Investigating the effects of frailty on six-month outcomes in older trauma patients admitted to UK major trauma centres: a multi-centre follow up study

Traumatic injury is a substantial health burden for older patients with a significant proportion having ongoing functional and psychological difficulties long after hospital discharge [1]. Populations are ageing, and there is a corresponding increase in the proportion of older trauma patients [2]. Trauma systems are also better at identifying previously under-recognised populations of older trauma patients [3]. Accordingly, trauma systems have adapted their practice and guidance for older patients, focussing on differing mechanisms of injury and age-related physiological variations, to ensure timely assessment and management in the early phases of care [4, 5]. However older trauma patients are a heterogeneous group in relation to age, pre-injury functional status and comorbidities, all of which might impact on longer-term recovery [6]. Recent evidence suggests that in older trauma patients, frailty rather than chronological age is strongly associated with in-hospital and 30-day mortality and adverse discharge disposition [2, 7, 8]. Frailty is a state of reduced physiological and cognitive reserve and may affect an individual’s ability to respond to a stressor event such as traumatic injury [9]. Between six and twelve months post trauma-discharge, frail patients are also more likely to have a trauma-related readmission [10] and poor functional status [11], yet the effect of frailty on longer term survival and quality of recovery is not widely reported.

Characterising pre-injury frailty may help to support prognostication in the acute phases of trauma care [12, 13] but it may also identify those who are vulnerable to longer term mortality, poor recovery and reduced quality of life. Health related quality of life (HRQoL) describes an individual’s perception of their physical and mental health, cognitive and emotional status [14]. Patient reported HRQoL measures are used to quantify the effect of injury on health, assessing functional, psychological, social and mobility dimensions to provide an insight into recovery [15]. There are a variety of HRQoL measures and in the UK the Euroquol EQ-5D-5 L measure has been implemented within a national programme of patient-reported outcomes after major trauma for all ages of adult patients [16]. A recent systematic review identified that EQ-5D-5 L had good feasibility and application when administered to older people, including when support to complete was required [17]. Understanding HRQoL following trauma is also important in planning and supporting ongoing care, and to ensure that patients and families or carers are given realistic information on their expected recovery [14]. In older, frail trauma patients HRQoL appears to deteriorate after discharge from hospital. At 30 days after injury, frailty is reported to negatively impact on the recovery of health-related quality of life compared to the discharge baseline [18]. Beyond this time-point, the effects of frailty on longer-term HRQoL after trauma are not widely reported. Frail survivors of a trauma admission may suffer a markedly worse longer-term outcome and HRQoL compared with those who are not frail. Therefore, we aimed to investigate health related quality of life at six months following discharge from hospital, in older trauma survivors with and without pre-injury frailty. We also aimed to compare longer-term care needs between these two patient groups.

This research comprised the longer-term outcome evaluation of the ‘Frailty in Major Trauma’ (FRAIL-T) multi-centre prospective observational study, carried out at five Major Trauma Centres (MTCs, Level 1 equivalent hospitals) in England [19]. The study was approved by the UK Social Care Research Ethics Committee (REC: 19/IEC08/0006).

Eligibility criteria included participants aged 65 or over, requiring trauma team activation and subsequently admitted to an MTC, irrespective of their injury severity score. All MTCs use a tiered trauma team activation (TTA) system depending on whether a full trauma team is required or not. Patients were approached for enrolment regardless of the level of TTA. Pre-injury frail status was determined during the in-patient phase of FRAIL-T and confirmed by geriatricians using the Clinical Frailty Scale (Frail defined as Clinical Frail Scale ≥ 6) [19]. Participants were consented for follow-up at the time of enrolment into the study. If a participant lacked capacity, then in line with the Health Research Authority for England and Wales guidance, a personal consultee who was able to advise on whether the patient should be included was approached. Quality of life data were collected via a questionnaire at two time points: on the day of discharge from hospital and at follow up, namely six months following discharge from hospital. When providing consent participants were asked to state their preference for receiving a follow-up questionnaire either by post or email, or administered over the telephone by a researcher trained in undertaking telephone interviews. Prior to follow-up at six-months the patient’s survival status was checked via the national digital record system (National Health Service Spine) to ensure it was appropriate to make contact. Participants who were known to be alive at follow up but who did not respond to the initial request were contacted by telephone up to three times, at which point if there no response they were assumed ‘lost to follow-up’.

The primary outcome was patient reported health related quality of life (HRQoL) at follow up compared to hospital discharge. HRQoL was assessed using two elements: the Euroqol EQ-VAS (visual analogue scale) which rates overall health on 0–100 scale where endpoints are labelled ‘The worst health you can imagine’ through to ‘The best health you can imagine’ [20]. A poor HRQoL was defined as a VAS rating ≤ 50 [21, 22]. Descriptions of the individual’s health status was measured using the EQ-5D-5 L system, a preference-based measure of self-reported health, which has been successfully validated in geriatric populations [23] and in frail older patients [24]. EQ-5D-5 L consists of five dimensions: mobility, self-care, usual activities, pain and discomfort, and anxiety and depression, and respondents report each dimension as no problems, slight problems, moderate problems, severe problems and extreme problems (or unable to complete) [25]. It has been developed for use either in face-to-face or telephone/online interviews when respondents are capable of self-reporting their health-related quality of life but unable to self-complete a paper/digital questionnaire [21]. We also recorded patient reported alteration in dependence status and care needs at follow up.

On day of hospital discharge participants were approached by a research team member and asked to complete the questionnaire. If a patient lacked capacity at follow-up, a relative or proxy (if available) was asked to complete the questionnaire on their behalf [26]. At six-month follow-up participants or their proxy were either sent an HRQoL for self-completion or contacted by telephone to complete the interviewer administered version of the tool.

Data were analysed using Stata (version 16.1). Continuous data were tested for normality using Shapiro Wilk tests. Parametric comparisons were conducted using t-tests or ANOVA with Tukeys post hoc testing, and reported as mean (standard deviation). Non-parametric comparisons were tested using Mann Whitney U or Kruskall Wallis tests with Dunn post hoc testing, and reported as median (interquartile range). All tests were two-sided. Analysis of categorical data was conducted using Fisher’s exact test and reported as number and percentage. The five EQ-5D-5 L dimensions were dichotomised into ‘no problems’ and ‘any problems’ [20] for comparison between cohorts at discharge and follow up. Multivariable linear regression analysis was conducted to evaluate the association between EQ-5D-5L VAS at follow up and predictor variables including frailty, age, sex, comorbidities, mechanism of injury, critical care admission, injury severity, length of stay and discharge destination. Robust standard errors were used and a p-value of < 0.05 was considered statistically significant. Potential non-linear effects of predictors were explored using fractional polynomial terms. Model specification was also checked by the link test and Ramsey’s regression specification-error test. Missing data was addressed using maximum likelihood method [27].