Parkinson disease (PD) is the second most common neurodegenerative disorder worldwide and the 14th leading cause of death in the United States (Marras et al., 2018; Kochanek et al., 2019). There are no disease-modifying therapies for PD, and progressive symptoms contribute to significant morbidity and mortality. Not surprisingly, age and advancing symptoms increase the risk of hospitalization (Woodford and Walker, 2005; Hobson et al., 2012; Hassan et al., 2013), and persons with PD are hospitalized at higher rates, have longer lengths of stay, and are more likely to die in the hospital than age-matched peers (Aminoff et al., 2011; Gerlach et al., 2012; Low et al., 2015; Shahgholi et al., 2017). Despite these data, little is known about PD hospitalizations in the end-of-life (EoL) period, commonly defined as the last 6 months of life (Lamont, 2005; Hui et al., 2014).

A recent study exploring inpatient treatment intensity among hospitalized patients with PD found that the final 6 months of life are a critical period in PD care (Aamodt et al., 2023a). Almost two-thirds of all United States Medicare beneficiaries with PD are hospitalized at least once in their last 6 months of life with high rates of aggressive care (Aamodt et al., 2023a). For example, of 33,107 beneficiaries with PD who were hospitalized in the EoL period in 2017, 16,266 (49%) were transferred to the intensive care unit (ICU) and 7,970 (25%) died in the hospital. Of those surviving hospitalization, 9,892 (30%) were discharged to inpatient or home hospice and 10,046 (30%) were readmitted within 30 days. These inpatient outcomes are considered markers of inappropriate EoL care quality in persons with Alzheimer’s disease (AD) and may suggest inappropriate EoL care practices in PD (De Schreye et al., 2017). However, it is currently unknown whether increased rates of hospitalization or inpatient treatment intensity among persons with PD are unique to the EoL period or occur throughout the disease course.

Building upon our existing data, the primary objectives of the current study were to (1) describe and compare the demographic and clinical characteristics of United States Medicare beneficiaries with PD who were hospitalized at least once in their last 6 months of life (e.g., hospitalized decedents) or hospitalized at least once before the EoL period (e.g., hospitalized non-decedents), (2) compare inpatient resource utilization and total inpatient charges between hospitalized decedents and non-decedents, and (3) determine whether rates of ICU admission and 30-day readmission differ between hospitalized decedents and non-decedents after adjusting for key demographic and clinical variables.

In this study using comprehensive Medicare data from 2017, we found that persons with PD who were hospitalized in the last 6 months of life were slightly older and had more medical comorbidities than hospitalized persons who survived 6 or more months after discharge. Compared to non-EoL hospitalizations, EoL hospitalizations among persons with PD were also longer and more costly. Notably, PD patients hospitalized in their last 6 months of life had twice the odds of ICU admission and 30-day readmission as those hospitalized at earlier time points after controlling for key demographic and clinical variables. Consistent with prior studies, these data suggest that the last 6 months of life are a critical period in PD care and associated with significant, potentially preventable healthcare utilization (McKenzie et al., 2022; Aamodt et al., 2023a).

When comparing the clinical characteristics of hospitalized decedents and hospitalized non-decedents, decedents in our PD cohort were more likely to have 6 or more medical comorbidities and had increased rates of dementia, which likely contributed to group differences in hospitalization rates (Phelan et al., 2012). Not surprisingly, age and comorbidity burden are associated with an increased risk of all-cause hospitalization and mortality in older adults (Miller et al., 1998) and persons with PD (Braga et al., 2014). Hospitalizations among PD patients can also lead to hospital-related complications that may hasten death (Aminoff et al., 2011; Gerlach et al., 2012; Low et al., 2015), underscoring the need to ensure that all hospitalizations are medically necessary and aligned with care preferences. Thus, a multi-disciplinary approach involving general, neurological, and palliative care providers is required to prevent inappropriate hospitalizations whenever possible.

