Preventing anxiety and depression in gynaecological cancer: a randomised controlled trial

BJOG: an International Journal of Obstetrics and Gynaecology April 2002, Vol. 109, pp. 386 –394 Preventing anxiety and depression in gynaecological cancer: a randomised controlled trial Rodney W. Petersen, Julie A. Quinlivan* Objective To examine the effect of counselling and relaxation intervention on psychological symptoms in patients with gynaecological cancer between the post-operative period and the six-week review. Design Randomised controlled trial. Participants Fifty-three patients with gynaecological cancer. Setting Three Australian tertiary referral hospitals. Methods Fifty-three patients were randomised to control or intervention and completed the baseline Hospital Anxiety and Depression Scale (HADS) and General Health Questionnaire-28 (GHQ-28) questionnaires. The intervention consisted of a relaxation and counselling session performed by a senior doctor. Follow up questionnaires were completed at six weeks. Demographic and tumour data were collated independently. Results Complete data were available on 50 patients. There were no significant differences in demographic, social support or tumour characteristics between the two groups. Multivariate analysis determined that only the intervention and baseline score were significant predictors of outcome. The intervention was associated with a significant reduction in total HADS score ( P ¼ 0.002). The reduction was seen in both anxiety and moderate depression subscales ( P ¼ 0.001 and P ¼ 0.02). The intervention was also associated with a significant reduction in total GHQ-28 score and in three of the four subscale scores (somatisation, anxiety and personality development; all P < 0.02). However, no significant difference was found in the fourth subscale of major depression. Conclusion A relaxation and counselling intervention performed by a treating doctor reduces psychological symptoms in women with a new diagnosis of gynaecological cancer. INTRODUCTION Evidence suggests that women confronted with a new diagnosis of gynaecological cancer experience a period of crisis1. This phase is most marked immediately following diagnosis and gradually dissipates over a period of 100 days 2. Prevalence studies have estimated that in this period 47% –70% of women experience psychological symptoms of sufficient degree to merit a diagnosis of moderate or severe depression or anxiety 3,4. Failure to address psychological symptoms results in immediate morbidity to the patient. As newer treatment modalities improve survival, quality of life issues take on increasing importance for survivors1. Furthermore, unresolved psychological morbidity may have an adverse effect on survival. The psycho-neuroimmunological literature provides strong evidence that psychological factors can have an impact on the function of the immune system and the regulation of host defences; these factors can have an impact upon disease progression and possibly on survival5,6. Despite the immediate morbidity and the possibility for adverse longer term outcomes, only two randomised trials have assessed the impact of strategies designed to reduce or alleviate psychological distress in women with gynaecological cancer7,8. The two trials have provided conflicting evidence about the efficacy of such interventions. Formal assessment of any psychological intervention is important, as there is evidence that ‘debriefing’ strategies can be associated with a significant increase in adverse outcomes compared with control 9 – 11. We have assessed the impact of a one-hour relaxation and counselling intervention upon changes in psychological symptoms between the immediate post-operative period and the six-week review. The intervention was performed by a treating doctor to ensure that the counselling component of the interview would be able to provide accurate information to the patient about her specific disease state, in addition to emotional support and care. Department of Obstetrics and Gynaecology, The University of Melbourne, Victoria, Australia METHODS * Correspondence: Dr J. Quinlivan, The University of Melbourne, University Department of Obstetrics and Gynaecology, Royal Women’s Hospital, 132 Grattan St, Carlton, Victoria, Australia 3053. Multi-institutional ethics committee approval and informed patient consent were obtained. Patients were recruited between April and November 2000 from three D RCOG 2002 BJOG: an International Journal of Obstetrics and Gynaecology PII: S 1 4 7 0 - 0 3 2 8 ( 0 2 ) 0 1 2 7 1 - 5 www.bjog-elsevier.com PREVENTING ANXIETY IN GYNAECOLOGICAL CANCER 387 Table 1. Relaxation exercises in the structured interview. Patients were instructed through the exercise series and were then given a written copy of the exercise to take home. tertiary hospitals providing gynaecological oncology services. The planned trial population included