Psychometric properties of the Canine Symptom Assessment Scale, a multidimensional owner-reported questionnaire instrument for assessment of physical symptoms in dogs with solid tumors

Michelle A. Giuffrida Department of Clinical Studies–Philadelphia, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA 19104.

Search for other papers by Michelle A. Giuffrida in
Current site
Google Scholar
PubMed
Close
 VMD, MSCE
,
John T. Farrar Center for Clinical Epidemiology and Biostatistics, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA 19104.

Search for other papers by John T. Farrar in
Current site
Google Scholar
PubMed
Close
 MD, PhD
, and
Dorothy Cimino Brown Department of Clinical Studies–Philadelphia, School of Veterinary Medicine, University of Pennsylvania, Philadelphia, PA 19104.

Search for other papers by Dorothy Cimino Brown in
Current site
Google Scholar
PubMed
Close
 DVM, MSCE

Click on author name to view affiliation information

Abstract

OBJECTIVE To describe development and initial psychometric testing of the Canine Symptom Assessment Scale (CSAS), a multidimensional owner-reported questionnaire instrument, in a population of dogs with solid tumors enrolled in clinical trials.

DESIGN Questionnaire development and validation study.

ANIMALS 238 client-owned dogs with solid tumors.

PROCEDURES A 14-symptom questionnaire was developed. Symptoms were defined as subjective physical disturbances dogs experienced during the course of daily living as assessed through proxy reports of pet owners. For each symptom, owners reported frequency and severity of the symptom and extent of distress caused by the symptom for the dog and the owner. Questionnaire content, symptom prevalence and dimensionality, internal consistency, and factor structure were examined. Construct and criterion validity were examined via comparison with the Canine Brief Pain Inventory (CBPI).

RESULTS Symptom prevalence was high, with pain and lack of energy reported in most dogs. Severity, versus frequency, was most highly correlated with both dog and owner distress. Two symptoms were removed from consideration because of poor performance. Analysis of the remaining 12 symptoms revealed that they could be grouped into 3 factors: malaise, anxiety, and digestive upset. The CSAS factor and total scores demonstrated predictable relationships with quality of life and pain scores as measured by the CBPI, including a significant association between increasing symptom burden and decreasing quality of life. The Cronbach α for the CSAS was 0.77.

CONCLUSIONS AND CLINICAL RELEVANCE The 12-item CSAS was a psychometrically sound owner-reported instrument for assessment of symptom frequency and characteristics in client-owned dogs with solid tumors. Potential applications include clinical research and practice settings.

Abstract

OBJECTIVE To describe development and initial psychometric testing of the Canine Symptom Assessment Scale (CSAS), a multidimensional owner-reported questionnaire instrument, in a population of dogs with solid tumors enrolled in clinical trials.

DESIGN Questionnaire development and validation study.

ANIMALS 238 client-owned dogs with solid tumors.

PROCEDURES A 14-symptom questionnaire was developed. Symptoms were defined as subjective physical disturbances dogs experienced during the course of daily living as assessed through proxy reports of pet owners. For each symptom, owners reported frequency and severity of the symptom and extent of distress caused by the symptom for the dog and the owner. Questionnaire content, symptom prevalence and dimensionality, internal consistency, and factor structure were examined. Construct and criterion validity were examined via comparison with the Canine Brief Pain Inventory (CBPI).

RESULTS Symptom prevalence was high, with pain and lack of energy reported in most dogs. Severity, versus frequency, was most highly correlated with both dog and owner distress. Two symptoms were removed from consideration because of poor performance. Analysis of the remaining 12 symptoms revealed that they could be grouped into 3 factors: malaise, anxiety, and digestive upset. The CSAS factor and total scores demonstrated predictable relationships with quality of life and pain scores as measured by the CBPI, including a significant association between increasing symptom burden and decreasing quality of life. The Cronbach α for the CSAS was 0.77.

CONCLUSIONS AND CLINICAL RELEVANCE The 12-item CSAS was a psychometrically sound owner-reported instrument for assessment of symptom frequency and characteristics in client-owned dogs with solid tumors. Potential applications include clinical research and practice settings.

Owner reports of their pets' subjective physical disturbances are fundamental to evaluating disease burden, treatment tolerability, and overall QOL in animals with cancer. Subjective physical disturbances experienced by human patients are termed symptoms,1 whether they are reported by the patients themselves or by proxies such as family members or other caregivers.2–6 The word symptom is conventionally avoided in the veterinary scholarly literature in favor of clinical sign, but this is often not an apt substitution because the latter term pertains to objective indications of disease discoverable on examination7 and as interpreted by the clinician rather than by the patient or a lay person. In fact, the assessment of dogs' health often includes subjective evidence that fits the definition of symptoms,8 typically in the form of observations made by pet owners that are not apparent on examination by a veterinarian. For example, a veterinarian can identify physical signs of dehydration; however, the examination may not reveal a history of diarrhea, vomiting, or decreased water intake, let alone the frequency or severity of such disturbances. A growing body of evidence9–15 indicates that subjective experiences not only exist in dogs, but can be measured with proxy reports from pet owners. As such, subjective physical disturbances dogs may have experienced during the course of daily living as assessed by pet owners are referred to hereafter as symptoms.

