Survey Description
Survey years: Surveys were conducted in 1993, 1996, 1999, 2002, 2005, 2008, 2011, 2014/2015, 2017 and 2020. The 2017 and 2020 surveys were conducted with a survey sample half as large as in previous cycles. No further Member Health Surveys are planned after the 2020 cycle.
Survey administration: Prior to 2014, a print questionnaire was mailed up to three times to enhance survey response rate; starting in 2014, we only sent two mailings. In some survey years, the final questionnaire was a slightly shorter version of the questionnaires. Beginning in 2005, individuals were also able to complete the survey online at the Kaiser Permanente Division of Research’s (DOR) secure website. (Note: the DOR website is totally separate from the Health Plan’s kp.org website so that survey data cannot be directly accessed by anyone outside the DOR).
See Methodology section for more detail.
Survey content:
-
Sociodemographic characteristics
-
Health status and health conditions
-
Health-related behaviors, psychosocial risks, and social/economic risk factors
-
Functional health, falls, quality of life, use of community-based services, and usual type of transportation (members aged 65 and older)
-
Use of selected over-the-counter (OTC) and prescription medicines
-
Use of dietary supplements and herbal medicines (starting in 1999)
-
Use of complementary and alternative medicine (CAM) (starting in 1999)
-
Use of selected preventive services
-
Access to and use of digital communication tools (computer, mobile phone, Internet, email)
-
Preferred methods for receiving health information and health education
Survey sample: Independent stratified random samples of adult members are selected for each survey cycle, which means there are very few members who have data from more than one survey cycle. For the 1993 through 2011 survey samples, we selected age-sex stratified random samples of 2100-2400 men and women aged 20 and over from each of our 19 largest medical center service populations. To be eligible for the survey, individuals needed to be English speakers, members for at least 6 months as of the time of the survey, and not have a history of dementia. In 2008-2015, the total number of men and women in the starting survey sample was approximately 42,000, with final respondent samples of over 16,000 adults. In 2017, the starting sample was 22,000, but due to experimental oversampling of young adults and racial/ethnic minorities, the final respondent sample was only 5000. In 2020, the starting sample was approximately 26,000 adults, yielding 7,947 respondents.
Response rate: Information about past response rates can be found in the profile reports for individual survey years.
Respondent weighting factors: Similar to other population health surveys, survey respondents are assigned weighting factors. These are used to make the survey results better reflect the age-sex composition of the medical center service population and the age-sex-geographic composition of the KPNC region. When we changed to using a regional rather than medical center-based survey sample, we created post-stratification weighting factors based on the age-sex-racial/ethnic composition of the adult membership. Regional survey weighting factors can also be created for datasets that combine two or more survey years.