women with a new diagnosis of gynaecological cancer who were Englishspeaking, did not have a previous psychiatric history and who had undergone surgery as their primary form of treatment. Patients were provided with information sheets in the post-operative period and were invited to participate in the study. Once consent was obtained, baseline questionnaires were completed before randomisation. Patients were randomised either to intervention or control groups by selecting from a block of ten numbered sealed opaque envelopes. Baseline questionnaires included the Hospital Anxiety and Depression Scale (HADS) 12 and the General Health Questionnaire-28 (GHQ-28)13. The patient also completed questions relating to their level of social support, and of the time periods between the onset of symptoms, presentation D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394 for investigation, diagnosis and surgical treatment. Questionnaires were completed independently by the patient without input from staff. Demographic data and tumour codings were collated independently after discharge. If patients were allocated to the intervention group, the session took place within 24 hours of discharge from hospital. The intervention consisted of a one-hour relaxation and counselling interview performed by a senior medical practitioner involved in the patient’s care. The structured interview consisted of listening to a tape of relaxation music for approximately five minutes, followed by a series of relaxation exercises that lasted 20 minutes. Patients were provided with an instruction sheet outlining the exercises and were encouraged to perform the relaxation measures at home if they felt stressed or anxious (Table 1). The remaining 30 to 35 minutes were spent discussing the patient’s condition. Patients were encouraged to ask 388 R.W. PETERSEN & J.A. QUINLIVAN specific questions relating to their disease, their emotional state and their treatment. The interview was performed in a quiet room and every effort was made to avoid interruptions during the counselling process. To ensure the absence of disturbance, pagers were switched off, ward staff were notified that an intervention was taking place, and a ‘do not disturb’ notice was put on the door of the room. Patients in the control and intervention groups were provided with all normal post-operative support, counselling and information services provided by the respective hospitals and their multi-disciplinary care teams. In the immediate post-operative period, before recruitment into the study, patients received a daily ward round visit from the resident doctor of the gynaecological oncology unit (Monday to Friday) or from rotation within the combined gynaecological specialties (Saturday to Sunday). Senior medical staff attended rounds variably according to the protocols at the respective hospitals. After discharge, patients did not receive contact with medical staff unless indicated on a clinical ground. Six weeks following surgery, at the time of the first medical outpatient review, subjects completed follow up questionnaires (HADS and GHQ-28), which were distributed in envelopes by clinic staff, who were blinded to treatment allocation. Patients who received the intervention were asked to provide short comments of their impression of the session. Main outcome measures The primary outcome was the difference between baseline and follow up HADS score12,14. The HADS has been utilised in over 200 clinical studies since it was developed in 1983 and is designed for use in physically ill patients. The HADS provides clinically meaningful results as a psychological screening tool, in clinical group comparisons and in correlational studies with several aspects of disease and quality of life. It is sensitive to course of time changes and to the assessment of response to psychotherapeutic and psycho-pharmacological intervention. Internal consistencies (Cronbach alphas) are high for the anxiety and depression subscales at 0.80– 0.93 and 0.81 –0.90, respectively. Re-test reliability is high (r > 0.8) and sensitivity and specificity has always exceeded 0.80 in 17 English version studies14. The HADS has been used extensively in cancer patients15,16. Secondary outcomes included the anxiety and depression subscales of the HADS questionnaire, the GHQ-28 total and four subscale scores and the patient’s overall impression of the intervention. probably disorders for anxiety and depression. In healthy controls (pooled studies of over 2500 subjects), 90% will score under 8, with the majority scoring under 4. In contrast, in psychiatric patient populations, over 78% will score 8 or higher14. As previously published prevalence figures suggested that 50% of gynaecological oncology patients had measurable psychological morbidity, we estimated that the median HADS score at baseline would be at least 83,4. A clinically significant