Psychiatric Disorders and Mental Health Service Use in Patients with Advanced Cancer

Abstract

BACKGROUND. Psychological morbidity has been proposed as a source of distress in cancer patients. This study aimed to: 1) determine the prevalence of diagnosable psychiatric illnesses, and 2) describe the mental health services received and predictors of service utilization in patients with advanced cancer.

METHODS. This was a cross-sectional, multi-institutional study of 251 eligible patients with advanced cancer. Eligibility included: distant metastases, primary therapy failure, nonpaid caregiver, age ≥20 years, stamina for the interview, English or Spanish-speaking, and adequate cognitive ability. Trained interviewers administered the Structured Clinical Interview for the Diagnostic Statistical Manual IV (DSM-IV) modules for Major Depressive Disorder, Generalized Anxiety Disorder, Panic Disorder, Post-Traumatic Stress Disorder, and a detailed questionnaire regarding mental health service utilization.

RESULTS. Overall, 12% met criteria for a major psychiatric condition and 28% had accessed a mental health intervention for a psychiatric illness since the cancer diagnosis. Seventeen percent had discussions with a mental health professional; 90% were willing to receive treatment for emotional problems. Mental health services were not accessed by 55% of patients with major psychiatric disorders. Cancer patients who had discussed psychological concerns with mental health staff (odds ratio [OR] = 19.2; 95% confidence interval [95% CI], 8.90-41.50) and non-Hispanic white patients (OR = 2.7; 95% CI, 1.01-7.43) were more likely to receive mental health services in adjusted analysis.

CONCLUSIONS. Advanced cancer patients experience major psychiatric disorders at a prevalence similar to the general population, but affected individuals have a low rate of utilizing mental health services. Oncology providers can enhance utilization of mental health services, and potentially improve clinical outcomes, by discussing mental health concerns with their patients.

Over 500,000 individuals in the U.S. die from cancer each year.¹ Relief of suffering, while an essential component of cancer care throughout all phases of the disease trajectory, achieves central importance as the patient approaches the end of life. In addition to physical suffering, psychiatric syndromes are increasingly recognized as major components of distress at the end of life. Several factors contribute to psychological distress in this population,² including the grief about current and anticipated losses, fear of death, concerns about loved ones, the effect of certain chemotherapeutic drugs on mood,^3-6 and the biology of the malignancy.⁷ Untreated psychological distress in terminal cancer patients is associated with amplified pain,⁸ increased desire for hastened death,⁹ increased disability,¹⁰ impaired ability to participate in end-of-life planning,² and diminished psychosocial functioning of caregivers.¹¹

The prevalence of single or multiple psychiatric disorders in advanced cancer patients is still poorly understood. Estimates for depression range from 3-38%.^9,12-18 Early studies using rigorous methods reported lower rates of psychiatric disorders in early-stage cancer compared to subsequent studies that were conducted using less sophisticated methods or smaller sample sizes.^19,20 The present study sought to use state-of-the-art interview methodology to assess prevalence rates in patients with advanced cancer. Generalized anxiety disorder (GAD) has only been examined previously from symptom scales,^9,21 not using a standardized diagnostic interview applying DSM-IV criteria. The true prevalence of GAD and its co-occur-rence with major depressive disorder (MDD) and other psychiatric disorders is not currently known. In order to quantify and compare the rates of psychiatric disturbance in the advanced cancer population, the prevalence of these and other major psychiatric disorders, including posttraumatic stress disorder and panic disorder, must be determined using rigorous diagnostic methodology.

Although psychiatric disorders in terminal cancer patients are highly treatable,²² the available studies suggest that healthcare providers do not adequately address mental illness in this population.²³ Ford et al.²⁴ audiotaped patient—doctor outpatient clinic encounters for 117 patients who received “bad news” about a cancer diagnosis at a tertiary care center. Although patients spent 14% of their total utterances giving information about their psychosocial concerns, clinicians only devoted 3% of their verbalizations to addressing them. In a large study of 1109 cancer patients, oncologists were concordant with patient-completed scales of depression in only 13% of patients reporting the severe range of symptoms.⁸ In the same study sample, nurses tended to underrecognize severe depressive symptoms at similar rates.²⁵ Research has not yet examined the extent to which cancer patients meeting diagnostic criteria for a major psychiatric disorder discuss and/or receive help for their mental health issues.

