Improving Response Rates and Representation of Spanish-Preferring Patients in Surveys Assessing Provider Quality: Results from a Randomized Experiment of National Disenrollment Survey

• Activity Number: 358
• Type: Contributed
• Date/Time: Tuesday, August 11, 2015 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Marc Elliott* and Ann Haas and Megan Beckett and Cheryl Damberg
• Companies: RAND Corporation and RAND Corporation and RAND Corporation and RAND Corporation
• Keywords: Disparities ; Quality Measurement

Abstract

Response rates (RRs) for Spanish-preferring (SP) patients are often low in population surveys. We examined the effectiveness of additional language and mode response options in improving the response rate and representation of such patients in a national survey of disenrollees from Medicare plans (MA and PDP types). High-probability SP disenrollees were randomized to (1) Standard survey protocol (2) English-language mailings; reference (n=4,546)) (2) Bilingual (English/Spanish) mailings (n=2001); or (3) Bilingual mailings + phone follow-up (n=2001). RRs in the control group for SP beneficiaries (25.6% MA, 15.5% PDP) were much lower than for others (45.8% MA, 43.4% PDP). The bilingual-mailing arm increased RRs (+3.9% PDP ns, +4.7% MA), and bilingual-mailing + phone more than doubled gains (+11.5% PDP, +10.4% MA) (p=0.03 additional effect of phone within PDPs, p< 0.001 throughout except where noted). The experimental arms particularly increased the representation of lower-income and less-healthy members of the population relative to the standard arm (p< 0.05). Alternate arms thus improved RRs and representativeness, more fully capturing this patient population's experiences.