Risk prediction instruments to guide perioperative care in elderly patients with advanced disease

High-income countries are faced with increasingly ageing populations because of rising life expectancy and decreasing fertility rates.1 This relates to a growing number of elderly patients (commonly defined for research purposes as those older than 60 or 65 years of age) being treated in acute care hospitals.2 These elderly patients, many of whom are frail, can have considerable comorbidities, reduced physical activity and cognitive impairment. Pre-existing comorbidities are related to postinterventional complications. Therefore, even minor surgical interventions or nonsurgical procedures may have considerable postinterventional complications, including death.3,4 Most of these elderly patients can be treated successfully with an adequate strategy, including early identification of patients at risk, interdisciplinary management, adequate preparation, well established clinical pathways, adapted minimally invasive interventions, preventive perioperative procedures and monitored structures with sufficient numbers of well trained nurses. These supplementary care processes and interventions during the perioperative period, increasingly referred to as ‘perioperative medicine’, may change perioperative outcomes, reducing the high complication rate and hospital stay in this population.5 These ‘geriatric, peri-interventional pathways’ can include a peri-interventional stay in an intermediate care or high-dependency unit for physiological stabilisation6,7 and/or specialised geriatric consultations during the perioperative period,8 measures which have been identified as effective. Potentially, even long-term survival could be improved, as mortality remains elevated for years in patients who develop early postinterventional complications.9,10 It is ethically inappropriate to manage the elderly with poorer care quality than that provided to younger patients. On the contrary, if we manage the most vulnerable elderly patient with the highest standards and success, we will probably also treat less vulnerable patients more successfully. However, there are limits to perioperative care in elderly patients with advanced disease. To exceed physiological and cognitive limitations is associated with an excessive postinterventional mortality.4,11–13 Therefore, considerations regarding personal perioperative goals, advance directives (including resuscitation status and limitations on invasive and intensive treatment) and potential end-of-life care must be part of the peri-interventional care of the elderly. Surgical or nonsurgical indications for an intervention in elderly patients should include an estimation of the risk of major complications or death based on the assessment of frailty, and on physical and cognitive capacities. Importantly, care should be taken to avoid over-optimistic prognosis and inappropriate indications for burdensome interventions in the last months of life, which happens too often, as suggested by a recent investigation in which 18% of patients in a cohort of elderly Medicare beneficiaries had surgery in the last month of life.14 High-risk interventions often performed at the end of life are emergency laparotomy, hip fracture repair or minimally invasive valve repair in patients with advanced dementia, and drainages in advanced and metastatic tumours often combined with a history of recent sepsis.15 Similar risks have been described in patients with end-stage liver cirrhosis16,17 or advanced neurological diseases.18 It is not unusual to witness disagreement between interventional physicians estimating the potential benefits of an intervention (e.g. cure of disease or improved mobility) and physicians responsible for preventing and treating major complications, such as anaesthesiologists and intensive care physicians. Reliable, quantitative risk assessment of comorbidities, cognition and frailty for realistic prognosis could nuance the complex preinterventional decision-making process in these elderly patients with advanced disease, contributing to a broader consensus of all healthcare personnel involved in the perioperative period for the best care of their patients. Realistic estimates of outcomes and postinterventional complications with avoidance of therapeutic illusions could optimise patient selection, personalised diagnostic and therapeutic preparation, and timing of intervention in these high-risk patients. The use of these instruments is likely to facilitate peri-interventional decisions in this group of patients, who often live in nursing homes, and may reduce burdensome postoperative complications such as uncontrolled infection, including pneumonia, decubitus ulcers and delirium with the use of mechanical restraints. At least, such instruments should improve communication between interventional physicians and anaesthesiologists by providing a common ground of solidly documented quantitative evidence. Individualised quantitative risk assessment in elderly patients with advanced disease is seldom performed, even though it is a legal obligation for doctors to provide accurate risk estimates, and patients need such information to develop realistic expectations and take de facto informed decisions about the interventions proposed to them.19 There are a number of validated risk prediction tools with high accuracy of discrimination for specific outcomes in general surgical populations20–23 or specific interventions,24,25 but very few studies on their actual implementation and use. There may be many reasons why quantitative risk assessment is not performed routinely in clinical practice. First, validated risk prediction tools may not be available for a specific intervention or a particular peri-interventional complication (e.g. bleeding or postinterventional sepsis). Second, validated risk prediction tools are available but the accuracy of discrimination is only moderate, as in the prediction rule of Marcantonio et al.26 for postoperative delirium. Third, validated prediction tools are available and accurate, but too complex or laborious for routine clinical use.27 Fourth, clinicians in anaesthesiology may have limited training in interpretation of studies testing predictive tools and inadequate understanding of concepts such as derivation and validation cohorts, calibration and external validation,28 which are the basics for high-quality quantitative risk assessment. Fifth, cultural factors and medical paternalism might be a barrier to enhanced patient autonomy and shared decision-making based on objective and accurately predicted complications and outcomes.29 Finally, robust evidence is still lacking on how preinterventional decision-making, perioperative care and outcomes can be influenced by the use of these prediction tools. The complex peri-interventional decision-making in elderly patients with advanced disease should be an interdisciplinary process, centred on each patient's individual expectations and overall life plan. It should be supported by estimated quantitative benefits of the intervention (e.g. prolongation of life or predicted functional recovery and regained gait) and validated prediction tools for complications (e.g. death, prolonged hospital stay, cognitive loss, loss of independence) delivered by anaesthesiologists. Geriatric physicians should provide objective frailty and life expectancy estimations, such as Global Deterioration Scale and Cognitive Impairment Scale,30 together with a realistic plan for physical, psychological and social re-adaptation – in a similar way to that in which, for many years, multidisciplinary teams have determined the most appropriate treatment for patients suffering from advanced tumours. Importantly, personnel in palliative and comfort care wards should participate in the elaboration of the perioperative care plan.31 Effective communication on preinterventional medical decisions with individualised concepts and goals targeted to each patient's educational, cognitive and cultural contexts is an essential part of successful medical management.32 This individualised plan should be clear and made available to all healthcare personnel directly involved in caring for the patient. Such multidisciplinary decision-making should decrease withholding of postinterventional strategies or withdrawal of life-prolonging treatments, and optimise patient satisfaction and comfort. Potentially, such preinterventional and well balanced planning based on sound scientific evidence of prediction tools toward high-quality care may also be cost-effective. Acknowledgements relating to this article Assistance with the editorial: none. Financial support and sponsorship: none. Conflicts of interest: none. Comment from the editor: this editorial was checked and accepted by the editors but was not sent for external peer review. BW is a Deputy Editor-in-Chief of the European Journal of Anaesthesiology.