Person-centredness in the care of older adults: a systematic review of questionnaire-based scales and their measurement properties

‘Person-centredness’ is internationally regarded by many as a foundation for modern health and social care services [1‐3], with the World Health Organization recently calling for a ‘fundamental paradigm shift’ in strategy and delivery in accordance with its principles [4]. It has widespread appeal as a philosophy of care that emphasizes the need for services to be responsive to individual needs, and promotes the rights of recipients in achieving a greater influence over decisions that affect them [5‐7]. Tracing its origins back to the 1950s, person-centredness can draw upon a spectrum of well-established conceptual frameworks, including personhood; normalisation; the social model of disability; citizenship; and new public management. In England, person-centredness is championed throughout the care system, from the National Health Service Constitution, through legislative programmes and individual policy initiatives, national clinical standards, regulation of care quality, indicators of performance and, ultimately, front-line practice [8‐10]. That person-centredness has come to hold such a prominent position in the care system is no accident. In addition to ethical arguments based on human rights and public service accountability, evidence suggests that it is strongly associated with service satisfaction; is linked with better engagement with, and adherence to, treatment plans; and is broadly associated with improved health and quality-of-life outcomes [11‐13].

Despite attaining such prominent status, `person-centredness' is notoriously difficult to define and conceptualise. Reviews commonly regard person-centredness as a composite [1, 7, 14] in combining care attributes that themselves are independently recognized components of quality. Different traditions of ‘centredness’ can be identified within the literature, using varied prefixes (eg ‘patient’, ‘client’, or ‘consumer’) with each giving different emphasis to its necessary and sufficient attributes. Nevertheless three themes are common to each, together forming an operational definition of person-centredness used in this review. First, it gives primacy to understanding the person and their unique interpretation and experience of illness or disability, in particular by taking a holistic view through recognition of psycho-social factors beyond presenting symptoms [12]. Second, service user empowerment in decision-making has been described as the ‘pinnacle’ of person-centredness [15], with greater delegation of control over choices to the service user, guided by a practitioner through appropriate information sharing [16]. Third, the importance of relationships in care and treatment is prioritized, since positive and respectful interpersonal exchanges and the development of trust built on continuity and coordination in care are viewed as therapeutic vehicles to successful support [2].

The importance of person-centredness for older people with long-term conditions may be at least as great as for other patient groups. The prevalence of multi-morbidity and long-term health problems increases with age, requiring many older adults to draw upon a wider range of support often from multiple care professionals and providers, and so increasing the risk of fragmented care relationships. Further, older people may prioritise the affective characteristics of the care exchange as much as the achievement of specific outcomes [17, 18]. This may reflect the value placed by older people on maintaining personal identity and usual routines in the context of cognitive or physical decline which can, in part, be achieved through positive care interactions and attention to the whole person [19]. Older people may also prioritise different facets of person-centredness, or require them to be achieved in a different manner. For example, preferences for autonomy and engagement in decision-making vary between cross-sections of younger and older adults [20].

Given conceptual ambiguity, and potential subtle differences in priorities and articulation amongst older people, it is unsurprising that no clear set of measures is available to assist service planners, regulators or researchers in assessing person-centred qualities. Yet recent international appeals to improve and standardise approaches to measurement have brought renewed attention to the need for appropriate instruments [4]. A Cochrane Review relating to clinical consultations [21] found that none of the included studies used direct measurement of person-centredness, precluding an understanding of cause-effect pathways, and adding to claims that current scales are either not fit for purpose or inconsistently used [7, 11, 12, 22]. Instead, broad satisfaction surveys are commonly adopted, which routinely identify positive experiences amongst older people but which are doubted both conceptually and empirically [23, 24].

No systematic review of measures of person-centredness relevant to the long-term care of older adults has yet been conducted. Most importantly, narrative descriptions of available measures [16, 25] have made reference to measurement properties without critically appraising the quality of research underpinning them. As in any research field, the acceptance of empirically-derived estimates without critical appraisal undermines the evidence-base [26]. The purpose of this review is to address this gap. Specifically, the review aimed to identify, describe and critically appraise measures of person-centredness relevant to the long-term care of older people.

Of the 11 instruments included in the review, four stood out as having been the subject of tests of measurement properties in three or more studies, and together accounted for over half of the 22 references: the Individualised Care Instrument (ICI) [29‐31]; Person-Centred Care Assessment Tool (P-CAT) [32‐35]; the Person-centred Climate Questionnaire (PCQ) (comprising both staff [36, 37] and patient) [38] versions); and the Client-Centred Care Questionnaire (CCCQ) [39‐42]. The remaining seven instruments had not been as extensively tested in an older adult population, although the Individualized Care Scale – Nurse (ICS-N) [43] and Measures of Processes of Care – Adult (MPOC-A) [44] were more widely used outside specialist old-age services. An ‘Untitled’ measure was also included in this review [45], but differed from others since establishing measurement properties was not the main focus of the associated reference.

