UCHC Pathology Informatics Facility
Nascent Pathology Informatics at UCHC began around 1978 with
the Immunology BBS & ImmunoMail, a Bulletin Board & Fidonet
based mail system running in CP/M on an Intel 8080 on a S100 bus and using POTS
dialups. Lack of users and the
availability of UCHC’s internet connection led to its abandonment. More successful about this time was LabX, our tool for generating frame based expert
systems. LabX was central in our offering a medical school
elective/selective in Artificial Intelligence.
Students used LabX to develop
expert systems in the clinical laboratory test domains. CaX
and ClotX were some of the expert
systems developed by our medical students that generated clinical diagnosis for
bone disorders and clotting disorders based on Bayesian inferences.
Pathology
Informatics at UCHC gave birth to its first production system about 1985
through the generous financial support of the then Chair Doctor Irving
Goldschnieder and the strong leadership of Doctor Bernard Gondos,
Director of Anatomic Pathology With
their support and 20,000+ lines of COBOL code the PRN Anatomic Pathology Management System,
a surgical pathology LIS running on DEC’s PDP 11/23 with TSX multi-user,
multi-tasking operation system on top of
RT-11 was placed into production. This system (Lindquist
RR. PRN Anatomic Pathology Management System. Computers in Biology and Medicine. 21(6):I, 1991) contained
patient demographics, diagnosis and written reports in btree data structures
that were fully searchable. Pathology
reports with diagnosis and SNOP codes where entered real time with Lex-11 word
processing software and the pathology reports were dumped to connected printers
which provided the paper report. The system communicated with the hospital’s central
billing computer by tapes delivered via sneakernet. After a successful run of several years, when
Dr Gondos left the institution, the computerphobic Luddites prevailed.
PathWeb (http://pathweb.uchc.edu ), our virtual pathology museum was our first web-base project, which went online viewable with Netscape v1.? in 1995, is a depository of mostly gross pathology images accompanied by specific descriptions of the images’ pathology and generic information about the etiology, pathogenesis, epidemiology, clinic presentation, general gross & microscopic descriptions, and references. (For a very brief visual tour of PathWeb goto http://mediasite.uchc.edu/Mediasite41/Viewer/?peid=229685f54c0b4b67a3fd26bd3c32b16e and look 18 minutes & 45 seconds into the Webinar) Included in PathWeb are specific links to the appropriate recent journal articles at PubMed to be sure the information keeps pace with rapid medical developments.
Pathweb is data centric and the content actively delivered to user queries. The data reside in a relational database (initially Butler SQL, then Microsoft SQL Server, and now mySQL). PathWeb interacted with the PathWeb database initially via Tango, then Cold Fusion and now PHP. The web servers delivering the content to users was initially WebStar, then IIS when Microsoft finally developed a server, and now Apache.
Users interact with the database via menus offing organ and/or pathological process. In the middle years full text search across all tables was offered via Verity, but was dropped when our Verity license could not be renewed for financial reasons.
Peer review of PathWeb’s content is offered via Rate It our interactive feedback module, which allows users not only to generically rate our content , but also to give us specific comments that may be incorporated into our content.
Our desire to provide PathWeb
content to our students via CD roms, lead to the development of PathWeb’s content being dumped into
HTML script from our database via our HTMLGEN & DB2HTML in 1997. With the HTML pages
our eAtlas of Pathology (http://radiology.uchc.edu/eAtlas/)
was created, which for historic reasons was placed on our radiology image
teaching site (infra vide). The eAtlas
contains the entire contents of the PathWeb
database in static form and is organized by Diseases by Organ and by Diseases
by General Disease Process. For a brief visual tour of eAtlas goto http://mediasite.uch
c.edu/Mediasite41/Viewer/?peid=229685f54c0b4b67a3fd26bd3c32b16e at 23
minutes 52 seconds into the webinar.
PathWeb et al. served as a focus for tweaking our content for search engine optimization (SEO). Currently (Jan 2010) a search for “pathology atlas” returns pathweb.uchc.edu in to top ten at Google (4th of 1,540,000) and at Yahoo (6th of 3,060,000). Our SEO efforts also are reflected in traffic to PathWeb and thereby opening up our content worldwide. Today (Jan, 2010) The eAtlas and the other PathWeb spin-off, eSynopsis of Pathology, is the second most visited web site in the Medical School after BlackBoard. Before BlackBoard more than half of all web traffic coming to uchc.edu landed on eAtlas et al. For all traffic coming to uchc.edu The order of most visited is Blackboad, dental school, and PathWeb et. al
The eSynopsis of Pathology (http://esynopsis.uchc.edu ), unlike the image centric eAtlas, is centered on specific diseases and presents all of the images for that disease along with generic information to the viewer.
PathCrawler is our effort to amass a large collection of pathology images from the internet. Our bot software crawled the internet seeking images of pathology specimens. It collected links to 55,000+ pathology specimen images and was fully searchable. Unfortunately the server housing PathCrawler was damaged during our recent move to a more economic location and has not yet been resuscitated.
Pathology Policy & Procedure Management System (Hopfer, SM and Lindquist, RR. Document Control: A College of American Pathologis Requirement. Annals of Clinical Laboratory Science 37:201, 2007) is in production. A webinar focusing on the front end of the system can be viewed at:
(http://mediasite.uchc.edu/Mediasite41/Viewer/?peid=229685f54c0b4b67a3fd26bd3c32b16e) It is a web based intranet system written to display policies & procedures and, most importantly, keep track of the various versions of policies & procedures and to exactly identify the specific dates of use for any given policy or procedure. This feature is especially important in possible malpractice actions. Moreover if an existing policy or procedure is replaced the system maintains to reason for the revision.
