The purpose of the Canadian Community Health Survey - Healthy Aging is to collect new information about the factors, influences and processes that contribute to healthy aging. The survey focuses on the health of Canadians aged 45 and over.
Data release – May 12, 2010 (first in a series of releases. Please refer to left sidebar under the heading "The Daily")
The Canadian Community Health Survey - Healthy Aging is part of the Canadian Community Health Survey program, but is a unique survey in that the target population, objectives, and many of the questions differ from those of earlier surveys administered under the umbrella of CCHS. The survey collects new information about the factors, influences and processes that contribute to healthy aging through a multidisciplinary approach focusing on health, social and economic determinants. The survey focuses on the health of Canadians aged 45 and over by examining the various factors that impact healthy aging, such as general health and well-being, physical activity, use of health care services, social participation, as well as work and retirement transitions.
Policy makers, researchers and planners are aware that planning for the future of an aging population is an important activity now and in the coming years. This survey will help them make informed decisions regarding health care, social services and income support programs that will affect all Canadians.
The specific objectives of the survey are:
1. To better understand the aging process of people aged 45 and over by collecting data on various aspects of their health and well being, use of health care services, social support and participation and work and retirement transitions.
2. To examine how lifestyle determinants affect health as people age.
3. To examine the links among healthy aging and social, demographic, geographic and economic variables or characteristics using a multidisciplinary approach.
4. To provide information on successful aging by age group and sex.
Data users will include Health Canada, the Public Health Agency of Canada, provincial ministries and departments of health, the Institut de la statistique du Québec, the Canadian Institute for Health Information, Human Resources and Social Development Canada, and academic researchers in a variety of disciplines such as gerontology and public health.
The Healthy Aging Survey covers the population 45 years of age and over living in the ten provinces. Excluded from the survey's coverage are: persons living on reserves and other Aboriginal settlements; full-time members of the Canadian Forces, persons living in collective dwellings and the institutionalized population. Altogether, these exclusions represent about 4% of the target population.
Health Statistics Division met with stakeholders from Health Canada, the Public Health Agency of Canada, the Provincial Health Ministries, Human Resources and Social Development Canada (HRSDC), and the Canadian Longitudinal Study on Aging (CLSA) to develop, refine and select questionnaire content.
Content was chosen using the following set of criteria:
-Issue(s) identified as data gaps from the stakeholder consultations;
-Strong evidence of aging as an important influence on health and quality of life, morbidity and premature mortality;
-Significant number of people affected by the targeted issue;
-Significant impact on family, community, and health care costs;
-Data has potential for health improvement with policy intervention;
-Issue(s) identified as priority for the support/development of programs and policy, surveillance requirements and/or research; and
-Issues(s) identified as relevant to the Healthy Aging Framework.
The questions were designed for computer-assisted interviewing (CAI), meaning that, as the questions were developed, the associated logical flow into and out of the questions was programmed. This included specifying the type of answer required, the minimum and maximum values, on-line edits associated with the question and what to do in case of item non-response.
In collaboration with Statistics Canada's Questionnaire Design Resource Centre, the questionnaire was subjected to three phases of qualitative testing which took place in March, May and June, 2007 and consisted of one-on-one interviews and focus groups. The objective was to evaluate respondent reactions to and understanding of the survey, willingness to respond to the questions and to obtain estimates of time for the various sections.
A Pilot Survey was conducted in November and early December of 2007 in Winnipeg, Toronto, Montreal and Halifax. The main objectives of the pilot test were to observe respondent reaction to the survey, to obtain estimates of time for the various content modules, to determine the effectiveness of the training of interviewers and the communication strategy, to test the training procedures and material, to provide a preliminary indication of the response rates, and to further test the computerized questionnaire, which was also extensively tested in-house to identify any error in the questionnaire flow and text.
The content of the Pilot Survey was split into two questionnaires (Pilot A and Pilot B), each with some common content and some independent content. The selection of the final content modules for the Main Survey was based on several factors, including:
-Analysis of results from the Pilot Survey including feedback received from our interviewers and field operation personnel;
-Constraints associated with the administration of the questionnaire;
-Input from experts and feedback received from all stakeholders. Health Canada (HC) and Public Health Agency of Canada (PHAC) provided specific input and priorities to help guide this process; and
-The fact that the Main Survey is designed as a cross-sectional survey.