To assist in synthesis, measures were organised according to whether they were specifically designed for application in older adult services (n = 6, hereafter ‘specific’), or else were originally designed for other/generic services, but had since been applied to older adult services, or in a predominantly older sample (n = 5, hereafter ‘generic’). Table 3 illustrates that all except one of the specific instruments (Patient-Centered Family Focused Care (PCFC) [46]) were initially designed for completion by practitioners, whilst the pattern was largely reversed for generic measures. This tallies with the service settings of the former; predominantly relating to long-stay care designed for people with dementia, thus likely to preclude self-completed questionnaires. Six instruments were designed primarily for use in residential or long-stay nursing care settings (ICI, P-CAT, Person-Directed Care (PDC) [47], 'Untitled', PCQ, ICS-N); four were designed for ambulatory hospital- or clinic-based services (PCFC, Person-Centred Health Care for Older Adults (PCHC) [48], MPOC-A, Client-Centred Rehabilitation Questionnaire (CCRQ) [49]) and with just one measure (CCCQ) designed explicitly for home-based services. The measures drew on a range of different traditions of ‘centredness’ to inform their development, such as a Kitwoodian analysis [19] of respect for personhood in dementia (P-CAT, 'Untitled'), and client-centredness in rehabilitation (CCRQ). The origins of two other specific measures (PCHC, PDC) lay with policy-makers rather than clinicians and academia: for example, the PDC measure supported a programme to improve the standing and attractiveness of work in the long-term care of older adults in Oregan [47]. Just three measures (ICI, CCCQ, CCRQ) sourced items empirically from primary exploratory fieldwork with service users, using a range of qualitative techniques (observation of care exchanges; qualitative semi-structured interviews; and focus groups). The CCRQ was unique in undertaking formal cognitive interviews as a mechanism for testing with service users how items were interpreted and how responses formulated [49].

Eight instruments (see Table 4) were multidimensional and formed distinct subscales, typically identified through factor analyses, enabling an assessment of how well the three broad themes of person-centredness (outlined above) were represented. Items relating to 'understanding the person' formed a distinct subscale of the ICI and PDC (both labelled 'knowing the person'); PCHC ('getting to know the individual'); and the ICS-N (comprising separate subscales assessing how well practitioners attended to the personal experiences and interpretation of their current ‘clinical situation’; and a second focused on understanding the patient’s wider ‘life situation’). Items relating to ‘empowerment in decision-making’ were distinct in the ICI and PDC (both labelled 'autonomy'); PCHC ('involvement in care planning'); the ICS-N ('decisional control over care'); MPOC-A ('enabling and partnership') and CCRQ ('decision making'). Furthermore, the MPOC-A and CCRQ additionally included subscales directed at the quality of information-sharing in supporting decisions; and the P-CAT included a subscale assessing how care decisions are tailored to the individual. Finally, features of 'relationships in care' were evident through items in the ICI ('communication with residents'); PCHC ('supporting relations'); MPOC-A ('respectful and supportive care') and CCRQ ('emotional support'). The extent of continuity and coordination in care was also a feature of the PCHC, MPOC-A and CCRQ. Whilst the labels of the PCQ subscales do not obviously correspond with the themes of person-centredness used in this paper, a closer inspection of individual items finds clear resonance. For example, within ‘a climate of safety’ are several items relating to caregiver interpersonal and relationship skills.

It has been argued that efforts to objectively measure person-centred care have not matched its rapid promotion amongst health service priorities [8]. Researchers seeking to evaluate interventions against these standards, and managers aiming to monitor and improve quality, have a limited evidence-base to support their choice of measurement instruments. Dow et al. [48] developed their own measure after a literature review found “no previously published measures of person-centered care in health settings” (p1066) was suited to their research in old age psychiatric services. No systematic review has hitherto been conducted and no formal quality appraisal has been undertaken. The present review aimed to fill this gap.