Further research is needed before intervention strategies can be designed to improve recognition and treatment of psychiatric illnesses in patients with advanced cancer. In the current study, a baseline interview was administered to patients with advanced cancer to determine 1) the prevalence of major psychiatric disorders meeting DSM-IV criteria; 2) the frequency of communication of mental health issues with oncology clinic staff, referrals for mental health care, and receipt of mental health interventions; and 3) the factors that predict receipt of mental health services.

MATERIALS AND METHODS

Patients

Patients were recruited from August 1, 2002 to August 1, 2004 as part of an ongoing multi-institutional longitudinal evaluation (National Institutes of Health [NIH] grants MH63892, CA106370) of the prevalence of mental illness and patterns of mental health service utilization in advanced cancer patients and their primary informal (nonpaid) caregivers in the Coping with Cancer Study. This report focuses on the mental health of the patients from the baseline evaluation; the caregiver experience is reported elsewhere.²⁶ Inclusion criteria for the study were: 1) diagnosis of advanced cancer (presence of distant metastasis and failure of first-line chemotherapy); 2) diagnosis at a participating site; 3) age ≥ 20 years; 4) identified unpaid, informal caregiver; and 5) adequate stamina to complete the interview. Excluded were patient—caregiver dyads in which either the patient or the caregiver met criteria for dementia or delirium (by neuro-behavioral cognitive status examination²⁷) or could not speak English or Spanish. The participating centers were the Yale Cancer Center (CT), the Veterans Affairs Connecticut Healthcare System Comprehensive Cancer Clinics (CT), Memorial Sloan-Kettering Cancer Center (NY), and the Parkland Hospital Palliative Care Service (TX).

Members of the oncology clinic staff at participating centers identified potentially eligible patients based on the criteria described above. All potentially eligible patients were then approached by trained research staff and offered participation in the study. Interested patients and caregivers were contacted by telephone or a hospital room visit within the week to obtain informed consent. After receipt of the patients' informed consent, eligibility was confirmed through review of medical records and consultation with clinic staff.

Separate patient and caregiver interviews were then scheduled to occur in the hospital room, clinic, or home. Patients and caregivers each received $25 for completing the interview. Interviewers from each site were trained by research staff at Yale University, where they were required to achieve a high standard of accuracy and reliability based on concordance with the Yale Project Director's rating of the Structured Clinical Interview for the DSM-IV (SCID) diagnoses (target kappa > 0.85). Interviewers at each site were required to participate in an initial 2-day training session followed by annual review training sessions. The interview duration was an average of 45 minutes for patients. Patients who declined participation were asked to complete a brief questionnaire regarding their reasons for refusal, amount of emotional and physical distress, and demographics.

Of the 397 patients who were approached for participation and confirmed to be eligible, 146 (37%) declined participation. The most common reasons for nonparticipation included “not interested” (n = 66), “caregiver refuses” (n = 22), and “too upset” (n = 19). Compared with participants, nonparticipants were more likely to be white (chi-square = 4.0; P = 0.05), younger (mean age of 58.2 yrs vs. 61.6 yrs, P = 0.03), and report more distress on a 5-point Likert scale whose extremes ranged from 1 (“minimal/nonexistent”) to 5 (“distraught”) (mean score of 3.0 vs. 2.2; P < 0.0001). They did not differ significantly from participants by gender or education.

All study protocol and contact documents were reviewed and approved by the human subjects committee at Yale University and at each of the participating institutions.

