Impact of Psychiatric Illness on Decreased Survival in Elderly Patients with Bladder Cancer in the United States

Abstract

Objective

Treatments for muscle-invasive bladder cancer are multimodal, complex and often carry significant physical and psychological morbidity risks. We sought to define the incidence and types of psychiatric illnesses diagnosed following treatment and determine the impact on survival outcomes.

Methods

A total of 3,709 patients diagnosed with clinical stage T2-T4a bladder cancer from January 1, 2002 to December 31, 2011 from the Surveillance, Epidemiology, and End Results (SEER)-Medicare were analyzed. We used multivariable analysis and cox proportional hazards models to determine predictors associated with psychiatric diagnosis and impact on survival outcomes.

Results

Of the 3,709 patients, 1,870 (50.4%) were diagnosed with post-treatment psychiatric disorders. Patients who underwent radical cystectomy were found to be at a significantly greater risk of having a post-treatment psychiatric illness in comparison to patients who underwent radiotherapy and/or chemotherapy (Hazard Ratio (HR) 1.19, 95% CI = 1.07 - 1.31, P = 0.001). In adjusted analyses, diagnosis of a psychiatric disorder resulted in significantly worse overall (HR 2.80, 95% CI 2.47 - 3.17, P < 0.001) and cancer-specific (HR 2.39, 95% CI 2.05 - 2.78, P < 0.001) survival, respectively.

Conclusions

Half of muscle-invasive bladder cancer patients who underwent treatment were diagnosed with a psychiatric disorder which resulted in worse survival outcomes as compared to patients without a post-treatment psychiatric diagnosis. This information can be used to inform interventions to educate patients with muscle-invasive bladder cancer regarding the impact of different treatments on mental health.

Introduction

There will be an estimated 79,030 new cases and 16,870 deaths from bladder cancer in the United States in 2017¹. Despite the multimodality treatment paradigm for muscle-invasive bladder cancer (MIBC), patients have estimated 5-year survival rates ranging from 26-64%^{2, 3}. Moreover, the morbidity and mortality of associated treatments for MIBC are non-negligible⁴. Taking this into account, survival outcomes have remained unaltered over the last three decades⁵.

Cancer diagnosis and treatment carries significant physical and emotional stress which may affect survival outcomes. Cancer patients are at increased risk for psychiatric disorders from the pre-diagnostic to the post treatment period as approximately one-third of cancer patients have been shown to be at risk of developing a psychiatric disorder^{6, 7}. Disorders, such as depression, cause serious suffering and distress, reduced adherence to cancer care, and increased risk of hospital readmission which in turn may lead to worse survival^7–9. Different cancers present with varying degrees and types of psychiatric disorder risks. Patients diagnosed with lung cancer are more likely to be diagnosed with depression or anxiety disorders than patients with prostate cancer^{10, 11}. Moreover, while studies that investigate post-treatment psychiatric disorders for any type of cancer are lacking one recent meta-analysis identified receipt of high-dose chemotherapy to be associated with an increased risk of cognitive dysfunction among breast cancer patients¹².

Bladder cancer patients have been previously shown to be at higher risk of suicide compared to the general population, with the highest risk of suicide in the first 5 years after diagnosis¹³. In addition, patients undergoing radical cystectomy (RC) for MIBC were found to be at even greater risk for suicide than patients with no intervention¹⁴. Despite this knowledge, there is limited data regarding the incidence and type of psychiatric disorders following treatment for MIBC. Moreover, the impact of having a post-treatment psychiatric disorder on survival outcomes among patients with MIBC remains to be determined. Against this backdrop, we sought to investigate the incidence of post-treatment psychiatric diagnoses following treatment with RC and radiotherapy/chemotherapy (RTX and/or CTX). We hypothesized that having a post-treatment psychiatric diagnosis would be associated with worse survival outcomes as compared to patients without a post-treatment psychiatric diagnosis.

Patients and methods

Data Source

We used the Surveillance, Epidemiology, and End Results (SEER)-Medicare linked data from the National Cancer Institute. SEER collects data on new cancer cases from 18 locations and sources in the United States in the most recent release, which allows the results to be generalized to the US population. Cancer cases reported in the SEER database conformed to the standards of the North American Association of Central Cancer Registries, and case ascertainment in the SEER data was 98% complete¹⁵. The SEER database reports data on patient tumor characteristics (including stage, grade, histology), demographics, and follow-up information. The Medicare database comprises data on inpatient and outpatient claims. The institutional review board of The University of Texas Medical Branch regarded this study exempt from any obligation or liabilities.

