The 2024 Commonwealth Fund International Health Policy Survey of Older Adults was conducted from February 29 to June 20, 2024. The survey was administered to a nationally representative sample of adults age 65 and older in Australia, Canada, France, Germany, the Netherlands, New Zealand, Sweden, Switzerland, the United Kingdom, and the United States. The Commonwealth Fund contracted with SSRS, a U.S.-based survey research firm, to field the survey in the U.S. and six additional countries, as well as collaborate with fieldwork partners and oversee survey administration in the other three countries. A total of 16,737 interviews of adults age 65 and older were completed for the 2024 survey. Final country sample sizes ranged from 300 to 3,989. Interviews were completed via landline telephone, mobile telephone, or online administration. For this analysis, U.S. respondents were limited to those with some form of Medicare coverage. Respondents who reported they were dually eligible for Medicare and Medicaid were considered having “Medicare Advantage” if they reported they had a Medicare Advantage plan or “traditional Medicare” if they reported they did not have a Medicare Advantage plan. The total sample of U.S. respondents for this analysis was 1,882 (traditional Medicare = 845 and Medicare Advantage = 1,037).
In Australia, Canada, France, Germany, the Netherlands, and New Zealand, a random-digit dial (RDD) overlapping-frame telephone design was used to obtain all interviews. A large portion of the interviews in both the U.K. and the U.S. were also obtained using an overlapping-frame telephone design. The sample design in both the U.K. and the U.S. also included interviews via Verian’s Public Voice panel and the SSRS Opinion Panel, respectively. In the U.S., the SSRS Opinion Panel sample was used to target subgroups of analytical interest to the Fund, namely low-income, Black, Hispanic, and rural respondents. Sweden and Switzerland both used population-based registries to draw their sample. Additional country-specific information on the sampling frames is below:
- For Australia and New Zealand, SSRS procured landline and cell phone RDD samples from its sampling partner, Sample Solutions. For Australia, the landline RDD frame was based on the phone number blocks used in the telephone numbering plan provided by the Australian Communications and Media Authority, and the landline RDD sample was stratified by Australia’s eight regions to ensure geographic representativeness. The selection of cell phone RDD sample in Australia used roughly the same approach as the landline sample, though geographic information is not available for this frame. The shares of each cell phone service provider for the entire market were balanced to ensure that all providers had proper representation when selecting the cell phone RDD sample. For New Zealand, the landline RDD sample was based on the numbering plan provided by Telecom of New Zealand and was stratified by New Zealand’s 16 regions plus the Chatham Islands, while the cell phone RDD sampling was essentially the same as in Australia.
- RDD landline and cell phone sample for Canada was provided by Dynata, a premier global provider of sampling solutions. The landline sample was drawn from Dynata’s database by geography after extensive cleaning and validation. The cell phone sample was drawn from the most recent monthly Telcordia TPM (Terminating Point Master) Data file, sorted by province, carrier name, and 1,000-block to provide a stratification that would yield a representative sample, both geographically and by large and small carriers.
- Sample Solutions provided the landline and cell phone RDD samples for France, the Netherlands, and the United Kingdom. The landline RDD frame for France was generated using the national numbering plan provided by L’Autorité de Régulation des Communications Électroniques et des Postes, an independent French agency in charge of regulating telecommunications in France. The landline RDD frame for the Netherlands was generated using the national numbering plan provided by the Ministry of Economic Affairs. The landline RDD frame for the U.K. was generated based on the phone number blocks used in the national telephone numbering plan, provided by the Office of Communications (OFCOM), London, the British Federal Network Agency, using precodes by region. Based on the numbering plan for each country, Sample Solutions stratified the landline RDD samples by official NUTS2 regions according to the population distribution in each country. For the cell phone RDD samples, the phone numbers were randomly generated similar to the landline RDD sample for each country. Since it is not possible to identify precodes by region on cell phones in France, the Netherlands, or the U.K., Sample Solutions identified providers used for residential services and excluded those used for commercial sample. Online interviews in the U.K. were completed via Verian’s Public Voice Panel, a probabilistic panel recruited via address-based online surveying and face-to-face interviews. Both recruitment protocols use probability sampling drawn to ensure the entire population of the U.K. is represented.