Evaluation

Mental health

The SCID Axis I Modules²⁸ were used to ascertain Major Depressive Disorder (MDD), Generalized Anxiety Disorder (GAD), Post-Traumatic Stress Disorder (PTSD), and Panic Disorder (PD). Reliability and validity has been established for this instrument²⁹ (kappa = 0.92 for GAD and 0.58 for PD). Patients only completed the full SCID modules for disorders for which they screened positive (i.e., only those who endorsed one of the two depression screening items of sadness for most of the day, nearly every day, for at least 2 weeks, or feeling a lot less interested in things or unable to enjoy the things they used to enjoy, went on to complete the full module for MDD; those who did not screen positive to either of these items proceeded to the next set of questions). Mental health outcomes were analyzed as dichotomous outcomes (i.e., met criteria versus did not meet criteria for a diagnosis). Because there has been debate regarding the most valid method to diagnose MDD in advanced cancer patients whose complaints may be difficult to interpret,^30,31 the diagnosis of MDD was also determined by the Endicott symptom substitution method for comparison. Endicott has recommended that DSM criteria be modified such that weight loss be substituted with depressed appearance, insomnia with social withdrawal or decreased talkativeness, loss of energy with brooding, self-pity, or pessimism, poor concentration with lack of reactivity/cannot be cheered up.³⁰ Minor depression, which is described among DSM-IV “Criteria and Axes Provided for Further Study,” was defined as meeting two to four impairment criteria (rather than the five or more for MDD), one of which was depressed mood or lack of interest/pleasure in activities. In calculating the prevalence of minor depression, we excluded all cases in which the subject endorsed prior history of MDD to be consistent with DSM-IV Minor Depression exclusion rules.

Mental health service utilization

Patients underwent a structured interview that asked about communication about mental health concerns and mental health care utilization. Patients were directly asked if they discussed their mental health with a “mental health professional” before and after the cancer diagnosis. They were also asked whether referrals for mental health services were made, whether referred mental health services were received, source of mental health services, and types of therapies received (i.e., psychotherapy, anxiolytics, or antidepressants) after the cancer diagnosis.

Other measures

The Zubrod Performance Scale was completed³² as a summary measure of current health status based on four categories: activity, pain, food intake, and nausea. Zubrod scores ≥2 are considered poor performance.³³ Race/ethnicity was analyzed as a potential determinant of mental health utilization because seriously ill cancer patients of different race/ethnic groups have been found to vary in overall healthcare utilization.^34,35 Race/ethnicity was determined by patient self-report in response to the open-ended question: “What race or ethnicity do you consider yourself to be?” Patients who identified themselves as Hispanic were analyzed as such, regardless of whether they reported white or black status. Hispanic patients were classified as white if they endorsed this as their racial/ethnic group; Hispanic patients not indicating they were white were classified as ‘nonwhite.’

Data analysis

Descriptive statistics were used to characterize the demographics of the study sample. The prevalence of the major psychiatric disorders was calculated. Frequencies of responses to the mental health care utilization interview were calculated, overall and stratified by the presence of a major psychiatric disorder. Chi-square or Fisher exact test statistics were used to determine if responses to the mental health care utilization interview varied by the presence of a major psychiatric disorder. The association between patient attributes and receipt of mental health care services was determined using bivariate logistic regression analyses. A multiple logistic regression model was constructed to include all the covariates that were significantly associated with receipt of mental health care in unadjusted analysis. Colinearity was assessed by cross-tabulations and plots of the analyzed covariates before proceeding to multiple logistic regression; no appreciable colinearity was detected. Data were analyzed with the SAS System for Windows v. 8.2 (SAS Institute, Inc., Cary, NC) with 2-tailed statistical tests.

RESULTS

Patient Characteristics

The characteristics of the 251 patients who enrolled and completed the evaluation are displayed in Table 1. In general, the sample tended to be white (71%), elderly (median age, 60 yrs), at least high school-educated, of relatively good health status (70% with Zubrod Score <2), and insured (80%).

TABLE 1.

