Table 4.

Examples of claims based on potentially biased evidence

Type	Drug	Claim	Reference	Reasons for potential bias
Abstract	Pantozol® (Pantoprazol)	Similar healing of reflux disease as omeprazol 20 mg	Bardnan KD et al. Gastroenterology 116(Suppl.4II): A 118, 1999	Abstract, open study, no blinded outcome assessment
Selection Bias	Neurodol® (Lidocaine)	Neurodol tissugel reduces pain and allodynia in postherpetic neuralgia	Meier T et al., Pain 106: 151–158, 2003	Selection bias (only patients who had already successfully been treated with lidocaine for more than 1 month were included in study), very few patients
Publication bias	Lamictal® (Lamotrigine)	Well tolerated	Bowden CL et al. Drug Safety 2004; 57: 173–184	Narrative review, possible publication and selective reporting of evidence bias
At least 3 quality criteria missing	Zoloft® (Sertraline)	Increases cognitive capacity	Newhouse PA et al. J Clin Psych 2000;61:559–68	High drop out rate(>30%), no descriptions of losses to follow-up, selective reporting of positive outcomes
Post hoc analysis	Jarsin® (St. John's word)	Similar efficacy as synthetic antidepressants	Brenner R et al. Clinical Therapeutics 22: 411–419, 2000	Post-hoc analysis, only 15 patients in each group, no sample size calculation