Palliative Care and Mental Health among Pancreatic Cancer Patients in the United States: An Examination of Service Utilization and Health Outcomes

Introduction: Palliative care (PC) utilization remains low among pancreatic cancer patients. This study explores the association of PC with mental health service and pharmacotherapy utilization among pancreatic cancer patients.Methods: Retrospective analysis was conducted on a sample of patients in the United States with newly diagnosed pancreatic cancer using Electronic Health Record data from Optum’s Integrated Claims-Clinical data set. Subsequent diagnoses of anxiety and depression and PC consultation encounters were determined using ICD-9/10 codes. Adjusted associations of mental health treatments with PC and patient characteristics were quantified using multiple logistic regression. Results: Among newly diagnosed pancreatic cancer patients (n = 4029), those with PC consultations exhibited a higher prevalence of anxiety (33.9% vs. 22.8%) and depression (36.2% vs. 23.2%). Mental health service use and pharmacotherapy varied, with the highest utilization among patients having both anxiety and depression. Treatment pattern was also influenced by age (aOR 1.832 for age <55 vs. 65–70 years). Notably, PC consultations showed no significant effect on the likelihood of documented treatment. Discussion: Our study emphasizes underutilization of PC and MH treatment for pancreatic cancer patients. These findings imply a crucial need for further investigation into palliative care’s role in addressing mental health concerns among pancreatic cancer patients.


Introduction
Cancer is a complicated health condition with a wide range of symptoms and treatments that can have tremendous effects on a patient's mental health and overall wellbeing [1].For that reason, the management of cancer and associated symptoms presents a substantial challenge for oncology healthcare providers [1].In recent years, palliative care has emerged as a promising approach to address these challenges [1][2][3] as it has shown positive outcomes in enhancing the quality of life for various types of cancer patients [2,[4][5][6][7][8][9].Despite the established benefits of palliative care for cancer patients, its adoption and utilization among healthcare providers have been relatively low [5,10], leading to potentially harmful consequences for patients with cancer due to a lack of palliative care-associated health interventions [11].Numerous factors influence treatment and symptom management, including type of cancer [3,7,[12][13][14]; however, the impact of cancer type on anxiety and depression remains a topic of debate in the medical community [10,12,15].
Existing literature has highlighted that the presence of anxiety and depression can exacerbate physical symptoms, such as pain, fatigue, weakness, and functional status [1,[13][14][15][16], among patients with advanced cancer, therefore, diminishing their overall quality of life [13].
These mental health issues also affect end-of-life care decisions, potentially leading to more aggressive treatments, such as chemotherapy, intensive care unit admission, and ventilation [3].Additionally, anxiety and depression can further exacerbate other symptoms, creating a compounding effect [1,4,[13][14][15].The complexity associated with comorbid cancer and anxiety/depression underscores the need for a holistic, interdisciplinary approach to supportive care for cancer patients.Despite the well-documented benefits of palliative care in cancer treatment [2,4,5,[7][8][9][10]17], there remains a lack of research on the intersection of palliative care and mental health in the context of supportive cancer care.
Pancreatic cancer is medically complex, often diagnosed at later stages [8,12,15], and documents a higher prevalence of severe anxiety and depression symptoms among its patients [3,4,[12][13][14], who were less likely to improve after palliative care consultation [7] compared to patients with other types of cancer.This demonstrates that the type of cancer can profoundly influence both symptom burden and palliative outcome.Patients with pancreatic cancer form a vulnerable cohort that necessitates further investigation into the effects of palliative care and mental health.To address this existing gap in research, our study aims to explore the impact of palliative care consultations on the utilization of mental health services and pharmacotherapy among pancreatic cancer patients diagnosed with anxiety and depression within 12 months of their initial cancer diagnosis.

