This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
The quality of physician-patient communication is a critical factor influencing treatment outcomes and patient satisfaction with care. To date, there is little research to document the effect of telemedicine (TM) on physician-patient communication.
The objectives of this study are to measure and describe verbal and nonverbal communication during clinical TM consultations and to compare TM with in-person (IP) consultations in terms of the quality of physician-patient communication.
Veteran patients (n = 19) requiring pulmonary medicine consultations were enrolled into the study. The study group included 11 patients from the Iron Mountain Veterans Affairs Hospital (VAMC) remote site. Patients had individual TM consultations with a pulmonary physician at the Milwaukee VAMC hub site. A control group of 8 patients had IP consultations with a pulmonary physician at the Milwaukee VAMC. Video recordings of medical consultations were coded for patient-physician verbal and nonverbal communication patterns using the Roter Interaction Analysis System (RIAS).
There were no differences in the length of TM consultations (22.2 minutes) and IP consultations (21.9 minutes). Analysis of visit dialogue indicated that the ratio of physician to patient talk was 1.45 for TM and 1.13 for IP consultations, indicating physician verbal dominance. Physicians were more likely to use orientation statements during IP consultations
The study findings indicate differences between TM and IP consultations in terms of physician-patient communication style
Physician-patient communication problems are a common cause for patient dissatisfaction during in-person (IP) consultations. Dissatisfied patients are less likely to return for physicians’ appointments, more likely to switch physicians, and more likely to be noncompliant with recommendations [
Clinical telemedicine (TM) includes applications that use communication technologies to link specialists at a tertiary center to patients and primary providers at a remote site. It is unclear how the physical separation between patient and physician inherent in TM consultations affects physician-patient communication. A number of comprehensive reviews have identified the need to evaluate physician-patient interactions during TM consultations [
One objective of this pilot study is to describe the use of the “Roter Interaction Analysis System” (RIAS) as a comprehensive content coding methodology for assessing verbal and nonverbal communication during clinical TM consultations [
We conducted an observational pilot study to compare the pattern of verbal and nonverbal physician-patient communication during TM and IP consultations. The study was conducted at the Milwaukee VAMC and was approved by the Institutional Review Board. The Milwaukee VAMC, a teaching affiliate of the Medical College of Wisconsin, is a tertiary care facility providing a full range of primary care and specialty services for veterans residing in the greater Milwaukee area. The Milwaukee VAMC provides TM services in medical subspecialties (eg, pulmonary medicine, rheumatology, and infectious disease) to the Iron Mountain VAMC in Michigan [
In the present study, 3 pulmonary specialists participated in conducting TM and IP consultations at the Milwaukee VA. These board certified specialists each had at least two years of experience in conducting TM consultations. To control for physician variability, the same group of physicians (n = 3) conducted both the IP and TM consultations. One of two nurses was always present during the TM consultations at the Iron Mountain remote site. These nurses are trained to assist with TM examinations and are familiar with the equipment and technology used during TM.
TM consultations were performed between the pulmonary specialists located at the Milwaukee VAMC telemedicine site and the patient and trained nurse located at the Iron Mountain VAMC telemedicine site. Consultations were conducted over a live two-way audio and video conferencing system using a high-speed (384 kbps) telecommunication connection. The Milwaukee VAMC used a Tandberg videoconference system with 27-inch Sony monitors; the Iron Mountain VAMC used a VTEL FRED unit with a 27-inch monitor. The established sound and video quality of TM was near perfect, and both parties were able to change the field of view by zooming in or out as needed. The nurse at the Iron Mountain VA assisted the physician during the physical examination. Electronic medical records of each patient were available to the Milwaukee physicians by accessing the VA’s computer network, and an electronic progress note for each consultation was entered into the patient’s medical record.
Patients undergoing IP care at the Milwaukee VAMC are from the Milwaukee metropolitan area. In our study, IP patients were checked in by a clinic nurse and placed in an examination room previously set up with the video recording equipment. Physicians conducted an IP medical interview and a hands-on physical examination with the patients. No nurse was present in the room during these consultations. Physicians had access to electronic medical records of each participating patient and entered an electronic progress note for each consultation.
For TM examinations, a VCR was used to record the picture and sound as viewed from the TM physician’s TV monitor in Milwaukee. This 27-inch monitor displayed the image of the remote site examination room in Iron Mountain, including a full-body frontal view of the seated patient and an upper body frontal view of the nurse seated behind a desk next to the patient. In addition, a 5-inch diagonal picture-in-picture frontal view image of the physician was projected in the upper-left corner of the monitor. The IP consultations were recorded by a digital camcorder equipped with a wide-angle lens mounted on a tripod placed in the examination room. The camera and microphone were placed in an unobtrusive manner, and both the patient and physician were alerted to the presence of these devices. The camcorder was set up in a standardized fashion, and the resulting image included a full body, oblique view of both the seated patient and the seated physician.
