a web-based correctional telemedicine system with ... · a web-based correctional telemedicine...

10

Click here to load reader

Upload: phunglien

Post on 30-Aug-2018

212 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

A Web-Based Correctional Telemedicine System with Distributed Expertise

Siva R. SankaranCalifornia State University

Northridge, CA [email protected]

Tung BuiUniversity of Hawaii, Manoa

Hawaii, HI [email protected]

h

s

d

d

Abstract

Due to increasingly tougher and mandatorysentencing laws, the U.S. has experienced a sharpincrease in the prison population during the last decade.Correctional facilities face unique challenges inproviding health care to a population that has a history ofphysical, mental and psychological abuse. Telemedicineis an innovative way to bring health care to prisons in asafe and cost-effective manner. The steady proliferation ofthe Internet, its versatility and low cost are ushering innew possibilities for providing telemedicine through sucha cyberspace medium. This paper discusses the workflowsinvolved in correctional health care environments anddevelops an object-oriented architecture for a Web-centric telemedicine system. The prototype uses relationaltables to represent expert knowledge and ActiveXprogramming for initiating and coordinating theconsultation sessions over the Internet. The prototype hasthe ability to deal with single physician consultations aswell as with distributed multiple experts in more complexdiagnostic sessions.

1. Introduction

The progress of technology over the last threedecades has touched many aspects of human life, fromassisting with the simplest of tasks at the individual level,to modern complex organizations, and to the mostcomplicated of national defense efforts. Among themultitude of computer applications, none has contributedmore directly to improving the quality of our daily livesthan health care systems. In the past, many of thesesystems were limited to patient billing, insurance,prescription and medical history databases (Snyder,1997). Further, these applications have largely beenlocalized to either a single hospital or a chain of hospitals

0-7695-0493-0/00

in a confined geographic area. Recent advances intelecommunications and Internet, however, has now madeit possible to break geographical barriers and allowphysicians deliver health care to patients on an any timeand any place basis (Mason, 1998). This has led to a newbranch of computer applications called telemedicinesystems.

Today, telemedicine allows patients in remoteareas to receive consultation rapidly by specialists whoare often located in urban areas (Lugg, 1998). Care maybe provided at local health centers instead of having totransport patients to large referral hospitals. Often patientscan receive early and more accurate diagnoses througconsultation with experts and the treatment interventionsmay be easier, safer and cheaper. Families don’t have toundertake the financial burden of paying for gasoline,food, lodging, and time lost from work to take the patientto a distant referral center for an outpatient consultation orhospitalization. The telemedicine system also decreaseduplication of laboratory and radiology studies, whichfrequently takes place when a patient is referred toanother medical facility. Hence, insurers, social agencies,and the patient’s family save costs (Armstrong, 1998).Telemedicine also permits greater access to health relateinformation for patients and families to learn more abouthealth problems, treatment options, and preventionstrategies. Another benefit is telementoring - thecontinuing medical education and training the fieldprofessionals spontaneously receive throughteleconsulting with experts (Franczyk, 1998). Finally,telemedicine has the potential to tie distant practitionersmore closely to experts and colleagues in other areas anbring new discoveries and treatments into practice fasterthan ever.

$10.00 (c) 2000 IEEE 1

Page 2: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

2

l

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

In the past, telemedicine has mainly been limitedto telephone consultations. Physicians, nurses and otherpersonnel talk with each other, patients and familiesproviding information and offering reassurance withoutthe expense of an office visit for the patient. Many healthinsurance organizations have established telephoneadvisement programs to reduce unnecessary office visits.The fax and e-mail have recently enhanced the availablealternatives to share medical data among consultingphysicians (Stoloff, 1998). There are also sensorsavailable that can transmit information such as pulse,breathing and blood pressure collected directly from thepatient. The most sophisticated telemedicine systemsinvolve interactive videos. These systems requireextensive investments towards the design, implementationand maintenance of telecommunications infra structure. Inaddition to laying long distance fiber optic networks andsetting up satellite links, these facilities are furnished withspecialized equipment and software at all interfacinglocations where human experts interact (Binius, 1999).However, the widespread growth of the Internet isbeginning to give us an alternative. Many patients todayhave Internet connections already set up in their homes.The versatility of the Internet in terms of being able tocarry multimedia such as voice, video and data combinedwith its low operating cost features makes it an attractivemedium for telemedicine consultations. Advances in datacompression, security, cross platform software anddevelopment of modeling techniques in desginingdistributed networks also hold the Internet as a promisingnew tool for future telemedicine applications.

