earch 64 (2008) 383–391

Journal of Psychosomatic Res

Psychological comorbidity and health-related quality of life and itsassociation with awareness, utilization, and need for psychosocial supportin a cancer register-based sample of long-term breast cancer survivors

Anja Mehnert⁎, Uwe Koch

Department of Medical Psychology, University Medical Center Hamburg-Eppendorf, Hamburg, Germany

Received 8 September 2007; received in revised form 3 December 2007; accepted 11 December 2007

Abstract

Objective: Psychosocial comorbidity and quality of life (QOL)and its association with knowledge, utilization, and need forpsychosocial support have been studied in long-term breast cancersurvivors. Methods: One thousand eighty-three patients wererecruited through a population-based cancer registry an average of47 months following diagnosis (66% response rate). Self-reportmeasures (e.g., Hospital Anxiety and Depression Scale, Posttrau-matic Stress Disorder Checklist—Civilian Version, and Short-FormHealth Survey) were used. Results: Thirty-eight percent of patientshad moderate to high anxiety, and 22% had moderate to highdepression; posttraumatic stress disorder was observed in 12%. Theoverall psychological comorbidity was 43% and 26% for a possibleand probable psychiatric disorder. Disease progress, detrimentalinteractions, less social support, a lower educational level, andyounger age were predictors of psychological comorbidity(Pb.004). Lower QOL (Pb.01) and higher levels of anxiety

⁎ Corresponding author. Department of Medical Psychology, UniversityMedical Center Hamburg-Eppendorf, Martinistr. 52-S35, 20246 Hamburg,Germany. Tel.: +49 40 42803 6203; fax: +49 40 42803 4940.

E-mail address: mehnert@uke.uni-hamburg.de (A. Mehnert).

0022-3999/07/$ – see front matter © 2008 Elsevier Inc. All rights reserved.doi:10.1016/j.jpsychores.2007.12.005

(Pb.001) were observed in cancer survivors compared to age-adjusted normative comparison groups. Time since diagnosis had nosignificant impact on psychological comorbidity as well as QOL.Forty-six percent of women felt insufficiently informed aboutsupport offers. Insufficient knowledge was associated with older ageand lower education (Pb.05). Since diagnosis, 57% had participatedin cancer rehabilitation and 24% in other psychosocial supportprograms. Fifteen percent of all patients and 23% of those with apossible psychiatric disorder expressed their need for psychosocialsupport. Women with distress and perceived support needs who didnot participate in past support programs were older, less educated,and less informed (Pb.05). Conclusion: Findings show the long-term impact of breast cancer and indicate need for patient education,screening for psychosocial distress, and implementation of psycho-logical interventions tailored in particular for older women.© 2008 Elsevier Inc. All rights reserved.

Keywords: Anxiety; Depression; Long-term breast cancer survivors; Psychosocial distress; Psychosocial support; PTSD; Quality of life

Introduction

Breast cancer is the most common cancer among womenin Germany as well as in developed countries with anestimated annual incidence of 1,152,161 new cases world-wide [1,2]. The increase of breast cancer incidence with ageand the improvement in survival rates through use ofenhanced treatment regimes within recent years lead to a

growing number of older patients and long-term survivors[2,3]. The consequences that evoke from this developmentdo not merely include only physical health components suchas a variety of late side effects during follow-up care [4], butalso a wide range of potentially persistent psychosocialproblems and needs [5,6].

Over the past decades, a considerable amount of researchhas shown the significant emotional and social impact ofbreast cancer and its treatment on patients and their families[6]. Previous studies using self-report screening measureshave reported distress, anxiety, and depression in, on average,one quarter to one third of breast cancer patients with levelsup to 50% following diagnosis and treatment [6–12]. While

384 A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

high levels of psychosocial distress have been found withinthe first year postdiagnosis, there is evidence for animprovement in distress, psychosocial well being, and qualityof life (QOL) in long-term survivors [13–15], even thoughfew research has shown no improvement or even a decline inphysical and/or mental health (MH), specifically in olderwomen [6,11,16,17]. Risk factors for psychosocial distressand lower levels of QOL include fewer personal resourcesandmaladaptive coping strategies, poorly controlled physicalsymptoms, lack of social support, psychiatric medical history,poor doctor–patient communication, lower educational level,as well as younger or intermediate age [6,18,19]. However,although younger women seem to have higher levels ofpsychosocial distress and report more unmet needs [20], theywere also found to be more aware of psychosocialinterventions and more likely to participate in support offersthan older women [21].

Research has shown that older women [22–24] and thosewith lower socioeconomic status [25] are less likely toreceive standard treatment and information about breastcancer care. Whether this might also be the case forpsychosocial care has not been widely investigated so far.A study in breast cancer patients aged from 55 to 97 years[26] found that older women do highly value the informationabout breast cancer care provided by their breast cancerphysician. However, while younger women did also receivemore written information from other sources such as primarycare physicians, family, and friends, specifically, womenaged 57 years or older were found to get less writteninformation from other sources. Furthermore, women whorated their ability to communicate with their physicians lesshighly were found to have lower levels of emotional health,suggesting that, in particular, patients with substantial needfor support are less likely to actively seek help. However,although this study focused on oncological care information,it can be assumed that information about psychosocial careand support might also be less given to older women.

