Specialist Physicians' Knowledge and Beliefs about Telemedicine: A Comparison of Users and Nonusers of the Technology

487

TELEMEDICINE AND e-HEALTHVolume 13, Number 5, 2007© Mary Ann Liebert, Inc.DOI: 10.1089/tmj.2006.0091

Original Research

Specialist Physicians’ Knowledge and Beliefs aboutTelemedicine: A Comparison of Users and Nonusers of

the Technology

PHOEBE LINDSEY BARTON, Ph.D.,1 ANGELA G. BREGA, Ph.D.,2PATRICIA A. DEVORE, B.S.,2 KEITH MUELLER, Ph.D.,3 MARSHA J. PAULICH, M.S.P.H.,2

NATASHA R. FLOERSCH, B.A.,2 GLENN K. GOODRICH, M.S.,2SYLVIA G. TALKINGTON, R.N., B.A.,2 JEFF BONTRAGER, M.S.P.H.,2

BILL GRIGSBY, Ph.D.,4 CAROL HRINCEVICH, M.A.,2 SUSANNAH NEAL, M.A.,2JEFF L. LOKER, B.S.,2 TESFA M. ARAYA, M.S.,2 RACHAEL E. BENNETT, M.A.,2

NEIL KROHN, Ph.D.,2 and JIM GRIGSBY, Ph.D.2

ABSTRACT

Telemedicine as a technology has been available for nearly 50 years, but its diffusion hasbeen slower than many had anticipated. Even efforts to reimburse providers for interactivevideo (IAV) telemedicine services have had a limited effect on rates of participation. The re-sulting low volume of services provided (and consequent paucity of research subjects) makesthe phenomenon difficult to study. This paper, part of a larger study that also explorestelemedicine utilization from the perspectives of referring primary care physicians andtelemedicine system administrators, reports the results of a survey of specialist and subspe-cialist physicians who are users and nonusers of telemedicine. The survey examined self-as-sessed knowledge and beliefs about telemedicine among users and nonusers, examining alsothe demographic characteristics of both groups. Statistically significant differences werefound in attitudes toward telemedicine between users and nonusers, but in many respectsthe views of the two groups were rather similar. Physicians who used telemedicine were awareof the limitations of the technology, but also recognized its potential as a means of provid-ing consultation. Demographic differences did not explain the differences in the knowledgeand beliefs of user and nonuser consultant physicians, although some of the differences maybe explained by other aspects of the professional environment.

1Department of Preventive Medicine and Biometrics, University of Colorado at Denver and Health Sciences Cen-ter, Denver, Colorado.

2Division of Health Care Policy and Research, Department of Medicine, University of Colorado at Denver andHealth Sciences Center, Aurora, Colorado.

3Nebraska Center for Rural Health Research, and Department of Preventive and Societal Medicine, University ofNebraska Medical Center, Omaha, Nebraska.

4Department of Anthropology and Sociology, Eastern Oregon University, LaGrande, Oregon.

INTRODUCTION

TELEMEDICINE TECHNOLOGY has been avail-able since at least 1959,1 but the rate of

adoption of the technology as a means of pro-viding health services has been slow. Its use re-mains concentrated in certain geographic areas,and is limited to a small percentage of thephysician population. In the mid-1990s, it wasthought that the capacity of telemedicine to in-crease access to healthcare would make it at-tractive to providers and patients alike, but theanticipated proliferation of telemedicine failedto materialize. This state of affairs was widelyattributed to concerns about liability, interstatelicensure of providers, and the unavailabilityof coverage. Medicare payment policy wasviewed as a particularly important obstacle.2–8

Experience with telemedicine, however, sug-gested that factors other than these also mightplay an important role in constraining diffu-sion. In some interactive video (IAV) programs,even the availability of payment for specialistphysicians to provide consultation failed to en-courage wider participation. Attention subse-quently turned to a number of human and or-ganizational variables affecting the acceptanceand dissemination of new technologies.9–12 Thereasons identified for nonuse of telemedicineare many: the equipment often is thought to beinconvenient or is inconveniently located,13

sometimes the technology is unreliable,13–15

many providers believe that their participationrequires too much time,13,16,17 reimbursementgenerally is inadequate or unavailable,18 andthe technology is considered by some to be notyet equivalent to in-person care.18 There have,however, been no large-scale surveys ofproviders to assess their stated reasons for ei-ther embracing or avoiding the technology.

We were interested in determining what dif-ferences in attitudes and knowledge might dis-tinguish users and nonusers of telemedicine. Inthis study of physician knowledge and beliefsabout telemedicine, we hypothesized that spe-cialist and subspecialist physicians who usetelemedicine (1) are likely to be younger thannonusers, (2) are likely to be more recent med-ical school graduates, (3) have convenient ac-cess to telemedicine facilities, and (4) are ear-

lier adopters of new technologies. Apart fromthese hypotheses, the study, which representsa preliminary exploration of why consultingphysicians do, or do not, use telemedicine, wasdescriptive in nature.

MATERIALS AND METHODS

Sampling frame

The objective of the study was to examine theuse of telemedicine among consulting physi-cians. The specialists and subspecialists in-cluded in the survey were recruited by twostudy sites: the University of Colorado at Den-ver and Health Sciences Center (UCDHSC),and the Center for Rural Health Research at the University of Nebraska Medical Center(UNMC). As described below, network andphysician recruitment used slightly differentapproaches at the two sites. The data from theColorado and Nebraska samples were aggre-gated for analysis.

Project staff at UCDHSC targeted 17telemedicine networks for study recruitment.These networks were identified from the largernational sample of programs that participatedin Grigsby and Brown’s 1998 study of telemed-icine activity.5 From this sample, 17 relativelylarge networks that had been in operation forseveral years and had well-established IAVteleconsultation programs were targeted for re-cruitment. Of the nine networks that agreed toparticipate, eight were based in urban aca-demic medical centers with affiliated, mostlyrural, healthcare organizations. The ninth net-work was a private, nonprofit hospital.

