A closer look at the baseline-observation-carried-forward (BOCF)

Figures & data

Table 1 Patient disposition

	Study 1		Study 2
	Placebo (N = 121) n (%)	Duloxetine 60/120 mg/day (N = 115) n (%)	Placebo (N = 128) n (%)	Duloxetine 60/120 mg/day (N = 128) n (%)
Completers of the 13-week treatment	98 (81.0)	84 (73.0)	111 (86.7)	93 (72.7)**
Discontinuation for any reason	23 (19.0)	31 (27.0)	17 (13.3)	35 (27.3)**
Adverse event	7 (5.8)	16 (13.9)*	7 (5.5)	24 (18.8)**
Subject decision	10 (8.3)	11 (9.6)	2 (1.6)	4 (3.1)
Protocol violation	2 (1.7)	2 (1.7)	2 (1.6)	3 (2.3)
Physician decision	2 (1.7)	1 (0.9)	0 (0.0)	2 (1.6)
Lost to follow-up	1 (0.8)	1 (0.9)	0 (0.0)	1 (0.8)
Lack of efficacy	1 (0.8)	0 (0.0)	5 (3.9)	1 (0.8)
Entry criteria not met	0 (0.0)	0 (0.0)	1 (0.8)	0 (0.0)

Notes:

* P < 0.05 versus placebo;

** P < 0.01 versus placebo.

Abbreviations: n, number of patients in the specified category; N, total number of randomized patients.

Table 2 Mean change analysis of Brief Pain Inventory average pain for all randomized patients in the 13-week treatment phase

Study	Analysis	Treatment groupa	LS mean change (SE)	Effect size	P valueb
Study 1c	MMRMd	DLX 60/120 mg QD	–2.32 (0.22)	0.36	0.004
		Placebo	–1.50 (0.21)
	BOCF	DLX 60/120 mg QD	–1.86 (0.20)	0.28	0.019
		Placebo	–1.25 (0.20)
	LOCF	DLX 60/120 mg QD	–2.09 (0.21)	0.28	0.019
		Placebo	–1.45 (0.21)
	mBOCF	DLX 60/120 mg QD	–1.91 (0.21)	0.25	0.041
		Placebo	–1.35 (0.21)
	aeBOCF	DLX 60/120 mg QD	–1.94 (0.21)	0.21	0.075
		Placebo	–1.46 (0.21)
Study 2e	MMRMd	DLX 60/120 mg QD	–2.72 (0.20)	0.41	<0.001
		Placebo	–1.88 (0.18)
	BOCF	DLX 60/120 mg QD	–2.23 (0.20)	0.28	0.013
		Placebo	–1.63 (0.19)
	LOCF	DLX 60/120 mg QD	–2.51 (0.20)	0.39	<0.001
		Placebo	–1.72 (0.18)
	mBOCF	DLX 60/120 mg QD	–2.29 (0.20)	0.32	0.005
		Placebo	–1.61 (0.19)
	aeBOCF	DLX 60/120 mg QD	–2.29 (0.20)	0.31	0.005
		Placebo	–1.62 (0.19)

a Study 1: N (DLX 60/120 QD) = 109, N (placebo) = 115.

Study 2: N (DLX 60/120 QD) = 121, N (placebo) = 127.

b P value comparison with placebo.

c Baseline mean (standard deviation): DLX 60/120 mg QD = 5.91 (1.61), placebo = 5.93 (1.67).

d Primary efficacy analysis in study 1.

e Baseline mean (standard deviation): DLX 60/120 mg QD = 6.09 (1.38), placebo = 6.16 (1.26).

Abbreviations: aeBOCF, discontinuation due to adverse events only; BOCF, baseline observation carried forward; DLX, duloxetine; LOCF, last observation carried forward; LS Mean, least-squares mean; mBOCF, modified BOCF; MMRM, mixed-model repeated measures; QD, once daily; SE, standard error.

Table 3 Thirty percent response rate to Brief Pain Inventory average pain

Study	Analysis	Treatment groupa	Responderb n (%)	P valuec
Study 1	BOCF	DLX 60/120 mg QD	50 (45.9)	0.056
		Placebo	38 (33.0)
	LOCF	DLX 60/120 mg QD	58 (53.2)	0.060
		Placebo	46 (40.0)
	mBOCF	DLX 60/120 mg QD	52 (47.7)	0.137
		Placebo	43 (37.4)
	aeBOCF	DLX 60/120 mg QD	53 (48.6)	0.226
		Placebo	46 (40.0)
Study 2	BOCF	DLX 60/120 mg QD	69 (57.0)	0.031
		Placebo	54 (42.5)
	LOCF	DLX 60/120 mg QD	79 (65.3)	<0.001
		Placebo	56 (44.1)
	mBOCF	DLX 60/120 mg QD	72 (59.5)	0.008
		Placebo	54 (42.5)
	aeBOCF	DLX 60/120 mg QD	72 (59.5)	0.008
		Placebo	54 (42.5)

a Study 1: N (DLX 60/120 mg QD) = 109, N (placebo) = 115.

Study 2: N (DLX 60/120 mg QD) = 121, N (placebo) = 127.

b Response was defined as at least a 30% reduction in BPI average pain.

c P value comparison with placebo.

Abbreviations: aeBOCF, discontinuation due to adverse events only; BOCF, baseline observation carried forward; BPI, Brief Pain Inventory; DLX, duloxetine; LOCF, last observation carried forward; mBOCF, modified BOCF; MMRM, mixed-model repeated measures; N, number of patients in the specified category; QD, once daily.

Table 4 Mean change analysis of BPI average pain (LOCF approach) for all randomized patients – by disposition status

	Study 1				Study 2
	Placebo		Duloxetine 60/120 mg/day		Placebo		Duloxetine 60/120 mg/day
	n	LS mean change (SE)	n	LS mean change (SE)	n	LS mean change (SE)	n	LS mean change (SE)
Completers	98	–1.51 (0.20)	84	–2.47 (0.22)***	111	–2.02 (0.17)	92	–3.04 (0.19)***
Discontinuation due to
Adverse events	4	0.85 (1.46)	13	–0.65 (1.06)	7	–0.88 (0.72)	20	–0.93 (0.45)
Lack of efficacy	4	–0.79 (1.44)	3	–0.79 (1.30)	6	–0.36 (0.23)	1	–1.62 (0.54)
Other reasons	9	–1.48 (0.95)	9	–0.49 (0.93)	3	0.07 (1.08)	8	–1.64 (0.69)

*** P ≤ 0.001 versus placebo.

Abbreviations: BPI, Brief Pain Inventory; LOCF, last observation carried forward; LS, least squares; n, number of patients in the specified category; SE, standard error.