Open Veterinary Journal, (2026), Vol. 16(4): 2128-2141

Research Article

10.5455/OVJ.2026.v16.i4.16

Client experience and owner-reported functional recovery after feline orthopedic surgery: A structural equation model of loyalty and recommendation intentions

Afif Yuda Kusumah^*, Cecilia Marliani, and Innocentius Bernarto

Master of Hospital Administration and Global Business Management Program, Faculty of Economy and Business, Pelita Harapan University, Jakarta, Indonesia

*Corresponding Author: Master of Hospital Administration and Global Business Management Program, Faculty of Economy and Business, Pelita Harapan University, Jakarta, Indonesia. Email: afifyuda25 [at] gmail.com

Submitted: 15/12/2025 Revised: 27/02/2026 Accepted: 11/03/2026 Published: 30/04/2026

---

ABSTRACT

Background: As client experience becomes an established quality domain in veterinary care, owner appraisals of postoperative recovery may influence downstream behavioral intentions. In feline medicine, postoperative recovery in daily life is commonly inferred from owner observations using structured owner-reported outcome measures, where patients cannot self-report.

Aim: To examine how client experience (CE), modeled as a reflective-reflective higher order construct, relates to owner-perceived postoperative functional recovery and how CE and owner-reported outcomes (ORO) predict client loyalty (CL) and recommendation likelihood (RL), including ORO’s mediating role.

Methods: Owners of cats who underwent orthopedic surgery at an Indonesian feline referral clinic completed a cross-sectional online questionnaire (N=153). ORO was assessed using the 9-item Feline Musculoskeletal Pain Index short form (FMPI-sf) as an owner-reported proxy of postoperative function and comfort (not a direct measure of radiographic healing). The data were analyzed using partial least squares structural equation modeling in SmartPLS 4, applying a disjoint two-stage approach for the CE higher-order construct and bootstrapping for inference.

Results: CE positively predicted ORO (β=0.677, p < 0.001) and had direct effects on CL (β=0.486, p < 0.001) and RL (β=0.709, p < 0.001). ORO predicted CL (β=0.451, p < 0.001) but not RL (β=0.144, p=0.089). ORO partially mediated the effect of CE on CL (β=0.305, p < 0.001), whereas it did not mediate RL (β=0.097, p=0.118). The model explained substantial variance in ORO (R2=0.458), CL (R2=0.737), and RL (R2=0.662).

Conclusion: In feline orthopedic pathways, perioperative communication, empathy, pain management guidance, and shared decision-making are clinically relevant care-process targets associated with stronger owner-perceived recovery and higher continuity intentions. Recommendation intention appeared to be primarily experience-driven rather than recovery-driven. The findings should be interpreted in light of the cross-sectional, single-clinic design and the use of FMPI-sf as a postoperative proxy measure.

Keywords: Client experience, Client loyalty, Feline orthopedic surgery, Owner-reported outcomes, Recommendation likelihood.

---

Introduction

Over the past two decades, the human–companion animal relationship has shifted toward a "family member" paradigm, strengthening owners’ emotional investment in their cats and dogs (Applebaum et al., 2021). Consequently, veterinary care expectations increasingly extend beyond clinical success to include owners’ experience of the full care pathway. In line with established healthcare quality frameworks, client/patient experience is recognized as a core dimension of service quality alongside clinical effectiveness and safety (Institute of Medicine, 2001).

Feline orthopedic surgery is a particularly demanding companion animal practice area. Procedures such as long-bone fracture repair (forelimb or hindlimb), pelvic fracture stabilization, and hip surgery require high technical proficiency, careful perioperative pain management, and substantial owner participation in post-surgical care. Traditionally, surgical success has been judged primarily using clinician-reported outcomes and objective indicators, such as radiographic bone healing or joint stability. However, these endpoints may not fully reflect what owners perceive as meaningful recovery in their cat’s everyday life—namely, the ability to walk, jump, climb, play, and interact comfortably at home.

Two complementary research streams inform post-surgical outcomes from the owner’s perspective. First, service research shows that clients’ service encounter evaluations are associated with downstream behavioral intentions, including loyalty-related behaviors and positive WOM/recommendation (Cronin and Taylor, 1992; Zeithaml et al., 1996). Second, outcomes research emphasizes patient-reported outcomes and owner-reported outcomes (ORO) as structured measures capturing perceived function, symptoms, and health-related status after an intervention in veterinary contexts (Gagnier, 2017). In companion-animal orthopedics, ORO instruments have been increasingly evaluated and used to capture outcomes that matter to owners, complementing clinician- and imaging-based endpoints (Essner et al., 2017; Evangelista et al., 2019; Gruen et al., 2021; Radke et al., 2022a; Glenn et al., 2024). As in human health care, ORO instruments should demonstrate validity, reliability, and responsiveness to clinical change (Reeve et al., 2013; Gagnier, 2017; Van Lieshout and Wijffels, 2020).

Integrating these streams, ORO can be theorized as a plausible explanatory pathway through which the service encounter translates into post-service behavioral intentions. This positioning is consistent with Donabedian’s structure–process–outcome logic, which links care processes to outcomes that subsequently inform decisions (Reinharth, 1988), and expectancy–disconfirmation theory, whereby perceived performance/outcomes relative to expectations shape downstream responses (Oliver, 1980). Loyalty and recommendation are related but not identical behavioral intentions: perceived functional recovery may reinforce the decision to return to the same provider, whereas active recommendation can represent a higher-threshold advocacy behavior. Whether perceived recovery similarly contributes to loyalty versus recommendation remains an empirical question in feline orthopedic surgery.

From a service and consumer-behavior perspective, client experience (CE) reflects owners' appraisal of how care is delivered, including information exchange, emotional support, and involvement in decision making. Effective veterinarian–client communication and empathic interactions are linked to greater trust, adherence, and satisfaction in veterinary settings (Coe et al., 2008; Coe et al., 2009; Adams and Kurtz, 2017). Contemporary approaches, such as trauma-informed care, further emphasize emotional safety, transparency, and empowerment, which have been associated with improved client trust and satisfaction in veterinary contexts (Rohlf et al., 2025). Benefit-focused communication in digital channels may also enhance client engagement with recommended care (Sutherland et al., 2025). Taken together, this evidence supports CE as a multidimensional construct that plausibly shapes how patients interpret and report recovery outcomes following orthopedic surgery.

