Dungeons Dragons Taught Me How to Make Better

Dungeons & Dragons Taught Me How to Make Better VPs Randomizing Physiological Data to Rapidly Produce 97 Clinically Realistic VPs Mike Paget, Janet Tworek, Kevin Mc. Laughlin, Bruce Wright University of Calgary, Alberta, Canada

Overview • 97 VPs Requirements • Clinically relevant Solution • Gaming experience • d. B design Results • 20 mins SME time per VP • 97 clinically relevant VPs

Requirements 1. Create 97 VPs for full year course – Clerkship (Year 3) students – Little SME time – Past experience with templates • Time consuming • Overlooked data – No preexisting cases for Uof. C context

Requirements 2. Clinically accurate values and variety in VP - Names History Laboratory Physical Exam

Solution - Gaming Dungeons and Dragons

Solution - Gaming Dungeon Master’s Guide • Randomized realism • Qualities related to numbers

Solution - Gaming Dungeon Master’s Guide • Randomized realism • Qualities related to numbers Virtual Patients • Assign text strings to numbers • Treat numeric data as numbers that fall within a pre-specified range.

Solution: Process of Creating VPs • All content areas • Ranges for normal results 1. VP Template 2. Create Data Tables • Table for each text field • ID number per row • Randomized normal values • Variety of text responses 3. Prepared Normal Template 4. SME Proofread • Change key nodes • Approve -, + • Complete content • SME approved 5. Case ready for upload

1 a. Create Template for VPs Node Patient Response Any diseases run in the family? Lab: Anti Hepatitis A Virus Total Lab: Chemistry Na – Sodium Knee exam: Do you have any pets? • Hx. Ex. Ix. Dx. Rx model of VPs • Set questions

1 b. Create d. B tables of VP data ID Diseases that run in the family 1 Cancer ID Positive or Negative 2 Stupidity 1 Positive 3 Heart attacks 2 Negative 4 Lung cancer 5 Arthritis 6 Accident prone 7 Asthma 8 Strokes 9 Dementia 10 Pneumonia ID Pet 1 No, I don’t like animals 2 I have 3 dogs 3 I have a cat I took in 4 Just a goldfish 5 I have a dog, but I think I might be allergic

2 a. Create Specific Data Range by Node Low High Additional Text Any diseases run in the family? 1 10 Not applicable Lab: Anti Hepatitis A Virus Total 1 2 Lab: Chemistry Na – Sodium 133 145 Knee exam: 1 1 Do you have any pets? 1 5 Normal range between 133 to 145 mmol/L

2 b. Randomly Assign ID Values in Data Range to specific VPs Node Patient 1 Patient 2 Patient 3 Additional Text Any diseases run in the family? 3 8 1 Lab: Anti Hepatitis A Virus - Total 1 1 2 Lab: Chemistry Na – Sodium 137 139 141 Knee exam: Active movement: Flexion 140 degrees, Ext…. Do you have any pets? 3 5 2 Normal range between 133 to 145 mmol/L

3 a. Create report by linking randomized ID values to text fields in appropriate tables

3 b. Convert IDs to actual values & text Node Patient 1 Patient 2 Patient 3 Additional Text Any diseases run in the family? Heart attacks Strokes Cancer Lab: Anti Hepatitis A Virus - Total Positive Negative Lab: Chemistry Na – Sodium 137 139 141 Knee exam: Active movement: Flexion 140 degrees, Ext…. Do you have any pets? I have a cat I took in I have a I have 3 dog, but I dogs think I might be allergic Normal range between 133 to 145 mmol/L

Proofing VPs • All content areas • Ranges for normal results 1. VP Template 2. Create Data Tables • Table for each text field • ID number per row • Randomized normal values • Variety of text responses 3. Prepared Normal Template 4. SME Proofread • Change key nodes • Approve -, + • Complete content • SME approved 5. Case ready for upload

4 a. Normal Template to SME

4 b. SME records edits Node John Smith Any diseases run in the family? Kidney disease in a few relatives Lab: Anti Hepatitis A Virus - Total Positive Lab: Chemistry Na – Sodium 160 Knee exam: Active movement: Flexion 140 degrees, Ext…. Do you have any pets? I have a dog, but I think I might be allergic Additional Text Normal range between 133 to 145 mmol/L

5. Upload content to OLab • SME time

Multiple cases at once • Insert picture of multiple cases

Results • Time: – 20 mins per VP – 45 mins per side-by-side VPs (2) • Salient negatives, positives/normals • Side-by-side – Contextual cues & patterns by presentation or diagnosis – Complexity of medicine

Result – Clinically accurate content

Conclusions • Education as driver • Realistic VPs = data issue – Leverage gaming, d. B experience – Existing software + open source VP software • Multiple clinically accurate VPs – Appropriate personnel use – Complexity of Medicine • Sharing

Conclusions • Gaming “nerd” = VP programming superhero

Thank You Janet Tworek jktworek@ucalgary. ca