This is a sample survey with a cross-sectional design.
Sample Size and Allocation
To meet the survey objectives of estimating healthy aging among Canadians aged 45 and over for specific domains of interest for each province, and given the budget allocated to the survey, a sample of 32,005 responding units was desired. A two-step strategy was used to allocate the sample to the provinces. First, 125 sample units were allocated to each domain of interest (10 age/sex groups) in each province. Thus, 1,250 units were assigned to each province in the first step for a total of 12,500. The remaining 19,505 units were allocated to the provinces using a power-allocation method using a power q=0.7 (Bankier M., 1988). The total sample size of any given province is found by adding the sizes obtained in the two steps.
Moreover and in order to have a good urban and rural representation in each province, the sample was subsequently allocated to two strata: urban and rural. The provincial sample was proportionally allocated to the urban and rural strata according to the number of dwellings having people aged 45 and over in each stratum. Then sample sizes were inflated before data collection to take into account out-of-scope and vacant dwellings and anticipated non-response.
Sampling of Dwellings from the 2006 Census
The CCHS - Healthy Aging uses as its sampling frame the 2006 Census. All dwellings within the 10 Canadian provinces containing at least one household member aged 43 and over (to become 45 and over in 2008) were included in the sampling population.
All dwellings within the same Census dissemination area block (CB), identified as either urban or rural, were grouped together. In each province, clusters of CBs were created having a fixed number of dwellings with a minimum number of people in the 75-84 and 85 or over age groups. Clusters were composed entirely of urban or rural CBs and could not cross provincial boundaries. Some remote and empty CBs were excluded from this process.
Each urban cluster will yield approximately 35 dwellings, and each rural cluster will yield approximately 20 dwellings. The number of each type of cluster to select in each province was determined by the urban/rural proportion of dwellings from the 2006 Census having a household member aged 85 or over.
For each province, the selection of clusters was done with probability proportional to size (number of dwellings in each cluster having a household member aged 45 or over) without replacement using the Hanurav-Vijayan algorithm.
Dwellings in each cluster were further stratified into three groups: first, dwellings having a household member aged 85 or over; second, dwellings with no household members aged 55 or over; and third, all other dwellings. The number of dwellings to select in each stratum is fixed for all provinces, with a slight adjustment made for Quebec and Ontario.
Sampling of Interviewees
Upon visiting a dwelling, all members of the household are listed and one person aged 45 years or over is automatically selected using various selection probability factors based on age.
There are five selection probability factors dependent on age: 45-54, 55-64, 65-74, 75-84 and 85 or over. The selection probabilities vary by province in order to achieve the targeted number of respondents in each age group.
Data collection for this reference period: 2008-12-01 – 2009-11-30
Responding to this survey is voluntary.
Data are collected directly from survey respondents.
Collection for the Canadian Community Health Survey - Healthy Aging took place between December 2008 and November 2009. Over the collection period, a total of 30,865 valid interviews were conducted using a type of computer-assisted interviewing (CAI) called computer assisted personal interviewing (CAPI).
Computer-assisted interviewing (CAI) offers two main advantages over other collection methods. First, CAI offers case management system and data transmission functionality. Second, CAI allows for custom interviews for every respondent based on their individual characteristics and survey responses.
Before the start of each collection period, introductory letters and brochures explaining the purpose of the survey were sent to the sampled households. These explained the importance of the survey and provided examples of how the CCHS - Healthy Aging data would be used.
Units selected from the area frame were interviewed by decentralized field interviewers using CAPI. CAPI interviewers were trained to make an initial personal contact with each sampled dwelling. Every effort was made to conduct the interviews face-to-face, and 94% of interviews were conducted exclusively in person. Collection by telephone was authorized only when a respondent requested an interview in the other official language but no bilingual interviewer was available in the area, or when the respondent spoke neither official language but another interviewer was available to translate for the respondent.
Interviewers were instructed to make all reasonable attempts to obtain interviews. Numerous repeat visits to the dwelling were made at different times on different days.
