Compassion through the lens of patients with life-limiting health conditions has been described as being both dispositional and relational. It is nonetheless insufficient for healthcare providers (HCPs) to simply possess virtues, as these good and noble qualities need to be coupled with understanding and action, aimed at alleviating the suffering of another person to be considered compassion [1]. These virtuous motivators and orientation toward action distinguishes compassion from sympathy, in that compassion is not a visceral, pity-based response towards a person’s circumstance or misfortune [2,3,4]. Furthermore, sympathy is a concept that is often coupled with a negative connotation, particularly from patients who find it unhelpful and demoralizing [3, 4]. While bearing a number of similarities with the construct of empathy, compassion requires action and is motivated by virtues such as love, kindness, and genuineness, thereby rendering it more dynamic and distinct [3,4,5,6].
Although various tools exist for measuring compassion in healthcare, they often conflate these surrogate constructs with compassion, and lack the rigorous establishment of psychometric properties [7]. A recent comprehensive and critical review identifying and evaluating existing compassionate care measures found a paucity of tools that theoretically conceptualize compassion as a multidimensional construct and perhaps most remarkably failed to incorporate the perspective of the recipients of compassion—patients [7]. This review identified one tool that measured compassion as an emotion in HCPs, failing to account for the crucial, observable dimensions such as behaviour or action, which are among its key defining features [8, 9]. Three tools measured the HCPs ability to provide compassionate care [10,11,12], and only one measured compassion in relation to the patient’s care experience [13]. Another tool, the Geriatric Attitudes Measure [14], despite being validated, measures limited facets of compassion. Due to the nascent nature of compassion measures and the underutilization of patients’ perspectives, there is a need to develop a patient reported compassion measure based on a patient-informed, theoretical model of compassion.
In recognizing these limitations, a Patient Compassion Model (PCM) (Fig. 1) was developed by our research group from the perspective of advanced cancer inpatients (n = 53) using a rigorous grounded theory approach [1]. A complex relationship was outlined between six mutually informing dimensions of the construct of compassion: virtues, as the good and noble qualities of HCPs that served as motivators of compassion; relational space, highlighting the highly relational nature of compassion where patients initially sensed their HCP capacity for compassion; virtuous response, the initial response of a HCP towards the person in suffering; seeking to understand the patient as a person and their needs; relational communicating, consisting of verbal and non-verbal displays of compassion within the clinical encounter; and attending to needs, the quintessential outcome of compassion—action aimed at addressing patients multifactorial suffering [1]. This multi-dimensional compassion construct represents, to our knowledge, the first rigorously established, patient informed, clinically relevant, and empirically derived, theoretical model [1]. While representing a comprehensive model, patients nonetheless recognized fluidity, overlap, and variance across and within the aforementioned dimensions, affirming the personalized nature of expressions (HCPs) and experiences (patients) of compassion, the dynamic nature of compassion, and the constructs adaptability to context and circumstance [1].
In qualitative research, an imperative means of ensuring that the theoretical foundation of a new construct stands on a solid foundation and is ‘trustworthy’ is to determine the credibility and/or transferability of the construct within the same or other patient populations [15,16,17]. Determining the credibility and transferability of an empirically derived theoretical model addresses two interrelated methodological issues limiting the initial qualitative, and subsequent measure development studies—the lack of follow up studies to verify the results of qualitative studies and the lack of initial content validity with end users within measure development studies, respectively. While qualitative researchers recognize the non-generalizability of their results at the outset of their research projects, and are quick to dismiss criticism from well-intended (yet misguided) quantitative reviewers who suggest otherwise, there is a tendency to treat their findings as fait accompli after their study is completed. To determine the extent to which the theory is in fact grounded in other contexts, Grounded Theorists encourage further development of the theory in subsequent studies and different settings, thereby cautioning researchers from taking a fait accompli stance to their theory [18]. In doing so, researchers can clarify additional understandings by modifying their initial findings and theoretical model accordingly [15, 18].
An adjacent issue hindering subsequent measure development studies after the theory has been established, is the need to conduct initial content validation to fortify the theoretical foundation that the measure purports to assess. While further steps of validation such as exploratory and confirmatory factor analysis can demonstrate that the questions or ‘items’ of a measure are aligned with researchers’ understandings of a construct, this understanding assumes that the construct is fully comprehensive and accurate. Confirming the content validity of the theory or model can enhance the content validity of a measure directly from those whom the measure will be administered to, as there is “great value of adding [an additional] layer of feedback to the elaboration of a conceptual model before its implementation” (p. 561) [19]. As it’s been argued by others, prior to measure development and its subsequent validation, conducting effective qualitative research is considered a crucial component to ensuring that the instrument possesses adequate content validity and to ensuring fidelity between the proposed measure and its theoretical foundation [20, 21].
The objective of the present study was to assess the credibility and transferability of the PCM among a convenience sample of 20 non-cancer patients living with life-limiting illnesses in multiple healthcare settings, a method of model validation utilized in a previous study [19]. In doing so, we also aimed to establish additional content/face validity for an ongoing study to develop a patient reported compassion measure for palliative patients living with diverse, incurable and life-limiting illnesses.