The data

View the most recent version.

This article is based on data from the 24-hour dietary recall component of the 2004 CHMS. Respondents were asked to list all foods and beverages consumed during the 24 hours before the day of their interview (midnight to midnight). Interviewers used the Automated Multi-pass Method,^9, 10 with a five-step approach to help respondents remember what they had to eat and drink:

• a quick list (respondents reported all foods and beverages consumed in whatever order they wished);
• questions about specific food categories and frequently forgotten foods;
• questions about the time of consumption and type of meal (for example, lunch, dinner);
• questions seeking more detailed, precise descriptions of foods and beverages and quantities consumed;
• and a final review.

A subsample of the population responded to a second 24-hour recall a few days later to help assess day-to-day variations in food and beverage intake. The energy and nutrient content of the food and beverages was derived from Health Canada's Canadian Nutrient File 2001b, Supplement.¹¹

A total of 35,107 people completed the initial 24-hour dietary recall, and a subsample of 10,786 completed the second recall. Response rates were 76.5% and 72.8%, respectively. Thirty-eight invalid or "null" recalls were excluded.

The first 24-hour recall for 20,159 adults aged 19 or older was used to estimate average beverage consumption. Usual intakes of caffeine and alcohol were based on both recalls, excluding 244 pregnant or breastfeeding women, and were estimated using the NCI method⁸ in a two-part model: first, estimating the probability of consuming in a logistic regression model; second, estimating the amount consumed in a non-linear mix model. Both parts of the model were estimated simultaneously to account for the correlation between the probability of consuming and the amount consumed. As well, both parts of the caffeine model were adjusted by weekend/weekday and by smoking status (daily smoker, occasional smoker, non-smoker). Alcoholic beverages were listed in the 24-hour dietary recall of the CHMS, and in addition, a question in the general component of the survey asked about the frequency of alcohol consumption. Both parts of the alcohol model were adjusted by weekend/ weekday and by frequency of drinking alcohol in the past 12 months (never, less than once a month, once a month, 2 to 3 times a month, once a week, 2 to 3 times a week, 4 to 6 times a week, everyday).

The bootstrap method, which accounts for the complex survey design, was used to estimate standard errors, coefficients of variation, and confidence intervals.^12, 13 The significance level was set at p < 0.05.

The beverage categories are based in groupings created by Health Canada's Bureau of Nutritional Sciences. Recipes and basic foods have separate categories. The categories were revised to eliminate double-counting. Beverages used in food recipes belonging to a non-beverage food category (milk in a cake recipe, for example) are excluded from the beverage categories.

The water category refers to municipal, bottled, well and distilled water consumed as such. It excludes water required to prepare another beverage (for instance, water in coffee is included in the coffee category).

The milk category includes all milk regardless of fat content, evaporated milk and milk added to tea or coffee. It also includes goat milk and infant formulas. Condensed milk and milk added to ready-to-eat or hot cooked cereals (on average, 45 grams a day) are excluded.

The fruit juice category refers to 100% pure juice, and includes the juice portion of alcoholic beverages and juice recipes (concentrate and water), whereas the fruit drinks category comprises beverages that contain less than 100% fruit juice.

The tea and coffee category excludes added items such as cream and sugar.

Alcoholic beverages are split into three categories: alcoholic beer, wine and spirits and liquor. In the last category, only the alcohol ingredient in a drink is included.

The methods used to gather information about food and beverage consumption are generally associated with some under-reporting. The 24-hour dietary recall is not exempt from this problem, even when the Automated Multiple-pass Method, which maximizes respondent recall, is used. Another report¹⁴ has estimated calorie under-reporting at close to 10% for CHMS respondents aged 12 or older. Under-reporting of beverages strongly correlates with that of calories.

It is possible that some traces of a beverage category are found in another category in cases where it is impossible to separate certain ingredients in a mixture.