For each participating network, permissionto contact consulting physicians on hospitalmedical staff rosters was obtained from thehead of the medical staff and the telemedicineadministrator. Telemedicine users and non-users were identified as such by the programadministrator based on a history of use (ornonuse) of the available technology. Althoughrecruitment employed administrator identifi-cation to classify physicians as users ornonusers of the technology, all study analyseswere based on the physicians’ self-report of

BARTON ET AL.488

use. Consulting physicians then were sampledrandomly from the lists of users and nonusers.Out of the 1,461 physicians targeted for re-cruitment from these nine networks, completedsurveys were received from 544, a response rateof 37.2% for the UCDHSC sample.

Project staff at UNMC sent the survey to atotal of 268 specialist physicians licensed topractice in the state of Nebraska. All physiciansaffiliated with Nebraska’s largest telemedicinenetwork (the Mid-Nebraska Telemedicine Net-work) were targeted for recruitment, as was arandom sample of 1/3 of the remaining spe-cialists in the state. Data for the 152 consultingphysicians (56.7%) who responded to the sur-vey were included with the UCDHSC data foranalysis. As with the UCDHSC sample, usersand nonusers were identified by self-report foranalytic purposes.

Survey instruments and procedures

UCDHSC investigators sent paper-and-pen-cil surveys to all potential participants by reg-ular mail. Surveys were completed betweenJune 2001 and February 2003. The survey in-struments also were made available for com-pletion by means of a Web-based database. Upto two follow-up requests were sent by mail toindividuals who were slow to respond (morethan 2 weeks beyond the initial request and thefirst follow-up invitation). After two follow-uprequests had been sent to providers with no re-sponse, project staff returned to the samplingpool and randomly selected additional indi-viduals who had not previously been invitedto participate. Physicians who took part in thesurvey received a small honorarium ($40) fortheir time, which was estimated at 15 to 30 min-utes.

At UNMC, data collection began in July 2000and continued through October 2000. An ini-tial mailing was followed 1 week later with apostcard. In August, the survey was sent a sec-ond time to nonresponders. Follow-up phonecalls and a final mailing of the complete sur-vey were conducted a month later. Participantswere not compensated for their participation.

For the analyses reported in this paper, twoinstruments were used: (1) a survey for spe-

cialist and subspecialist physicians who weretelemedicine users (the Telemedicine User Sur-vey), and (2) a survey for specialist/subspe-cialist physicians who were not telemedicineusers (the Telemedicine Nonuser Survey).

Areas of investigation included the follow-ing: (1) demographic and practice information,(2) physician attitudes toward and knowledgeof telemedicine, (3) perceived advantages forpractice, (4) telemedicine and referral patterns,(5) perceived convenience/inconvenience oftelemedicine, (6) effects of the technology onpatients, (7) perceived financial investment, (8)concerns regarding malpractice and liability,and (9) reimbursement issues, especially re-lated to Medicare. For all analyses, physicianswere identified as users and nonusers based onself-reported utilization of telemedicine tech-nology.

The majority of attitude and knowledge sur-vey questions were presented on a four-pointLikert scale: strongly agree, agree, disagree, andstrongly disagree. Items were analyzed using thefour-point scale, and subsequently were col-lapsed into dichotomous agree/disagree cate-gories. We conducted descriptive analyses ofthe sample in addition to comparisons of userswith nonusers. Analytic methods includedmeasures of association (correlations), t-tests, z-tests, chi-square, and logistic regression. Wecontrolled for multiple comparisons by usingthe Holm test.19

RESULTS

Response rate and description of the physician samples

The final sample reported in this paper con-sisted of 696 consulting physicians. Of theseparticipants, 202 were users and 494 werenonusers of telemedicine. This sample size rep-resents 40.3% of the physicians targeted for re-cruitment in the study across the two sites. Inthe UCDHSC sample, response rates wereslightly higher among users (40.5%) thannonusers (35.9%). However, because projectstaff at the UNMC did not know in advance theuser/non-user status of the physicians targeted

SPECIALIST PHYSICIANS’ KNOWLEDGE ABOUT TELEMEDICINE 489

for enrollment in their sample, it is not possi-ble to compute the response rate for users andnonusers across sites. Response rates were notsignificantly different statistically by specialty,although cardiology and dermatology wererelatively overrepresented among the telemed-icine users, as might be expected given thatthese specialties have typically been more fre-quent users of the technology.5 Small groupsizes and a large number of specialties limitedour ability to detect differences across special-ties. The distribution of physicians by spe-cialty/subspecialty may be found in Table 1.

Demographic and practice characteristics ofthe sample are provided in Table 2. Across thesample, 21% of all respondents were female, andthe median age was 48.0 years. The mean num-ber of years practicing medicine was 18.4; themean number of years practicing in the currentcommunity was 11.7; the mean number of yearssince medical school graduation was 21.8. Noneof these demographic variables was associatedwith a greater likelihood of using telemedicine,contradicting our first and second hypotheses.The only significant differences found were inpractice site variables: a higher proportion oftelemedicine users practiced in community

health centers (6.5% vs. 1.4%, p � 0.0007), and ahigher proportion of nonusers were in privatepractice (44.5% vs. 28.5%, p � 0.0001).

Users’ experience with telemedicine

A series of survey questions asked physicianswho had used telemedicine to report on their ac-tual utilization experiences (see Table 3). Use oftelemedicine technology was fairly modestamong these physicians. On average, users inthe sample had received nine referrals for tele-consultations in the preceding 6 months. Nearly64% felt they had not used telemedicine enoughto make it a regular part of their practices.

