Advertisement

DOES CLINICAL USE OF AMYLOID-PET AFFECT PHYSICIANS’ BELIEFS ON THE PATHOGENETIC ROLE OF AMYLOID-β AND THE CLINICAL USAGE OF AMYLOID-PET?

      Background

      Previous data suggest that beliefs on the pathogenic role of amyloid-beta in AD do not affect the intended clinical use of amyloid-PET (Boccardi et al., 2016). Here, we evaluate whether practice in clinical use of amyloid-PET, in turn, affects physicians’ beliefs on the role of amyloid-B (AB) in AD and the intended clinical usage of amyloid-PET.

      Methods

      We administered a questionnaire at the beginning, in the middle and at the end of a study using amyloid-PET. Physicians indicated their belief regarding the role of AB in the AD pathogenesis ('belief'), and then expressed the probability of diagnostic change after amyloid-PET in three hypothetical scenarios: ‘AD-negative', ‘NonAD-positive' and ‘NonAD-negative'. Differences in the probability of diagnostic change between scenarios and over time were evaluated through a linear mixed model with 'belief' as covariate. Effect Size indices (ES; Cohen’s d for repeated measures) were also computed. We analyzed the seven dementia experts who both filled out the questionnaire at all time points and enrolled a minimum of 15 patients.

      Results

      ‘Belief' did not affect the probability of diagnostic change in different scenarios and over time (p=0.655). The probability of diagnostic change did not change significantly throughout the study for all scenarios (p=0.970). Due to the very small sample size, we considered ES greater than 0.20 (‘small'). These were consistent with: i) a trend in increase in the belief on a causal role of AB in AD and ii) a growth of the positive predictive value of the amyloid-PET in ‘NonAD-positive' over the whole study (Table).

      Conclusions

      Beliefs on the pathogenic role of amyloid-B do not affect the intended clinical use of amyloid-PET in the course of a naturalistic study using amyloid-PET. ES suggested an increase of the belief in a causal role of AB in AD pathogenesis over the study. Moreover, an increased use of the positive predictive value of amyloid-PET emerged after practical usage of amyloid-PET. These results will be compared with the actual use of amyloid-PET by the same physicians.
      TableChanges in the diagnostic confidence between scenarios and over the time.
      VariablesT00T06T12T06-T00T12-T06T12-T00
      mean (SD)mean (SD)mean (SD)mean_diff (SD_diff)Cohen's dmean_diff (SD_diff)Cohen's dmean_diff (SD_diff)Cohen's d
      Befief (0-10)5.86 (1.47)6.29 (1.59)7.00 (2.02)0.43 (1.35)0.320.71 (1.60)0.451.14 (1.87)0.61
      Scenario 1 (0-100%)54.29 (31.55)66.21 (31.85)55.50 (30.30)11.93 (31.57)0.38-10.71 (26.84)-0.401.21 (38.72)0.03
      Scenario 2 (0-100%)67.14 (25.17)55.00 (30.57)71.07 (29.96)-12.14 (26.10)-0.4616.07 (32.74)0.493.93 (11.45)0.34
      Scenario 3 (0-100%)23.33 (26.71)20.48 (23.34)18.09 (18.06)-2.86 (33.66)-0.08-2.38 (25.19)-0.09-5.24 (24.06)-0.22
      Values for the three scenarios refer to percentages. Cohen's d is an index of effect size: defined as the difference between two means divided by a standard deviation for the data. Mean_diff is the difference between two time points, SD_diff is calculated with the following formula: SD_pre*sqrt(2*(1-rho)), where rho is the correlation coefficient. Values are in italics or bold when the Cohen's d is at least ‘small’ (≥ 0.20): italics for positive difference and bold for negative.