Goals of care and end-of-life conversations

As individuals' progress through the stages of dementia, cognition, and the capacity to make informed decisions is often lost, with few individuals creating advanced care plans outlining their specific wishes, families are often left to make decisions regarding goals of care and end-of-life care. Palliative care places a focus on having these vitally important discussions addressing goals of care and end-of-life care. Hanson et al. (Reference Hanson, Zimmerman and Song2017) noted that using a goals of care video aids along with structured end-of-life discussions as part of a palliative care intervention led to a statistically significant increase in concordance between family decision-makers and clinicians at 9 months of follow-up or at death (133 [88.4%] vs. 108 [71.2%], p = 0.001). Hendriks et al. (Reference Hendriks, Smalbrugge and Hertogh2017), however, highlighted that creating advanced care plans may not be enough. They noted that 21% of nursing home patients with dementia were hospitalized despite a “do not hospitalize” order, underscoring the importance of healthcare provider education around respecting an advanced care plan or specific palliative goals (Hendriks et al., Reference Hendriks, Smalbrugge and Hertogh2017). Sampson et al. (Reference Sampson, Jones and Thune-Boyle2011) noted another interesting theme surrounding advanced care planning; despite its benefits for those with advanced dementia, throughout their study, there was a reluctance from patients and family members in making these advanced care plans (Sampson et al., Reference Sampson, Jones and Thune-Boyle2011). Perhaps, this reluctance could be mitigated by discussing disease trajectory (Figure 4). Evidence shows that discussing disease trajectory and what to expect helps patients and their caregivers be more willing to accept palliative care in advanced stages and improves satisfaction with the care (Teno et al., Reference Teno, Lynn and Phillips1994).

Fig. 4. Disease trajectory of a chronic life limiting illness like dementia (Murray et al., Reference Murray, Kendall and Boyd2005).

Prognostication

Adding complexity to end-of-life planning is the lack of a reliable prognostication tool in dementia. Brown et al. (Reference Brown, Sampson and Jones2013) reviewed prognosticators in patients with advanced dementia and found that the Functional Assessment Staging 7c criterion was likely not a reliable predictor of 6-month mortality. Their systematic review concluded a lack of prognosticator concordance across the literature (Brown et al., Reference Brown, Sampson and Jones2013). Similar to the Functional Assessment Staging 7c criterion, the palliative performance scale is an assessment tool used by clinicians to assess functional status and to prognosticate. The tool was initially validated amongst oncology patients where it is a reliable predictor of survival. Unfortunately, this was not reflective in the dementia community. When compared, a low palliative performance scale (<20) indicated short prognoses in malignant, nonmalignant, and dementia patients (survival <3 days, sensitivity 75%, and specificity 93%) but with higher palliative performance scales (>20), medium-term life expectancy was not reliably predicted in nonmalignant or dementia patients (Linklater et al., Reference Linklater, Lawton and Fielding2012). The Gold Standards Framework Proactive Identification Guidance (PIG) supports the use of the “surprise question” (Figure 5) in conjunction with general indicators of decline (Figure 6) and specific clinical indicators (Figure 7) to identify patients nearing the end of life (The Gold Standards Framework, 2010). The “surprise question” in a study did predict death better than intuition (53.2%, intuition = 33.7%; p = 0.001) but with a higher false positive rate. Unfortunately, similar to the palliative performance scale, the “surprise question” has been shown to predict death more accurately in patients with cancerous illness vs. noncancerous illnesses [95% confidence interval (CI): 0.79–0.87 vs. 95% CI: 0.73–0.81, respectively; p = 0.02] (Downar et al., Reference Downar, Goldman and Pinto2017; Mitchell et al., Reference Mitchell, Senior and Rhee2018). Echoing indicators of decline, such as hypoalbuminemia and weight loss, the Functional Assessment Staging 7c criterion did find anorexia [hazard ratio (HR) 2.22; 95% CI: 1.52–3.44], dry mouth (HR 1.81; 95% CI: 1.23–2.67), and cachexia (HR 1.27; 95% CI: 1.3–1.55) to be associated with increased mortality at 6 months (Brown et al., Reference Brown, Sampson and Jones2013). The PIG framework challenges the notion of exact prognostication and instead proposes predicting need as being more valuable than prognostication of exact time remaining. Whether prognostication or identifying need helps patients and/or families living with dementia, placing a palliative care focus in the early stages of the illness is essential to ensure that patients and families are aware of the importance of goals of care, end-of-life planning, and disease trajectory (The Gold Standards Framework, 2010).

Fig. 5. The surprise question (The Gold Standards Framework, 2010).

Fig. 6. General indicators of decline (The Gold Standards Framework, 2010).

Fig. 7. Specific clinical indicators for dementia (The Gold Standards Framework, 2010).

Symptom management

A common theme in the literature was patients and family's impression that symptoms were generally poorly managed in patients with dementia, particularly advanced dementia. When the symptoms of patients dying with cancer were compared to that of patients dying of dementia, there were no statistically significant differences in moderate to severe dyspnea (dementia: n = 28, 49% vs. cancer: n = 33, 61%; p = 0.28) and moderate to severe agitation (dementia: n = 4, 7% vs. cancer: n = 12, 23%; p = 0.09) (Soares et al., Reference Soares, Japiassu and Gomes2018). It is standard practice to involve palliative care for the management of end-of-life symptoms in cancer patients; however, it is not for patients with dementia despite evidence that both groups experience severe and comparable symptoms. Mean scores in Sternberg et al.’s (Reference Teno, Lynn and Phillips2014) retrospective cohort study indicated poor management of patients with dementia in areas of pain, shortness of breath, fear, and skin breakdown. Sternberg et al. noted that pain and agitation were the most common symptoms, advanced dementia being associated with more pain. They also noted that symptom management only intensified at the end of life (Hendriks et al., Reference Hendriks, Smalbrugge and Hertogh2017). Both studies indicate that a priority should be placed not only on educating healthcare providers but also on providing focused symptom control throughout all stages of dementia management.