Knowledge about psychosocial support offers is animportant requirement for their use. Eakin and Strycker[27] led a study in 2001 on the knowledge about andparticipation in psychosocial support offers among patientswith breast, colon, and prostate cancer. The findings indicatedthat only a percentage of patients were informed of existingsupport offers, and that knowledge of these offers was notconsequently accompanied by participation in them. Theauthors were able to show that, in addition to a higher degreeof education, a referral by the physician played an importantrole for the utilization of psychosocial support offers.

While many psychosocial and QOL studies have beenconducted with breast cancer patients, few studies haveconcentrated specifically on the association between psy-chological comorbidity and the awareness and utilization ofpsychosocial support offers in long-term survivors. Hodg-kinson et al. [28] found a high percentage (86%) of reportedunmet needs in breast cancer survivors and a significantlyhigher percentage in those patients with high levels of

anxiety or depression. However, research including olderwomen has been relatively rare so far.

The purpose of this cross-sectional cancer registry-basedstudy is to provide information about the prevalence ofanxiety, depression, and posttraumatic stress disorder(PTSD), as well as about health-related QOL in long-termbreast cancer survivors. Group comparisons with population-based normative values for anxiety, depression, and QOLhave been undertaken. We furthermore investigate perceivedknowledge, utilization, and subjective need for psychosocialsupport. The study examines predictors of psychologicalcomorbidity as well as predictors of psychosocial supportneeds, taking into account specifically the time sincebreast cancer diagnosis, age, and educational level. Wehypothesize that psychological comorbidity and awarenessof and need for psychosocial support are significantlyassociated with age, educational level, time since diagnosis,and disease stage.

Patients and methods

Study design and participants

Patients older than 18 years with a clinical diagnosis ofbreast cancer confirmed between 6.5 years and 18 monthsprior to our survey were recruited through the population-based Cancer Registry of the Federal German state ofHamburg. The period of 18 months postdiagnosis was theearliest feasible time due to the cancer registration procedure.The Hamburg Cancer Registry has a case completeness forbreast cancer higher than 95% [29]; however, registrationand participation in research projects are voluntary forpatients. The protocol received local research ethicscommittee approval. Written informed consent was obtainedfrom all patients.

For each year postdiagnosis (18–24 months, up to 36, 48,60, and more than 60 months to less than 78 months), thedata sets of 400 women were randomly selected by thecancer registry. The upper time limit of 78 monthspostdiagnosis was determined by the cancer registry due tothe fact that several studies in breast cancer patients usingself-report measures were conducted at the same time as ourresearch. The limit was defined to avoid possible distress inpatients caused through repeated requests for study partici-pation. The data sets contained name, address by date ofregistration, age, month and year of cancer diagnosis of eachpatient, as well as complete TNM status and histopatholo-gical grading for the majority of patients (77.1%). A set ofself-report questionnaires, consent forms, and a self-addressed stamped envelope were mailed to a total of 2000patients. A reminder letter was sent after 4 weeks. Amongthese women, 321 (16.1%) had moved to an unknownaddress and did not receive our set of questionnaires, 39(2.0%) had recently died, and 7 (0.4%) were not able toparticipate due to severe cognitive or physical problems

385A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

(e.g., dementia, actively dying) noted by their relatives. Ofthe remaining eligible 1633 patients, evaluable question-naires were returned by 1083 patients (66.3%). Participants(1083) and nonparticipants (550) did not differ in termsof age (mean=61.8±9.8 vs. mean=58.9±10.6) (P=.46),months since initial diagnosis (mean=46.5±17.5 vs. mean=46.9±17.0) (P=.66), and tumor stage (P=.51). Sociodemo-graphic and medical characteristics of the participants arelisted in Table 1.

Measures

Sociodemographic and cancer treatment-related informa-tion was obtained using a standardized self-report question-

Table 1Demographic and medical characteristics of the participants (n=1083)

n %

Mean age in years±S.D. (range) 61.8±9.8 (31–81)Marital statusMarried 661 61.0Never married 102 9.4Divorced 161 14.9Widowed 159 14.7Partnership 725 66.9

Educational levelLess than elementary school 76 7.0Elementary school 445 41.1Junior high school 375 34.6High school/university degree 187 17.2

Employment statusRetired 691 63.9Employed 233 21.5Housewife 114 10.5Unemployed/other 45 4.1

Disease phaseRemission 942 87.0Recurrence/progress 141 13.0

Disease stage at diagnosis (TNM)I 392 46.9IIA/B 372 44.6IIIA/B 57 6.8IV 14 1.7(Missing information) (248) (22.9)Mean months since initial

diagnosis±S.D. (range)46.5±17.5 (18–77)

18–24 months 151 13.9Up to 36 months 217 20.0Up to 48 months 225 20.8Up to 60 months 194 17.9More than 60 months after diagnosis 296 27.3

Type of surgery and primary treatmentBreast-conserving surgery 704 65.0Mastectomy 379 35.0Chemotherapy 118 10.9Radiotherapy 444 41.0Both 371 34.3

naire. Years and months of breast cancer diagnosis as well asTNM classification were received through the cancer registry.

The Hospital Anxiety and Depression Scale (HADS)—German Version is a validated screening instrument foranxiety and depression in somatically ill patients [30].According to the German Manual of the HADS, a score of0 to 7 is categorized as normal, a score of 8 to 10 is consideredto indicate a possible anxiety or depressive disorder, and ascore of 11 or above is considered to indicate a probableanxiety or depressive disorder. Representative population-based standardized values for women at different age groupswere computed through regression analyses as recommendedby Hinz and Schwarz [31]. The following equations wereapplied: 0.026×age+0.69×1(female gender)+3.09 for anxietyand 0.079×age+0.15×1(female gender)+0.73 for depression.