We used the Roter Interaction Analysis System (RIAS) to code the medical dialogue. The RIAS treats a complete thought, defined as a simple sentence, a sentence clause, a sentence fragment, or single word, as the unit of analysis. A complete thought may be categorized as one of 38 mutually exclusive and exhaustive codes. Coding is done directly from video recordings without transcription. In the current study, a nurse was always present during the TM consultations, and her communication was coded separately from that of the TM physician. In 8 of the 19 consultations, the patient was accompanied by a companion, and his or her speech was also coded separately. RIAS coding was performed by trained coders under the supervision of Dr. Roter (co-investigator) at Johns Hopkins University. Dr. Roter trained coders in the use of the RIAS System over several weeks, using a coding manual with detailed definitions and annotated examples, and training tapes demonstrating standardized techniques. Inter-rater reliability was calculated based on double coding of 20% sample by 2 independent and blinded coders. In the current study, reliability ranged across categories roughly from 0.7 to 0.9, based on Pearson correlation coefficients.
In addition to verbal communication, the RIAS is also used for coding nonverbal communication. In this study, we used RIAS for coding of physician and patient global affect ratings. These ratings were based on the emotional tone of each speaker for the following dimensions: (1) angry or irritated, (2) anxious or nervous, (3) depressed or sad (patient only), (4) emotionally distressed (patient only), (5) dominant or assertive, (6) interested or attentive, (7) friendly or warm, (8) responsive or engaged, (9) sympathetic or empathetic, (10) hurried or rushed (physician only), and (11) respectful. The global affect ratings were performed by trained coders under the supervision of co-investigator Dr. Roter at Johns Hopkins University. The coders assigned an overall Likert score (1 = lowest and 6 = highest) for each dimension for the whole interaction. Inter-rater reliability, calculated as percentage agreement by double coding of 20% random sample by 2 independent and blinded coders, ranged from 87% to 100%.
Visit length is associated with communication quantity (number of utterances) and was measured for each IP and TM consultation. There were no significant differences in the length of consultations (
Descriptive analyses of the visit dialogue can also be presented as various ratios to capture relative amounts of talk. The ratio of total number of provider utterances to total number of patient utterances is called the “verbal dominance ratio” and is a summary measure of “patient-centered” versus “physician-centered” style of communication. A verbal dominance ratio of 1 indicates equal participation by patient and physician and is indicative of a patient-centered interview style, whereas a ratio of greater than 1 indicates a physician-centered interview style. In this study, the physician dominated the interview more during TM as compared to IP visits (TM = 1.45 vs. IP = 1.13; t = -1.25,
For the purposes of this study, each utterance made during the visit by the patient, physician, and nurse was coded into one of 38 mutually exclusive and exhaustive RIAS categories. During analysis, the 38 mutually exclusive content categories are often combined into larger “subsuming categories” that share common meaning [
Physician-patient verbal communication during TM and IP visits
Speaker | Physician | Patient | ||||||
Type of Visit | TM | IP | TM | IP | ||||
RIAS Subsuming Communication Category | N (SD) | N (SD) | t | N (SD) | N (SD) | t | ||
Information-gathering, biomedical | 29.1 (18.9) | 26.5 (23.1) | -0.26 | .80 | 4.9 (4.3) | 4.6 (2.4) | -1.84 | .86 |
Information-gathering, psychosocial | 4.9 (5.6) | 4.8 (4.4) | -0.07 | .95 | 0.2 (0.6) | 0.1(0.4) | -0.26 | .80 |
Information-giving, biomedical | 71.1 (56.6) | 51 (29.8) | -1.00 | .33 | 86.3 (75.8) | 97 (58.9) | 0.35 | .73 |
Information-giving, psychosocial | 1.4 (2.4) | 4.8 (9.2) | 1.02 | .34 | 11.8 (15.2) | 12.8 (17.3) | 0.12 | .91 |
Positive talk | 31 (23) | 27 (23) | -0.36 | .73 | 24 (26) | 32 (19) | 0.80 | .44 |
Negative talk | 0.2 (0.4) | 0.4 (0.5) | 0.88 | .40 | 1.5 (2) | 0.8 (1) | -.98 | .34 |
Social talk | 3.7 (3) | 2.1 (1.8) | -1.46 | .16 | 2.6 (2.8) | 1.3 (1.4) | -1.32 | .21 |
Rapport building/emotional responsiveness | 8.1 (9.1) | 10.4 (9.8) | 0.52 | .61 | 5.4 (5.4) | 6.6 (7.2) | 0.42 | .69 |
Partnership building | 14.2 (11.5) | 12.3 (7.5) | -0.44 | .66 | 1.9 (3.1) | 2.9 (2.4) | 0.76 | .46 |
Orientation statements | 9.7 (8.9) | 19 (9.4) | 2.17 | .047 | 0.9 (1.25) | 1.6 (2.7) | 0.78 | .46 |
These RIAS categories include closed- and open-ended questions asked by patient and physician related to (1) medical condition and therapeutic regimen (biomedical), and (2) lifestyle and psychosocial topics. Overall, both patients and physicians engaged in more biomedical than psychosocial information gathering during both TM and IP visits. When comparing the two consultation modes, there was no significant difference in the amount of biomedical information gathered by physicians (TM = 29.1 vs. IP = 26.5; t = -0.26,
These RIAS categories include information sharing by patient and information sharing plus counseling by physician related to (1) medical condition and therapeutic regimen (biomedical), and (2) lifestyle and psychosocial topics. It was noted that both physician and patient predominantly exchanged information regarding biomedical versus lifestyle and psychosocial topics during both IP and TM consultations. Physicians provided more patient counseling and information sharing regarding biomedical issues during TM as compared to IP visits (TM = 71.1 vs. IP = 51; t = -1.00,