The most common uses of telemedicine iden-tified by the 202 telemedicine-using specialistphysicians were diagnosis (58%) and patientfollow-up (53%). Less common uses includedcontinuing medical education (40%), providingsecond opinions (28%), and chronic diseasemanagement (23%). IAV telemedicine was themodality most frequently used by specialistphysicians (84%), which was expected in thissample as IAV was the predominant mode ofthe participating telemedicine networks.Thirty-two percent had used shared computer

BARTON ET AL.490

TABLE 1. DISTRIBUTION OF PHYSICIANS IN THE SAMPLE GROUPS BY SPECIALTY

Telemedicine users Telemedicine nonusers

Specialty n % of Total sample n % of Total sample

Anesthesiology 1 0.50 2 0.41Cardiology 28 13.86 23 4.67Dermatology 14 6.93 16 3.25Emergency medicine 5 2.48 21 4.27Family practice 9 4.46 25 5.08Internal medicine 3 1.49 11 2.24Neurology 16 7.92 20 4.07Obstetrics/gynecology 5 2.48 17 3.46Occupational medicine 0 0.00 1 0.20Oncology 8 3.96 37 7.52Ophthalmology 0 0.00 18 3.66Orthopedics 9 4.46 37 7.52Pathology 2 0.99 4 0.81Pediatrics 17 8.42 30 6.10Preventive medicine 0 0.00 1 0.20Psychiatry 20 9.90 62 12.60Radiology 2 0.99 19 3.86Rehabilitation medicine 6 2.97 7 1.42Surgery 21 10.40 56 11.38Other medical subspecialty 36 17.82 85 17.28Totals 202 492.a

aInformation about medical specialty was missing for two participants.

screen images (in real-time) while communi-cating by telephone. Only 20% had used store-and-forward telemedicine.

Access to telemedicine equipment, and thetechnical capabilities of the telemedicine system,play an important role in specialists’ use of thetechnology. Although nearly half (48%) re-ported that technical problems with equipmentoccasionally interfered with consultations,nearly two-thirds (64%) stated that their use oftelemedicine had increased over time. Only 14%indicated that the telemedicine system was ren-dered inconvenient by other activities using thesystem during times when the providers wouldlike to offer services (e.g., administration, con-tinuing education programs). Ninety-one per-cent reported adequate technical assistance dur-ing consultations, and 87% reported that thesound quality was adequate for clinical pur-poses—an especially important consideration.Although most participants were satisfied withthe technical quality of telemedicine services,they were not satisfied with Medicare coveragefor these services. Only about one-quarter (26%)of respondents agreed that Medicare reim-bursement for telemedicine usage was adequatefor their level of participation.

A substantial majority of consulting physi-

cians indicated that the telemedicine technologyavailable to them was well suited to the needsof their patients (74%), and reported that theywere satisfied with the quality of patient care at-tainable through the use of telemedicine (83%).However, only a minority of respondents (38%)indicated that most of the cases on which theyconsult could be handled by means of telemed-icine alone. Consistent with this, opinions varied on the effectiveness of telemedicine forpatient examination and diagnosis. About one-third of respondents (32%) stated that they couldconduct a thorough physical exam of the patientusing telemedicine. Two-thirds (66%) reportedthat they found IAV telemedicine more accept-able for rendering second opinions or offeringinformal consultations than for diagnosing newpatients. Although this group of 202 respon-dents classified themselves as users of telemed-icine, nearly two-thirds (64%) stated that theydo not use telemedicine frequently enough tomake it a regular part of their practices.

Telemedicine knowledge reported by users and nonusers

Users and nonusers differed in their self-re-ported telemedicine knowledge. Users were

SPECIALIST PHYSICIANS’ KNOWLEDGE ABOUT TELEMEDICINE 491

TABLE 2. DEMOGRAPHIC AND PRACTICE CHARACTERISTICS OF THE SAMPLE

Variable

Gender Male Female Male Female79.1% 20.9% 77.9% 22.1%

Mean SD Mean SD

Age 47.6 10.0 48.2 10.1Years since medical school graduation 21.6 10.6 21.9 10.4Years practicing medicine 18.1 10.9 18.5 11.6Years practicing in community 12.1 9.5 11.5 9.8Number of states in which 2.0 1.4 1.9 2.5

physician practicesNumber of patients in physician’s practice 3,800 4,145

% Affirmative % Affirmative

Practice site � hospital 80%.0 71.5% (ns)Practice site � community health center 6.5% 1.4% (p � 0.0007)a

Practice site � rural health clinic 4.5% 3.2% (ns)Private practice 28.5% 44.5% (p � 0.0001)a

aThese two variables are the only ones that remained significantly different at an equivalent p � 0.05 after the Holmtest of significance was applied.

Users (n � 202) Nonusers (n � 494)

more likely to describe themselves as “some-what knowledgeable” or “knowledgeable”about telemedicine, whereas nonusers de-scribed themselves as “not at all knowledge-able” to “somewhat knowledgeable” abouttelemedicine (p � 0.0001; see Table 4). Signifi-cant differences (p � 0.0001) also were found inthe extent to which respondents agreed withthe statement, “I do not know enough abouttelemedicine technology and its applications touse it in my practice.” Only 18.6% of usersagreed with this statement, whereas 58.9% ofnonusers agreed (OR � 0.15; CI � 0.11–0.24;p � 0.0001). Even among the users who agreedthat they were knowledgeable, use of telemed-icine was modest (an average of 7.1 times in theprevious 6 months). Although the utilizationrate was low even among knowledgeablephysicians, it is important to note that the

physicians included in this study had a limitednumber of referrals. The low rate of referralsmakes it difficult to assess the relationship be-tween knowledge and utilization rates.

