Materials and Methods

This study was conducted from March to October 2006. The study protocol was reviewed and cleared by the University of Guelph Research Ethics Board and the Institutional Review Board of the Johns Hopkins Bloomberg School of Public Health. Full details of the study protocol have been previously published.²⁵

Study population—In brief, the target population from which veterinarians were recruited consisted of companion-animal practitioners from 14 counties across eastern Ontario, Canada. All veterinarians in this geographic area were identified by use of the College of Veterinarians of Ontario database, which includes all licensed veterinarians in the province. A randomized list of all veterinarians practicing ≥ 50% of their time in companion-animal practice was compiled. Veterinarians were contacted in sequence through an initial letter of introduction, followed by telephone recruitment until 20 had agreed to participate in the study. Because individual veterinarians were selected, it was possible to have multiple veterinarians from a single practice participate in the study.

For veterinarians agreeing to participate in the study, 2 initial study visits were scheduled to begin data collection. As part of the consent process, veterinarians 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 the outcomes of veterinary care. Each veterinarian's written, informed consent was obtained at the onset of the first study visit.

Written, informed consent was obtained from all clients, including permission for audio-video recording of the appointment, completion of a short demographic survey, a postinteraction appointment-specific CSQ, a 2-week follow-up telephone interview, and authorization to review the patient's medical record within 1 year after the original study visit. Clients were recruited in each clinic when data were being collected for the participating veterinarian. For compassionate reasons, appointments scheduled for a planned euthanasia were excluded from the study.

Data collection—A minimum of 14 veterinarian-client-patient interactions were videotaped/veterinarian. A video camera was strategically mounted in a corner on the ceiling of the examination room designated for participating clients. The camera was turned on prior to the first appointment and ran continuously for the duration of each study visit to minimize the chances of losing critical data. A short demographic survey was completed once by each participating veterinarian and client. An exit questionnaire designed to evaluate the impact of videotaping was administered to participating veterinarians at the end of each study day.

Measuring client adherence to medical recommendations was attempted by use of 2 methods. The first, self-reporting, involved telephoning participating clients approximately 2 weeks after their visit to assess their adherence. The second involved reviewing each patient's medical record approximately 6 months after their visit to document adherence. This process involved accessing the medical record for each patient evaluated during the initial study visits at the practice employing each participating veterinarian. The medical records were reviewed; the adherence data were documented on-site by one of the authors (JBC). Each patient's medical record was thoroughly examined from the date of the original study visit to the time of the medical record review for written documentation that a dental or surgical procedure had been performed.

RIAS coding of videotaped interactions—The RIAS is a framework that has been used extensively in human medicine to study physician-patient interactions by dividing talk into discrete communication variables. Subsequently, the RIAS has been adapted to veterinary medicine.^22,23,25,26 All interactions captured by audio-video recording during the study visits were analyzed by use of the RIAS.^b Trained coders evaluated both verbal and nonverbal communication used during the veterinarian-client-patient interactions. Verbal communication was broken down into individual utterances and coded into 1 of 48 mutually exclusive communication variables. An utterance is defined as the smallest discriminable speech segment (statement) to which classification may be assigned.³⁶ The present study assessed veterinarian-client-patient communication in the following 3 directions: veterinarian to client, client to veterinarian, and veterinarian to pet. For interactions involving ≥ 2 clients, all client talk was combined.

Identification of dentistry and surgery recommendations—From the RIAS-coded interactions, visits containing a dentistry recommendation, surgery recommendation, or both were identified. First, 2 proficiency codes were added to the RIAS coding framework to capture the presence of discussion related to dental care (first proficiency code) or surgical procedures (second proficiency code). Proficiency codes are distinct markers made in the coding of a visit that designate the occurrence of a predefined type of communication.

Next, all video-recorded interactions containing dentistry discussion, surgery discussion, or both were subsequently reviewed from the beginning by the principal author (NK) to identify more specifically visits where a dentistry recommendation, surgery recommendation, or both was made. A dentistry recommendation was defined as any proposal requiring the animal to undergo anesthesia for a procedure to improve the condition or prevent further wound, injury, or disease to the animal's mouth, including teeth and soft tissue. A surgery recommendation was defined as requiring the animal to undergo an invasive operative procedure under anesthesia where normal anatomy is substantially altered through the incision of skin or mucous membranes with connective tissue or where a body cavity is invaded. In addition, the type of surgical procedure was assigned to 1 of 4 categories: spay or neuter, lump removal, luxating patella, and other types of surgeries.

