Table 2 Overview of limitations and biases of the cited studies
From: Effects of proton pump inhibitor (PPI) use on migraine – a critical review
Data on sex | Possible biases and limitations | |
|---|---|---|
Claessens et al | none | - no contact with a medical professional - low response rate – reproducibility in the general population? - participation was voluntarily and probably biased - only voluntarily reported diagnoses |
Liang et al | available | - survey performed on known risk factor for headache events - precise diagnosis of headache disorder was not available - reported coded diagnoses – no in patient confound factors, no clinical data - only patients with contact with a medical professional were reported - possible deviations in the claims data - case-crossover design does not account for within-person confounding (f.e. acute indication of PPI use, fluctuations in disease severity) - could not delineate whether the clinic visits for headache were because of headache exacerbation or new onset headache based on the claim data |
Makunts et al | none | - survey performed on known risk factor for headache events - adverse event reporting is voluntary, so frequencies etc. could not be referred to general population - occasionally missing demographic variables - lack of comprehensive medical record data - no in-patient confound factors - some concurrent medications and comorbidities may be also underreported, which in turn may affect the cohort composition and statistical analysis - over-the-counter medication and supplement use still remains a significant unknown variable as it relies on patient self-reporting - physiological mechanisms cannot be derived from this study due to its observational nature – no causality can be implied |
Pisanu et al | available | - survey performed on known risk factors for headache events - over-the-counter medication and supplement use still remains a significant unknown variable as it relies on patient self-reporting |
Kang et al | available | - survey performed on known risk factor for headache events - reported coded diagnoses – no in patient confound factors - only patients with contact with a medical professional were reported - no causality could be extracted from this study due to its retrospective nature - medication adherence could not be monitored - unmeasured confounding influences could not be eliminated (f.e. NSAID intake) - confounding effect of missing data (like f.e. genetic background, family history) was not considered - no evaluation between migraine and different kinds of PPI |
Slavin et al | available | - survey performed on known risk factor for headache events - reported coded diagnoses – no in patient confound factors - only patients with contact with a medical professional were reported - cannot imply causality due to study design - no data on confounding factors |