Interestingly, dysphagia was sparsely discussed as a priority in symptom management through the literature reviewed in this study. Dysphagia is a symptom which can lead to aspiration pneumonia, laryngospasm, and lung abscess. Given that aspiration rates are significantly higher in the dementia population [35.6% of patients with dementia vs. 6.7% of controls (χ ²-test; χ ² = 9.895; df = 1; p < 0.01)] (Rosler et al., Reference Rosler, Pfeil and Lessmann2015), and aspiration pneumonia is the leading cause of death in the dementia population [odds ratio (OR) 2.891; 95% CI: 1.459–5.730] (Manabe et al., Reference Manabe, Mizukami and Akatsu2017), it is critical that a palliative care approach be taken where managing symptoms like dysphagia and dyspnea would be a priority.

Dementia is a particularly devastating diagnosis for it not only causes an onslaught of physical symptoms but also responsive behavior of dementia which is not as thoroughly discussed in the literature. We know that responsive behavior of dementia can be challenging to treat and distressing to patients and their loved ones. Passmore et al. (Reference Passmore, Ho and Gallagher2012) suggested that a palliative care approach which recognizes the importance of personhood and preservation of dignity would bridge the current gap which exists in providing care to patients with the complex biopsychosocial constellation of symptoms caused by advancing dementia (Passmore et al., Reference Passmore, Ho and Gallagher2012).

Emergency room visits

Emergency room visits can not only be lengthy and tedious for patients, but they place a major burden on our healthcare system. Rosenwax et al. (Reference Rosenwax, Spilsbury and Arendts2015) compared the use of the hospital emergency room by people with dementia who did and did not have access to community-based palliative care. Through their retrospective cohort study, they concluded that community-based palliative care was associated with significant reductions in hospital emergency room use. In fact, patients without access to community-based palliative care visited the emergency room 1.4 times more often in the first 3 months of study (95% CI: 1.1–1.9) with 6.7 times more visits in the weeks immediately preceding death (95% CI: 4.7–9.6). The most frequent emergency room presenting complaints included shortness of breath, altered conscious state, confusion, and nausea/vomiting (Rosenwax et al., Reference Rosenwax, Spilsbury and Arendts2015). Given the evidence that aspiration pneumonia is the leading cause of death in the dementia population, the leading emergency room chief complaint of shortness of breath was suspicious for possible aspiration pneumonias, another indication that perhaps dysphagia is not being optimally managed without standardized palliative care interventions in place (Manabe et al., Reference Manabe, Mizukami and Akatsu2017). By improving access to palliative care services for patients with dementia, it could not only spare patients and family members the time and distress of an emergency room visit but would better utilize healthcare resources.

Prescribing behavior

Prescribing behavior surrounding patients with advanced dementia was another theme identified through two studies in this review. Prescribing patterns are particularly important given the potential for polypharmacy and prescribing cascades that can occur with this medically complex population (Figure 8).

Fig. 8. Prescribing cascade (Rochon and Gurwitz, Reference Rochon and Gurwitz1995, Reference Rochon and Gurwitz1997).

Holmes et al. (Reference Holmes, Sachs and Shega2008) questioned the safety and appropriateness of medications prescribed to those with advanced dementia. They found that, on average, 6.5 medications were prescribed per person with 6 of 34 patients (18%) prescribed 10 or more daily medications. The study evaluated medications being used at the end of life and determined appropriateness based on how the medication affected quality of life and improved pain and symptom management. Twenty-nine percent of patients in this study had been prescribed a medication considered to be “never appropriate” based on the modified Delphi process (Holmes et al., Reference Holmes, Sachs and Shega2008). Common medications deemed to be “never appropriate” included cholesterol-reducing agents, bisphosphonates, certain antiarrhythmics, and anticholinergics (Holmes et al., Reference Holmes, Sachs and Shega2008). Palliative care specializes in streamlining and discontinuing medications that do not align with patients' identified goals of care. With more standardized involvement of palliative care consultation, discontinuation of these “never appropriate” medications at end of life could not only improve the quality of life but would reduce the potential for unnecessary side effects and adverse reactions experienced by the patient.

There is also a financial benefit to palliative care consultation in the dementia population. Araw et al. (Reference Araw, Kozikowski and Sison2015) compared the pharmacy cost before and after palliative care consultation in patients with end-stage dementia and found a statistically significant decrease in medication cost following palliative care consultation ($31.16 + 24.71 vs. $20.83 + 19.56; p = 0.003). There was a shift in the medications prescribed with a significant difference in the number of patients taking analgesics before and after palliative care consultation (55 vs. 73.3%; p = 0.009), and an average daily antiemetic cost showing moderate increase from pre- to post-palliative care consultation ($0.08 + 0.37 vs. $0.23 + 0.75, p = 0.047). These findings show that palliative care consultation aims to optimize the types of medications prescribed to patients, with more emphasis on symptom control but overall leading to a reduction in pharmacy costs (Araw et al., Reference Araw, Kozikowski and Sison2015). This supports the multifaceted benefit on prescribing practices shown when palliative care consultation is implemented in the management of patients with dementia.