Next, when comparing inpatient resource utilization between hospitalized decedents and hospitalized non-decedents admitted for all causes, inpatient treatment intensity was significantly greater among decedents, with longer lengths of stay and more costly admissions in the last 6 months of life. Our findings support results from prior studies which also showed that persons with PD have longer lengths of stay than individuals in the general population (Gerlach et al., 2011). Longer hospital stays in PD patients have been associated with delayed or missed administration of dopaminergic drugs (Martinez-Ramirez et al., 2015), administration of dopamine receptor blocking agents (Martinez-Ramirez et al., 2015), elective surgery (Kleiner et al., 2019), and post-operative delirium and other complications (Pepper and Goldstein, 1999; Abboud et al., 2020). However, further research is needed to determine risk factors for prolonged length of stay in the EoL period. With regard to cost, hospitalizations account for the greatest proportion of EoL healthcare spending relative to other spending categories (French et al., 2017). In the United States, Medicare spending in the last year of life accounts for up to 25% of total Medicare costs (Duncan et al., 2019), and increased spending at EoL may reflect inappropriate care practices. Hospitalizations in the last 6 months of life may also be incongruent with care preferences, reiterating the importance of advance care planning for persons with PD to improve quality of life and reduce spending at the national level.

Lastly, we found that hospitalized decedents with PD were twice as likely as hospitalized non-decedents with PD to be admitted to the ICU or experience a 30-day readmission after controlling for age, sex, race/ethnicity, rural/urban residence, and comorbidities. ICU admissions among persons with PD are often unrelated to PD duration or severity (Paul et al., 2019), and ICU mortality data are overall mixed. In one study comparing ICU admissions between persons with and without PD in Brazil, those with PD had longer hospital stays but did not experience an increased mortality risk resulting from their ICU admission (Réa-Neto et al., 2021). However, in a second study involving a random sample of elderly Medicare beneficiaries, ICU length of stay was associated with an increased risk of 1-year mortality among ICU survivors (Moitra et al., 2016). Hospital readmission is also common among persons with PD (Shahgholi et al., 2017). Although we could not determine the exact reasons for readmission in our cohort, other studies have demonstrated that readmissions among PD patients are associated with medical comorbidities, elective surgeries, and caregiver strain (Shahgholi et al., 2017; Fullard et al., 2020). For example, persons with advanced PD become increasingly reliant on caregivers for support, and increasing physical dependence can lead to caregiver burden (Lo Monaco et al., 2021; Aamodt et al., 2023b). Caregiver burden and depression are associated with a higher risk of emergency department visits (Rashid et al., 2023) and re-hospitalization among persons with PD (Shahgholi et al., 2017), presumably when caregivers are overwhelmed and require additional support. In the general population, 30-day readmission is also associated with age, male sex, comorbidities, polypharmacy, and length of stay at the initial hospital visit (Glans et al., 2020). Because ICU admission and 30-day readmission among persons with PD are significantly more common in the last 6 months of life, these hospital outcomes may signify a transition to the EoL period and aid in prognostication. Interestingly, the principal admission diagnosis was similar between hospitalized decedents and hospitalized non-decedents and may not be a reliable indicator of the EoL period.

Numerous studies have advocated for improved EoL care in the United States (Teno et al., 2013; Committee on Approaching Death: Addressing Key End of Life Issues, Institute of Medicine, 2015). Although rates of EoL hospitalization declined among United States Medicare beneficiaries between 2000 and 2015 (Emanuel, 2018; Teno et al., 2018), progress is slow and EoL interventions are needed for those at greatest risk of death. Despite the prognostic uncertainty associated with heterogeneous symptoms and rates of disease progression in persons with PD, our findings have important clinical implications. Because PD patients hospitalized in their last 6 months of life have twice the odds of ICU admission and 30-day readmission, these events could trigger automatic inpatient palliative care consultations to clarify goals of care. While advance care planning can reduce unwanted, invasive, and potentially inappropriate EoL care (Kwak et al., 2014), most discussions occur in the outpatient setting and may miss critical opportunities to intervene after hospital admission. Thus, inpatient admissions can offer time-sensitive opportunities to discuss EoL preferences and prompt earlier referral to hospice care when indicated and desired.