change was deemed to be a change in median HADS score from 8 (possible disorder) to 3 (no disorder) in the intervention group. In order to have greater than 80% power to detect at a two-sided 0.05 level, the change in HADS from 8 to 3, 25 patients were required in each arm. A two-sided 0.05 level was selected, as there have been reports of debriefing interventions resulting in adverse outcomes compared with control9 – 11. Analysis was performed on an intention-to-treat basis. Demographic and clinical data were compared between the two groups. Descriptive statistics were performed to categorise data. Continuous data were compared with the Mann – Whitney U test for non-parametric data and Student’s t test (two-tailed) or ANOVA for parametric data. Categorical variables were compared with the x2 test or Fisher’s exact probability test. In order to apportion the share of each factor in the determination of the outcome, multivariate analysis was performed using logistic regression with backward elimination, with P for exclusion set at 0.10. A P value of 0.05 was considered significant. RESULTS Figure 1 summarises the study protocol. Of 57 patients approached to participate in the study, 53 provided consent Statistical analysis In their original study, Zigmond and Snaith12 referred to cutoff values of 7/8 and 10/11 to indicate possible and Fig. 1. Patient recruitment and follow up. D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394 PREVENTING ANXIETY IN GYNAECOLOGICAL CANCER 389 Table 2. Patient demographics, social supports and surgical care. Continuous data are presented as mean and standard deviation or median and interquartile range. Discrete data are presented as n (%). Cohort demographics Control Intervention P Hospital No 1 No 2 No 3 11 (42) 10 (39) 5 (19) 12 (50) 4 (17) 8 (33) 0.11 61.2 [13.5] 63.0 [9.6] 0.59 Mean [SD] age (years) Median (IQR) patient perception of support level (score 0 to 5) 5 [4 – 5] 5 [4 – 5] 0.53 Anaesthetic coding (ASA score) 1 2 3 12 (50) 8 (33) 04 (17) 11 (42) 11 (42) 4 (16) 0.63 Surgical procedures undertaken* Transabdominal hysterectomy Bilateral salpingo-oophorectomy Unilateral salpingo-oophorectomy Modified or radical hysterectomy Pelvic lymph node dissection Omentectomy Large bowel resection Local radical excision vulva Vulvectomy Groin node dissection Exenterative surgery 16 16 2 5 15 11 1 2 1 2 0 18 18 0 4 18 6 0 0 1 1 1 (75) (75) (0) (17) (75) (25) (0) (0) (4) (4) (4) 0.51 122 [33] 112 [35] 0.30 Post-operative complications Yes No 8 (31) 18 (69) 9 (38) 15 (62) 0.43 Mean [SD] post-operative bedstay (days) 7.8 [2.0] 7.2 [3.6] 0.47 Adjuvant therapy required Yes No 19 (73) 7 (27) 14 (58) 10 (42) 0.07 Mean [SD] length of surgery (min) (62) (62) (8) (19) (58) (42) (4) (8) (4) (8) (0) * Sum > 100% as patients often had more than one surgical procedure. and were randomised into control (n ¼ 26) and intervention groups (n ¼ 27), respectively. The principle reason for declining participation in the study was fatigue and a reluctance to complete the questionnaires. In the intervention group, one patient withdrew from the study after completing the baseline questionnaires and before the intervention was undertaken; one patient withdrew after the intervention. In the control group, one patient moved interstate and was lost to follow up. Therefore, a total of 50 patients were available for analysis. Table 2 summarises the demographic data and surgical characteristics of the two groups. No significant differences were observed in the hospital recruitment patterns, patient age or pre-existing social supports. Of note, the study cohort reported high levels of social supports with median scores of 5 out of 5 in both groups. Pre-existing medical status, measured by ASA scores, were also similar between D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394 the two groups. Patients underwent a variety of surgical procedures, often multiple in nature. There were no significant differences in length of surgery, incidence of postoperative complications, post-operative bedstay or need for adjuvant therapy. Table 3 summarises tumour details and delays in presentation and treatment of the two groups. There were no significant differences in tumour site, stage or grade. The types of presenting symptoms were also similar between groups. The delays between the onset of symptoms and presentation for medical care or diagnosis were similar, as was the presentation to treatment interval. There were no significant differences in the time interval between completion of baseline and follow up questionnaires (control 47 days, intervention 51 days, P ¼ 0.21). Figures 2 and 3 summarise the baseline, follow up and difference in HADS and GHQ-28 scores in the two groups. 