Data Source and Study Participants
This study utilized Electronic Health Record (EHR) data from Optum's de-identified Integrated Claims-Clinical data set, including healthcare encounters between 1 January 2010 and 31 December 2018.The Integrated data set is an all-payer data set, i.e., data was captured from records of patients who were commercially, publicly, or uninsured, including Medicare and Medicaid beneficiaries.The study examined adults aged 18 years and older, in the United States, with newly diagnosed pancreatic cancer within the studied period.To identify newly diagnosed pancreatic cancer, all the included patients were required to have at least 6 months of activity in the EHR data prior to the date of first pancreatic cancer diagnosis.Among the patients with newly diagnosed pancreatic cancer, the diagnosis of anxiety, depression, or both was determined, and the number of days from the date of first pancreatic cancer diagnosis to the dates of first diagnosis of these mental health issues.Anxiety and depression diagnoses were identified using ICD-9-CM and ICD-10-CM codes (Appendix A. Table A1).Patients who had anxiety or depression diagnosed prior to their diagnosis of pancreatic cancer were excluded from further analysis.

Ethical Considerations
The Saint Louis University Institutional Review Board deemed this to be Non-Human Subjects Research.

Measures
The primary outcome of interest was utilization of mental health services and pharmacotherapy for anxiety and/or depression.Mental health services were identified by examining the encounters of healthcare services provided by the following specialties: Clinical Neuropsychology, Marriage and Family Therapy, Psychiatric Hospital, Psychiatry, Psychiatry and Neurology, Psychoanalysis, Psychology, and Clinical Social Work.In addition, encounters with mental health services were identified by CPT and HCPCS codes (Appendix A. Table A2).Utilization of pharmacotherapy was determined by examining prescriptions of medications for anxiety and depression within 3 years following the diagnosis of pancreatic cancer (Appendix A. Tables A3 and A4).The generic and brand names and drug classes were consistent with previous studies and informed by an academic psychiatrist collaborator [18].The absence of either a mental health services visit or pharmacotherapy was considered no treatment.As treatment for anxiety and depression overlaps, no attempt was made to determine the diagnosis that indicated the treatment (particularly in the case of those with both depression and anxiety diagnoses), but rather whether treatment was documented.
The exposure of interest was a palliative care (PC) consultation, determined by encounters with PC services in the EHR records (ICD-9: V66.7, ICD-10: Z51.5).The V66.7 code has been validated in multiple studies on PC for patients with chronic life-threatening illnesses in the U.S (United States), which find that the code is highly specific (>90%) and moderately sensitive (45-89%) [19].The Z51.5 code needs further validation but is equivalent to V66.7 [20].The covariates examined at the time of cancer diagnosis included the patients' age, gender, race, primary payer, and the Charlson-Deyo Comorbidity Index [21,22].

Statistical Analysis
Chi-Square tests were used to assess the distribution of categorical demographics among groups, as well as the prevalence of mental health service utilization, pharmacotherapy, or any mental health treatment among those with and without palliative care consultation.Continuous demographic variables (i.e., age) by exposure were examined with Student t-tests.A multiple regression analysis was conducted to quantify the effects of palliative care consultation, gender, age at time of cancer diagnosis, race, and the Charlson comorbidity index on the use of mental health services and pharmacotherapy.Statistical analyses were performed in SAS 9.4 (SAS Institute, Cary, NC, USA).All tests were two tailed with α = 0.05 to determine statistical significance.

