Definitions

Depression is characterized by a depressed mood and/or lack of interest in most things, along with other symptoms, all lasting at least two weeks.  These symptoms include appetite or sleep disturbance, decreased energy, difficulty concentrating, feelings of worthlessness, and/or suicidal thoughts.   The National Population Health Survey (NPHS) measures depression with a subset of questions, administered by lay interviewers, from the Composite International Diagnostic Interview.^37,38  These questions cover a cluster of symptoms for a depressive disorder, which are listed in the Diagnostic and Statistical Manual of Mental Disorders (DSM-III‑R).³⁹  For this article, the presence of depression refers to the 12 months before the date of the survey interview. The NPHS questionnaire is available at http://www.statcan.gc.ca/cgi-bin/imdb/p2SV.pl?Function=getSurvey&SDDS=3225&lang=en&db=IMDB&dbg=f&adm=8&dis=2. Overall scores were totaled, and the results were transformed into a probability estimate of a diagnosis of depression.  For this article, if the estimate was 0.9 or more (that is, a 90% likelihood of a positive diagnosis of depression, corresponding to a value of 5 or more), the respondent was considered to have experienced depression in the previous 12 months.  For the algorithm and questions used to measure depression, see Appendix B.  

At cycle 6 only, respondents were asked if they had ever been diagnosed with depression by a health professional, and their age at diagnosis.  This information was combined and cross-referenced against the respondent's age at each baseline cycle.  Respondents who had been diagnosed with depression before their age at baseline were considered to have a history of depression.  Data on history of depression were missing for 14.4% of records.  A missing category was included to retain as many observations as possible for the multivariate analyses.  Additional analyses were done using the probability estimate of a diagnosis of depression from earlier cycles of the NPHS (data not shown).  Because the association between marital dissolution and depression was virtually unchanged, this information was not included in the models.

Marital status was categorized as:  remained married or experienced marital dissolution.  Respondents were considered to have remained married if they reported their marital status as “married,” “common-law” or “living with a partner” at baseline and again two years later.  Respondents were considered to have experienced marital dissolution if they reported their marital status as “married,” “common-law” or “living with a partner” at baseline, and two years later reported their marital status as “separated,” “divorced” or “single.”  Marital status definitions were not provided to respondents.

Total household income from all sources in the previous 12 months was adjusted for the number of people in the household and for the low-income cutoff (LICO) specific to the household and community size.  Adjusted household incomes were then grouped into deciles (10 groupings each containing one-tenth of Canadians).  A two-decile (one-quintile) change in the ranking between two consecutive NPHS cycles was defined as a change in adjusted household income.  Because of missing values, the change in adjusted household income could not be calculated for 10.2% of records.  A missing category was included to retain as many observations as possible for the multivariate analyses.

Number of children in household was based on the number of children aged 15 or younger in the household at baseline and the number aged 17 or younger in the household at follow-up two years later.  Because many older adolescents leave home for postsecondary education, the ages of youth included in this calculation were restricted.  Only households in which children were reported to reside at baseline were considered to have them.  Households were defined as those from which children had departed if the number of children decreased between consecutive NPHS cycles.  An additional variable for no children in household at baseline was included to retain individuals without children in the analyses.

Four questions measured social support across all six NPHS cycles, using an abridged version of measures in the Medical Outcomes Study (MOS).⁴⁰  Respondents were asked if they had someone to confide in, to give them advice, to count on in a crisis, and to make them feel loved and cared for or to show them love and affection.  In 1994/1995 and 1996/1997, the possible responses to these questions were “yes” or “no.”  In the remaining cycles, responses were structured on a five-point scale:  “all of the time,” “most of the time,” “some of the time,” “a little of the time,” and “none of the time.”  Respondents who answered “no” (in 1994/1995 and 1996/1997) and “none of the time” or “a little of the time” (in subsequent cycles) to at least one of the four questions were considered to have low emotional support in that cycle.  Respondents were grouped into four categories depending on their level of support and any change between cycles:  support increased, support remained high, support decreased, and support remained low.

Work status was assigned one of four possible values:  working at baseline and follow-up; working at baseline, but not at follow-up; not working at baseline, but working at follow-up; and not working at baseline or follow-up.  Respondents who reported having a job last week and those who reported currently working were classified as working; those who did not have a job or who were permanently unable to work were considered not to have been working.  

Respondents were grouped into three education categories based on the highest level attained at baseline:  secondary graduation or less, some postsecondary and postsecondary graduation.

Age at baseline was used as a continuous variable and ranged in value from 20 to 64.