For individuals who at first refused to participate in the survey, a letter was sent from the nearest Statistics Canada Regional Office to the respondent, stressing the importance of the survey and the household's collaboration. This was followed by a second visit (or call) from a senior interviewer, a project supervisor or another interviewer to try to convince respondent of the importance of participating in the survey.
In cases where the selected respondent was, for reasons of physical or mental health, incapable of completing an interview, another knowledgeable member of the household supplied information about the selected respondent. While proxy interviewees are often able to provide accurate answers to most of the survey questions, the more sensitive or personal questions were beyond the scope of knowledge of a proxy respondent. This resulted in some questions from the proxy interview being unanswered. Every effort was taken to keep proxy interviews to a minimum.
The CCHS - Healthy Aging sample was divided into six non-overlapping two-month collection periods. Regional collection offices were instructed to use the first 4 weeks of each collection period to resolve the majority of the sample, with the next 4 weeks being used to finalise the remaining sample and to follow up on outstanding non-response cases. Cases initially coded to "No contact" or "Absent for duration of survey" were resent in two additional collection periods: 1) August 15 to November 30, 2009 and 2) October 15 to November 30, 2009.
Most editing of the data was performed at the time of the interview by the computer-assisted personal interviewing (CAPI) application. It was not possible for interviewers to enter out-of-range values and flow errors were controlled through programmed skip patterns. For example, the CAPI application ensured that questions that did not apply to the respondent were not asked.
In response to some types of inconsistent or unusual reporting, warning messages were invoked but no corrective action was taken at the time of the interview. Where appropriate, edits were instead developed to be performed after data collection at Head Office. Inconsistencies were usually corrected by setting one or both of the variables in question to "not stated".
Pre-coded answer categories were supplied for all suitable variables. Interviewers were trained to assign the respondent's answers to the appropriate category.
In the event that a respondent's answer could not be easily assigned to an existing category, several questions also allowed the interviewer to enter a long-answer text in the "Other-specify" category. Following collection, all such questions were reviewed in head office processing. For some of these questions, write-in responses were coded into one of the existing listed categories if the write-in information duplicated a listed category.
In order for estimates produced from survey data to be representative of the covered population and not just the sample itself, users must incorporate the survey weights into their calculations. A survey weight is given to each person included in the final sample, that is, the sample of persons having responded to the survey. This weight corresponds to the number of persons in the entire population that the respondent represents. Seven separate adjustments are part of the weighting strategy.
1. Initial weight
2. Removal of out-of-scope units
3. Household non-response
4. Creation of person level weight
5. Person non-response
The final CCHS -- Healthy Aging weight can be found on the data file with the variable name WTS_M.
In order to supply coefficients of variation that will be applicable to a wide variety of categorical estimates produced from a Public Use Microdata File (PUMF) and that can be readily accessed by the user, a set of Approximate Sampling Variability Tables will be produced with each PUMF. These "look-up" tables allow the user to obtain an approximate coefficient of variation based on the size of the estimate calculated from the survey data.
The coefficients of variation (CV) are derived using the variance formula for simple random sampling and by incorporating a factor (design effect) that reflects the multi stage, clustered nature of the sample design.
The design effects, sample sizes and population counts used to produce the Approximate Sampling Variability Tables as well as the tables themselves are presented in a document, which is included on the PUMF CD.
All coefficients of variation in the Approximate Sampling Variability Tables are approximate and, therefore, unofficial. The computation of exact coefficients of variation is not a straightforward task since no simple mathematical formula can account for all CCHS sampling frame and weighting aspects. Therefore, other techniques such as resampling methods must be used in order to estimate measures of precision. Among these methods, the bootstrap method is recommended for analysing CCHS data.
The computation of coefficients of variation (or any other measure of precision) using the bootstrap method requires access to information that is considered confidential and thus not available on the PUMF. This computation must be done using the Master file.
A macro program called Bootvar was developed to give users easy access to the bootstrap method. The Bootvar program is available in SAS and SPSS formats, and is made up of macros that calculate the variances of totals, ratios and differences between ratios, as well as linear and logistic regressions.
