This prospective analysis of 592 adults enrolled into the Western Sydney Clinical Frailty registry demonstrates that pre-morbid frailty assessed using the simple, pictorial CFS is a powerful independent predictor of rehospitalisation and/or mortality within 12 months post-discharge. A recent systematic review examining the association between CFS and adverse health outcomes in older adults in acute care settings supports our findings, reporting that all studies (n = 29) demonstrated that a CFS score could independently predict multiple adverse outcomes, such as rehospitalisation, mortality, length-of-stay, functional decline [17]. Another recent study evaluating the CFS in adults ≥ 70 years admitted to geriatric acute care reported that the CFS showed high inter-rater reliability between consultant doctors, nurses and other medical officers (intraclass correlation coefficient = 0.859, 95% CI: 0.827-0.885, P < 0.001) [21]. Furthermore, the CFS can be rapidly completed, which is helpful in a busy clinical setting.
A higher Charlson comorbidity score (≥ 3) was also an independent predictor of the composite endpoint. Higher frailty scores were significantly associated with higher Charlson comorbidity index scores. Frailty and comorbidity are closely related, with both occurring due to aging-related processes, yet they are distinct, with frailty reflecting physiological vulnerability and comorbidities reflecting specific disease burden [22]. These results highlight the importance of including an assessment of comorbidity alongside a frailty assessment. Including frailty and comorbidity measures could provide a more comprehensive understanding of overall health status, allowing for more accurate risk stratification and personalised care planning [23].
Polypharmacy was significantly associated with higher frailty scores in this cohort of older adults admitted to acute geriatric services. People classified as frail were significantly more likely to be prescribed analgesics, antacids, loop diuretics, and antipsychotics. However, they were significantly less likely to be prescribed a statin, antiplatelet or proton pump inhibitor. Our results provide real-world information regarding the use of medicines in older, frail populations. Currently, there is a lack of guidelines to inform prescribing (and deprescribing) for frail older adults and limited evidence about the pharmacokinetics and pharmacodynamics in the context of frailty [24]. The International Union of Basic and Clinical Pharmacology Geriatric Committee recently recommended that frailty be assessed in clinical trials involving older adults, both at baseline and as an outcome for efficacy and safety [25]. Therefore, it is important to routinely assess frailty in the clinical setting and clinical trials to inform future practice on the quality use of medicines.
Our results suggest that routine frailty screening using the CFS for older adults presenting to acute geriatric settings is not only feasible but clinically useful. For example, frailty scores could help inform shared decision-making conversations on the likelihood of readmission and/or death after discharge. These data could also improve decision-making in the context of invasive procedures (e.g. endoscopy), de-prescribing, and advanced care planning at a local level in Western Sydney and more broadly at a national level. There is future hope for real-time frailty assessment via Electronic Medical Records and dashboards [26,27,28], which would eliminate costly human resources needed to undertake more time-consuming performance-based frailty measures.
Clinical implications and future directions
Recently, the HARMONY model (acHieving dAta-dRiven quality iMprovement to enhance frailty Outcomes using a learNing health sYstem), a new frailty learning health system model of implementation science and practice improvement, was applied to the Western Sydney Frailty Clinical Frailty Registry [29]. Clinicians at the study site were presented with interim results from the frailty registry, and in general, there was surprise at the high mortality rate in those with severe frailty. This has important implications for acute geriatric medicine care. On the one hand, this could represent missed opportunities for preventative care but may also support the idea that these patients are in the final years of their life. In such cases, the patient’s preferences for goals of care should be discussed, including advanced care planning. It also demonstrates the importance of routinely collecting post-discharge outcome data that might be important in managing future patients.
The Western Sydney Clinical Frailty Registry recently had ethical amendments approved for ‘opt-out’ consent procedures. A consumer advisory group with Aged Care and Rehabilitation services consumers revealed that consumers believed the post-discharge phone calls provided within the registry follow-up should be standard practice. Having opt-out consent is less burdensome for participants, researchers, and clinicians. It also allows greater access to the increased follow-up post-discharge for all people admitted to the Aged Care and Rehabilitation services at Blacktown Mount Druitt Hospitals. Further, this consent model aims to reduce study selection and recruitment bias.
As per a recent recommendation by the Australian Registry of Clinical Quality Registries, we have also been approved to collect additional patient-reported outcome measures (PROMS) on willing and able participants. The PROMs include quality of life, self-reported frailty, and depression and will be collected at baseline and repeated at the 12-month follow-up.
Readmission was common in this cohort, with one in two patients rehospitalised within 12 months. The research team is conducting qualitative research with consumers, clinicians, and expert stakeholders to explore hospital transition for older adults with frailty. Consumer priorities were brain health and functional independence. Further, the use of the hospital was often viewed as an entitlement of older Australians, contrasting hospital management and policy priorities of reducing readmissions and emergency department presentations. We intend to co-design and pilot an intervention to improve the hospital transition experience.
Strengths and Limitations.
The oldest and most frail patients are often left out of clinical trials. We have demonstrated the feasibility of prospectively recruiting and following up a large cohort of frail older people from a busy acute geriatric medicine service. As noted, follow-up data was available on all participants because this was permitted from the participant’s electronic medical record. This ongoing study presents an opportunity for trials within a cohort study design to evaluate frailty interventions and data linkage studies. Our consumer advisory panel supported the research and helped revise our registry procedures.
The major weakness of our registry was the inevitable selection bias; for example, the eligibility criteria for the Western Sydney Clinical Frailty Registry stipulates that participants must speak English or have a family member who can provide consent on their behalf, which has inevitably resulted in a selected population. We also did not have the resources to recruit all admitted patients. However, we have assembled a large cohort of older adults with differing levels of frailty that has allowed us to explore the effects of frailty on important outcomes. As mentioned previously if participants were unable to be contacted for follow-up, information regarding rehospitalisation and death were extracted from medical records or publicly available death registries. While every effort was made to avoid reporting errors, using two different data collection methods may have resulted in discrepancies. Finally, as the frailty registry utilises routinely collected data, it is possible that misclassification bias and underreporting may have impacted data quality. These limitations may have affected our results and reduced the generalisability of our findings.
link