The data

View the most recent version.

The longitudinal National Population Health Survey (NPHS) followed 17,276 individuals who were aged 12 or older in 1994/1995. Interviews were conducted every two years. The NPHS design, sampling procedures, data collection, and response rates by cycles are described in detail elsewhere.¹² Data for this study are from cycles 1 to 7 (1994/1995 to 2006/2007) and pertain to respondents aged 50 or older at the first cycle (N=5,404) who initially reported no chronic condition, but who, in a subsequent cycle, reported a new diagnosis of one of five chronic conditions. The mean age of the sample was 65.5 (SD = 10.3), and slightly more than half (54.1%) were women. The primary focus is a prospective comparison of pre- and post-diagnosis (two years) health behaviour among people reporting a new diagnosis during the 12 years covered by the study. Cycles 2 and 3 were used to examine health behaviour changes among a healthy comparison group who did not have any of the five health conditions.

At each cycle, respondents were asked, "Do you have any of the following long-term (lasting or expecting to last at least six months) conditions that have been diagnosed by a health care professional?"  The five conditions selected for this analysis were "heart disease," "cancer" (skin cancer excluded), "effects of stroke," "chronic bronchitis or emphysema" (asthma excluded), and "diabetes."

The smoking measures used in this study were self-reported smoking status (never, former or current) and number of cigarettes per day.

Physical activity was measured by the reported frequency and duration of participation in 20 leisure-time activities (for example, walking for exercise, yoga) in the previous three months. Participation in any activity at least three times per week was considered "physically active." A continuous measure of energy expenditure (kcal/kg/day) was derived from the frequency and duration of activities.¹³

Respondents' alcohol consumption was defined as never or low (never drinks to less than 1 drink per week), moderate (2 or fewer drinks per day and 1 to 14 drinks per week for men; 2 or fewer drinks per day and 1 to 9 drinks per week for women), and excessive (more than 2 drinks per day or more than 14 drinks per week for men; more than 2 drinks per day or more than 9 drinks per week for women).¹⁴

Responses to six questions about the frequency of fruit and vegetable consumption were used to determine what percentage attained minimum recommended daily intake. The recommendation for servings per day is based on an average portion size of 80g, which, studies suggest, approximates the amount consumed on each occasion.¹⁵ The current target is seven servings of fruit and vegetables per day for Canadians older than age 50,¹⁶ but fewer than 12% of the sample reported consuming this amount. Therefore, respondents were classified as meeting the recommended minimum if they reported consuming fruit and vegetables five or more times per day. This is consistent with Canada's Food Guide recommendations before 1992,¹⁷ as well as the minimum recommended by the World Health Organization (WHO) and the Food and Agriculture Organization (FAO),¹⁸ and in several other countries.^17-20 Diet information was available only for cycles 5, 6 and 7 of the NPHS.

Analyses were weighted and adjusted for the complex sampling design using SAS 9.2 or SUDAAN 10.0 with balance repeated replication. Rao-Scott chi-square²¹ obtained from SAS PROC SURVEYFREQ using discordant cells (0-1 versus 1-0 responses) was used to test for significant change between pre- and post-diagnosis.²² (The McNemar test, which is typically used for this comparison, is not available in SAS PROC SURVEYFREQ or SUDAAN.) Paired t-tests were used to compare pre- and post-diagnosis health behaviours with continuous measures. Logistic regression models controlling for pre-diagnosis levels of health behaviours were used to compare sex and age differences in health behaviour pre- and post-diagnosis. A significant coefficient for one of these socio-demographic variables indicates greater change from pre- to post-diagnosis in one of the groups.²³

This study has a number of limitations. Chronic conditions and health behaviours were based on self-reports. While it is unlikely that respondents would report having been diagnosed with a major condition if it did not exist,^24-28 underreporting is possible.²⁹ Self-reported smoking status is generally considered accurate. The NPHS leisure-time physical activity measure was more extensive than those available in most large health surveys, but was, by definition, limited to leisure time. Other physical activity, for example, at work, for transport and for rehabilitation, was excluded. Fruit and vegetable consumption was based on the reported frequency of consumption, and may not correspond precisely to the number of servings. NPHS data were collected every two years, so behaviour changes between cycles may have been missed; for example, smoking cessation attempts and remission. Moreover, changes observed over a two-year period may not be maintained over the long-term.