PAPER 31 Clinical Paper Session 4 General e Friday, September 8, 2017  10:48e10:53 AM Treatment; Prognosis/Outcomes; Patient Education A Prospective Randomized Study Analyzing the Effect of Pre-Operative Opioid Counseling on Post-Operative Opioid Consumption after Hand Surgery Level 2 Evidence Asif Ilyas, MD Todd Alter, BS COI: Royalty: Jaypee Medical Publishers (Ilyas) Consulting Fee: Globus (Ilyas) Speakers Bureau: DepuySynthes (Ilyas) Hypothesis: Prescription opioid abuse has become increasingly prevalent in the United States. Opioid counseling has been proposed to decrease opioid consumption after surgery. This study aimed to evaluate the effect of pre- operative opioid counseling on patients’ pain experience and opioid con- sumption in hand surgery, using a carpal tunnel release (CTR) surgery. A hypothesis was made that patients receiving pre-operative opioid coun- seling would use less of their prescribed opioid and terminate its use sooner as compared to patients who do not receive any counseling. Methods: A prospective randomized comparison of consecutive patients scheduled to undergo CTR surgery was conducted. Patients were ran- domized to either receiving formal pre-operative opioid counseling or no counseling. All operations were performed with the same mini-open CTR surgical technique and the same number of opioids were prescribed post- operatively. Daily opioid pill consumption, pain levels, and any adverse reactions were recorded. Pre-study power analysis indicated that a mini- mum of 20 patients were needed in each group, which was achieved. Results: On the day of surgery, patients in the group with counseling reported significantly fewer prescribed opioid pills consumed, 0.65 versus 1.90, compared to patients in the group without counseling (P < 0.05), while experiencing no significant different in pain level experience. The same was found on the first postoperative day, patients in the group with counseling reported significantly fewer prescribed opioid pills consumed, 0.45 versus 1.50, compared to patients in the group without counseling (P < 0.05), again with no significant difference in pain level experience. In addition, patients in the group with counseling reported a significantly lower number of total pain pills consumed over the course of the study than the group without counseling, 1.40 vs. 4.20 (P < 0.05) (Fig. 31-1). No major adverse reactions were noted in either group. Summary Points:  Pre-operative opioid counseling was found to result in a significant decrease in overall opioid consumption post-operatively.  Surgeons should consider routine pre-operative counseling of their patients to help minimize opioid use and potentially theoretical opioid abuse or diversion.  Surgeons should also consider recommend prescribing no more than 5-10 opioids post-operatively after CTR surgery. Figure 31-1: Postoperative pill consumption for counseling versus no counseling. PAPER 32 Clinical Paper Session 4 General e Friday, September 8, 2017  10:55e11:00 AM Evaluation/Diagnosis A Convolutional Neural Network Automatically Detects and Localizes Fractures of the Distal Radius with Greater Accuracy than Inexperienced Clinicians Level 2 Evidence Aaron Daluiski, MD Robert N. Hotchkiss, MD Doug Hanel, MD Sumit Chopra, PhD Rob Lindsey, PhD COI: Ownership Interest: Stock in Imagen Technologies (Daluiski, Hotchkiss, Hanel, Chopra, Lindsey) Salary: Imagen Technologies (Chopra, Lindsey) Hypothesis: Deep convolutional neural networks (CNNs) can accurately detect and localize distal radius fractures in radiographs better than inex- perienced clinicians. Methods: Deep CNNs have had tremendous success in detecting, local- izing, and identifying objects in many computer vision domains in which large labeled datasets are available for model training. We developed a novel CNN model and trained it on a dataset of 36,408 wrist radiographs labeled by two experienced hand surgeons (AD, RH) for the presence and location of fractures of the distal radius. The model treats the problem as a semantic segmentation task which first localizes regions of interest in a radiograph and then uses a deep CNN to output a heatmap which represents pixelwise probabilities of a fracture. The labeled dataset was divided into training, development, and test sets consisting of 80%, 10%, and 10% of the images, respectively. An additional held-out set was prepared consisting of 2,215 consecutive wrist radiographs from a hospital collected between July and September, 2016. The “ground truth” labels of this dataset were determined by the majority opinion of three authors (AD, RH, DH). The ROC AUC (area under the curve) was calculated to report the diagnostic accuracy of the model. To further asses the impact of our model on clinicians who could potentially be aided by this technology, we conducted a controlled experiment using two radiology trainees, two physician extenders, and two practicing urgent care physicians. Each clinician was asked to indicate the presence or absence of distal radius fractures in 250 radiographs randomly sampled from the held-out set, and we compared their performance against the model. Results: Of the 36,408 radiograph films 31% had fractures. Children less than 10 years of age, females over 50 years old, and left wrists were more likely to have a fracture. Our model had an AUC of 0.97 on the test set. On the held-out set our model had an AUC of 0.98, and a specificity of 93% at an operating point of 95% sensitivity, which is significantly more accurate than the tested trainees, physician extenders or practicing urgent care physicians (Figs. 32-1, 32-2). Summary Points:  Our proposed CNN-based semantic segmentation model produce heat- maps which accurately identify the presence and location of distal radius fractures.  The accuracy of the model was better than the tested trainees, physician extenders, and urgent care physicians, suggesting the proposed technol- ogy may be helpful in both residency training and urgent care clinical settings. S20