Skip to main content

Table 3 Detailed information of what had been reported in the included studies regarding the inconsistency between protocols or registrations and full reportsa

From: A systematic review of comparisons between protocols or registrations and full reports in primary biomedical research

First author, year

Levels of inconsistent reporting between protocols or registrations and full reports

Participant eligibility criteria

Sample size

Randomization

Interventions and their delivery

Blinding

Outcome measure

Study duration

Statistical analysis choice (e.g., model selection, model adjustment, missing data handling, intention-to-treat, et al)

Subgroup analysis

Funding source

Othersb

Al-Marzouki, 2008 [17]

–

–

–

–

–

30%

–

–

36%

–

–

Chan, 2004 (from CMAJ) [19]

–

–

–

–

–

40%

–

–

–

–

–

Chan, 2004 (from JAMA) [7]

–

–

–

–

–

62%

–

–

–

–

–

Boonacker, 2011 [18]

–

–

–

–

–

–

–

–

75%

–

–

Chan, 2008 [3]

–

53%

–

–

–

60%

–

Handling missing data: 80%

Interim analyses: 62%

Model adjustments: 82%

100%

–

 

Dekkers, 2015 [11]

–

–

–

–

–

–

–

Noninferiority margin definitions reported: 9%

Noninferiority margins and confidence intervals reported: 28%

–

–

–

Hahn, 2002 [20]

–

–

–

–

–

33%

–

Analysis plans reported: 88%

–

–

–

Hannink, 2013 [21]

–

–

–

–

–

49%

–

–

–

–

–

Hartung, 2014 [22]

–

–

–

–

–

15%

–

–

–

–

 

Hernandez, 2005 [23]

–

–

–

–

–

–

–

–

100%

–

–

Kasenda, 2014 [9]

–

–

–

–

–

–

–

–

38%

–

–

Killlen, 2014

–

–

–

–

–

29%

–

–

–

–

–

Korevaar, 2014 [25]

12%

–

–

–

–

21%

–

 

–

–

Result presentations: 6%

Li, 2013 [26]

–

–

–

–

–

14%

–

–

–

–

–

Mathieu, 2009 [27]

–

–

–

–

–

31%

–

–

–

–

–

Maund, 2014 [28]

–

–

–

–

–

–

–

Primary efficacy analyses reported: 14%

–

–

Adverse events reported: 100%

Melander, 2003 [29]

–

–

–

–

–

–

–

Intention-to-treat or per-protocol analyses reported: 93%

–

–

–

Mhaskar, 2012 [30]

–

59%

77%

72%

53%

–

–

Intention-to-treat analyses reported: 70%

–

–

 

Milette, 2011 [31]

–

–

–

–

–

100%

–

–

–

–

–

Nankervis, 2012 [32]

–

–

–

–

–

72%

–

–

 

–

–

Norris, 2014 [33]

–

–

–

–

–

83%

–

–

20%

–

Adverse events reported; 43%

Redmond, 2013 [34]

–

–

–

–

–

29% (primary outcomes: 17%; secondary outcomes: 33%)

–

–

–

–

–

Riehm, 2015 [35]

–

–

–

–

–

33%

–

–

–

–

–

Rising, 2008 [36]

–

–

–

–

–

31%

–

–

–

–

Conclusions reported: 9%

Riveros, 2013 [37]

–

–

–

–

–

87%

–

–

–

–

Adverse events reported: 37%

Rongen, 2016 [38]

–

–

–

–

–

54%

 

–

–

–

–

Rosati, 2016 [39]

45%

40%

–

–

–

100%

Studies discontinued early without justifications:65%

Intention-to-treat analysis: 10%

–

45%

 

Rosenthal, 2013 [40]

–

27%

2%

–

23%

Primary outcomes: 45%

Secondary outcomes: 67%

Start of patient enrolment reported: 43%

End of enrollment reported: 71%

–

–

45%

Ethical committee approval: 2%

Saquib, 2013 [41]

–

–

–

–

–

–

–

Analysis plans reported: 47%

–

–

–

Smith, 2013 [42]

–

–

–

–

–

79%

–

–

–

–

–

Soares, 2004 [48]

–

60%

–

59%

–

–

–

Intention-to-treat analysis: 14%

–

–

Dropouts reported: 9%

Su, 2015 [43]

–

–

–

–

–

45%

–

–

–

–

–

Turner, 2012 [44]

–

–

–

–

–

25%-

–

–

–

–

Effect sizes reported: 8%

Vedula, 2009 [45]

–

–

–

–

–

67%

–

–

–

–

–

Vedula, 2013 [10]

–

–

–

–

–

–

–

Intention-to-treat analysis definitions: 67%

Safety analysis definitions: 50%

–

–

–

Vera-Badillo 2013 [46]

–

–

–

–

–

23%

–

–

–

–

–

You, 2012 [47]

–

–

–

–

–

14%

–

–

–

–

–

  1. aCells with a ‘-’ indicated the information was not reported in the included studies
  2. bOther inconsistency measures including comparisons of effect size, sample size, control, ethics, key finding reporting, and/or conclusion reporting