Strategies for Surveying Families of Medicaid-Insured Children by Telephone

Brief communication 17 STRATEGIES FOR SURVEYING FAMILIES OF MEDICAID-INSURED CHILDREN BY TELEPHONE The increasing use of managed care for Medicaid populations has raised concerns that quality of care will suffer due to cost containment efforts. Collecting data directly from members is an important component of monitoring quality of care.1 Medicaid populations are notoriously difficult to survey, with response rates of 50 percent considered acceptable.2,3 In research or quality monitoring, budget limitations may preclude in-person interviews. To our knowledge, scant information has been published about effective strategies for surveying Medicaid members by telephone or mail. This study's aim was to describe the strategies used for locating families and completing telephone and mailed surveys with parents of Medicaidinsured children in a study of the quality of pediatric asthma care. Method We conducted a cross-sectional review of a stratified random sample of telephone interview completion records from the Asthma Care Quality Assessment (ACQA) Study. The ACQA Study is an ongoing study of Medicaid-insured children with asthma insured by five health plans (designated HMOs A, B, C, D, and E) and served by seven primary care networks in California, Massachusetts, and Washington. HMOs A, B, and C were groupmodel health maintenance organizations in which a single large provider group contracted exclusively with the health insurance plan. In contrast, HMOs D and E were mixed-model Medicaid managed care organizations in which the insurer contracted with a variety of provider groups, including community health centers. The proportion of members insured by Medicaid was 5 percent or less for HMOs A, B, and C, compared with 89 and 100 percent for HMOs D and E. For the current study, we identified a random sample of 50 completed interviews from each of the seven primary care networks in the study. This resulted in 150 interviews from HMO D and 50 interviews from the other health plans. For each interview sampled, we reviewed the interview completion record and tallied the accuracy of the information originally provided by the health plan and the source of the telephone number that led to the completed Received February 2,2001; revised January 15, 2002; accepted February 1,2002. Journal of Health Care for the Poor and Underserved · Vol. 14, No. 1 · 2003 DOI: 10.1177/1049208902238820 18 Surveying by Telephone interview. We analyzed the proportion of completed interviews contributed by each strategy, stratified by health plan. We calculated the cost per completed interview by multiplying the number of interviewer hours by the hourly salary plus benefits and dividing by the number of completed interviews. This cost did not include telephone charges or time spent for coding, editing, mailing, or data entry. The overall completion rate for the ACQA Study was 63 percent and is further detailed in a related article.4 In the ACQA Study, Medicaid-insured children aged 2 to 16 years with asthma-related health care use were identified using computerized health plan files. The child's and parent or guardian's names, address, and telephone numbers were drawn from health plan databases . A missing address was cause for exclusion; missing or incorrect telephone numbers were not. The research team mailed each parent or guardian a contact letter on health plan letterhead and offered a $10 incentive for participation in the interview. The telephone interview averaged 27 minutes and was conducted in English or Spanish by interviewers at the Kaiser Permanente Northern California Division of Research from February through October 1999. In August and September 1999, an identical version of the survey was mailed to families who could not be reached by telephone. In addition to the contact information supplied by the health plan, interviewers used multiple sources of information. These included telephone messages left by parents responding to the initial mailing; directory information; a digital database; Internet searches; reverse directories; updated information from the health plan; and referrals from relatives, employers, and others. The research team also requested telephone numbers from large community health clinics and from a local government agency that handled Medicaid eligibility . Contact information was not sought from individual primary care providers for varying reasons, including resource limitations. Interviewers were instructed to make up to 16 attempts per...