Reporting methodological issues of the mendelian randomization studies in health and medical research: a systematic review

Table 1 Percentage Reporting According to Suggested Guidelines in a Review of IV Publications Assessing Effects of Medical Interventions (n = 97)

Guideline	Count	Percentage
Empirically verified 1st assumption
Yes	66	68.0
No	31	32.0
Strength of the 1st assumption
Verified in data using F-statistic	28	28.9
Verified in data using F-statistic and R²	11	11.3
Verified in data using odds ratio	1	1.0
Not reported	57	58.8
Provided theoretical justifications for 2nd and 3rd assumption
Clearly Stated & Discussed	34	35.1
Lacked Clear Discussion	24	24.7
No Acknowledgment	39	40.2
Clearly reported falsification tests for 2nd and 3rd assumption
Reported two or more types	1	1.0
Reported exactly one type	7	7.2
Did not report any tests	89	91.8
Detection of pleiotropy
Yes	46	47.4
No	51	52.6
Clearly stated the effect to be estimates
The effect in the population (Average treatment effects, ATE)	1	1.0
Effect in the compliers (Local average treatment effects, LATE)	6	6.2
Both stated (ATE & LATE)	1	1.0
Not stated	89	91.8
Estimated causal effect bounds, under the 1st, 2nd, and 3rd assumption
Yes	18	18.6
No	79	81.4
Discussed theoretical justification for the pertinent fourth assumption
Stated and discussed homogeneity assumption (4 h)	1	1.0
Stated and discussed monotonicity assumption (4 m)	4	4.1
Stated and discussed both (4 h) and (4 m)	0	0.0
Stated but not discussed (4 h)	3	3.1
Stated but not discussed (4 m)	2	2.1
No acknowledgment of the 4th assumption	87	89.7
Modeling approach for the estimation was clearly described
The modeling approach clearly described	74	76.3
Lack of adequate description of the modeling approach	23	23.7
Conduct Sensitivity Analysis
Yes	43	44.3
No	54	55.7
Discussed Linkage Disequilibrium
Yes	25	25.8
No	72	74.2

ISSN: 1471-2288