Skip to main content

Table 3 Data analysis and missing data

From: Adherence to participant flow diagrams in trials on postoperative pain management after total hip and knee arthroplasty: a methodological review

 

Trials with flow diagrams (240)

Trials without flow diagrams (330)

p value

Strategy for primary analysis

n (%)

n (%)

 

 Intention-to-treat

81 (34%)

23 (7%)

< 0.0001

 Modified intention-to-treat

3 (1%)

4 (1%)

1

 Per-protocol

15 (6%)

1 (0%)

< 0.0001

 Not mentioned

141 (59%)

302 (92%)

< 0.0001

Handling of missing outcome data

 Not mentioned

207 (86%)

309 (94%)

0.001

 Excluded from analysis

15 (6%)

14 (4%)

0.27

 Imputed

14 (6%)

3 (1%)

0.0007

 Last observation carried forward

4 (2%)

4 (1%)

0.32

Reasons for post-randomisation exclusions

 Non-adherence to the protocola

90 (35%)

106 (32%)

N/A

 None excluded

81 (32%)

43 (13%)

N/A

 Withdrawal of consent incl. dropouts/attrition

79 (31%)

38 (12%)

N/A

 Did not receive intervention

51 (20%)

26 (8%)

N/A

 Intractable pain/non-protocolled analgesicsb

34 (13%)

44 (13%)

N/A

 Wrongful inclusion

25 (10%)

12 (4%)

N/A

 No outcome data

24 (9%)

19 (6%)

N/A

 Group size too small to analyse

1 (0%)

0 (0%)

N/A

 Not mentioned

0 (0%)

147 (45%)

N/A

  1. RCTs may report multiple reasons for exclusion, why the sum is more than 100%
  2. aStopped taking medication/protocol violation, etc.
  3. bRequired rescue block, other opioids, etc.
  4. p values calculated by chi-square test