Materials and Methods

The study protocol was reviewed by the University of Guelph Research Ethics Board. A detailed description of the overall study design has been previously described.³

Study participants—In brief, a list of veterinarians identified as spending at least 50% of their time in small animal practice from 14 counties in Eastern Ontario was compiled from the College of Veterinarians of Ontario database. By use of a random number generator, veterinarians on the list were randomized and contacted sequentially until 20 veterinarians were recruited to participate. Veterinarians were initially contacted via a letter of introduction. In a follow-up telephone call 2 to 3 weeks later, further details of the study were discussed by one of the authors (JBC) and verbal consent to participate was obtained from the veterinarian. Two initial study visits were scheduled, at which time written consent was obtained.

During each study visit, all clients who had appointments scheduled with a participating veterinarian, with the exception of those for arranged euthanasia, were approached by one of the authors (JBC) immediately prior to their appointment. After being informed about the study, clients were invited to participate and written consent was obtained. During the consent process, veterinarians and clients were informed that the purpose of the study was to describe veterinarian-client-patient interactions in companion animal practice and to examine the impact of these interactions on outcomes in veterinary care.

Data collection—Between 14 and 28 appointments were recorded for each participating veterinarian by use of a video camera mounted to the ceiling of an examination room. The video camera ran continuously during each study visit to minimize data loss. Only video recordings that captured the beginning of the veterinarian-client-patient interaction were included in the study. Demographic information was collected on the participating veterinarians and clients. To assess the perceived impact of video recording on the veterinarian's performance, a brief exit questionnaire was administered to each veterinarian at the end of each study visit, asking whether he or she thought being videotaped interfered with his or her clinical performance.

Appointments were classified as wellness or problem appointments.⁴ A wellness appointment was defined as one in which the animal was presumed to be healthy at the onset of the appointment; a problem appointment was defined as one in which the patient was to be evaluated for a clinical concern.

Coding of veterinarian solicitations of client concerns—A coding system was developed by 2 of the authors (LMAD and JBC) on the basis of previous research conducted in human medicine in which physician solicitation of patient concerns was analyzed.^5–7 All video-recorded veterinarian-client-patient interactions included in the present study were coded by the principal author (LMAD). Although solicitations for concerns by a veterinarian were identified and coded throughout each veterinarian-client-patient interaction, the focus of the present study was on solicitations made during the opening segment of the interview. The opening segment of the interaction was defined as the period from the first verbal interaction between the veterinarian and client to the first observed transition in the interview (eg, opening segment to history gathering).

A solicitation was defined as any question that probed the reason for the visit or for a client's concerns. Once identified, a solicitation was categorized as either an open-ended question, which is a broad inquiry into a client's concerns that does not shape or direct the content of the response¹¹ (eg, “What can I do for you today?”), or a closed-ended question. Closed-ended questions (eg, “Any concerns?” or “Is Rory in today for his vaccinations?”) are structured to elicit a specific and often 1-word response (eg, “Yes.” or “No.”).¹¹ A concern was defined as any question or statement expressed by the client and related to the health or behavioral well-being of the animal.

Frequency of interruptions by veterinarians and the length of client responses prior to interruption—Coding for the completion of the client's response was performed on the basis of the methods described by Marvel et al.⁶ A client's opening statement was considered complete if 1 of 3 criteria were met: a statement of completion was made by the client (eg, “That's all.”), the client asked a concern-related question (eg, “Is this a normal reaction?”), or the client responded in the negative to a solicitation for concerns from the veterinarian without the subsequent expression of a concern by the client (eg, “No, I don't have any concerns.”).

The method used for coding each type of interruption was based on the descriptions used in human medicine by Beckman and Frankel.⁵ On the basis of the intent of the interruption being made, interruptions were classified as closed-ended question, elaborator, recompleter, or noninterrogative statement used by the veterinarian (Appendix). An additional category, interruption made by veterinary patient, was added to the present study to account for possible interruption caused by the patient or by the veterinarian directing her or his attention to the patient. Following verbal interruptions, an observation was made and recorded of whether the client returned to and completed his or her statement after the interruption.