Ascertainment of study cohort

Our analysis was restricted to elderly (age ≥66 years) bladder cancer patients diagnosed with stage II, III, and IV transitional cell or urothelial carcinoma from January 1, 2002 to December 31, 2011 for whom claims data were available up to December 31, 2013. Patients who did not receive treatment for bladder cancer may be significantly different than those who received the treatment. Therefore, the study was restricted to patients who received treatment for bladder cancer and did not have a psychiatric diagnosis within 12 months prior to receipt of treatment¹⁶. Only patients who had Medicare Fee-for-Service coverage and for whom Medicare Part A and Part B claims data were available were included in the study. Patients with any other non-bladder cancer diagnosis were excluded. The final cohort consisted of 3,709 patients (Supplementary Figure 1).

Identification of bladder cancer treatments

Patients who underwent RC were identified using procedure codes, including both open and robot-assisted laparoscopic surgery with or without pelvic lymph node dissection. Patients who received radiation were classified based on the Medicare diagnosis and procedure codes consistent with radiotherapeutic procedures. Patients who received chemotherapy were identified using HCPCS codes consistent with chemotherapeutic agents that are commonly used to manage bladder cancer in the absence of a concomitant code for radical cystectomy. The RC group included patients who underwent surgery alone or in combination with radiation or chemotherapy¹⁷. Patients who underwent either chemotherapy alone, radiotherapy alone, or both chemotherapy and radiotherapy were placed in one group, as bladder-sparing protocols for invasive bladder cancer often combine the utilization of chemotherapy and radiation.

Identification of psychiatric illnesses

All patients were followed for 1 year from the bladder cancer treatment date to determine psychiatric illnesses. Patients who received RC were followed after they were discharged from the hospital. Psychiatric illnesses were identified using ICD-9 codes, excluding developmental psychiatric diagnoses and organic mental disorders (Supplementary Table 1). Patients having any ICD-9 code for psychiatric illness following treatment start date until 1 year were classified as having a post-treatment psychiatric illness. Each treatment group (RC, RTX and/or CTX) was stratified to two groups: those with post-treatment psychiatric diagnosis and those without post-treatment psychiatric diagnosis. Patients who had a post-treatment psychiatric diagnosis were further stratified to discern the incidence of each psychiatric diagnosis type that we identified. Psychiatric diagnosis types were obtained from the International Statistical Classification of Diseases and Related Health Problems – WHO Version for 2016¹⁸. Supplementary Table 1 includes the list of all the specific psychiatric diagnoses that fall under each psychiatric diagnosis type we identified in this study.

Survival Outcome

Patients were followed for one year to determine all-cause survival and cancer-specific survival. Cancer-specific survival was defined as patients who died due to cancer. Overall and cancer-specific survival was first determined for all patients with a psychiatric disorder(s) and compared to patients without a post-treatment psychiatric diagnosis. Patients with post-treatment psychiatric disorders were then stratified based on the number of psychiatric diagnoses they received during the post-treatment period (1-year) and the overall and cancer-specific survivals were then investigated. The SEER cancer registry uses algorithms to process causes of death from death certificates in order to identify a single, disease-specific, underlying cause of death. In order to accurately classify cause of death, SEER takes into account causes of deaths in conjunction with tumor sequence (i.e., only one tumor or the first of subsequent tumors), site of the original cancer diagnosis, and comorbidities (e.g., AIDS and/or site-related diseases).

Study covariates

Using the SEER-Medicare database we extracted information on patient age, sex, race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, and non-Hispanic other races), marital status (single, married, and unknown), and SEER region (Northeast, South, Midwest, and West). We determined socioeconomic characteristics for patients from SEER-Medicare. County level education was defined according to percentage of residents with at least 4 years of college education. County-level median household income was acquired through linkage to the area health resource file and then divided into quartiles. Comorbidity was assessed using the Klabunde modification of the Charlson index during the year before cancer diagnosis¹⁹. The Klabunde modification uses comorbid conditions identified by the Charlson comorbidity index and incorporates the diagnostic and procedure data contained in hospital claims and physician/outpatient claims. Diagnosing providers were identified via the AMA Physician Masterfile linked with outpatient claims.

