Teaching clinical skills and decision makin in the di ital a e making in the digital age: Development of closed loop simulators for training and evaluation of clinical professionals.
Kanav Kahol PhD, Asst Professor, Professor Dept of Biomedical Informatics, Informatics Arizona Ari ona State University Research Faculty, Simulation Education and Training Center, Banner Good Samaritan Medical Center
[email protected] 6028272546
Objectives • To show how to measure and train both psychomotor and To show how to measure and train both psychomotor and cognitive skills of surgeons. • To enable understanding of solutions for feedback while To enable understanding of solutions for feedback while maintaining the current levels of faculty involvement. • To acquire an awareness of the potential for the use of simulation gaming platforms in training. • To develop strategies for effective remediation.
Surgical Simulation Training Surgical Simulation Training “Virtual Virtual reality reality–trained trained residents : residents : performed the procedure 30% faster, and performed the procedure 30% faster, and made six times fewer intraoperative errors p when dissecting the gallbladder from the liver bed.”
Virtual Reality training improves operating room performance; Results of a randomized double performance; Results of a randomized, double blinded study. Seymour et al; Ann Surg., 2002
New Directions New Directions N i Novice
Focused Simulation Training At‐Home Trainingg Psychomotor Cognitive
Embodiment Embodiment Training
Pre Trained Novice
Initiatives • Cognitive Simulators: Simulators that enhance critical thinking g g and hone cognitive skills. • At‐Home Simulator Systems: Simulator systems and monitoring architecture to support offline learning. it i hit t t t ffli l i • Embodiment Simulators: Enabling clinical professionals to p practice learnt skills in actual environments. • Intelligent Tutoring Systems: Programs that adapt and offer real time feedback • Use of Simulators Beyond Education: Measuring the effect of Fatigue • Simulators Driven by real world: Simulators to measure and Simulators Driven by real world: Simulators to measure and train team behaviors
Cognitive Simulators Cognitive Simulators • A key aspect of surgery and for other skill based for other skill based procedures lies in applying decision making skills while accomplishing psychomotor li hi h t skills. • When people perform two When people perform two or more tasks simultaneously, the tasks are often e ec ted slo er are often executed slower and with more errors than when they are carried out as single tasks.
Multitasking Environments Multitasking Environments • SSignificant research has been conducted to g ca t esea c as bee co ducted to understand the relation between task interference, learning and experience. In general, task interference is severe during learning periods k f d l d but reduces dramatically with practice. (Ruthruff 2006) • Hence if simulators present multitasking environments with cognitive tasks as well as environments with cognitive tasks as well as psychomotor tasks they will provide an adequate basis for training.
Framework For Cognitive Simulators
cognitive sensors psychomotor sensors
Information Inte egration
Sensory Module
Simulation Module
Cognitive layer
Conventional Simulators (psychomotor evaluation)
(cognitive evaluation)
neuropsychologically inspired variations py g y p (cognitive load in addition to psychomotor load)
FFeedback and db k d Evaluation Module
cognitive proficiency feedback psychomotor proficiency feedback
Weighting Mechanism
Universal score
Objective Proficiency Measures Objective Proficiency Measures • Employ neurological and kinesiological features to analyze surgical proficiencyy p • Constructive task de‐ composition based feedback – Breaks a complex motion into simpler units that are easy to analyze and more easy to analyze and more importantly easy to comprehend and change by the user.
Tool movements N i Novice
Intermediate
Rosen 2002
Expert
Hand Motion Hand Motion
0.0042 moved 0.023 1.00 0.34 0.0043 moved 0.0023 1.04 0.37
Skill Analysis Systems • Gesture Segmentation (Naïve Bayesian Classifiers) • Coupled Hidden Markov M d l ith hi Models with hierarchical hi l hand representations • Validated across
– Level of experience in surgical activities. – Level of Fatigue.
Computer Vision for Surgical Movement Analysis Movement Analysis •
•
NSF Sponsored grant to enable webcam based analysis of surgical movements. American Board of Surgery.
Initial Results Initial Results In nstrument P Path Inefficciency
12 1.2
1
Expert p
Intermediate
Novice 0.8
0.6
0.4
0.2
0 Smoothness Between Groups ptying}
Adapt gaming Adapt gaming scores to our needs
Monitor progress through mechanism that work in an ambient manner
Matching observational Matching observational Parameters in the real world And virtual world
Wii and fine motor skills
Fine motor skills based games are very suitable
Very high correlation with b i basic gestures of surgery t f
Quantitatively we found that hand movement acceleration, and joint angles showed 0.78 to 0.91% correlation. correlation
Cons: doesn’t have the fulcrum effect and significant weight.
