| Study |
Differences at baseline |
Risk of bias judgement |
| Campbell 2008 |
The numbers of participants malnourished at baseline differed between groups. |
Unclear risk |
| Cano‐Torres 2017 |
Difference in haemoglobin levels between groups, but unlikely to affect outcomes |
Low risk |
| Forster 2012 |
Intervention group 1 (food group) had significantly higher alcohol intake than micronutrient or control groups. |
Unclear risk |
| Gu 2015 |
Gender imbalance (more males than females), but unlikely to affect outcomes |
Low risk |
| Imes 1988 |
Participants in the no advice group were younger and in better clinical condition than those in the group receiving dietary advice. |
Unclear risk |
| Salva 2011 |
Intervention group were frailer and more of these participants were malnourished or at risk of malnourishment.This trial also has potential sources of bias related to being cluster‐randomised |
High risk |
| Stow 2015 |
Control group (food‐based) were heavier, had higher energy, protein and fluid intake as well as higher VAS score.This trial also has potential sources of bias related to being cluster‐randomised. |
High risk |
