Prof Kenneth Rockwood is Director of Geriatric Medicine Research at Dalhousie University, Canada and serves on the International Advisory Panel of Age and Ageing journal.
People who are frail have accumulated more health deficits than have others of the same age. Deficit accumulation occurs when damage either goes unremoved (or unavoided) or unrepaired. That’s why rational management of frailty needs to pay attention both to reducing the exposure to damage, and to increasing its removal or repair. The same can be brought to mitigating the consequences of acute illness and its treatment. A recent study of patients admitted to an Intensive Care Unit (ICU) shows how.
In the before/after controlled trial of a bundled set of interventions to improve outcomes of critically ill patients (296 aged 19-late 80s, 187 of whom were intubated) Balas et al. were able to reduce delirium duration and mobilize patients early. The study was conducted in several critical care units in a large tertiary care teaching hospital. It evaluated a “bundle”, being a series of evidence-based interventions to improve patient outcomes. Some aimed to reduce damage, such as by not prolonging sedation or mechanical ventilation. Others aimed to increase repair, such as by early mobilization. Usefully, the outcome measures included undertakings that might arise as adverse consequences of the intervention – for example, reintubation for patients who failed a trial of spontaneous breathing. This strikes me as an important point for dealing with frail older adults: imagine if other interventions recorded adverse effects as part of the outcome measures.
Before the evidence-based vandals clobber this study with their checklist cudgels, let’s consider the issue of the before/after design, compared with a randomized one. Before /after designs are subject to bias that arises from temporal trends – for example, if delirium was becoming less frequent anyway in critical care settings, then the after results would not necessarily reflect the effect of the intervention. Similarly, unspecified confounding can arise in relation to the time in which the study was done (hypothermia being less likely in the summer, for example). The Hawthorne effect (where people perform better simply in consequence of being observed) can operate. Such factors favour a randomized design. But not necessarily in an educational intervention: there, so-called “contamination bias” is a key concern. In other words, if you teach people how to do something better, you cannot expect them to do it to some patients and not to others. For randomization to work, it would need to be at the level of the group of people being educated, not at the level of the patient. In single institutions, this becomes problematic, so that the before/after design is the best design for that circumstance.
Two additional considerations make its interpretation appropriate in this study. First, the “before” observation period included the training period, so that even though contamination bias was not eliminated, it makes the interpretation of the effects of the intervention conservative. Second, given that each element in the bundle had been tested, the study is an exercise in knowledge translation. The discipline of Knowledge Translation is young but evolving, and it is getting to grips with the fact that the point of a KT study is not to replicate the controlled trials evidence, but to study implementation and uptake. Here, for example, the authors noted that some parts of the intervention were taken up less often than others: continuous sedation was not avoided as much as might have been expected. The failure of exhortation is balanced against the consideration that there might be merit to the concern that pain relief trumps sedation in many patients. Would that we knew as much about the uptake of various components of geriatric assessment!
If we are to make headway in relation to the management of frail older adults, especially when they become acutely ill, it will not be enough for us to use frailty assessment to reduce the risk of common interventions by having fewer frail people exposed to them. Frail older adults in many ways are the canaries in the coal mine: they are showing us many of the adverse effects of routine hospitalization. No one benefits from sleep deprivation, or malnutrition, or inadequate pain control, or not having goals of care made clear: frail people show more than others just how harmful this is. Better care for frail older adults involves making care better in general and not just less available. And making care better can start with doing less harm, and seeking out to repair what harm we find, especially that which we have caused.