In Indonesia and many other low- and middle-income countries, postoperative monitoring in veterinary orthopedic practice remains largely focused on clinic-based objective assessments, whereas routine ORO measurement and structured surveillance using owner questionnaires are less commonly implemented. Owners’expectations and digital literacy are rising, and online reviews and recommendations increasingly influence veterinary service selection. Evidence from human health care suggests that patient experience, satisfaction, and perceived outcomes are associated with loyalty and electronic word-of-mouth, shaping provider reputation and service utilization (Cheung and Thadani, 2012; Doyle et al., 2013). In veterinary research, studies of service quality and satisfaction (e.g., in Ghana) report positive associations between perceived service quality, satisfaction, and loyalty, but typically do not incorporate ORO or model CE as a higher-order construct (Turkson, 2011; Turkson et al., 2025). Consequently, integrated models connecting CE, ORO, and client behavior—particularly client loyalty and recommendation likelihood—within a single structural framework in feline orthopedic surgery remain scarce.

To address this gap, this study develops and tests a conceptual model (Fig. 1) in which CE is specified as a reflective–reflective higher-order construct manifested through four first-order dimensions: Communication, Empathy, Pain Management, and Shared Decision Making. This specification aligns with patient-experience frameworks that emphasize the informational, emotional, and participatory elements of care (Doyle et al., 2013; Gagnier, 2017). According to the service-quality theory, CE is expected to be positively associated with post-service behavioral intentions (Cronin and Taylor, 1992; Zeithaml et al., 1996). ORO captures owner-perceived functional recovery and is increasingly emphasized in orthopedic outcome assessment (Gagnier, 2017; Radke et al., 2022a). Drawing on Donabedian’s process–outcome logic and expectancy–disconfirmation theory, we propose that CE as a service-process appraisal shapes owners’ perceived functional recovery ORO, which may subsequently inform post-service behavioral intentions client loyalty (CL), recommendation likelihood (RL), warranting empirical testing in feline orthopedic surgery (Oliver, 1980; Donabedian, 1988).

Fig. 1. Conceptual framework of the study. H1: CE positively influences ORO. H2: CE exerts a direct positive effect on CL. H3: CE exerts a direct positive effect on RL. H4: ORO positively influences CL. H5: ORO positively influences RL. H6: ORO mediates the CE–CL relationship (CE → ORO → CL). H7: ORO mediates the CE–RL relationship (CE → ORO → RL).

This study offers both scientific and clinical contributions to the research on veterinary outcomes. Conceptually, it extends patient-experience and proxy-reported outcome frameworks to feline orthopedic services by modeling CE as a higher-order construct and explicitly testing ORO as a mediator between experience and client behavior. Clinically, it provides evidence from an Indonesian referral-clinic setting that perioperative communication, empathy, pain management guidance, and shared decision-making are not only service-process attributes but also clinically relevant components of postoperative care pathways because they shape how owners understand, monitor, and report recovery at home. An integrated post-surgical model was tested in which CE—specified as a reflective-reflective higher-order construct, was linked to ORO and two behavioral intentions (Client Loyalty and Recommendation Likelihood). Mediation via ORO was assessed to clarify how perceptions of experience translate into post-service intentions.

Aims and hypotheses

Grounded in the preceding theoretical discussion and the limited empirical evidence on post-service behavioral intentions in feline orthopedic pathways, this study sought to explain how the service encounter, as perceived by owners, translates into perceived postoperative functional recovery and downstream relationship outcomes. The first aim of this study was to determine whether CE—specified as a reflective–reflective higher-order construct—is positively associated with ORO following feline orthopedic surgery. The second aim was to evaluate the extent to which CE and ORO independently contribute to two central post-service intentions, namely, CL and RL. The third aim was to test the underlying mechanism by examining whether ORO mediates the effects of CE on CL and RL. Seven hypotheses were formulated consistent with this logic model.

---

Materials and Methods

Study design and setting

This research employed a quantitative, observational, cross-sectional design. Data were collected using a structured questionnaire completed by owners of cats that underwent orthopedic surgery at Anugerah Satwa Cat Care Center, a feline referral clinic in South Tangerang, Banten Province, Indonesia. Most surgical cases involved trauma-related fractures (often vehicular) affecting the forelimb, hindlimb (including unilateral/bilateral involvement), and pelvic region (Fig. 2). Fractures were stabilized using pins, wires, plates, and screws as indicated by the fracture configuration and clinical requirements. A cross-sectional approach was considered suitable for estimating structural relationships among latent constructs in a veterinary health service context using partial least squares structural equation modeling (PLS-SEM).

Fig. 2. Fracture cases among 153 responses, including single- and bilateral-limb involvement.

Population, eligibility criteria, and sampling

A total of 153 eligible cases comprised client-owned cats that underwent orthopedic surgery involving the forelimb, hindlimb, and pelvic region within the preceding 12 months (December 2024; to November 2025). Data collection was conducted in the first week of December 2025. This inclusion window was selected to balance (i) sufficient time for biological healing and functional stabilization and (ii) feasible owner recall of perioperative experiences and early recovery, consistent with published orthopedic healing timelines and owner-reported outcome measurement considerations (Montavon et al., 2009; Yap et al., 2014; Yap et al., 2015; Decamp et al., 2016; Meeson and Geddes, 2017; Bird and De Vicente, 2020; Radke et al., 2022b; Innes, 2023). Recovery trajectories vary by fracture type and case complexity; Figure 3 shows the distribution of reported recovery timelines across the 153 cases. Because recruitment relied on completed responses during the survey period, the sample may under-represent owners who were difficult to contact, less digitally engaged, dissatisfied, or no longer in follow-up, and may also under-represent unfavorable outcomes (including loss to follow-up or death), which could introduce selection or survivorship bias.

Fig. 3. Recovery process after surgery in 153 cats.

Data-collection procedures

The questionnaire was administered online using Google Forms. Owners accessed the survey link on their mobile phones during a 7-day data-collection period, yielding 153 eligible responses. All items were set as mandatory in the form; therefore, no missing data were observed for the focal constructs (CE, ORO, CL, and RL). The questionnaire used clear, concise wording and straightforward response options to minimize satisficing and reduce respondent burden. Participation was voluntary and had no impact on current or future care. The first page presented an electronic informed-consent statement, and only owners who provided consent were allowed to proceed. Any clarifications requested via the provided contact number were limited to explaining the item’s meaning without steering responses. The analytical dataset contained no direct personal identifiers, and each record was stored under an anonymous code. The survey link was voluntarily distributed via the clinic’s communication channels and completed. The total number of owners who received and viewed the link could not be determined with certainty; thus, no formal response rate could be calculated. Consequently, self-selection and nonresponse bias cannot be ruled out, as participation may have been more likely among owners who were more engaged, more satisfied, or more confident using digital forms (Groves, 2006). In addition, both CE and ORO were measured retrospectively in a single survey, so recall error and social desirability (particularly toward the treating clinic) may have influenced responses. Accordingly, the findings should be interpreted as reflecting the experiences of respondents and may not be fully generalizable to all owners of cats undergoing orthopedic surgery at the clinic.