Statistical analysis

We conducted univariate analyses to assess the association between bladder cancer treatments (RC, RTX and/or CTX) and the covariates described above using the Pearson χ² test. Unadjusted cumulative incidence curves were generated to compare the incidence of post-treatment psychiatric disorders for each treatment modality. To investigate factors related to post-treatment psychiatric diagnosis among patients newly diagnosed with stage II, III, or IV bladder cancer, we performed Fine and Gray’s competing risk hazards regression model, while treating death as a competing risk. The model included following independent variables: bladder cancer treatment, year of diagnosis, age, sex, race/ethnicity, marital status, income, education, county population, census region, disease grade, and comorbidity.

To determine the association of psychiatric illnesses with all-cause survival, time-dependent Cox proportional hazards regression was used, and Fine and Gray’s competing risk hazards regression model was used for cancer-specific survival; deaths due to non-cancer causes were treated as a competing risk. Psychiatric illness was assessed using cumulative effects. Psychiatric illness was modeled as a time-varying variable in both regression models, and all other covariates such as patient demographics and clinical characteristics were modelled as fixed variables. Modeling of psychiatric illness as a time varying variable in time-dependent Cox regression models is a better alternative to traditional landmark analysis approach²⁰.

Statistical significance was defined as p < 0.05. All statistical tests were two-sided, and all analyses were done using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Results

Patient demographics according to treatment type are summarized in Table 1. Of the patients who underwent treatment, 1,861 (50.2%) underwent RC and 1,848 (49.8%) underwent RTX and/or CTX. Patients who underwent surgery were younger, married, more educated, had fewer comorbidities, and had less advanced stage disease than those who underwent RTX and/or CTX. Of those patients who underwent RC, 225 (12.1%) and 374 (20.1%) received neoadjuvant and adjuvant chemotherapy, respectively (P <0.001).

Table 1.

Patient demographic and clinical characteristics

		Treatments
		RC		RTX and/or CTX
Characteristic	Total	N	%	N	%	P-Value
	3709	1861	50.2	1848	49.8
Age group						<0.001
66-69	584	379	20.4	205	11.1
70-74	858	512	27.5	346	18.7
75-79	950	528	28.4	422	22.8
≥80 yr.	1316	441	23.7	875	47.4
Sex						<0.001
Male	2487	1186	63.7	1301	70.4
Female	1222	675	36.3	547	29.6
Race/ethnicity						0.021
Non-Hispanic White	3247	1641	88.2	1606	86.9
Non-Hispanic Black	195	78	4.2	117	6.3
Hispanic	124	63	3.4	61	3.3
Other	143	79	4.3	64	3.5
Marital Status						<0.001
Single	486	239	12.8	247	13.4
Married	2227	1179	63.4	1048	56.7
Unknown	996	443	23.8	553	29.9
CTX Status						<0.001
None	2562	1262	67.8	–	–
Neoadjuvant	606	225	12.1	–	–
Adjuvant	541	374	20.1
Census Region						0.140
West	1453	761	40.9	692	37.5
Northeast	903	437	23.5	466	25.2
Midwest	477	226	12.1	251	13.6
South	876	437	23.5	439	23.8
Median household income, $						0.061
≤42992	854	397	21.3	457	24.7
42993 -56188	923	468	25.2	455	24.6
56189 -73827	948	477	25.6	471	25.5
≥73828	984	519	27.9	465	25.2
County-level education, % with at least 4 years of college						0.009
≤20.58%	960	507	27.2	453	24.5
20.59-27.36%	930	476	25.6	454	24.6
27.37-34.83%	917	417	22.4	500	27.1
≥34.84%	902	461	24.8	441	23.9
Comorbidity						<0.001
0	2093	1145	61.5	948	51.3
1	948	450	24.2	498	27.0
2	367	162	8.7	205	11.1
3+	301	104	5.6	197	10.7
Stage						<0.001
II	1789	725	39.0	1064	57.6
III	801	568	30.5	233	12.6
IV	1119	568	30.5	551	29.8
Grade						<0.001
Low	165	73	3.9	92	5.0
High	3374	1746	93.8	1628	88.1
Unknown	170	42	2.3	128	6.9
Year of Diagnosis						0.021
2002	395	209	11.2	186	10.1
2003	370	216	11.6	154	8.3
2004	414	210	11.3	204	11.0
2005	414	215	11.6	199	10.8
2006	408	192	10.3	216	11.7
2007	363	179	9.6	184	10.0
2008	348	178	9.6	170	9.2
2009	332	153	8.2	179	9.7
2010	319	146	7.9	173	9.4
2011	346	163	8.8	183	9.9

