Table 2 Methodological standards for quality assessment of included studies

From: Diagnostic accuracy systematic review of rectal bleeding in combination with other symptoms, signs and tests in relation to colorectal cancer

Study ID Were selection criteria clearly described? (Whiting et al, 2006) Was the spectrum of patients representative of the patients who will receive the test in practice? (Whiting et al, 2006) Were all outcome events and predictors clearly defined? (Laupacis et al, 1997) Did the whole sample or a random selection of the sample receive verification using a reference standard? (Whiting et al, 2006) Did patients receive the same reference standard regardless of their symptoms and signs? (Whiting et al, 2006) Is the reference standard likely to correctly classify the target condition? (Whiting et al, 2006) Were the reference standard results interpreted without knowledge of the symptoms and signs? (Whiting et al, 2006) Was there a 100% follow-up of those enrolled? (Laupacis et al, 1997)
Were withdrawals from the study explained? (Whiting et al, 2006)
Du Toit et al (2006) Yes Yes No Unclear
“A small number of patients may not have entered the diagnostic protocol, despite of frequent reminders.” (p 69)
No
-Rigid sigmoidoscopy
-Flexible sigmoidoscopy
-Colonoscopy
Suboptimal
Suboptimal
Best method available
Unclear -No
-Withdrawals not explained
Ellis and Thompson (2005) Yes Yes No Yes No
-Flexible sigmoidoscopy (219 pt)
-Patient questionnaire (47 pt)
-Flexible sigmoidscopy & questionnaire (53 pt)
-Barium enaema (37 pt)
-Colonoscopy (24 pt)
Suboptimal
Not suitable
Suboptimal
Suboptimal
Best method available
Unclear Unclear
Fijten et al (1994) Yes Yes Yes Yes No
-Sigmoidoscopy (8 pt)
-Colon-roentenography (8 pt)
-Proctoscopy (7 pt)
Sonography (5 pt)
-Colonoscopy (2 pt)
Suboptimal
Suboptimal
Suboptimal
Suboptimal
Best method available
Yes -No
-Yes
“21 patients excluded because lost of follow-up. (moved to an unknown destination)”
Heintze et al (2005) Yes Yes Yes No
The selection of patients having further investigation was not at random
No
-Sonography (52 pt)
-Rectoscopy (29 pt)
-Sigmoidoscopy (26 pt)
-Colonoscopy (195 pt)
Suboptimal
Suboptimal
Suboptimal
Best method available
Unclear Unclear
Mant et al (1989) Yes Yes No Yes No
-Total colonoscopy (104 pt)
-Endoscopy to at least 30 cm and an air-contrast barium enaema (32 pt)
- Investigations not complete, but an obvious source was found, for example, rectal cancer at proctoscopy. (9 pt)
Best method available
Suboptimal
Suboptimal
Unclear Unclear
Metcalf et al (1996) Yes Yes Yes Yes Yes
Colonoscopy 98 pt
Barium enaema 1 pt (because colonoscopy was impossible)
Best method available Unclear
“The questionnaire was re-administered by the colonoscopist before the procedure” (p162)
Unclear
Norrelund and Norrelund (1996) (1) Yes No
Selection made during recruitment. GPs were allowed to include a maximum of 3 patients.
Yes Yes No
GPs were asked to arrange either a barium enaema or a colonoscopy at the first consultation. (p161)
Suboptimal
“Although the authors asked the GP to refer all patients for a full colon examination, but this was no inclusion criterion.” (p165)
Unclear Unclear
Norrelund and Norrelund (1996) (2) Yes No
Selection made during recruitment. GPs were allowed to include a maximum of 4 patients.
Yes Yes No
GPs were asked to arrange either a barium enaema or a colonoscopy at the first consultation. (p161)
Suboptimal Unclear Unclear
Wauters et al (2000) Yes Yes No
Predictors clearly defined, but outcomes only colorectal cancer reported.
Yes No
“Our reference standard was colorectal cancer diagnosed during a clinical follow-up of 18–30 months. Investigations, such as endoscopy, were not systematically performed.” (p 998)
Suboptimal Unclear Unclear
  1. Abbreviations: GP=general practitioners; pt=patient. The page numbers refer to the original text of the included studies.