Results
In this study, we identified 4029 adult patients who had newly diagnosed pancreatic cancer out of 5266 patients with any record of pancreatic cancer diagnosis from 2010 to 2018.Within this sample, patients were mostly Caucasian (80.5%), while 9.7% were African American, and 9.8% identified as other.The mean age at the time of cancer diagnosis was 68 years old (Standard Deviation: 12.3 years), while the mean age at the time of anxiety or depression diagnosis was 65.2 and 64.7 years old, respectively.Pancreatic cancer patients with and without PC consultation had statistically significant differences in the distribution of race, region, Charlson Comorbidity Index scores, concurrent anxiety, or depression compared to those with PC consultation (Table 1).Notably, a higher proportion of patients with anxiety or depression had a PC consultation, compared to those who were not diagnosed with anxiety or depression (36.2% vs. 23.2%,33.9% vs. 22.8%, respectively, Table 1).Regardless of mental health diagnosis, most patients in the sample resided in the Midwest and identified as Caucasian (Table 2).A higher proportion of patients with a diagnosis of depression, without anxiety, were over 70 years old, compared to patients with a diagnosis of anxiety alone or both depression and anxiety (Table 2).Notably, for patients consulted by PC who had a diagnosis of anxiety alone, there was a higher prevalence of those who identified as African American, compared to those who were not consulted (10.82% vs. 5.08%, Table 2).This relationship is not observed for those diagnosed with depression and the effect was slightly reversed for those diagnosed with both anxiety and depression (Table 2).When examined by PC consultation within mental health group, the demographic trends remained comparable to those seen with pancreatic cancer patients overall, including the significant differences in comorbidity representation.There were significant differences in treatment documented by the mental health diagnosis group.For patients with both anxiety and depression, 64.2% documented pharmacotherapy, compared to 52.7% of those with anxiety alone and 39.7% of patients diagnosed with depression alone (p < 0.0001, Appendix A. Table A2).As for the utilization of mental health services among the different mental health diagnoses, 35.3% of patients diagnosed with both depression and anxiety documented use of mental health services, compared to 21.8% of those with depression alone and 20.6% of patients with anxiety alone (p = 0.0003, Appendix A. Table A2).Regarding the prevalence of treatment among the various mental health diagnoses, stratified by PC consultation, 45.7% of patients diagnosed with both anxiety and depression who were consulted by PC documented the receipt of mental health services, whereas 28.5% of patients with both anxiety and depression but without PC consultation documented mental health services (p = 0.0118).
PC had no significant effect on the odds of receiving documented pharmacotherapy, mental health services, or any treatment.Patients diagnosed with anxiety alone or depression alone had lower odds of documented pharmacotherapy (aOR = 0.624 (0.429, 0.907), aOR = 0.365 (0.247, 0.540), respectively), mental health services (aOR = 0.411 (0.266, 0.637), aOR = 0.499 (0.320, 0.779), respectively), and any treatment (aOR = 0.551 (0.371, 0.819), aOR = 0.376; (0.251, 0.564), respectively), compared to patients diagnosed with both anxiety and depression (Table 3).Compared to patients 55-64 years old, patients less than 55 years old had around two times higher odds of documented treatment of any modality and five times higher odds of documented mental health services.A similar trend was seen for patients 55 to 64 years old, who had three times higher odds of documented mental health services than patients 65 to 70 years old.There was no observable association for patients over 70.Patients scoring over 4 on the Charlson Index had higher odds of documented pharmacotherapy (aOR = 1.975 (1.223, 3.188), aOR = 1.583 (1.011, 2.479), respectively, Table 3), and those scoring 4 to 8 had higher odds of any treatment being documented (aOR = 1.866 (1.147, 3.035), Table 3) than those scoring 0 to 3.

