Table 9.

Outcomes and consequences of contacting authors of eligible studies

Author	Number of contacting attemptsa	Willingness of authors to reply	Number of research questions answered	Additional research data provided by the contacted authors and its consequences
Motoyoshi et al. [6]	5 attempts	Unclear	0 of 6 questions	• No additional research data were provided. • No consequences for the risk of bias scores were therefore assigned.
Chen et al. [84]	3 attempts	Yes	1 of 1 question	• Outcome assessors were blinded. This information changed the risk of bias score for the domain ‘Bias in measurement of outcomes’ from ‘No information’ to ‘Low’ risk of bias.
Brisceno et al. [25]	7 attempts	Yes	6 of 6 questions	• Insertion torque was measured at complete insertion of the 8 mm implant length. This information was not sufficient to lower the risk of bias score for the domain ‘Bias in measurements of interventions’. • Personnel and outcome assessors were not blinded. This information changed the risk of bias score for the domain ‘Bias in measurement of outcomes’ from ‘No information’ to ‘Serious’ risk of bias. • The sample consisted of 23 implants with and 23 without root contact. This information changed the risk of bias score for the domain ‘Bias due to missing data’ from ‘Serious’ risk to ‘Moderate’ risk of bias. This information also permitted the calculation of various statistics and list them in a forest plot.
Wilmes et al. [33]	5 attempts	Yes	2 of 6 questions	• Animals were 8–10 months old. Most of our questions were not answered by the contacted authors and no consequences were therefore applied.
McEwan [32]	2 attempts	Yes	7 of 7 questions	• Animals were approximately the same age. Different screw types were randomly assigned to the mandibles. This information changed the risk of bias score for the domain ‘Bias due to confounding’ from ‘Moderate’ to ‘Low’ risk of bias. • Outcome assessors and personnel were not blinded. This information changed the risk of bias score for the domain ‘Bias in measurement of outcomes’ from ‘No information’ to ‘Serious’ risk of bias.

^aThis number refers to the total number of attempts by email to get an answer from a contacted author

This number also includes the number of attempts to contact a co-author(s). An initial attempt or a subsequent reminder attempt was each counted as one attempt. As soon as authors replied, successive emails were not counted as additional attempts. Ideally, only two attempts are made: (1) the email to request the ‘willingness to reply’ and (2) the email to get additional data from the contacted authors

Attempts of sending emails from other email addresses were not counted as additional attempts. Sending such emails could at times be indicated because our initial email could be identified as ‘spam mail’ and could then be deleted by the receiving internet provider