Sample size adequacy

Sample size adequacy was rigorously justified. An a priori power analysis conducted with G*Power 3.1 for a linear multiple regression (fixed model, R² deviation from zero) indicated a minimum requirement of 68 cases (effect size f²=0.15, α=0.05, power=0.80). The final sample of 153 far exceeds this requirement, ensuring sufficient statistical power for the analysis. Furthermore, the sample size comfortably surpasses common heuristic guidelines for PLS-SEM (e.g., ten times the largest number of predictors for any endogenous construct, which was 2 in this model: CE and ORO for CL/RL).

Measures and instrument development

The model comprised 4 latent constructs, namely, CE, ORO, CL, and RL, which were specified as reflective measurement models. CE was modeled as a reflective-reflective higher-order construct manifested by four reflective first-order dimensions representing the core aspects of the service encounter: Communication (COM), Empathy (EMP), Pain Management (PM), and Shared Decision Making (SDM). ORO captured owners’ perceptions of postoperative functional recovery in daily life, CL reflected intentions to continue using the same clinic, and preference for it over alternatives, and RL assessed advocacy-oriented intention (likelihood to recommend), consistent with widely used Likelihood to Recommend operationalisations (Reichheld, 2003). Communication was assessed using 5 items (COM1-COM5). Empathy was assessed using 4 items (EMP1-EMP4), with EMP5 excluded. Pain management was assessed using four items (PM2-PM5), reflecting owners’ confidence in perioperative pain control and clarity of analgesic guidance consistent with contemporary small-animal pain management guidelines (Mathews et al., 2014; Steagall et al., 2022), with PM1 excluded. SDM was assessed with three items focusing on involvement in discussing options, understanding risks/benefits, and participation in decisions, with SDM1 and SDM2 excluded. Owner-Reported Outcomes were operationalized using the 9-item Feline Musculoskeletal Pain Index short form (FMPI-sf), a refined version of the original 17-item feline musculoskeletal pain index (FMPI), developed to improve responsiveness while maintaining reliability for monitoring chronic musculoskeletal pain-related functional impairment in cats (Gruen et al., 2015; Enomoto et al., 2022). The FMPI-sf was used in this study as a structured owner-reported proxy for postoperative function and comfort because it captures observable daily activities (e.g., movement, jumping, mobility-related behavior, and comfort) that owners monitor during home recovery, which are clinically relevant domains after orthopedic surgery. This choice does not imply that the FMPI-sf is a surgery-specific instrument or a direct measure of fracture union, implant stability, or clinician-assessed surgical success; rather, it was used to quantify owner-perceived functional recovery in a standardized way within a single post-surgical survey. Owner/observer-reported outcome measures are widely used in companion-animal musculoskeletal care because owners consistently observe many clinically meaningful changes during routine daily activities rather than during brief in-clinic assessments (Lascelles et al., 2007; Gruen et al., 2015). Earlier feline work also combined owner-reported outcome measures with activity monitoring to evaluate pain relief in cats, supporting the value of structured owner-reported assessment in feline musculoskeletal conditions (Lascelles et al., 2007). An established feline metrology instrument is the FMPI, which conceptualizes musculoskeletal outcome as a multidimensional construct spanning everyday activity/mobility, comfort/pain-related impact, and global quality of life (Benito et al., 2013). In selecting nine items, we prioritized content validity (relevance and adequate coverage of key recovery domains for the target population and context) while keeping the instrument sufficiently concise to minimize respondent burden and support completeness in a single-administration survey (Terwee et al., 2018; Aiyegbusi et al., 2022). Refinement of the FMPI demonstrated that reducing the original 17-item scale to a nine-item short form improved responsiveness while maintaining reliability in client-owned cats, supporting the feasibility of nine well-chosen items for owner-reported feline musculoskeletal outcomes (Enomoto et al., 2022). Client loyalty was measured using 4 items (CL1-CL4). RL was measured using three items (RL1, RL2, RL4); the original RL3 item was removed during model development due to inadequate psychometric performance, resulting in a more parsimonious RL construct consistent with PLS-SEM measurement refinement guidance (Hair et al., 2019; Hair et al., 2022). All items were rated using a seven-point scale. For CE, CL, and RL, anchors ranged from 1 ("strongly disagree") to 7 ("strongly agree"). For ORO, anchors captured perceived change (e.g., 1="much worse" to 7="much better") while retaining the 1–7 numeric coding for analysis. The questionnaire was drafted in English, translated into Indonesian, and reviewed by veterinary orthopedic clinicians for clinical relevance, clarity, and cultural/linguistic appropriateness. Minor wording refinements were made to align professional terminology with language typically used by cat owners.

Common method bias (CMB) mitigation and assessment

CMB was screened using the full collinearity assessment. Specifically, four auxiliary regression models were estimated in which each Stage-2 construct (CE, ORO, CL, and RL) was alternately treated as an endogenous variable and regressed on the remaining constructs. The resulting variance inflation factor (VIFs) represent FCVIFs that capture both vertical and lateral collinearity. The FCVIF values were assessed, with all values below 5. Although some values slightly exceeded the conservative threshold of 3.3, they did not indicate severe collinearity (Kock and Lynn 2012; Kock, 2015).