The traditional method of symptom assessment in human clinical trials involves a continuum of information transfer—from patient or proxy to clinician to medical record to study database to manuscript—with reinterpretation at each step.16,17 Clinicians filter symptom information subjectively and tend to interpret physical symptoms as being less severe or less impactful, compared with what patients actually experience.18 The result is systematic underreporting of early and mild symptoms17,19,20 and a body of information that reflects the perspective of medical personnel rather than that of the patients themselves.21 A parallel situation likely exists in veterinary medicine. Recently, our group22 conducted a systematic review of adverse event reporting in veterinary clinical trials of cancer treatments. Results indicated that only 19% (32/168) of recently published trials actively collected any data pertaining to symptoms of the patients studied.22 Furthermore, the methodology used to ascertain and grade animals' symptoms almost exclusively relied on clinician adjudication. These findings suggest that, as for human patients, underreporting of symptoms is commonplace in veterinary clinical trials of cancer treatments. Importantly, when symptoms are not accurately represented, biased assessments of treatment risks and benefits may occur.23

In an effort to address this problem in human clinical trials, researchers have advocated for more widespread use of patient-reported symptom assessments.24 With this approach, study subjects provide firsthand information about their symptoms, often via structured questionnaires; questionnaire answers are converted to numeric values and entered directly into trial databases.21,24 Patient reports are considered an adjunct to clinician reports, enabling earlier, more accurate, and more comprehensive detection of symptoms.21 Such patient data can also improve understanding of how symptoms interfere with patients' daily activities or QOL.25 In veterinary medicine, pets cannot provide direct reports of their subjective experiences, but pet owners are considered valid proxies for reporting observable events such as physical symptoms.9,10 Although there are validated owner-reported instruments to measure subjective states in companion animals, including pain9,11,12 and QOL,13 little attention has been paid to methods of symptom assessment. Three owner-reported questionnaire instruments for measurement of adverse gastrointestinal events have been described.14,15 These instruments are best suited to clinical rather than research applications because of inclusion of few symptoms, lack of psychometric testing, and reliance on open-ended responses that cannot easily be quantified. Existing questionnaires focus on frequency and treatment of symptoms, but do not attempt to quantify how severe a symptom is or how much distress (or lack thereof) that an animal or pet owner experiences as a result of it. These unmeasured dimensions are likely to impact pet QOL and health-care decisions, particularly when the pet's long-term prognosis is guarded. Additionally, whereas many useful symptom assessment instruments exist for human patients, none are immediately applicable to veterinary patients because of differences in cognition and activities of daily living between species. In our view, the ability of veterinary clinical trials to capture comprehensive, objective measures of pet symptomology and its impact on QOL has been hindered by a lack of validated symptom assessment tools.

Therefore, the objective of the study reported here was to describe the development of the CSAS, an owner-reported questionnaire instrument adapted from the MSAS26 for human patients, and to report the results of initial psychometric testing in pet dogs with solid tumors. We hypothesized that the CSAS could be used to describe the prevalence and dimensions (ie, frequency, severity, and distress) of dogs' symptoms and that predictable relationships between symptom dimensions could be identified. We further hypothesized that the CSAS would have good internal consistency (Cronbach α > 0.70), that exploratory factor analysis would reveal a clinically meaningful structure involving 1 or more factors, and that CSAS scores would be highly and predictably correlated (r > 0.6) with validated measures of pain and QOL.

Materials and Methods

CSAS development

The CSAS was designed to provide quantitative, multidimensional assessment of common physical symptoms experienced by pet dogs with a variety of underlying disease conditions. Standard methods were followed for the development of questionnaire instruments intended to measure subjective health states.27 Item structure, questionnaire format, and response scaling were based on the MSAS, a validated multidimensional instrument that assesses the frequency and severity of and distress associated with physical and psychological symptoms in human patients.26 It has been used extensively in studies of human cancer patients28 and is considered a valid measure when used by patient proxies, particularly for assessment of objective physical symptoms.29

An initial pool of symptoms commonly associated with cancer in pet dogs was selected following a review of the literature pertaining to assessment of physical symptoms in dogs and humans. Key-informant interviews with pet owners and veterinarians were used to refine the pool to symptoms that met the following criteria: relevant to dogs with a variety of different diseases and treatment plans, indicative of disease presence or progression or the adverse effects of treatment, and potentially associated with distress, suffering, or diminished QOL. A pet owner–completed questionnaire instrument was constructed from the symptom list. Respondents were asked to indicate (yes or no) whether the dog experienced each symptom during the past 10 days. For each symptom the dog experienced, 4 dimensions were evaluated with separate 4- or 5-point Likert scales: frequency with which the symptom occurred, severity of the symptom, how much the symptom distressed or bothered the dog, and how much the symptom distressed or bothered the pet owner. Space was provided for respondents to write in additional symptoms and to rate each symptom according to the 4 dimensions.

Patient selection and questionnaire administration

Client-owned dogs with spontaneously occurring solid tumors enrolled in 5 single-center clinical trials of cancer treatments at the Veterinary Clinical Investigations Center of the University of Pennsylvania School of Veterinary Medicine between November 2006 and November 2014 were eligible for inclusion in the present study. The study protocol, including all questionnaires, was approved by the Institutional Review Board and the Institutional Animal Care and Use Committee of the University of Pennsylvania. An independent group of dogs was enrolled in each clinical trial, and all patients were evaluated longitudinally by means of consecutive study visits from the time of enrollment until death, euthanasia, or loss to follow-up. All pet owners provided written informed consent prior to inclusion of their dogs in the trial. Patient signalment data including date of birth, sex (including sterilization status), breed, and body weight were recorded with standardized forms. At each study visit, the pet owner completed the CSAS and the CPBI. The CPBI is an owner-reported questionnaire instrument for measurement of the severity and impact of chronic pain in pet dogs.9,30 It consists of 4 items pertaining to the severity of the dog's pain, 6 items pertaining to how pain interferes with the dog's daily activities, and 1 global QOL item. Both questionnaires were self-administered by pet owners in pen-and-paper format. Trained clinical trial nurses distributed questionnaires to pet owners, instructed pet owners on how to properly complete the instruments, collected completed forms, and were available to answer questions if required. Nurses checked each completed questionnaire for missing or incomplete items and returned forms to pet owners for completion when necessary.

