Validating and updating a risk model for pneumonia – a case study

Table 2 Estimated regression coefficients: original expert model (Approach 1) compared to Approaches 4 and 5

Variable	Approach 1	Approach 4	Approach 5
Variable	Expert model	Random g	Maximum likelihood
Intercept	−79.19	−37.50	−36.31
Age	−0.0054	−0.0017	−0.0079
Duration of new/worsened cough	0.15	0.027	0.033
Maximum temperature	1.89	0.39	0.37
Dyspnea	−1.34	0.51	0.59
Dyspnea at effort only	2.63	−0.078	−0.11
Rigors	0.52	−0.14	−0.017
Number of cigarettes / day	0.023	0.017	0.0054
Current temperature	0.14	0.54	0.53
Signs of upper respiratory infection	1.23	−0.43	−0.61
Prolonged expiration	−1.40	0.12	0.16
Percussion dullness	0.84	1.44	1.60
Auscultation friction rub	0.97	1.39	1.41
Auscultation diminished insp. sound	0.41	0.69	0.44
Auscultation abnormality breath sound	0.52	0.66	0.73
(Age – 45)²	0.00017	0.00050	0.00058
(Duration of cough – 10)²	−0.00912	−0.0010	−0.00083
(Maximum temperature – 38.5)²	−0.83	0.057	0.046
(Current temperature – 38.5)²	−0.33	0.033	0.072

ISSN: 1471-2288