The two groups differed with respect to theirsources of information about telemedicine. Ap-proximately 1 in 10 (9.8%) nonusers reportedthat formal telemedicine training was a source ofinformation for them, as opposed to 21.5% ofusers (p � 0.0001). Nonusers (16.1%) were sig-nificantly more likely (p � 0.0001) to reportmass media as a source of information than wereusers (1.5%). Users were significantly morelikely (p � 0.0001) than nonusers to report thatsources of information other than those listedin the survey were important to them (33% vs.12.6%). Other sources mentioned included a lo-cal telemedicine program, physician mentors,and on-the-job training. Differences between

BARTON ET AL.492

TABLE 3. SPECIALIST TELEMEDICINE UTILIZATION (n � 202)

Variable % Who agree

Most common uses of telemedicineDiagnosis 58.19Follow-up 52.54Continuing medical education 39.55Second opinion 27.68Chronic disease management 23.16

Type of telemedicine modality usedInteractive video 83.52Shared computer screen images with audio 32.39Store and forward 19.89

Access to and technical capacity of telemedicineAccess limited by other uses of system 13.66Inconvenience greater than benefits 30.27Telemedicine use has increased over time 63.48Technical problems with equipment interferes with consultations 48.39Sound quality adequate for clinical purposes 87.29Adequate technical assistance available during consultations 91.16Most of the consultations I do could be accomplished by using 38.33

interactive videoAppropriateness of use with patients

Satisfied with patient quality of care using telemedicine 82.94Telemedicine technology available to me is well-suited to the needs 73.77

of my patientsIn most cases, I can conduct a thorough patient physical exam using 31.58

telemedicineInteractive video is more acceptable for second opinions or 65.92

informal consultations than for diagnosing new patientsReimbursement

Current Medicare reimbursement rate for consultations is adequate 26.00for my level of participation

Regularity of use of telemedicineI do not use telemedicine enough to make it a regular part of my 63.74

practiceAverage number of times a patient has been referred for a 9.15

telemedicine consultation in the last 6 months

groups with respect to other sources of knowl-edge (e.g., colleagues, medical literature, med-ical or postgraduate training, professional as-sociation meetings and conferences, grandrounds, and electronic media) were not signif-icant.

Beliefs about telemedicine reported by usersand nonusers

Specialists were asked about their beliefs re-garding telemedicine in several areas: patientcare; time and convenience of telemedicine use;whether they considered themselves earlyadopters of technology; concerns about licen-sure, credentialing and malpractice; reim-bursement; and other factors.

As illustrated in Table 5, a higher proportionof nonusers believe patients are likely to getbetter care in person than through telemedicine(85.2% vs. 71.1%, OR � 0.43, CI � 0.29–0.64,p � 0.0001). Respondents differed significantlyin their beliefs about the necessity of a patient’spresence for an adequate physical exam.Eighty-seven percent of nonusers thought a pa-tient’s physical presence was necessary for anadequate examination, whereas only 69.7% ofusers believed that the patient’s presence wasnecessary (OR � 0.33, CI � 0.22–0.50, p �

0.0001) (see Table 6.) Nearly twice the propor-tion of users to nonusers would considertelemedicine for initial office visits, and moreusers than nonusers (72.9% vs. 58.2%, OR �1.93, CI � 1.33–2.82, p � 0.0006) believe tele-medicine might be effective for acute non-emergency care.

Time and convenience in telemedicine

Respondents differed in their beliefs aboutthe time involved in telemedicine, and in theextent to which telemedicine was convenient.Statistically significant differences were foundin respondents’ willingness to put up withsome inconvenience to use telemedicine, withtheir assessment of the convenience/incon-venience of telemedicine facilities, and with the importance of reduced travel time thattelemedicine affords for specialist consultants.Users reported greater willingness to toleratesome inconvenience (OR � 2.88, CI � 2.02–4.10,p � 0.0001), that telemedicine facilities wereconvenient for their use (OR � 8.56,CI �4.59–15.95, p � 0.0001), and that the ability toreduce their travel time by using telemedicinewas important to their practices (OR � 2.97,CI � 1.90–4.63, p � 0.0001). There also was asignificant difference in beliefs about the level

SPECIALIST PHYSICIANS’ KNOWLEDGE ABOUT TELEMEDICINE 493

TABLE 4. KNOWLEDGE ABOUT TELEMEDICINE, USERS AND NONUSERS

Nonusers UsersVariable n � 494 n � 202 p valuea

Knowledge sourceColleagues 55.86% 56.50% 0.9324Medical literature 34.94% 25.50% 0.0189Formal telemedicine training 9.83% 21.50% 0.0001a

Medical or postgraduate training 11.72% 5.50% 0.0157Grand roundsMass media 16.11% 1.50% 0.0000a

Professional meetings/conferences 25.10% 17.00% 0.0211Electronic media 15.90% 7.50% 0.029Other sources # 12.55% 33.00% 0.0000a

Knowledge level about telemedicine 1.84 2.53 0.0000a

1 � not at all; 2 � somewhat;3 � knowledgeable; 4 � very

Agree that “I do not know enough about 58.93% 18.59% 0.0000a

telemedicine technology and itsapplications to use it in my practice”

aThese variables were found to be significant at the p � 0.05 level after the Holm test of significance was applied.Other sources of knowledge about telemedicine were reported as: local telemedicine program at my institution;

name of a particular physician mentor; on the job learning; local hospital; informal conversations; reading ECHO; es-tablished a program for prison telemedicine use; consultation with a friend who makes telemedicine equipment; grantactivities.

of disruption of office routine associated withscheduling telemedicine appointments, withnonusers more likely to see this as a problemthan users (67.4% vs. 48.9%, p � 0.0001). Thesedata provided support for our third hypothe-sis.