Data analysis: PLS-SEM with a reflective higher-order construct

Data were analysed using PLS-SEM in SmartPLS (version 4) (Ringle et al., 2024). PLS-SEM was selected because it accommodates complex latent variable models, including reflective higher-order constructs, and is suitable for moderate sample sizes without requiring multivariate normality. A disjoint two-stage approach was applied to the higher-order construct. In Stage 1, the first-order dimensions (COM, EMP, PM, and SDM) were estimated with their indicators, and the scores of the latent variables were saved. In Stage 2, these dimension scores served as indicators for the CE higher-order construct within the full structural model (CE, ORO, CL, RL), and all hypothesized relationships were estimated. All constructs were modeled as reflective, and each indicator was expected to act as a consistent manifestation of its latent variable. Following established PLS-SEM measurement procedures, indicators were refined iteratively by examining indicator reliability (outer loadings), internal consistency reliability (Cronbach's α/ρA/composite reliability), convergent validity (AVE), and discriminant validity (HTMT), while retaining each domain’s content validity (Hair et al., 2019; Hair et al., 2022). Indicators that showed inadmissible or weak measurement behavior, including low/unstable loadings and/or problematic overlap with other constructs, and that adversely affected reliability and convergent validity, were removed to ensure a clean measurement foundation before estimating the reflective–reflective higher-order construct using the disjoint two-stage approach.

Accordingly, RL3, SDM1–SDM2, EMP5, and PM1 were removed at Stage 1 prior to estimating the reflective–reflective higher-order construct. The final measurement retained RL1, RL2, and RL4 and ensured adequate representation of each CE dimension (SDM3–SDM5; EMP1–EMP4; PM2–PM5). Structural model evaluation included collinearity assessment using inner VIF. Path significance (direct, indirect, and total effects) was tested using bias-corrected bootstrapping with two-tailed 5,000 resamples. Explanatory power was evaluated with R², effect sizes with f²: 0.02 (small), 0.15 (medium), and 0.35 (large) (Cohen, 1988), and predictive relevance with blindfolding-based Q² values. Out-of-sample predictive performance was assessed using PLSpredict with cross-validation and CVPAT in SmartPLS in addition to explanatory assessment (Shmueli et al., 2019; Liengaard et al., 2021). Predictive relevance was examined via Q²_predict values at the construct level and by comparing prediction error metrics (RMSE and MAE) against benchmark models.

Ethical approval

The ethical approval date for this study is December 2, 2025. The ethical approval number is 054/MARS/EC/XII/2025. All procedures complied with institutional and national ethical guidelines for research involving human participants in a veterinary health service context. Participants could discontinue the study at any time before submitting the anonymous questionnaire. Data were securely stored, analyzed in aggregate form, and reported without identifying individual owners or animals.

---

Results

Measurement model assessment

All constructs were specified as reflective. Client Experience was modeled as a reflective–reflective higher-order construct (HOC) represented by Communication, Empathy, Pain Management, and Shared Decision Making, and estimated using a disjoint two-stage approach. Indicator reliability was supported in the final model, with all retained indicators exhibiting satisfactory and statistically significant outer loadings (Table 1). Specifically, loadings ranged from 0.825–0.857 for CL, 0.631–0.826 for CE indicators from 0.631 to 0.826, ORO from 0.766 to 0.870, and recommendation likelihood from 0.840 to 0.890. RL was initially operationalized with 4 indicators; however, RL3 showed an inadmissible pattern and weakened measurement quality. Therefore, RL3 was removed, resulting in a parsimonious three-indicator RL construct (RL1, RL2, RL4) with strong loadings (Table 1). In addition to the refinement of the RL scale, Stage-1 screening of the CE first-order dimensions resulted in the removal of SDM1–SDM2, EMP5, and PM1 due to the reflective specification’s weak measurement behavior. These deletions improved the psychometric quality of the first-order dimensions used to form the CE higher-order construct while ensuring sufficient coverage through the retained indicators (SDM3–SDM5; EMP1–EMP4; PM2–PM5).

Table 1. Outer loadings of reflective indicators (disjoint two-stage HOC model).

The full-collinearity screen produced VIFs ranging from 1.907 to 3.844. The construct-level FCVIFs (maximum VIF per auxiliary model) were CE=3.837, ORO=3.535, CL=3.331, and RL=3.844 (overall maximum FCVIF=3.844). Using the conservative cutoff of 3.3, these results indicate that the FCVIF criterion cannot completely rule out CMB because several values slightly exceed 3.3 (Kock, 2015). However, Kock noted that VIF thresholds may be set somewhat higher in contexts where algorithms incorporate measurement error (e.g., using 5 as a pragmatic threshold), and conventional PLS-SEM structural collinearity guidance often treats VIF values below 5 as not critical. Accordingly, while the conservative FCVIF screen suggests possible CMB, the magnitudes observed (all < 5) do not indicate severe collinearity (Kock, 2015; Ringle et al., 2024).

Internal consistency, reliability, and convergent validity met the established criteria for all constructs (Table 2). CL demonstrated good reliability and convergent validity (Cronbach's α=0.864, rho_A=0.865, composite reliability=0.908, AVE=0.711). CE showed strong internal consistency (Cronbach's α=0.944, rho_A=0.946, composite reliability=0.950) with adequate convergent validity [adjusted convergent validity (AVE)=0.546]. The ORO demonstrated high reliability (Cronbach's α=0.943, rho_A=0.945, composite reliability=0.951, and AVE=0.686). RL also showed satisfactory reliability and convergent validity (Cronbach's α=0.838, rho_A=0.839, composite reliability=0.902, AVE=0.755) (Table 2). The higher-order CE construct (Stage 2) exhibited excellent reliability and convergent validity (Cronbach's α=0.940, composite reliability=0.957, AVE=0.848) (Table 2).

Table 2. Construct reliability and convergent validity (Stage 1 and higher-order CE constructs).

The heterotrait–monotrait ratio (HTMT) was used to evaluate discriminant validity (Henseler et al., 2015). All HTMT values were 0.95, indicating acceptable discriminant validity for conceptually related constructs, with the highest association observed between CE and RL (0.905) (Table 3).

Table 3. Discriminant validity (HTMT).

Structural model and hypothesis testing

Bootstrapping results for the direct effects are presented in Table 4. CE had a positive effect on ORO (CE → ORO: β=0.677, t=8.661, p < 0.001, f²=0.846), explaining R²=0.458 of the variance in ORO. CE also had a direct positive effect on CL (CE → CL: β=0.486, t=5.477, p < 0.001, f²=0.487) and RL (CE → RL: β=0.709, t=9.467, p < 0.001, f²=0.806) (Table 4). ORO positively predicted CL (ORO → CL: β=0.451, t=4.274, p < 0.001, f²=0.419), whereas its effect on RL was not supported (ORO → RL: β=0.144, t=1.700, p=0.089, f²=0.033) (Table 4). The model explained substantial variance in behavioral outcomes (CL: R²=0.737; RL: R²=0.662) (Table 5b).

Table 4. Results of the structural model and hypothesis testing.

Table 5. Effects of mediation and predictive assessment.