Questionnaires were administered at multiple hospital visits; however, for the purposes of initial validation of the CSAS, data from a single visit per dog were analyzed. For dogs undergoing radiation therapy, data from the study visit that corresponded with completion of 75% of the planned radiation protocol were used. For dogs undergoing other types of cancer treatment, data from the third study visit were analyzed. These study visits were chosen in an effort to avoid response errors that could occur when a respondent (ie, pet owner) was completing the questionnaire for the first time and to ensure that all pets had begun treatment and could potentially have experienced adverse effects. If data for the desired study visit were not available, questionnaire data from the most recent prior visit were used. Dogs with data for only 1 visit were excluded from the study.

Statistical analysis

Descriptive statistics were calculated for patient demographic variables. Categorical data were reported as the number (percentage) of patients. Continuous data were evaluated for normality with histograms and quantile-quantile plots and reported as either the mean (SD) or median (range). Symptom prevalence in the entire study population was calculated for each questionnaire item. Distributions of frequency, severity, and dog and owner distress were calculated for each symptom among the dogs that experienced that symptom. Spearman rank correlation coefficients were used to evaluate the independence and importance of the 4 symptom dimensions (ie, frequency, severity, dog distress, and owner distress) for each symptom. Responses to open-ended questionnaire items about additional symptoms were tabulated. Pet owner and dog distress scores for each symptom were compared with the Wilcoxon signed rank test.

Questionnaires were scored by means of a simple method adapted from the MSAS.26 Briefly, for each symptom, an item score between 0 and 4 was generated according to whether the symptom occurred and pursuant to the frequency, severity, and distress responses. Symptoms that did not occur were scored as 0. Symptoms that did occur were scored on the basis of equally weighted frequency, severity, pet distress, and owner distress responses. Frequency and severity were measured on a 4-point Likert scale and scored as 1, 2, 3, or 4, with higher scores indicating increasing symptom frequency and severity. Dog and owner distress were measured on a 5-point Likert scale and scored as 0.8, 1.6, 2.4, 3.2, or 4, with higher scores indicating increasing distress. The item score for each symptom was the sum of its 4 dimension scores divided by 4. The total CSAS score was the sum of the item scores. The Cronbach α was used to evaluate how item scores correlated with one another and with the total CSAS score. The interitem correlation matrix and item-total score correlations were used to screen for items with weak or negative correlations. Low-prevalence items were omitted from further analysis if their removal improved internal consistency and was supported by results of content analysis. Exploratory principal components factor analysis of symptom scores with subsequent oblique oblimin rotation was used to evaluate factor structure for meaningful groupings of patient symptoms.31,32 Individual scores for identified factors were computed with the sum of scores for items loading on each factor. Construct validity was assessed by calculating Spearman rank correlation coefficients (ρ) for factor-factor, factor-CSAS, factor-CBPI, and CSAS-CBPI correlations. Specific hypotheses for construct validation testing were formulated after the factor structure was explored, but were determined on the basis of a working theory suggesting that groups of symptoms corresponding to different body systems or syndromes would not exhibit strong correlations with one another, that all groups of symptoms would display an inverse relationship with QOL, and that only pain-related symptoms would correlate strongly with results of the CPBI. Correlations between CSAS and CBPI total and subscale scores also served as measures of criterion validation. The nonparametric test for trend was used to test the hypothesis that CSAS total and factor scores were inversely related to QOL, as reported for the CBPI.9,30

To examine the influence of owner distress on the performance and psychometric properties of the CSAS, a second set of scores was generated with identical methodology, but only frequency, severity, and dog distress were considered. Psychometric testing as described was also performed with this set of dog-only scores. All analyses were performed with statistical softwarea; values of P < 0.05 were considered significant.

Results

CSAS development

An initial pool of symptom items was generated and refined with the methods described. Fourteen observable symptoms were selected for inclusion in the CSAS: pain, lack of energy, panting, cough, pacing, difficulty sleeping, aggression, sleepiness, vomiting, diarrhea, lack of appetite, constipation, yelping-crying out, and moaning-groaning-whining. The questionnaire was formatted as a single-page pen-and-paper instrument with a Flesch Reading Ease score of 80.9 and Flesch Kincaid grade level score of 5.8; these scores were indicators that the questionnaire was easy to read and understand.33

Patient selection

There were 241 eligible dogs enrolled in 5 active clinical trials of cancer treatments taking place at the Veterinary Clinical Investigations Center between November 2006 and November 2014, but 3 dogs with oral tumors were withdrawn from their trials prior to initiation of treatment and were excluded. Therefore, data from 238 dogs were analyzed for the present study (Table 1).

Table 1—

Characteristics of 238 client-owned dogs with solid tumors enrolled in a study describing the development and initial psychometric testing of a multidimensional owner-reported questionnaire instrument (CSAS) for assessment of frequency and severity of and dog and owner distress associated with various physical symptoms in dogs with cancer.

VariableNo. (%)
Age (y)*9 (7–10)
Male150 (58.8)
Neutered217 (91.2)
Breed or type of dog
  Labrador Retriever32 (13.5)
  Golden Retriever27 (11.3)
  Rottweiler21 (8.8)
  German Shepherd Dog10 (4.2)
  Boxer8 (3.6)
  Other toy or small purebred dog29 (12.2)
  Other large or giant purebred dog47 (19.7)
  Mixed breed64 (27.9)
  Weight (kg)*35 (27–42)
Type of solid tumor
  Malignant bone tumor117 (49.2)
  Malignant oral tumor82 (34.4)
  Splenic hemangiosarcoma39 (16.4)
Clinical trial treatment intervention
  Medical84 (35.3)
  Radiation82 (34.4)
  Palliative72 (30.3)

Dogs enrolled in 5 single-center clinical trials of cancer treatments at the Veterinary Clinical Investigations Center of the University of Pennsylvania between November 2006 and November 2014 were eligible for study inclusion.

