Anxiety and depression after cancer diagnosis: Prevalence rates by cancer type, gender, and age Wolfgang Lindena, b, , 1, , Andrea Vodermaiera, c, , 1, Regina MacKenzieb, Duncan GreigaShow moredoi:10.1016/j.jad.2012.03.025

AbstractBackgroundReported prevalence of emotional distress in cancer patients varies widely across studies. The present study determined prevalence of anxiety and depression (separated for presence of symptoms versus clinical levels) in a large, representative sample of cancer patients after diagnosis.MethodDuring the years 20042009, 10,153 consecutive patients were routinely screened with the Psychosocial Screen for Cancer questionnaire at two major cancer centers.ResultsPatients' mean age was 59years and 45% were men. Across cancer types, 19.0% of patients showed clinical levels of anxiety and another 22.6% had subclinical symptoms. Further, 12.9% of patients reported clinical symptoms of depression and an additional 16.5% described subclinical symptoms. Analyses by cancer type revealed significant differences such that patients with lung, gynecological, or hematological cancer reported the highest levels of distress at the time point of cancer diagnosis. As expected, women showed higher rates of anxiety and depression, and for some cancer types the prevalence was two to three times higher than that seen for men. In some cancer types emotional distress was inversely related to age. Patients younger than 50 and women across all cancer types revealed either subclinical or clinical levels of anxiety in over 50% of cases.LimitationsFindings describe levels of emotional distress after diagnosis but cannot inform about trajectories of anxiety and depression over time.ConclusionGiven that levels of anxiety and depression varied widely by cancer type, gender, and age, these results inform which cancer patients are most likely in need of psychosocial support.Keywords Anxiety; Depression; Cancer diagnosis; Cancer type; Gender; Age

1. IntroductionEmotional distress in cancer patients (operationally defined here as anxiety or depression) reduces patients' quality of life, negatively impacts compliance with medical treatment (DiMatteo et al., 2000andGreer et al., 2008), and carries an elevated risk of mortality (Pinquart and Duberstein, 2010andSatin et al., 2009). For this reason, emotional distress is recognized as the sixth vital sign in cancer care (Bultz and Carlson, 2005), calling for systematic allocation of supportive resources. Cost-effective resource allocation, however, requires precise knowledge of the extent of a problem. Correspondingly, researchers have reported the prevalence rates of anxiety and depression in the cancer population but this literature can be difficult to interpret because:[a]researchers do not always clarify whether an actual clinical diagnosis has been made or whether symptoms were present;[b]diagnostic cut-offs are sometimes not empirically validated and diagnoses are at times based on different diagnostic systems, with the latter resulting in prevalence estimates that range from 25% to 38% in the same sample (Kathol et al., 1990);[c]the quality of measurement tools varies greatly (Vodermaier et al., 2009); and[d]prevalence rates are often assessed at varying time points in the trajectory of the disease although emotional distress is known to change as patients transition through stages of diagnosis, acute treatment, and post-treatment (Stommel et al., 2004). Prevalence rates for distress also depend on whether or not patients have responded positively to treatment or not (Hopwood et al., 2009andvan't Spijker et al., 1997).When researchers report on the prevalence of emotional distress they frequently acknowledge a wide range of observed prevalence rates. Massie (2004) reported that major depression in cancer patients ranged from 0 to 38%. Variability in reported prevalence rates is echoed in a meta-analysis (van't Spijker et al., 1997) where prevalence rates ranged from 0% to 46% for depression and 1% to 49% for anxiety. Reporting such wide ranges is not helpful to administrators who must allocate finite resources to help alleviate emotional distress in patients. A more recent meta-analysis restricted to cancer inpatients and high-quality prevalence estimates based on structured clinical interviews (as compared to self-report scales or physician judgment) demonstrated that one third of cancer patients suffer from some type of mental disorder during active treatment (Singer et al., 2010).We posit that large-sample studies are more likely to sample broadly and to reveal replicable findings and as shown in Table1, a number of large sample studies of distress prevalence exist (where large is arbitrarily defined as n>1000) (Brintzenhofe-Szoc et al., 2009, Caminiti et al., 2004, Carlson et al., 2004, Hinz et al., 2010, Hopwood et al., 2009, Sharpe et al., 2004andZabora et al., 2001). Together, these studies possess many strengths, including the noted large samples, no restrictions regarding cancer types (except for one study), and use of standardized tools. Noted weaknesses, however, include non-consecutive recruitment and variable timing of assessment within the disease and treatment trajectory.Table1. Prevalence studies of patients with cancer, N>1000.Study and sample sizeTypes of cancer included[a] Standardized measure?[b] Cut-offs empirically determined?Differentiation of disorder from presence of symptomsConsecutive recruitmentAssessment conducted at fixed time in disease trajectory?