Mediation effects and predictive assessment

Indirect effects are summarized in Table 5a. The effect of CE on CL was significantly mediated by ORO [CE → ORO → CL: β=0.305, t=4.185, p < 0.001, 95% CI (0.149, 0.457)]), supporting H6 (Table 5a). In contrast, the indirect effect of CE on RL via ORO was not significant (CE → ORO → RL: β=0.097, t=1.564, p=0.118; the confidence interval included zero). Therefore, H7 was not supported (Table 5a).

As shown in Table 5b, the model demonstrated predictive relevance based on blindfolding, with Q² values of 0.307 ORO, 0.510 CL, and 0.482 RL. The out-of-sample predictive capability was also supported, with Q²_predict values of 0.423 ORO, 0.617 CL, and 0.640 RL (Table 5b).

---

Discussion

This research examined how CE, conceptualized as a reflective–reflective higher-order construct, influences ORO after feline orthopedic surgery and how CE and ORO subsequently shape C and RL. The model demonstrated meaningful explanatory power for perceived recovery and post-service behavioral intentions, indicating that owners’ appraisals of the care encounter are consequential not only for perceived functional recovery but also for retention and advocacy-related intentions. This aligns with health care quality frameworks that place experience alongside effectiveness and safety as a core dimension of quality (Institute of Medicine, 2001) and with service research linking perceived encounter quality to downstream behavioral intentions (Cronin and Taylor, 1992; Zeithaml et al., 1996).

Client experience as a driver of perceived recovery (H1)

Supporting H1, CE showed a strong positive association with ORO. Conceptually, this is consistent with the logic that the perioperative experience shapes owners’ interpretation of recovery through expectation-setting, understanding of the clinical trajectory, and confidence in home-care execution. In orthopedic pathways where outcomes may evolve over weeks to months, owners rely heavily on communication clarity, empathic support, and structured guidance to interpret progress and cope with uncertainty. This positioning echoes Donabedian’s structure–process–outcome framework, in which care processes influence outcomes that subsequently inform decisions and evaluations (Donabedian, 1988), and expectancy–disconfirmation theory, in which perceived performance relative to expectations drives downstream responses (Oliver, 1980). In veterinary settings, relationship-centered communication and empathy have repeatedly been linked to satisfaction, trust, and adherence—mechanisms that support better owner-perceived recovery through improved follow-through at home (Coe et al., 2008; Coe et al., 2009; Adams and Kurtz, 2017).

Why does CE predict loyalty and recommendation, but not identically (H2–H5)

As hypothesised, CE directly predicted both CL (H2) and RL (H3), with a notably stronger effect on recommendation. This pattern is consistent with service and relationship marketing perspectives in which the perceived quality of the service encounter shapes both repeat-use intentions and positive WOM (Cronin and Taylor, 1992; Zeithaml et al., 1996). In high-involvement services, such as surgery—where technical competence is difficult for owners to evaluate directly—experience-based cues (clarity, responsiveness, empathy, and involvement in decisions) may act as salient signals of reliability and professionalism. The higher-order structure of CE remained robust: the 4 CE dimensions loaded strongly on CE, supporting CE as a coherent but multidimensional quality domain. This implies that improving "experience" likely requires consistent performance across communication, empathic engagement, pain management processes, and shared decision-making rather than a single isolated tactic. ORO behaved differently across the two behavioral outcomes. Supporting H4, ORO significantly predicted loyalty, suggesting that perceived functional recovery and progress reinforce the decision to continue care with the same clinic—consistent with the idea that "staying" is strengthened by perceived outcomes and confidence in ongoing follow-up. In contrast, H5 was not supported: although ORO → RL was positive, it was not statistically significant at the conventional two-tailed p-value of 0.05. Rather than being merely a null result, this divergence is theoretically informative. Recommendation is a socially consequential act that carries reputational risk; therefore, owners may base advocacy more on trust cues and perceived integrity of the care process than on perceived recovery alone—particularly because orthopedic outcomes can vary with injury severity, adherence to home care, and rehabilitation time course. Accordingly, owners may recommend a clinic because the team was transparent, empathic, and supportive even when recovery was slow or complicated; conversely, favorable recovery may not automatically translate into advocacy if the experience involved confusion, inadequate updates, or friction in the service process. This aligns with broader insights that advocacy reflects the overall encounter and emotional tone of the relationship, beyond outcomes alone (Reichheld, 2003; Cheung and Thadani, 2012; Doyle et al., 2013).

Mediation: When recovery perceptions carry the effect of experience (H6–H7)

The mediation analysis clarifies how CE translates into behavioral intentions. Supporting H6, ORO significantly mediated the CE → CL relationship, and the direct CE → CL path remained significant, indicating partial mediation. This suggests that loyalty is shaped by both (i) experienced encounter quality and (ii) the owner’s interpretation of recovery. This mechanism is consistent with patient-reported (and proxy-reported) outcomes frameworks, where perceived outcomes function as an explanatory bridge between care processes and downstream behaviors (Reeve et al., 2013; Gagnier, 2017; Van Lieshout and Wijffels, 2020; Innes, 2023). In contrast, H7 was not supported: the indirect pathway CE → ORO → RL was not significant, reinforcing the interpretation that recommendation intention in this context is primarily experience-driven once CE is accounted for. Operationally, investments in trust-building behaviors across the encounter (clear updates, empathic support, structured shared decisions) may yield a more direct payoff in recommendation intentions than focusing solely on perceived recovery.

Interpreting the CE–RL proximity

The measurement results indicate that CE and RL are conceptually close—as recommendation intention often expresses the overall encounter evaluation—but the constructs remain theoretically and empirically distinguishable. Empirically, all HTMT values were 0.95, with the highest association between CE and RL (HTMT=0.905). This suggests a strong relationship, yet it remains within the < 0.95 threshold commonly used for conceptually related constructs. Importantly, the structural results still show differentiated nomological patterns—most notably, ORO predicts loyalty but not recommendation—supporting the practical distinctness of CL versus RL as outcomes.