Coding clear versus ambiguous recommendations—Each identified recommendation was coded by the principal author (NK) as either clear or ambiguous. A clear recommendation was defined as one in which the veterinarian clearly proposed a need for a dental or surgical procedure to the client. For example, “Given the hard tartar and plaque build-up, I recommend we do a dentistry [procedure], to scale it off and prevent further gingivitis and gum recession” or “Unfortunately, the only sure thing to do is surgery.” An ambiguous recommendation was defined as one in which the veterinarian suggestively promoted a dental or surgical procedure in the form of an insinuation or vague proposition. This included occasions on which a number of alternatives were presented but no definitive medical recommendation was made, thus introducing a source of uncertainty. The following examples of ambiguous recommendations were not followed by any other clear recommendations: “We could drain it again and see what happens, or the other way to deal with a cyst would be to have an epidural under anesthetic and actually make an incision to take it out” and “Looking at her teeth, it wouldn't hurt to consider a dental [procedure] in the near future.”

Measuring appointment-specific client satisfaction—Appointment-specific client satisfaction was measured by use of the 15-item CSQ immediately administered following each veterinarian-client-patient interaction. The response format used a positively skewed, 6-point adjectival scale (1 = poor, 2 = fair, 3 = good, 4 = very good, 5 = excellent, and 6 = could not be better). The cumulative mean of all 15 satisfaction items making up the CSQ was calculated to produce an overall satisfaction score for each interaction. The CSQ has been validated³⁷ previously within the population of the present study.

Measuring RCC—The RCC score was calculated as a proportion of client-centered talk to veterinarian-centered talk, reflecting the balance of dialogue within an appointment.^c Client-centered talk is the sum of veterinarian and client information gathering and information giving regarding lifestyle or social matters (ie, discussion of the pet's daily activities and social interactions with humans and other animals), activation and partnership by the veterinarian, veterinarian rapport building, and gathering of biomedical information (ie, discussion relating to the medical condition, diagnosis, treatment, or prognosis) by the client. Veterinarian-centered discussion comprises the sum of veterinarian information gathering and information giving with respect to biomedical topics, veterinarian orientation, and client biomedical information giving.

A high RCC score represented visits with a greater indication of client centeredness, in which dialogue regarding lifestyle and social topics, activation and partnership and rapport building, and opportunity for the client to ask questions were evident. Conversely, visits corresponding to a low RCC score indicated a greater emphasis on the veterinarian agenda, in which veterinarians sought out biomedical data and clients primarily provided biomedical information in return.

Assessing the emotional tone during each study visit—During the initial coding of the veterinarian-client-patient interactions, the RIAS coders assigned global affect ratings to both the veterinarian and client by use of a rating scale of 1 (low) to 5 (high), capturing a number of positive (eg, interested and attentive, friendly and warm, responsive and engaged, sympathetic and empathetic, and respectful) and negative (eg, angry and irritated, anxious and nervous, depressed and sad, emotionally distressed and upset, dominant and assertive, and hurried and rushed) visit-level emotional tones. Emotional tones are assigned by the RIAS coders on the basis of their overall affective impressions of the veterinarian and client and are related to vocal qualities rather than literal content.³⁶ In the present study, 10 emotional tone ratings for veterinarians and 12 emotional tone ratings for clients were assessed. Because of the lack of distribution in data across each emotional tone category, each variable was dichotomized on the basis of frequency count to report a low and a high score.

Statistical analysis—Demographic variables were compared between the 20 respondent and 13 nonrespondent veterinarians.²⁵ Contingency table analyses were used for all discrete variables (gender; associate vs practice owner or partner; full-time vs part-time employment; urban, suburban, or rural practice location; and number of veterinarians employed by the practice) with P values calculated on the basis of the Fisher exact test. The Kolmogorov-Smirnov test was used to determine whether continuous variables were normally distributed. The independent t test was used to compare years of clinical experience between groups, and because data were not normally distributed, the Wilcoxon-Mann-Whitney test was used to compare base cost of a regular office visit between groups.