The Posttraumatic Stress Disorder Checklist—CivilianVersion (PCL-C) [32,33] assesses the PTSD symptomcriteria intrusion, avoidance, and hyperarousal according tothe Diagnostic and Statistical Manual of Mental Disorders,Fourth Edition. Participants estimate the degree, to whichthey have been concerned, of each symptom during the pastmonth. The term “distressing event” was replaced by theterm “cancer or cancer treatment.” Participants are consid-ered likely to have a PTSD if they meet at least one intrusion,three avoidance, and two arousal symptoms (rated as“moderately” or above).

The Short-Form Health Survey [34] assesses dimensionsof QOL: general health (GH), physical functioning (PF), rolephysical (RP), bodily pain (BP), vitality (VT), socialfunctioning (SF), MH, role emotional (RE), and two summaryscores for physical (PCS) andmental health (MCS). Each itemscore or summary measurement ranges from 0 to 100, higherscores indicating better health. Representative population-based reference data for women at different age groups forGermany were published by Ellert et al. [35].

Social support was assessed using the 24-item validatedGerman version of the Illness-specific Social Support Scale[36], which measures positive support (15 items) anddetrimental interaction (9 items). Items are scored on a5-point Likert scale ranging from 1 (never) to 5 (always).

Perceived need for and utilization of professionalpsychosocial support offers were assessed by single-itemquestions using Likert or categorical answer scales.

Statistical analysis

Statistical analyses were performed using the StatisticalPackage for the Social Sciences (SPSS; Chicago, IL, USA)version 12.0. Associations between variables were calcu-lated using Pearson Product–Moment Correlation or Spear-man's correlation coefficient. Uni- and multivariate analysesof variance (ANOVA/MANOVA) were conducted for groupcomparisons in metric data. Group differences werecalculated using t tests in metric and χ2 tests in categoricalvariables. Logistic regression analyses utilizing forwardconditional method were carried out. To provide an estimate

386 A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

of the magnitude of the group differences, we calculatedCohen's standardized effect size (ρ, d, η2). The valuesρ=0.1, d=0.2, and η2=0.009 represent the cutoff for a smalleffect size; ρ=0.3, d=0.5, and η2=0.0588 represent the cutofffor a moderate effect size; and ρ=0.5, d=0.8, and η2=0.1379represent the cutoff for a large effect size. Two-tailedsignificance tests were conducted using a significance levelof P≤.05.

Results

Psychological comorbidity and QOL

Eighteen percent of the total samples were classified ashaving moderate anxiety and 20% as having high levels of

Table 2Anxiety, depression, PTSD, and QOL at different months postcancer diagnosis (n

Totalsample(n=1083)

Time postdiagnosis

18–24 months(n=151)

Up36 m(n=2

n (%) n (%) n (%

AnxietyMean (S.D.) 6.66 (4.5) 6.70 (4.3) 6.4Normal 669 (61.8) 92 (60.9) 12Moderate 196 (18.1) 28 (18.5) 4High 218 (20.1) 31 (20.5) 3

DepressionMean (S.D.) 4.58 (4.2) 4.42 (4.1) 4.4Normal 843 (77.8) 121 (80.1) 17Moderate 118 (10.9) 12 (7.9) 2High 122 (11.3) 18 (11.9) 2

PTSDIntrusion 397 (36.7) 59 (39.1) 8Avoidance 232 (21.4) 38 (25.2) 4Hyperarousal 360 (33.2) 60 (39.7) 7PTSD diagnosis 131 (12.1) 20 (13.2) 2

TotalPTSD, anxiety and/or depression(HADS cutoff ≥8)

465 (42.9) 64 (42.4) 9

PTSD, anxiety and/or depression(HADS cutoff ≥11)

286 (26.4) 42 (27.8) 5

QOLPF 42.76 (9.1) 42.51 (9.1) 42.2RP 40.95 (9.4) 40.18 (9.3) 40.7BP 47.73 (10.0) 47.43 (9.4) 47.5GH 45.33 (6.8) 45.53 (6.9) 45.8VT 50.28 (8.8) 49.60 (9.6) 50.3SF 47.12 (9.0) 47.46 (8.8) 46.8RE 43.47 (8.3) 42.93 (8.6) 43.9MH 44.85 (9.6) 44.35 (9.6) 45.0PCS 43.39 (10.2) 42.93 (10.0) 43.1MCS 46.84 (10.8) 46.37 (11.0) 47.2

⁎ P (MANOVA).⁎⁎ P (χ2).

anxiety. Eleven percent of the patients were classified ashaving moderate depression and further 11% as havinghigh levels of depression. Cancer-related PTSD wasobserved in 12% of the patients. No group differenceswere found in the frequency of anxiety, depression, as wellas PTSD between patients at different months postdiagno-sis. Furthermore, ANOVA for QOL between the differenttime groups gave nonsignificant results for all QOLdimensions (Table 2).