Predictive relevance and robustness of the model

Beyond the explanatory power, the model exhibited meaningful predictive performance. The construct-level Q2 and Q2_predict values were positive across all endogenous constructs, indicating that the model has both in-sample and out-of-sample predictive relevance. Out-of-sample performance was further examined using PLSpredict and the CVPAT in SmartPLS. The CVPAT results showed that the PLS-SEM model consistently outperformed the indicator-average benchmark, as evidenced by the negative average loss differences and significant p-values across the constructs. When benchmarked against a linear model, CVPAT indicated significantly better predictive performance for ORO and RL, whereas the overall comparison was not significant at the p-value of 0.05. Taken together, these findings suggest that the proposed model provides practical predictive power for postoperative monitoring, clinical follow-up prioritization, and pathway-level quality improvement planning in feline orthopedic care (Shmueli et al., 2019; Liengaard et al., 2021).

Practical interpretation of feline orthopedic pathways

Taken together, the findings position CE as a clinically actionable care-process domain in orthopedic services for cats. Clinics can translate experience quality into concrete postoperative care behaviors by standardizing perioperative communication (milestones, analgesia plans, implant/rehabilitation rationale, and warning signs), strengthening shared decision tools, and reinforcing proactive follow-up and responsiveness (Table 6). Enhancing owner-perceived recovery appears to be particularly important for continuity and recheck adherence, whereas recommendation intention seems to be particularly sensitive to trust cues and expectation alignment across the pathway. In this sense, the findings support the improvement of postoperative orthopedic care rather than a narrow service-marketing interpretation. This finding is consistent with veterinary communication evidence that relationship-centered care and empathy are central to perceived quality, trust, and adherence (Coe et al., 2008; Coe et al., 2009; Adams and Kurtz, 2017).

Table 6. Clinical action points for postoperative orthopedic care and outcome monitoring after orthopedic surgery in cats.

Limitations and future research directions

Several limitations should be considered when interpreting these findings. First, despite procedural precautions, common method variance may still be present because all focal constructs were measured from the same respondents at a single time point; therefore, statistical screening (reported in the Results) should be interpreted as diagnostic rather than definitive (Podsakoff et al., 2003; Kock, 2015; Podsakoff et al., 2024). Residual CMB was assessed using Kock’s full collinearity approach; although several FCVIFs slightly exceeded the conservative 3.3 guideline, all values remained below 5, suggesting that collinearity was not severe but cannot be completely ruled out. Second, the cross-sectional, retrospective design precludes causal inference and is vulnerable to recall bias because owners evaluated both perioperative experience and postoperative function after the surgery had already occurred. Third, ORO was measured using the FMPI-sf as a proxy for postoperative function and comfort. Although the FMPI-sf covers clinically relevant observable mobility and comfort domains, it was originally refined for musculoskeletal pain-related impairment and is not a surgery-specific postoperative instrument for fracture healing or implant success. Accordingly, the ORO construct in this study should be interpreted as an owner-perceived functional recovery rather than a direct clinical efficacy endpoint. Fourth, recruitment used a voluntary online survey in a single referral clinic, and no denominator of recipients/viewers was available; thus, the response rate could not be calculated, and self-selection/nonresponse bias is possible. The sample may overrepresent owners who remained engaged with follow-up, were more digitally literate, or had more favorable experiences and outcomes. Fifth, the study did not incorporate objective postoperative indicators (e.g., radiographic union, complication grading, surgeon-assessed lameness/function, rehabilitation adherence, or standardized injury-severity measures) and did not model time-since-surgery or case complexity as covariates, which may confound owner-reported recovery perceptions and recommendation behavior. Finally, the single specialized clinical setting may constrain the generalizability to other practice types, regions, and case mixes. Future studies should use longitudinal and/or multi-source designs, include objective orthopedic recovery indicators alongside ORO, and evaluate moderators such as injury severity, complication burden, time since surgery, and home-care adherence. Future work should also test additional mechanisms linking recovery to advocacy, such as trust, perceived value, and communication quality during complications, given the non-significant ORO -> RL pathway.

---

Conclusion

This research clarifies how CE, specified as a reflective-reflective higher-order construct, relates to owner-perceived postoperative functional recovery and post-service behavioral intentions in feline orthopedic surgery. The results show that CE is strongly associated with owners’ recovery assessments and with both CL and RL. However, ORO primarily strengthened loyalty rather than recommendation, indicating that continuity and advocacy are related but distinct behavioral intentions shaped by partly different factors in this clinical setting. Mediation findings further suggest that owner-perceived recovery partially carries the effect of CE on loyalty, whereas recommendation remains more directly linked to the experience of the care process. Clinically, the findings support the prioritization of communication quality, empathy, pain management guidance, and shared decision-making as part of the postoperative orthopedic pathway design and follow-up monitoring. Because the study was cross-sectional, single-center, and used FMPI-sf as a postoperative proxy measure, the results should be interpreted as evidence on owner-perceived recovery and care-process quality rather than definitive proof of surgical outcome efficacy.

---

Acknowledgments

The authors would like to thank Anugerah Satwa Cat Care Center Veterinary Clinic for facilitating participant recruitment by distributing the questionnaire through the clinic’s routine client communication channels. The authors would also like to thank all participating cat owners for their time and responses.

Conflict of interest

The authors declare no conflict of interest.

Funding

This study received no specific grant.

Authors' contributions

AYK conceptualized and designed the study. AYK performed the clinical work and coordinated data collection. AYK and CM drafted the manuscript. AYK and IB performed all statistical analyses. All authors have critically revised the manuscript and approved the final version for submission.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

---

References

Adams, C.L. and Kurtz, S. 2017. Skills for communicating in veterinary medicine. Oxford, UK: Otmoor Publishing.

Aiyegbusi, O.L., Roydhouse, J., Rivera, S.C., Kamudoni, P., Schache, P., Wilson, R., Stephens, R. and Calvert, M. 2022. Key considerations to reduce or address respondent burden in patient-reported outcome (PRO) data collection. Nat. Commun. 13(1), 6026; doi:10.1038/s41467-022-33826-4

Applebaum, J.W., Maclean, E.L. and Mcdonald, S.E. 2021. Love, fear, and the human–animal bond: on adversity and multispecies relationships. Compr. Psychoneuroendocrinol. 7, 100071; doi: 10.1016/j.cpnec.2021.100071

Benito, J., DePuy, V., Hardie, E., Zamprogno, H., Thomson, A., Simpson, W., Roe, S., Hansen, B. and Lascelles, B.D.X. 2013. Reliability and discriminatory testing of a client-based metrology instrument, feline musculoskeletal pain index (FMPI), for the evaluation of degenerative joint disease-associated pain in cats. Vet. J. 196(3), 368–373; doi:10.1016/j.tvjl.2012.12.015