Characteristics of the Study Population of Advanced Patients with Cancer

Attribute
No. of cases	251
Women, no. (%)	114 (45)
Race/ethnicity, no. (%)
White, non-Hispanic	178 (71)
Black, non-Hispanic	37 (15)
Hispanic	25 (10)
Asian	9 (4)
Other	2 (1)
Age at evaluation in yrs
Mean (SD)	58 (12)
Median (range)	60 (22%-88)
Education in yrs
Mean (SD)	13 (4)
Median (range)	13 (0%-24)
Treatment center, no. (%)
Yale Cancer Center	134 (54)
VA CT Healthcare System	24 (10)
MSKCC	46 (18)
Parkland Hospital	45 (18)
Cancer site, no. (%)
Gastrointestinal	114 (46)
Thoracic	51 (20)
Breast	17 (7)
Leukemia/lymphoma	15 (6)
Genitourinary	13 (5)
Musculoskeletal	11 (4)
Other	24 (10)
Unknown	6 (2)
Zubrod Score, no. (%)a
< 2	175 (70)
≥ 2	73 (29)
Missing	19 (8)
Insurance status, no. (%)
Insured	200 (80)
Uninsured	44 (18)
Missing	7 (3)

SD: standard deviation; VA CT: Veterans Affairs Connecticut Healthcare System Comprehensive Cancer Clinics; MSKCC: Memorial Sloan-Kettering Cancer Center; Parkland Hospital: Parkland Hospital Palliative Care Service.

^aZubrod performance scores of ≥2 are considered indicative of poor current health status.

Prevalence of Psychiatric Conditions

Overall, 11.6% (n = 29; 95% confidence interval [95% CI], 7.9 -16.1%) of the 251 participants met DSM-IV criteria for at least one of the major psychiatric conditions of MDD, GAD, PD, and PTSD, as displayed in Table 2. The most frequently identified condition was MDD (6.8%; 95% CI, 3.8-10.2%). The prevalence of major depression by the Endicott substitution method was 6.0%, similar to that by DSM-IV criteria. An additional 7.2% (n = 18; 95% CI, 4-10.4%) of patients met criteria for minor depression, totaling 14% of the sample with minor or major depression. The prevalence of major psychiatric disorders was not higher in patients enrolled from the Parkland Hospital Palliative Care Service, among those with a Zubrod score of 2, or within any race/ethnicity groups (data not shown).

TABLE 2.

Prevalence of Psychiatric Conditions Meeting Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) Criteria

Diagnosis	No. (%)	95% CI
Major depressive disorder	17 (6.8)	3.8-10.2
Generalized anxiety disorder	8 (3.2)	1.1-5.6
Panic disorder	12 (4.8)	2.2-7.7
Posttraumatic stress disorder	6 (2.4)	0.5-4.4
At least one of above diagnoses	29 (11.6)	7.9-16.1

95% CI: 95% confidence interval.

Thirty-five percent (n = 10) of the 29 patients with a psychiatric diagnosis had two or more psychiatric diagnoses. Among the 17 patients who met diagnostic criteria for MDD, 23.5% (n = 4) also had GAD, 17.6% (n = 3) had PD, and 29.4% (n = 5) had PTSD.

A subanalysis stratified patients according to whether they had a history of lifetime MDD from the SCID (data not displayed). We found that among the 29 patients with a current psychiatric disorder, 12 (41%) had a history of MDD by the lifetime SCID and 17 (59%) had no history of MDD. These results indicate that a higher percentage of cases of MDD after the patient's cancer diagnosis were new onset cases than a perpetuation or recurrence of a preexisting condition. Patients with a positive history of lifetime MDD had a 29.3% (12 of 41 patients) prevalence of a current major psychiatric disorder versus 8.4% (17 of 202 patients) in patients without lifetime MDD (P = 0.001). Thus, preexisting psychopathology heightened risk for postcancer diagnosis psychiatric illness.

Mental Health Service Utilization

Discussion with healthcare team

Twenty-four percent (60 of 251 patients) of the sample reported retrospectively that, prior to the cancer diagnosis, they had discussed mental health concerns with a mental health professional (Table 3).