2. Telemedicine and Correctional Setting

An important development in telemedicineduring the last decade has been its appeal among state andfederal correctional institutions. Correctional facilitieshave unique characteristics that make them well suited fortelemed applications. The prison population in the UnitedStates has risen exponentially over the past two decadesdue to tougher laws and mandatory sentencing andcurrently stands at more than 1.5 million (Thorburn, 1995;Vitucci, 1999). This growth has resulted in higher healthcare costs. Health care providers practicing within thewalls of prisons are faced with the challenge of caring forpatients who had long abused their health with drugs,unsafe lifestyle and poor nutrition. Many suffer fromchronic physical and mental illnesses as well. Diseasessuch as Hepatitis and Aids, along with heart conditionsand cancer among the older prisoners are increasinglyplacing pressure on prison administrators to findinnovative ways to provide health care in a cost-effectivemanner (Spaulding, et al., 1999). Telemedicine reducesthe cost of bringing medical specialists to the correctionalfacility. Treating inmates through distance consultation

0-7695-0493-0/00

helps to avoid the high costs of maximum security-transportation and security-supervision of the patient atthe hospital. The transportation and supervision costs arehigh because at least two guards and a state vehicle arerequired for security reasons. It is estimated that theaverage prisoner transportation cost is over $700 for oneoff-grounds appointment (Swanson, et al., 1993). Whenhelicopter transportation is involved the cost rises to$5000. In addition, correctional programs also areexpected to ensure public safety by preventing potentialescapes during transportation to and from the hospitals.Furthermore, incorrect diagnosis and treatment also havelegal and monetary consequences (Gilbert, 1999). In somesituations when inmates require medical attention, it canbe difficult to decide whether to treat the patient inhouseor send to the emergency room for off-grounds hospitalcare. For example, if an inmate becomes unconsciousbriefly during a prison fight but maintains normal vitalsigns, and no other obvious sign of injury, it may be adifficult task to decide whether to send the patient to thehospital or treat the patient within the prison infirmary.Where inmates have died due to incorrect medicaldiagnoses under state care, multimillion dollars lawsuitsfrom families are not uncommon. Telemedicine offers aneffective solution in the diagnosis and treatment of theprison population in these situations.

It can be seen from the discussion so far thattelemedicine involves a complex interplay amongpatients, local health personnel, remote expert physicians,and delivery technologies in space and time. Modelinghealth exams, diagnostic workflows, expert consultationand negotiation sessions, incorporating medicalpolicies/procedures and patient data into the decisionprocesses are complex but yet are the requiredfundamental blocks in conceiving and implementing anytelemedicine system. The purpose of this paper is toaddress these issues with emphasis on the correctionasettings. The paper is organized as follows. Section 3covers the literature review of telemedicine with some ofthe early systems both in standard as well as prisonenvironments. Section 4 discusses the workflows intypical telemedicine sessions and develops a processmodel. The object model for a prototype telemedicinesystem consisting of the suggested objects, attributes,relationships and methods are illustrated in Section 5along with examples of how the distributed expertise ofthe various consultants can be represented on the Web. InSection 6, the prototype implementation itself is explainedusing a sample scenario. Finally, we conclude with adiscussion of future challenges in telemedicine.

3. Literature Review

$10.00 (c) 2000 IEEE 2

Page 3: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

3

t

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

3.1 Telemedicine Systems in StandardEnvironments

Telemedicine has been in existence for almost 40years ( Grigsby and Sanders, 1998). The first knowntelemedicine project involved the transmission ofneurological records across the campus of the Universityof Nebraska in 1959 (Benschoter, 1971). Five years later,the same university established a link with a state mentalhospital 112 miles away. The Nebraska program andmany of the other early telemedicine applications aroseout of concerns about the limited access of remotepopulations to a variety of health services. Since thattime, many state and federal agencies as well as privateinsurers, health maintenance organizations (HMOs),software companies, and medical device manufacturershave created their own telemedicine initiatives. Althoughthe total amount of money spent on telemedicine researchand development is unknown, a recent report by the U.S.General Accounting Office (GAO) places telemedicineinvestments by nine federal departments and independentagencies at $646 million for fiscal years 1994-1996.According to the report, the Department of Defense(DOD) is by far the biggest investor, allotting $262million to telemedicine initiatives (Mun, et al., 1998).The departments of Veterans Affairs, Health & HumanServices, and Commerce follow, with each investingapproximately $100 million. Like the military, the healthcare system operated by the Department of VeteransAffairs (VA) has characteristics that are attractive to thoseevaluating telemedicine. In addition, the VA has, overseveral years, developed a fairly comprehensive andflexible patient information system that can integrate text,test results, and images (Dayhoff, 1996).