Further analysis was primarily conducted at the appointment level (eg, CSQ scores, RCC scores, and appointment duration) and occasionally at the veterinarian and client level (eg, emotional tone ratings). Dentistry and surgery recommendations were combined to increase the power of the study. Descriptive statistics (mean ± SD, median, and range) were calculated. A Fisher exact test was used to test the significance of clear versus ambiguous recommendations on client adherence. Appointment-specific CSQ score, RCC score, and appointment duration for visits were tested for normality by use of the Shapiro-Wilk test statistic. The Wilcoxon-Mann-Whitney exact test, with the exact permutational distribution, was used to compare median overall CSQ score, RCC score, and duration of appointment with client adherence and nonadherence. The Pearson χ² exact test, with the exact permutational distribution, was used to compare differences between high and low scores of veterinarian and client emotional tone ratings for appointments of clients who adhered to a dentistry recommendation, surgery recommendation, or both and those who did not. Spearman correlations were tested between similar veterinarian and client emotional tones when both were found to be significant. On a random subset of approximately 50% of the audio-video recordings, intrarater agreement was assessed for the coding of clear and ambiguous recommendations. The audio-video recordings were recoded by the principal author (NK) approximately 6 months following the initial coding, and the Cohen κ test was used to test agreement.

Twenty-one client-, pet-, veterinarian-, practice-, and appointment-related demographic variables (Appendix 1) along with 30 RIAS-coded communication composites capturing all 3 directions of veterinarian and client discussion (Appendix 2) were tested as potential determinants of client adherence to a dentistry recommendation, surgery recommendation, or both. The communication composites were composed of like variables from the 48 RIAS communication variables.²² Because of the multilevel structure of the data set, random effects were explored. In the end, binary logistic regression was used to avoid overfitting the data, considering that the data contained similar numbers of veterinarian participants (n = 19) and positive responses of adherence (25).

Using descriptive statistics, a cutoff was chosen on the basis of frequency count to dichotomize 1 RIAS-coded communication composite (ie, client-to-veterinarian activation and partnership). This newly categorized variable classified clients as offering either no activation and partnership or some activation and partnership statements to their veterinarian. The purpose of this grouping was to compensate for the small distribution of data. Univariate analyses were performed initially to screen all potential determinants for an association with client adherence to a dentistry recommendation, surgery recommendation, or both. All predictor variables (Appendices 1 and 2) for which the P value in univariate analyses was < 0.20 were retained.

Because of the limited size of the data set, initially, all retained variables (P < 0.20) were divided on the basis of demographic variables and RIAS-coded communication composites and modeled separately without interaction terms. Backward elimination techniques were used for model building. In addition, an alternative model-building process was used in which all retained variables were also incorporated into a series of 2-term subset models, including the interaction. Because of the limited size of the data set, each model was restricted to a maximum of 5 df.³⁸ Linearity was assessed for all continuous variables prior to testing the interaction models by use of graphic methods and by adding squared terms (quadratics). Backward elimination techniques were used.

All analyses were performed with standard software.^d,e Values of P ≤ 0.05 were considered significant.

Results

The follow-up telephone interview, conducted 2 weeks after the initial visit, was deemed unsuccessful at determining client adherence to a dentistry recommendation, surgery recommendation, or both. A number of clients had not reached the point at which their adherence could be measured, given that sufficient time had not elapsed since the initial study visit. On the other hand, dentistry and surgery recommendations could be consistently identified for client adherence or non-adherence by documentation in each patient's medical record. As a result, this was used as the sole method of measuring client adherence for the present study.

Study—The study involved 20 veterinarians, with a response rate of 60.6% (20/33). The 20 veterinarians represented 19 veterinary practices. Demographic information was obtained from all 13 veterinarians who declined to participate. The primary reason given for not participating was discomfort with having client interactions videotaped. In 3 instances, associate veterinarians who initially agreed to participate subsequently declined because they were unable to secure permission from their employers. In all 3 instances, the employer expressed concern about the potential disruption the study might have on the practice. Participants and nonparticipants did not differ significantly with regard to any of the demographic variables examined.