Early adoption of new technologies

More than three-quarters (77.7%) of users de-scribed themselves as early adopters of new

technologies, whereas 60.4% of nonusers char-acterized themselves this way, a difference thatremained significant (p � 0.003) after applica-tion of the Holm test, providing modest sup-port for our fourth hypothesis.

Concerns about licensure, credentialing, and malpractice

The practice of medicine is governed by in-dividual states’ Medical Practice Acts. Physi-

BARTON ET AL.494

TABLE 5. SPECIALISTS’ BELIEFS ABOUT TELEMEDICINE

Nonusers UsersVariable n � 494 n � 202 p valuea

Beliefs about patient carePatient likely to get better in-person care 85.15% 71.07% 0.0000a

Would consider telemedicine use for follow-up care 62.00% 73.74% 0.0054Telemedicine likely to be more effective for emergent care 43.76% 55.43% 0.0086Telemedicine more likely to be effective for chronic condition 87.34% 93.26% 0.0279

managementWould consider telemedicine for initial office visits 22.27% 46.47% 0.0006a

Telemedicine might be effective for postsurgical follow-up 57.05% 71.76% 0.0008Telemedicine might be effective for acute nonemergency care 58.21% 72.93% 0.0006a

The patient’s presence is necessary for an adequate physical exam 87.34% 69.70% 0.0000a

Time and convenience in use of telemedicineUse of telemedicine would not be an effective use of time 52.24% 38.97% 0.0021Would put up with some inconvenience in order to use telemedicine 44.78% 70.00% 0.0000a

Telemedicine facilities are convenient for use 64.31% 93.91% 0.0000a

Scheduling telemedicine appointments would be disruptive to 67.42% 48.94% 0.0000a

office routineReduced travel for consultants that is possible with the use of 64.38% 84.29% 0.0000a

telemedicine is importantEarly adopter of technology

Respondent views self as an early adopter of technology 60.37% 77.66% 0.0030a

Concerns about licensure, credentialing, malpracticeConcerned about liability issues if telemedicine is used 66.11% 50.00% 0.0001a

Use of telemedicine would increase the risk of malpractice suits 50.44% 28.79% 0.0000a

Credentialing and licensure issues discourage telemedicine use 70.35% 33.71% 0.0000a

ReimbursementCompensation for use of telemedicine should be on a par with 78.90% 64.74% 0.0003a

in-person treatmentOther factors influencing use of telemedicine

Use of telemedicine would expand network of colleagues 68.00% 44.62% 0.0000a

Colleagues influence use of such technologies as telemedicine 69.61% 48.99% 0.000a

Would use interactive telemedicine if it were available in 66.95% 80.15% 0.0005a

their officesPrefer use of store and forward over interactive telemedicine 38.44% 24.74% 0.0014a

Percent of patient population that provider believes could be 23.33% 29.19% 0.0124treated using store and forward technology

Percent of patient population that provider believes could be 25.22% 40.44%a

treated using interactive technologyPrefer interactive over store and forward technology where 1 � 3.20 3.55 0.0007a

strongly agree and 5 � strongly disagreeDislike loss of personal contact with patient that results from use 83.72% 65.50% 0.0000a

of telemedicineMore research on telemedicine is needed 59.28% 37.37% 0.0000a

aThese variables were found to be significant at the p � 0.05 level after the Holm test of significance was applied.

cians whose practices cross state lines typicallyare licensed by all states in which they practice.The use of telemedicine, which readily permitsproviders to see patients in many other geo-graphic jurisdictions, has raised questionsabout licensure.20–24 We asked physiciansabout their concerns regarding across-state-lines issues of licensure, credentialing, and malpractice. Significant differences exist in re-sponses between specialists who use telemed-icine and those who do not. A higher propor-tion of nonusers believed that credentialingand licensure issues discourage telemedicineuse (33.7% vs. 70.4%, OR � 0.21, CI � 0.15–0.31, p � 0.0001). The same relationship heldfor individuals who reported concerns about li-ability issues if telemedicine is used (66.2% vs.50.0%, OR � 0.51, CI � 0.37–0.72, p � 0.0001).Nonusers also were more likely to believe that

the use of telemedicine would increase the riskof malpractice suits (28.8% vs. 50.4%, OR �0.40, CI � 0.28–0.57, p � 0.0001). These resultsare summarized in Table 6.

Reimbursement

In response to a question about disparitiesbetween in-person care and teleconsultation,more nonusers than users (78.9% vs. 64.7%, p �0.0003) agreed that compensation for telemed-icine should be on a par with in-person treat-ment.

Other factors influencing the use of telemedicine

Several other beliefs are related to the use oftelemedicine by specialist and subspecialistphysicians. Although both users and nonusers

SPECIALIST PHYSICIANS’ KNOWLEDGE ABOUT TELEMEDICINE 495

TABLE 6. LIKELIHOOD OF TELEMEDICINE USE AS A FUNCTION OF ATTITUDES TOWARD TELEMEDICINE

Attitude statement Odds ratio 95% CI p value

Beliefs about patient careWould consider use of telemedicine use for application of 2.28 2.08–4.41 0.0001

initial office visitTelemedicine might be effective or is effective for 2.01 1.07–3.75 0.029

chronic condition managementTelemedicine might be effective or is effective for 1.91 1.30–2.81 0.0009

postsurgical follow-upTelemedicine might be effective or is effective for 1.93 1.33–2.82 0.0006nonemergency care

I do not think an adequate physical examination can be 0.38 0.22–0.50 0.0001conducted without the patient being present physically

