TY - JOUR
T1 - Educational interventions to improve screening mammography interpretation
T2 - A randomized controlled trial
AU - Geller, Berta M.
AU - Bogart, Andy
AU - Carney, Patricia A.
AU - Sickles, Edward A.
AU - Smith, Robert
AU - Monsees, Barbara
AU - Bassett, Lawrence W.
AU - Buist, Diana M.
AU - Kerlikowske, Karla
AU - Onega, Tracy
AU - Yankaskas, Bonnie C.
AU - Haneuse, Sebastien
AU - Hill, Deirdre
AU - Wallis, Matthew G.
AU - Miglioretti, Diana
PY - 2014/6
Y1 - 2014/6
N2 - OBJECTIVE. The objective of our study was to conduct a randomized controlled trial of educational interventions that were created to improve performance of screening mammography interpretation. MATERIALS AND METHODS. We randomly assigned physicians who interpret mammography to one of three groups: self-paced DVD, live expert-led educational seminar, or control. The DVD and seminar interventions used mammography cases of varying difficulty and provided associated teaching points. Interpretive performance was compared using a pretest-posttest design. Sensitivity, specificity, and positive predictive value (PPV) were calculated relative to two outcomes: cancer status and consensus of three experts about recall. The performance measures for each group were compared using logistic regression adjusting for pretest performance. RESULTS. One hundred two radiologists completed all aspects of the trial. After adjustment for preintervention performance, the odds of improved sensitivity for correctly identifying a lesion relative to expert recall were 1.34 times higher for DVD participants than for control subjects (95% CI, 1.00-1.81; p = 0.050). The odds of an improved PPV for correctly identifying a lesion relative to both expert recall (odds ratio [OR] = 1.94; 95% CI, 1.24-3.05; p = 0.004) and cancer status (OR = 1.81; 95% CI, 1.01-3.23; p = 0.045) were significantly improved for DVD participants compared with control subjects, with no significant change in specificity. For the seminar group, specificity was significantly lower than the control group (OR relative to expert recall = 0.80; 95% CI, 0.64-1.00; p = 0.048; OR relative to cancer status = 0.79; 95% CI, 0.65-0.95; p = 0.015). CONCLUSION. In this randomized controlled trial, the DVD educational intervention resulted in a significant improvement in screening mammography interpretive performance on a test set, which could translate into improved interpretative performance in clinical practice.
AB - OBJECTIVE. The objective of our study was to conduct a randomized controlled trial of educational interventions that were created to improve performance of screening mammography interpretation. MATERIALS AND METHODS. We randomly assigned physicians who interpret mammography to one of three groups: self-paced DVD, live expert-led educational seminar, or control. The DVD and seminar interventions used mammography cases of varying difficulty and provided associated teaching points. Interpretive performance was compared using a pretest-posttest design. Sensitivity, specificity, and positive predictive value (PPV) were calculated relative to two outcomes: cancer status and consensus of three experts about recall. The performance measures for each group were compared using logistic regression adjusting for pretest performance. RESULTS. One hundred two radiologists completed all aspects of the trial. After adjustment for preintervention performance, the odds of improved sensitivity for correctly identifying a lesion relative to expert recall were 1.34 times higher for DVD participants than for control subjects (95% CI, 1.00-1.81; p = 0.050). The odds of an improved PPV for correctly identifying a lesion relative to both expert recall (odds ratio [OR] = 1.94; 95% CI, 1.24-3.05; p = 0.004) and cancer status (OR = 1.81; 95% CI, 1.01-3.23; p = 0.045) were significantly improved for DVD participants compared with control subjects, with no significant change in specificity. For the seminar group, specificity was significantly lower than the control group (OR relative to expert recall = 0.80; 95% CI, 0.64-1.00; p = 0.048; OR relative to cancer status = 0.79; 95% CI, 0.65-0.95; p = 0.015). CONCLUSION. In this randomized controlled trial, the DVD educational intervention resulted in a significant improvement in screening mammography interpretive performance on a test set, which could translate into improved interpretative performance in clinical practice.
KW - Interpretive performance
KW - Intervention
KW - Mammography
KW - Screening
UR - http://www.scopus.com/inward/record.url?scp=84904128840&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84904128840&partnerID=8YFLogxK
U2 - 10.2214/AJR.13.11147
DO - 10.2214/AJR.13.11147
M3 - Article
C2 - 24848854
AN - SCOPUS:84904128840
SN - 0361-803X
VL - 202
SP - W586-W596
JO - American Journal of Roentgenology
JF - American Journal of Roentgenology
IS - 6
ER -