Development and preliminary psychometric evaluation of an owner-completed measure of feline quality of life

Highlights

• Health-related quality of life (HRQoL) is increasingly studied in companion animals but less so in cats.

• Forty-five pet owners completed an online survey confirming feline behavioural concepts from the literature.

• A 23-item feline QoL measure was drafted, capturing the identified concepts, and tested in qualitative pet owner interviews with pet owners.

• The measure was reduced to 22 items following qualitative pet owner interviews with pet owners, since items overlapped or were not relevant to pet-owners.

• The final measure demonstrated strong content and psychometric validity and reliability in a sample of healthy cats in the USA.

Abstract

Due to improved healthcare and pet longevity, measurement of health-related quality of life (HRQoL) is increasingly important in companion animal medicine. The aim of this study was to develop and evaluate the content and psychometric properties of an owner-completed assessment of health and quality of life (QoL) in cats for use in general veterinary clinical practice. A 23-item feline QoL measure, drafted based on findings from an online survey completed by 45 pet owners, was revised following qualitative interviews with 10 pet owners of healthy cats to assess content validity. The resulting 22-item measure was completed twice by 199 owners of healthy cats to assess the reliability and validity of the measure via psychometric evaluation, including assessment of missing data, item response distributions, item correlations, factor analysis, internal consistency, test–retest reliability, multi-trait analysis, known groups analyses and estimation of minimally important differences. There were no missing data. Responses for all items were heavily skewed due to the sample being healthy. Analysis of items and factor analysis supported deletion of six items and calculation of two domain scores and a total score. Internal consistency and test–retest reliability were strong for all domains (0.70–0.80), indicating good reliability. All but three items demonstrated strong item convergent validity (item-scale correlation > 0.40) and correlated highest with their respective domain (item discriminant validity). Significant between-group differences in scores differing according to a global impression of feline health item provided evidence of discriminative validity. Findings provide evidence that the final 16-item feline QoL measure has strong cross-sectional psychometric properties.

Introduction

Due to improved healthcare and longevity, health-related quality of life (HRQoL) has been increasingly studied in companion animals (Niessen et al., 2010, Lynch et al., 2011, Lavan, 2013). HRQoL is considered a multi-dimensional concept (Lavan, 2013, Mullan, 2015), representing both physical (including comfort/discomfort) and non-physical factors (feelings felt by animals). Assessing these concepts in animals in a valid and reliable way is challenging (McMillan, 2000). However, measuring HRQoL can help to influence treatment decisions, and aid veterinarians and pet owners in tracking changes in HRQoL over time. The Food and Drug Administration (FDA) patient-reported outcome (PRO) guidance¹ outlines robust methods for the development and validation of human PROs and HRQoL measures (Campbell and Fiske, 1959, Nunnally, 1994, Marquis et al., 2006, Streiner et al., 2008, Lasch et al., 2010). Such methods have been applied to companion animal HRQoL measures, using in-depth qualitative research in the target population (i.e. pet owners), to identify relevant concepts (McMillan, 2000, Niessen et al., 2010, Lynch et al., 2011), and to inform question (item) development and comprehensiveness of concepts assessed, as well as reliability and validity using statistical methods. Validity establishes whether measures assess concepts they were designed to measure; reliability indicates the accuracy of assessing responses over time (Yeates and Main, 2009).¹

Arguably, challenges in developing an animal HRQoL assessment are similar to those in developing observer-reported outcome (ObsRO) measures in non-verbal human paediatric populations (Patrick et al., 2011, Arbuckle and Abetz-Webb, 2013); there is a need to measure observable behaviours that can be accurately and consistently rated (Yeates and Main, 2009, Arbuckle and Abetz-Webb, 2013, Matza et al., 2013).

Numerous disease-specific HRQoL measures have been developed for dogs and cats (Freeman et al., 2005, Yazbek and Fantoni, 2005, Brown et al., 2007, Budke et al., 2008, Hielm-Bjorkman et al., 2009, Favrot et al., 2010, Niessen et al., 2010, Lynch et al., 2011, Noli et al., 2011, Iliopoulou et al., 2013, Lavan, 2013, Belshaw et al., 2015, Bijsmans et al., 2015); however, there is a paucity of HRQoL measures for the broad feline population (Bijsmans et al., 2015, Freeman et al., 2016). This paper describes the development and psychometric evaluation of an owner-completed measure to assess feline QoL in veterinary clinical practice.