Clinical Complexity and Use of Antipsychotics and Restraints in Long-Term Care Residents with Parkinson’s Disease

INTRODUCTION

Parkinson’s disease (PD) is a progressive movement disorder that is often complicated by neuropsychiatric manifestations (including dementia, depression, anxiety, sleep disturbance and psychosis) and autonomic dysfunction (including constipation, urinary frequency, erectile dysfunction, and orthostatic hypotension) [1]. PD is the second most common neurodegenerative disorder after Alzheimer’s disease, with an overall lifetime risk of 1.5% [1]. PD primarily affects older individuals, with a sharply rising incidence after the age of 60 years, and its prevalence continues to increase with population aging [2]. PD results in impaired quality of life, disability, caregiver burden, greater health care utilization and costs (drug costs, home care, physician visits and hospitalization), and an increased risk of mortality [3–11]. The care of persons with PD is further complicated by a rising burden of age-associated comorbidities such as cardiovascular disease, diabetes, and arthritis, and geriatric syndromes such as falls and fractures, urinary incontinence, and frailty [3, 12–17].

Ultimately, many persons with PD are admitted into Long Term Care (LTC) homes. Between 5 and 10% of LTC residents have a diagnosis of PD, with admission often associated with cognitive impairment, hallucinations and functional impairment [2]. Residents with PD are typically younger on average than others, suffer from a greater burden of comorbidities, cognitive impairment and psychosis, and are more likely to experience poorer quality of life and accelerated functional decline [2]. The mortality rate of LTC residents with PD is approximately 50% at three years, risk factors for which include cognitive and functional impairment, pneumonia and heart failure [2, 18]. There is evidence to suggest that poor outcomes experienced by persons with PD may, at least in part, reflect suboptimal management and associated comorbidities [2, 19].

The management of PD can lead to therapeutic competition, an adverse drug event whereby the treatment of one symptom or disease feature leads to a side-effect, the treatment of which exacerbates the initial symptom or feature [20]. In PD, dopaminergic agents, prescribed to improve mobility, may precipitate psychosis, leading to the prescription of an antipsychotic, which then may counter the motor effect of dopaminergic agents. Up to 50% of persons with PD are eventually prescribed an antipsychotic drug, and studies among LTC residents with PD show that prescribing rates range from 15% to almost 40% [21–27]. Justification for the use of antipsychotics is often lacking, with some studies showing that psychosis is documented in as few as a quarter of recipients, who are thus unnecessarily exposed to a high risk of adverse events [27–30].

The use of restraints is also common in LTC, usually with the aim of preventing falls and managing behavioural disturbances among residents [31]. While evidence of benefit is lacking, substantial evidence shows that restraint use can result in harm, including agitation, anxiety, depression, delirium, functional decline, pressure ulcers, aspiration, falls, fractures, and death [32–39].

The optimal management of PD in LTC homes may thus be complicated by the clinical complexity of affected residents, resulting in excessive use of antipsychotics and restraints. To date, the small number of studies to examine characteristics of PD in LTC homes have been limited by small samples or have only been analyzed on a regional level. Moreover, available data on antipsychotic use are limited and there have been no studies to our knowledge on restraint use in LTC settings in persons with PD. The objectives of this paper are to 1) characterize the clinical complexity of older persons with PD living in LTC and Complex Continuing Care (CCC) facilities in Canada; 2) estimate the prevalence of antipsychotic prescribing and restraint use in this population; and 3) identify correlates of antipsychotic and restraint use in this population.

METHODS

RESULTS

Characteristics of LTC and CCC residents with PD are shown in Table 1. There were 12,206 LTC residents and 7,851 CCC residents with a recorded diagnosis of PD, for prevalence figures of 6.5% and 4.3%, respectively. The mean ages of LTC and CCC residents with PD were 82.6 years (95% confidence interval 82.4, 82.7) and 79.6 (95% confidence interval 79.4, 79.6) years, respectively, with the overwhelming majority being over the age of 75 years. There were significantly fewer CCC compared to LTC residents in the 85 years-and-over age category, including those with PD, and fewer women with PD compared to those without PD.

The most frequent comorbidities in the entire sample, also shown in Table 1, were dementia, followed by cardiovascular conditions and stroke, depression, arthritis, diabetes mellitus and COPD. Residents received on average over 10 medications each. Dementia, hypertension, arthritis, and depression were more common in LTC residents than among CCC residents, whereas major cardiac diseases, cancer and renal insufficiency were more common in CCC. Compared to residents without PD, those with PD were significantly more likely to be diagnosed with dementia, depression, and less likely to be diagnosed with stroke, arteriosclerotic heart disease, heart failure, arrhythmia, PVD, CRF and COPD.