TABLE 3.

Mental Health Service Utilization among Study Sample, Stratified by the Presence of Major Psychiatric Condition^a

	Full sample (n = 251)	Psychiatric diagnosis (n = 29)	No psychiatric diagnosis (n = 222)	P valueb
Discussed mental health before the cancer diagnosis with a mental health professional	60 (23.9%)	9 (31.0%)	51 (23.0%)	0.34
Discussed mental health since the cancer diagnosis with a mental health professional	43 (17.1%)	7 (24.1%)	36 (16.2%)	0.30
Accessed mental health services since the cancer diagnosisc	70 (27.9%)	13 (44.8%)	57 (25.7%)	0.03
— Psychotherapy	32 (12.7%)	8 (27.6%)	24 (10.8%)	0.02
— Antidepressants	56 (22.3%)	12 (41.4%)	44 (19.8%)	0.009
— Anxiolytics	21 (8.4%)	7 (24.1%)	14 (6.3%)	0.005
Would pursue help if aware of emotional problem	225 (89.6%)	27 (93.1%)	198 (89.2%)	0.75

^aMajor Depressive Disorder, Generalized Anxiety Disorder, Panic Disorder, or Posttraumatic Stress Disorder.

^bPatients with and without a psychiatric diagnosis compared using the chi-square test or Fisher exact test, as appropriate.

^cCategories are not mutually exclusive. The mental health services analyzed were restricted to psychotherapy, antidepressants, and anxiolytics.

Seventeen percent of patients (n = 43) reported having had discussions about mental health since their cancer diagnosis; 24% of those with a current major psychiatric disorder and 16% of those without a current psychiatric disorder reported such a discussion. Discussing mental health concerns before the cancer diagnosis was significantly associated with discussing mental health issues after diagnosis (chi-square 17.7; P < 0.0001), as shown in Table 4. Nevertheless, among the 60 cases who had discussions with a mental health professional before the patients' cancer diagnosis, only 35% (n = 21) had such discussions after the diagnosis.

TABLE 4.

Discussion with Mental Health Professional before and after Cancer Diagnosis

		Discussed mental health concern with mental health professional after patients' cancer diagnosis
		Yes (n = 43; 17.1%)	No (n = 208; 82.9%)	Chi-square statistic	P value	Degrees of freedom
Discussed mental health concern with mental health professional	Yes (n = 60; 23.9%)	21 (8.4%)	39 (15.5%)	17.7331	< 0.0001	1
before patients' cancer diagnosis	No (n = 191; 76.1%)	22 (8.8%)	169 (67.3%)

Among patients with a positive history of lifetime MDD by the SCID, 44% (13 of 41 patients) had discussed mental health issues with a mental health professional before the cancer diagnosis compared with 24% (10 of 41 patients) after the diagnosis. In contrast, 19% of patients (39 of 202 patients) with no history of lifetime MDD had mental health discussions before their cancer diagnosis and 14% (31 of 202 patients) had such discussions after diagnosis.

Referral and receipt of mental health services

In the entire sample, 90% of participants stated that they would seek help if they were aware that they had an “emotional problem.” Of the 86 patients who discussed mental health issues with a health professional after the cancer diagnosis, 54 (62.8%) were offered treatment or given a referral for mental health services; of those 54 individuals, 48 (88.9%) received the treatment offered (data not shown).

Overall, through referral and nonreferral mechanisms, mental health services were accessed by 28% (n = 70) of patients since the diagnosis of cancer (Table 4). This proportion was 45% in patients who met criteria for a psychiatric disorder, indicating that more than half of those who met criteria for psychiatric illness did not access mental health services. Among the 70 individuals who had accessed mental health services, only 18.6% (n = 13) met criteria for one of the psychiatric disorders evaluated. The following are the proportions of individuals in the entire sample receiving mental health services: antidepressants (22%), psychotherapy (13%), and anxiolytics (8%).