In May 1996, telemedicine was used during anexpedition to Mt.Everest. To ensure the safety of theseclimbers, who were 29,000 feet above sea level andexperiencing a harsh environment, a team of remotephysicians monitored them and reported back on theirprogress. From the base camp clinic, video and data weretransmitted via the INMARSAT Indian Ocean satellite toa land station in Malaysia. Calls were routed fromMalaysia over the SATCOM Global ISDN network toSanta Paula, California where they were picked up by theAT&T ISDN network. Video and data received from themedical teams and the Everest site was placed directly onthe Web site. Without the telemedicine network, theclimbers could not have been successfully monitoredfrom remote locations. (Larkin, 1999).

It was mentioned in the earlier section that someof the new telemedicine systems involve the use of video.These systems let specialists examine patients by lookingat computer video images from remote hospitals. They

0-7695-0493-0/00

are being increasingly used in emergency rooms. This isbecause it is not cost-effective to have a trauma specialistpresent 24 hours a day. Patients can still get proper care athe hospital because they have a link with a traumaexpert. The local physician receives assistance from atrauma doctor who watches, listens and offers advicefrom afar via an interactive video link. To use thetelemedicine system, the physician speaks into a wireless,remote headset while the video is rolling and assessespatients while the whole event is broadcast live to anemergency department doctor, who can direct cameras inthe far-away examining room and zoom in on thepatient’s injury. This entire scenario enables the physicianto have an expert consultation and guidance right in theroom. Any patient having straightforward problems suchas flu symptoms, sinus infections, and allergic reactionscan be treated without transfer to a full-service hospital.Symptoms that generally prompt transfer in adults areserious chest pain, abdominal pain, and infections andhigh fever and serious illness in children. Thetelemedicine approach is not to burden the primary careproviders in rural communities; it is intended to act morelike a back up to community doctors.

3.2 Telemedicine Systems in Prison Settings

Health care clinics within a correctional facilityare often designed with limited goals in mind and are notequipped to function as full-service hospitals. In general,many of the inmates are not physically in need ofemergency care but suffer from past neglect, chronicillness and other psychological deficiencies (Spaulding, etal., 1999). Thus, clinical care in correctional settings isnot resource, equipment and personnel intensive. Oftenthere is only 24-hour nursing care with one fieldphysician who works during the day at the facility andwho remains on call the rest of the time. Emergencyrequiring hospitalization is rare, but such situations doarise. Riots, fights, heart attacks, illnesses among the oldinmates often are the most common cause. If severity ofinjury is borderline between sending a person to hospitaland retaining at the facility, telemedicine systems can beof great help. It can ensure expert consultation and best ofcare for the inmate without the accompanying escape-riskand cost of transportation (Raines, et al., 1998).

Two early telemedicine systems implemented incorrectional facilities settings and considered largest involume of consultations handled are the Texas TechUniversity Helath Sciences Center (TTUHSC) and theUniversity of Texas Medical Branch (UTMB). Both wereborn of actions by the Texas legislature in 1993 and areresponsible for providing health care for the inmates ofthe Texas Department of Criminal Justice. This legislationin effect created a health maintenance organization

$10.00 (c) 2000 IEEE 3

Page 4: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

4

t

n

s

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

(Tinstman and McCaughan, 1995). It is makingsignificant differences in the delivery of inmate healthcare, giving providers a new avenue to provide care, aswell as breaking the isolation and stigma associated withpracticing medicine within a prison (Rendleman, 1999).TTUHSC is responsible for the inmate health care in thewestern half of Texas, where the inmate population hasreached more than 30,000. TTUHSC created thedepartment of Correctional Health Care to facilitate thedelivery of inmate health care. Telemedicine isconducted on a weekly basis with the Texas Departmentof Criminal Justice at the Clements Unit in Amarillo, theRobertson Unit in Abilene, the Allred Unit in WichitaFalls, and the Laughan Unit in Fort Stockton, Texas.Amarillo, the pilot site for the telemedicine project, ishome to the Clements Unit, housing 3300 male inmates,and the neighboring women’s unit, housing additional1300 female inmates. The National Commission onCorrectional Health Care since February 1993 hasaccredited the Clements Unit. The infirmary includes a17-bed inpatient facility, emergency room, and multipleclinical settings throughout the unit (Armstrong, 1998).