Time and convenience in use of telemedicineIf interactive video were available in my office, I 2.04 1.36–3.05 0.0005

would use itI am willing to put up with some inconvenience for my 2.88 2.02–4.10 0.0001

patients to receive telemedicine servicesIf I must present the patient for teleconsultation, current 8.56 4.59–15.95 0.0001

telemedicine equipment location is convenient for meI would use telemedicine if it allowed me to significantly 2.97 1.90–4.63 0.0001

reduce the time I spend traveling to other communitiesto see patients

Early adopter of technologyI am generally one of the first among colleagues to 2.28 1.56–3.34 0.0001

try new technologiesKnowledge statement

I do not know enough about telemedicine to use it 0.15 0.11–0.24 0.0001in my practice

Licensure, credentialing, and malpracticeI am concerned with possible liability issues associated 0.51 0.37–0.72 0.0001

with use of telemedicineUse of telemedicine would increase my risk of being 0.40 0.28–0.57 0.0001

sued for malpracticeCredentialing and licensure for telemedicine are 0.21 0.15–0.31 0.0001

burdensome/discourage use

reported that they dislike the loss of personalpatient contact that results from the use oftelemedicine, a significantly higher proportionof nonusers endorsed this concern (p � 0.0001).A higher proportion of nonusers believed thatthe use of telemedicine would expand their net-work of colleagues (p � 0.0001). More nonusersalso hold the opinion that colleagues influencetheir use of new technologies such as telemed-icine (p � 0.0001), and that more research ontelemedicine is needed (p � 0.0001). More usersthan nonusers reported that they would em-ploy IAV telemedicine if it were available intheir offices (p � 0.0005). Nonusers were morelikely than users to favor the use of store-and-forward technology over IAV (p � 0.0014). Nosignificant difference was found in the propor-tion of their patients that nonusers and usersbelieve could be treated using store-and-for-ward technology (23.3% vs. 29.2%); by contrast,a significantly higher proportion of users be-lieved that they could treat more patients intheir practice with IAV telemedicine (40.4% vs.25.2%, p � 0.0007).

DISCUSSION

The results of this survey suggest that de-mographic characteristics do not explain theuse or nonuse of telemedicine. The only statis-tically significant demographic difference be-tween the two groups was that more userspracticed in public clinics and more nonuserswere in private practice. Public clinics, includ-ing federally qualified health centers, commu-nity and migrant health centers, and evenhealth clinics that serve the prison populationmight well depend on the use of telemedicineto serve their diverse and sometimes scatteredpopulations. The proportion of users in suchsettings was nonetheless modest, under 10% ofthe reporting users. This lack of demographicdifferences between users and nonusers is incontrast with the conventional wisdom intelemedicine, which is that younger, more re-cent medical school graduates are likely to beusers of telemedicine.

The results of the study indicate that usersand nonusers differ somewhat in their self-re-ported knowledge and beliefs about telemedi-

cine. Users were more likely to report beingknowledgeable about telemedicine, and weremore likely to perceive the potential advan-tages of telemedicine, and less likely to ac-knowledge barriers to its use. It may be that theperception of greater knowledge and greaterendorsements of the advantages of telemedi-cine is a causal factor that predisposes physi-cians to use telemedicine. Alternately, the suc-cessful provision of teleconsultation servicesmay lead specialists and subspecialists towardmore positive perceptions of their knowledgeand telemedicine’s advantages. Because thestudy was correlational in nature, it is not pos-sible to identify causation in the relationshipsunder investigation.

Some differences between what users andnonusers know or believe about telemedicinemay be explained by other aspects of their pro-fessional environments. The influence and roleof colleagues seemed greater for nonusers, whoreported that they depend on colleagues as asource of information and attitudes about top-ics such as telemedicine; users saw their col-leagues as having less influence over their (theusers’) decision making. Nonusers also weremore likely than users to believe that the useof telemedicine would increase their networkof colleagues, a finding that runs counter to theconventional wisdom that expanding one’s col-legial network is an incentive for telemedicineuse.

According to Berwick’s25 discussion of dis-seminating innovations in healthcare, in whichhe applies the work of Rogers26 on diffusion ofinnovations, the first group to adopt an inno-vation (“innovators”) represents a small pro-portion (2.5%) of the population under con-sideration; the second group to adopt aninnovation (about 13.5% of the population) arereferred to as early adopters. Across this sam-ple, roughly two-thirds of participants de-scribed themselves as early adopters of tech-nology. Telemedicine users were somewhatmore likely than nonusers to describe them-selves this way (albeit over 60% of nonusers de-scribed themselves as early adopters as well).

Although these findings may suggest a dis-crepancy between self-perception and behav-ior, it is likely that the threshold for identify-ing oneself as an early adopter is different

BARTON ET AL.496

among physicians than among telemedicine re-searchers. Further, it may be reasonable to ex-pect higher rates of participation by earlyadopters in this study than in the general pop-ulation of physicians. Participants (both usersand nonusers) may have self-selected into thestudy due to an interest in technology in gen-eral and telemedicine technology in particular.These findings may also suggest that a ten-dency toward early adoption of technologymay not cut across all technologies. Partici-pants may, in fact, be early adopters of sometechnologies, but not others.

The economic motivations for physician par-ticipation in telemedicine were explored to alimited extent in this survey. Respondentsagreed that reimbursement of telemedicine ser-vices should be at least on a par with that forin-person care. Questions pertaining to thetypes of payers and their proportion in eachphysician’s practice did not yield statisticallysignificant differences. Other factors beyond re-imbursement that likely need to be more fullyconsidered in assessing a physician’s economicmotivation for participating in telemedicine in-clude the location and convenience of theequipment, its availability and ease of sched-uling, the age of the equipment, availability oftechnical support, and other factors that influ-ence a physician’s time expenditure. Physiciansin private practice are less likely to use telemed-icine than those practicing in other settings, es-pecially public clinics, perhaps because theyare not on a salary and telemedicine is per-ceived as an opportunity cost in many cases.

