To advance and promote the discipline and professional practice of epidemiology in Ontario public health units
Please click here to visit our new website











 

 

 

10 The Canadian Community Health Survey (CCHS)

The Canadian Community Health Survey (CCHS)

The Canadian Community Health Survey (CCHS)

Original source: Statistics Canada
Distributed by:
1. Ontario Ministry of Health and Long Term Care (MOHLTC)
2. Statistics Canada
Suggested citation (see Data Citation Notes):
1. Canadian Community Health Survey [year], Statistics Canada, Share File, Ontario MOHLTC
2. Canadian Community Health Survey [year], Statistics Canada, Public Use Microdata File, Statistics Canada

Data Notes

  • The CCHS is a national survey designed to provide health information at the regional and provincial levels. The CCHS is the data source for many of the Health Indicators generated by Statistics Canada and the Canadian Institute for Health Information.
  • Prior to 2005, the CCHS consisted of two separate cross-sectional surveys conducted over a two-year, repeating cycle:
    • In the first year of the two-year collection cycle, a general health survey was administered that provides data at the health region level
    • In the second year, a topic specific survey was administered that provides data at the provincial level.  The names of each of the cycles and the data contained in them are as follows:
      • Cycle 1.1 – Provides general health data collected in 2000/01 at the health region level
      • Cycle 1.2 – Provides provincial level data collected in 2002 that focuses on mental health.
      • Cycle 2.1 – Provides general health data collected in 2003 at the health region level
      • Cycle 2.2 – Provides provincial level data collected in 2004 that focuses on nutrition.
  • In 2005, data collection methodology changed. Only one survey was collected over one year, and was released about 6 months after collection. So data release occurred every other year.
      • Cycle 3.1 – Provides general health data collected in 2005 at the health region level.
      • Cycle 4.1 (now referred to as CCHS 2007, see below)  – Provides general health data collected in 2007 at the health region level.
  • In 2007, data collection methodology changed again.  Data collection that provides data at a health region level now occurs on an ongoing basis.  Statistics Canada releases a file that contains annual data covering a period from January to December. In addition, a file that combines two years of data is also released. To prevent confusion with previous cycles of the CCHS, surveys are no longer designated by x.1 and x.2. Regional data files are now simply named by the year of collection (e.g., “Canadian Community Health Survey, 2008”). Topic specific data files are named by the focus of the data file (e.g., “Canadian Community Health Survey on Healthy Aging”).
  • With the establishment of the CCHS, the National Population Health Survey (NPHS) stopped including a cross-sectional component but focused instead on providing longitudinal data. This data consists of approximately 17,000 people nationally followed since 1994/95 for up to 20 years.
  • The target population of the CCHS includes household residents in all provinces and territories, with the principal exclusion of populations on Indian Reserves, Canadian Forces Bases, and some remote areas. There is one randomly selected respondent per household, with an oversampling of youths resulting in a second member of certain households being interviewed. For CCHS 2000/02, only those 12 years of age and over were eligible for selection. The CCHS 2002 on mental health included respondents aged 15 years and over.  The CCHS 2004 Nutrition survey included respondents aged 6 and over (Not all data are available for every age group).
  • The CCHS sample is primarily a selection of dwellings drawn from the Labour Force Survey area sampling frame. For the regional level survey, the sample is supplemented with a random digit dialing (RDD) sample in some health regions.
  • Data collection for the CCHS is done by either computer assisted personal or telephone interviewing for the area sample or telephone interviewing for the RDD sample. The interview for the health region-level survey takes approximately 45 minutes: 30 minutes of common content to be asked in all health units, 10 minutes of optional content determined by each health region from a predefined list of questionnaire modules, and 5 minutes of socio-economic and demographic content.  In Ontario, the Ministry of Health and Long-Term Care (MOHLTC) co-ordinates a process to co-operatively select a single set of optional modules for all the health regions in the province.
  • There are three types of CCHS files: master files, sharing files, and Public Use Microdata Files (PUMF). 
  • CCHS master files can be accessed at Statistics Canada’s regional offices or regional data centres.
  • The CCHS sharing file is a subset of the CCHS master file.  The CCHS sharing file consists of all the respondents who agreed to share their health card number and data with the provincial ministries and health regions. The Health Analytics Branch, MOHLTC distributes the Ontario portion of the sharing file to public health units (PHUs) that agree to the privacy conditions for access
  • Statistics Canada makes available a portion of the CCHS public use microdata file (PUMF), which is designed to preserve statistical analytic capabilities while ensuring respondent confidentiality. Thus, the PUMF contains fewer variables and less detailed response categories than the master file. 
  • Bootstrapping:
    • Bootstrapping is more precise than using the approximate sampling variability (C.V.) tables because it estimates the variance by generating a random sample with replacement 500 times from within the CCHS sample and estimating the variance from these 500 estimates.
    • Statistics Canada provides bootstrap weights and SPSS or SAS programs to calculate the bootstrap variable for the CCHS sharing file but this is not provided in the Public Use Microdata File (PUMF).
    • Bootstrap weights can be used for quantitative variables and complex analyses, such as regression.

 

Release Guidelines

Users of the CCHS Sharing File - Ontario sample, must adhere to the corresponding guidelines when publishing or otherwise releasing any data derived from the microdata files supplied by the MOHLTC. 

 

Calibration: Issues with Household-Level Analysis

As noted in the CCHS_Household_Weights.pdf, the last step in creating the CCHS household weight is calibration.  The calibration step is performed to ensure that, in each household size group, the sum of the final weights approximates the number of households at the provincial level. It controls the distribution of the survey weights at the provincial level so they reflect the proportion of the households in each of the different household size groups.

Because calibration is done only at the provincial level, when we use household variables at the health region level, there is a potential problem if the variable is related to household size. Health region estimates calculated using the household weights may result in over- or under-representation of households in the health regions. As a result, an estimate using a calibrated household weight may not accurately represent the underlying population of interest - even for proportions.

In most cases, the weighted distribution of households by size should be close to the actual distribution of households by size, but unfortunately without knowing the actual counts of the number of households of various sizes in each health region, it is not possible to know how close the estimates are to reality.  Where the variable of interest is related to household size, users should be aware of the implications for interpretation of the results.  See more detailed notes in the Food Insecurity Core Indicator. If the variable of interest does not relate to household size, it is appropriate to do analysis of proportions at the PHU level.  As for all CCHS variables, it is not recommended to report counts at the PHU level.

 

References and Resources

  1. Beland Y. Canadian Community Health Survey – Methodological Overview. Health Reports 2002;13(3):9-14.
  2. Statistics Canada. Population Health Surveys http://www.statcan.ca/english/concepts/hs/
General reference:Statistics Canada Canadian Community Health Survey, available at http://www.statcan.gc.ca/cgi-bin/imdb/p2SV.pl?Function=getSurvey&SDDS=3226⟨=en&db=imdb&adm=8&dis=2Note the links on the left of the webpage for Summary of changes over time and for Other reference periods.

Changes Made

DateType of Review (Formal Review or Ad Hoc?)Changes made byChanges
April 2011Formal Healthy Eating and Active Living sub-group
  • Data Notes updated
  • Release Guidelines updated
  • References and Resources updated
Acknowledgements
Lead AuthorsCarma Lynn Wylie, Niagara Region Public Health, Elizabeth Rael, Ministry of Health Promotion and Sport
Contributing Author(s)Jennifer Skinner, Ahalya Mahendra, Elsa Ho
Reviewers 
Page last updated:  April 15, 2011
Treasurer/Secretary | Admin | Members Login

BrickHost