CTX: chemotherapy; RC: radical cystectomy; RTX: radiotherapy

Overall 998 (53.6%) RC patients developed post-treatment psychiatric disorder(s) compared to 872 (47.2%) RTX and/or CTX patients (P <0.001) (Table 2). Among patients who underwent RC, the most common psychiatric disorder types were mental and behavioral disorders due to psychoactive substance use (19.0%), depressive episode (13.5%), and symptoms and signs involving cognitive functions and awareness (13.1%). In comparison, the most common psychiatric disorder types in RTX and/or CTX in order of incidence were mental and behavioral disorders due to psychoactive substance use (13.7%), symptoms and signs involving cognitive functions and awareness (13.5%), dizziness and giddiness (11.9%), followed by depression (9.2%), respectively. Out of the 1,132 (30.5%) patients with identifying provider information, most common type of diagnosing provider included internal medicine (18.1%), family medicine (8.5%), emergency medicine (8.0%), urology (6.8%), psychiatry (3.9%) and hematology/oncology (3.5%).

Table 2.

Psychiatric Diagnoses According to Treatment Type

	Patients with Post-Treatment Psychiatric Diagnosis, n (%)
	RC		RTX and/or CTX
Psychiatric Diagnosis Type	N	%	N	%	P - Value
Any Psychiatric Diagnosis	998	53.6	872	47.2	<0.001
Depressive Episode	251	13.5	169	9.2	<0.001
Recurrent Depressive Episode	23	1.2	18	1.0	0.446
Persistent Mood (Affective) Disorders	39	2.1	28	1.5	0.184
Anxiety Disorders	177	9.5	127	6.9	0.003
Reaction to Severe Stress, and Adjustment Disorders	88	4.7	49	2.7	<0.001
Mental and Behavioral Disorders due to Psychoactive Substance Use	354	19.0	254	13.7	<0.001
Symptoms and Signs Involving Cognitive Functions and Awareness	244	13.1	250	13.5	0.709
Dizziness and Giddiness	221	11.9	219	11.9	0.982
Symptoms and Signs Involving Emotional State	134	7.2	146	7.9	0.420
Delirium, Not Induced by Alcohol and Other Psychoactive Substances	164	8.8	160	8.7	0.868
Psychotic Disorders	91	4.9	95	5.1	0.726
Other*	61	3.3	53	2.9	0.470

CTX: chemotherapy; RC: radical cystectomy; RTX: radiotherapy

^*Cells <11 cannot be reported due to confidentiality concerns using SEER-Medicare dataset. Due to cells<11, rates of suicide/suicidal and other categories were combined

Multivariable results from the competing risks regression model discerning predictors for psychiatric diagnosis are shown in Table 3. Patients who underwent surgery were significantly more likely to be diagnosed with a post-treatment psychiatric disorder in comparison to patients who underwent RTX and/or CTX (Hazard Ratio (HR) 1.19, 95% CI 1.07 – 1.31, P = 0.001) (Figure 1). The median time to psychiatric diagnosis for RC was 270 days, however, median time was not reached for those who underwent RTX and/or CTX. Overall, patients who underwent adjuvant chemotherapy were found to be at less risk of developing a psychiatric disorder in comparison to patients without chemotherapy intervention (HR 0.87, 95% CI 0.76 - 0.99, P = 0.036) (Table 3). Interaction of psychiatric diagnosis with treatment type was not significant. Patients who were younger, single and had advanced clinical stage disease were at a greater risk of being diagnosed with a post-treatment psychiatric disorder (all P < 0.05). More recent year of treatment was associated with greater likelihood of being diagnosed with a psychiatric disorder (2011 vs. 2002, HR 1.59, 95% CI 1.28 - 1.98, P < 0.001). Patients who underwent RC were significantly more likely to be diagnosed with depression than those who underwent RTX and/or CTX (HR 1.43, 95% CI 1.15 - 1.77, P = 0.001) (Supplementary Table 2).

Table 3.