Zabora et al. (2001)N=4496All[a] Yes, BSI 53[b] Not in cancer patientsNo, caseness was defined as T-score >62NoNo

Carlson et al. (2004)N=3095All[a] Yes, BSI-18[b] YesNo, caseness was defined as T-score >62NoNo

Caminiti et al. (2004) N=3293All[a] HADS[b] YesNo, only clinical caseness reportedNoYes, during chemotherapy

Brintzenhofe-Szoc et al. (2009)N=8265All[a] BSI-53[b] Not in cancer patientsNoNoNo

Hopwood et al. (2009)N=2208Early stage breast cancer[a] Yes, HADS[b] YesYesYesYes, in acute care, before radiation

Hinz et al. (2010)N=1529All[a] Yes, HADS[b] YesNo, only caseness was reportedNoNo

Sharpe et al. (2004)N=5613All[a] Yes, HADS.[b] YesYesYesNo

Linden et al. (this article)N=10,153All[a] Yes, PSSCAN[b] YesYesYesYes, post-diagnosis but pre-treatment

Table optionsThe present study adds data on subclinical and clinical symptoms of anxiety and depression in another large sample and uses a standardized instrument with good psychometric properties and empirically validated cut-offs. While acknowledging the substantial overlap of these constructs, we chose not to aggregate data into a single distress index because anxiety and depression may have different trajectories from diagnosis to end of first-line treatment to long-term follow-up (Ando et al., 2009, Den Oudsten et al., 2010, Kangas et al., 2007andThomas et al., 2011). Anxiety tends to reflect a reaction to the diagnosis and the anticipated aversive treatment and is often transient, depression is more likely to reflect a stable predisposition. Furthermore, given earlier reports that distress may vary not only by cancer type but also by age and gender, we explicitly conducted analyses on these moderator variables as well.2. Methods2.1. ProceduresIn 2004, all British Columbia Cancer Agency centers implemented a routine screening program for emotional distress. Since then, patients have routinely completed the Psychosocial Screen for Cancer (PSSCAN) (Linden et al., 2005andLinden et al., 2009) during the first visit to a provincial cancer center, prior to beginning treatment. Patients are free to refuse completion but record checks during data acquisition revealed that 73% of eligible patients provided usable data. Unfortunately, the cancer clinics do not systematically record reasons for non-completion, but clinicians report that an estimated 1015% of the total patient pool speak English as a second language and do not complete standardized tests due to poor comprehension. Other prominent reasons for non-participation are a high level of medical crisis at arrival in the cancer clinic and/or lack of lucidity typically due to old age (Linden et al., 2005).For this study, the psychosocial data obtained at the two largest centers were merged with electronically archived medical and demographic data. Given that breast cancer in men is a very rare disease and difficult to compare with breast cancer in women, the few men were excluded from analysis. Similarly, prostate cancer as a tumor type was distinguished from the large genitourinary category given that it represents a highly prevalent disease in older men and carries a different prognosis than other genitourinary cancers. Also, three age groups were formed: [a] patients below 50 (where cancer is relatively rare), [b] 50 to 69 (where cancer is on the rise), and [c] >70 (where cancer is often superimposed on pre-existing health problems). This study was approved by the Institutional Review Board of the British Columbia Cancer Agency.2.2. MeasureThe 21-item Psychosocial Screen for Cancer assesses anxiety and depressive symptoms, perceived social support, desired social support, and quality of life. The scale was specifically developed and validated for use with cancer patients and was designed to serve as a