Access to and experience with the use oftelemedicine appear to be factors in specialistutilization. In this sample, users of telemedi-cine were more likely than nonusers to reportthat their access to telemedicine facilities is convenient. The frequently cited barriers totelemedicine use (credentialing, licensure, andmalpractice; inconvenience; disruption to officeroutines caused by telemedicine scheduling)were perceived as barriers by nonusers to agreater extent than by users. Likewise, userswere more likely to report being knowledge-able about the use of telemedicine. However, itwas noteworthy that substantial proportions ofboth groups cited these factors as problems.Some users even reported having too little

knowledge of telemedicine to incorporate itinto their practices. Perhaps telemedicine usersare less likely than nonusers to be deterred bythese variables because they believe in the util-ity of telemedicine. However, it also may bethat these concerns do, in fact, influence theparticipation of users, who reported providingonly a small number of teleconsultations. Per-haps users only participated in teleconsultationin cases where the limitations were perceivedas nominal.

Users were open to a wider range of appli-cations of telemedicine technology in patientcare than nonusers. This may be because theirexperience with telemedicine has been moresatisfying than has been the case for nonusers,or that past satisfaction with the technologypredisposes them to have generally favorableexpectations toward future use. The opportu-nity to use telemedicine in a convenient settingseems to be associated with a positive per-spective on telemedicine’s potential as an ad-junct to their practice of medicine. Hands-onexperience may be required to persuade clini-cians of telemedicine’s utility.

The Centers for Medicare & Medicaid Services(CMS) has sought, through demonstration proj-ects and research studies, to determine whethertelemedicine services should be reimbursed, andif so, whether the reimbursement rates should becomparable to those for in-person care.20,21 Theoriginal Medicare payment waiver approved forCMS demonstration sites provided substantiallyless payment for providers who used telemedi-cine. The majority of respondents in this study,both users and nonusers alike, reported thatcurrent reimbursement remains a problem.Payers who want to encourage telemedicineutilization as one way to increase provider ef-ficiency, reduce patient and provider travel, orfor other reasons may need to reconsider theirtelemedicine reimbursement policies.

Dissemination of telemedicine technology atthe present appears to be driven more by ad-ministrators on the supply side than from anyintrinsic demand on the part of physicians orpatients. Yellowlees27 makes the case that a ten-dency exists for central bureaucracies—at dif-ferent levels of the hierarchy—to decide that atelemedicine system can solve a range of de-livery system problems. Such bureaucracies

SPECIALIST PHYSICIANS’ KNOWLEDGE ABOUT TELEMEDICINE 497

then design and establish systems, often withlittle input from the prospective clinician–usersabout their preferences, knowledge, and beliefsabout telemedicine. This top-down approachhas contributed to dissatisfaction with and fail-ure of some telemedicine systems. This is a crit-ical finding, strongly suggesting that the po-tential users of telemedicine must be not onlyconsulted prior to implementation of the tech-nology, but significantly involved in the de-velopment of telemedicine programs and applications. Such involvement is likely to pro-vide some of the incentive necessary for the es-tablishment of routines of practice that may be-come habitual.

The results of this survey are particularly in-teresting on several counts. First, nonusers dif-fer little from users on most demographic andpractice variables. The major exception is thatnonusers are more likely than users to be in pri-vate practice. This may reflect economic factors(e.g., capital expenditures and opportunitycosts for private practitioners), or it may be thatit is easier to use telemedicine technology whensomeone else (e.g., hospital administration)takes responsibility for establishing a system,especially one with convenient access.

Also of considerable interest was the fact thatusers and nonusers of telemedicine alike viewthe technology as having both potential ad-vantages and downsides. Even after experiencewith the technology, many users have a ten-dency to view telemedicine as somewhat in-convenient, less effective than in-person care,and less than optimal. The recognition of limi-tations and barriers to telemedicine servicesmay explain the relatively low utilization oftelemedicine even among those who haveadopted the technology. Telemedicine usersnevertheless appear to find the technology ofsufficient value that they are willing to take ad-vantage of it.

Because they demonstrate association andnot causality, the data do not provide a cleardirection with regard to mechanisms for in-creasing telemedicine utilization by physicians.A qualitative analysis of data obtained fromtelemedicine administrators as part of a sepa-rate component of this same study28 suggestedthat the issue is in part a function of habit. Thatis, physicians develop practice routines that are

efficient, and changing these habitual routinesinvolves at least temporary disruption of an ef-ficient pattern of workflow, irrespective of at-titudes, beliefs, and knowledge.29

Behavior may precede belief in this case, assuggested by social psychological research onpersuasion.30 Perhaps if physicians can be in-duced to use telemedicine on a regular basis,they will come to believe in its value, and in sodoing, become more knowledgeable about it.The more convenient telemedicine is for usersand the more exposure they have to the tech-nology, the more familiar it will become. Inconjunction with adequate reimbursement,over time the diffusion of telemedicine is likelyto increase. As long as physicians feel no com-pelling need to use telemedicine, however, thisprocess may occur slowly.

Although this study provides an importantexamination of knowledge and beliefs amongusers and nonusers of telemedicine, employinga large sample of physicians, the study has twoimportant limitations. It is possible that physi-cians with particular interest in telemedicineand/or technology in general were more likelyto complete and return the survey. Because wewere not able to obtain information about thosephysicians who chose not to participate, wecould not conduct formal analyses examiningthe potential for self-selection bias in this sam-ple. If more technologically minded physiciansparticipated in the study, this self-selectionmay have reduced the size of the differencesbetween the groups of users and nonusers, andlimited our ability to detect demographic dif-ferences between groups. In addition, becausethe study was correlational in nature, we werenot able to identify causal relationships. Futureresearch in this area should examine the causalimpact that variables such as demographiccharacteristics, knowledge, and beliefs abouttelemedicine advantages and barriers mayhave on the use of the technology.