This study has several limitations. First, Medicare is an administrative dataset, and studies using Medicare claims data are limited to the diagnoses and treatments documented in the medical record and subject to misclassification bias. Second, although comorbid dementia was determined using a validated indicator variable for Alzheimer’s disease and related disorders (Chronic Conditions Data Warehouse, 2023), the exact dementia diagnosis was unknown. Similarly, the cause of death among hospitalized decedents in our cohort was unknown, and hospitalizations among persons with PD may stem from other illnesses. However, sepsis, aspiration pneumonitis, UTI, acute kidney failure, and PD comprised the top 31% of inpatient principal diagnoses among hospitalized decedents. By contrast, these conditions only comprised the top 20% of inpatient principal diagnoses among hospitalized non-decedents (Supplementary Table S1), suggesting that hospitalizations in the EoL period are more often associated with PD-related system failure. Because UTIs and pneumonia are leading causes of acute hospitalization in persons with advanced PD (Okunoye et al., 2020), further work is needed to minimize these risk factors. Future studies should also explore inpatient outcomes at EoL using more granular clinical data. In addition, we could not account for a do-not-resuscitate (DNR) order or pre-existing advance directives that may have influenced resource utilization and treatment intensity. While hospitalized patients with PD have higher odds of DNR utilization than other hospitalized patients (Mahajan et al., 2017), hospitalized decedents in our cohort still had high rates of treatment intensity and readmission, suggesting that new approaches to advance care planning may be needed in acute care settings. Lastly, our study presents data from a large, nationally representative sample of persons with PD who were hospitalized before the COVID-19 pandemic. Future studies should also utilize more recent data, when available, to account for post-pandemic changes in healthcare delivery, acute care utilization, and life expectancy (Balser et al., 2021; Pujolar et al., 2022; Schöley et al., 2022; Huang et al., 2023). Despite these limitations, this study is among the first to describe differences in inpatient outcomes across the PD lifespan.

In conclusion, intensive and frequent hospitalizations among persons with PD are a unique characteristic of the EoL period. Although hospitalizations may stem from non-PD-related illnesses, inpatients with PD are twice as likely to experience an ICU admission and hospital readmission in the last 6 months of life than at other time points. Because hospitalizations among persons with PD may portend a poor prognosis and reflect poor EoL care quality, providers must ensure that all hospitalizations are reasonable, necessary, and aligned with care preferences. In addition, because caregiver burden is a risk factor for re-hospitalization among persons with PD, inpatient admissions may provide important opportunities to address caregiver strain. Further work is needed to reduce hospitalizations in the EoL period, which can lead to improved quality of life, more efficient resource allocation, and reduced healthcare spending for all persons with PD.

The data analyzed in this study is subject to the following licenses/restrictions: all datasets are available to purchase at resdac.org/. Aggregated de-identified data may be shared on request. Requests to access these datasets should be directed to https://resdac.org/.

The studies involving humans were approved by the University of Pennsylvania Human Research Protections Office and the Centers for Medicare and Medicaid Services via a Data Use Agreement and waiver of consent. The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent for participation was not required from the participants or the participants’ legal guardians/next of kin in accordance with the national legislation and institutional requirements.

WA: Conceptualization, Data curation, Formal analysis, Methodology, Writing – original draft, Writing – review & editing. ND: Conceptualization, Formal analysis, Methodology, Writing – review & editing. WB: Data curation, Methodology, Writing – review & editing. JF: Conceptualization, Methodology, Supervision, Writing – review & editing. AW: Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Writing – review & editing.