Predictors of receipt of mental health services

Table 5 displays factors analyzed as potential determinants of receipt of mental health interventions. In unadjusted analysis, the following factors were each associated with receiving mental health treatment after the diagnosis of advanced cancer: 1) discussion of mental health concerns before the cancer diagnosis (odd ratio [OR] = 6.71; 95% CI, 3.63-12.41); 2) discussion of mental health concerns after the cancer diagnosis (OR = 22.98; 95% CI, 11.17-47.29); 3) currently meeting criteria for a major psychiatric illness (OR = 2.35; 95% CI, 1.07-5.19); and 4) white race/ethnicity (OR = 3.71; 95% CI, 1.73-7.95). Also examined, but not found to be significant, were insurance status, gender, age, education, and Zubrod score. In adjusted analysis, discussion of mental health concerns with a healthcare professional after the cancer diagnosis was the strongest predictor of receiving mental health services (OR = 19.21; 95% CI, 8.90-41.50). White patients remained significantly more likely than nonwhite patients to receive services (OR = 2.74; 95% CI, 1.01-7.43). Currently meeting criteria for a major psychiatric disorder did not significantly predict use of services in adjusted analysis.

TABLE 5.

Determinants of Receipt of Mental Health Treatment among Advanced Cancer Patients^a

	Unadjusted		Adjustedb
Variable	Odds ratio	95% CI	Odds ratio	95% CI
Discussed mental health before cancer diagnosis	6.71	3.63-12.41	4.31	1.96-9.48
Discussed mental health after cancer diagnosis	22.98	11.17-47.29	19.21	8.90-41.50
Presence of current major psychiatric diagnosis	2.35	1.07-5.19	2.22	0.71-6.93
White non-Hispanic race/ethnicity	3.71	1.73-7.95	2.74	1.01-7.43
Insured status	1.88	0.83-4.29
Female	0.72	0.41-1.24
Age	1.01	0.99-1.03
Education (yrs)	0.94	0.87-1.01
Zubrod score ≥ 2c	1.09	0.80-1.49

95% CI: 95% confidence interval.

^aMental health treatment was restricted to psychotherapy, antidepressants, and anxiolytics.

^bAll covariates found to be significant in the unadjusted regression analysis were included in the final adjusted analysis.

^cA Zubrod score ≥ 2 was indicative of a decreased health performance.

DISCUSSION

This study provides a comprehensive application of DSM-IV criteria to the determination of the prevalence of psychiatric disorders in advanced cancer patients. It is also among the first to examine and relate these diagnoses to the utilization of mental health services at the end-of-life. Unlike prior studies that have used symptom checklists that focus primarily on symptoms of generalized distress or depression,^36-39 trained interviewers used current DSM IV psychiatric diagnostic criteria to determine the prevalence of the full range of mental disorders in a relatively large, ethnically and geographically diverse sample of end-stage cancer patients. The results indicated that 12% of advanced cancer patients met established criteria for at least one major psychiatric disorder, which is a rate that is comparable to that of the general population,⁴⁰ and consistent with older studies in cancer patients that relied on rigorous interview criteria.^9,12,15,41 However, our study identified the under-utilization of mental health services among advanced cancer patients. Fifty-five percent of patients who had a major psychiatric disorder had not received mental health services.

Our results indicated that members of the oncology team frequently failed to address mental health concerns of patients with advanced cancer. Only 17% of our sample of advanced cancer patients had discussions of mental health issues with a mental health care professional after the cancer diagnosis. This proportion is essentially the same as the rate reported prior to the cancer diagnosis overall; the proportion is decreased by half (from 44% to 24%) in patients with a positive history of lifetime MDD by the SCID. This suggests that many of the patients most vulnerable to psychiatric illness (i.e., those who had discussed mental health concerns with a clinician prior to their cancer diagnosis and those with a positive history of lifetime MDD) do not discuss these concerns after they learn that they have cancer. This also suggests that the powerful effect of postcancer diagnosis mental health discussions on accessing mental health services is not a function of tapping into a preexisting pattern of help-seeking behavior.

