Introduction

Delirium, characterised by an acute disturbance of cognition, arousal and inattention, affects around 1 in 4 older inpatients [1]. It is substantially underdiagnosed, in part due to its fluctuating nature and diversity of clinical manifestations. Knowledge of patients’ baseline cognition is critical for delirium detection, and can be a barrier for health professionals if this is lacking [2]. Symptoms of delirium may be misattributed to causes, such as depression, dementia or normal ageing [3]. Hypoactive delirium is most commonly missed and is associated with worse outcomes [4]. Missed diagnoses may contribute to excess mortality [5, 6], making systematic detection of delirium essential in any setting.

Families and carers are generally well placed to identify changes in an older person’s cognitive state and they too can under-appreciate delirium symptoms if the term or concept is unfamiliar. Yet this initial recognition and response to delirium is essential to prompt medical assessment and treatment, and perhaps reduce associated morbidity and mortality. To understand this, we investigated the awareness, recognition and response to delirium symptoms in adults admitted to an acute medical unit.

Methods

Participants

We included adults with delirium admitted from the community to the acute medical unit at a large university hospital over four months in early 2019. Patients were approached on a convenience basis and excluded if they were under the care of another primary team (e.g. surgery, haematology, oncology). Patients who developed delirium during admission, or who were admitted from another hospital or rehabilitation unit were excluded. All data were collected by a team (KG, AK, RS and AJ) who received standardised training by KG.

Diagnostic measures

Delirium diagnoses were made by the consultant geriatrician in charge of that patient’s care and diagnoses were confirmed by KG using a protocol which included the 4AT [7].

Outcome measures

A delirium recognition questionnaire (Appendix 1) was developed by a clinical research fellow (KG) and a consultant geriatrician (DD). It comprised three sections: (1) process of delirium detection by health professionals; (2) recognition of delirium by the person who sought medical help, termed the responder (which may have been the patient themselves); (3) responder knowledge of delirium.

We determined vital status at four months through chart review. Deaths occurring outside of hospital were captured through daily updates on the NHS Spine, a collection of local and national databases and systems containing demographic information.

Other variables

We recorded basic demographic and clinical information. Cognitive status was classified as: dementia, if they had a clear diagnosis on GP records or previous hospital documents; undiagnosed cognitive impairment, if there was evidence of cognitive impairment on previous records or collateral history but no formal diagnosis of dementia; or no cognitive impairment. We noted the timing of any documentation of delirium, along with any associated symptoms. The responder was identified from paramedic notes and physician history, and they were approached either in person or by phone if not available.

Data analysis

Data were mainly descriptive, though reported symptoms were categorised in relation to length of time to response (< 6 h, 6–24 h, 1–2 days, 3–7 days, 1–2 weeks). Symptoms documented by health professionals were recorded by day of admission (day 1, 2, 3, 4 or later) as well as diagnosing service (emergency department, acute medicine, specialist geriatrician). For the symptoms used to recognise delirium, we assessed differences between health professional and responder recognition using McNemar’s test.

Results

Participant characteristics

On each day, data were collected, all eligible patients were included. Sixty patients were included (mean age 85 years, SD 6.77). There were 27% (n = 16) living with dementia, 30% (n = 18) with undiagnosed cognitive impairment, and 43% (n = 26) with no cognitive impairment.

Delirium recognition

Delirium was documented in the notes in 88% of cases, mostly on Day 1 by the acute medical team (Fig. 1).

Fig. 1
figure 1

The documentation of delirium by day and team. ED Emergency Department, AMU Acute Medical Unit, AG acute geriatrics

Full size image

The responders

Most (63%) responders were family members; 10% were paid carers, three were friends, one neighbour, one health professional and in one case, the patient phoned the ambulance. In 15%, we were unable to identify the responder. Responders were either contacted by phone (39%) or face-to-face (37%), and 24% were unable to be contacted after three attempts. A total of 37 responders were available to complete the questionnaire.