390 R.W. PETERSEN & J.A. QUINLIVAN Table 3. Patient presentation and tumour characteristics. Continuous data are presented as median and interquartile range. Discrete data are presented as n (%) of group. Cohort demographics Control Tumour type Ovarian Cervix Uterine Other 9 3 10 4 (35) (12) (38) (15) 5 4 13 2 (21) (17) (54) (8) 0.10 Tumour stage Stage 1 Stage 2 Stage 3 Stage 4 16 3 7 0 (61) (12) (27) (0) 14 4 6 0 (58) (17) (25) (0) 0.87 Tumour grade Well differentiated Moderately differentiated Poorly differentiated 14 (54) 6 (23) 6 (23) 8 (33) 7 (29) 9 (38) 0.12 Presenting symptom Postmenopausal bleeding Abdominal distension Routine screening (assymptomatic) Pelvic discomfort Post coital bleeding Intermenstrual bleeding Other 10 4 2 2 1 1 6 12 (50) 3 (12.5) 4 (17) 1 (4) 0 (0) 1 (4) 3 (12.5) 0.40 (38) (15) (8) (8) (4) (4) (23) Intervention P Median [IQR] time from onset of symptoms to medical presentation (days) 60 [30 – 98] 40 [28 – 65] 0.57 Median [IQR] time from onset of symptoms to diagnosis (days) 62 [42 – 134] 60 [36 – 76] 0.44 Median [IQR] time from medical presentation to treatment (days) 15 [14 – 21] 16 [14 – 21] 0.79 There were no significant differences in baseline HADS and GHQ-28 scores (HADS: control 9.72, intervention 12.79, P ¼ 0.15; GHQ-28: control 6.28, intervention 9.42, P ¼ 0.14). Logistic regression determined that tumour site was a significant predictor of the baseline score (r ¼ 1.6, P ¼ 0.04; Table 4). At follow up, HADS and GHQ-28 scores were significantly lower in the intervention group compared with control (HADS: control 7.96, intervention 3.58, P ¼ 0.003; GHQ-28: control 5.65, intervention 2.09, P ¼ 0.03). Table 5 summarises the univariate analysis of the primary outcome. The intervention was associated with a significant reduction in the total HADS score between baseline and follow up (control 2.04, intervention 9.17; P ¼ 0.002). In order to be complete, the analysis was repeated assessing the difference in HADS score on a case/noncase basis (discrete variable: case score 8, noncase score <8, x2) and a significant outcome was also observed ( P < 0.0001). Table 6 outlines the results of the regression analysis for the difference in HADS score. Despite the inclusion of Fig. 2. Baseline, follow up and differences in total HADS score in the two groups. Control and intervention are white and black, respectively. An asterisk indicates a significant difference between control and intervention. At follow up, scores in the intervention group were significantly lower compared with the control group; this translated into a significant reduction in score from baseline in the intervention, but not the control group. tumour stage, grade and type, patient age, social supports, and need for additional adjuvant therapy, only the intervention and baseline HADS score were significantly associated with outcome (intervention: r 5.1, P ¼ 0.008; baseline HADS: r 0.52, P < 0.0001). Thus, the intervention was associated with a significant reduction in total HADS score. A high baseline HADS score was also associated with a tendency to improve over time. Secondary outcomes are summarised in Table 5. Subscale analysis of the HADS demonstrated that the intervention was associated with a significant reduction in both anxiety and mild/moderate depression subscale scores compared with control (anxiety P ¼ 0.001; mild to moderate depression P ¼ 0.02). The intervention was also associated with a significant reduction in the total GHQ28 scores and in three of the four subscales, namely anxiety, Fig. 3. Baseline, follow up and differences in total GHQ-28 scores in the two groups. Control and intervention are white and black, respectively. An asterisk indicates a significant difference between control and intervention. At follow up, scores in the intervention group were significantly lower compared with the control group; this translated into a significant reduction in score from baseline in the intervention, but not the control group. D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394 PREVENTING ANXIETY IN GYNAECOLOGICAL CANCER Table 4. Baseline and follow-up HADS score by tumour site in the two groups. Data are median [interquartile range]. Groups Baseline Follow up Control Ovary (n ¼ 9) Cervix (n ¼ 3) Uterus (n ¼ 10) Other (n ¼ 4) 15 2 5 9 [7 – 17] [2 – 7] [3 – 15] [5 – 13] 12 1 4 7 [6 – 13] [0.5 – 6] [3 – 9.75] [3.5 – 10.75] Intervention Ovary (n ¼ 5) Cervix (n ¼ 4) Uterus (n ¼ 13) Other (n ¼ 2) 19 7.5 14 8.5 [16 – 21] [5.5 – 9.5] [10 – 16] [5 – 11] 3 4 3 1.5 [3 – 3] [3.5 – 4.25] [1 – 4] [1.25 – 1.75] Table 5. Differences in total and subscale scores of the HADS and GHQ28 in the intervention and control groups between the post-operative period and the six-week surgical review. Results are presented as mean, standard deviation, range and P values. Outcomes somatisation and personality development compared with control (GHQ-28 total P ¼ 0.001; anxiety P ¼ 0.0007 somatisation P ¼ 0.008, personality development P ¼ 0.02). However, there was no significant effect in the fourth subscale relating to major