Among mental health indicators, residents with PD were more likely to have more severe cognitive impairment than those without, according to CPS scores, with those in LTC being more severely impaired than those in CCC. In all, approximately 90% of residents in LTC and CCC had some degree of cognitive impairment. Further, over half of residents had at least moderate to severe cognitive impairment, particularly among CCC facilities (61% of PD versus 46% all other residents) as indicated by a CPS score of at least 3 or greater. Those with PD were more likely to manifest indicators of worsening cognitive decline than those without. Clinically meaningful symptoms of depression, as indicated by DRS scores of 3 or greater, were common with rates of 28.8% in LTC and 23.1% in CCC facilities, respectively. Over half of LTC residents and 70% of those in CCC had significant behavioural symptoms. Residents with PD were significantly more likely than those without PD to exhibit delusions and hallucinations, though overall the prevalence of these symptoms was low at 7.9% and 11% in LTC and CCC, respectively.

Among functional and health indicators, fewer than 5% of residents in either LTC or CCC facilities had no documented ADL impairment, with a large majority of those in both sectors having moderate to severe ADL impairment. Residents with PD were significantly more likely to exhibit more severe functional impairment, and were particularly more likely to have urinary incontinence than those without PD. Daily pain was documented in 45% of LTC residents with PD and over two thirds of those in CCC, though only a small proportion were documented to suffer from severe daily pain. There were no significant differences in health instability between residents with or without PD according to CHESS score distributions, though residents of CCC generally exhibited a greater degree of medical complexity than those in LTC. Finally, residents with PD were more likely than those without PD to experience falls and gait abnormalities, and more so in CCC than in LTC.

DISCUSSION

The results of this large cross-sectional study, conducted across multiple Canadian jurisdictions, illustrate well the substantial clinical complexity of LTC and CCC residents with PD. With population ageing and changing admission criteria for LTC homes, this complexity will continue to grow. The prevalence of PD in this sample is consistent with previous literature, as are the findings related to age and gender [2, 3]. Not unexpectedly, health instability, as indicated by the CHESS scale, was greater among CCC residents than those in LTC.

Cognitive and psychiatric comorbidities were common and were characterized by significant symptom burden. These findings are also consistent with previous literature and reflect the neuropsychiatric complications of PD [1]. Medical comorbidities, including major cardiovascular disease, arthritis, diabetes, and COPD were also very common, although less common than those without PD. These conditions are associated with a significant burden of disability and health service utilization among seniors, both in the community and in institutional settings [54, 64–66]. Cardiovascular comorbidities may be particularly problematic in persons with PD, in whom a predisposition to orthostatic hypotension and falls may be especially exacerbated by comorbid cardiovascular conditions and associated medications [67–69]. Further, comorbid COPD may predispose to further respiratory complications in a population with significant pulmonary compromise arising from PD [37, 66].

More than one third (34%) and nearly a quarter (24%) of LTC and CCC residents with PD were prescribed antipsychotics, respectively. A 2003 study of LTC homes in 5 countries found a prevalence of antipsychotic use ranging from 11 to 38%, with no consistent associations shown to explain this variability. Recently published data from the 2004 US National Nursing Home Survey showed that the prevalence of antipsychotic use in LTC residents with PD was 31% [28]. In that study, the use of two or more psychotropic medications was associated with the presence of depression, schizophrenia, psychosis and anxiety. Specific predictors of antipsychotic use were not identified. Despite using more recent data, we found that antipsychotic use remains very common still. However, several provincial and national initiatives to reduce inappropriate antipsychotic use are now underway based on interRAI quality indicators drawn from the RAI 2.0 [70].

Guidelines recommend that the use of antipsychotic medications in persons with dementia be limited to those exhibiting agitation, aggression or psychosis of such severity that there is risk of harm to them or to others [71], and only after non-pharmacological measures have failed to reduce this risk. In our sample, the presence of delusions or hallucinations was a strong predictor of antipsychotic use, as was the presence of aggressive behaviour. However, the prevalence of antipsychotic use was far greater than that of severe aggression and psychosis combined, and was associated with other resident characteristics including cognitive impairment, depression, greater comorbidity and medication counts, male gender, younger age and preserved functional capacity. Given the risks associated with the use of antipsychotics in patients with PD, which include worsening mobility, falls, cognitive impairment, infections and mortality, the prescription of antipsychotics without clear justification of benefit is concerning [22, 29].

The prevalence of restraint use in this study is also cause for concern. A 2003 study of 14,504 LTC homes in 5 countries found that rates of restraint use varied from 6 to 31%, with much of the variability remaining unexplained [72]. Whereas evidence suggests that rates of restraint use of no more than 5% should be achievable in LTC homes [38], our study found that one quarter of LTC and CCC residents with PD were restrained, rates that were substantially higher than those among residents without PD. In addition to falls, multiple resident characteristics were associated with restraint use, including male gender, functional and cognitive impairment, depression, urinary incontinence, psychotic symptoms and aggressive behaviour. There was marked inter-jurisdictional variability and again, physician visits were modestly protective.

