Notes

Outline

Design Heuristics for Creating Patient Education Materials for the Web

""It is more important..." "It is more important to know what sort of patient has a disease that what sort of disease a patient has."                               Sir William Osler

Goal PETTT seeks to enhance the effectiveness of                                     the University of Washington's faculty and thus of the institution itself, by creating a campus framework to promote the thoughtful                                     exploration, development, assessment, and                                    dissemination of next-generation technologies                                     and strategies for teaching and learning. Source: http://depts.washington.edu/pettt

Philosophy Respond to learner expectations and information needs Technology in the service of good teaching Instructional goals and needs should drive applications of technology Development should be part of the workflow, not extra-ordinary

Approach

Slide 6

Slide 7

Slide 8

"Site Value" Site Value Analysis Survey Feedback Email QuickPoll

Dissemination Conducting logfile analysis Conducting online survey Conducting interview with developer

Slide 11

Formative Evaluation Scientific-experimental approaches remain dominant and utilize methods that emphasize objective data gathering procedures and quantitative data analysis. Recent approaches include theory-driven evaluation, and model guided program implementation studies. Qualitative/anthropological approaches emphasize the importance of observation in social contexts, and the value of subjective human interpretation in the evaluation process.  ‘Grounded theory' has been used recently to study patient education processes.

Formative Evaluation We are integrating quantitative (experimental) and qualitative methods. Our focus is on process evaluation investigating the process of delivering a program or technology, including alternative delivery procedures. http://trochim.human.cornell.edu/kb/intreval.htm We are building qualitative and quantitative methods to: - integrate user feedback to content authors - develop an infrastructure for continuous quality improvement

Our methods are informed by Judy Ramey’s guidelines and strategies for analyzing audiences and patterns of use of Web sites. Ramey (2000) Guidelines for Web Data Collection Conducting separate analysis of: - Log files - Simple Utility Tool (web poll) - User survey - Feedback emails to the physician content author (free text queries)

Log files Logs of session connections to the web server requesting specific pages.  The chart below shows the times of day when users most frequently request web pages from the server

Simple Utility Tool (web poll) For selected pages the user is given an option to rate that specific page on a simple measure of satisfaction A five point scale to evaluate “usefulness”.

User survey

Feedback emails to the physician content author (free text queries) -Data comes from the Shoulder Source of the Arthritis Source -Prototyped a qualitative analysis of 176 messages -“Binning” procedure  to obtain 5 to 10 content categories for user questions

Slide 19

Slide 20

Content Authoring Cycle

-Focus on the analysis of email to physician content author (free text queries) -Describe how it will help us to integrate feedback to the design of authoring guides (templates) -Describe distinction between authoring templates and content management templates (XML/RDF schema)

Slide 23

Zope, XML, and 20-50-300 Good data design from the ground up

Meta-data in Articles Article Title Description Pages Subheadings -> Page titles Topic sentences -> Description Body text -> Keywords, search terms

Advantages of Standardizing Consistency Ease of use Accessibility Visibility to search engines

Tools for Standardizing 20-50-300: Text formatting guidelines XML: Extensible data format Zope: Content Management System 4XSLT: Centralized stylesheet processing

Formatting Guidelines 20-50-300 Each article has multiple pages For each page: ≤20 word topic sentence ≤50 character title ≤300 words in body text Note: These limits were targeted for Inktomi

Formatting Guidelines Set limits based on search engine “robot” preferences Too many/few words = Spam More pages = More relevance Improves: Ease of reading & searching Modularity As search engine ranking improves, hits will increase

Hits

Search Engine Ranking

Results

eXtensible Markup XML allows usage of 20-50-300 Allows descriptive coding of data <article type=“disease” title=“Rheumatoid Arthritis”> <description> … </description> <page type=“symptoms” title=“Symptoms of Rheumatoid Arthritis”> <keywords> … </keywords> <description> … </description> <paragraph> … </paragraph> … </page> <page> … </page> </article>

Advantages of XML Modular Separates data from formatting Standards-compliant (W3C) Inherently Meta-data friendly Extensible, customizable

Content Management Reduces complexity Increases consistency Allows implementation of standards Permits management by multiple content providers

Zope Open source Constant improvement Online community support, free Modular and Expandable Powerful scripting and external methods using Python XSLT parsing with 4XSLT Web interface

4XSLT XML to HTML Conversion using: XML stylesheet templates (XSLT) XSLT Parser (4XSLT, Xalan, etc.) May also serve other protocols (WAP) Python-based Easily incorporated into Zope through external methods

XML to HTML

Diffusion Convey a relative advantage over existing ways of doing things Are compatible with expectations, experience and needs of faculty Reduce complexity associated with new technologies Are triable in that they can be experimented with on a limited basis Are widely observable to the campus community

PETTT: depts.washington.edu/pettt Catalyst: www.catalyst.washington.edu Arthritis Source: www.orthop.washington.edu/arthritis