With regard to the overall psychological comorbidity, wefound 43% of the sample having at least one psychologicaldiagnosis (anxiety and/or depression and/or PTSD) whenusing a HADS cutoff ≥8 and 26% when using a HADScutoff ≥11 (Table 2). Again, using HADS cutoff ≥8, 21.2%of the patients had one diagnosis, 13.9% had two diagnoses,and 7.8% had three diagnoses. Using HADS cutoff ≥11,

=1083)

P

toonths17)

Up to48 months(n=225)

Up to60 months(n=194)

More than60 months(n=296)

) n (%) n (%) n (%)

3 (4.3) 6.68 (4.6) 6.77 (4.5) 6.71 (4.6) .94 ⁎

9 (59.4) 147 (65.3) 116 (59.8) 185 (62.5) .62 ⁎⁎

9 (22.6) 33 (14.7) 38 (19.6) 48 (16.2)9 (18.0) 45 (20.0) 40 (20.6) 63 (21.3)

6 (4.0) 4.67 (4.3) 4.61 (4.1) 4.66 (4.3) .53 ⁎

6 (81.1) 172 (76.4) 146 (75.3) 228 (77.0) .65 ⁎⁎

1 (9.7) 28 (12.4) 27 (13.9) 30 (10.1)0 (9.2) 25 (11.1) 21 (10.8) 38 (12.8)

1 (37.3) 88 (39.1) 74 (38.1) 95 (32.1) .43 ⁎⁎

1 (18.9) 50 (22.2) 40 (20.6) 63 (21.3) .69 ⁎⁎

3 (33.6) 73 (32.4) 60 (30.9) 94 (31.8) .45 ⁎⁎

1 (9.7) 27 (12.0) 25 (12.9) 38 (12.8) .77 ⁎⁎

5 (43.8) 88 (39.1) 92 (47.4) 126 (42.6) .55 ⁎⁎

1 (23.5) 56 (24.9) 52 (26.8) 85 (28.7) .70 ⁎⁎

5 (8.7) 43.38 (9.4) 42.46 (9.5) 42.98 (9.0) .69 ⁎

4 (9.0) 41.70 (9.6) 40.68 (9.7) 41.11 (9.4) .59 ⁎

5 (10.3) 48.32 (10.3) 48.03 (9.8) 47.40 (10.2) .83 ⁎

7 (6.2) 45.55 (6.7) 44.95 (6.6) 44.92 (7.3) .49 ⁎

7 (8.8) 50.63 (8.4) 50.10 (8.7) 50.39 (8.9) .84 ⁎

1 (9.0) 46.87 (9.2) 47.05 (9.1) 47.43 (8.9) .91 ⁎

9 (8.0) 43.44 (8.5) 43.40 (8.5) 43.41 (8.2) .82 ⁎

1 (9.3) 44.62 (10.1) 45.06 (9.4) 45.02 (9.8) .94 ⁎

5 (9.8) 44.29 (10.3) 43.10 (10.5) 43.30 (10.2) .67 ⁎

7 (10.6) 46.45 (11.1) 46.99 (10.5) 46.96 (10.9) .91 ⁎

Table 3Predictors for psychological comorbidity (n=835)

β S.E. Odds ratio95% confidenceinterval P

Education −0.25 0.09 0.78 0.65–0.94 ≤.008Disease stage 1.08 0.25 3.00 1.83–4.79 ≤.001Social support −0.69 0.12 0.50 0.40–0.63 ≤.001Detrimental

interactions0.93 0.15 2.53 1.90–3.39 ≤.001

Age −0.02 0.01 0.98 0.96–1.00 ≤.005

387A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

13.9% of the sample had one diagnosis, 7.8% had twodiagnoses, and 4.7% had three diagnoses.

The sample was divided into three age groups: (a) up to50 years (n=153), (b) 51 to 65 years (n=519), and (c) 66years and older (n=411). Analysis of variance for anxiety,depression, and PTSD symptoms between these three agegroups gave significant results. Older patients were found tohave less anxiety (P=.004) (η2=.01), less intrusion (P=.05)(η2=.01), less avoidance (P=.04) (η2=.01), and lesshyperarousal (P=.001) (η2=.01), but higher levels ofdepression (P=.004) (η2=.01) than the younger patients.

Predictors of psychological comorbidity

We hypothesized that overall psychological comorbidity(anxiety/depression cutoff ≥8, PTSD) (n=465) would bepredicted by age, level of school education, disease stage andcancer progress, social support, and detrimental interactions,as well a time since initial breast cancer diagnosis. Theresults of the logistic regression analysis utilizing forwardconditional method are shown in Table 3. Disease progress, ahigher level of detrimental interactions, a lower level ofschool educational level, less social support, and youngerage were significant predictors of psychological comorbidity(Nagelkerke's R2=.23).

Anxiety, depression, and QOL in comparison to representa-tive population-based standardized values for women atdifferent age groups

Table 4 displays anxiety and depression in the breastcancer sample compared to representative population-based

Table 4Anxiety and depression in breast cancer patients compared to representative popu

Age group

Anxiety

Breast cancer patients Female norm populati

Mean (S.D.) Mean (S.D.) P

Up to 50 years (n=153) 7.22 (4.31) 4.87 (0.13) ≤.0051–65 years (n=519) 6.94 (4.69) 5.24 (0.12) ≤.0066 years and older (n=411) 6.09 (4.19) 5.54 (0.11) ≤.00Total (n=1083) 6.66 (4.47) 5.30 (0.25) ≤.00

⁎ P (t test).⁎⁎ Effect size d (mean difference/S.D.).

standardized values for women at different age groups [31].We found breast cancer patients at all age groups atsignificant higher levels of anxiety compared to theequivalent group values for the German normative sample.In contrast, women of the middle and older age groups weresignificantly less depressed than women of the normativesample, whereas no significant result was found for thegroup of younger breast cancer patients compared to theirnormative age cohort.