clinical as well as a research tool. The anxiety and depression subscales were developed to map onto DSM-IV TR defined disorders, i.e., major depression and generalized anxiety disorder and these scales were of particular interest for the present study. Both subscales are comprised of 5 items each. Items are scaled on a 5-point Likert scale (from 1 = not at all to 5 = very much so), with a potential range of scores from 5 to 25. An example item of the anxiety subscale is I felt nervous and shaky inside. And for depression a sample is In the past year I have had 2weeks or more during which I felt sad, blue or depressed. These subscales have satisfactory internal consistency (=.83, for anxiety; =.79 for depression), and are sensitive to change (testretest reliability over 2months: r=.67 for anxiety, r=.61 for depression). The anxiety and depression subscales were highly sensitive and specific when compared to the Hospital Anxiety and Depression Scale (sensitivity 92% and specificity 98%, for anxiety; sensitivity 100% and specificity 86%, for depression; Zigmond and Snaith, 1983). As well, normative data comparing cancer patients with healthy controls exist (Linden et al., 2009). Results regarding discriminant validity indicate that scores of 11 and greater are suggestive of a clinical diagnosis, and scores greater than 8 but less than 11 indicate the presence of symptoms for both the anxiety and depression subscales alike. PSSCAN psychometrics were also judged to meet a satisfactory standard in a comprehensive comparison of screening tools and available data indicate that it is suitable as a research tool and not just a screening tool (Vodermaier et al., 2009).3. Results3.1. Sample descriptionPatient mean age was 58.9 (14.6) years. 4553 (44.8%) patients were male, and 5600 (55.2%) were female. Absolute and relative probabilities by cancer types were n=2430 (23.9%) breast, n=1598 (15.7) prostate, n=1334 (13.1%) gastrointestinal, n=949 (9.3%) gynecological, n=673 (6.6%) lung, n=550 (5.4%) neuroendocrine, n=501 (4.9%) skin, n=454 (4.5%) head and neck, n=302 (3.0%) genitourinary (excluding prostate), n=211 (2.1%) bone, n=180 (1.8%) hematological, and n=971 (9.6%) other unclassified primary tumors.3.2. Anxiety mean scoresAs shown in Fig.1, on average cancer patients experienced anxiety just above the subclinical threshold (8.1(3.8)). Relative to the total sample, aggregated across cancer types, patients with gynecological (d=.29***), hematological (d=.15) and lung cancer (d=.22***) reported the highest levels of anxiety, whereas patients with skin (d=.23***) and prostate (d=.43***) cancer were less anxious than the average cancer patient (Fig.1). These differences were partially attributable to gender effects (d=.38***), indicating that female patients with gynecological, hematological, head and neck, and lung cancers report the highest levels of anxiety (Fig.3).

Fig.1.Levels of anxiety by cancer type.Figure options

Fig.2.Levels of depression by cancer type.Figure options

Fig.3.Levels of anxiety by cancer type and gender.Figure optionsThe visual display of age effects in Fig.5 was supplemented with inferential testing via computation of Pearson-Product moment correlations given that age is a continuous variable. Overall, age was not strongly linked to differences in distress levels, never explaining more than 2.5% of the variance. Nevertheless, younger patients (below 50) reported higher anxiety than older patients (70 and above). Within tumor groups, visual inspection of graphed data indicates that middle-aged patients with genitourinary cancers were most anxious. No age differences for anxiety were evident for lung, bone, skin and genitourinary cancers. Older age was on average associated with less anxiety (r=.15).