The length (measured in seconds) of a client's response to the veterinarian's solicitation before interruption was recorded from the start of the client's first utterance following the solicitation to the point of verbal disruption of the client's response by the veterinarian. The length of time was measured by use of an audio and video player.^a

Occurrence of concerns arising at the end of appointments—Concerns were initially observed and recorded throughout the entire veterinarian-client interaction. Following the initial coding of concerns, appointments that contained a concern in the last half of the appointment were reexamined for the presence of concerns arising during the closing segment of the appointment. Based on the methods of White et al,¹² the closing segment begins at the transition phrase (eg, “All right, so that's everything for today.”) and is concluded with a closing statement (eg, “Bye now.”). To code late-arising concerns, the statement used to transition the conversation into the closing segment was first identified. Any new concerns expressed by the client during this period were then coded as late-arising concerns. Appointments in which the veterinarian solicited for a final concern immediately prior to or during the transition into the closing segment (eg, “Well, that's everything unless you had any other concerns or questions?”) were categorized as not having a late-arising concern.

Statistical analysis—Descriptive statistics were calculated. A comparison of the effect of open-ended versus closed-ended questioning in eliciting the expression of 1 or more concerns by the client was tested by use of a 2-tailed exact permutation test. A Fisher exact test was used to test for an association between veterinarian solicitation of concerns at the beginning of the interview and the presence of concerns arising during the closing segment of the interview.

The relationships among potential determinants, veterinarian solicitation, and the completion of a client's response were tested. Generalized linear mixed modeling was used to address the complex multilevel structure of the data set. Initially, univariable analyses were performed to screen potential determinants for unconditional associations (P < 0.20) with the presence of a solicitation by the veterinarian and for completion status of a client's response to a solicitation. On the basis of previous research evidence from human and veterinary medicine demonstrating the effect of certain demographic characteristics on clinical communication, a number of specific demographic variables were analyzed, including 2 veterinarian-level variables (male vs female¹³ and number of years in clinical practice³), 3 client-level variables (male vs female, age,¹⁴ and regular veterinarian, occasional veterinarian, or first-time encounter¹⁵), and 1 appointment-level variable (wellness vs problem appointment^4,16,17). For each outcome of interest, all predictor variables demonstrating an unconditional association with the dependent variable were placed into a maximum model, including all 2-way interactions. Backward elimination was used to reduce terms in the model. Random-effect terms were retained if the estimate equaled or was greater than one-quarter of the SE.

By use of a randomly selected subset (10%) of all video-recorded interactions, an analysis of intrarater agreement was performed on the coding of veterinarian solicitation occurrence and client statement completion. The video recordings were recorded by the first author (LMAD), and κ statistics were calculated.

All analyses were performed with statistical software.^b Values of P < 0.05 were considered significant.

Results

Study population—The overall response rates for veterinarians and clients contributing data to the present study were 61% (20/33) and 91% (334/366), respectively. As has been previously reported,³ 13 (65%) of the participating veterinarians were female and 7 (35%) were male. The median number of years in practice for the veterinarians participating in the study was 12.5 (range, 2 to 31 years). Nineteen clinics were represented by the participating veterinarians. Nine clinics were located in urban areas, 5 in suburban areas, and 5 in rural areas. On the exit questionnaire administered at the end of each study visit, 17 (85%) veterinarians reported that the video recording did not interfere with their clinical performance and 2 (10%) indicated that the presence of the video camera interfered with their clinical performance during the first visit but not the second. One veterinarian believed that the video recording interfered with clinical performance throughout the study.

For the 334 interactions analyzed in the present study, 245 (73%) clients were female and 89 (27%) were male. The median age among participating clients was 45 years (range, 18 to 80 years). Overall, 208 visits had a dog, 117 visits had a cat, 2 visits had a dog and cat, and 7 visits had another type of animal (eg, iguana or guinea pig).

The median duration of the 334 appointments was 15.12 minutes (range, 2 to 63 minutes). On the basis of the reason given by clients for evaluation of their pet, wellness appointments constituted 51% (n = 171) of the appointments analyzed and problem appointments made up the remaining 49% (163).

