Table 4.

Detailed analysis of risk scoring system for ≥ grade 2 diarrhea and skin rash

Score cut point	Observed riska (%)	Model estimated riska (%)	Sensitivity (%)	Specificity (%)	Likelihood ratioc
Diarrhea
≤75	0.5	≤0.8	100	0	1.0
>75 to ≤125	2.4	0.8–4.2	99	14	1.2
>125 to ≤175a	8.8	4.2–18.8	82	56	1.9
>175	25	>18.8	17	97	5.5
Skin rash
≤20	0.7	<1	100	0	1.0
>20 to ≤30	1.0	1–1.2	79	37	1.3
>30 to ≤40	1.9	1.2–1.6	32	92	3.9
>40c	6	>1.6	26	95	5.5

^aPatients with a risk score of >125 to ≤175 had a diarrhea prevalence of approximately 8.8 % during that cycle of lapatinib therapy as observed in the patient sample. Patients with scores of > 125 would have a model estimated diarrhea risk of > 4.2 %. Therefore in our analysis, we considered a diarrhea risk score of > 125 to be “high risk”

^bPatients with a risk score of > 40 had a rash prevalence of approximately 6 % during that cycle of lapatinib therapy. Therefore in our analysis, we considered a skin rash risk score of > 40 to be “high risk”

^cThe ratio of the probability of a positive test result, in the case of diarrhea, a risk score of 125 units or more among patients who actually developed ≥ grade 2 diarrhea to the probability of a positive test result among patients who did not develop such an event. Therefore, patients who developed ≥ grade 2 diarrhea were 1.9 times more likely than patients who did not develop diarrhea to have a risk score of 125 or more