The Ohio Department of Rehabilitation andCorrections (ODRC) implemented a telemedicine pilotproject in 1995 (Brunicardi, 1998). The project broughttogether the southern Ohio correctional facility inLucasville, the corrections medical center in Columbus,and the Ohio State university medical center also locatedin Columbus. With the two-way interactive video,physicans in one location established audio and videolinks with inmates hundreds of miles away. The ODRCexperienced considerable savings for telemedicine usageAnother telemedicine project in correctional environmentconnects Powhatan Correctional Center (PCC) of theVirginia Department of Corrections and the MedicalCollege of Virginia Commonwealth University (McCue,et al., 1997). The project saved 25% of expenses throughsavings in transportation and medical costs. Non-dollarbenefits included implementing more consistent andtimely treatment of inmates and reducing security risk.

There are other correctional telemedicinesystems as well that are currently in use. A telemedicinesystem connects the Waupaun prison and the Universityof Wisconsin Medical School which are fifty miles apart.The system has already saved by cutting in half thenumber of trips between the prison and the hospital(Appleby, 1995). Similarly, there are nine prisons in Iowathat have access to medical facilities through telemedicine(Zollo, 1999).

In summary, with rising prison health care costsand shrinking state budgets, correctional facilities have todevise innovative ways to provide health care. Of the

0-7695-0493-0/00

.

several methods being explored, telemedicine stands ouprominently. Rabinowitz (1997) estimates that over thelast five years, telemedicine systems have saved theprison health costs by over $125 million. Currently, manyinstitutions are using audio, video and teleconferencinglinks that have been specially installed for telemedconsultations. As the Internet and the related technologiesmature, there is an opportunity to investigate how it canbe adapted to correctional care. This offers a potential tobring down the costs of communications even lower. Bymodeling the workflows and diagnostic processes andcapturing physician expertise, this paper investigates howa Web-centric telemed system can be designed andimplemented.

4. Modeling Telemedicine Workflow

At its most basic level, in the telemedicinecontext, workflow is the automation of medicalexamination, consulting, diagnosis and treatmentprocesses. Workflow applications decompose processesinto a number of steps called tasks. In addition tospeeding up execution of tasks wherever possible inparallel, understanding workflow enables tracking thestatus of tasks in progress. Any workflow model shouldprovide a comprehensive workflow diagram similar toFigure 1 for describing the various paths through which adecision process may flow. This provides aninfrastructure for programmers who later would be codingthe system. An incomplete or an incorrect model couldlead to expensive alterations to the program later. Ifundetected, it might have serious consequences during aactual medical emergency.

The workflow model divides the telemedicineprocess into three phases. These are: i) Examination, ii)Diagnosis and iii) Recommendation/Treatment Plan.These are described below.

Examination

The problem-solving process begins with datacollection from the patient user who has a health problemthat needs to be solved. Typically, when a patient arrivesas an emergency case to the clinic, the patient is firstevaluated to determine if he/she needs to be sent to ahospital on an emergency basis. Subjective data such apatient’s description of symptoms and feelings as well asobjective data like vital sign measurements are collected.Some of the symptoms/criteria that are used in evaluatingsuch emergency includes: if the patient heart rate is below40 or above 60, if the patient is suffering from fever withbody temperature more than 104o F, uncontrollablebleeding and others as shown in the Figure 1. If thepatient meets one or more of these criteria, then the

$10.00 (c) 2000 IEEE 4

Page 5: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

5

t

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

security is notified and the emergency number 911 iscalled. It is standard practice for two custodial officers toaccompany the patient to the hospital to prevent prisonerescapes during the transportation and stay at the hospitalThe resident doctor is also notified of the actions for the

0-7695-0493-0/00

.

follow up diagnostic processes. If the patient, however,does not face a life-threatening emergency, then the nexstep is to perform a complete diagnosis workflow in-house itself.

Figure 1: Correctional NursingWorkflow Model

Y

Life threateningSigns & symptomsSystolic BP <80 | >200Heart Rate <40 | >160Fever >104FUncontrolled BleedingUnconsciousSevere AbdominalP i

Patient walk-in

Emergency?

Notify

Notify Doctor

Call 911

NursingTreatment

DressingIce packingOver the counter MedicinesReassurance

Order RN to:Draw bloodMonitor vital signsIV treatmentKeep in infirmary

Send outtohospital?

MinorSymptoms?

DocumentEnco

Moderateto severesymptoms?

Need ExpertConsultation?