depressive symptoms. Patients who had undergone the intervention were asked to comment about it separately at the end of completing their questionnaires. The majority of patients who experienced the intervention stated that the experience was positive (positive 79%, neutral 14%, negative 7%, P < 0.0001). Of note, most felt that a treating doctor should perform the intervention, rather than another health professional (doctor 57%, no preference 29%, social worker 7%, blank 7%, nurse 0%, psychologist 0%; P < 0.0001). Patients made several specific positive comments regarding the intervention. These comments related to the themes of an appreciation of the opportunity to talk to a doctor without the pressure of time constraints (n ¼ 6), of the intervention making them feel important as an individual (n ¼ 2), of an appreciation at the effort that was taken to make them feel relaxed and comfortable before starting the session (n ¼ 2), of receiving specific information about their disease and options for care (n ¼ 1), and achieving motivation to succeed (n ¼ 1). Eight patients made comments about the beneficial impact of spending time with their doctor. The only negative comments were that the intervention was too short (n ¼ 1) and that the bed in the counselling room was too hard (n ¼ 1). DISCUSSION The key finding was that a one-hour relaxation and information session conducted by a treating doctor was associated with a significant reduction in psychological symptoms between the post-operative period and follow up at six weeks. Reductions were seen in all psychological subscales with the exception of severe depression. Patients with high baseline HADS scores demonstrated a tendency to improve over the six-week period. However, the intervention was additive to the time effect. D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394 391 HADS Difference in total score Mean (SD) Range P value Control Intervention 2.04 (7.9) 20 to 25 9.17 (7.2) 23 to 5 0.002 Difference in anxiety subscale score Mean (SD) 1.16 (4.6) Range 13 to 14 P value 6.04 (5.1) 15 to 6 0.001 Difference in depression subscale score Mean (SD) 0.72 (3.7) Range 8 to 11 P value 3.00 (2.9) 9 to 1 0.02 GHQ-28 Difference in total score Mean (SD) Range P value 7.17 (7.6) 24 to 2 0.001 0.64 (5.1) 18 to 12 Difference in anxiety subscale score Mean (SD) 0.2 (1.9) Range 6 to 5 P value 2.52 (2.4) 7 to 1 0.0007 Difference in somatisation and sleep disturbance subscale score Mean (SD) 0.12 (1.5) 2.00 (2.9) Range 4 to 3 7 to 5 P value 0.008 Difference in personality development subscale score Mean (SD) 0.28 (2.5) Range 6 to 7 P value 1.96 (2.4) 7 to 1 0.02 Difference in severe depression subscale score Mean (SD) 0.04 (1.0) Range 2 to 4 P value 0.70 (1.7) 5 to 2 0.08 The effect was independent of tumour stage, grade and type, patient age, social supports, and need for adjuvant therapy. As a result of randomisation, these variables were not significantly different between the two groups and therefore do not account for the differences in outcome. Furthermore, these demographics were not independent predictors of the primary outcome in the regression analysis. Table 6. Logistic regression: only the intervention and baseline HADS score were significant predictors of outcome. Variables Intervention Baseline HADS score Coefficient 5.13 0.53 t P 95% CI 2.79 0.008 1.43 to 8.83 4.51 <0.0001 0.76 to 0.29 392 R.W. PETERSEN & J.A. QUINLIVAN However, tumour site was a factor in determining the baseline, but not follow up or difference in score. Previously published prevalence figures suggest that 47% –70% of women experience psychological symptoms of sufficient degree to merit a diagnosis of moderate or severe depression or anxiety3,4. Our findings support these figures with median scores of 10– 13 in the study cohort at baseline. Scores above 10/11 are thought to indicate probable anxiety and depression disorders12. At follow up, the median score in the control group was 8, suggesting that 50% still had HADS scores suggestive of underlying morbidity. In contrast, median HADS score in the intervention group was 4, well inside the range seen in community controls14. Psychosocial factors interact with daily life. Individuals are distressed in different ways in response to situations that endanger them2. The term ‘existential plight’ has been used to describe the 100-day period that occurs immediately following the new diagnosis of cancer. The concept was first described in a longitudinal study of 120 newly diagnosed cancer patients who were repeatedly interviewed at four to six week intervals from the time of diagnosis until three to four months had elapsed. In the study, almost all patients stated that they experienced a phase in this time period in which they felt that their existence was endangered and that they were going to die 2. Since this time, data suggest that up to 85% of newly diagnosed cancer patients