ACKNOWLEDGMENTS

This research was supported by the Centersfor Medicare & Medicaid Services (CMS), Co-operative Agreement Number 18-C-90617/8(J.G.).

BARTON ET AL.498

REFERENCES

1. Benschoter R. Multipurpose television. Ann NY AcadSci 1967;142:471–478.

2. Bashshur RL, Puskin D, Silva J. Conference report:Second invitational consensus conference on telemed-icine and the National Information Infrastructure.Telemed J 1995;1:321–375.

3. Gilbert F. How to minimize the risk of disclosure ofpatient information used in telemedicine. Telemed J1995;1:91–94.

4. Granade PF. Malpractice issues in the practice oftelemedicine. Telemed J 1995;1:87–89.

5. Grigsby B, Brown N. 1998 Report on U.S. TelemedicineActivity. Portland, OR: Association of Telehealth Ser-vice Providers, 1998.

6. Grigsby J, Sanders JH. Telemedicine: Where it is andwhere it’s going. Ann Intern Med 1998;129:123–127.

7. Physician Payment Review Commission: Annual Re-port to Congress. Washington, DC: Physician PaymentReview Commission, 1995.

8. Shannon GW. The Atlantic Rim Telemedicine Sum-mit. Telemed J 1997;3:269–296.

9. Grigsby J, Rigby M, Hiemstra A, House M, Olsson S,Whitten P. The diffusion of telemedicine. Telemed J e-Health 2002;8:79–94.

10. Helitzer D, Heath D, Maltrud K, Sullivan E, AlversonD. Assessing or predicting adoption of telehealth us-ing the diffusion of innovations theory: A practicalexample from a rural program in New Mexico.Telemed J e-Health 2003;9:179–187.

11. Hu PJ, Chau PYK, Sheng ORL, Tam KY. Examiningthe technology acceptance model using physician ac-ceptance of telemedicine technology. J Manage InformSyst 1999;16:91–112.

12. Mitchell BR, Mitchell JG, Disney AP. User adoptionissues in renal telemedicine. J Telemed Telecare 1996;2:81–86.

13. Walker J, Whetten S. The diffusion of innovation: Fac-tors influencing the uptake of telehealth. J TelemedTelecare 2002;8(Suppl 3):S73–S75.

14. Mills OF, Tatqarko M, Bates JF, Hunsberger TA, Ever-hart-Yost E, Pendleton V. Telemedicine precepting ina family practice center. Fam Med 1999;31:239–243.

15. Guilfoyle C, Wootton R, Hassall S, Offer J, Warren M,Smith D, Eddie M. User satisfaction with alliedhealth services delivered to residential facilities viavideoconferencing. J Telemed Telecare 2003;9(Suppl 1):S52–S54.

16. Campbell JD, Harris KD, Hodge R. Introducingtelemedicine technology to rural physicians and set-tings. J Fam Pract 2001;50:419–424.

17. Larsen F, Gjerdrum E, Obstfelder A, Lundvoll L. Im-plementing telemedicine services in northern Nor-way: Barriers and facilitators. J Telemed Telecare 2003;9(Suppl 1): S17–S18.

18. Sicotte C, Lehoux P. Teleconsultation: Rejected andemerging uses. Methods Inf Med 2003;42:451–457.

19. Holm S. A simple sequentially rejective multiple testprocedure. Scand J Stat 1979;6:65–70.

20. Grigsby J, Grigsby B, Barton PL, Neal S, DeVore P,Brown N. Report to Congress on Telemedicine. Writtenfor the Health Care Financing Administration(HCFA), U.S. Department of Health and Human Ser-vices, in response to Section 4206(c) of the BalancedBudget Act of 1997, 1999.

21. Grigsby J, DeVore P, Barton PL, Talkington S, PaulichM, Floersch N, Araya T, Brown N, Loker J, GrigsbyB, Brooks E, Krohn N. Report to Congress on Expansionof Medicare Coverage of Telehealth Services. Written forthe Centers for Medicare and Medicaid Services(CMS), U.S. Department of Health and Human Ser-vices, in response to Section 223(d) of the Medicare,Medicaid, and State Children’s Health Insurance Pro-gram (SCHIP) Benefits Improvement and ProtectionAct of 2000 (BIPA), 2002.

22. Jacobson PD, Selvin E. Licensing telemedicine: Theneed for a national system. Telemed J e-Health 2000;6:429–439.

23. Johnson LJ. Legal risks of telemedicine. Med Econ2003;80:101.

24. Silverman RD. Current legal and ethical concerns intelemedicine and e-medicine. J Telemed Telecare 2003;9(Suppl 1):S67–S69.

25. Berwick DM. Disseminating innovations in healthcare. JAMA 2003;289:1969–1975.

26. Rogers EM. Diffusion of Innovations, 4th ed. New York:Free Press, 1995.

27. Yellowlees P. How not to develop telemedicine sys-tems. Telemed Today 1997;5:6–7,17.

28. Grigsby B, Brega AG, Bennett RE, DeVore PA, PaulichMJ, Talkington SJ, Floersch NR, Barton PL, Araya TM,Loker JL, Grigsby J. The slow pace of interactive videotelemedicine adoption: The perspective of telemedi-cine program administrators on physician participa-tion. Telemed e-Health 2007;13(6). In press.

29. Grigsby J, Stevens D. Neurodynamics of Personality.New York: Guilford Publications, 2000.

30. Cialdini RB. Influence: Science and Practice, 4th ed.Boston: Allyn & Bacon, 2000.

Address reprint requests to:Jim Grigsby, Ph.D.

Department of MedicineUniversity of Colorado at Denver

and Health Sciences Center13611 East Colfax Avenue, #100

Aurora, CO 80045-5701

E-mail: [email protected]

SPECIALIST PHYSICIANS’ KNOWLEDGE ABOUT TELEMEDICINE 499