Invited commentary for International Psychogeriatrics on Tieges Z, Stíobhairt A, Scott K, Suchorab K, Weir A, Parks S, Shenkin S and MacLullich A. Development of a smartphone application for the objective detection of attentional deficits in delirium. International Psychogeriatrics. 2015 Mar 6:1–12.

April 26, 2015

People get sick the only way they know how. This fact largely is lost in contemporary healthcare, especially when it comes to old people. Having done so well for so long with ever greater specialization, and having streamlined that care to be not just effective but efficient, we modern healthcare providers much prefer that people get sick on our terms. Patients should present with single system illness (and if not that, then with many symptoms that could be ascribed to a single cause). They should get better quickly as a result of our ministrations and they should then be on their way. And if they cannot do that, they should at least be able to give a history. But the cheek of it: many people come to hospital with only the vaguest account of why they are here, and even then half of what they tell us can be wrong. Surely, they would be better off being treated somewhere else. Failing that, something should be done.

At the hospital where I work, what we did was first to show that an educational intervention could improve outcomes (Rockwood et al., Reference Rockwood1994). We then spent years proving exhaustively that even its evangelical exhortation could not make routine better. After hearing that it was too complicated, we then tried to try and make things as simple as possible. When encountering a patient who could not give a history, healthcare providers were meant simply to ask of an informant “is this new?”. Like many such interventions, (Hendry et al., Reference Hendry, Hill, Quinn, Evans and Stott2015) no lasting effect was proven.

In this issue, Tieges and colleagues (Reference Tieges2015) from Edinburgh present extensive and successful pilot data for a different approach. Reflecting, for better or for worse, (Adamis et al., Reference Adamis, Rooney, Meagher, Mulligan and McCarthy2015a) DSM-5's return from clouded consciousness to emphasize inattention as the essence of delirium, they re-cast an electronic method for testing attention (the Edinburgh Delirium Test Box) as a smart-phone app. The “DelApp” is a software application for detecting deficits of visual sustained attention in delirium. The authors report two studies: a small series that was obliged to exclude many patients of interest in delirium (the more cognitively impaired and more acutely ill) of feasibility and acceptability, and a second of discriminative ability in relation to delirium, dementia, and no cognitive impairment. It too was obliged to exclude many patients of clinical interest: two dozen who clearly fit none of the three defined categories, and three dozen who were too ill or showed too little arousal when first approached. Even so, the results were impressive, with high ROC statistics reported in relation to the ability of the software to discriminate between the three clinical groups, whose group assignment was made blind to the DelApp results.

This report has much merit. Not the least is its description of just a few of the challenges of conducting clinical research, especially on older people, and more especially still on older people with cognitive impairment. In contrast to what apparently is achievable in the Elysian fields of the typical critical appraisal journal club, where all patients consent, none is lost to follow-up, all retain their diagnostic categories, no researcher ever has a hunch that means that they might accidentally become unblinded, researchers and patients are always mutually available to each other and no one is ever too ill, Tieges and colleagues take us through some of the experience of research as she is actually done.

No study is perfect. Many experienced clinicians have their doubts about whether delirium is always distinguishable from dementia. It is not just the evident overlap between delirium and Lewy body dementia (Gore et al., Reference Gore, Vardy and O’Brien2015) or the fact that delirium for many never recovers and so merges into dementia (Adamis et al., Reference Adamis, Devaney, Shanahan, McCarthy and Meagher2015b). Instead, there is the troubling possibility that we have got delirium wrong in how we think about the brain in relation to the rest of the body. Perhaps, delirium is not best conceived of as a specific neuropsychiatric syndrome with, for any individual, a short list of predisposing and a shorter one still of precipitating factors. Delirium often reflects brains that have accumulated many neuropathological deficits (Davis et al., Reference Davis2015).

It now seems that we must go beyond this. In all likelihood, most people with delirium have more than just brain-based deficits. Just as late life cognitive impairment occurs most often in people with many health deficits, whether or not those deficits are known as traditional dementia risk factors (DeCarlo et al., Reference DeCarlo, Tuokko, Williams, Dixon and MacDonald2014; Song et al., Reference Song, Mitnitski and Rockwood2014), so we must get to grips with the fact that delirium is much more common in frail older adults, with or without known prior dementia (Eeles et al., Reference Eeles, White, O’Mahony, Bayer and Hubbard2012; Joosten et al., Reference Joosten, Demuynck, Detroyer and Milisen2014). Incomplete recovery from delirium is associated with many other adverse health outcomes that befall frail older adults (Dasgupta and Brymer, Reference Dasgupta and Brymer2015). Indeed, single core features of delirium forecast a range of adverse health outcomes, even in people who do not otherwise meet delirium criteria (Cole et al., Reference Cole2014; Lam et al., Reference Lam, Tay, Chan, Ding and Chong2014). That delirium, however defined, is clearly tied up with frailty, however defined, should give us pause. As argued elsewhere, an alternative account of delirium, at least in many old people, might simply be its manifestation as the failure of a high-order function (chiefly attention, but also time sequencing) in complex systems close to failure (Bhat and Rockwood, Reference Bhat and Rockwood2007). Presumably, for these patients, the truth of it lies a little farther away from delirium-is-all-about-the-brain pole and a little more towards the pole of delirium signifying an at-risk brain in an at-risk body.

This however is a mere quibble compared with the central challenge raised by the work of Tieges et al. For their intervention to be effective, physicians must still get the app out, as it were. Will they do it? There is reason to worry: under-recognition of delirium is a persisting problem, and attitudes towards old people are implicated (Bellilli et al., Reference Bellelli2014; Fisher et al., Reference Fisher2015). A report from Canada is telling: even when reviewing cases in which delirium was present in general medical patients, and directly relevant to the presenting illness, teaching discussions focused almost exclusively on the management of single system problems even when they were less related to why the patient had come to hospital (Diachun et al., Reference Diachun, Charise, Goldszmidt, Hui and Lingard2014). Will inattention be recognized enough to cause physicians to test for it? Or will the bad attitudes still prevail?

Tieges et al. have presented us a technically useful solution that has high acceptability. It should be taken up by young doctors who use apps routinely, and its use might even intrigue the older ones for whom smart phones remain vaguely exotic. It might even encourage young doctors to adopt approaches which can indicate that old people who are sick need to be treated properly.

Failure to engage a patient enough to recognize that they are cognitively impaired is surely a low bar; such failure likely represents poor routine care of older adults (Inouye et al., Reference Inouye, Westendorp and Saczynski2014). To realize its potential, however, we have to find ways to motivate physicians to actually use the app. An effectiveness trial is urgently needed, with specific attention to how the app might also help fix healthcare provider attitudes towards what we expect of people when they are ill.