To ensure the survey met its objectives (see "Description"), the CCHS - Healthy Aging content was developed based on a multi-stage consultation process. A Steering Committee and an Advisory Board comprised of authorities from the provincial and territorial Ministries of Health, the Canadian Institute for Health Information, Health Canada and the Public Health Agency of Canada determined the concepts and focus. Experts and stakeholders advised on the selection of content and provided recommendations on appropriate and proven collection instruments and indices. The resulting data are recognized as valid measures of concepts such as nutritional risk, loneliness and cognition. The cognition module is the subject of a validation study published in Health Reports.
Throughout the collection process, control and monitoring measures were put in place and corrective action was taken to minimize non-sampling errors. These measures included response rate evaluation, reported and non-reported data evaluation, on site observation of interviews, improved collection tools for interviewers and others.
Once processing steps are completed, three data validation steps were undertaken. First, a validation program was run in order to compare estimates for the health indicators taken from the common content with the annual Canadian Community Health Survey. This validation was performed by age group, sex, and province. A similar comparison was done for content in common with the General Social Survey. Significant differences were examined further to find any anomalies in data.
The work of analysts who use the CCHS data to publish analytical articles allows for an in-depth look at many variables of the survey and represents a very effective way to find errors.
Last, an external validation step was also part of the validation process. Share files were sent before release to provincial and federal partners for a ten day examination period. They could then scrutinize the data and inform Statistics Canada of any concerns or anomalies related to data quality.
Statistics Canada is prohibited by law from releasing any information it collects which could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.
Users requiring access to information excluded from the microdata files may purchase custom tabulations, or access the master files through the Research Data Centres program or the Remote Access program. Outputs are vetted for confidentiality before being given to users.
Public use microdata files
The public use microdata files (PUMF) are developed from the master files using a technique that balances the need to ensure respondent confidentiality with the need to produce the most useful data possible at the provincial or regional level. The PUMF must meet stringent security and confidentiality standards required by the Statistics Act before they are released for public access. To ensure that these standards have been achieved, each PUMF goes through a formal review and approval process by an executive committee of Statistics Canada. Variables most likely to lead to identification of an individual are deleted from the data file or are collapsed to broader categories.
In total, 41,496 of the selected units in the CCHS -- Healthy Aging were in-scope for the survey. Out of these, 33,517 households agreed to participate in the survey, resulting in an overall household-level response rate of 80.8%. Among these responding households, 33,517 individuals (one per household) were selected to participate in the survey, out of which a response was obtained for 30,865 individuals, resulting in an overall person-level response rate of 92.1%. At the Canada level, this yields a combined (household and person) response rate of 74.4% for the CCHS -- Healthy Aging.
Since it is an unavoidable fact that estimates from a sample survey are subject to sampling error, sound statistical practice calls for researchers to provide users with some indication of the magnitude of this sampling error. The basis for measuring the potential size of sampling errors is the standard deviation of the estimates derived from survey results. However, due to the large variety of estimates that can be produced from a survey, the standard deviation of an estimate is usually expressed relative to the estimate to which it pertains. This resulting measure, known as the coefficient of variation (CV) of an estimate, is obtained by dividing the standard deviation of the estimate by the estimate itself and is expressed as a percentage of the estimate.
Statistics Canada commonly uses CV results when analyzing data and urges users producing estimates from the CCHS -- Healthy Aging data files to do so as well.
Over a large number of observations, randomly occurring errors will have little effect on estimates derived from the survey. However, errors occurring systematically will contribute to biases in the survey estimates. Considerable time and effort was devoted to reducing non-sampling errors in the CCHS -- Healthy Aging. Quality assurance measures were implemented at each step of data collection and processing to monitor the quality of the data.
A major source of non-sampling errors in surveys is the effect of non-response on survey results. The extent of non-response varies from partial non-response (failure to answer one or a few questions) to total non-response. Partial non-response to the CCHS -- Healthy Aging was minimal; once the questionnaire was started, it tended to be completed with very little non-response. Total non-response occurred either because a person refused to participate in the survey or because the interviewer was unable to contact the selected person. Total non-response was handled by adjusting the weight of persons who responded to the survey to compensate for those who did not respond.
Non-response to any particular question (item non-response) is generally low, but may be higher in some modules for various reasons. Users are cautioned to account for the larger variations during analysis, for example, by including a dummy variable for income non-response in regression analysis.