To determine the effectiveness of a workplace wellness programme intervention in improving participants’ behaviour towards choosing a healthy diet and the correlation with health indicators.

A retrospective cohort study.

Wellness programme in the Midwest, USA.

Employees (n 12 636) who participated in a wellness programme for three consecutive years during years 2004 to 2013 and who completed web-based health risk questionnaires. The wellness programme included annual health screening, laboratory measures, health risk questionnaire and personalized health-care programme. Participants’ food group intakes, BMI and health indicators were compared between the first and last year of participation. McNemar’s non-parametric test was used for paired nominal data. Pearson correlations were computed for paired food and health indicator measurements. Correlations between dietary intake and BMI, cholesterol and TAG were computed using Pearson correlations and McNemar’s test.

There were negative correlations between intakes of fruits, vegetables, grains, dairy, healthy eating pattern and health outcome indicators such as BMI and TAG levels. Additionally, the percentage of employees who increased their consumption of fruits (16·88 v. 12·08 %, P<0·001), vegetables (15·20 v. 11·44 %, P<0·001) and dark green leafy vegetables (12·03 v. 7·27 %, P 0·001) was significantly higher than the percentage of participants who decreased their intake of these food groups during the third-year follow-up.

The wellness programme improved some health indicator parameters and had a positive impact on increasing participants’ intakes of fruits, vegetables and whole grains at the third year of follow-up.