Eleven instruments were identified, spanning both general and gerontological nursing, rehabilitation and occupational therapy, and palliative care. However, the breadth and methodological quality of research underpinning these measures was generally poor, and none can be recommended without significant reservations. Two measures, the P-CAT and CCCQ, have been subject of the most attempts to test measurement properties. The former was designed for completion by staff in long-term care facilities to self-assess the person-centred quality of their service. However, the dimensionality of the instrument is subject to some uncertainty, with an unstable two or three factor solution in different language versions. Further development work and confirmatory factor analyses would bolster the measure. The CCCQ, by contrast, is a unidimensional measure developed primarily for home care services. The items were formulated directly from quotations from service user interviews, although these were mostly younger adults, and none aged over sixty [50]. The measure has been subject to the most rigorous testing of all the instruments, although there are doubts over its construct validity since it failed to conform to hypotheses tested in one study. There is some evidence that the items may have proved challenging for an elderly population to answer, and the authors have recommended the development of instruments using items tailored to the experiences and abilities of the particular client group being researched [42]. Cognitive interviewing is one method for rigorously testing the applicability and comprehension of instruments prior to wider piloting, and was adopted by only one reference in this review [49].

Quality appraisal was attempted for seven measurement properties across the 12 instruments (treating the two PCQ versions separately), permitting 84 possible synthesis results. Yet many measurement properties had not been successfully tested in any study meeting the eligibility criteria. Of the 51 populated cells, the review found only eight were supported by evidence rated as ‘strong’, with many ‘undetermined’ results due to poor methodological quality. The evidence-base would be better served by more studies of higher quality, even if that is at the expense of fewer measurement properties being investigated, at least in early development. For example, efforts to translate existing instruments into a multitude of new languages appear wasteful if the instrument has yet to demonstrate solid validity and reliability in its original language. Furthermore, using exploratory analysis with a fledgling instrument, and applying a post-hoc rationalisation of how results support validity, should be avoided. Not all such studies explicitly sought to formally establish a measurement property and might have been excluded from the review. Regardless, by not specifying clear hypotheses a priori, they did not achieve a good rating. Use of modern toolkits, such as COSMIN, may assist researchers in reaching the standards expected in modern measurement studies.

In common with Edvardsson et al. [32], this review also finds a notable lack of service user or carer perspectives in selecting items for inclusion in questionnaires. In addition to being poor practice, it is ironically incompatible with person-centredness. Researchers have also shied away from supplying evidence of measurement error, and its related concept, ‘minimal important change’ [26]. It is essential to understand what the smallest change is in any given measure that has meaning and value to service users. This can then be compared to estimated error within the measure, and it can then be determined if meaningful change is within or beyond what can be reliably detected. Thus, greater consideration to service user perspectives in the development and testing of future instruments is required.

The review strived to include measures of person-centredness in care, but given definitional ambiguities in the construct, the instruments synthesised are diverse in their content and intended application. Three domains of person-centredness - used as an operational definition for this review - were well represented in the measures; however it is evident that some implicitly used wider interpretations. Examples include an extension of person-centredness into patient perceptions of personal safety and physical comfort. There is nothing inherently problematic in this as long as researchers’ own interpretations of person-centredness are clearly articulated when developing their instruments, and, of course, that those using the questionnaires are alert to this. That said, instrument developers should at least ensure that the items reflect their intent: to achieve a high rating for content validity, the COSMIN framework demands that studies provide evidence that they assessed how comprehensively the items spanned their construct, and this was rarely demonstrated.

Other methodological considerations of this review should be borne in mind when interpreting the results. First, the COSMIN framework is relatively new, and its implementation is not without some travails. Some rating decisions remain a matter of significant judgement as to what constitutes violations of particular standards. Examples include whether hypothesis testing was guided by “adequate” or “poor” descriptions of comparison instruments, without a guide to expectations. Future development and testing of the COSMIN framework would be welcome. Second, the review is limited by its focus on questionnaire-based instruments. Observation-based measures, such as Dementia Care Mapping [51], may be suited to some research circumstances and could form a basis for criterion validity assessments. Further, the review has restricted its focus to measures developed and/or tested in older adult services, though this is not to say that other measures are necessarily inappropriate. However, before using such instruments, it would be imperative to inspect and test their validity with older people using long-term care services.

Person-centredness is now regarded as a central component of any high quality long-term care service for older people. However, those seeking to evaluate change and improve standards have limited evidence to support their choice of measurement instruments. This review aimed to identify, describe and critically appraise relevant measures. Eleven instruments were included. The review found that references testing measurement properties were generally of low methodological quality. Two measures (the P-CAT and CCCQ) stand-out as having been tested beyond the initial development stages, though concerns remain over the structural validity of the former, and construct validity of the latter. The review recommends closer attention to methodological quality in testing measurement properties, and greater inclusion of service users and families in item development and validation.