Parkinson’s disease	13	−6.51	3.56	0.0679
Psychiatric condition	110	−8.00	1.29	<0.0001
Heart failure	14	−9.62	3.90	0.0138
Dementia	66	−16.62	1.96	<0.0001

Regression coefficients, standard errors (SE), and p-values resulting from the linear additive model predicting the value of the health metric in wave 4 based on the incidence of health conditions within the last two years and controlled for the value of the health metric in wave 3 and other covariates. For the health conditions, the number of cases with incidence in the last two years is provided. Health conditions are sorted by increasing effect. The nonlinear effect of age is shown in Additional file 6

The reference categories are male, low education, low income and no incidence of the respective health condition within the last two years