Description of veterinarian solicitation of client concerns—From the 334 interviews analyzed, 123 (37%) interviews included a veterinarian solicitation at the beginning of the interview. In the remaining 63% (n = 211), no solicitation was made. An open-ended solicitation was observed in 76% (n = 93) of interviews with solicitations; the other 24% (30) of solicitations took the form of a closed-ended solicitation.

The use of an open-ended solicitation resulted in no expressions of client concerns in 24% (22/93) of the appointments, 1 concern in 65% (60/93) of the appointments, and ≥ 2 concerns in 11% (10/93) of the appointments. In comparison, the use of closed-ended solicitations resulted in no client concerns being expressed in 60% (18/30) of the appointments, 1 concern in 27% (8/30) of the appointments, and 2 concerns in 13% (4/30) of the appointments. Following the use of a closed-ended question, > 2 client concerns were never expressed. The use of open-ended solicitations resulted in significantly (P < 0.01) more client concerns being elicited than the use of closed-ended solicitations.

Overall, when veterinarians used open-ended solicitations, clients spoke a mean of 13.2 seconds (median, 8.0 seconds; range, 1 to 139 seconds), whereas with the use of closed-ended solicitations, clients spoke a mean of 5.0 seconds (median, 2.0 seconds; range, 1 to 24 seconds).

Frequency of interruptions by veterinarians and the length of client responses prior to interruption—Of the 123 appointments containing a solicitation by the veterinarian, 68 (55%) included the client's response being interrupted by the veterinarian prior to completion. Several reasons for incompletion of client responses were identified (Table 1), with the most frequent cause for interruption being a closed-ended question asked by the veterinarian (39/68) and the second most common being a noninterrogative statement made by the veterinarian (18/68). Most often (58/68), interruptions by the veterinarian occurred after the client had expressed at least 1 concern.

Among the 68 appointments that contained an incomplete response to the veterinarian's solicitation, the median length of time clients spoke before being interrupted by their veterinarian was 11 seconds (range, 1 to 139 seconds; mean, 15.6 seconds; SD, 12.1 seconds). However, it appears that the longer a client was allowed to speak prior to being interrupted, the more concerns arose (Table 2). Clients who expressed 1 concern prior to being interrupted by their veterinarian spoke a median time of 11.5 seconds (range, 1 to 139 seconds; mean, 16.5 seconds; SD: 21.1 seconds), whereas clients who did not express a concern prior to being interrupted spoke a median time of only 5 seconds (range, 1 to 26 seconds; mean, 8.7 seconds, SD, 8.87 seconds). In instances in which the client was interrupted by the veterinarian, clients were provided the opportunity to return to and complete their response in 28% (19/68) of appointments.

Occurrence of concerns arising at the end of appointments—Of the 334 appointments observed, 114 involved a concern raised in the last half of the appointment; 29 of these were classified as a late-arising concern. Of the 29 appointments that involved a late-arising concern, 86% (25/29) of the appointments did not contain a veterinarian solicitation at the beginning of the appointment. Appointments in which the veterinarian did not solicit client concerns at the beginning of the interview were significantly (P < 0.01) more likely to have a concern arise during the closing segment of the interview than were appointments in which a solicitation for the client's concerns were made (odds ratio, 4.0; 95% confidence interval, 1.4 to 11.8).

Factors associated with the occurrence of veterinarian solicitations—Following the model-building process, the final model when adjusted for random effects contained 2 main effects: appointment type (P < 0.001) and number of years in clinical practice (P = 0.03). From this model, the odds of a solicitation occurring in a wellness appointment were 2.8 times as high (95% confidence interval, 1.7 to 4.7) as the odds of a solicitation occurring in a problem appointment when accounting for random effects and the number of years the veterinarian had been in clinical practice. Taking into account random effects and appointment type, there was a negative association between the number of years a veterinarian had been in clinical practice and the likelihood of a solicitation (coefficient, 0.94; 95% confidence interval, 0.90 to 0.99).