experience traumatic episodes of anxiety or depression17. In studies of gynaecological oncology patients, the data have suggested that up to 70% experience symptoms of anxiety or depression of sufficient severity to meet the criteria of a psychiatric disorder 3,4. In view of this widespread morbidity, it is surprising that relatively few controlled trials have assessed interventions designed to ameliorate these symptoms. In 1980 a prospective study of 56 patients with a new diagnosis of gynaecological cancer that underwent a counselling intervention, were compared with a non-randomised control group of 41 patients. Counselled patients reported less confusion and problems in self-perception and were more likely to return to work than controls. However because there were only 34 working women in the sample, this later outcome was not statistically significant18. In the two randomised controlled trials investigating psychological interventions, the outcomes were contradictory. In the first study, 80 women newly diagnosed with gynaecological cancer were randomised to standard group counselling (control, n ¼ 31), themed individual counselling (n ¼ 21) or themed group counselling (n ¼ 28). Only 60 patients completed follow up. Patients undergoing themed individual or grouped counselling were less depressed and anxious, had better relationships with carers and increased participation in leisure activities than controls who received standard group counselling7. In the second study, 37 patients with early gynaecological cancer were randomised to group therapy focussing on cancer problem solving, goal setting and relaxation, while controls were sent a booklet from the National Cancer Institute. The key findings were that immediately after completing the intervention, patients receiving counselling reported significantly higher levels of mood disturbance than controls. However, by six to 10 weeks and at 12 – 24 months, no significant differences existed between the two groups8. Of note, the study has been criticised because of the poor recruitment rate, with only 37 of 107 eligible patients agreeing to participate: it also had a high dropout rate, with only 15 subjects completing the long term follow up at 12 –24 months19. However, concern has been expressed that many counselling or ‘debriefing’ interventions are not effective in reducing psychological symptoms, or worse yet, may actually cause harm9 – 11. Males with testicular cancer have similar levels of anxiety and depression as female patients with gynaecological malignancies. However, a randomised trial to determine the efficacy of adjuvant psychological therapy in 73 patients with testicular cancer reported no difference in outcome measures between treatment groups. Of note, 111 of the 184 men approached to participate in the trial declined. Compared with participants, nonparticipants in the study had less anxious pre-occupation with their disease9. Adverse outcomes have been reported following debriefing. In a large randomised trial, after operative childbirth, women were randomised to receive a midwife-led debriefing and counselling intervention or control. The debriefing group reported higher levels of depression and of depression-related problems six months postpartum compared with controls, although significant differences in outcomes were only noted in the emotional role functioning subscale of the SF-3610. A recent Cochrane review of eight randomised trials found no evidence that debriefing had any effect on psychological morbidity. Based upon their findings, the authors recommended that compulsory debriefing should cease11. It has been suggested that ‘debriefing’ interventions may increase immediate feelings of anxiety because they serve to remind the patient of an unpleasant or traumatic event. However, patients with cancer are reminded of their disease every time they attend clinics for adjuvant or palliative chemotherapy, radiotherapy or general follow up care. Therefore, a possible detrimental impact of counselling serving to remind the patient of their disease may be less relevant. Indeed, research data in this field have been conflicting, with evidence supporting both an adverse and a protective role of early denial in relation to cancer patients1,2,20. It is also possible that debriefing interventions fail because they are not performed by an individual that is readily identified by the patients as an integral team member such as a treating doctor or primary care nurse. Patients may feel reluctant to ask wide ranging questions relating to their care. As such, a key ingredient in respect to the provision of patient specific information and answers may be missing from the psychological intervention. There is good evidence that information can be beneficial in relieving anxiety 21. D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394 PREVENTING ANXIETY IN GYNAECOLOGICAL