Fig.4.Levels of depression by cancer type and gender.Figure options

Fig.5.Levels of anxiety by cancer type and age.Figure options3.3. Prevalence rates of anxietyOn average, 19.0% of patients showed levels of anxiety in the clinical range, and another 22.6% reported subclinical symptoms (Fig.1). Female cancer patients were almost two times more likely than males (24.0% versus 12.9%) to report clinical levels of anxiety (Fig.3). Prevalence rates of clinical anxiety exceeding 30% were evident for females with hematological and lung cancers. In terms of age, prevalence rates varied across cancer types; they were 25.7% for the youngest, 18.9% for the middle, and 11.8% for the oldest age group (Fig.5). Within tumor groups, younger patients with bone cancer, breast cancer and prostate cancer were almost three times more likely to report clinically relevant anxiety than the oldest age group. Patients with gastrointestinal, hematological and neuroendocrine cancers were more than two times more likely to report clinical levels of anxiety than their older counterparts.3.4. Depression mean scoresAs shown in Fig.2, cancer patients on average experienced levels of depression below subclinical thresholds. Patients with lung (d=.21***), hematological (d=.20*) and gynecological (d=.18***) cancers were the most distressed, whereas patients with skin (d=.37***) and prostate (d=.18***) cancers were less depressed than the average cancer patient. Women scored higher than men (d=.33***; Fig.4). Women with genitourinary, hematological, and lung cancers reported levels of depression at or above the subclinical threshold.In terms of age differences across cancer types, there emerged again a linear but weak relationship such that younger cancer patients experienced the highest and older cancer patients the lowest levels of depressive symptoms across all cancers when aggregated (r=.12), but visual inspection suggested that patients with genitourinary, gynecological, and hematological cancer types were the most depressed when they fell into the middle-aged group (Fig.6). Middle-aged patients with unspecified primary tumors seemed less depressed than both their older and younger counterparts.

