Table 7.

Limitations of the included studies

Study	Limitations of the included studies
Zawawi 2014 [85]	• No published or registered protocol of the research study (risk of selective reporting) • The research design (survey) • Questionnaires with only closed-ended questions • Mostly non-defined questions • No pilot testing of research procedures • No reference was made to a reference standard for assessing implementation constructs. Survey instruments were not validated • Unclear selection procedures but were clarified through contacting the author • Unclear whether the answers of parents and siblings (13.3 % of the respondents) were representative for those of the stakeholders • The surveyed population was recruited from 2 different settings • Prior knowledge on OMIs of 12.7 % (21/165) of respondents • Additional unclear issues in Tables 2, 3, 4, and 5 and among the critical appraisal scores (Additional file 4)
Meeran 2012 [8]	• No published or registered protocol of the research study (risk of selective reporting) • The research design (survey) • Questionnaires with only closed-ended questions • Mostly non-defined questions • No pilot testing of research procedures • No reference was made to a reference standard for assessing implementation constructs. Survey instruments were not validated • Unclear selection procedures • Summing the total numbers of respondents in the different settings does not add up to the total number of respondents referred to in other parts of the results • The surveyed population was recruited from 2 different settings and were not equally distributed. Urban respondents were overrepresented in the total sample and had less non-users of OMIs compared with the non-urban population. This issue could have skewed outcomes • Additional unclear issues in Tables 2, 3, 4, and 5 and among the critical appraisal scores (Additional file 4)
Bock 2015 [5]	• No published or registered protocol of the research study was identified (risk of selective reporting) • The research design (survey) • Questionnaires with only closed-ended questions • Mostly non-defined questions • No pilot testing of research procedures • No reference was made to a reference standard for assessing implementation constructs. Survey instruments were not validated • Unclear terminology in the questionnaire had created confusion among respondents on the correct interpretation of questions. This misunderstanding could have resulted in an overestimation of all around users and OPI-only users and therefore an underestimation of MSC-only users, but this issue did not affect our populations of interest • The numerators and denominators were not completely clear in the published article and were confirmed through contacting the authors of this research study • Both strict non-users of OMIs and users of osseointegrated palatal implants were included among the respondents, i.e., 417 (95 %) strict non-users of OMIs and 22 (5 %) osseointegrated palatal implants users. Only strict non-users of OMIs would have been our preferred target population • Various types of experience among subgroups of non-users of OMIs • The 439 non-users were adequately described, but the characteristics of the 345 respondents on implementation constructs of these 439 non-users were not described • The response rate of the non-users of OMIs was 78.6 % (345/439). The rationale and the consequences of this dropout were not described • Additional unclear issues in Tables 2, 3, 4, and 5 and among the critical appraisal scores (Additional file 4)