Treat in-houseCardiologist

Neurologist

Orthopedic…

Consult Experts

Unanimousrecommendation

InmateMedicalDB

Negotiate treatmentplan

Treatment Plan

Treat at infirmary in-houseSend to hospital

N

Y

Y

Y

Y

N

N

Y

N

N

N

EXAM

DIAGNOSI

RECVRY&TREATMNT

P

$10.00 (c) 2000 IEEE 5

Page 6: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

6

:

s

,

y

e

e

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

Diagnosis

Diagnosis is the process of identifying the causeof the illness or injury. In simple cases it is often possiblefor the in-house nurse to make a diagnosis without aphysician's help. In case of minor symptoms, such asminor cuts, bruises and mild fever, the diagnosis phase isfairlystraightforward. In such situations, the physician is notcontacted but the nurse treats the patient and releasehim/her to the prison officers. The medical encounter isfiled for record purposes.

When the symptoms rangefrom moderate tosevere, the resident doctor is notified. If the doctor'sdiagnosis indicates a life threatening situation, s/he mightnotify the security and order the patient's transportation toa hospital. On the other hand, if the symptoms are notsevere enough, the doctor might decide to admit thepatient to the prison infirmary. If the doctor is unsure ofthe diagnosis, she/he might want to make a consultationwith outside experts via telemedicine. The doctorprovides these experts with the medical informationregarding the patient. The medical database is also madeavailable to the experts if they want to access it to find outmore information.

Recommendation/Treatment Plan

In a simple scenario, the nurse her/himself willbe able to make an appropriate treatment without havingto contact the physician. Example nursing treatments ofthis category would include: dressing the wound, icepacking, giving over the counter medicines and reassuringthe patient.

In situations where the patient symptoms aremore severe, the physician may follow two possible leadsin arriving at a recommendation. In the first alternative,the physician may conclude that he/she has conclusivelydiagnosed the malady and is certain about the treatmenmethod that would be adopted. If, however, the physicianis unsure about the treatment strategy, he/she may decidto consult other experts in the field using the telemedicineapproach for advice. Once the experts have formed anopinion and recommended a solution, the next step is todetermine if the recommendations from various expertsare unanimous or compatible with one another. In casethe recommendations are consistent, a treatment plan ismade out. If they are not, then the experts negotiate aconsensual treatment plan that is then carried out by theprison physician. The treatment plan might prescribe thetreatment to be administered in-house or the patient mightbe sent to a hospital.

0-7695-0493-0/00

s

t

e

5. Object Modeling for a Web-basedCorrectional Telemedicine System

This section describes strategies for transformingthe workflow model described earlier into an object-oriented architecture that is implementable on theInternet. Several objects have been represented forcompleteness. Some sample ones shown in Figure 2 arefrmSmptms (request object), MedHistory (data object),MetaDiagnos (data object), Dr_A (data object), Dr_B(data object), and frmRcmnd (response object). The dataobjects can reside as part of a single or multiple databaselocated any node in the Internet. Their attributes andcardinal relationships are shown in Figure 2.

Figure 2 Objects in Correctional Telemedicine

The above objects have methods associated withthem that can be invoked when appropriate eventconditions arise during a diagnosis session. For examplethe frmSmptms form has a procedure that can check thevalidity of user input values to the form that is invokedwhen the Submit button is pressed (see Figure 8 in thenext section). It has also a Request method which allowsthe form to bundle all the property values submitted by auser and forward them tothe Internet server. The MetaDiagnosis has Connectionmethods to set up a new connection to a database on anof the node on the Internet and run Structure QueryLanguage (SQL) queries through the use of ODBCdrivers. All results are obtained through the Responsemethod. Figure 3 summarizes the web-based architecturefor the correctional telemedicine system in more detail.

There were several reasons why sucharchitecture was used. First, the suggested objects can brepresented using relational structures which are wellsuited to capture the expert rules of individual doctors. Instandard expert systems rules are represented in the thformat, "IF <conditions> THEN <conclusion>". Forexample, "If <sneeze=yes and eyes=teary> Then<diagnosis=allergy>" is a production rule. In thetelemedicine system, we propose that the relational record

$10.00 (c) 2000 IEEE 6

Page 7: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

7Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

is the logical equivalent of the production rule byadopting a relation of the format R (Condition1,Condition2, …, ConditionN, Conclusion). Thus, theearlier rule can be represented by an instance of the recotype viz., R ( "sneeze=yes", "eyes=teary","diagnosis=allergy"). By using the relational records, itcan be seen that the entire knowledge base of an expecan be stored into the table. In our system a relationatable is used for each expert /consulting doctor. Thestables in our prototype are called as Dr_A and Dr_B. Wehave also a table called MetaDiagnos which containsinformation on which experts are appropriate forconsultation depending different combinations ofpotential symptoms.