Data are given as median (range).

Analysis of CSAS symptoms and dimensionality

Pain, lack of energy, and panting were the most prevalent symptoms, occurring in > 50% of dogs; constipation and aggression were the least prevalent symptoms, occurring in < 10% of dogs (Table 2). There was substantial variability in the frequency, severity, and distress associated with each symptom. Among dogs that experienced a symptom, frequency and severity were moderately to highly correlated for all symptoms except diarrhea (Table 3). In general, symptom severity, rather than frequency, was more highly correlated with dog and owner distress. Dog and owner distress were highly correlated for all symptoms. Owners reported feeling greater distress about the symptoms than they perceived their dog to feel; owner distress scores were significantly (P ≤ 0.001) higher for all symptoms except aggression (P = 0.046) and constipation (P = 0.145).

Table 2—

Prevalence and dimensionality of symptoms assessed with the CSAS for the patients in Table 1.

  Prevalence (%) when symptom was present
SymptomOverall prevalence (%)Frequency*IntensityDog distressOwner distress
Pain75.243.619.654.273.2
Lack of energy74.437.915.832.252.5
Panting59.735.96.319.034.5
Sleepiness42.444.611.914.932.7
Whining, moaning, and groaning28.622.17.432.457.4
Difficulty sleeping27.323.19.238.556.9
Lack of appetite27.333.823.129.270.8
Cough23.17.37.312.730.9
Pacing20.620.48.230.649.0
Diarrhea16.820.015.027.540.0
Yelping, crying out14.72.914.351.465.7
Vomiting14.35.911.835.352.9
Constipation6.320.06.740.040.0
Aggression5.50.07.738.561.5

Percentage of dogs in which the symptom was frequently or constantly present.

Percentage of dogs in which the symptom was of moderate or severe intensity.

Percentage of dogs in which the dog or owner was somewhat, quite a bit, or very much distressed because of the symptom.

Table 3—

Correlations (Spearman rank correlation coefficients) among dimensionality of symptoms, as measured by the CSAS, for the dogs in Table 1.

SymptomNo. of dogs with symptomFrequency vs severityFrequency vs dog distressFrequency vs owner distressSeverity vs dog distressSeverity vs owner distressDog distress vs owner distress
Pain1790.730.620.620.740.650.75
Lack of energy1770.740.530.580.620.630.73
Panting1420.620.270.350.580.530.66
Sleepiness1010.650.390.450.480.580.65
Whining, groaning, moaning680.550.300.270.590.400.67
Difficulty sleeping650.660.710.640.710.710.82
Lack of appetite650.610.250.610.410.680.57
Cough550.530.530.530.600.570.62
Pacing490.660.340.430.500.560.73
Diarrhea400.330.150.120.440.510.73
Yelping, crying out350.520.380.430.820.690.84
Vomiting340.580.520.630.600.660.83
Constipation150.740.350.390.400.650.71
Aggression130.600.280.150.520.500.54
Mean 0.610.390.440.570.590.70

Additional symptoms were written in by owners of 47 of the 238 (19.7%) dogs. These included the following: hind limb weakness (n = 5), limping (5), swelling of the limb or paw (4), abnormal urination pattern (4), sneezing (4), excessive drinking (3), snorting (2), rubbing face (2), eating grass (1), anxiety (1), blood in stool (1), depressed (1), difficulty picking up toys or food (1), drooling (1), dropping stool (1), eye irritation (1), inverted breathing (1), leg shaking (1), looks unstable (1), odor (1), not playing with other dogs (1), sudden panting in the middle of the night (1), nausea (1), seizure (1), gas (1), epistaxis (1), nasal swelling (1), retching (1), and wheezing (1). On the basis of these data, weakness, limping, abnormal urination, excessive drinking, sneezing, and face-rubbing were considered candidates for inclusion in future versions of the CSAS.

Psychometric testing

Because performance and results of psychometric testing were similar for total versus dog-only scores, with almost identical internal consistency and factor solution, only results obtained with total scores were included. Two items, aggression and constipation, performed poorly, with low prevalence and low item-test and inter-item correlations (ρ < 0.20). These items were removed from further analysis. Exploratory factor analysis was performed with scores for the remaining 12 symptoms. A 3-factor solution emerged in which each factor contained meaningful groupings of symptoms and had an Eigenvalue > 1.0, and each item loaded on a single factor (Table 4). The total amount of the variance accounted for by the factors was 57%. The 3 factors were termed malaise (pain, lack of energy, sleepiness, moaning-groaning-whining, and yelping-crying out), anxiety (panting, pacing, and difficulty sleeping), and digestive upset (lack of appetite, vomiting, diarrhea, cough). The Cronbach α was 0.774 for the CSAS instrument, indicating a high degree of internal consistency. The Cronbach α was 0.724 for the malaise factor but was less robust for the anxiety (0.610) and digestive (0.559) factors, suggesting that the anxiety score and digestive score may not be reliable as independent subscales. Individual item α values and item-test correlations indicated that further omission of items would not have improved internal consistency of the overall instrument or its factors. Factors correlated strongly with the total score (ρ = 0.61 to 0.89) and moderately with each other (ρ = 0.25 to 0.48). Once the factor structure of the 12-item CSAS (Supplementary Appendix S1, available at avmajournals.avma.org/doi/suppl/10.2460/javma.251.12.1405) was identified, hypotheses were formulated for construct and criterion validation. Because the CBPI is a measure of pain, construct and criterion validation would be supported by positive correlations between CBPI pain scores and total CSAS and malaise factor scores, but weaker correlations between CBPI pain scores and anxiety and digestive factor scores. Construct validation would also be supported by a negative correlation between total CSAS score and global QOL score and between each CSAS subscale score and global QOL score. As hypothesized, strong positive correlations were observed between CBPI subscale scores and the CSAS total score and malaise factor score, whereas weaker correlations were observed between CBPI subscale scores and anxiety and digestive factor scores (Table 5). Global QOL had a strong negative correlation with CSAS total and malaise factor scores and moderate negative correlations with anxiety and digestive factor scores. With the nonparametric test for trend, QOL, as assessed on a 5-item Likert scale, was inversely related to CSAS total score (P < 0.001) and to each factor score (P < 0.001; Figure 1).

