The study took place in Nithsdale, south west Scotland. It focused on schizophrenic patients living in accommodation provided by the Dumfries and Galloway Mental Health Association. Their position in the community had been assessed by social services as sufficiently precarious for them to need additional support. The residents, however, are encouraged to be responsible for their own domestic chores, including shopping and cooking. Each patient was matched with a normal control for sex, age, smoking status (smoker v non-smoker), and employment status—variables that affect a person’s diet. All patients were unemployed.
Patients and controls were interviewed by a psychiatrist. The current average weekly food intake was obtained through a modified version of an established food frequency questionnaire.2
Also recorded were patients’ and controls’ height and weight. Patients’ mental state was examined using the positive and negative syndrome scale for schizophrenia. A blood sample was taken to measure serum concentrations of cholesterol and vitamin E.3
We studied 30 patients (17 men; mean age 44 (SD 15, range 20-79) years). Twenty three patients smoked. More patients (20) than controls (11) were overweight or obese, as assessed by body mass index (weight (kg)/(height(m)2
)); McNemar’s test, χ2
=4.27; P=0.04). The patients consumed significantly less energy, total fibre, retinol, carotene, vitamin C, vitamin E, and alcohol (table). In all, 83% of the patients consumed less fibre, 71% of the male and 69% of the female patients consumed less vitamin E, and 70% of the patients and 73% of the controls consumed more energy from saturated fats than the suggested UK estimated average requirements (the amounts that any stated group of people will, on average, need).4
The patients, when compared with the controls, consumed fewer fruit portions (median weekly intake 2.3 (range 0-20) v
7.0 (range 0-33); Wilcoxon matched pairs signed rank test, median difference 3.5 (95% confidence interval 0.5 to 7.5); P=0.03) and vegetable portions (10.0 (1-23) v
19.0 (4-34); 8.5 (4.0 to 12.0); P=0.001).
Fewer patients than controls (8 v 18; McNemar’s test, χ2=6.7; P=0.01) had a ratio of serum vitamin E concentration to cholesterol concentration of over 5 (said to be necessary to protect against cardiovascular disease).
Where dietary measurements in the patients differed significantly from those in the controls, correlations between these measurements and scores in the positive, negative, and total symptom scales were measured. In female patients, a positive correlation was found between positive symptoms and alcohol intake (rho=0.75, P=0.006).