Skip to main content

Table 4 Effect of added reference items and the discrepancy rule on the success rate of treatment

From: The index ‘Treatment Duration Control’ for enabling randomized controlled trials with variation in duration of treatment of chronic pain patients

Occasion of evaluation

Mode of adding reference items

n TDC ≤ −0.379

n TDC > −0.379

n discrepancy rule

n S-Tx

n U-Tx

Success rate (%)

PEM

A

100

18

8

92

26

78.0

PEM

NA

100

18

8

92

26

78.0

EM

s-A

93

25

4

89

29

75.4

EM

c-A

96

22

8

88

30

74.6

EM

NA

98

20

6

92

26

78.0

LM

s-A

70

48

4

66

52

55.9

LM

c-A

78*

40*

9*

69*

49*

58.5*

LM

NA

79*

39*

6*

73*

45*

61.9*

  1. Occasion of evaluation of treatment (Tx) success rate: PEM (clinician); EM, (assessor); LM, (assessor; cf legend of Figure 2). Mode of adding reference items: A, added (by clinician); s-A, separately added by clinician and by investigator (based on data from assesor), and only the added reference items from the assessor are considered in the Tx evaluation; c-A, continually added by clinician and subsequently by the investigator (based on data from assessor), and all added reference items are considered in the Tx-outcome; NA, no addition. n TDC ≤ −0.379 and n TDC > −0.379: number of patients for which TDC ≤ −0.379 or TDC > −0.379 respectively. n discrepancy rule: number of patients with application of the ‘discrepancy rule’ (see text, section ‘treatment procedure’). n S-Tx and n U-Tx: number of patients with a successful Tx and a unsuccessful Tx respectively. Note that n S-Tx = [(n TDC ≤ −0.379) – (n discrepancy rule)], and n U-Tx = [(n TDC > −0.379) + (n discrepancy rule)]. Note also that the application of the discrepancy rule was occasional, i.e. for 3.4-7.6% of the patients.
  2. Success rate (%) = (n S-Tx/118)×100% (118 = total number of patients).
  3. *, values based on n = 89 patients who entered the follow-up at EM according to separately added reference items (s-A) from the assessor (the mode of addition used in the RCT of the present study). Because the number of patients entering the follow-up at EM would have been slightly larger according to NA (n = 92) than actually occurring according to s-A (n = 89), the success-rate (n S-Tx) at LM might be slightly underestimated for the mode NA. The success-rate at LM is approximately correctly estimated for the mode c-A as the number of patients entering the follow-up was nearly the same for c-A (n = 88) as for s-A (n = 89). Note that regardless of a possible underestimation of n S-Tx for LM, NA, the value of success-rate is the smallest for LM, s-A.
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us