The second reason for the proposed architecturis that it is consistent with the Active X objects andODBC technologies currently available for developingInternet applications. Active Server Paging (ASP) comeswith several built in methods for dealing with the server,application, session and other objects. Further, in ASPprogramming model, there is a wide range offunctionality that is accessible to the programmer. ASPallows us to track the state of a user, dynamically generatHTML, output, and take data from forms to be insertedinto a database. All of this functionality makes ASP wellsuited for distributed processing over the Internet that thetelemedicine system demands.

6. Prototype Implementation

Based on the workflow and object modelsdescribed in earlier sections, a prototype has beeimplemented using Microsoft Personal Web Server andAccess database files. Its operation is described belowusing (i) a sample admission and (ii) a consultationscenario.

6.1 Sample Admission: Input of PatientMedHistory

When a prisoner is brought into the correctionalfacility for the first time, a complete medical history andphysical exam of the patient is recorded in the databaseThis helps in future medical decisions. The medicalhistory includes any serious medical condition that thepatient is suffering from e.g. heart disease, diabetesallergies and the like. This ensures that he/she is noadministered any medication that would be harmful in asubsequent medical encounter.

The information is inserted into the table using aweb page form (Figure 3). Once the information has beefilled by the health professional and the Submit button ispressed, the web form executes the corresponding AS

0-7695-0493-0/00

rd

rtl

e

e

e

n

.

,t

n

P

page. This results in setting up the database connectionand inserting the information into the table (Figure 4).

Figure 3 Registration form for taking medical history ofa new inmate

Figure 4 Creating connection object and SQL forinserting record

6.2 Consultation Scenario

Set Conn = Server.CreateObject("ADODB.Connection")Conn.Open("dsn=TempAsp")sql="insert into Master(ID,FName,LName,Age,Sex,Status,Cond1,Cond2,Cond3)"sql=sql + "values ('" + Id + "','" + Fname + "','" + Lname +"'," + Age + ",'" + Sex + "','" + Status + "','" + Cond1 + "','"+ Cond2 + "','" + Cond3 + "')"Conn.Execute(sql)Conn.closeSet Conn = nothing

Figure 5 Steps in the telemedicine consultation

Research

Assessment

Publication

Input

Negotiation

Meta-Knowledge

Local DBUser Requestto the system

INTERNET

$10.00 (c) 2000 IEEE 7

Page 8: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

8

n

s

et

t

n

es

-

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

Each telemedicine consultation session isdivided by the prototype system into five sequentialprocesses: input, assessment, research, negotiation anpublication (Figure 5). Tasks in each of these processesare discussed separately.

Input

Whenever a situation arises that requires medicalintervention, the patient is brought into the infirmarywithin the correctional facility. The advice nurse performsan examination on thepatient. This is when our telemedicine system comes intoplay. The nurse inputs the symptoms that the patient issuffering from. Similar to the patient medical history, thisinput is done on a web form (Figure 6). The prototype asimplemented at this time can accept only three symptoms.In practice, this is sufficient. The possible symptoms havebeen coded into the form as list boxes. Not only does thismake the data entry easier for the nurse evaluator, but alsoensures uniformity in the input data string when the samesymptom has to be entered in different sessions. This isimportant because during assessment phase, the systemmust be able to match symptom description string thatwas input to this form with the symptom string stored inthe knowledge base.

For the purpose of illustration of the input phase,let us assume the nurse has input for an imaginary patientthree symptoms consisting of nose bleeding, light-headedness and bruised face as shown in the figure andpressed the submit button. The control is then passed onto the Assessmment module.

Figure 6 Input form used for diagnosis

While accepting inputs, the form object the system does avalidation check to insure the integrity of the data beingentered. When the patient ID is being entered, the system

0-7695-0493-0/00

d

makes sure that the field is not left blank. Also, thesystem does not allow the user to input characters that arenot valid ID field characters. Users will receive errormessages if incorrect data are entered into the form.

Assessment

Assessment is the second step in the consultationscenario. Once the symptoms have been entered on aASP page, then the Visual Basic server-side scriptsconnect to the MetaDiagnos table in the database andretrieves diagnosis for the particular problem along withthe recommendation. Using the same illustrative scenariomentioned in the Input phase earlier, the system searchein the MetaDiagnos relational table for the record with thethree symptoms of nose bleed, light-headedness andbruised face. Once the matching record is located, thecorresponding value in the Diag attribute for this recorddetermines the diagnosis for these symptoms and makesthe recommendation to use ice packs.

The system also usess the MedHistory table andtakes patient’s medical history into consideration beforegiving a recommendation. For this, the value of thesymptoms gathered from the input are formulated into aSQL statement and executed to identify recommendationsof local and/or distributed expert nodes on the Internet.The system also shows the email links to the other expertsin the query results so that if need be the experts may becontacted for further consultation.