Multivariate Results of the Competing Risk Regression Model of Predictors for Any Post-Treatment Psychiatric Diagnosis

Characteristic	Competing Risks Regression
	Time-to-first psychiatric diagnosis
	HR	95% CI		p Value
Treatment
RTX and/or CTX	ref	–	–
RC	1.19	1.07	1.31	0.001
Age Group
66-69	ref	–	–
70-74	0.87	0.75	1.01	0.070
75-79	0.93	0.81	1.07	0.324
80+	0.80	0.69	0.92	0.002
Sex
Male	ref	–	–
Female	0.98	0.88	1.10	0.772
Race/ethnicity
White	ref	–	–
Black	1.16	0.95	1.42	0.143
Hispanics	0.80	0.61	1.05	0.111
Other	0.95	0.75	1.20	0.662
CTX Status
None	ref	–	–
Neoadjuvant	0.89	0.79	1.01	0.073
Adjuvant	0.87	0.76	0.99	0.036
Marital Status
Single	ref	–	–
Married	0.69	0.60	0.79	<.0001
Unknown	0.83	0.71	0.96	0.015
Census Region
West	ref	–	–
Northeast	1.04	0.90	1.21	0.558
Midwest	0.98	0.83	1.15	0.761
South	0.94	0.81	1.08	0.362
Stage
II	ref	–	–
III	1.16	1.03	1.30	0.018
IV	1.14	1.02	1.28	0.021
Grade
Low	ref	–	–
High	0.87	0.70	1.09	0.228
Unknown	0.85	0.62	1.16	0.294
Comorbidity Score
0	ref	–	–
1	1.05	0.94	1.17	0.354
2	1.21	1.04	1.41	0.015
3+	0.96	0.80	1.14	0.614
Median household income, $
≤42992	ref	–	–
42993 -56188	0.87	0.76	1.01	0.061
56189 -73827	1.01	0.87	1.17	0.937
≥73828	0.97	0.81	1.17	0.773
County-level education, % with at least 4 years of college
≤20.58%	ref	–	–
20.59-27.36%	1.11	0.97	1.28	0.142
27.37-34.83%	1.03	0.88	1.22	0.687
≥34.84%	1.17	0.98	1.40	0.092
Year of Diagnosis
2002	ref	–	–
2003	1.30	1.05	1.62	0.017
2004	1.37	1.11	1.70	0.004
2005	1.42	1.15	1.76	0.001
2006	1.41	1.14	1.75	0.002
2007	1.27	1.03	1.58	0.027
2008	1.67	1.35	2.07	<.0001
2009	1.69	1.35	2.10	<.0001
2010	1.57	1.26	1.95	<.0001
2011	1.59	1.28	1.98	<.0001

CTX: chemotherapy; RC: radical cystectomy; RTX: radiotherapy

Figure 1.

Unadjusted Cumulative Incidence of Psychiatric Diagnosis According to Treatment Type (P < 0.001).

Adjusted analyses for overall and cancer-specific survival were performed (Table 4). Diagnosis of a psychiatric disorder resulted in significantly worse overall (HR 2.80, 95% CI 2.47 - 3.17, P < 0.001) and cancer-specific (HR 2.39, 95% CI 2.05 - 2.78, P < 0.001) survival, respectively (Figure 1). Increasing number of psychiatric diagnoses was associated with worse overall (3+ vs. none, HR 2.21, 95% CI 1.85 – 2.66, P < 0.001) and cancer-specific (3+ vs. none, HR 2.14, 95% CI 1.73 – 2.65, P < 0.001) survival, respectively. Median time to overall or cancer-specific survival according to psychiatric diagnosis was not reached.

Table 4.

Adjusted Hazard Ratios for Overall and Cancer-Specific Survivals.

		Overall Survival				Cancer-Specific Survival
Treatment	N	HR	95% CI		P-value	HR	95% CI		P-value
Psychiatric diagnosis*
No	1839	ref	–	–		ref	–	–
1	1085	1.29	1.11	1.49	0.001	1.29	1.08	1.53	0.004
2	431	1.97	1.65	2.35	<0.001	1.92	1.57	2.36	<0.001
3+	354	2.21	1.85	2.66	<0.001	2.14	1.73	2.65	<0.001

Adjusted for treatment, age, sex, race/ethnicity, marital status, census, stage, grade, comorbidity, income, education and year of diagnosis.

^*Psychiatric diagnosis was modelled as time-varying variable. Interaction of psychiatric diagnosis with treatment type was not significant.