Medical record data were not available for 1 of the 20 veterinarians recruited; therefore, only 19 of the 33 (58%) veterinarians initially contacted to participate were included in analyses. The actual duration of time between the initial videotaped interactions and the completion of the medical record reviews ranged from 5.5 to 8 months (median, 6 months). Among the participants, females represented a majority of the veterinarians included (14/19 [74%]). The mean time in practice for the participating veterinarians was 13 years (range, 2 to 25 years). Eighteen of the 19 (95%) veterinarians practiced in clinics where ≥ 2 practitioners were employed. In 1 instance, 2 veterinarians from the same practice participated in the study; however, they did not encounter the same client during the study. Of the 18 veterinary clinics within eastern Ontario, 9 were located in urban areas, 5 in suburban areas, and 4 in rural areas.

Overall, 350 of 366 (96%) clients consented to participate in the research project; however, as a result of loss of battery power to the video camera, termination of videotape recording for compassionate reasons, admittance of client participants to an examination room without the mounted video camera, and the unavailability of medical record data for 1 veterinarian, 316 of the 350 (90%) consenting clients contributed data for this investigation. Based on the process used to identify dentistry and surgery recommendations, 83 videotaped interactions were ultimately included in the analyses of the present study.

Of these 83 interactions, 64 (77%) clients were female. The mean age of clients was 47 years (range, 22 to 78 years). Eleven (13%) clients had a household income ≤ $35,000, 18 (22%) had an income between $36,000 and $60,000, 22 (26.5%) had an income between $61,000 and $100,000, and 22 (26.5%) had an income > $100,000 (Canadian dollars). Annual household income was not provided for 10 (12%) clients. In addition, 52 of 83 (63%) clients had completed a college or university degree. Among the pet population, 51 (61.4%) visits involved appointments with only dogs, 31 (37.3%) with only cats, and 1 (1.2%) with a guinea pig. All 83 visits involved a single species of pet. As well, 8 of 83 (10%) visits were first-time interactions between clients and veterinarians.

Data quality—The RIAS coders rated the quality of the audio portion of the videotapes as good (n = 67 [81%]), fair (15 [18%]), or poor (1 [1%]). None of the videotaped interactions had abrupt beginnings. Only 1 (1%) interaction had an abrupt ending where the veterinarian exited the examination room to perform a procedure on the animal and did not return to formally conclude the appointment.

Impact of videotaping on participating veterinarians—The veterinarian exit questionnaire examining the impact of videotaping revealed that 17 of the 19 (89%) veterinarians felt that the videotaping did not hinder their clinical performance and that they were able to perform and behave normally throughout the appointment. However, 1 veterinarian reported that the videotaping disrupted clinical performance during the first study visit but not during the second visit, yet the veterinarian indicated being unable to behave as normal in front of the video camera during both study visits. Another veterinarian reported that the act of being videotaped affected clinical performance throughout the study.

Description of the veterinarian-client-patient interaction—Of the 83 study visits, the mean number of interactions contributed by each veterinarian was 4 (range, 1 to 7). The mean appointment duration was 18 minutes (range, 6 to 51 minutes), and a mean of 407 utterances (range, 95 to 1,040 utterances) were coded for each visit. Veterinarians contributed 62% of the total utterances during study visits, with 55% being directed at the client and 7% at the pet. In contrast, clients contributed 38% of the utterances.

Frequency of dentistry and surgery discussions—Of the 83 analyzed study interactions relating to a dentistry recommendation, surgery recommendation, or both, 49 (59%) visits contained discussion associated with a surgery recommendation, 33 (40%) consisted of discussion regarding a dentistry recommendation, and 1 (1%) included discussion of both a dentistry and a surgery recommendation.

Regarding surgical procedures, 28 of the 49 (57%) appointments dealt with talk related to spaying or neutering, 6 (12%) consisted of talk regarding lump removal, 4 (8%) involved the repair of luxating patella (dogs), and 11 (23%) involved other types of surgical procedures, each occurring once (eg, reducing hernia, surgical exploration, and caesarean section). The appointment dealing with discussions of both a dental and a surgical procedure consisted of discussion related to a dental cleaning as well as a lump removal.

Effect of clear versus ambiguous recommendations on client adherence—Overall, 25 of 83 (30%) clients adhered to veterinarian recommendations. Fifty-two of the 83 (63%) interactions consisted of a clear recommendation by the veterinarian regarding a dental or surgical procedure, whereas 31 (37%) were coded as ambiguous recommendations. From the 25 positive outcomes of client adherence detected in the study, most adhering clients (n = 22 [88%]) were provided with a clear recommendation, whereas 3 (12%) adhering clients were provided with an ambiguous recommendation. A significant difference was found between the clarity of recommendation given by the veterinarian (ie, clear or ambiguous) and client adherence to the procedure recommendation. In particular, the odds of a client adhering to a dentistry recommendation, surgery recommendation, or both when provided with a clear recommendation was 7 times as great as when clients were provided with an ambiguous recommendation (P = 0.001; 95% confidence interval, 1.8 to 25.4).

