The feasibility of web surveys for obtaining patient-reported outcomes from cancer survivors: a randomized experiment comparing survey modes and brochure enclosures

Table 2 Response outcomes by survey mode and brochure enclosure, survey of Utah cancer survivors

	Sample	Responded¹		Adjusted Risk Ratio²	95% CI
	n	n	%	Adjusted Risk Ratio²	95% CI
Overall	446	209	46.9
Survey mode
Paper	226	114	50.4	1.00	Ref.
Web	220	95	43.2	0.88	0.72, 1.07
Brochure
No	225	103	45.8	1.00	Ref.
Yes	221	106	48.0	1.03	0.85, 1.25
Brochure by mode³
Paper, no brochure	115	58	50.4	1.00	Ref.
Paper, brochure	111	56	50.5	0.97	0.75, 1.25
Web, no brochure	110	45	40.9	1.00	Ref.
Web, brochure	110	50	45.5	1.08	0.81, 1.45

1. The percentage of cases that responded equates to response rate (AAPOR RR1)
2. Risk ratios from multivariable models predicting survey response that include design features and adjust for time since diagnosis, sex, age, race, ethnicity, geography, cancer site, and cancer stage at diagnosis. In these models, none of the demographic or cancer variables significantly predicted response
3. When conducting the model using a single reference group to compare all four mode-brochure combinations simultaneously, the adjusted risk ratios were as follows: Paper, no brochure: Ref.; Paper, brochure: RR = 0.98 (95% CI: 0.76, 1.27); Web, no brochure: RR = 0.83 (95% CI: 0.62, 1.10); Web, brochure: RR = 0.91 (95% CI: 0.70, 1.18)

ISSN: 1471-2288