CANCER In designing an intervention for the present study, efforts were made to distinguish clearly the intervention from a debriefing activity. The intervention focus was to create a relaxed and intimate environment, with an adequate time allotment in which to explore issues and provide accurate information to the patient about matters of personal concern. In view of this aim, it was necessary for the intervention to be performed by individuals with sufficient expertise to be able to answer any possible questions about risks, side effects, outcomes and prognosis, in addition to issues concerning the psychological adjustment to cancer. It was felt that this was most easily achieved by the involvement of senior medical staff who had also undertaken basic training in counselling, relaxation and listening skills, rather than an alternative member of the multidisciplinary care team of nursing, social worker and psychologist staff. This enabled patients to address a wide variety of concerns relating to their disease without limiting the discussion to psychological aspects of care. Patient feedback indicated that the intervention’s aims were achieved. Patient’s comments focused on the beneficial impact of spending time with a doctor in a context without the pressure of time constraints and which provided an opportunity for the patient to become relaxed and comfortable before the discussion. The issues discussed in the intervention session were wide-ranging and included social, work, and interpersonal problems through to issues of death, prognosis, follow up, alternative treatments and the rationale for various treatments. In light of this, it was perhaps not surprising that a majority of patients stated that they felt that a doctor should perform the intervention. However, it is also possible that, having only had experience of a doctor performing the counselling intervention, patients chose this option. Experience with another health professional in a similar setting might lead to an alternative preference. The findings reinforce the age-old impression that patients value spending time with their doctor. As time constraints increase in clinical practice, there have been increasing pressures to outsource aspects of care to allied health professionals. A key component of this has been the tendency to refer counselling to social work and psychologists without formal evaluation of the impact upon or preference of patients in this regard. While such measures largely have been introduced for patient benefit, few of the changes in practice have been subjected to randomised trial to evaluate their efficacy to achieve a reduction in psychological symptoms. It is important that such assessments occur. Reducing psychological symptoms is a desirable outcome as it improves the quality of life of cancer patients. However, there is increasing evidence that reduction in anxiety and depressive symptoms might also be necessary to optimise the body’s immunological response to cancer. The psycho-neuroimmunological literature provides good evidence that chronic stress has a detrimental effect upon immune function that may retard the ability of an individual with cancer to resist disease progression. A study D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394 393 of 116 patients with breast cancer reported that the psychological effects of stress were associated with inhibition of the cellular immune responses relevant to cancer prognosis, including natural killer cell toxicity and T-cell responses22. Furthermore, methods to reduce psychological stress can enhance some aspects of the cellular immune response. A study in 61 adults receiving relaxation training found evidence of a 30% increase in natural killer cell lysis compared with controls23. In a study of melanoma patients, those receiving psychological interventions exhibited higher levels of natural killer cell activity augmented by interferon alpha than controls24. These benefits have also been reported in patients with gynaecological cancers. A study of 22 patients with ovarian cancer receiving chemotherapy reported that those undergoing relaxation training had higher lymphocyte counts and higher total white cell counts than controls25. Further research is warranted to explore whether patient outcomes of post-diagnosis interventions in the setting of gynaecological cancer are altered by the nature of the provider of the intervention, namely doctor, nurse, social worker or psychologist. We also need further trials to compare the efficacy of different types of interventions at achieving resolution of patient’s symptoms. Finally, while senior medical staff spending quality time with their patients in the post-diagnosis phase may have an immediate cost in terms of staffing, it is possible that such an allocation of resources may prove to be more cost effective than other areas of expenditure. In the present