Figure 1—
Figure 1—

Box-and-whisker plots for client-owned dog enrolled in the study in Table 1. The lower and upper limits of the boxes represent the 25th and 75th percentiles, the horizontal line within each box represents the median value, the whiskers indicate the 10th and 90th percentiles, and the dots represent outliers. The graphs illustrate evaluation of construct and criterion validity of the CSAS by means of correlation of total and individual factor (ie, malaise, digestive, and anxiety) scores with the global QOL score (assessed with a 5-item Likert scale) from the CBPI.9,30 Lower CSAS total and factor scores were significantly (P < 0.001) associated with higher QOL scores for all graphs as indicated by the nonparametric test for trend.

Citation: Journal of the American Veterinary Medical Association 251, 12; 10.2460/javma.251.12.1405

Table 4—

Factors, item loadings, and item and factor correlations for the 12-item CSAS as measured for the patients in Table 1.

FactorFactor loadings*UniquenessItem-factor correlationFactor-scale correlation
Malaise factor   0.89
  Pain0.7670.3710.82 
  Lack of energy0.8160.3120.79 
  Sleepiness0.5950.6070.60 
  Moaning, groaning, and whining0.5310.4660.54 
  Yelping, crying out0.5450.5460.49 
Anxiety factor   0.73
  Panting0.4850.6030.58 
  Pacing0.8070.3780.81 
  Difficulty sleeping0.5940.3940.73 
Digestive factor   0.64
  Lack of appetite0.6160.3720.73 
  Vomiting0.6510.5420.51 
  Diarrhea0.4890.6680.51 
  Cough0.5490.4630.57 

Factor loadings are the relationships between syndromes and their factors, accounting for the correlations between factors.

Uniqueness is the proportion of variance that is specific to the syndrome (not the factor).

Values are Spearman rank correlation coefficients.

Table 5—

Correlations (Spearman rank correlation coefficients) between CSAS total and individual factor scores and CBPI subscale and global QOL scores for the patients in Table 1.

 CBPI scores
CSAS scoresPain subscaleInterference subscaleGlobal QOL
12-item total0.680.66−0.68
Malaise factor0.740.71−0.68
Anxiety factor0.390.35−0.35
Digestive factor0.270.28−0.37

Discussion

The CSAS developed in the present study was a multidimensional questionnaire instrument that measured frequency and severity of and patient and owner distress associated with symptoms experienced by pet dogs with cancer, as assessed by means of proxy reports of pet owners. The 12-item CSAS was psychometrically sound and exhibited good internal consistency, construct validity, and criterion validity in a population of dogs with solid tumors enrolled in clinical trials of cancer treatment. On the basis of results of the present study, we suggest that the CSAS is a valuable instrument that addresses an unmet need in veterinary research and may also serve as a useful clinical tool. We anticipated results of this initial study (eg, reliability and validation testing data) would help guide changes to the instrument and assist formulation of hypotheses for future psychometric testing.

Overall symptomology was high for patients evaluated in the present study. Pain and lack of energy were each reported in approximately 75% of the 238 dogs included in the study, and these are also the most frequently reported physical symptoms among human cancer patients as determined with the MSAS.26,28 Pain is underdiagnosed and consequently undertreated in dogs with cancer.34 Cancer itself can cause pain, or pain can occur as a consequence of diagnostic testing, treatment, or the presence of coexisting conditions such as osteoarthritis. Whereas valid instruments exist to measure pain in dogs, clinicians and investigators might not think to use them unless severe pain is anticipated. As a general symptomology instrument, we suggest that use of the CSAS would ensure that pain is evaluated even when veterinarians are not anticipating it. Its validity in this capacity was supported by strong correlation with scores for the CPBI, a pain measurement tool known to be valid and reliable in dogs with pain associated with cancer. Cancer-related fatigue is also highly prevalent in human patients during and after cancer treatment.35 Cancer can cause fatigue as a result of direct effects on metabolism and muscle strength, but fatigue can also occur because of sleep disruption, nutritional deficiencies, emotional distress, and pain.35 As for human patients, proactive identification and treatment of these factors could improve QOL in dogs with cancer. The CSAS developed in the present study provides a means to evaluate not only whether fatigue is present, but also whether pet owners have observed other treatable symptoms that could be contributing to fatigue (eg, pain, inappetence, and difficulty sleeping).