Responder responses

Common themes for the first indication of illness were: lack of responsiveness, drowsiness, poor appetite, confused speech, and loss of mobility or fall (Table 1). Thirteen (35%) responders reported feeling concerned, but did not know the cause for the patient’s symptoms, 7 thought the patient had an infection, 6 thought they had had a stroke, 9 thought there was another cause and 2 responders reported recognising a change, but not being concerned.

Table 1 First change in patient noticed, according to time taken to seek medical advice
Full size table

The initial response of most responders was to seek medical advice: 32% phoned an ambulance and 27% phoned the urgent help line or the general practitioner. 22% did nothing and three commented, “She was managing okay”, “I was just hoping it would get better" and "She seemed okay and would recover".

Twenty-two (60%) sought medical advice within six hours of the initial change in the patient, and 76% within 24 h (Fig. 2). Two responders took 1–2 weeks to respond, commenting "Very tired, confused. Just old age” and “Not eating or drinking properly” (Table 2).

Fig. 2
figure 2

Time to seek medical advice

Full size image
Table 2 Response of 1–2 week responders
Full size table

Knowledge of delirium

Twenty-five (68%) responders had heard the term delirium. However, their understanding of delirium was variable (Table 3).

Table 3 Responder's knowledge of delirium
Full size table

Responder and healthcare professional-reported symptoms

Compared with healthcare professionals, responders were less likely to recognise drowsiness, agitation or hallucinations as being part of the delirium symptoms of in the index presentation (Table 4). However, responders did consider confusion and decreased mobility as delirium symptoms.

Table 4 Responder and healthcare professional-reported symptoms
Full size table

Follow-up outcomes

At follow-up, 20% (12/60) of patients had died (mean age 86 years, SD 4.87). In seven, the responder was unable to be contacted. Among the five deceased patients whose responder did complete the questionnaire, four had uncharacteristic drowsiness as identified by the responder, compared with 33% of the presentations as a whole. All responders sought medical advice within 2 days (three within 6 h), and four responders had heard of delirium.

Discussion

We found that while family and carers recognised symptoms of delirium, their interpretation and response to them were variable. Most responders sought advice promptly, however, some took considerably longer despite identifying symptoms that gave them concern. A third of responders could either not be identified or contacted which represents a significant barrier for physicians to collate a clear account of events leading to admission. Taken together, these results suggest targeting poor awareness and understanding of delirium among families and carers could improve more prompt recognition and management.

Our data are limited to an urban population at one hospital, which may not generalise to other settings. Socioeconomic status and ethnicity of the responder were not recorded, nor were previous episodes of delirium, which may have impacted on responder behaviour. It is possible that some cases of delirium may have been missed due to reliance on ward round lists and the duty consultant to identify delirious patients. However, our study was able to capture a typical population presenting to acute care and we were systematic in our approach to standardising data collection.

Few other studies have previously examined the pre-hospital appreciation and response of delirium in patients. The awareness and recognition of delirium among family and caregivers has been reported to be low when presenting theoretical scenarios [8]. Others report that nearly 97% of family caregivers had not heard of delirium [9]. Family educational interventions have tried to improve prevention and early recognition of delirium during hospital admission [10, 11]. Such strategies could be broadened to include families’ future recognition of delirium.

Our findings highlight a need for delirium education in family and carers and for greater public awareness. Uncharacteristic drowsiness appeared to be more common in people that died within four months, suggesting this could be a particular feature worth targeting. Memory clinics are an opportunity to offer delirium education to family and carers of at-risk patients. Other potential routes of education could include local pharmacies, care staff agency, nursing home staff, paramedics and online training for health professionals. Public health initiatives, such as a delirium equivalent of the FAST test for stroke, may increase delirium awareness and response [12]. A key research priority is to investigate the time of symptom onset and its relationship to the time to medical intervention and outcomes. All education interventions would benefit from an emphasis on the recognition of hypoactive delirium with the aim of improving recognition and outcome.