Table 4.

Change in Logistic Regression (LR) coefficients when misclassification errors were modeled as Bayesian priors.

Condition (Yes vs. No)	LR coefficient	LR Bayes small errors	LR Bayes large errors
Behavior change	1.75	2.56	5.4
Third party consultation	0.65	0.82	1.43
Depression	0.58	0.72	1.21
Possible falls	0.41	0.51	0.81
GP home visit	0.40	0.42	0.67
Did not attend	0.36	0.43	0.67
Stroke	0.33	0.46	0.79
Cerebrovascular disease	0.26	0.25	0.41
Receives home care	0.18	0.24	0.41
Attended emergency room	0.18	0.22	0.36
Anxiety	0.18	0.18	0.32
Depressive symptoms	0.14	0.22	0.49
Constipation	0.09	0.13	0.23
Lower limb fracture	0.01	−0.001	0.006
Urinary tract infection	−0.02	0.03	0.06
Impaired mobility	−0.03	−0.003	0.05
Non-urgent hospital admission	−0.03	−0.03	−0.03
Social services involvement	−0.13	−0.21	−0.41
Living in a nursing home	−0.14	−0.14	−0.2
(Intercept)	−0.72	−0.82	−1.16