Owners of dogs enrolled in the present study reported that both they and their dogs were considerably bothered or distressed by the dogs' symptoms. Pain, lack of appetite, and yelping or crying out were the most distressing events, whereas panting, sleepiness, and cough were the least distressing. Although cough might seem out of place in the digestive upset factor of the CSAS, it is in fact a common sign of gastroesophageal reflux disease in dogs and humans.36,37 The high degree of correlation between dog and owner distress suggested anthropomorphizing, but response patterns nevertheless indicated that owners were able to distinguish between the dog's experience and their own psychological state. Symptoms such as lack of energy and sleepiness were reported to be minimally distressing to dogs, but worrisome to pet owners. Presumably, this occurred because humans, but not dogs, recognized these symptoms as potential indicators of disease progression, triggering psychological distress in anticipation of the pet's future suffering, need for treatment, and death or a combination of these factors. The generally heightened level of distress reported by pet owners, compared with that for their dogs, also likely reflected owner emotional and psychological responses to pet symptoms.

Aggression and constipation were omitted from the CSAS because of poor performance during initial psychometric testing. Although we expected that constipation was probably perceived as meaningful to dogs, it was not an easily observable symptom by pet owners unless severe (eg, nonproductive straining). As such, owners of patients enrolled in the present study may not have recognized that subtle changes in bowel movement patterns were symptoms of constipation. Aggression is a behavioral trait that relates strongly to temperament, which could explain its poor performance during testing. Some dogs will never behave aggressively regardless of health status, and other dogs will behave aggressively despite apparently perfect health. Symptom assessment scales for human patients include dozens of items, but a large proportion includes psychological or emotional symptoms that are not applicable to dogs or proxy assessments. For example, it is difficult to know whether dogs experience symptoms such as worry and loss of self-image, and owners cannot directly observe other symptoms such as dry mouth or numbness and tingling. Although limited to observable physical symptoms, the CSAS could be expanded to include additional relevant items identified with the open-ended questions. No individual symptom was voluntarily disclosed by a large number of respondents, but the results of the open-ended questions for pet owners in this study do not accurately gauge symptom prevalence in the study population. Human cancer patients reported experiencing a significantly greater number of symptoms when data were obtained with questionnaires versus open-ended questions.38 The present study identified several symptoms that we suggest may be valuable additions to future versions of the CSAS, including weakness, lameness, excessive drinking, and abnormal urination. Replicating this study in different populations of dogs could identify other relevant symptoms.

The 12-item CSAS had good face and content validity, which means that the questionnaire appeared to be measuring symptoms and that the items and scales were relevant for measuring general symptomology in the population of dogs studied. Construct validation, the process of determining whether the instrument truly reflected the subjective concepts we were attempting to measure, was achieved by comparing the instrument's actual performance in different scenarios with how a theoretically valid instrument should perform. Factor analysis examines construct validity by considering whether groups of items in the questionnaire clustered together according to known or expected relationships. An exploratory factor analysis is performed in instances such as this study, where there is no prior evidence demonstrating how the items relate to one another. The CSAS factor analysis demonstrated construct validity because the symptoms in each factor group reflected accepted syndromes observed in actual patients. Construct validity was also demonstrated by showing that the CSAS and its factors related to the CBPI in the ways we would have expected, given the partial overlap in measurement objectives between the instruments. As hypothesized, all symptoms had a negative impact on QOL, but only malaise factor scores correlated strongly with CBPI subscale scores. We also identified an important dose-response relationship between symptoms and QOL: as symptomology increased, QOL decreased, and vice versa. Among the CSAS subscales, the malaise factor included symptoms that were rated most prevalent and most distressing and that correlated most strongly with QOL. The significant dose-response pattern in the anxiety and digestive upset factors suggested that less prevalent symptoms were also meaningful to pets and pet owners and contributed to QOL assessments. Comparing the CSAS and CPBI also provided a measure of criterion validation, which is demonstrated when a new scale correlates strongly with an existing measure or gold standard. There is no existing gold standard for measuring the symptoms included in the CSAS. However, the strong correlation between malaise factor symptoms and CBPI indices supported criterion validation for the aspects of the CSAS for which validated measures do exist.

The present study demonstrated high internal consistency of the CSAS, a measure of reliability. An internally consistent instrument measures different parts of the same underlying trait (in this case, symptomology), not different traits. The aggression item was removed because we believed that it was tapping into the trait of temperament, not symptomology. The malaise factor subscale had excellent internal consistency and could be useful in the population studied to screen for overall QOL. The anxiety and digestive factor subscales had only moderate internal consistency, but performed well in tests of construct and criterion validity. Given the small number of symptoms for these factors, and the generally low prevalence of several symptoms, these findings were expected and helped to define appropriate use of the scores without compromising the overall validity of the instrument. On the basis of results of the present study, insufficient evidence exists to consider the anxiety and digestive factors as valid subscales independent of the whole instrument. In a population of dogs in which symptom patterns vary from those in the present study, the subscales might perform differently.

To function as an outcome measure in clinical trials, the CSAS ideally would demonstrate stability (test-retest reliability) on repeated administration to dogs with stable symptoms, responsiveness to relevant health status changes, and generalizability to other populations of dogs. The use of CSAS score as a clinical trial outcome measure will also depend on the research question. Total score could be used to compare treatments on the basis of relative symptom burden by defining a threshold score beyond which dogs are withdrawn from the study; another option would be to report subjects' cumulative or mean scores over a set time period. Change in CSAS score between 2 time points could be used to measure treatment response. In each of these scenarios, meaningful interpretation of results depends on defining a clinically relevant CSAS value on which comparisons can be based. We chose to scale the instrument according to the methodology of the MSAS. A high total score could reflect a few distressing symptoms or several mild ones. Future study is needed to determine whether the patient burden is equivalent under these scenarios and to define clinically meaningful scores and score differences required to determine the responsiveness of the CSAS to clinically relevant changes in symptom burden.

