Development and validity of owner-based screening tools for canine cognitive decline

Canine cognitive dysfunction (CCD) is a progressive, age-related neurodegenerative disorder of dogs diagnosed by exclusion of alternative causes of the non-specific behavioural signs. This study validated owner-friendly versions of instruments for the assessment of CCD based on: The Canine Cognitive Dysfunction Rating (CCDR), Canine Dementia Scale (CADES), and Age-Related Cognitive and Affective Disorders (ARCAD). While scale development focuses on maximising specificity, clinical tools need to balance this with sensitivity, considering the risks of false positives e.g. treating some subjects unnecessarily versus those related to false negatives e.g. not starting treatment. The scales were adapted for greater owner-comprehensibility to create CCDRa, CADESa, ARCADa. 959 completions of CCDRa, CADESa and ARCADa were obtained via an online questionnaire. Two groups of respondents were created for analysis: dogs with an owner-declared veterinary diagnosis for CCD (CCD dogs) and those without an owner-declared veterinary diagnosis for CCD and being “very unlikely” to have CCD (Non-CCD dogs). Individual total scale scores were calculated to examine how these changed with age group. Receiver Operating Curves and Youden’s Indices were used to identify the peak trade-off between sensitivity and specificity, alongside Positive Predictive Values (PPV) and Negative Predictive Values (NPV). These were used to create recommendations for screening value (SV) thresholds for each scale, for different categories of dog, namely: older dogs (seniors of unknown specific age) CCDRa ≥ 40, CADESa ≥ 25, and ARCADa ≥ 17; dogs aged 7 – 12 years old, CCDRa ≥ 40, CADESa ≥ 25, ARCADa ≥ 19; dogs aged 13 years old and above, CCDRa≥ 42, CADESa ≥ 25, ARCADa ≥ 21. CCDRa and CADESa showed the greatest level of convergent validity, and ARCADa least convergence when using published thresholds (CCDR ≥ 50, CADES ≥45, ARCAD ≥ 31) or SV thresholds. Using published thresholds, only 64/103 (62%) of CCD cases were identified by at least one of the scales, but this rose to 102/103 (99%) when the revised SV thresholds were applied. Given that CCD is an irreversible disease, early intervention is crucial for patient health and welfare. SV thresholds (in the absence of other potential medical explanations for the signs) represent a value at which we recommend veterinarians consider discussing with owners the value of intervention for CCD, including a quantified risk of false positives. Our adapted version of CADES (CADESa) performed best and allows monitoring of disease progression. Accordingly, CADESa is our preferred tool for the assessment and monitoring of CCD. • Owner-friendly adaptations of CCDR, CADES and ARCAD created • 959 completions of all 3 instruments assessed for convergent validity • Convergent validity generally moderate for all three instruments • Optimal threshold values for screening lower than optimal for development • CADES performed best of the three scales with a screening value of 25 recommended