These findings suggest that LTC residents with PD do not receive optimal care and that many are exposed to serious risks associated with antipsychotics and restraint use despite the absence of clear indications. Residents with less access to physicians in LTC and CCC facilities are more likely to receive antipsychotics and be restrained. In addition, the data suggest that men of younger age are more likely to be prescribed antipsychotics or restraints. Younger men may be perceived as more dangerous than older men, and in the absence of behavioural support programs, may be prescribed antipsychotics or restraints with the aim to protect staff, other residents or individuals themselves. Similar gender findings were seen in a cross-sectional study of restraint use in inpatient mental health settings [73]. Further, other symptoms such as incontinence may be exacerbated, significantly decreasing quality of life.

A key concern appears to be related to physician access. The majority of physician services for LTC residents are provided by family physicians, with a recent Ontario study showing that over 90% of care to over 50,000 LTC residents was provided by 628 family physicians, most of whom are also aging [74, 75]. However, most Canadian family physicians receive limited exposure to geriatric medicine or neurology during residency training, and furthermore, the number of family physicians practicing in Canadian LTC settings is at risk of declining [75, 76]. Other studies have shown that the involvement of specialists in the care provided to persons with PD may result in a more accurate diagnosis and potentially slower functional decline and fewer hip fractures [77, 78]. However, recent evidence suggests that, at least in the ambulatory care setting, ageing seniors with multiple comorbidities are less likely to receive care from specialists [79]. Access to specialist care may be even more restricted in institutional settings [80]. In Canada, only 32% of the approximately 240 geriatricians in the country provide services to LTC homes [81, 82]. However, greater access to specialist care alone may be insufficient to improve the quality of care provided to institutionalized seniors with PD. For example, evidence suggests that neurologists may be limited in their ability to recognize the neuropsychiatric complications of PD [83].

There is widespread recognition of the importance of generalist oversight in the management of complex patients with multiple comorbidities, and that primary care providers can play an important and effective role in the management of these patients, including those with PD [84–88]. The management of patients with complex chronic conditions must be rooted and sustained in primary care and involve a truly collaborative shared-care approach between primary care clinicians, specialists, and inter-professional health providers [89]. Emerging data from other settings suggest that shared care approaches involving specialized and primary care, ideally co-located, can lead to more optimal management and improved outcomes among community-dwelling patients with complex conditions such as heart failure, dementia, or frailty [89–94]. Similar approaches must be explored for institutionalized seniors with PD. An example of a successful inter-professional shared care model, in which full use was made of the clinical planning capabilities of the interRAI LTCF (the updated version of the MDS 2.0 used in this study) [95], was evaluated in a cluster randomized controlled trial in Dutch LTC homes [96]. Intervention homes saw an overall improvement in care quality, as measured by a sum score of 32 risk-adjusted quality indicators, and overall better pain control, reduced frequency of problematic behaviours, and less antipsychotic use.

The examination of a large population-based sample of LTC and CCC residents and the comprehensive nature of the RAI 2.0 data are important strengths of this study, and the findings reflect the true complexity of the health characteristics of a real world sample of institutionalized persons with PD. However, some limitations should be noted. The contents of the CCRS database are limited to representing jurisdictions that routinely collect RAI 2.0 data. This precludes a truly representative national sample. Further, as the RAI 2.0 has been mandated in Ontario since 2003, the majority of cases in our sample are from this province. Nevertheless, the prevalence and similarities between our sample and those from previous studies does attest to the representativeness and generalizability of our results. The RAI 2.0 item “Parkinson’s disease or Parkinsonism” does not permit the distinction between idiopathic Parkinson’s disease and other etiologies of Parkinsonism. However, work done earlier in the context of this ideasPNC project compared this RAI 2.0 item to the validated Discharge Admission Database maintained by the Canadian Institute of Health Information, and showed a very high specificity and very good sensitivity for Parkinson’s disease [40]. Moreover, the use of antipsychotics and restraints is concerning for any patient with Parkinsonism. Finally, the cross-sectional design limits any definite interpretation regarding the direction of observed associations. Specifically, the quality of care provided in LTC or CCC reflects more than simply physician availability. However, the apparent inverse relationship between the extent of physician contact and the use of restraints strongly suggests the importance of access to physicians in the quality of care provided.

CONCLUSIONS

Residents with PD and who reside in LTC and CCC are affected by a substantial burden of comorbidities, both associated to PD itself, and also associated with ageing itself. Our data suggest that the care of these patients may be suboptimal, potentially due to limited access to physicians. Integrated and inter-professional shared-care approaches, centered on the full clinical care planning capabilities of interRAI instruments, may offer a solution towards improved care of this complex patient population.

CONFLICT OF INTEREST

Drs. Heckman, Kergoat and Hirdes are fellows of interRAI, an international not-for-profit research consortium. Otherwise, the authors have no conflict of interest to report.