- The sample for Germany was sourced from the ADM sampling system (Arbeitsgemeinschaft ADM-Telefonstichproben). The ADM master sample is based on the range of numbers available in the German telephone network as updated, monitored, and published by the Federal Network Agency (the government agency in charge of the German telephone network). Since about 99 percent of the population can be reached via at least one telephone number, the ADM system provides near-full coverage of the German population. The sample frame for Sweden utilized the Total Population Registry (RTB). The RTB includes more than 2.1 million adults age 65 and older and covers 99 percent of the Swedish population. To create the sample frame, personal identification numbers were matched with addresses so that invitations to partake in the survey could be sent to the respondents selected from the sample. Four variables were used to stratify the sample frame into a total of 36 strata: degree of urbanization (three groups), Swedish/foreign background (two groups), level of education (three groups), and age (two groups). In general, proportional allocation was used, with the exception of one stratum which oversampled individuals over 80 years of age with postsecondary education, who were born outside of Sweden, and reside in sparsely populated areas. In Switzerland, an individual sample of people age 65 and older was drawn by the Swiss Federal Statistical Office (SFSO), using Switzerland’s nationwide population registry. This registry covers nearly 100 percent of the Swiss population and is updated on a quarterly basis. The sample was stratified by the three linguistic regions: German-, French-, and Italian-speaking. The cantons of Zürich, Schaffhausen, Valais, and Basel Stadt were oversampled and extracted separately as their own strata, for a total of seven strata.
- Three different sample frames were used for data collection in the United States: 1) landline RDD, 2) cell phone RDD, and 3) the SSRS Opinion Panel to maximize the number of interviews among subgroups of analytical interest. The landline and cell phone RDD frames were generated by Marketing Systems Group (MSG), with the cell phone RDD sample being prepared using the Advanced Cellular Frame (ACF). Both the landline and cell phone RDD samples were disproportionately stratified, based on flagging records on both RDD frames with appended data. The landline RDD sample was matched against Neustar’s Pure Consumer Premium Database to identify phone numbers that are more likely to be assigned to households with residents who are age 65 and older. The cell phone RDD sample contained an ACF flag that identified phone numbers that are more likely to belong to individuals age 65 and older. The strata containing phone numbers that were flagged across both of the RDD samples were oversampled. To reach sufficient sample sizes among subgroups of analytical interest — including Black and Hispanic adults — as well as to target adults ages 65 to 74 and males age 65 and older, online interviews in the U.S. were completed via the SSRS Opinion Panel. SSRS Opinion Panelists are recruited randomly based on a nationally representative ABS (address-based sample) probability design (including Hawaii and Alaska), yielding a nationally representative panel of U.S. adults age 18 and older.
A common questionnaire was developed, translated, adapted, and adjusted for country-specific wording as needed. Interviewers in each country were trained to conduct interviews using a standardized protocol. Response rates varied from 3 percent in the United States to 50 percent in Switzerland.
International partners joined with the Commonwealth Fund to sponsor surveys, and some countries supported the use of expanded samples to enable within-country analyses. Data were weighted to ensure that the final outcome was representative of the population of adults age 65 and older in each country. Weighting procedures considered sample design, probability of selection, and systematic nonresponse across known geographic and demographic parameters including region, sex, age, education, and other characteristics relevant to the population of each country. In the U.S., the variables used for calibration also included race and ethnicity.
The margin of sampling error for the 2024 Commonwealth Fund International Health Policy Survey of Older Adults ranged from +/– 2.3 percentage points for Canada’s sample to +/– 7.1 percentage points for France’s sample, all at the 95 percent confidence interval.