In addition, a complete audit trail of the College of American Pathologists (CAP) mandated & enforced annual review of each policy and procedure by the department chair’s designated reviewer is maintained. Automatic e-mail is generated by the system to remind reviewers of upcoming review deadlines. Modules were written that also keep audit trail for the CAP mandated reviews by technologists of their relevant policies & procedures.
The CAP requires immediate remedy in the event of server and/or network failure so that there is never any time when the laboratory is deprived of access to policies or procedures. Programs were written so that the entire laboratory document management system resides on CD or portable USB drives that can be loaded into any of the multiple Windows computers and be up and running within 2 mouse clicks whether connected to the network or not.. Thus the laboratory will never be deprived of up to date policies and procedures.
Finally to conform to CAP guidelines that the data must be verified free of corruption or malicious tampering, computer programs were written to digitally fingerprint all data and to verify the fingerprints at will. Addition computer programs were written to insure compliance with title 21 Code of Federal Regulations (CFR) so that results generated by the laboratory can be used in support of new drug applications to the Food & Drug Administration.
Having all policies & procedures located in a central database completely eliminates chasing after policies & procedures which often turn out to be out of date and or un-reviewed and insures the correct, current procedure is applied to laboratory testing. It also facilitates the accreditation inspections by the CAP.
Electronic Specimen Collection Manual: A significant problem with the paper manuals that existed prior to this project was that whenever changes/corrections were made to the manuals all of the manuals at the many nursing stations, clinics and doctors offices had to be tracked down and replaced. This required much expense in printing the new manuals and a lot of leg work in physically replacing the manuals. Moreover, it was never certain that all the manuals had been replaced. These problems were eliminated by placing all of the information in a database and serving them up on the web. This project is in production on the intranet. Unfortunately because of current IT policy, this system could not be made available on the internet. For the internet we have static html pages server up. Similarly the public website for the laboratory is completely static html pages. In contrast the intranet home for pathology is all database driven. It can be viewed from within UCHC at http://labmed.uchc.net
A current project is the Visual Learning Cloud. It is a resurrection of an old (12 year old?) salami slicing project: the curricular knowledge domain can be digested if made finely granular and presented grain by grain. Just like salami eating: difficult to eat it whole but easy slice by slice. This salami slicing project was not supported by the school and it lost out to the school funding a “DICOM database” project which never materialized.
The VLC is also based conceptually on our old PathX in that each grain of knowledge is represented as a frame or object or “flash card in the deck.” Because I wanted a convenient testing environment that could be quickly populated the demo version demo is image centric; however, text can replace the images without any modification of the current code.
The VLC has 2
frame collections: one common/public to all student cohorts and one private
frame collection belonging to each individual student. An individual student amasses his/her private
frames either by creating his/her own or by selecting one from the
common/public frame collection. When collecting a public frame to be included
in his/her private frame collection the student can edit the content of the
frame.
Importantly, when the student selects one of the common/public (or creates his/her own) the student rates his/her knowledge of the frame’s content on a 1 -5 scale with 5 being good grasp of the content/knowledge and 1 poor. This rating is also done on the self created frames. The reason for the rating is that the student is reminded on logon and by email that certain frames should be reviewed at a specific times and the rating re-evaluated. This way allows the frequent review of shaky knowledge and infrequent review of frames well assimilated by the student. The algorithm determining the review intervals is based on Leitner's selective learning principle.
In reviewing the frames’ knowledge, the student has several choices for the review: review only frames with certain knowledge ratings; how many frames to review; sequential review; random review; and/or review with scoring. With the scoring option the student is presented at the end with those needing more work and the opportunity to review them. The student also has the option of having his/her response instantly evaluated so that the student can review the material again.
VLC allows for the creation of study groups so that a student can share their private frames with his/her study group for feedback and peer review and those frames passing peer review can be added to the common/public frame collection or added to the private frame collections of other students. --
To see some screen shots of VLC: http://informatics.uchc.edu/PathInformatics/VCL_screen.htm
Social networking is an important component of VLC.
The basic code for social networking component has been written for
another project, Faculty Professional Network (Infra vide) and is being
incorporated into VLC with some
slight modification to be more appropriate to students rather than
faculty. In short the social networking
component is FaceBookoid. It provides
for joining groups such as study groups, finding classmates who have some
expertise in a particular area, blogging, etc. Click here to see some screen
shots in progress. http://informatics.uchc.edu/PathInformatics/snet.htm
eSyllabus is a current project attempting to bring and encourage student-faculty interactivity to the teaching syllabi . Currently the lecture syllabi are posted on blackboard for students to view or download and provide no means to interact with the faculty author. In an attempt to facilitate student-faculty interactions, each syllabus paragraph by paragraph is entered into a database. When the syllabus is viewed on the web by the student each paragraph has a space for student comment/question. The faculty member sees the comment/question on logging on to the system and can immediately respond to the student’s question/comment. The original student or other students can follow up on the faculty response. Student comment/questions are anonymous to encourage student participation. The faculty member has the option to eliminate comments that are inappropriate. For a screenshot of this project in its early stage see this link: http://informatics.uchc.edu/PathInformatics/Syllabus.htm
eQuiz
(http://pathology.uchc.edu/elab/labh/exam_find.php?row=1&ls=6&order=49|43|53|39|45|40|54|52|41|47|46|42|55|48|50|51|56|9999 ) is a quizzing program for images and is in
production awaiting more content.