PURPOSE     To measure US and Canadian (CDN) consumer preferences for colorectal cancer (CRC) screening using conjoint analysis (CA), and compare with primary care physicians' perceptions of those preferences.

   METHODS    CRC screening reduces mortality but uptake remains poor in the US and Canada.  Nine attributes of CRC screening (process, pain, frequency, preparation, follow-up, specificity, sensitivity, complication rate, and cost) were identified based on focus groups and results from a previous CA survey in Canada.  The fractional factorial CA survey design used to maximize D-efficiency included three blocks with 11 choice sets and one repeat scenario.  A follow-up question asked if respondents would prefer no CRC screening at all.    Survey invitations were emailed to a random sample of US and CDN consumers (aged 45-70 years) and US MDs who were online panel members.  CDN MDs were posted an invitation to complete the survey online.    ß-coefficients from bivariate probit regression estimated the marginal utilities of the attributes including the no screening option and were scaled from 0 to 10. 

   RESULTS    Response rates ranged from 7.5% (CDN MDs) to 47.9% (CDN consumers).

   Order of attribute importance (rescaled ß-coefficient).

 

US consumers n=1087	CDN consumers n=501	US MDs n=500	CDN MDs n=500
sensitivity (8.19)	sensitivity (7.12)	no screening (6.48)	no screening  (8.65)
process (6.90)	process (6.61)	sensitivity (5.35)	sensitivity (7.32)
pain (3.61)	specificity (3.33)	process (4.46)	specificity (3.05)
risk (3.60)	pain (3.32)	frequency (2.07)	frequency (2.89)
specificity (3.15)	risk (3.28)	specificity (2.03)	process (2.56)
no screening (2.09)	preparation (1.63)	risk (1.86)	risk (1.77)
preparation(1.98)	frequency (1.49)	pain(1.63)	preparation (1.44)
frequency (1.81)	no screening (1.19)	preparation(0.42)	pain (1.23)
follow-up (0.29)	follow-up (0.16)	follow-up (0.16)	follow-up (0.29)

   The top two attributes were the same for US and CDN consumers (sensitivity and process) and US and CDN MDs (no screening and sensitivity).  MDs overestimated consumer preferences for no screening and underestimated preferences for test process (e.g., stool sample vs. having an endoscopy).  Both consumers and MDs assigned low preference to the need for confirmatory follow-up tests. 

   CONCLUSIONS    These results indicate how preferences of consumers differ from perceptions of providers about consumer preferences.  This can be used to inform how CRC screening programs are presented to consumers to help increase screening uptake.