The category “positive talk” includes laughter, agreement, and approval. Positive talk constituted a considerable portion of total talk for both physicians and patients (TM = 17.9%, SD = 5.3; IP = 18%, SD = 4.8), there was no difference in physician talk (TM = 31 vs. IP = 27; t = -0.36,
The RIAS category “rapport building/emotional responsiveness” includes instances in which the patient or physician shows concern or asks for or provides an opinion or reassurance. There was no difference between TM and IP visits for physicians (TM = 8.1 vs. IP = 10.4; t = 0.52,
This category includes physician statements that tell the patient what is expected during the consultation or what is about to happen (eg, “I am going to check your pulse now.”), as well as statements that serve to orient the patient to major topics of discussion or the physical flow of the visit. Physicians used fewer orientation statements during TM (9.7) as compared to IP (19) visits (t = 2.17,
Requests for repetition are statements (eg, “What?” “Come again?” “How is that?” “Would you repeat that?”) in response to instances in which one participant has not clearly heard or understood another’s words or statements. Such requests are indicative of perceptual difficulties. Utterances are coded as “unintelligible” when the coder is unable to understand what a participant (eg, patient or physician) has said. These two categories are not part of the subsuming Roter categories (
The RIAS was used for global ratings of the emotional tone (nonverbal communication) of both the patient and physician. Based on vocal qualities, coders captured the global ratings (ie, a single rating for the entire visit) of emotional affect for both the patient and physician on 11 affective dimensions (anger, anxiety, sadness, distress, dominance, interest, friendliness, responsiveness, sympathy, hurried, and respectful) for each IP and TM visit. There were no significant differences noted in global affect ratings for physicians (TM = 2.4 vs. IP = 2.4; t = -0.15,
A popular conceptual model used to describe physician-patient communication defines communication styles as either “physician-centered” or “patient-centered.” Physician behaviors, such as gathering of information via closed-ended questions, testing hypotheses to make a diagnosis, giving medical directions, and controlling the visit, represent a “physician-centered” style of communication. The patient’s role is to listen, follow the physician’s directions, and play a passive role during the medical encounter. This type of communication is less successful in addressing the needs of the patient [
The findings of this study suggest that the use of TM does influence patient-physician communication style. During TM visits, physicians were more likely to dominate the dialogue, as evidenced by a higher verbal dominance ratio (TM = 1.45 vs. IP = 1.13; t = -1.25,
Clinical TM consultations fundamentally differ from IP encounters due to the physical separation of physician and patient. Direct physical examination and interview are not possible, and a virtual environment replaces the familiar physician’s office. It is possible that the physical separation and the lack of the hands-on physical examination reinforce a “physician-centered” interview style observed during TM consultations. Whether such differences in communication, particularly in the absence of the hands-on examination, have an effect on quality of care, health outcomes, and future patient utilization of TM is not known.
We observed patients to be less engaged (less talkative) and more likely to take on a passive role during TM as compared to IP visits. It is possible that poor patient participation and communication during TM visits are due to the lack of familiarity with technology and the perception of physician detachment due to the inherent physical separation. Patient concerns about privacy and confidentiality during TM consultations may further inhibit patient participation, especially if the dialogue involves collecting data on sensitive or personal topics (eg, sexual history).
Clinical TM consultations frequently involve a three-party communication exchange between a specialist at a tertiary site, and a patient and second provider at a remote site. In Street’s study, the presence of a referring primary care physician was linked to inhibited patient communication and lack of direct engagement with the consulting physician [
The present study results suggest that TM visits are less patient centered than IP visits. In contrast, results of studies that use patient self report to measure patient satisfaction often indicate high patient satisfaction with TM [
The study is small and exploratory with obvious limitations in terms of experimental design and statistical power. Nevertheless, the study provides a framework for detailed observations and description of patient-physician communication during TM consultations that can be useful in the design of future research in this area. Undoubtedly, the quality of interaction during TM and its potential impact on health care outcomes is an area of growing importance. Further research is needed to help fill the current gaps in the literature and to develop specific interventions that can improve the quality of communication during TM encounters.
This material is based upon work supported in part by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development—Health Services Research and Development and by resources and the use of facilities at the VA Medical Center Milwaukee, WI.
None declared.
telemedicine
in-person
Veterans Affairs Medical Center
veterans affairs
Roter Interaction Analyses System