Factors associated with the completion of a client's opening statement—Following the model-building process, only appointment type (P < 0.001) remained in the final model. The odds of a client's opening statement being completed in a wellness appointment were 6 times as high (95% confidence interval, 2.9 to 12.6) as the odds of a client's opening statement being completed in a problem appointment.

Thirty-four video recordings were recoded for intrarater agreement. The κ statistics calculated for the occurrence of veterinarian solicitations and the completion of client statements were 0.93 and 0.87, respectively, demonstrating high intrarater agreement.

Discussion

The results of the present study suggest that veterinarians are more likely to solicit concerns at the beginning of an interview in wellness appointments than in problem appointments. When solicitations did occur, clients were provided the opportunity to complete their responses in just under half (45%) of the situations, with the odds of completing their opening statement being significantly higher in wellness appointments than in problem appointments. Finally, this study found that late-arising concerns were more likely to occur in appointments that lacked a veterinarian solicitation at the beginning of the interview.

Veterinarians solicited for concerns at the beginning of the interview in a little over a third of (37%) appointments, with solicitations occurring more frequently in wellness appointments than in problem appointments. The relationship between solicitations and appointment type may be related to the intent of the visit: wellness appointments are often visits intended to identify any changes in the health of the animal and where the solicitation of client concerns can be a natural progression for the interaction. In problem visits, a major concern or focus for the interview is often established prior to the appointment, as the reason for the evaluation of the patient and solicitation for further concerns may be overlooked or not considered to be a priority. However, research in human medicine has shown that disparity can exist between the presenting complaint and the principal problem defined by the clinician at the end of the interview.¹⁸ Regardless of the type of appointment, veterinarians should be careful of assuming that the reason for evaluation or even the first expressed concern is always the most important. With the possible exception of an emergency situation, taking a moment to solicit for all of a client's concerns at the beginning of an interview is likely to assist practitioners in acquiring the full range of a client's concerns to prioritize the most critical and, in turn, determine the most appropriate agenda for the interview.

With closed-ended questioning accounting for one-quarter of solicitations observed in the present study, examination of the differences between the use of open- and closed-ended solicitations by veterinarians is important. Open-ended questioning was found to significantly increase the number of concerns expressed by the client, compared with the use of closed-ended questioning. The results of a previous study⁴ in veterinary medicine found that veterinarians commonly use closed-ended questioning, with 25% of the 300 studied interactions not containing the use of a single open-ended question by the veterinarian. The findings from the present study suggest that open-ended questions are beneficial in identifying a client's concerns. Furthermore, research in human medicine suggests that the use of predominantly closed-ended questioning during clinical interactions is more likely to result in frequent interruption on the part of the clinician, which subsequently reduces rapport and empathetic connection with the client.¹⁹ On the basis of results of the present study, the use of open-ended solicitations to elicit client concerns should be promoted, as such solicitations encourage clients to talk longer and share more concerns.

In human medicine, Silverman et al¹¹ suggested taking a 4-part approach to soliciting for concerns during medical encounters. First, solicitation of the client's concerns should begin with an open-ended enquiry followed by active listening. Second, during the client's response, veterinarians should be aware of and respond to verbal and nonverbal cues expressed by the client. Third, after the client expresses his or her concerns, veterinarians should screen for additional concerns with further open-ended questioning. Finally, confirming the list of expressed concerns through summarization and collaborative agenda setting between the veterinarian and client should be done before moving on to information gathering, which is considered the next stage of a well-structured interview.

Allowing clients to complete their response to a solicitation for concerns is likely to improve the probability of discovering all of a client's concerns. The present study found the odds of a client completing his or her response to a solicitation for concerns by the veterinarian to be significantly greater during a wellness appointment than during a problem appointment. Previous research in veterinary medicine found veterinarians to be considerably more hurried or rushed during problem appointments than during wellness appointments.¹⁷ Time pressure has been cited as a challenge to effective communication by veterinarians²⁰ and may influence whether a veterinarian solicits for further concerns or interrupts a client during his or her opening statement. It is possible that the higher number of incomplete responses during problem appointments is a result of veterinarians feeling more time demands or restrictions in problem appointments, resulting in a greater number of client interruptions. Soliciting all of a client's concerns at the beginning of an interview, regardless of the type of appointment, allows the veterinarian and client to prioritize the concerns, gather a more detailed history, and avoid the inefficiency that can occur when concerns are raised late in an interview.