The comparison of the breast cancer sample with therepresentative data for German women in QOL [35] showsthe following results. In both the youngest and the middleage groups, significant lower levels of QOL could beobserved in breast cancer patients in PF, RP, SF, RE, MH, aswell as physical and mental QOL (Pb.001). No differenceswere found in BP, GH, and VT. Older cancer patients over65 years had significant lower levels in all QOL dimensions(Pb.01), except for GH and VT.

Perceived knowledge about professional psychosocialsupport

Table 5 shows the distribution of perceived knowledgeabout professional psychosocial support offers. A total ofn=502 patients (46.3%) feels inadequately informed aboutsupport offers. There was no statistically significantdifference in perceived knowledge between the youngerand the middle age groups. However, patients older than65 years reported being significantly less informed. Besidesyounger age, higher perceived knowledge about psychoso-cial support was associated with a higher level of education(P=.001) (η2=.04), less depression (P=.01) (η2=.01), andbetter social support (P=.01) (η2=.02). No associations werefound with anxiety, PTSD symptoms, and time sincediagnosis as well as QOL.

Participation in psychosocial support offers

Since their initial cancer diagnosis, 57% (n=619) of thewomen had participated in a cancer rehabilitation program,which is basically provided to cancer patients in Germany toregain physical and psychosocial functioning. Besidesphysical exercises, services comprise patient education,

lation-based standardized values for women at different age groups

Depression

on Breast cancer patients Female norm population

d Mean (S.D.) Mean (S.D.) P d

1 1.10 4.02 (3.59) 4.18 (0.40) ≤.59 –1 0.71 4.33 (4.11) 5.32 (0.36) ≤.001 0.448 0.26 5.11 (4.11) 6.23 (0.34) ≤.001 0.501 0.58 4.58 (4.40) 5.50 (0.77) ≤.001 0.36

Table 5Perceived knowledge about professional psychosocial support (n=1083)

Perceivedknowledge

Age groups

P ρ

Total(n=1083)

Up to 50 years(n=153)

51–65 years(n=519)

66 years and older(n=411)

n (%) n (%) n (%) n (%)

Not at all informed 258 (23.8) 25 (16.3) 101 (19.5) 132 (32.1) ≤.001 0.17A little bit informed 244 (22.5) 39 (25.5) 114 (22.0) 91 (22.1)Somewhat informed 250 (23.1) 38 (24.8) 125 (24.1) 87 (21.2)Quite a bit informed 260 (24.0) 40 (26.1) 138 (26.6) 82 (20.0)Very much informed 71 (6.6) 11 (7.2) 41 (7.9) 19 (4.6)

⁎ P (χ2).⁎⁎ Effect size ρ.

388 A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

relaxation training, and psychosocial support groups, as wellas individual psychotherapy. Apart from cancer rehabilita-tion, a total of 24% (n=258) had participated in additionalsupport programs: 16% (n=174) had received ambulatoryprofessionally led support such as psychotherapy orpsychosocial counseling, 7% (n=79) had attended a self-help cancer support group, 4% (n=42) had receivedpsychosocial counseling during hospital admission, and3% (n=33) had received support through a cancer supportinformation centre (multiple responses). Thirty-eight percentof the women (n=408) had not participated in anyrehabilitation or psychosocial support program. Womenthat participated in rehabilitation and/or psychosocial sup-port offers were younger (Pb.001) (d=.35), were morefrequently single or divorced (Pb.001) (ρ=0.17), and morefrequently reported a progress of the disease (P=.02)(ρ=0.07). No group differences were observed in terms oftime since diagnosis (P=.84), tumor stage (P=.21), and levelof education (P=.12). However, women that participated inpsychosocial support offers apart from cancer rehabilitationalso had a significant higher level of education (Pb.001)(ρ=0.22).

Predictors of perceived need for psychosocial support

Fifteen percent of all patients (n=165) expressed theirneed for psychosocial support. To determine predictors forperceived need for psychosocial support (n=165), weincluded age, level of education, disease stage, socialsupport, time since breast cancer diagnosis, informationabout support offers, past participation in a psychosocialsupport program, psychological comorbidity, and physical as

Table 6Predictors for perceived need for psychosocial support (n=835)

β S.E.Oddsratio

95%confidenceinterval P

Information aboutsupport offers

−0.36 0.09 0.70 0.59–0.83 ≤.001

MCS −0.06 0.01 0.94 0.93–0.96 ≤.001

well as MH scores in a logistic regression analysis. A lowerMH status and a lower level of information were observed asthe only two significant predictors (Nagelkerke's R2=.14)(Table 6).

However, of those women classified as having at least onepossible anxiety, depressive, or PTSD (n=465), 23% (n=107)expressed the subjective need for support. In contrast, 9%(n=58) of those patients with no psychological comorbidityhad the subjective need for support (Pb.001) (ρ=0.19).

Women with significant levels of distress and perceivedneed for support (n=107) differ significantly in the pastutilization of support offers except cancer rehabilitation.Those patients who did not participate in additional supportprograms were significantly older (mean=62.8±9.3 vs.mean=57.8±11.3) (P=.03) (d=0.49), less educated (P=.04)(ρ=0.24), and less informed about psychosocial supportprograms (Pb.001) (ρ=0.44).

Discussion

This cancer registry-based study examined the prevalenceof anxiety, depression, and PTSD as well as health-relatedQOL in long-term breast cancer survivors. The secondobjective was to investigate perceived knowledge, utiliza-tion, and subjective need for professional psychosocialsupport offers.