Research

One of the greatest advantages of using a web-basedtelemedicine system is that we are not confined to a localdatabase. We can contact other databases over thInternet and by passing out parameter values, we can gerecommendations from other experts as well. Thus, whenthe situation is more complicated than the local databasecan handle, we contact other expert’s databases and gethe recommendation from them by passing patientsymptoms to these databases.

The methodology for identifying the experts'recommendations is same as the one employed in theAssessment phase. In addition to the symptoms, diagnosisand recommendation attributes, the MetaDiagnos tablehas also attributes with names of expert databases that cabe further contacted for consultation whereverappropriate. The system uses these expert database namas 'links' to make ActiveX Data Object (ADO)connections. Each expert database has a DiagnosisRecommendation table which have heuristics stored asdescribed in Section 5. This allows the user to receiveindependent diagnosis and recommendations from

$10.00 (c) 2000 IEEE 8

Page 9: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

9

t

e

r

he

e

Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

multiple experts distributed over the Internet for the sameset of symptoms online.

Negotiation

We can have situations where we contact morethan one expert’s databases to get the recommendation.In such cases, we can be faced with results where expertrecommendations do not agree with each other. In thiscase, we need to negotiate among the experts and comeup with a consensus on the diagnosis and treatment plan.Currently work is being done in choosing the appropriatetheoretical framework and related algorithms towardstheir implementation into the prototype.

Publication

Finally, all the recommendations are publishedon the Internet browser interface for the user to see andtake appropriate medical action. Following is an examplesystem response (Figure 7). It can be seen from thesample response, that the local node recommendationfrom the MetaDiagnos table for the symptoms of nosebleeding, red eyes and back pain is heat stroke. However,the response screen also recommends the user to consultwo other experts, viz., Dr_A and Dr_B. Since these aredisplayed as hyperlinks, all the user has to do to receivethese expert consultations is to click on these hyperlinks.

Figure 7 Web page showing results of diagnsis

User Authorization

One other area that our system has capabilities isthe security aspect of the telemedicine system. Thepersonal medical information has always consideredprivate from the legal point of view. This is especiallyimportant to the prison system where the population has adisproportionately large number of ailments such asHepatitis, Aids and other mental illnesses. When data ismade available on the Internet, the potential for hacking is

0-7695-0493-0/00

t

even higher and the system developers should implemenfoolproof firewalls.

In our prototype, only authorized users can enterthe system. This is done by having the input pagedynamically created when the user name and a validpassword has been read by the system and verified in thLogin table. Thus, since the information does not evenexist in a format that could be read and is only createdafter the verification of the user input information, theprototype provides a great measure of security for thedatabase. The following screen shot shows the useauthorization form.

Figure 8 Web page showing the administrator's screen

7. Summary and Future Challenges

The workflow and object models discussed inthis paper should provide a developer with sufficientknowledge to design a Web-centric telemedicine system.Our experience with the prototype shows the concept ofimplementing a telemedicine on the Internet astechnically feasible and the use of ActiveX technologyquite powerful. Enhancements of the prototype arecurrently under way to incorporate additional expertdatabase nodes and increase the rules represented in tsystem. The interface is being expanded to incorporateaudio and video. Another important aspect of the systembeing augmented is the power of its negotiation module.Finally, software agents will be added so that the systemcan search proactively and provide preventive advice topatients and physicians before health problems becomeout of hand. Prisoners may not have to be hospitalized ifaccurate and expert diagnosis can catch and treat thproblem quickly. Telemedicine technology has thepotential to change the face of health care for the betterand reducing the cost of health care services for thecorrectional facilities everywhere.

$10.00 (c) 2000 IEEE 9

Page 10: A Web-Based Correctional Telemedicine System with ... · A Web-Based Correctional Telemedicine System with Distributed Expertise ... This paper discusses the workflows ... 2000 IEEE

10Proceedings of the 33rd Hawaii International Conference on System Sciences - 2000

Telemedicine, like most other advancedinformation and communications technologies, dependson complex technical and human infrastructures thatoperate both within discrete institutions and acrossorganizational and geographic boundaries. The harshrealities of the current health-care environment sometimeshinder the innovative strides being made by thevisionaries in the field of telemedicine. Those responsiblefor creating information and telecommunications systemsand programs face a bewildering and constantly changingarray of hardware and software options, many of whichare not tailored to health care uses. Different telemedicineapplications may involve different combinations oftechnologies and because each Telemedicine programreflects different organizational objectives, theinfrastructure on which they depend will vary from placeto place. Despite these and many other challenges toovercome, the future of telemedicine looks verypromising. The last ten years have seen a steady increasein the number of telemedicine projects throughout theU.S. and internationally. As the benefits of telemedicinesystems become more manifest, more funding becomesavailable, and as the technology costs continue to decline,the use of telemedicine systems is bound to become anormal part of our daily lives.