Discussion
Pancreatic cancer ranks as the seventh leading cause of cancer-related deaths worldwide [23].Due to the advanced stage at the time of diagnosis and the severity of symptoms, palliative care is of utmost importance for pancreatic cancer patients, as it addresses the physical, emotional, and psychosocial needs of patients.Despite known benefits, palliative care utilization remains low among pancreatic cancer patients, consistent with other studies of cancer patients [5,13,24,25].Our study aims to explore the impact of palliative care consultations on the utilization of mental health services and pharmacotherapy.Our study found significant differences in the distribution of race, region, Charlson Comorbidity Index scores, and concurrent anxiety or depression between pancreatic cancer patients with and without palliative care consultation.Patients consulted for palliative care had a higher proportion of comorbidities and a higher prevalence of anxiety or depression, suggesting that patients with more complex medical and psychological needs are more likely to be referred to palliative care services [26].
Patients with a diagnosis of new-onset depression, without anxiety, were more likely to be over 70 years old, while patients with a diagnosis of new-onset anxiety or both depression and anxiety were younger (Appendix A. Table A5).Moreover, these age-related patterns in mental health diagnoses can also inform further research into the underlying causes and risk factors for these conditions in different age groups.The prevalence of mental health treatment varied depending on the mental health diagnosis.Patients with anxiety alone had a higher prevalence of pharmacotherapy compared to patients with depression alone, whereas patients with anxiety and depression had the highest prevalence of mental health treatment, for both pharmacotherapy and mental health services.
In the multiple regression analysis, we found that patients diagnosed with anxiety alone or depression alone had lower odds of receiving documented pharmacotherapy, mental health services, or any treatment compared to patients diagnosed with both anxiety and depression.These findings suggest that patients with comorbid anxiety and depression may have more severe symptoms and therefore be more likely to receive treatment [27,28].Studies suggest that depression present in palliative care patients often goes untreated, given that pancreatic cancer patients have a poor prognosis [29].With a poor prognosis, determining the right dose and appropriate drug for pancreatic cancer patients can pose challenges with limited time.Moreover, we found that younger patients had higher odds of receiving treatment, with patients less than 55 years old having around two times higher odds of documented treatment compared to patients 65 to 70 years old, consistent with existing literature surrounding mental healthcare delivery to younger and older patients [30].
The absence of a significant association between palliative care and the likelihood of any documented treatment may be attributed to several factors.It is possible that palliative care consultations primarily addressed physical symptoms rather than focusing on mental health symptoms.A study by Hatano et al. in 2018 found that pain management was the primary reason for PC deferral, and depression was a secondary, yet less common reason.Furthermore, treatment was often withheld because of very late referrals [31].Additionally, patients might not have prioritized mental health as a primary aspect of their care during these PC consultations.
Despite these important findings, there are several limitations to be noted.We could not determine the staging of pancreatic cancer given the nature of the administrative data source utilized in this study.Moreover, the specific focus on anxiety and depression, while important, might exclude other pertinent mental health conditions, such as acute stress reactions following cancer diagnosis, post-traumatic stress disorder, or substance use disorders, therefore reducing the comprehensive assessment of mental health in this context.Lastly, the actual receipt of pharmacotherapy and mental health is challenging to ascertain in the EHR database, leaving some possibility of misclassification.
Regardless of these limitations, this study has several notable strengths.First, we utilized a comprehensive approach to ICD-9/10 codes, which have been demonstrated in prior work to be highly sensitive [19].The data source used for the study has both commercially and Medicare-insured patients from diverse regions across the United States and depicts a high degree of representativeness for our study, thereby rendering our findings generalizable to the broader population.These findings also provide invaluable insights into the demographics and mental health attributes of this patient group, offering healthcare professionals a deeper understanding to develop tailored interventions for pancreatic cancer patients.
The overall comprehensive approach, exploring sociodemographic factors, like race, age, gender, region, comorbidity index scores, and mental health diagnoses, yields a depiction of demographics and mental health profiles of pancreatic cancer patients.Notably, this study underscores the importance of addressing both anxiety and depression together within the scope of pancreatic cancer treatment, emphasizing the fundamental necessity for a more comprehensive and integrated approach to patient-centered care.

Conclusions
Our study highlights the need for further investigation of palliative care in relation to mental health treatments and services utilization for pancreatic cancer patients.Fur-  Abbreviations: selective-serotonin reuptake inhibitor (SSRI), serotonin and norepinephrine reuptake inhibitor (SNRI).Abbreviations: selective-serotonin reuptake inhibitor (SSRI), serotonin and norepinephrine reuptake inhibitor (SNRI), tricyclic antidepressant (TCA), monoamine-oxidase inhibitor (MAO-i).

Table 1 .
Sample Characteristics Stratified by Receipt of Palliative Care Consultation.

Table 2 .
Sample Characteristics, Grouped by Mental Health Diagnoses and Stratified by Receipt of Palliative Care Consultation.

Table 3 .
Results of Multiple Logistic Regression for Mental Health Service Use, Pharmacotherapy, and Any Treatment.

Table A2 .
CPT Codes Used to Measure Utilization of Mental Health Services.

Table A2 .
Cont. of results of psychiatric, other medical examinations and procedures, or other accumulated data to family or other responsible persons, or advising them how to assist patient 90889 Preparation of report of patient's psychiatric status, history, treatment, or progress for other individuals, agencies, or insurance carriers 90899 Unlisted psychiatric service or procedure * Code discontinued.