A Randomized Experiment to Increase Response Rates to a Health Care Survey Among Individuals with a High Predicted Probability of Preferring Spanish

• Activity Number: 560
• Type: Topic Contributed
• Date/Time: Wednesday, August 1, 2012 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Marc Elliott*+
• Companies: RAND Corporation
• Keywords: Indirect Estimation ; Survey Mode ; Randomized Experiment ; Spanish Language

Abstract

Spanish-preferring seniors are an underserved group, immunized at 1/2 the rate of English-preferring non-Hispanic White counterparts. Also, response rates that are ~ 1/2 of overall response rates in national surveys of patient experience limit our ability to assess needs and improve care. We describe a prediction algorithm and randomized experiment addressing these issues. CAHPS surveys of Medicare beneficiaries are conducted by mail, with phone followup. Mail surveys are only in English, with a rarely-used option to request a Spanish language instrument, so responses in Spanish are few and mainly via phone followup. We used CMS administrative data, including address and surname, to predict the probability of Spanish-phone response in a previous year with 98% concordance. In the subsequent year, 1/2 of beneficiaries with the highest predicted probability of preferring Spanish were randomized to a bilingual mailing. This intervention substantially increased mail and total response rates in the target population, with a majority of mail response in Spanish in the intervention condition, so that both representation and data quality may have increased.