Although designed as a research outcome measure, the CSAS also has clinical applicability. Most veterinary care seeks to minimize symptoms, support normal function, and maintain good QOL, outcomes that the pet owner is best situated to judge. Veterinarians rely on pet owners to report bothersome symptoms observed in their animals, either spontaneously or in response to directed questioning. The use of an owner-reported questionnaire such as the CSAS has several advantages over depending on client interviews alone. The CSAS was efficient, compared with the amount of time a veterinarian or technician would need to obtain and record the same information. The questionnaire can be directly incorporated into the patient medical record, limiting paperwork and interpretive biases. It ensures that data on important symptoms are consistently requested, especially in circumstances when detailed client interviews are impractical because of time constraints. The CSAS also provides an opportunity for pet owners to actively engage in the health-care process, which may have health benefits for the pet in itself.39,40 Although scoring the CSAS by hand is likely to be impractical for busy clinicians, this could be mitigated in the future by administering the questionnaire through a software-based application. Even without being scored, the CSAS provides detailed at-a-glance symptom information that can be immediately incorporated into veterinarians' clinical assessments.

The population evaluated in the present study was not representative of all dogs with cancer nor all dogs enrolled in clinical trials of cancer treatments. A limited number of cancer types were represented, disease stage was not evaluated at the time of questionnaire administration, and most dogs were undergoing local or palliative treatments. The 12 CSAS symptoms are potentially relevant to all dogs with cancer, yet probably insufficient to describe the full symptom burden experienced by many. Reproducing these findings in other populations of dogs would allow us to demonstrate wider generalizability and to identify additional symptoms warranting inclusion in the questionnaire. Incorporation of additional general symptoms and perhaps disease- or treatment-specific subscales could improve the CSAS's ability to collect comprehensive symptom data in a variety of clinical scenarios. The CSAS provides a wealth of descriptive data but a moderate respondent burden because of the multidimensional format.

In the present study, CSAS completion rate was excellent; pet owners were familiar with the questionnaire from previous visits, and study nurses ensured questionnaires were completed correctly. Because the instrument format could be confusing for first-time users, we recommend that respondents be given verbal instructions prior to using the instrument and that accurate completion be confirmed before responses are transferred to a database. Demographic information about pet owners was not collected in this study, but might reasonably influence proxy assessments and responses on the CSAS. Although we suggest that the pet owners of this study should be a representative sample of those who enroll dogs in clinical trials of cancer treatment at our institution, they are likely not representative of all dog owners. Owners who seek treatment for their dogs' cancer and participate in clinical trials could differ in terms of education, financial means, philosophy toward pet ownership, or other important characteristics, compared with those who do not elect treatment or who are not offered a chance to or decline to enroll their dogs in research studies. Additional research could elucidate how such differences might affect the performance of the CSAS; further examine reliability, validity, and responsiveness with repeated administrations; and improve interpretation of total and subscale scores. If additional psychometric testing is successful, it would provide even more support for wide applicability of the CSAS in veterinary clinical trials and practice.

Acknowledgments

Supported in part by the National Institutes of Health (grant No. R21DE018178), Pfizer Animal Health, Integrated Chinese Medicine Holdings Limited, and the University of Pennsylvania Companion Animal Research Fund.

The CSAS can be downloaded at www.CanineSAS.com.

ABBREVIATIONS

CBPI

Canine Brief Pain Inventory

CSAS

Canine Symptom Assessment Scale

MSAS

Memorial Symptom Assessment Scale

QOL

Quality of life

Footnotes

a.

Stata Statistical Software, release 12, StataCorp LP, College Station, Tex.

References

  • 1. Merriam-Webster dictionary. Symptom. Available at: www.merriam-webster.com/dictionary/symptom#medicalDictionary. Accessed Nov 20, 2015.

  • 2. Nekolaichuk CL, Magiure TO, Suarez-Almazor M, et al. Assessing the reliability of patient, nurse, and family caregiver symptom ratings in hospitalized advanced cancer patients. J Clin Oncol 1999;17:36213630.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3. Lobchuk MM. Symptom experiences: perceptual accuracy between advanced-stage cancer patients and family caregivers in the home care setting. J Clin Oncol 2002;20:34953507.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4. Oi-Ling K, Man-Wah D & Kam-Hung D. Symptom distress as rated by advanced cancer patients, caregivers and physicians in the last week of life. Palliat Med 2005;19:228233.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5. Baggott C, Cooper BA, Marina N, et al. Symptom assessment in pediatric oncology: how should concordance between children's and parents' reports be evaluated? Cancer Nurs 2014;37:252262.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6. Zhukovsky DS, Rozmus CL, Robert RS, et al. Symptom profiles in children with advanced cancer: patient, family caregiver, and oncologist ratings. Cancer 2015;121:40804087.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 7. Stedman's medical dictionary. 27th ed. Baltimore: Lippincott Williams & Wilkins, 2000.

  • 8. Smith RD. Veterinary clinical epidemiology. 3rd ed. Boca Raton, Fla: Taylor and Francis Group, 2006.

