Table 3. Predictors of Overestimating Expected Benefit
PredictorBreast Cancer ScreeningBowel Cancer ScreeningHip Fracture PreventionCVD Prevention
% (95% CI)ORa (95% CI)% (95% CI)ORa (95% CI)% (95% CI)ORa (95% CI)% (95% CI)ORa (95% CI)
Age, y
 50-5486 (77-92)0.6 (0.3-1.0)89 (81-95)0.6 (0.3-1.0)80 (71-88)1.2 (0.7-2.2)70 (59-79)0.7 (0.4-1.3)
 55-5990 (82-95)0.6 (0.3-1.1)95 (88-98)0.5 (0.3-0.8)83 (73-90)1.0 (0.6-1.8)62 (51-72)0.5 (0.3-1.0)
 60-6498 (93-100)0.9 (0.5-1.6)98 (93-100)0.7 (0.4-1.3)87 (79-92)1.3 (0.7-2.2)75 (66-83)0.9 (0.5-1.7)
 65-7095 (88-98)1.0100 (96-100)1.082 (72-89)1.074 (63-83)1.0
P value .12 .05 .73 .13
Sex
 Male92 (88-95)1.094 (90-97)1.083 (78-88)1.067 (60-73)1.0
 Female93 (89-97)1.0 (0.7-1.6)97 (93-99)1.5 (1.0-2.2)84 (77-89)1.0 (0.7-1.6)73 (66-80)1.3 (0.9-2.0)
P valueb .81 .058 .84 .18
Educationc
 Lower97 (93-99)1.098 (94-100)1.086 (79-92)1.078 (69-85)1.0
 Higher91 (87-94)0.6 (0.4-1.0)94 (91-97)0.5 (0.3-0.8)82 (77-87)0.8 (0.5-1.1)66 (60-72)0.6 (0.4-0.9)
P valued .03 .002 .19 .026
Past experiencee
 No96 (93-98)1.0 (0.6-1.4)84 (79-98)0.9 (0.5-1.7)69 (64-74)1.0 (0.6-1.8)
 Yes94 (89-97)1.080 (66-90)1.076 (61-86)1.0
P valued .85 .82 .98
  • CVD = cardiovascular disease; OR = odds ratio.

  • a OR calculated from multivariable logistic regression with the proportional odds assumption, across all response levels excluding unknown. OR >1.0 indicates higher estimates of effectiveness.

  • b df = 3.

  • c df = 1.

  • d Lower = primary or secondary school only. Higher = post-school qualification.

  • e Prior bowel cancer screening, diagnosis of osteoporosis or current medication for osteoporosis, and history of vascular disease. Breast cancer screening—only 5 women had never been screened so not included in the mode.