MS in Public Health (M.S.P.H.)
Degree Granting Department
Alfred Mbah, Ph.D.
Rachel Williams, M.D., M.S.P.H
Jared Jeffries, M.D., M.S.P.H
Decision Support Tool, Musculoskeletal Injuries, POCUS, Rotator Cuff
This descriptive retrospective cohort study utilized a large Workers Compensation insurance database. Included in this study were 481 shoulder MRI's performed during calendar year 2017 that were (a) at least 2 weeks after the initial clinic visit, or (b) greater than 6 weeks after injury, but (c) not more than 3 months after the date of injury. Of the 481 cases, 432 were used to evaluate potential associations between the measured variables and generate a Bayesian network prediction model where only a few variables were required to accurately guide clinical decision making for rotator cuff and biceps tendon injuries. The other 49 cases were randomly selected as a validation set and were not included in the model. After removing unnecessary variables, the model was condensed into a clinical decision support tool based. The generated clinical support tool in conjunction with limited point of care ultrasound examination of the supraspinatus and infraspinatus tendons would guide clinicians to triage patients into one of three clinical management groups: no follow up needed, conservative management, or MRI and orthopedic referral. Internal validation of the generated model yielded 96% accuracy in placing patients in the correct clinical management group which could help improve provider and patient confidence in the diagnosis and treatment plan as well as minimize delays in patient care, prevent unnecessary referrals, and expedite return to work.
Scholar Commons Citation
Ayers, Gwen Marie, "Limited Point of Care Ultrasound Clinical Decision Support Model for Work-related Injuries of the Shoulder Utilizing Bayesian Network" (2021). Graduate Theses and Dissertations.