The κ test results for the recoding of 42 of the 83 (51%) visits containing a clear or ambiguous recommendation revealed high intrarater agreement (κ = 1), demonstrating the overall probability of random agreement by chance between author assessments to be small.

Effect of appointment-specific client satisfaction on client adherence—Overall, CSQ scores for the 83 veterinarian-client-patient interactions ranged from 3.0 to 6.0 (mean ± SD, 5.2 ± 0.7; median, 5.1). Adhering clients (median CSQ score, 5.8) were found to be significantly (P = 0.023) more satisfied with the veterinarian following the interaction than were nonadhering clients (median CSQ score, 5.0).

Effect of RCC score on client adherence—The RCC score ranged from 0.3 to 2.6 (mean ± SD, 1.0 ± 0.5; median, 0.8). Similar to CSQ score, the RCC score was significantly (P = 0.048) higher for adhering clients (median, 1.0) than for nonadhering clients (median, 0.8).

Effects of visit-level emotional tone on client adherence—For veterinarians, high sympathetic and empathetic tone (P = 0.004) and low hurried and rushed tone (P = 0.012) were found among visits where clients subsequently adhered to the recommendation. Likewise for clients, high sympathetic and empathetic tone (P = 0.027) and low hurried and rushed tone (P = 0.012) were found among appointments where clients subsequently adhered to the recommendation. Correlation analysis performed between the veterinarian and client sympathetic and empathetic tones detected no significant (P = 0.158) collinearity (ρ = 0.16), indicating 2 independent variables. Conversely, a significant (P < 0.001) correlation (ρ = 0.81) was found between the veterinarian and client hurried and rushed tone ratings. Appointment duration was also found to be significantly (P = 0.033) longer where there was subsequent adherence to a dentistry recommendation, surgery recommendation, or both (median, 18 minutes) than where clients did not adhere to a dentistry recommendation, surgery recommendation, or both (median, 14 minutes).

Factors associated with client adherence to a dentistry recommendation, surgery recommendation, or both—Few specific demographic- or communication-related variables were found to be significantly associated with client adherence to a dentistry recommendation, surgery recommendation, or both. Exploration of a possible random component was not suggestive, which is consistent with findings from the same study population when used to study the outcome of appointment-specific client satisfaction.³⁷ In that study,³⁷ it was found that only 7.5% of the variation in appointment-specific client satisfaction was at the veterinarian level, with the remainder of the variation explained at the visit level.

As a result of the univariate analyses, 12 variables (6 demographic variables and 6 RIAS-coded communication composites) with a value of P < 0.20 were retained for modeling. The 6 retained demographic variables were related to the client (pet primary care-giver and relationship with pet), veterinarian (years of clinical practice), practice (scheduled appointment duration), and appointment (verbal dominance score and measured duration of appointment). The 6 retained RIAS-coded communication composites were relationship building (client-to-veterinarian and veterinarian-to-client positive talk), activation and partnership (client to veterinarian, veterinarian to client, and veterinarian to pet), and orientation (veterinarian to pet). The initial model-building process, not involving interaction terms, resulted in a single model containing the variable representing veterinarian-to-client positive rapport-building statements (P = 0.04; coefficient, 0.032; 95% confidence interval, 0.001 to 0.063). Furthermore, the alternative model-building process produced a single model, again containing the RIAS-coded communication variable representing veterinarian-to-client positive rapport-building statements (P < 0.01; coefficient, 0.265; 95% confidence interval, 0.076 to 0.455) and its squared term (P = 0.02; coefficient, −0.003; 95% confidence interval, −0.006 to 0.001). Graphically, this model demonstrated that the probability of client adherence to a dentistry recommendation, surgery recommendation, or both initially was higher with veterinarian use of positive statements directed toward the client; however, once a threshold of positive statements (n = 45) was surpassed, the probability of client adherence began to decrease.