Fig.6.Levels of depression by cancer type and age.Figure options3.5. Prevalence rates of depressionOn average, 12.9% of patients showed levels of depression in the clinical range. Another 16.5% reported subclinical symptoms. Similarly, female cancer patients were almost two times more likely than males (16.4% versus 8.6%) to report clinical levels of depression. Female patients with lung cancer demonstrated the highest prevalence rate of depression with 24.7% being clinically depressed, followed by hematological (23.2%), and bone cancers (19.4%). Regarding age, younger age was associated with higher rates of depression but within lung cancer no age differences emerged. Again, for genitourinary, gynecological, and hematological cancers an inverse U-shaped relationship emerged with the middle-aged group being the most depressed.4. DiscussionThe present study adds a large sample of patients with all types of cancers to the extant literature on the prevalence of anxiety and depression but is the first to provide representative data on anxiety and depressive symptoms for all cancer types at the same time point in the disease trajectory, namely after diagnosis but prior to treatment. The sample is larger than the samples in either one of three earlier meta-analyses (Mitchell et al., 2011, Singer et al., 2010andvan't Spijker et al., 1997), or another large systematic review (Ng et al., 2010). Comparison of observed prevalence rates between the current study and previous studies calls for cautious interpretation because no other study using all cancer types had assessed patients at a standard point in the cancer trajectory. On the other hand, the prevalence rates reported here do fit well into the range of prevalence rates reported elsewhere (Brintzenhofe-Szoc et al., 2009, Caminiti et al., 2004, Carlson et al., 2004, Hinz et al., 2010, Hopwood et al., 2009, Mitchell et al., 2011, Ng et al., 2010andZabora et al., 2001) indicating that roughly 19% and 13% respectively met the threshold for a level of anxiety or depression within the clinical range.Also of interest is that mean levels of anxiety and depression were on average not much greater than those of a healthy population sample (Linden et al., 2009), but that they varied considerably as a function of cancer type, gender, and age. This implies that there are distinct risk groups covering a full range of prevalence rates from essentially population norm for prostate and skin cancer to majority depressed or anxious for lung and hematological cancer. In contrast to van't Spijker et al.'s meta-analytic findings (1997), where women had lower rates of emotional distress, our results showed that across all cancer types female cancer patients showed higher prevalence rates of anxiety and depression than men. This finding is consistent with higher rates of anxiety and depression in the general healthy female population as compared to men (Piccinelli and Wilkinson, 2000). This gender difference may reflect a gender difference in willingness to report distress but could also arise because women tend to use emotional approach coping (Goldzweig et al., 2009, Jacobs-Lawson et al., 2010andStanton et al., 2000). Our data suggest that men initially experience cancer as less threatening and this may arise in part from the high prevalence of prostate cancer which indeed has a good overall prognosis. Given the timing of our participant recruitment, we do not, however, know how men respond when cancer progresses and a prognosis may worsen.Interestingly, a clear inverse relationship between emotional distress and age was seen. Prevalence rates of anxiety and depression were higher in the youngest age group and lowest in older adults, likely due to more disruption of everyday living in younger cancer patients, whereas older patients may already have impairments in physical function and are cognitively and emotionally better prepared to accept illness. However, for a number of cancer types no age effect emerged suggesting that cancers with unfavorable prognosis (i.e., gynecological, hematological, lung) affect all age groups equally. This, in turn, may reflect a ceiling effect.Which factors account for divergent levels of emotional distress between cancer groups? The highest levels of emotional distress were reported from women patients with lung and hematological cancer and this confirms results of smaller studies (Castelli et al., 2009andNron et al., 2007), which showed very high prevalence rates of depression. Advanced stage, poor prognosis, and invasive treatment, therefore, are plausible sources for elevated levels of emotional distress (Vodermaier et al., 2011). To illustrate the range of varying prognoses, 5-year survival rates are 25% for lung cancer and 57% for leukemia in British Columbia (www.bccancer.bc.ca).What do psychological symptoms at the time point of cancer