Google Glass

Incorporated with Evidence at the bedside

Glass use cases

• Glass has a lack of civilian applications

• However, there are many useful applications of Glass in the medical field

Dr. Selene Parekh, one of the first surgeons to use Glass in  surgery

"the conscientious, explicit and judicious use of current best evidence in making decisions about the care of individual patients"

Currently nurses use their own experience to treat patients, but many people want to have nurses follow evidence based medicine

Nurses can follow evidence based medicine by using "best pratice sheets" which reside in the nursing station

Evidence Based Medicine

Best Practice Sheets

These sheets give Nurses the ability to treat patients using Evidence Based Medicine

Problem!

Requires many trips to the Nurses station to retrieve. 

Would be much better if nurses had sheets on hand

Solution! Use Google Glass

• Does not require constant trips to nurses station
• Is mostly handsfree (less germ spreading)
• Heads up display allows nurses to continue working with information displayed

What do we need?

• Web Server
  • To send the information to glass
• Web Scraper
  • To download the pdfs from the databse
• PDF text extractor
  • To grab the Best Pratice items and send them to glass
• Glass Java Application
  • Allows user to specify which item they'd like to see information for

GET /teach/<searchQuery>/<searchBy>{title,keyword}

When a get request is sent here, the server uses the searchQuery item to search the database(using the Webscraper), and the returns a list of articles related to what the user is interested in. The response to this GET request is a Javascript Object, aka JSON. 

GET /choose/<searchQuery>/<choiceIndex>/<searchBy>

A GET request is sent here once the user chooses from the list of articles. The web scraper retrives the url for the pdf file associated with the article, and then the PDF text extractor retreives the items from the PDF and then sends them to glass as JSON.

Web Server

GET http://localhost:8080/teach/pressure%20ulcers/title

{
  "options": [
    "Pressure Ulcers (Prevention)", 
    "Pressure Ulcers: Nutrition.", 
    "Pressure Ulcers: Prevention and Management."
  ]
}

GET http://localhost:8080/choose/pressure%20ulcers/1/title

{
  "bullets": [
    " A high-protein oral nutritional supplement may be recommended along with standard\npressure ulcer care to reduce pressure ulcers in high risk...     persons. (Grade B)\n", 
    " Adults with pressure ulcers should receive nutritional screening. (Grade B)\n", 
    " Adults with pressure ulcers should be referred to a dietitian for assessment. (Grade B)\n", 
    " Individualized interventions prescribed by the dietician to improve nutritional status should\nbe implemented for those with pressure ulcers. (Grade B)\n", 
    " While nutritional interventions to prevent pressure ulcers in high risk patients have been\nshown to be effective and are recommended it needs to be borne in mind that research has\nhighlighted their cost relative to other interventions. Therefore ways to reduce the cost of this\nintervention should be sought. (Grade B)"
  ], 
  "title": "Pressure Ulcers: Nutrition"
}

Implemented in Python using the Flask Web Framework

Web Scraper

Implemented in Java using the HtmlUnit Library

<form id="sfovidclassic" name="sfovidclassic" method="POST" action="ovidweb.cgi" >

<input id="ovidclassic_focus" class="msp-ovidclassic-textbox"
       type="text" tabindex="1" name="textBox"
       value=""
       title="Enter keyword or phrase (* or $ for truncation)"/>


<input class="msp-ovidclassic-search-button" type="submit" tabindex="2"
       value="Search" name="submit:Perform Search|1"/>

 HtmlForm form = page.getFormByName("sfovidclassic");

 // Enter the search Query into the textbox
HtmlTextInput textBox = form.getInputByName("textBox");
textBox.setValueAttribute(searchQuery);




HtmlSubmitInput button = form.getInputByName("submit:Perform Search|1");
	
HtmlPage resultPage = null;
try {
    resultPage = button.click();
			
} catch (IOException e1) {
    // TODO Auto-generated catch block
    e1.printStackTrace();
}
		   

Using the OvidSP Databse

Web Scraper

PDF Text Extractor

Implemented in Java using Apache Tika Library

PDF Text Extractor

<div class="page"><p/>
<p>Copyright (c) 2014 The Joanna Briggs Institute licenced for use by the corporate member during the term of membership 3 of 4
</p>
<p>essential minerals, may yield deficiencies therein and provide an avenue for treatment.
• Many simple assessment tools are available and these are effective in highlighting the
discrepancies between a patient’s nutritional status and the state they should be in to
prevent or overcome a pressure ulcer.
</p>
<p>Best Practice Recommendations
</p>
<p>• A high-protein oral nutritional supplement may be recommended along with standard
pressure ulcer care to reduce pressure ulcers in high risk persons. (Grade B)
• Adults with pressure ulcers should receive nutritional screening. (Grade B)
• Adults with pressure ulcers should be referred to a dietitian for assessment. (Grade B)
• Individualized interventions prescribed by the dietician to improve nutritional status should
be implemented for those with pressure ulcers. (Grade B)
• While nutritional interventions to prevent pressure ulcers in high risk patients have been
shown to be effective and are recommended it needs to be borne in mind that research has
highlighted their cost relative to other interventions. Therefore ways to reduce the cost of this
intervention should be sought. (Grade B)</p>

The PDF Text Extractor takes the PDF and turns it into XML with <div> seprating pages and <p> seperating sections on the page

Glass Android App

Build like a regular Android App, but using some glass specific features in the Android SDK

Mostly a dumb app, just listens for responses to GET requests and displays them

Putting it all together

1. User says to "OK Google Teach me About"
2. Glass asks "What about?"
3. User says "Cancer"
4. Glass does a Get Request -- GET /teach/cancer/title
5. The Server runs the Web Scraper and retrieves a list article related to cancer
6. The Server then sends that list to Glass and displays it to the user.
7. The user taps Glass on the article that wants to be view Another GET request is sent -- GET /choose/cancer/1/title
8. Web Scraper returns to the list of article and download the appropriate PDF
9. The PDF Text extractor takes then pdfs and remove the best pratice text from it
10. Server sends the best pratice items to glass and displays them to the user