The term long-term survivor has been used differentlywithin the current literature [6]. According to the NationalCancer Institute in the United States [3], “an individual isconsidered a cancer survivor from the time of cancerdiagnosis through the balance of his or her life” [37].Other definitions determine the term long-term survivorshipat a minimum of 5 years postdiagnosis [3]. Other authorssuch as Mullan [38] and Bloom [39] have suggested thatsurvivorship can be divided into different temporal stagessuch as (a) acute survival, including the first year afterdiagnosis and primary cancer treatment, (b) extendedsurvival, which includes the period from the first throughthe third years after treatment when recurrence is most likely,and (c) permanent survival, which extends from the thirdyear throughout the life and includes the longer-term

389A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

adjustment period. Although all women in our samplewere at least 18 months postdiagnosis and had completedtreatment primary cancer treatment, according to thelatter definitions, our sample represents the majority ofwomen in the acute and extended survival phase, whereas27% of women can be seen as 5-year postdiagnosispermanent survivors.

Overall, high levels up to 38% for moderate to severeanxiety and up to 22% for moderate to severe depressionhave been found independently from time since initial breastcancer diagnosis. Previous studies in breast cancer survivorsassessing long-term psychological comorbidity havereported noticeable lower rates of moderate and high levelsof anxiety or depression [10,11,28], although similar orhigher rates have been following diagnosis and primarybreast cancer treatment [6,40,41]. The prevalence rates foundhere are twice as high as the rates observed in a recent studyby Hodgkinson et al. [28]. Also, the total number of womensuffering from at least one possible or probable psychiatricdisorder (anxiety and/or depression and/or PTSD) can beconsidered as remarkably high with 43% and 26%. As theused measures (HADS and PCL-C) are sensitive self-reportquestionnaires with very good psychometric properties,these high prevalence rates indicate an insufficient profes-sional psychosocial support system for long-term breastcancer patients in Germany. In particular, anxiety, fear ofprogression, and cancer relapse are not well investigated andclinically addressed in cancer patients and seem to be apersistent burden [42], possibly leading to depression andimpaired QOL.

The comparison with population-based data revealssignificant higher levels of anxiety in all patient age groups.However, the effect size is specifically large in youngerwomen. These results provide evidence that the experienceof breast cancer has a long-lasting impact on anxiety. Inparticular, fear of progression and cancer-related worriesmight be salient problems that have been found as major andpersistent sources of stress in cancer patients [42,43].Interestingly, similar findings have been publicized byHodgkinson et al. [28], showing higher anxiety scores inlong-term breast cancer survivors compared to communitylevels, but no differences in depression and QOL. In ourstudy, rates of depression were even significantly better inbreast cancer patients older than 51 years compared to age-adjusted population-based prevalence rates. Thus, breastcancer appears to induce higher levels of anxiety but couldconcurrently lead to a redefinition and reappraisal of lifeperspectives and life goals that might be a buffer againstfeelings of depression. The estimation of PTSD was 12%, anumber that corresponds with the 1% to 20% estimatesreported in the empirical literature [12,44] but is slightlyhigher than the PTSD lifetime prevalence rates of women inthe general population [45].

Consistent with the literature [6], we found cancerprogress, detrimental social interactions, and lower levelsof social support, as well as lower educational levels, to be

significant predictors for overall psychological comorbid-ity. Surprisingly, age was found to be a significant butrather weak predictor in this study. Furthermore, time sinceinitial diagnosis did not play a significant role inprevalence rates of psychological comorbidity, which wasalso observed in previous research [46]. An explanation forthe apparent lack of differences in psychological comor-bidity in women with rather large time differences sincediagnosis is that anxiety is a predominant feature followingcancer diagnosis and treatment that appears to be ratherstable over time with regard to the enduring threat of apossible disease recurrence. The specification of subtypesof anxiety in long-term survivors such as panic disorder,general anxiety disorder, or fear of progression requiresfurther investigation.

Breast cancer survivors did report impaired QOL in themajority of dimensions, which is only partly consistent withother findings in breast cancer survivors. Whereas themajority of studies reported improvement in QOL in breastcancer survivors [6,14,15], several research did not find anincrease in physical or mental QOL domains [16,17]. Onlymarginal differences were found between younger and olderage groups. However, it is noteworthy that a consistent lowerQOL in breast cancer survivors was found even years afterdiagnosis and treatment.

In our study, a relatively high percentage of breast cancerpatients (62%) participated in a cancer rehabilitationprogram provided in Germany to regain physical andpsychosocial functioning, and/or received other forms ofpsychosocial support. The findings emphasize the impor-tance of cancer rehabilitation following diagnosis andprimary cancer treatment. However, the percentage ofwomen that does participate in other psychosocial supportoffers or self-help organizations is much lower (24%).

Rather surprising was the relatively high rate (46%) ofwomen that feel insufficiently informed about psychosocialsupport offers, because Hamburg is a city with acomparatively well functioning network of psychosocialsupport offers for cancer patients, including inpatientsconsultation and liaison services in cancer centers andmajor hospitals, psychosocial ambulatory services, as wellas several cancer support information centers. In particular,patients older than 65 years and patients with lowereducational levels are less informed. Insofar, our resultsprovide emphasis to previous research that indicates that thequality of patient–physician interaction and informationgiving declines in older patients and in patients with lowersocioeconomic status [23,47].

In contrast to other research [28], we found a rather low(15%) percentage of perceived need for psychosocialsupport. Only 23% of those patients with a possiblepsychiatric disorder had the subjective desire for profes-sional support. Although it was not assessed in this study,due to other research [27], it can be assumed that referral byprimary care physician or oncologist may be very importantfor the participation in support offers. In particular, older

390 A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

women and women with lower school education mightbenefit from a valid psychosocial screening and referral.