In terms of intervention, only half the individuals who met the criteria for a major psychiatric disorder had accessed mental health services. The most robust predictors of receipt of mental health services by advanced cancer patients were 1) discussing mental health issues with a healthcare professional, and 2) white race/ethnicity; not whether one currently meets criteria for a psychiatric diagnosis. These results highlight the critical role of healthcare providers, especially oncologists who direct overall care, in assessing the mental health of their patients with life-threatening cancer.

The National Institutes of Health State-of-the Science Conference Statement regarding symptom management in cancer observed that there is considerable heterogeneity in published prevalence rates of depression in this population because of “lack of uniformity in measurement and methodology.”⁴² Our estimates of the MDD prevalence rates (6.8%) using DSM-IV criteria is comparable to the point prevalence of depression in healthy women (4.5-9.3%) and men (2.3-3.2%), suggesting that there is no excess depression associated with advanced cancer.³ Contrary to the prevailing opinion, our results are similar to reports which relied on stringent interview-based diagnostic criteria from previous versions of the DSM.⁴³ Derogatis et al.¹² concluded that 5% of 215 cancer patients had MDD based on DSM-III standards. A 2004 study by Akechi et al.¹⁵ assessed 129 patients with unresectable nonsmall cell lung carcinoma with DSM-IIIR criteria and also found a 5% prevalence of MDD.

Our study highlights the comorbidity of other major psychiatric conditions in the advanced cancer population, which has been previously identified in studies of the general cancer population.^43,44 The prevalence of GAD and PD in the cancer population together represent a burden that is similar to MDD. In our sample, among individuals with MDD, 24% concurrently met criteria for GAD, 18% concurrently met criteria for PD, and 30% also concurrently meet criteria for PTSD. These results are particularly striking because greater disability, more severe symptoms, and greater refractoriness to therapy have been observed in individuals with multiple psychiatric conditions, compared to those with a single disorder.^3,45

There is reason to believe that our reported rates of psychiatric disorders may be a low estimate of the true prevalence of clinically significant psychiatric distress. Our comparison of participants and nonparticipants indicates that we encountered the typical difficulty in recruiting the most distressed patients. Another potential explanation for the low rates of psychiatric disorders is that only patients who screened positive then completed the full SCID module for that particular disorder and perhaps the screening questions in the SCID were not sufficiently sensitive for this population. One finding suggestive of this is that the trained psychiatric interviewers identified 10 patients who failed to meet DSM-IV criteria for current MDD as clinically depressed based on their personal impressions (rated independently from the SCID). In addition, the poor correspondence between meeting criteria for psychiatric disorders and mental health service use (e.g., meeting criteria for a psychiatric disorder did not predict use of mental health services, many persons with a psychiatric disorder did not access mental health services and vice versa) suggests that treatment effectively reduced psychiatric morbidity and/or potential problems related to diagnostic accuracy in the identification of mental illness in advanced cancer patients. This finding also raises the question of whether patients may elect to utilize mental health services for distress that does not constitute a psychiatric disorder (e.g., fear about the future, anxiety about death, concerns about family members). Our measurement instruments are to our knowledge the best currently available, but may not be sensitive enough to the overlap of physical and psychological distress that are commonly seen among cancer patients. Our findings highlight the need to understand better the potential deficiencies in diagnostic precision in the identification of psychiatric disorders among advanced stage cancer patients, as well as what psychological issues for which patients believe they need help in the context of advanced cancer.

