Improving survey response rates among patients at a cancer center during a global pandemic

ABSTRACT

This study evaluated approaches for recruiting patients at a cancer center to an electronic survey about their experience with the COVID-19 pandemic. Eight groups (n = 2,750 patients total) were contacted via email to complete a 15-minute survey. Groups consisted of 1) a detailed email only, 2) a brief email only, 3) a mailed letter and brief email, 4) a mailed letter, small prize in the envelope, and brief email, 5) a mailed postcard and brief email, 6) a brief email and $10 gift card, 7) a mailed letter, brief email and $10 gift card, and 8) a mailed postcard, brief email, and $10 gift card. Patients were considered eligible if they had a valid email address, were seen at the cancer center since January 1, 2015, had English as a preferred language, a last known vital status of alive, an address inside the cancer center's catchment area, and were between 40 and 89 years old. Patients were over sampled for Hispanic ethnicity and African American race, then randomly sampled. Response rates were evaluated overall, by race/ethnicity, gender, time since visit to the cancer center and age. Multiple logistic regression was used to assess the odds of completing the survey. 259 patients (9.4%) completed the survey across all pilot groups. Response rates varied by pilot group, ranging from 2.6% response for a detailed email only, to 18.4% response for a brief email, postcard and gift card. The latter group was also among the highest for response rates among Hispanics (25.6%) and Non-Hispanic/Non-White patients (15.9%). In a multivariate model adjusting for race, ethnicity, age, and gender, we found those who received a gift card had 1.86 times (95% CI 1.40-2.48) higher odds of completing the survey than those who did not. Additionally, those who received a postcard or letter compared to those who only received an email had 1.46 times (95% CI 1.05-2.04) higher odds of completing the survey. In our study of cancer patients seen at a major cancer center, prompting potential study participants with a letter or postcard before an email improved response rates. Further including a gift card increased response rates, particularly for underrepresented minorities. Future analyses include evaluating response rates by insurance status and cancer type.