The results of the searches and study selection are presented in Figure 1. The initial and updated searches identified 16,434 unique records to be screened, and of these 15,574 papers were excluded as they did not fulfil the inclusion criteria. The full-text copies of the remaining 860 papers, plus an additional 6 papers identified through hand-searching, were obtained and the inclusion criteria were applied independently by at least 2 reviewers. At this stage, 728 papers were excluded as they did not fulfil the inclusion criteria. A further 52 studies were excluded once reviewers began to extract data. The reasons for exclusions during data extraction are shown in Appendix 4. A further 30 papers reported on ongoing studies that were yet to report their results.
FIGURE 1 Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram. a, Only applies to updated search where studies may have been identified in the earlier search. b, Some studies provided data that contributed to ≥ 1 review. (Figure not reproduced here; see the original article.)
Four papers62,63,64,65 reporting barriers and facilitators data were identified in the update searches, but as these papers did not contribute any new data to the review they were excluded but retained for inclusion in a future update. Overall, 52 papers (44 studies) were included in at least 1 of the 3 reviews (Reviews 1–3). The overlap between the reviews is shown in Figure 2 and a list of the studies included in each review is shown in Table 5.
| Review data | Included studies (N) | References |
|---|---|---|
| Individual data | ||
| Effects only | 21 | Austin 2019,66 Bienek 2019,67 Cunningham 2016,68 Garland 2014,69 Gersch 2021,70 Hudak 2020,71 Huffman 2017,72 Jacobs 2016,73 Krumova 2013,74 Kurita 2018,75 Laigaard 2020,76 Montgomery 2020,77 Murphy 2013,78 Panicker 2022,79 Rivich 2018,80 Seal 2020,81 Sharp 2018,82 Thakral 201883 (linked papers: Von Korff 2016,84 2017,85 Sherman 2018,86 Thakral 201887), Townsend 2008,88 Twillman 2018,89 Zhou 201790 |
| AEs only | 2 | Caldera 2020,91 Lalanne 201692 |
| Barriers and facilitators only | 15 | Benintendi 2021,93 Firemark 2021,94 Kuntz 2021,95 Frank 2016,96 Giannitrapani 201897 (linked paper: Giannitrapani 201898), Henry 2019,99 Henry 2019100 (linked paper: Henry 2017101), Kennedy 2018,102 Langford 2020,103 Magee 2021,104 Matthias 2017,105 McNeilage 2022,106 Quinlan 2021,107 White 2020,108 Wu 2019109 |
| Combined data | ||
| Effects and AEs | 5 | Capano 2020,110 Jackson 2021,111 Zheng 2008,112 Sullivan 2017113 (linked paper: NCT 2013114),115 Zheng 2019116 (linked paper: Xue 2012115) |
| Effects and barriers and facilitators | 1 | Westanmo 2015117 |
| Inequalities | 42 | Austin 2019,66 Benintendi 2021,93 Bienek 2019,67 Capano 2020,110 Cunningham 2016,68 Firemark 2021,94 Frank 2016,96 Garland 2014,69 Gersch 2021,70 Giannitrapani 2018,97,98 Henry 2019,99 Henry 2019,100,101 Hudak 2020,71 Huffman 2017,72 Jackson 2021,111 Jacobs 2016,73 Kennedy 2018,102 Krumova 2013,74 Kuntz 2021,95 Kurita 2018,75 Laigaard 2020,76 Langford 2020,103 Magee 2021,104 Matthias 2017,105 McNeilage 2022,106 Montgomery 2020,77 Murphy 2013,78 Panicker 2022,79 Quinlan 2021,107 Rivich 2018,80 Seal 2020,81 Sharp 2018,82 Sullivan 2017,113 Thakral 201883,87 (linked papers: Von Korff 2016,84 2017,85 Sherman 201886), Townsend 2008,88 Twillman 2018,89 Westanmo 2015,117 White 2020,108 Wu 2019,109 Zheng 2008,112 Zheng 2019,116 Zhou 201790 |
FIGURE 2 Overlap of studies between reviews. B&F, barriers and facilitators. (Figure not reproduced here; see the original article.)