To our knowledge, fewer than half of the individuals meeting stringent criteria for a major psychiatric condition had been offered a mental health intervention. All cancer patients deserve optimal therapy for psychiatric disorders, particularly to enhance quality of life at the end of life. Previous studies suggest that patients with breast cancer and lung cancer who experience more depression may have worse survival outcomes.^46,47 Therapeutic studies in cancer patients have demonstrated that this population is responsive to both psychotropic medications and psychotherapy.⁴² Pharmacotherapy has been termed the ‘mainstay' of depression treatment in advanced cancer patients and is well established.^48,49 Randomized controlled trials of antidepressants which were of adequate dose and duration show benefit in as many as 65-70% of patients.^49,50 Pirl⁵¹ summarized the results of three published meta-analyses designed to determine the efficacy of published psychosocial interventions of depression in cancer patients. He concluded that psychotherapy interventions result in moderate improvement in symptoms in the vast majority of studies. In addition to experiencing improved quality of life, cancer patients who receive adequate treatment of their mental health are more compliant with chemotherapy.⁵²

Numerous barriers to treating psychosocial distress in advanced cancer patients have been described. The American College of Physicians—American Society of Internal Medicine Consensus Panel recognized that “patients and clinicians frequently believe that psychological distress is a normal feature of the dying process.”² The National Institutes of Health (NIH) State-of-the-Science Panel highlighted the inability of oncology providers to recognize depression as being a “particularly important” impediment.⁴² The panel also speculates that patients may consider psychiatric diagnoses and/or mental health service use to be stigmatizing. Our study findings confirm that oncology healthcare providers do not adequately priori-tize mental health issues, given the low proportion of patients who discussed psychosocial issues with oncology clinicians as well as the high proportion of patients with diagnosed psychiatric illness who did not access mental health services (48%). However, 93% of the advanced cancer patients who met criteria for a psychiatric disorder stated that they would pursue mental health services if they were aware that they had an emotional problem, and 89% of those referred actually went, suggesting that current cohorts of patients appear to be receptive to treatment for mental health problems.

Another important barrier to receipt of mental health services identified by our study, previously unreported in cancer patients, is race/ethnicity. Given that no significant racial/ethnic differences in the prevalence of psychiatric disorders emerged, different mental health needs cannot account for racial/ethnic differences in mental health service use. As noted earlier, nonparticipants were more likely to be distressed and of white race/ethnicity than participants. However, further analyses were performed and there is no significant association noted between race/ethnicity and distress level among participants, nonparticipants, or the entire eligible sample. Past studies either did not report race or consisted of samples of more than 85% white patients.⁵² Our results found white patients to be almost three times more likely to receive a psychiatric intervention than nonwhites, after adjusting for other factors. Although the explanation for this discrepancy by race/ethnicity is not clear, there is reason to believe that access to mental health care (race/ethnicity was closely associated with insurance status), cultural attitudes about seeking assistance for mental health problems, and ingrained beliefs about the need to struggle when confronted with adversity may be operating.⁵³ In addition, clinicians may be less likely to explore these issues with non-white patients, or to attribute distress to other psychosocial circumstances (e.g., poverty) than to psychiatric disorder. Ongoing studies are being conducted by our research group to explore the associations between racial/ethnic status, acculturation, and the willingness to seek mental health services (NIH grant CA106370). Such studies will inform other intervention strategies to meet mental health needs of nonwhite cancer patients.

There are limitations to the current study. Given the 63% participation rates, there is the potential for selection bias. Individuals who reported more distress were less likely to participate. Therefore, as discussed more extensively earlier, the prevalence rates reported herein are likely an underestimate of the burden of psychiatric problems in this population. Another limitation is that we did not ascertain whether mental health services were accessed before the cancer diagnosis, a factor that may be associated with postcancer use of services.

As cancer treatment becomes more scientifically sophisticated, the focus is increasingly on the treatments and less on the patient. If quality of life is an important goal of treatment, psychological symptoms should be recognized as sources of suffering that can be as debilitating as physical complaints. Discussion of mental health issues with a healthcare provider allows distress to be detected, and is the most important predictor of whether services will be received. More of these discussions need to occur in the oncology clinical care setting. Further studies are needed to better understand why nonwhite patients are less likely to access mental health care services and how to detect the range of clinically significant psychological disorders in cancer patients in a meaningful way.