Bird, F.G. and De Vicente, F. 2020. Conservative management of sacroiliac luxation fracture in cats: medium- to long-term functional outcome. J. Feline Med. Surg. 22(6), 575–581; doi:10.1177/1098612X19867516

Cheung, C.M.K. and Thadani, D.R. 2012. The impact of electronic word-of-mouth communication: a literature analysis and integrative model. Decis. Support. Syst. 54(1), 461–470; doi:10.1016/j.dss.2012.06.008

Coe, J.B., Adams, C.L. and Bonnett, B.N. 2008. A focus group study of veterinarians’ and pet owners’ perceptions of veterinarian–client communication in companion animal practice. J. Am. Vet. Med. Assoc. 233(7), 1072–1080; doi:10.2460/javma.233.7.1072

Coe, J.B., Adams, C.L. and Bonnett, B.N. 2009. Prevalence and nature of cost discussions during clinical appointments in companion animal practice. J. Am. Vet. Med. Assoc. 234(11), 1418–1424; doi:10.2460/javma.234.11.1418

Cohen, J. 1988. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates.

Cronin, J.J. and Taylor, S.A. 1992. Measuring service quality: a reexamination and extension. J. Marketing. 56(3), 55–68; doi:10.2307/1252296

Decamp, C.E., Johnston, S.A., Déjardin, L.M. and Schaefer, S.L. 2016. Brinker, Piermattei and Flo’s handbook of small animal orthopedics and fracture repair. 5th ed. St. Louis, MO, USA: Saunders; doi: 10.1016/C2009-0-64185-4

Doyle, C., Lennox, L. and Bell, D. 2013. A systematic review of evidence on the links between patient experience and clinical safety and effectiveness. BMJ. Open. 3(1), e001570; doi:10.1136/bmjopen-2012-001570

Enomoto, M., Lascelles, B.D.X., Robertson, J.B. and Gruen, M.E. 2022. Refinement of the feline musculoskeletal pain index (FMPI) and development of the short-form FMPI. J. Feline. Med. Surg. 24(2), 142–151; doi:10.1177/1098612X211011984

Essner, A., Zetterberg, L., Hellström, K., Gustås, P., Högberg, H. and Sjöström, R. 2017. Validity and reliability properties of the canine brief pain inventory and the canine orthopedic index in dogs with osteoarthritis. Acta. Vet. Scand. 59, 44; doi:10.1186/s13028-017-0300-5

Evangelista, M.C., Watanabe, R., Leung, V.S.Y., Monteiro, B.P., O’Toole, E., Pang, D.S.J. and Steagall, P.V. 2019. Facial expressions of pain in cats: the development and validation of a Feline Grimace Scale. Scientific Rep. 9, 19128; doi:10.1038/s41598-019-55693-8

Gagnier, J.J. 2017. Patient-reported outcomes in orthopedics. J. Orthopedic Res. 35(10), 2098–2108; doi:10.1002/jor.23604

Glenn, O.J., Faux, I., Pratschke, K.M. and Bowlt Blacklock, K.L. 2024. Evaluation of a client questionnaire at diagnosing surgical site infections in an active surveillance system. Vet. Surg. 53(1), 184–193; doi:10.1111/vsu.14011

Groves, R.M. 2006. Nonresponse rates and nonresponse bias in household surveys. Public Opinion Quart. 70(5), 646–675; doi:10.1093/poq/nfl033

Gruen, M.E., Griffith, E.H., Thomson, A.E., Simpson, W. and Lascelles, B.D.X. 2015. Criterion validation testing of clinical metrology instruments for measuring degenerative joint disease associated mobility impairment in cats. PLos One. 10(7), 131839; doi:10.1371/journal.pone.0131839

Gruen, M.E., Myers, J.A.E., Tena, J.K.S., Becskei, C., Cleaver, D.M. and Lascelles, B.D.X. 2021. Frunevetmab, a felinized anti-nerve growth factor monoclonal antibody, for the treatment of pain from osteoarthritis in cats. J. Vet. Internal Med. 35(6), 2752–2762; doi:10.1111/jvim.16291

Hair, J.F., Hult, G.T.M., Ringle, C.M. and Sarstedt, M. 2022. A primer on partial least squares structural equation modeling (PLS-SEM. 3rd ed., Los Angeles, CA, USA: Sage.

Hair, J.F., Risher, J.J., Sarstedt, M. and Ringle, C.M. 2019. When to use and how to report the results of PLS-SEM. Eur. Bus. Rev. 31(1), 2–24; doi:10.1108/EBR-11-2018-0203

Henseler, J., Ringle, C.M. and Sarstedt, M. 2015. A new criterion for assessing discriminant validity in variance-based structural equation modeling. J. Acad. Marketing Sci. 43(1), 115–135; doi:10.1007/s11747-014-0403-8

Innes, J.F. How client-reported outcome measures can improve veterinary care. Today’s Veterinary Practice. Available via https://todaysveterinarypractice.com/practice-management/how-client-reported-outcome-measures-can-improve-veterinary-care/ (Accessed 13 February 2023).

Institute of Medicine. 2001. Crossing the Quality Chasm: a New Health System for the 21st Century. Washington, D.C. USA: The National Academies Press; doi: 10.17226/10027

Kock, N. 2015. Common method bias in PLS-SEM: a full collinearity assessment approach. Int. J. E-Collaboration. 11(4), 1–10; doi:10.4018/ijec.2015100101

Kock, N. and Lynn, G. 2012. Lateral collinearity and misleading results in variance-based SEM: an illustration and recommendations. J. Assoc. For Inf. Syst. 13(7), 546–580; doi:10.17705/1jais.00302

Lascelles, B.D.X., Hansen, B.D., Roe, S., Depuy, V., Thomson, A., Pierce, C.C., Smith, E.S. and Rowinski, E. 2007. Evaluation of client-specific outcome measures and activity monitoring to measure pain relief in cats with osteoarthritis. J. Vet. Internal Med. 21(3), 410–416; doi:10.1892/0891-6640(2007)21

Liengaard, B.D., Sharma, P.N., Hult, G.T.M., Jensen, M.B., Sarstedt, M., Hair, J.F. and Ringle, C.M. 2021. Prediction: coveted, yet forsaken? Introducing a cross-validated predictive ability test in partial least squares path modeling. Decis. Sci. 52(2), 362–392; doi:10.1111/deci.12445