diagnosis represent? For most patients they are a natural reaction to what the patient perceives to be great uncertainty and thus a severe stressor. In others, the larger emotional reaction may be superimposed upon a latent vulnerability, e.g., preexisting mental disorders. Also, anxiety may be more of a reaction to acute events and is likely to decrease after completion of primary treatment (Thomas et al., 2011) once patients have become familiar with the side effects of treatment and also may have received positive prognostic information. Depressive symptoms on the other hand appear to be more stable and indicative of a trait-type disposition (Ando et al., 2009, Den Oudsten et al., 2010andKangas et al., 2007).Study strengths are very large sample, and separately reported prevalence rates of reporting of symptoms versus more severe range. The sample is representative of a multi-ethnic, urban population served by a publicly funded, equal-access health care system. Finally, the data represent consecutive recruitment of patients conducted at the same time point in the disease trajectory whereas other large-scale studies that include different cancer types only report on convenience samples measured at varying points in the course of cancer. Furthermore, this dataset has also allowed us to study the effect of disease stage on distress levels, and these data are already published elsewhere (Vodermaier et al., 2011).One of the study's strengths is at the same time also a limitation. The current data describe levels of emotional distress at the time of diagnosis and cannot speak to changes in emotional distress over the disease trajectory. Therefore, findings do not reflect treatment-induced emotional disturbance or emotional adjustment during survivorship or palliation. Lastly, the findings need to be seen in light of their representativeness for the entire cancer population.It is a strength that recruiting was based on routinely collected clinical data and consecutive recruiting and as such has no known major selection bias (except English language proficiency). Nevertheless, it is worth comparing this sample to its underlying population, and for this effort we were able to draw on recent data for British Columbia (www.bccancer.bc.ca) and for all of Canada (www.cancer.ca/Canada-wide/Cancer research/Cancer statistics.aspx?sc_lang=en). The rank ordering of most to least prevalent cancer types for our sample reflects those for the province and the whole country, however, our data set has a notably lower prevalence of lung cancer cases than does the province and the country (6% versus 13% and 14%, respectively) and also has a somewhat higher prevalence of prostate cancer cases (36% versus 25% and 27%). We attribute these deviations to regional differences in how care is provided and where records are kept. For a number of years now, the charts of deceased patients are moved to a warehouse and are kept by a private company contracted by the regional health authority. These charts are not directly accessible to researchers and, given the high mortality associated with lung cancer, our data likely under-represent lung cancer patients because the charts of those who have died were likely removed from the hospital-based archives. Hence, the prevalence rates in our sample differ for a few cancer types from provincial data, and this also accounts for the fact that our sample has slightly more women than men (55% versus 45%) whereas the entire population of cancer patients has a 48% to 52% ratio for women/men.Although PSSCAN has satisfactory psychometrics, its anxiety and depression subscales are each only 5 items long and the tool has not been validated using gold standard structured interviews. We cannot rule out potential confounding of anxiety or depression with medical problems and/or pharmacological treatment side effects (Holland and Alici, 2010). Similarly, patient anxiety and depression need to be understood in the context of the patient's knowledge about the disease, treatment, and probable outcomes as well as the quality of the patients' social network. We also need to sensitize readers to the fact that the data reflect the population of one limited geographical area in which all patients receive a similar quality of universal, third-party paid health care. Cancer may be perceived as an even greater threat when patients worry about medical bills that may exceed their resources, or have reason to fear loss of a job if their country of residence does not provide illness leaves and disability pensions.The present study offers a clear picture of absolute prevalence rates of emotional distress and it identifies distinct at-risk groups. Armed with these data, policy-makers can now use a data-driven approach to allocate staffing resources to psychosocial care. Fortunately, many cancer clinics already offer