Fifty papers reporting on 46 studies (2,282 patients on supplements) were included: 32 RCTs (1,590 patients on supplements) and 14 non-randomised studies (five surveys and nine non-comparative studies). RCTs scored variably on the Jadad scale and only four papers scored 5.
Compliance overall: Pooled mean compliance with oral nutritional supplementation was 78.2% (SD 15) across 52 studies (range 37% to 100%). In a meta-analysis overall compliance was 71.6% (SEM 5%; 10 studies, 357 patients). Energy intake was higher in patients who received supplements than in controls (by 375 kcal/day, SEM 72; eight RCTs).
Compliance according to setting: Mean compliance was significantly greater in community setting studies (80.9%; 33 studies) compared with hospital setting studies (67.2%; 10 studies) but when an analysis weighted by sample size was used this difference became non-significant. Compliance was not significantly different between randomised and non-randomised studies.
Compliance with prescription: The correlation between prescribed and consumed nutritional supplementation energy was high (r²=0.833, p=0.0001; 41 studies) but there was no association between amount of supplementation prescribed and percentage compliance with the prescription.
Oral nutritional supplementation related factors: Compliance was not affected by duration of the intervention. Participants complied significantly more with higher energy density supplements (r²=0.093, p=0.05; 41 studies). Studies that used a variety of flavours of supplement (18 studies) reported significantly higher compliance than those studies that offered variety of supplement types (six studies), (81% versus 63%, p=0.027). Giving instructions on when to take the supplements did not appear to alter compliance.
There were no significant differences in compliance between patient groups (disease groupings) or based on BMI but younger patients tended to be associated with lower compliance levels (r²=0.148, p=0.01; 44 studies).