There are several important limitations of the study. Theresponse rate of 66% of eligible patients can be evaluated asrather positive given a cancer register-based survey where nopersonal clinical contact between patients and our institution(Department of Medical Psychology) was established before.Participants and nonparticipants did not differ in importantvariables such as age, time since diagnosis, and tumor stage.Nevertheless, there remains a certain level of uncertaintyregarding the representativeness and a possible sample bias.In the view of 34% of nonparticipants, our findings mayover- or underestimate the prevalence of psychologicalcomorbidity and the levels of QOL in breast cancersurvivors. As this is a cross-sectional study, importantinformation about pre- or posttreatment psychologicaladjustment and QOL is missing to specifically determinethe course of psychological morbidity and QOL in breastcancer survivors. Furthermore, perceived knowledge andneed for psychosocial support were assessed using single-item questions. The use of comprehensive questionnaireswould add substantial information here. Future studies mayneed to take account of breast cancer survivors five and moreyears postdiagnosis. However, despite these limitations,results of this study may have important implications forpatient education and information, screening for psychoso-cial distress, the promotion and implementation of psycho-logical interventions tailored in particular for older women,as well as further psychosocial research.

Acknowledgments

This research has been supported by grants from the CoraLobscheid Foundation and the Astrid-Haugstrup-Sörensen-Memorial Foundation within the Donors' Association for thePromotion of Sciences and Humanities in Essen, Nordrhein-Westfalen, Germany (grant numbers T 129-12.171 and T100-13.496). We thank Dr. Stefan Hentschel and NorbertBurkhard at the Hamburg Cancer Registry for their valuablesupport to this work.

References

[1] Gesellschaft der epidemiologischen Krebsregister in Deutschland e.V.und das Robert Koch Institut. Krebs in Deutschland. 5. überarbeitete,aktualisierte Ausgabe. Saarbrücken, 2006.

[2] Kamangar F, Dores GM, Anderson WF. Patterns of cancer incidence,mortality, and prevalence across five continents: defining priorities toreduce cancer disparities in different geographic regions of the world.J Clin Oncol 2006;24:2137–50.

[3] Brenner H. Long-term survival rates of cancer patients achieved bythe end of the 20th century: a period analysis. Lancet 2002;360:1131–5.

[4] Ganz PA. Breast cancer, menopause, and long-term survivorship:critical issues for the 21st century. Am J Med 2005;118:136–41.

[5] Hewitt M, Herdman R, Holland J. Meeting psychosocial needs ofwomen with breast cancer. Washington (DC): The National AcademiesPress; National Cancer Policy Board, 2004.

[6] Knobf MT. Psychosocial responses in breast cancer survivors. SeminOncol Nurs 2007;23:71–83.

[7] Aapro M, Cull A. Depression in breast cancer patients: the need fortreatment. Ann Oncol 1999;10:627–36.

[8] Zabora J, BrintzenhofeSzoc K, Curbow B, Hooker C, Piantadosi S.The prevalence of psychological distress by cancer site. Psychooncol-ogy 2001;10:19–28.

[9] Morasso G, Costantini M, Viterbori P, et al. Predicting mood disordersin breast cancer patients. Eur J Cancer 2001;37:216–23.

[10] Osborne RH, Elsworth GR, Hopper JL. Age-specific norms anddeterminants of anxiety and depression in 731 women with breastcancer recruited through a population-based cancer registry. Eur JCancer 2003;39:755–62.

[11] Burgess C, Cornelius V, Love S, Graham J, Richards M, Ramirez A.Depression and anxiety in women with early breast cancer: five yearobservational cohort study. BMJ 2005;330:702–6.

[12] Mehnert A, Koch U. Prevalence of acute and post-traumatic stressdisorder and comorbid mental disorders in breast cancer patientsduring primary cancer care: a prospective study. Psychooncology2007;16:181–8.

[13] Shim EJ, Mehnert A, Koyama A, et al. Health-related quality of life inbreast cancer: a cross-cultural survey of German, Japanese, and SouthKorean patients. Breast Cancer Res Treat 2006;99:341–50.

[14] Ganz PA, Desmond KA, Leedham B, et al. Quality of life in long-term,disease free survivors of breast cancer: a follow-up study. J Natl CancerInst 2002;94:39–49.

[15] Helgeson VS, Tomich PL. Surviving cancer: comparison of 5-yeardisease-free breast cancer survivors with healthy women. Psychoon-cology 2005;14:307–17.

[16] Ganz PA, Guadagnoli E, Landrum MB, Lash TL, Rakowski W,Silliman RA. Breast cancer in older women: quality of life andpsychosocial adjustment in the 15 months after diagnosis. J Clin Oncol2003;21:4027–33.

[17] Helgeson VS, Snyder P, Seltman H. Psychological and physicaladjustment to breast cancer over 4 years: identifying distincttrajectories of change. Health Psychol 2004;23:3–15.

[18] Baider L, Andritsch E, Uziely B, et al. Effects of age on coping andpsychological distress in women diagnosed with breast cancer: reviewof literature and analysis of two different geographical settings. CritRev Oncol Hematol 2003;46:5–16.

[19] Clough-Gorr KM, Ganz PA, Silliman RA. Older breast cancersurvivors: factors associated with change in emotional well-being.J Clin Oncol 2007;25:1334–40.