psycho-oncological services to which vulnerable patients can be referred, but just because a referral was made does not mean that all patients will accept this service offer. Also, patients who are not considered emotionally distressed as defined by cutoffs on the PSSCAN (or any other sensitive screening tool) may still ask for psychosocial support and this raises the thorny ethical question of whether or not, given generally scarce resources, these lower-risk groups should be offered professional help on demand.For cost-effective implementation of screening and treatment, some questions remain unresolved. What kind and how much treatment do subclinical levels of anxiety and depression require compared to clinical anxiety and depressive disorder? Do we need repeated screenings to assure that persistent anxiety and depression problems are treated? Should anxiety treatments differ as a function of what patients are specifically afraid of? They may have, for example, a longstanding, generalized anxiety disorder, or classically conditioned phobic anxiety responses to aversive cancer treatment, or may have high levels of fear of recurrence.Role of funding sourceSupported by CIHR Team for Supportive Cancer Care (#AQC83559).Conflict of interestThere are no conflicts of interest or financial interests associated with this work.AcknowledgmentsWe greatly appreciate the technical assistance of Colleen Wong, Leanne Fichtner, Tina May, and Sylvie Roy.References1. Ando et al., 2009 N. Ando, Y. Iwamitsu, M. Kuranami, S. Okazaki, M. Wada, K. Yamamoto, et al. Psychological characteristics and subjective symptoms as determinants of psychological distress in patients prior to breast cancer diagnosis Supportive Care in Cancer, 17 (2009), pp. 1361370 [SD-008]2. Brintzenhofe-Szoc et al., 2009 K.M. Brintzenhofe-Szoc, T.T. Levin, Y. Li, D.W. Kissane, J.R. Zabora Mixed anxiety/depression symptoms in a large cancer cohort: prevalence by cancer type Psychosomatics, 50 (2009), pp. 383391 [SD-008]3. Bultz and Carlson, 2005 B.D. Bultz, L.E. Carlson Emotional distress: the sixth vital sign in cancer care Journal of Clinical Oncology, 23 (2005), pp. 64406441 [SD-008]4. Caminiti et al., 2004 C. Caminiti, F. Campione, C. Sivelli, F. Diodati, R. Passalacqua Prevalence, predictors and risk factors of anxiety and depression in cancer patients A Nation-Wide Survey, J. Clin. Oncol. ASCO Annual Meeting Proceedings (Post-Meeting Edition), 22 (2004), p. 8121 [SD-008]5. Carlson et al., 2004 L.E. Carlson, M. Angen, J. Cullum, E. Goodey, J. Koopmans, L. Lamont, et al. High levels of untreated distress and fatigue in cancer patients British Journal of Cancer, 90 (2004), pp. 22972304 [SD-008]6. Castelli et al., 2009 L. Castelli, L. Binaschi, P. Caldera, R. Torta Depression in lung cancer patients: is the HADS an effective screening tool? Supportive Care Cancer, 17 (2009), pp. 11291132 [SD-008]7. Den Oudsten et al., 2010 B.L. Den Oudsten, G.L. Van Heck, A.F. Van der Steeg, J.A. Roukema, J. De Vries Clinical factors are not the best predictors of quality of sexual life and sexual functioning in women with early stage breast cancer Psycho-Oncology, 19 (2010), pp. 646656 [SD-008]8. DiMatteo et al., 2000 M.R. DiMatteo, H.S. Lepper, T.W. Croghan Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherence Archives of Internal Medicine, 160 (2000), pp. 21012107 [SD-008]9. Goldzweig et al., 2009 G. Goldzweig, E. Andritsch, A. Hubert, N. Walach, S. Perry, B. Brenner, et al. How relevant is marital status and gender variables in coping with colorectal cancer? A sample of middle-aged and older cancer survivors Psycho-Oncology, 18 (2009), pp. 866874 [SD-008]10. Greer et al., 2008 J.A. Greer, W.F. Pirl, E.R. Park, T.J. Lynch, J.S. Temel Behavioral and psychological predictors of chemotherapy adherence in patients with advanced non-small cell lung cancer Journal of Psychosomatic Research, 65 (2008), pp. 549552 [SD-008]11. Hinz et al., 2010 A. Hinz, O. Krauss, J.P. Hauss, M. Hckel, R.D. Kortmann, J.U. Stolzenburg, et al. Anxiety and depression in cancer patients compared with the general population European Journal of Cancer Care, 19 (2010), pp. 522529 [SD-008]12. Holland and Alici, 2010 J.C. Holland, Y. Alici Management of distress in cancer patients The Journal of Supportive Oncology, 8 (2010), pp. 412 [SD-008]13. Hopwood et al., 2009 P. Hopwood, G. Sumo, J. Mills, J. Haviland, J.M. Bliss The course of anxiety and depression over 5years of follow-up and risk factors in women with early breast cancer: Results from the UK Standardisation of Radiotherapy Trials (START) Breast, 19 (2009), pp. 8491 [SD-008]14. Jacobs-Lawson et al., 2010 J.M. Jacobs-Lawson, M.M. Schumacher, T. Hughes, S. Arnold Gender differences in psychosocial responses to lung cancer Gender Medicine, 7 (2010), pp. 137148 [SD-008]15. Kangas et al., 2007 M. Kangas, J.L. Henry, R.A. Bryant