Apparatus pp • Gaming Gaming Extensions to Wii Extensions to Wii can be modified for surgical probe based interactions. • WiiMote Extension • Movement Constrainer
Location of wiimote
Full System in Action y
Study
experimental group p g p control group
proficiency
time
errors
Cost of surgeons learning robotic surgery in OR Cost of Robotic Surgery 25000
20000
15000
10000
5000
0 1‐2 weeks
2‐4 weeks
4‐6 weeks
6‐8 weeks
8‐10weeks
This is over and above mandatory 1 day training by Intuitive@$5000 Per physician Estimated as $1429 initial cost and $535.50 per 15 minutes
Robotic Surgery Simulator Robotic Surgery Simulator
Embodiment Simulators Embodiment Simulators •
The statement that learnt psychomotor skills in pristine environments can transfer to real environments is a speculation. i i l i
1
The Effect of Noise on Proficiency
0.9 0.8 0.7
Percentage decrement 31%
33%
26%
125%
97%
0.6 0.5
No Noise Noise
0.4 0.3 0.2 0.1 0 Proficiency
Hand Movement Smoothness Tool Movement Smoothness
Time
Cognitive Errors
Framework For Embodiment Simulators Sensory Module
Simulation Module
cognitive sensors psychomotor sensors
Information Inte egration
Environmental Variations
Cognitive layer
Conventional Simulators (psychomotor evaluation)
(cognitive evaluation)
neuropsychologically inspired variations py g y p (cognitive load in addition to psychomotor load)
FFeedback and db k d Evaluation Module
cognitive proficiency feedback psychomotor proficiency feedback
Weighting Mechanism
Universal score
Noisy Simulators Noisy Simulators • In addition to cognitive variations added noise to simulations to simulations Difference in Groups Trained in Noisy Conditions and Noiseless Conditions 1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 Proficiency
Hand Movement Smoothness
Tool Movement Smoothness
Group Trained in Noisy Conditions
Time
Group Trained in Noiseless Conditions
Cognitive Errors
Integration into Curriculum Integration into Curriculum • Simulation cannot be successful in a parallel education model. • We need a strategy wherein simulation can be integrated into curriculum for residents and medical students. • Take home simulation helps in reducing overall time T k h i l i h l i d i ll i required for simulation. • An additional methodology can help A dditi l th d l h l • In our residency, residents spend a mandatory 1 month training at simulation center where they are month training at simulation center where they are also taught basics of research (symbiotic loop)
Focusing Simulation Training Focusing Simulation Training
Example Advanced Cardiac Life Support Example Advanced Cardiac Life Support Training Focused on EKG recognition and errors made in diagnosis.
Errors Training and its Effect Errors Training and its Effect nt Pearson ns Correllation Co oefficien
Correlation with Experts 0.9 0.8 0.7 06 0.6 0.5 0.4 0.3 0.2 01 0.1 0
Novices Before T i i Training
Conventional ACLS Low End Simulation High End Training T i i with Errors Training ith E T i i Simulation with Si l ti ith Errors Training
Bringing the real into virtual environments • A system to monitor real y environments and play them back in virtual environments. • Can capture group activities C t ti iti through RFID sensors, audio analytics, proximity information, process through Hidden Markov Models and Kalman Filters Kalman Filters • Outputs a scripted activity log which can be played back in http://www.slideshare.net/KanavKaho l/ i t l l/virtual‐worlds‐and‐real‐world ld d l ld Virtual Worlds like ActiveWorlds and SecondLife…
Virtual Playback and Analysis Tool Virtual Playback and Analysis Tool Demo
Conclusions An insitu design of simulators addresses the An insitu design of simulators addresses the problems of education in an effective manner. Many of the innovations are basic from Many of the innovations are basic from technology perspective but have huge impact on the surgical learning on the surgical learning. All the simulators are developed without the generally required graduate student add‐on. ll i d d d dd Measuring activities in real world is not an option but required to ensure validity.
TATRC PI $2 2 M $2.2 Socially relevant Telemedicine training Networks NSF co-PI $899,000 A Machine Learning Approach to Computational Understanding of Skill Criteria in Surgical Training LABORATORY Partner
http://symbiosis.asu.edu
James S McDonell Foundation co-PI co PI Cognitive Complexity In Critical Care $5M