  • 9. Brown DC, Boston RC, Coyne JC, et al. Development and psychometric testing of an instrument designed to measure chronic pain in dogs with osteoarthritis. Am J Vet Res 2007;68:631637.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 10. Wiseman-Orr ML, Nolan AM, Reid J, et al. Development of a questionnaire to measure the effects of chronic pain on health-related quality of life in dogs. Am J Vet Res 2004;65:10771084.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11. Wiseman-Orr ML, Scott EM, Reid J, et al. Validation of a structured questionnaire as an instrument to measure chronic pain in dogs on the basis of effects on health-related quality of life. Am J Vet Res 2006;67:18261836.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 12. Hielm-Björkman AK, Rita H & Tulamo RM. Psychometric testing of the Helsinki chronic pain index by completion of a questionnaire in Finnish by owners of dogs with chronic signs of pain caused by osteoarthritis. Am J Vet Res 2009;70:727734.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13. Reid J, Wiseman-Orr ML, Scott EM, et al. Development, validation and reliability of a web-based questionnaire to measure health-related quality of life in dogs. J Small Anim Pract 2013;54:227233.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 14. Malone EK, Rassnick KM, Bailey DB, et al. Comparison of two questionnaires to assess gastrointestinal toxicity in dogs and cats treated with chemotherapy. Vet Comp Oncol 2011;9:225231.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15. Hamilton MJ, Sarcornrattana O, Illiopoulou M, et al. Questionnaire-based assessment of owner concerns and doctor responsiveness: 107 canine chemotherapy patients. J Small Anim Pract 2012;53:627633.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16. Basch E, Iasonos A, McDonough T, et al. Patient versus clinician symptom reporting using the National Cancer Institute Common Terminology Criteria for Adverse Events: results of a questionnaire-based study. Lancet Oncol 2006;7:903909.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17. Basch E, Reeve BB, Mitchell SA, et al. Development of the National Cancer Institute's patient-reported outcomes version of the common terminology criteria for adverse events (PRO-CTCAE). J Natl Cancer Inst 2014;106:dju244.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 18. Atkinson TM, Li Y, Coffey CW, et al. Reliability of adverse symptom event reporting by clinicians. Qual Life Res 2012;21:11591164.

  • 19. Basch E, Jia X, Heller G, et al. Adverse symptom event reporting by patients vs clinicians: relationships with clinical outcomes. J Natl Cancer Inst 2009;101:16241632.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20. Di Maio M, Gallo C, Leighl NB, et al. Symptomatic toxicities experienced during anticancer treatment: agreement between patient and physician reporting in three randomized trials. J Clin Oncol 2015;33:910915.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21. Secord AA, Coleman RL, Havrilesky LJ, et al. Patient-reported outcomes as end points and outcome indicators in solid tumours. Nat Rev Clin Oncol 2015;12:358370.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 22. Giuffrida MA. A systematic review of adverse event reporting in companion animal clinical trials evaluating treatment. J Am Vet Med Assoc 2016;249:10791087.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23. Ioannidis JPA, Evans SJW, Gøtzsche PC, et al. Better reporting of harms in randomized trials: an extension of the CONSORT statement. Ann Intern Med 2004;141:781788.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 24. Banerjee AK, Okus S, Edwards IR, et al. Patient-reported outcome measures in safety event reporting: PROSPER Consortium Guidance. Drug Saf 2013;36:11291149.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25. Reeve BB, Mitchell SA, Dueck AC, et al. Recommended patient-reported core set of symptoms to measure in adult cancer treatment trials. J Natl Cancer Inst 2014;106:dju129.

    • Search Google Scholar
    • Export Citation
  • 26. Portenoy RK, Thaler HT, Kornblith AB, et al. The Memorial Symptom Assessment Scale: an instrument for the evaluation of symptom prevalence, characteristics and distress. Eur J Cancer 1994;30A:13261336.

    • Search Google Scholar
    • Export Citation
  • 27. Streiner DL & Norman GR. Health measurement scales: a practical guide to their development and use. Oxford, England: Oxford University Press, 2008.

    • Search Google Scholar
    • Export Citation
  • 28. Change VT, Hwang SS, Feuerman M, et al. The Memorial Symptom Assessment Scale short form (MSAS-SF). Cancer 2000;89:11621171.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 29. Kutner JS, Bryant LL, Beaty BL, et al. Symptom distress and quality-of-life assessment at the end of life: the role of proxy response. J Pain Symptom Manage 2006;32:300310.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 30. Brown DC, Boston R, Coyne JC, et al. A novel approach to the use of animals in studies of pain: validation of the Canine Brief Pain Inventory in canine bone cancer. Pain Med 2009;10:133142.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 31. Gorsuch RL. Factor analysis. 2nd ed. Hillsdale, NJ: Lawrence Erlbaum Associates, 1983.

  • 32. Costello AB & Osborne JW. Best practices in exploratory factor analysis: four recommendations for getting the most from your analysis. Pract Assess Res Eval 2005;10:19.

    • Search Google Scholar
    • Export Citation
  • 33. Flesch R. A new readability yardstick. J Appl Psychol 1948;32:221233.

  • 34. Fan TM. Pain management in veterinary patients with cancer. Vet Clin North Am Small Anim Pract 2014;44:9891001.

  • 35. Berger AM, Mitchell SA, Jacobsen PB, et al. Screening, evaluation, and management of cancer-related fatigue: ready for implementation to practice? CA Cancer J Clin 2015;65:190211.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 36. Münster M, Hörauf A, Lübke-Becker A, et al. Idiopathic esophagopathies resembling gastroesophageal reflux disease in dogs. Tierarztl Prax Ausg K Klientiere Heimtiere 2013;41:173179.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 37. Chung KF & Pavord ID. Prevalence, pathogenesis, and causes of chronic cough. Lancet 2008;371:13641374.

  • 38. Bent S, Padula A & Avins AL. Brief communication: better ways to question patients about adverse medical events: a randomized controlled trial. Ann Intern Med 2006;144:257261.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 39. Yeates JW, Mullan S, Stone M, et al. Promoting discussions and decisions about dogs' quality of life. J Small Anim Pract 2011;52:459463.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 40. Velikova G. Measuring quality of life in routine oncology practice improves communications and patient well-being: a randomized controlled trial. J Clin Oncol 2004;22:714724.

    • Crossref
    • Search Google Scholar
    • Export Citation

Supplementary Materials

All Time Past Year Past 30 Days
Abstract Views 286 0 0
Full Text Views 880 671 240
PDF Downloads 379 143 21
Advertisement