Although the majority of Americans receive mental health care from medical providers who are not mental health clinicians, little is known about the extent to which subspecialists provide such care (1). Most of the prior research on the delivery of mental health services in medical settings has taken place in primary care. With the recent publication of the Institute of Medicine's report Cancer Care for the Whole Patient: Meeting Psychosocial Health Needs (2), increased attention has been focused on psychosocial care by oncology providers.
The National Comprehensive Cancer Network (NCCN) coined "psychosocial distress" as a nonstigmatizing term that encompasses emotional, psychological, and social difficulties, as well as DSM-IV psychiatric disorders. As part of their Supportive Care Guidelines, the NCCN charged oncology providers with routinely screening for and treating psychosocial distress among their patients.
Although one-third to one-half of all patients with cancer have significant psychosocial distress, these patients report that oncology providers do not consider psychosocial support integral to their care and fail to recognize, adequately treat, or offer referral for psychosocial distress (2,3).
The goal of this study was to survey oncologists on the perceived prevalence and management of psychosocial distress in their practices, as well as mental health resource availability. The data were part of a larger survey study that investigated oncologists' awareness of the NCCN guidelines and their screening practices for psychosocial distress.
After the study received institutional review board approval, 1,000 oncologists from the American Society of Clinical Oncology (ASCO) were randomly selected from the association's annual membership directory listing of 17,000 U.S. members. These oncologists were sent an e-mail requesting their participation in an online survey. Details of survey instrument development, data collection, and analyses are available elsewhere (4). The final survey consisted of 21 questions about observed prevalence of distress, frequencies of types of distress, availability of mental health services, screening for distress, barriers to screening, management of distress, and knowledge of NCCN guidelines (5). The survey was conducted online between September 12, 2005, and November 12, 2005, with two subsequent postal mailings sent to nonresponders. The survey closed on July 15, 2006. Online and mail surveys were identical with the exception of the response format for one item: types of mental health clinicians associated with the respondent's practice. The online version allowed only one response and the mail version allowed multiple responses. Because of this difference, only data from the mail version were analyzed for this question.
Multiple regression analyses were used to identify independent associations between practice variables (practice site, level of experience, hours of patient care per week, and availability of mental health services) and aspects of psychosocial care.
Of the 1,000 oncologists contacted, nine e-mails or letters were returned because of invalid addresses, 24 respondents declined participation, and two surveys were returned after the survey closed. Of the remaining sample (N=965), 448 responded (46%). There were 189 online responses and 259 mail responses, with no differences in practice settings between the online and mail respondents. Demographic characteristics of respondents are fairly representative of the ASCO membership (see Table 1).
Respondents estimated that a mean±SD of 72%±23% of their patients experience psychosocial distress, with over one-third (38%±22%) having distress that warrants treatment. On a scale with possible scores ranging from 1, never, to 5, very often, respondents identified the frequencies of various types of distress. Practical problems were most common (mean=3.74±.86), followed by emotional problems unrelated to illness (mean=3.63 ±.76), psychological coping with illness and treatment (mean=3.53±.90), employment issues (mean=3.46±.80), dealing with a partner (mean=2.92±.77), dealing with children (mean=2.80±.85), and spiritual concerns (mean=2.40±.80). In multivariate linear regression models, none of the practice variables was associated with either rates of observed distress or distress requiring treatment.
Almost all respondents (441 of 447, or 99%), reported referring patients to psychosocial services or starting psychiatric medications. Nearly half said they refer patients to psychosocial services only (212 of 447, or 47%), and an approximately equal number said they both refer patients and start medications (214 of 447, or 48%); a few reported starting medications only (15 of 447, or 3%).
Half (225 of 447, or 50%) reported having mental health services affiliated with their practice, which varied by treatment setting as follows: 92 of 122 (75%) in cancer centers, 15 of 31 (48%) in hospitals, and 111 of 284 (39%) in the community (p<.001). Affiliated mental health professionals included social workers (122 of 258, or 47%), psychologists (67 of 259, or 26%), psychiatrists (44 of 259, or 17%), and psychiatric nurse practitioners (21 of 259, or 8%). Availability of social workers and psychologists was associated with practice setting (p<.001 and p= .008, respectively), with greater availability in cancer centers.
The most common interventions reported were referrals to support groups (307 of 447, or 69%) and to social workers (287 of 447, or 64%). Over half (241 of 447, or 54%) reported referring distressed patients to psychiatrists. Oncologists who reported not having mental health services affiliated with their practice reported lower rates of referrals to both social workers (p<.001) and psychiatrists (p=.009). However, the oncologists estimated that only about half of their patients (mean=58%±27%) followed through with these referrals. Oncologists who had mental health services affiliated with their practices reported higher rates of follow-up in referrals compared with those without such services (p<.001).
Although almost half of the respondents reported starting their patients on a psychotropic medication regimen, there were no differences in practice variables between oncologists who reported prescribing and those who did not. Selective serotonin reuptake inhibitors (SSRIs) were commonly prescribed (204 of 229, or 89%), as were benzodiazepines (158 of 229, or 69%). Oncologists also reported prescribing tricyclics (66 of 229, or 29%), stimulants (53 of 229, or 23%), other antidepressants (45 of 229, or 20%), antipsychotics (21 of 229, or 9%), and anticonvulsants (five of 229, or 2%). The prescription of SSRIs was associated with having less than 20 years of experience (odds ratio=.80, 95% confidence interval=.68–.96 p=.015) but not with any other practice variable.
The oncologists in this sample reported frequently encountering psychosocial distress in their practices. Their estimated rate of clinically significant distress, 38%, is similar to rates reported by patients in other studies (6). Fortunately, almost all reported providing some level of psychosocial care.
However, these findings also highlight two important issues in the delivery of psychosocial care to cancer patients: recognition of distress and the availability of mental health resources. Although nearly all the oncologists reported intervening once psychosocial distress was identified, many of their patients may not receive mental health treatment because of lack of recognition. For about half of oncology patients with psychiatric disorders, such as major depression, the disorders go unrecognized by their oncologists (7). The majority of oncologists, and even those in NCCN institutions, do not routinely screen with methods that are accurate for detection of psychosocial distress (4,8). This may not be so different from primary care, where patients report screening rates of 21% for anxiety and depression by their primary care providers, and depression is recognized for only 64% of patients with depression through usual care (9,10).
Only half of the oncologists reported having mental health services affiliated with their practices. As in primary care settings (9), lack of available services was associated with lower rates of referral. The fact that rates of referrals appear greater than the availability suggests that oncologists also refer their patients for mental health care in the community. However, patients with severe medical comorbidities may be less likely to follow through with referrals in the community. Without available mental health clinicians, developing screening tools and treatment protocols that oncologists themselves could implement may be useful (7).
Although this study provides a first look at the delivery of psychosocial care by oncologists, it has several limitations. Self-reports may not accurately reflect clinical practice. Also, the survey focused on "distress" and not on the identification and management of specific psychiatric disorders, such as major depression, anxiety, or delirium. With our modest response rate, there is also the possibility that oncologists who are more sensitive to psychosocial issues were more likely to complete the survey. Finally, patients are often cared for by a team of providers, and this survey could not capture multidisciplinary care.
Oncologists reported delivering psychosocial care to their patients. Efforts to improve this care would benefit from increasing access to mental health services and increasing oncologists' knowledge about identifying and treating psychosocial distress among their patients.
This research was supported by grants R03 CA113104 and K23 CA115908 from the National Cancer Institute.
Dr. Pirl has received research funding from Forrest Laboratories. The other authors report no competing interests.