Study population

Between October 2018 and May 2019, a total of 411 men aged 50–69 years attended for screening. Prior to the screening, 403/408 (98.7%) completed each domain of the pre-test EBQ. All men completed the PSA screening test and the post-test questionnaire (PBQ) and one participant did not complete the post-test MRI questionnaire (407/408, 99.7%). The mean duration in minutes for each screening test was 1.3 for PSA (SD 0.59), 19.7 for MRI (SD 3.83) and 8.4 for ultrasound (SD 3.17).

Outcomes

The perceived overall burden for MRI and PSA were compared. In total, 30.8% perceived MRI to be worse than PSA, 18.7% perceived MRI to be better than PSA, and 50.4% perceived MRI and PSA to be the same (Table 1); PSA was perceived to have a lower burden than MRI (p = 0.0007). In terms of the component scores, 108 (26.5%) had higher anxiety scores for MRI than PSA, 43 (10.6%) had a score for MRI lower than PSA, and 256 (62.9%) had equal scores (p < 0.001). For the burden component, 75 (18.5%) perceived MRI to be more burdensome than PSA, 15 (3.7%) perceived PSA to be more burdensome than MRI, and 316 (77.8%) perceived them to be similar (p = 0.001). The embarrassment component showed 21 (5.2%) perceived MRI as more embarrassing than PSA, 7 (1.7%) perceived PSA as more embarrassing than MRI, and 379 (93.1%) perceived both as similar embarrassment. The pain component was the only score which was higher for PSA, with 9 (2.2%) perceiving MRI to be more painful, 89 (21.9%) PSA to be more painful, and 75.9% to be similar (p < 0.0001) (Fig. 1).

Table 1 Output of non-parametric two-sample paired Sign test comparing PBQ scores for MRI vs. PSA.Fig. 1: Comparison of PBQ Scores for MRI vs PSA.

Diverging bar chart depicting the difference between participantʼs anxiety, burden, embarrassment and pain for PSA and MRI.

Mean scores and Wilcoxon comparison EBQ

(Table 2) The overall mean EBQ score for MRI and PSA was 1.45 (SD 0.65) and 1.46 (SD 0.54), respectively. Comparison between individual domain scores showed that a higher proportion having MRI expected it to have more anxiety (1.75 vs 1.63, p = 0.02), more burdensome (1.47 vs 1.33, p < 0.001) and more embarrassing. The expected pain score for PSA was higher than MRI (1.67 vs. 1.26, p < 0.001).

Table 2 Output of Wilcoxon signed-rank test to compare mean EBQ & PBQ scores. PBQ

The participants perceived the overall burden of MRI to be higher than PSA (1.21 vs. 1.16, p = 0.003) (Table 2). A comparison of the component scores showed that the difference in overall score was due to a higher degree of anxiety (1.46 vs. 1.26, p < 0.001), burden (1.29 vs. 1.11, p < 0.001) and embarrassment (1.08 vs. 1.03, p = 0.03). The pain score remained higher for PSA compared to MRI (1.25 vs. 1.04, p < 0.001). All component scores were lower in the PBQ compared to the EBQ.

MRI comparison of expected and perceived scores

The participants’ transition between EBQ (pre-test) and PBQ (post-test) for MRI are shown in Sankey Diagrams (Fig. S1). There was a consistent trend for men to have lower pre-test scores across all component scores. For anxiety, 16% expected some/rather/extreme anxiety-reducing to 7% (Spearman’s rank correlation 0.421, p < 0.001); no burden increased from 65 to 78% (Spearman’s rank correlation 0.227, p < 0.001); for embarrassment, 6% expected it to have some/rather/extreme embarrassment but reducing to 2%; and for pain 5 to 1% (Spearman’s rank correlation coefficients of 0.141 and 0.103 for embarrassment and pain, p < 0.001).

PSA comparison of expected and perceived scores

There was a similar trend for PSA, with post-test (PBQ) scores improving compared to pre-test scores (Fig. S2). For anxiety, the proportion of participants who reported no anxiety increased from 54 to 82% (Spearman’s rank correlation 0.376, p < 0.001); no burden increased from 74 to 92%; no embarrassment was 98% after the PSA test (Spearman’s rank correlation coefficients of 0.325 and 0.334 for burden and pain, p < 0.001); 56% who expected the phlebotomy procedure to be associated with slight/some/rather/extreme pain reduced to 23% (Spearman’s rank correlation 0.334, p < 0.001).

Preference for screening examination

Of 408 participants, 194 (47.5%) had no preference, 106 (26.0%) preferred MRI and 79 (19.4%) preferred PSA. This indicates that prior to screening, participants preferred MRI compared to PSA (+6.6%, 95% CI 4.4–8.4, p = 0.02). The proportion of participants who had no preference was higher than any category (p < 0.001) (Fig. 2). After undergoing all screening tests, 164 (40.2%) preferred MRI, 156 (38.2%) had no preference and 78 (19.1%) preferred PSA. The proportion of participants who preferred MRI compared to PSA was +21.1% (95% CI 14.9–27.1, p < 0.001).

Fig. 2: Overall preference of participants for the screening test.

Bar chart illustrating that participants indicated an overall preference for MRI after completing all screening tests.

Determinants of the overall burdenMRI

In the multivariable regression analysis, the presence of pre-test anxiety (odds ratio 2.16, p < 0.001) and Afro-Caribbean ethnicity (odds ratio 0.521, p = 0.048) were identified as significant determinants of the overall burden of MRI (Table S3).

PSA

Multivariable analysis demonstrated that the presence of pre-test anxiety (OR 1.12, p < 0.001) and pre-test expected burden (OR 1.14, p < 0.001) were identified as significant determinants of overall burden (Table S4).

Acceptability of ultrasoundExpected and perceived scores

For ultrasound, the majority of post-test domains were lower than the pre-test scores indicating participants found the ultrasound test more burdensome than expected (Fig. S3).

Comparison with other tests

A comparison of the post-test (PBQ) scores of the three screening tests showed a larger proportion of participants found ultrasound anxiety-inducing, burdensome, embarrassing and painful compared to both MRI and PSA (Fig. 3).

Fig. 3: Distribution of PBQ Scores for Ultrasound, MRI and PSA.

Bar chart depicting participantʼs perception of anxiety, burden, embarrassment and pain after having a Ultrasound, MRI and PSA.

Determinants of ultrasound burden

Multivariable regression analysis demonstrated Afro-Caribbean ethnicity (OR 2.10, p = 0.043) and pre-test anxiety (OR 2.16, p < 0.001) as significant determinants of the overall burden of ultrasound (Table S5).

Willingness to repeat ultrasound

In total, 34.6% (95% CI 17.1–24.9) either strongly disagreed, disagreed or were neutral with a recommendation for a repeat ultrasound in comparison to 22.9% (95%CI 19.1–27.2) and 24.6% (95% CI 20.7–29.0) for PSA and MRI, respectively (Fig. S7).