The end-of-study patient survey: methods influencing response rate in the AVID Trial

Patient surveys are commonly distributed at the end of a multicenter clinical trial. This Antiarrhythmics Versus Implantable Defibrillators (AVID) substudy prospectively explored the relationship between methods used in distributing a survey and the quantity of responses received. AVID was a multicenter, randomized trial comparing survival in arrhythmia patients treated with antiarrhythmic drugs versus implantable defibrillators. At study termination, a patient satisfaction survey was mailed to the 664 surviving participants. Questions included reasons for study participation, study benefits and problems and quality of care. Survey mailings were stratified by four factors in a 2x2x2x2 factorial design: delivery mode (overnight vs. regular mail), certificate of appreciation, timing of administration ("early" vs. "late") and cover letter signed by a physician versus coordinator. Patients were randomly assigned to received one of 16 combinations of these four factors. Clinical characteristics and response rates were evaluated. Patients were more likely to return surveys delivered by overnight mail (75% vs. 68%, p=0.04), with no certificate of appreciation enclosed (75% vs. 68%, p=0.05) and administered close to the time of study closeout (79% vs. 72%, p=0.085). Compared to the 184 nonrespondents, the 456 (71%) respondents were older, Caucasian, lived with others, were high school graduates and less likely to have Medicare/Medicaid or HMO insurance (p<0.03). Physician recommendation was the most common reason cited for trial participation. Main benefits included increased knowledge of their medical condition and improved health. Reported problems included parking, transportation and excess clinic wait time. This randomized study demonstrated that methods of patient survey distribution affect the survey return rate. Additional studies should explore mechanisms for maximizing return rates.