Correlates of acute stress disorder in cancer patients Journal of Traumatic Stress, 20 (2007), pp. 325334 [SD-008]16. Kathol et al., 1990 R. Kathol, A. Mutgi, J. Williams, G. Clamon, R. Noyes Jr. Diagnosis of major depression according to four sets of criteria The American Journal of Psychiatry, 147 (1990), pp. 10211024 [SD-008]17. Linden et al., 2005 W. Linden, D. Yi, M.C. Barroetavena, R. MacKenzie, R. Doll Development and validation of a psychosocial screening instrument for cancer Health and Quality of Life Outcomes, 3 (2005), pp. 5460 [SD-008]18. Linden et al., 2009 W. Linden, A. Vodermaier, R. MacKenzie, M.C. Barroetavena, D. Yi, R. Doll The Psychosocial Screen for Cancer (PSSCAN): further validation and normative data Health and Quality of Life Outcomes, 7 (2009), pp. 1623 [SD-008]19. Massie, 2004 M.J. Massie Prevalence of depression in patients with cancer Journal of the National Cancer Institute, 32 (2004), pp. 5771 [SD-008]20. Mitchell et al., 2011 A.J. Mitchell, M. Chan, H. Bhatti, M. Halton, L. Grassi, C. Johansen, et al. Prevalence of depression, anxiety, and adjustment disorder in oncological, haematological, and palliative-care setting: a meta-analysis of 94 interview based studies The Lancet Oncology, 12 (2011), pp. 160174 [SD-008]21. Nron et al., 2007 S. Nron, J.A. Correa, E. Dajczman, G. Kasymjanova, H. Kreisman, D. Small Screening for depressive symptoms in patients with unresectable lung cancer Supportive Care Cancer, 15 (2007), pp. 12071212 [SD-008]22. Ng et al., 2010 C.G. Ng, M.P.M. Boks, N.Z. Zainal, N.J. deWit The prevalence and pharmacotherapy of depression in cancer patients Journal of Affective Disorders, 13 (2010), pp. 17 [SD-008]23. Piccinelli and Wilkinson, 2000 M. Piccinelli, G. Wilkinson Gender differences in depression: critical review The British Journal of Psychiatry, 177 (2000), pp. 486492 [SD-008]24. Pinquart and Duberstein, 2010 M. Pinquart, P.R. Duberstein Depression and cancer mortality: a meta-analysis Psychological Medicine, 20 (2010), pp. 114 [SD-008]25. Satin et al., 2009 J.R. Satin, W. Linden, M.J. Phillips Depression as a predictor of disease progression and mortality in cancer patients: a meta-analysis Cancer, 115 (2009), pp. 53495361 [SD-008]26. Sharpe et al., 2004 M. Sharpe, V. Strong, K. Allen, R. Rush, K. Postma, A. Tulloh, et al. Major depression in outpatients attending a regional cancer centre: screening and unmet treatment needs British Journal of Cancer, 90 (2004), pp. 314320 [SD-008]27. Singer et al., 2010 S. Singer, J. Das-Munshi, E. Brhler Prevalence of mental health conditions in cancer patients in acute carea meta-analysis Annals of Oncology, 21 (2010), pp. 925930 [SD-008]28. Stanton et al., 2000 A.L. Stanton, S. Danoff-Burg, C.L. Cameron, M. Bishop, C.A. Collins, S.B. Kirk Emotionally expressive coping predicts psychological and physical adjustment to breast cancer Journal of Consulting and Clinical Psychology, 68 (2000), pp. 875882 [SD-008]29. Stommel et al., 2004 M. Stommel, M.E. Kurtz, J.C. Kurtz, C.W. Given, B.A. Given A longitudinal analysis of the course of depressive symptomatology in geriatric patients with cancer of the breast, colon, lung, or prostate Health Psychology, 23 (2004), pp. 564573 [SD-008]30. Thomas et al., 2011 B.C. Thomas, V. Nandamohan, M.K. Nair, M. Pandey Gender, age and surgery as a treatment modality leads to higher distress in patients with cancer Supportive Care in Cancer, 19 (2011), pp. 239250 [SD-008]31. van't Spijker et al., 1997 A. van't Spijker, R.W. Trijsburg, H.J. Duivenvoorden Psychological sequelae of cancer diagnosis: a meta-analytical review of 58 studies after 1980 Psychosomatic Medicine, 59 (1997), pp. 280293 [SD-008]32. Vodermaier et al., 2009 A. Vodermaier, W. Linden, C. Siu Screening for emotional distress in cancer care. A systematic review of assessment instruments Journal of the National Cancer Institute, 101 (2009), pp. 14641488 [SD-008]33. Vodermaier et al., 2011 A. Vodermaier, W. Linden, R. McKenzie, D. Greig, C. Marshall Post-diagnosis anxiety and depression: contribution of disease stage and cancer site in four common types of cancer British Journal of Cancer, 105 (2011), pp. 18141817 [SD-008]34. Zabora et al., 2001 J. Zabora, K. Brintzenhofe-Szoc, B. Curbow, C. Hooker, S. Piantadosi The prevalence of psychological distress by cancer site Psycho-Oncology, 10 (2001), pp. 1928 [SD-008]35. Zigmond and Snaith, 1983 A.S. Zigmond, R.P. Snaith The hospital anxiety and depression scale Acta Psychiatrica Scandinavica, 67 (1983), pp. 361370 [SD-008]Previous presentations: presented in part at the World Congress for Psycho-Oncology, Quebec City, Canada, May 2629, 2010.Corresponding authors: University of British Columbia, Department of Psychology, 2136 West Mall, Vancouver BC, Canada, V6T 1Z4. Tel.: +1 604 822 4156; fax: +1 604 822 6923.1Both Dr. Linden and Dr. Vodermaier share first authorship.