[20] Thewes B, Butow P, Girgis A, Pendlebury S. The psychosocial needsof breast cancer survivors; a qualitative study of the shared and uniqueneeds of younger versus older survivors. Psychooncology 2004;13:177–89.

[21] Plass A, Koch U. Participation of oncological outpatients inpsychosocial support. Psychooncology 2001;10:511–20.

[22] Newschaffer CJ, Penberthy L, Desch CE, Retchin SM, WhittemoreM. The effect of age and comorbidity in the treatment of elderlywomen with nonmetastatic breast cancer. Arch Intern Med 1996;156:85–90.

[23] Silliman RA, Troyan SL, Guadagnoli E, Kaplan SH, Greenfield S. Theimpact of age, marital status, and physician patient interactions on thecare of older women with breast cancer. Cancer 1997;80:1326–34.

[24] Enger SM, Thwin SS, Buist DS, et al. Breast cancer treatment of olderwomen in integrated health care settings. J Clin Oncol 2006;24:4377–83.

[25] Macleod U, Ross S, Fallowfield L, Watt GC. Anxiety and support inbreast cancer: is this different for affluent and deprived women? Aquestionnaire study. Br J Cancer 2004;91:879–83.

[26] Silliman RA, Dukes KA, Sullivan LM, Kaplan SH. Breast cancer carein older women: sources of information, social support, and emotionalhealth outcomes. Cancer 1998;83:706–11.

[27] Eakin EG, Strycker LA. Awareness and barriers to use of cancersupport and information resources by HMO patients with breast,

391A. Mehnert, U. Koch / Journal of Psychosomatic Research 64 (2008) 383–391

prostate, or colon cancer: patient and provider perspectives.Psychooncology 2001;10:103–13.

[28] Hodgkinson K, Butow P, Hunt GE, Pendlebury S, Hobbs KM,Wain G.Breast cancer survivors' supportive care needs 2–10 years afterdiagnosis. Support Care Cancer 2007;15:515–23.

[29] Hamburger Krebsdokumentation 2002–2004. Neuerkrankungen undSterbefälle. Freie und Hansestadt Hamburg, Behörde für Soziales,Familie, Gesundheit und Verbraucherschutz (Ed), 2007.

[30] Herrmann C, Buss U, Snaith RP. HADS-D—Hospital Anxiety andDepression Scale—Deutsche Version: Ein Fragebogen zur Erfassungvon Angst und Depressivitaet in der somatischen Medizin. Bern:Huber, 1995.

[31] Hinz A, Schwarz R. Anxiety and depression in the general population:normal values in the Hospital Anxiety and Depression Scale.Psychother Psychosom Med Psychol 2001;51:193–200.

[32] Weathers FW, Huska JA, Keane TM. PCL-C for DSM-IV. Boston:National Center for PTSD—Behavioral Science Division, 1991.

[33] Teegen F. Deutsche Ü bersetzung der Posttraumatic Stress DisorderChecklist (PCL-C) des National Center for PTSD. Hamburg:Universitaet Psychologisches Institut III, 1997.

[34] Ware JE, Kosinski M, Dewey JE, Gandek B. How to score andinterpret single-item health status measures: a manual for users of theSF-8 TM Health Survey. Lincoln, RI: QualityMetric Inc, 1999.

[35] Ellert U, Lampert T, Ravens-Sieberer U. Measuring health-relatedquality of life with the SF-8. Normal sample of the German population.Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz2005;48:1330–7.

[36] Ramm GC, Hasenbring M. Die deutsche Adaptation des Illness-specific Social Support Scale und ihre teststatistische Überprüfungbeim Einsatz an Patienten vor und nach Knochenmarktransplantation.Z Med Psychol 2003;12:29–38.

[37] National Institutes of Health websites (http://dccps.nci.nih.gov/ocs/)[38] Mullan F. Seasons of survival: reflections of a physician with cancer.

N Engl J Med 1985;313:270–3.[39] Bloom JR. Surviving and thriving. Psychooncology 2002;11:

89–92.[40] Bender CM, Ergun FS, Rosenzweig MQ, et al. Symptom clusters in

breast cancer across 3 phases of the disease. Cancer Nurs 2005;28:219–25.

[41] Deshields T, Tibbs T, Fan M, et al. Differences in patterns ofdepression after treatment for breast cancer. Psychooncology 2006;15:398–406.

[42] Herschbach P, Berg P, Dankert A, et al. Fear of progression in chronicdiseases: psychometric properties of the Fear of ProgressionQuestionnaire. J Psychosom Res 2005;58:505–11.

[43] Deimling GT, Bowman KF, Sterns S, Wagner LJ, Kahana B. Cancer-related health worries and psychological distress among older adult,long-term cancer survivors. Psychooncology 2006;15:306–20.

[44] Kangas M, Henry JL, Bryant RA. Posttraumatic stress disorderfollowing cancer. A conceptual and empirical review. Clin Psychol Rev2002;22:499–524.

[45] Frans O, Rimmo PA, Aberg L, Fredrikson M. Trauma exposure andpost-traumatic stress disorder in the general population. Acta PsychiatrScand 2005;111:291–300.

[46] Osborne RH, Elsworth GR, Sprangers MA, Oort FJ, Hopper JL. Thevalue of the Hospital Anxiety and Depression Scale (HADS) forcomparing women with early onset breast cancer with population-based reference women. Qual Life Res 2004;13:191–206.

[47] Siminoff LA, Graham GC, Gordon NH. Cancer communicationpatterns and the influence of patient characteristics: disparities ininformation-giving and affective behaviors. Patient Educ Couns 2006;62:355–60.