Discussion

The National Comprehensive Cancer Network Survivorship Guidelines recommend the assessment of late psychological and physical effects, as well as intervention for consequences of cancer and treatment (including psychological distress, financial and social concerns)²¹. The American Society of Clinical Oncology recommends routine screening, follow-up, and re-assessment of psychological distress to improve cancer patients’ quality of life and potentially improve survival²². A meta-analysis found psychosocial intervention may alleviate pain in cancer patients²³ and others found minimal resources are needed to improve quality of life through psychological interventions in cancer patients²⁴. Communication between caregivers and patients is imperative to improve the efficiency of screening and timely intervention for psychosocial distress in cancer patients. Bladder cancer treatment is multimodal, complex and carries significant risk of morbidity. In the present study, the diagnosis and treatment of bladder cancer carries a significant psychological burden with up to half of patients having a post-treatment psychiatric diagnosis. Moreover, patients who developed psychiatric disorders after receiving any type of treatment were found to have worse overall and cancer-specific survival.

Our study has several important findings. First, we observed that half of patients who underwent any type of treatment had a post-treatment psychiatric disorder. While prior research illustrated an increase in suicidal risk for patients undergoing RC versus patients who did not undergo any treatment¹⁴, no study thus far has assessed the incidence of the full spectrum of psychiatric disorders following treatment for MIBC. The incidence of developing post-treatment psychiatric disorders was significantly greater for patients who underwent surgery than RTX and/or CTX. An investigation into the types of psychiatric disorders revealed that there is a tendency for patients to develop a mental and behavioral disorder due to psychoactive substance use^{25, 26}. Patients who underwent surgery showed a markedly greater risk in the development of depressive disorders and were at increased risk to first depressive episode than patients who underwent RTX and/or CTX. Prior studies in other cancers have noted similar increased risk of depression following cancer treatment^{16, 27}. This increased risk of depression following surgery may be attributed to not only the physical and mental demands of undergoing surgical treatment but also the lack of proper screening and counseling prior to depression onset²⁷. Moreover, given the need to have a urinary diversion (orthotopic, continent or incontinent (i.e. ileal conduit)) may further increase the risk of psychiatric disorders than other bladder-sparing treatment options. Interestingly, RC patients who received adjuvant (and not neoadjuvant) chemotherapy were less likely to be diagnosed with a post-treatment psychiatric disorder. The impact of chemotherapy on cognitive function is well documented with some studies illustrating a dose effect with higher doses reported to have a greater impact on mental health in other malignancies^{28, 29}. However, the timing of chemotherapy use and mental illness among cancer patients remains to be determined. Given the limited numbers of RC patients who received chemotherapy in the present study, our findings warrant further investigation on the impact of type, dose, duration and timing of chemotherapy among these patients.

Second, we identified several predictors for a post-treatment psychiatric disorder. Older, married patients were less likely to be diagnosed with a post-treatment psychiatric disorder. Given the morbidity and mortality associated with MIBC despite treatment (approximately 50% at 5 years) one may argue the relatively young age and likelihood of being diagnosed with a post-treatment psychiatric disorder may be attributed to inherent selection bias. Marital status has been previously associated with survival outcomes in patients undergoing cancer-directed surgery³⁰, with worse overall and cancer-specific survival outcomes among single, divorced or widowed patients^{30, 31}. Rationale for these findings include improved social network among married patients with additional reinforcement of health and improved quality of life³⁰. Furthermore, as supported by the present study, married patients are less likely to be depressed which may contribute to improved survival outcomes³². Patient support groups and survivorship programs are an essential component of cancer care and these data support the effectiveness of these programs especially among unmarried patients³¹.

Third, patients with more advanced stage disease were significantly more likely to be diagnosed with a post-treatment psychiatric disorder. Cancer survivors are at significant risk for psychiatric illness after diagnosis and according to treatment type³³. Patients with more aggressive disease features are more likely to undergo multiple treatments which may carry significant physical and emotional burden to the patient¹². Our findings support a recent study which found patients with localized cancer to be at significantly increased risk for mental disorders following diagnosis³⁴. As in other cancer types, bladder cancer patients face a worse prognosis with more advanced stage of disease^{4, 35}. The combined burden of cancer diagnosis, prognosis, with multiple treatment options places these patients at greater risk for mental disorders which adversely impact the health of these patients³⁶. Patients with higher stage and localized disease represent an at-risk cancer population for mental disorders which warrants targeted preventative and intervention programs²¹.

