OBJECTIVE—To examine whether self rated health confounds or modifies the relation between a prudent food intake pattern and mortality and to study whether the prudent food intake pattern predicts subsequent changes in self rated health. DESIGN—A prospective cohort study with follow up of total mortality and changes in self rated health. Food intake patterns were identified by principal component analysis from a 28 item food frequency questionnaire, collected at baseline. SETTING—MONICA surveys, Copenhagen County, Denmark. PARTICIPANTS—A random sample of 3698 men and 3618 women aged 30-70 years were followed up from 1982 to 1998 (median 15years). MAIN RESULTS—Among participants with complete information on all variables 18% had rated their health as poor (average or bad) at the baseline examination. Poor self rated health was related to a low score on the prudent food intake pattern, which was characterised by a frequent intake of wholemeal bread, fruit and vegetables. Three hundred and seventy six men and 210 women died during follow up. Poor self rated health and a low prudent food score were associated with increased mortality in both men and women. Self rated health did not modify the relation between diet and mortality. Of the 1098 men and 1048 women with good self rated health at baseline, 243 men and 297 women reported poor health during follow up. Low prudent food score, smoking, and high BMI increased the risk of developing poor health in both men and women, but in multivariate analysis the associations attenuated and were only significant for BMI. CONCLUSION—Both prudent food intake pattern and self reported health are independent predictors of mortality. Self rated health does not seem to modify the relation between diet and mortality.