Mathews, K., Kronen, P.W., Lascelles, D., Nolan, A., Robertson, S., Steagall, P.V., Wright, B. and Yamashita, K. 2014. Guidelines for recognition, assessment and treatment of pain. J. Small. Anim. Pract. 55(6), E10–E68; doi:10.1111/jsap.12200

Meeson, R.L. and Geddes, A.T. 2017. Management and long-term outcome of pelvic fractures: a retrospective study of 43 cats. J. Feline. Med. Surg. 19(1), 36–41; doi:10.1177/1098612X15606958

Montavon, P.M., Voss, K. and Langley-Hobbs, S.J. 2009. Feline orthopedic Surgery and Musculoskeletal Disease. Edinburgh, UK: Mosby Elsevier

Oliver, R.L. 1980. A cognitive model of the antecedents and consequences of satisfaction decisions. J. Marketing Res. 17(4), 460–469; doi:10.2307/3150499

Podsakoff, P.M., Mackenzie, S.B., Lee, J.Y. and Podsakoff, N.P. 2003. Common method biases in behavioral research: a critical review of the literature and recommended remedies. J. Appl. Psychol. 88(5), 879–903; doi:10.1037/0021-9010.88.5.879

Podsakoff, P.M., Podsakoff, N.P., Williams, L.J., Huang, C. and Yang, J. 2024. Common method bias: it’s bad, it’s complex, it’s widespread, and it’s not easy to fix. Annu. Rev. Organizational Psychol. Organizational Behav. 11, 17–61; doi:10.1146/annurev-orgpsych-110721-040030

Radke, H., Joeris, A. and Chen, M. 2022a. Evidence-based evaluation of owner-reported outcome measures for canine orthopedic care: a COSMIN evaluation of 6 instruments. Vet. Surg. 51(2), 244–253; doi: 10.1111/vsu.13753

Radke, H., Zhu, T.Y., Knoll, C., Allen, M.J. and Joeris, A. 2022b. Owner-reported outcome measures in veterinary care for companion animal orthopedic patients: an international online survey of veterinarians’ expectations and practices. Vet. Surg. 51(6), 903–913; doi: 10.1111/vsu.13842

Reeve, B.B., Wyrwich, K.W., Wu, A.W., Velikova, G., Terwee, C.B., Snyder, C.F., Schwartz, C., Revicki, D.A., Moinpour, C.M., Mcleod, L.D., Lyons, J.C., Lenderking, W.R., Hinds, P.S., Hays, R.D., Greenhalgh, J., Gershon, R., Feeny, D., Fayers, P.M., Cella, D., Brundage, M., Ahmed, S., Aaronson, N.K. and Butt, Z. 2013. ISOQOL recommends minimum standards for patient-reported outcome measures used in patient-centered outcomes and comparative effectiveness research. Qual. Life Res. 22(8), 1889–1905; doi:10.1007/s11136-012-0344-y

Reichheld, F.F. 2003. The one number you need to grow. Harvard Bus. Rev. 81(12), 46–54. Available via https://pubmed.ncbi.nlm.nih.gov/14712543/

Reinharth, D. 1988. The quality of care: how can it be assessed?. J. Am. Med. Assoc. 260(12), 1743–1748; doi:10.1001/jama.260.12.1743

Ringle, C.M., Wende, S. and Becker, J.M. 2024. SmartPLS 4. Bönningstedt, Germany: SmartPLS.

Rohlf, V.I., Manfrenuzzi, N., Rehman, N. and Bennett, P.C. 2025. Exploration of client experiences of veterinary care for companion animals (dogs and cats) and the impact of trauma-informed care on client outcomes. Vet. Sci. 12(8), 709; doi:10.3390/vetsci12080709

Shmueli, G., Sarstedt, M., Hair, J.F., Cheah, J.H., Ting, H., Vaithilingam, S. and Ringle, C.M. 2019. Predictive model assessment in PLS-SEM: guidelines for using PLSpredict. Eur. J. Marketing. 53(11), 2322–2347; doi:10.1108/EJM-02-2019-0189

Steagall, P.V., Robertson, S., Simon, B., Warne, L.N., Shilo-Benjamini, Y. and Taylor, S. 2022. 2022 ISFM consensus guidelines on the management of acute pain in cats. J. Feline Med. Surg. 24(1), 4–30; doi:10.1177/1098612X211066268

Sutherland, K.A., Coe, J.B., Blais-Vaillancourt, K., Gober, M. and Berst, H. 2025. Veterinary clients prefer benefit-focused online communication while clinic websites uncommonly communicate benefits of preventive care services. J. Am. Vet. Med. Assoc. 263(1), 113–121; doi:10.2460/javma.24.09.0568

Terwee, C.B., Prinsen, C.A.C., Chiarotto, A., Westerman, M.J., Patrick, D.L., Alonso, J., Bouter, L.M., De Vet, H.C.W. and Mokkink, L.B. 2018. COSMIN methodology for evaluating the content validity of patient-reported outcome measures: a Delphi study. Qual. Life Res. 27(5), 1159–1170; doi:10.1007/s11136-018-1829-0

Turkson, P.K. 2011. Predictors of client satisfaction with veterinary services in a mixed veterinary practice in Ghana. Vet. Med. Int. 2011, 321369; doi:10.4061/2011/321369

Turkson, P.K., Coppson, C.N.D. and Turkson, A.J. 2025. Indicators and dimensions of clients’ satisfaction with veterinary services delivery performance in a companion animal hospital in Accra, Ghana. Vet. Med. Sci. 11(3), e70331; doi:10.1002/vms3.70331

Van Lieshout, E.M.M. and Wijffels, M.M.E. 2020. Patient-reported outcomes: which ones are most relevant?. Injury 51(Suppl 2), S37–S42; doi:10.1016/j.injury.2019.10.073

Yap, F.W., Dunn, A.L., Farrell, M. and Calvo, I. 2014. Trans-iliac pin/bolt/screw internal fixation for sacroiliac luxation or separation in cats: six cases. J. Feline. Med. Surg. 16(4), 354–362; doi:10.1177/1098612X13503650

Yap, F.W., Dunn, A.L., Garcia-Fernandez, P.M., Brown, G., Allan, R.M. and Calvo, I. 2015. Femoral head and neck excision in cats: medium- to long-term functional outcome in 18 cats. J. Feline. Med. Surg. 17(8), 704–710; doi:10.1177/1098612X14556848

Zeithaml, V.A., Berry, L.L. and Parasuraman, A. 1996. The behavioral consequences of service quality. J. Marketing. 60(2), 31–46; doi:10.2307/1251929