The recognition of delirium in hospice inpatient units

Background

Delirium is a common condition affecting up to 50% of terminally ill patients ¹ and up to 88% of patients with advanced cancer. ²

The characteristic features of delirium are an altered and fluctuating conscious level, inattention and disorganized thinking. These features overlap with those seen in depression and dementia making recognition and diagnosis difficult at times. The hypoactive form of delirium is thought to be more common than the hyperactive form and is often under recognized or misdiagnosed as depression. ³

Older patients with cancer are at increased risk of developing delirium, ⁴ as are those with pre-existing cognitive impairment, poor performance status and sensory impairment. Within palliative medicine delirium poses particular problems. It impairs communication making it difficult for professionals to assess symptoms, and difficult for patients to make autonomous decisions about future care. In addition, delirium can be distressing in itself and this may contribute to the level of emotional distress for families and carers. Previous work has shown that in patients who had recovered from an episode of delirium, 50% could recall the event, with 80% of these grading it as a severely distressing experience. ⁵

In vulnerable patients, an episode of delirium can be triggered by many factors, including infection, hypoxia, dehydration and medication changes. It is known that early diagnosis can lead to rapid improvement with decreased length of admission and reduced long-term complications.^6,7

Healthcare providers do not use a standardized assessment tool and there is a general failure by specialists and non-specialists to recognize and document the diagnosis of delirium. In Yorkshire, one hospital palliative care team noted that the features of delirium, when recognized, were often documented using synonyms, which they used to establish a prevalence of delirium. They felt that raised awareness, especially of hypoactive delirium, and increased confidence in diagnosis could enable teams to institute timely management. ⁸

Recently published NICE guidance recommends screening for those at risk, during initial assessment on admission to hospital and long-term care facilities. ⁹ The recommended screening tool is the short Confusion Assessment Method (CAM), which is an observational tool specific for delirium. The CAM-ICU is a modified version for use on the Intensive Care Unit. The full guideline acknowledges that the Abbreviated Mental Test Score (AMTS) also has a high sensitivity and specificity in screening for cognitive impairment. The British Geriatric Society ¹⁰ and The Royal College of Physicians ¹¹ also advocate screening on admission to hospital for all patients aged over 65 years using the 10-point AMTS or the 30-point Mini Mental State Examination (MMSE)

Despite the availability of many screening tools and the inclusion of the AMTS on some admission documentation, the tools are not completed in all patients and the reasons for this are often poorly documented. This issue was highlighted in a recent audit performed in two hospices in Yorkshire.

Method

A retrospective review of 101 sets of hospice case notes was performed for consecutive admissions to a 32-bed hospice in Leeds. The AMTS is printed on the medical admission document as the standard screening tool to be completed by the admitting doctor. This audit was then repeated at a 13-bed hospice within Yorkshire which also uses the AMTS as a standard tool on the medical admission paper work. Within both hospice units the admitting doctor could be of any grade ranging from first post-graduate year (Foundation Year 1) to consultant level. The induction programme for new doctors would cover being shown the admission proforma but no specific guidance or training on the use of the AMTS. It is assumed that doctors are already familiar with using this tool.

Results

These audits demonstrated low rates of completion of the screening tools. Of these 101 cases, in the first audit, 36 patients were documented as having altered mental state during inpatient stay. Twenty-four of these 36 patients had not been screened on admission. However, in 23 cases, altered mental state was noted within 1 day of admission and may have been evident on screening at the first assessment. Often no justification was documented for omitting a screening test. Following a change in management, 13 of the 36 patients experienced full resolution (4 of these patients were discharged from inpatient palliative care), 1 patient experienced partial resolution and was discharged to a nursing home. The second audit showed that of 69 consecutive admissions to another Yorkshire hospice, 17 (25%) patients developed delirium within 72 hours. Eight of these 17 had no documented record of cognitive assessment from the initial assessment. Justification was documented in 4 of the 8.

Discussion

Although delirium is often multi-factorial and not always reversible we propose that earlier detection and management with both pharmacological and non-pharmacological measures may improve the outcome for the patient and also improve the experience for those caring for them. Early detection may be hampered by a lack of awareness, and by specific barriers to the use of screening tools. It may be possible to address both of these issues with education.

There are often occasions, particularly in the hospice setting, where completing a formal cognitive assessment is not appropriate, such as in the terminal phase. In these cases, the reasons for incomplete assessment should be clearly documented.

These two audits show that cognitive screening is often omitted from hospice admission assessments, despite a prompt in the documentation. Although it is appropriate to omit cognitive screening in some cases (e.g. where a patient is imminently dying) we are potentially missing delirium in some cases where screening would be appropriate. In this first audit it is not known how many of those with altered mental state within 1 day of admission had an altered or fluctuating cognitive state on arrival. If this had been detected then appropriate management plans could have been instigated, potentially improving the outcome.

Reasons given for non-completion included ‘not obviously confused’ or ‘alert and orientated’. Whilst NICE recommend that for experienced doctors, assessing cognition may form part of an in-depth holistic assessment, it was not clear that the admitting doctor specifically considered the possibility of delirium.

Anecdotal evidence suggests that possible barriers to the use of screening tools by healthcare professionals include risk of upsetting patients or relatives, not remembering the questions to ask, cultural or language differences and the belief that cognitive screening is not appropriate during an already lengthy admission process with very unwell patients.

These audits were limited to one deanery within the UK. Further information from other locations would be useful to understand if this is an isolated problem or representative of a wider issue.

Future work

We are currently undertaking a study to explore the possible barriers to the use of cognitive screening tools on admission to a hospice unit, as perceived by healthcare professionals.