Kinesiophobia is not required to predict chronic low back pain in workers: a decision curve analysis

Table 4 Net Benefits of prediction Models 1 and 2 compared to the NB of treating nobody or everybody and the consequences for the number of TPs at probability thresholds ranging from 5 to 60%

Pt^a	NB Model 1^b	NB Model 2 ^b	NB treat all	TP M1 vs all^c	TP M2 vs all^c	TP M2 vs M1^d
0.05	0.513	0.513	0.511	0	0	0
0.1	0.493	0.493	0.484	1	1	0
0.2	0.444	0.441	0.419	2	2	0
0.3	0.410	0.411	0.336	7	7	0
0.4	0.398	0.402	0.225	17	18	0
0.5	0.329	0.353	0.071	26	28	2
0.6^e	0.288	0.265	−0.162	45	43	−2

^aP_t is threshold probability
^b NB is Net Benefit of the Model compared to assuming everybody does not develop chronic LBP
^c TP is the increase in number of true positive patients when the Model is compared to assuming everyone has chronic LBP
^d TP is the increase in number of true positive patients when Model 2 is compared to Model 1
^e Only comparison until 60% was feasible because the NB for treat all was negative at higher percentages

ISSN: 1471-2474