Fourth, we found more recent year of treatment was associated with an increased likelihood of having a post-treatment psychiatric diagnosis. Diagnosis and treatment of mental disorders are well-recognized and have become more commonplace in recent years^{16, 21}. Moreover, implementation of effective mental disorder screening programs including those for depression have likely increased the diagnosis of mental disorders in more recent years^{6, 21}. Nonetheless, increased screening, diagnosis and treatment of mental disorders among cancer patients are needed. As observed in the present study, heterogeneity in specialty of diagnosing provider highlights need for further awareness across all disciplines regarding the impact of psychiatric illness among cancer patients. Failure to address the psychosocial needs of cancer patients can compromise the effectiveness of health care and thereby adversely affect the health and survival of these patients³⁶.

Fifth, we found patients who developed post-treatment psychiatric disorders had significantly worse overall and cancer-specific survivals outcomes. Moreover, we observed a dose-response relationship with increased number of psychiatric diagnoses associated with worse overall and cancer-specific survival. These findings suggest the important interplay between psychiatric illness and survival outcomes following treatment for MIBC^{13, 14}. While oncologic outcomes and perioperative management are critically important to ameliorate the morbidity and mortality associated with surgery, we provide evidence to suggest the importance of addressing the non-oncologic psychosocial needs of these patients as well. Why patients that develop post-treatment psychiatric disorders have poorer survival outcomes is somewhat speculative. Previous studies from the breast cancer literature have reported that patients with mental illness do not receive guideline-consistent treatment and follow-up, resulting in increased risk of cancer-specific mortality compared to patients without mental illness^{37, 38}. If such disparities exist for patients with bladder cancer and post-treatment psychiatric disorders, additional measures to improve care and outcomes for these patients are required. These may include incorporating improved screening methods for psychiatric disorders such as depression, utilization of social workers to assist with maintaining follow-up appointments, and implementation of survivorship clinics dedicated towards assessing the non-oncologic needs of cancer patients.

While our findings are relevant for policy and clinical decision-making, they must be interpreted within the context of the study design. First, patients identified in the present study were older and we cannot comment on the incidence of post-treatment psychiatric disorders among younger bladder cancer patients. However, most patients diagnosed with bladder cancer for the first time before undergoing a treatment are in the sixth decade of life, and we provide a contemporary analysis of post-treatment psychiatric disorders in bladder cancer patients undergoing treatment. Second, mental disorders represent a chronic disease state characterized by remitting and recurring episodes which may have an antecedent presence in a significant portion of the population. While we excluded patients with a psychiatric diagnosis the year prior to treatment we cannot account for potential confounding of our results due to those patients with a mental illness greater than a year prior. Third, our analysis is retrospective and included psychiatric diagnoses derived from administrative claims data which have inherent limitations. Prior studies have determined accuracy in use of claims data to conduct research involving psychiatric diagnoses with positive and negative predictive values of 80-90% and >90%, respectively^{39, 40}. This supports the use of this data to provide a generalizable population-based assessment of post-treatment psychiatric illness and impact on survival outcomes. Fourth, we controlled for area-level socioeconomic (income and education) which limits individual level assessment. Fifth, treatment types included patients who received radiation with or without chemotherapy and this may include a large proportion of patients undergoing palliative treatment. Lastly, the use of large population-based data to identify patients with a psychiatric disorder and whether these administrative claims diagnoses account for clinically relevant psychiatric conditions remains to be determined⁴¹.

Conclusions

We identified 50.4% of patients with at least one psychiatric diagnosis following treatment for MIBC. Diagnosis of a post-treatment psychiatric disorder was associated with worse overall and cancer-specific survival thus underscoring the importance of addressing post-treatment psychiatric disorders which may improve survival in these patients. To optimize survival, non-oncologic interventions such as depression screening and survivorship clinics are critical. This information can be used to inform interventions to educate patients with muscle-invasive bladder cancer regarding the impact of different treatments on mental health.

Supplementary Material

Supp figS1

Supplementary Figure 1: Patient Selection Process

Precis (condensed abstract).

In a large population-based study among patients with muscle-invasive bladder cancer, we found 50.4% of patients who underwent treatment were diagnosed with a psychiatric disorder. Diagnosis of a psychiatric disorder resulted in significantly worse overall (HR 2.80, 95% CI 2.47 - 3.17, P < 0.001) and cancer-specific (HR 2.39, 95% CI 2.05 - 2.78, P < 0.001) survival, respectively. This information can be used to inform interventions to educate patients with muscle-invasive bladder cancer regarding the impact of different treatments on mental health.