Discussion

In the present study, the association between veterinarian-client-patient interactions and client adherence was explored. Relationship-centered care, client satisfaction, clarity of recommendation, appointment duration, and certain visit-level emotional tones were found to differ among appointments where clients subsequently adhered to a dentistry recommendation, surgery recommendation, or both versus appointments where clients did not adhere.

Overall, the present study found client adherence to dentistry and surgery recommendations to be poor (30%). Nonetheless, in the present study, when clients were given a clear recommendation, they were 7 times as likely to follow through with the dental procedure, surgical procedure, or both. This suggests that when an explicit recommendation is put forward by the veterinarian, it has positive connotations for client adherence. In light of the positive association that was also found between RCC and client adherence in the present study, it is important to emphasize that this finding does not imply that veterinarians should simply use a paternalistic, expert-in-charge approach to making medical recommendations. Rather, making clients aware of the options available is a required element of informed owner consent.^39,40 Studies^19,41 have shown that most pet owners expect to be provided all options regarding their pet's health care. In addition to providing clients with all available options, findings in the present study address an inherent responsibility of veterinarians to clearly communicate to clients the best medically based recommendation for the pet from the available alternatives on the basis of the veterinarian's current understanding of the animal's condition, standards of care, and evidence-based medicine.

Moreover, results of a recent study²⁰ indicate that client perception of a recommendation's value is based on how well the veterinarian explains the rationale for the recommendation. The study²⁰ found that 30% of clients had not followed through with a recommendation by their veterinarian because they felt it was not necessary. Therefore, it is also important for practitioners to ensure that clients understand the rationale for each option and the recommendation. Research has shown that veterinarians and clients have different perceptions of value.⁴² Veterinarians tend to focus on the time, service, and expertise they provide their clients, whereas pet owners focus on the benefit of these services to their pet's health and well-being. For each of the options, it is important for the veterinarian to convey the value to the client in a manner that addresses the benefits to the overall health and well-being of the animal.

Results of the present study support the a priori hypothesis that client satisfaction contributes to greater client adherence to medical advice in veterinary medicine. Satisfaction reflects meeting the expectations of the client, communicating effectively,³⁷ and building strong veterinarian-client-patient relationships.^10,43 In an effort to reinforce client satisfaction and promote greater client adherence in practice, veterinarians are encouraged to explore the client's perspective, including their specific expectations in relation to their pet's care. Potential questions to elicit client perspectives may include, for example, “What do you hope to achieve today?,” “What are your thoughts on how we can tackle this problem?,” or “What questions or concerns do you have regarding the procedure I just outlined?” The information gathered through such questioning positions the veterinarian to relay the importance of his or her recommendation in a manner that aligns with the expectations and needs of the client and in turn promote greater client adherence.

Additionally, the present study also confirmed the a priori hypothesis that RCC leads to greater client adherence to recommendations. An RCC approach represents a cooperative endeavor between the veterinarian and client, provides opportunities for the sharing of critical information through open communication and collaboration, and ensures that the client's concerns, ideas, and expectations are thoroughly explored and appreciated.⁴⁴ This mutual endeavor allows clients to participate in making key decisions regarding their pet's health. A 2006 study²³ identifying communication patterns used during clinical appointments in companion-animal practice described the use of an RCC approach in 42% of the appointment visits studied. Thus, to shift the balance of power toward greater client involvement and in turn encourage higher rates of client adherence, veterinarians may find it beneficial to enhance communication skills that are in agreement with the RCC model. This may include providing information in small portions and verifying understanding with the client (so-called chunk and check), explaining to the client the range of treatment options, discussing possible concerns the client may have, and actively encouraging the client to participate in the development of a mutually agreed upon plan.⁴⁵ Research in human medicine suggests that use of RCC in veterinary medicine may also have other important benefits in addition to improving adherence, such as higher veterinarian and client satisfaction,⁴⁶ fewer malpractice claims,^47,48 and improved patient health.¹⁷

In addition to these findings, when appointment visits led to client adherence, veterinarians and clients were independently rated by RIAS coders as being more sympathetic and empathetic than when appointment visits led to client nonadherence. Roughly 80% of all communication between individuals has been identified to be nonverbal.⁴⁹ Empathy is 1 such skill that is central to the development of the veterinarian-client-patient relationship. It can be directly communicated through body language such as posture, eye contact, facial expressions, and body movements and does not always require the expression of verbal communication.^45,49 Veterinarian nonverbal communication skills such as soft tone of voice, use of touch, and caring facial expressions demonstrate empathy. By expressing attentiveness and sensitive appreciation for client preferences and concerns through nonverbal support, client adherence may be improved.

