Effectiveness of diet quality indices in measuring a change in diet quality over time: a systematic review and meta-analysis of randomized controlled trials.

Abstract

Diet quality indices (DQIs) were developed to score and rank adherence to dietary patterns in observational studies, but their use to measure changes in diet quality in intervention trials is becoming common in the literature. This systematic review and meta-analysis aimed to assess the effectiveness of DQIs to measure change in diet quality in intervention trials. MEDLINE, CINAHL, Embase, and the Cochrane Central Register of Controlled Trials databases were searched from January 1994 to June 2020. Two reviewers independently completed full-text screening. Eligible studies were randomized controlled trials that used validated a priori DQIs to measure change in diet quality in adults. Data were extracted by an independent reviewer and reviewed by the research team. Risk of bias was assessed by the Cochrane Collaboration's Risk of Bias 2.0 tool. The 34 included studies (52% of reviewed studies, 0.6% of initially identified studies) used 10 different DQIs, 7 of which were able to measure significant change in diet quality. Meta-analyses of pooled results demonstrated change in the Healthy Eating Index (MD 5.35; 95%CI, 2.74-7.97; P < 0.001) and the Mediterranean Dietary Adherence Screener (MD 1.61; 95%CI, 1.00-2.23; P < 0.001) scores. DQIs were more likely to measure change in diet quality if they reflected the diet pattern being implemented, if the intervention was significantly different from the baseline and control diets, and if the study was adequately powered to detect change. DQIs are responsive to change in diet quality in intervention trials when the index used reflects the dietary changes made and the study is adequately powered. The appropriate selection of a DQI to suitably match dietary changes and study populations is important for future dietary intervention trials. PROSPERO registration no. CRD42020181357.