______Acknowledgement: The authors wish to thank TeddyPujji for his invaluable assistance with theimplementation of the prototype discussed in this paper.

REFERENCES

[1] Appleby, C., "A prison plugs in," Hospitals and HealthNetworks, 69(21), 1995, pp. 56-58.

[2] Armstrong, M., Telecommunications for HealthProfessionals, Springfield: New York, 1998.

[3] Binius, T., "Moving Pictures", American Medical News,42(12), 1999, pp 27-31.

[4] Brunicardi, B.O., "Financial analysis of savings fromtelemedicine in Ohio's prison system,", TelemedicineJournal, 4(1), 1998, pp. 49-54.

[5] Franczyk, A., “Telemedicine connects the area’s doctors,”Business First of Buffalo, 15(6), 1998, pp. 40-51.

[6] Grigsby, J. and Sanders, J.M., "Telemedicine: Where it isand where it's going?," Annual Internist Medicine, 129(2),1998, pp.123-127.

[7] Gilbert, F., "Legal issues in the use of the Internet forhealth care", Telemedicine Journal, 5(1), 1999, pp 34.

[8] Larkin, M., "Mt. Everest telemedicine expedition trackedon the Web," Lancet, May, 1999.

[9] Lugg, D.J., "Telemedicine: Have technological advancesimproved health care to remote Antarctic populations?,"International Journal of Circumpolar Health, 57(1), 1998,pp. 682-5.

[10] Mason, C., “The telemedicine breakthrough,” America’sNetwork, 102(2), 1998, pp. 38-42.

0-7695-0493-0/00

[11] McCue, M., Hampton, C., Marks, T.K., Fisher, E., &Parpart, F., "The case of Powhatan CorrectionalCenter/Virginia Department of Corrections and VirginiaCommonwealth Univeristy", Telemedicine Journal, 3(1),1997, pp. 11-17.

[12] Mun, S.K., Levine, B., Cleary, K., & Dal, H., "Deployableteleradiology and telemedicine for the US military,"Computing Methods and Programs in Biomedicine, 57(1),1998, pp. 21-27.

[13] Rabinowitz, E., "Correctional helath care: facilities look toreduce costs and improve the quality of inmate care,"Health Care Business Digest, 2(11), 1997, pp. 32-36.

[14] Raines, K., Toenjes, J. and Liebgott, A., "Telemedicine: it'snot just for rural jails," American Jails, 12(2), 1998, pp. 9-21.

[15] Reichale, D., “Telemedicine Privacy Hazards DivideExperts,” National Underwriter Property, Casualty-Risk &Benefits Management, 14(1), 1998, pp. 14-21.

[16] Rendleman, John., “Texas finds telemedicine cure forbudget woes.” PC Week, 16(11), 1999, pp. 49-51.

[17] Snyder, K., “Telemedicine: the new frontier,” Drug Topics, 140(12), 1997, pp. 60-65.

[18] Spaulding, A., Greene, C., Davidson, K., Scheidermann, M.& Rich, J., "Hepatitis C in state correctional facilities,"Preventive Medicine, 28(1), 1999, pp. 92-100.

[19] Stoloff, P.H., Garcia, F.E., Thomason, J.E., & Shia, D.S.,"A cost-effectiveness analysis of shipboard telemedicine,"Telemedicine Journal, 4(4), 1998, pp. 293-304.

[20] Swanson, J.W., Morrissey, J.P., Goldstrom, I., ?rudolph, L.& Manderscheid, R.W., "Funding, expenditures, andstaffing of mental helath services in state adult corrctionalfacilities: United States," Mental Health Statistics Note,July, 1993, pp. 1-20.

[21] Thorburn, K.M., "Health care in correctional facilities,"Western Journal of Medicine, 163(6), pp. 560-564.

[22] Tinstman, T.C. and McCaughan, W.T., "Telemedicine inrural correctional facilities,", Texas Journal of RuralHealth, 14(2), 1995, pp.70-77.

[23] Vitucci, N., Health and Wellness Education, Correct Care,13 (2), Spring 1999, pp 1-10.

[24] Zollo, S.A., "Correctional facilities in Iowa usetelemedicine," Health Connections, No.6, 1999, pp. 9.

$10.00 (c) 2000 IEEE 10