Both veterinarians and clients in the present study were perceived by the RIAS coders to be less hurried or rushed during appointment visits leading to client adherence than in appointments leading to client non-adherence. In addition, appointment duration was also found to be longer for interactions leading to client adherence to a dentistry recommendation, surgery recommendation, or both. Overall, considerable evidence was found that longer interactions are associated with a positive outcome of adherence. This finding is also consistent with that of an American Animal Hospital Association study,³² which found client adherence to be higher among longer appointments. It may be that when an appointment is rushed, the communication between a client and veterinarian is compromised, resulting in a negative impact on client adherence.⁴⁵ Interestingly, a strong correlation was identified between veterinarian and client hurried tone, which may be linked to a well-known social psychology theory called the reciprocity theory which holds that individuals respond to a positive action with another positive one and a negative action with another negative one.⁵⁰ In other words, it becomes fundamental that veterinarians be aware of the premise that when either the veterinarian or client appears to be rushed during the appointment, it has the potential to impact the overall veterinarian-client-patient interaction in a way that can negatively affect client adherence.

Interestingly, a common finding of both model-building approaches was that veterinarian use of positive rapport-building statements (ie, laughter, approval, compliments, and agreement) has a role in client adherence to dentistry and surgery recommendations. Assessing linearity for the RIAS-coded communication variables was an important part of the model-building process. Realistically, communication is a complex phenomenon that may not have linear relationships with factors associated with a particular outcome, the reason being that there normally is an optimum level or threshold up to which a relationship is linear and, thereafter, a negative relationship may be apparent. For example, in the present study, the use of positive relationship-building statements by veterinary practitioners toward clients initially had a positive association with client adherence; however, once a threshold of positive statements was reached, additional statements were observed to have an opposite effect on client adherence. Veterinarians should be encouraged to use positive rapport-building statements with clients but remain conscious that there may be a limit to the positive effects of this form of exchange.

In keeping with recommendations for measuring adherence,^28,51,52 2 approaches were initially used in the present study. Reviewing patient medical records proved to be a more valid measure of client's adherence to dentistry and surgery recommendations because Ontario licensure regulations⁵³ clearly define medical record requirements for companion animals, making it a legislative obligation for practitioners to record procedures requiring anesthetics (eg, dentistry and surgery) including detailed documentation of the procedure. Notwithstanding, it is still possible clients involved in the present study pursued the services of another veterinarian to perform the dentistry procedure, surgical procedure, or both that would not have been captured by this measure of adherence. To move adherence research forward, continued efforts to develop reliable and valid measures of adherence are important.

The present study included a small sample size of practitioners and their clients from a small region of eastern Ontario, and there may be disparities in the frequency and nature of dentistry and surgery discussions within different regions. Further, not all veterinarians approached to participate in the present study agreed. Although demographic differences between participants and nonparticipants were not found,²⁵ it is possible differences in the nature of the interactions for these 2 populations exist. The low incidence of dentistry and surgery recommendations available led to the analysis of a fairly small sample of only 83 veterinarian-client-patient interactions. For this reason, controlling for the multilevel structure of the data set was not possible during statistical analyses. Future studies should consider including a larger and more representative sample of veterinarian-client-patient interactions. Furthermore, the present study was limited to the investigation of client adherence specific to dental and surgical procedures. It is suggested that the nature of the veterinarian-client-patient relationship is likely to be influenced by the context (eg, the clinical problem and the physical space) within which the interaction occurs⁵⁴; therefore, future research that examines the findings of the present study in relation to the management of other clinical problems (eg, medication and diet) and environments (eg, emergency clinics) is needed. Finally, research studying veterinarian-client communication, particularly in relation to decision making and client adherence, would benefit from including a blend of quantitative and qualitative approaches. The role of veterinarian-client communication in client adherence is complex; future qualitative research would provide an inductive approach for further examining the intricate nature of these interactions and their potential role in client adherence. Including quantitative research would allow researchers to test and better understand findings within a broader population of veterinarians and their clients.