The use of effective, evidence-informed communication skills in veterinary medicine has only recently become emphasized in veterinary school curricula.²¹ The present study found that veterinarians with a greater number of years in clinical practice were less likely to solicit for client concerns at the beginning of an interaction than were veterinarians that had graduated more recently. These results support previous findings of a similar relationship between the number of years in practice and the likelihood of a cost discussion taking place between the veterinarian and client.³ The recent introduction of formal communication skills curricula to veterinary schools around the world^22–26 has resulted from an acknowledgement that communication is a core clinical competency required of all veterinarians.^27,28 Thus, it is possible that more recent graduates may have a greater awareness of the use and impact of their own communication skills and, as a result, be more likely to use communication skills that pursue a broader, more complete history, including solicitation and screening for a client's concerns. This is in contrast to the so-called traditional medical history, which may be more commonly used by practitioners with a greater number of years in clinical practice. This traditional framework for history gathering focuses primarily on the clinical information deemed relevant to formulating a diagnosis.^11,29 Regardless of the number of years in clinical practice, all veterinarians would benefit from actively pursuing a client's complete set of concerns at the beginning of an interaction to prioritize the concerns and reduce the likelihood of a concern arising during the closing moments of the appointment.

The most common barriers to completion of a client's initial response to veterinarian solicitations found in the present study were closed-ended questions posed by the veterinarian. Research in human medicine has shown that interruptions made by the clinician during the initial expression of a patient's concerns inhibit the identification of any additional concerns by halting the spontaneous flow of information.⁵ Specifically, any utterance made in an attempt to obtain or provide further information about a previously elicited concern becomes a barrier to the expression of subsequent concerns. In the present study, following an interruption made by the veterinarian, clients returned to and completed their response in only 28% of cases. Veterinarians may naturally want to probe concerns as they arise or respond to a concern with an explanation when one is available; however, withholding this response until clients have had the opportunity to disclose all of their concerns will contribute to structuring the interview appropriately and, in turn, improve the overall efficiency.

Our study found that the presence of a veterinarian solicitation at the beginning of an interview was associated with a lower frequency of concerns raised during the closing segment of the appointment. This suggests that veterinarian solicitation of client concerns at the beginning of the interaction is useful in avoiding concerns being raised at the endpoint of an appointment. A systematic review of the literature in human medicine examining the role of patient-physician relationships and communication skills on efficiency (ie, length of time of visit) identified up-front agenda setting, which includes soliciting patient concerns, to be 1 of 3 domains that improve clinical efficiency⁹ Attempting to solicit all of a client's concerns early in the interview should assist veterinarians in avoiding the dilemma that ensues following the introduction of a late-arising concern: clinicians are forced to make a choice between reopening the interview at a cost to their appointment schedule, ignoring the concern at a possible cost to client satisfaction, or deferring exploration of the late-arising concern to another visit with possible implications to the patient and the veterinarian-client relationship.⁵

The present study was restricted to veterinarian-client-patient interactions taking place within the examination room and did not consider the influence of support staff on the solicitation of client's concerns. Although support staff may be enlisted to solicit for client's concerns in advance before the interview with the veterinarian, this could not be examined in the present study. Regardless of whether support staff play a role in identifying client concerns, we believe it is important for all veterinarians to screen for additional concerns directly with the client, by use of open-ended questioning at the beginning of each client interaction. The impact of the use of support staff to elicit client concerns is an area that warrants further consideration and research.

Importantly, the sample population in the present study included only a small number of primary care veterinarians and their clients from a limited region in a single Canadian province. Differences may exist in relation to findings of the present study when applied to other regions. In addition, it is suggested that nonverbal cues can act as a solicitation¹²; however, the nonverbal aspects of veterinarian-client interactions were not included in the definition of a solicitation for the study reported here. Future research should be developed to better understand the nonverbal aspects of veterinarian-client-patient interactions.