study, the allocation of intervention time by senior medical staff resulted in significant reductions in adverse psychological sequelae. The intervention achieved an immediate improvement in patient’s symptoms, and as such was a desirable endpoint in its own right. However, it is possible that resolution of stress symptoms improves immunological responses necessary to retard tumour recurrence or progression. If this is the case, then spending time talking to our patients may represent better value for money than the incorporation of the newest and increasingly more expensive chemotherapy regimens into the treatment protocol. Acknowledgements The authors would like to thank Dr M. Davy for her support, the staff who assisted with the distribution of questionnaires and all the participating patients. References 1. Andersen BL. Psychological issues. In: Berek JS, Hacker NF, editors. Practical Gynecologic Oncology. Philadelphia: Lippincott Williams & Wilkins, 2000:887 – 913. 2. Weisman AD, Worden JW. The existential plight in cancer: significance of the first 100 days. Int J Psychiatry Med 1976;7:1 – 15. 394 R.W. PETERSEN & J.A. QUINLIVAN 3. Evans DW, McCartney CF, Nemeroff CB, et al. Depression in women treated for gynaecological cancer: clinical and neuroendocrine assessment. Am J Psychiatry 1986;143:447 – 452. 4. Cain EN, Kohorn EI, Quinlan DM, Schwartz PE, Latimer K, Rodgers L. Psychosocial reactions to the diagnosis of gynecologic cancer. Obstet Gynecol 1983;62:635 – 641. 5. Maier SF, Watkins LR, Fleshner M. Psychoneuroimmunology: the interface between behavior, brain, and immunity. Am Psychol 1994;49:1004 – 1017. 6. Herbert TB, Cohen S. Depression and immunity: a meta-analytic review. Psychol Bull 1993;113:472 – 486. 7. Cain EN, Kohorn EI, Quinlan DM, et al. Psychosocial benefits of a cancer support group. Cancer 1986;57:183 – 189. 8. Wenzel LB, Robinson SE, Blake DD. The effects of problemfocussed group counselling for early-stage gynaecologic cancer patients. J Mental Health Counseling 1995;17:81 – 93. 9. Moynihan C, Bliss JM, Davidson J, Burchell L, Horwich A. Evaluation of adjuvant therapy in patients with testicular cancer: randomised controlled trial. BMJ 1998;316:429 – 435. 10. Small R, Lumley J, Donohue L, Potter A, Waldenstrom U. Randomised controlled trial of midwife led debriefing to reduce maternal depression after operative birth. BMJ 2000;321:1043 – 1047. 11. Wessely S, Rose S, Bisson J. A systematic review of brief psychological interventions (‘debriefing’) for the treatment of immediate traumatic symptoms and the prevention of post traumatic stress disorder. Cochrane Collaboration: Cochrane Library, Issue 4. Oxford: Update Software, 1999. 12. Zigmond AS, Snaith RP. The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 1983;67:361 – 370. 13. Goldberg DP, Hillier VF. A scaled version of the General Health Questionnaire. Psychological Med 1979;9:139 – 145. 14. Hermann C. International experience with the Hospital Anxiety and Depression Scale: a review of validation data and clinical results. J Psychosomatic Res 1997;42:17 – 41. 15. Walker LG, Heys SD, Walker MB, et al. Psychological factors can predict the response to primary chemotherapy in patients with locally advanced breast cancer. Eur J Cancer 1999;35:1783 – 1788. 16. Hall A, A’Hern R, Fallowfield L. Are we using appropriate selfreport questionnaires for detecting anxiety and depression in women with early breast cancer? Eur J Cancer 1999;35:79 – 85. 17. Achte KA, Vanhkonen ML. Cancer and the psyche. Omega 1971; 2:46 – 56. 18. Capone MA, Good RS, Westie KS, et al. Psychosocial rehabilitation of gynecologic oncology patients. Arch Phys Med Rehabil 1980; 61:128 – 132. 19. NHS Executive. Guidance on Commissioning Cancer Services. Improving Outcomes in Gynaecological Cancers: The Research Evidence. London: HMSO, 1999. 20. Watson M, Greer S, Blake S, Shrapnell K. Reaction to a diagnosis of breast cancer. Relationship between denial, delay and rates of psychological morbidity. Cancer 1984;53:2008 – 2012. 21. Scott JT, Entwistle VA, Sowden AJ, Watt I. Recordings or summaries of consultations for people with cancer. Cochrane Database of Systematic Reviews, No. CD001539, 2000. 22. Andersen BL, Farrar WB, Golden-Kreutz D, et al. Stress and Immune Responses after Surgical Treatment for Regional Breast Cancer. J Natl Cancer Inst 1998;90:30 – 36. 23. Kiecolt-Glaser JK, Glaser R, Williger D, et al. Psychosocial enhancement of immunocompetance in a geriatric population. Health Psychol 1985;4:25 – 41. 24. Fawzy FI, Kemeny ME, Fawzy NW, et al. A structured psychiatric intervention for cancer patients. II. Changes over time in immunological measures. Arch Gen Psychiatry 1990;47:729 – 735. 25. Lekander M, Furst CJ, Rotstein S, Hursti TJ, Fredrikson M. Immune effects of relaxation during chemotherapy for ovarian cancer. Psychother Psychosom 1997;66:185 – 191. Accepted 4 February 2002 D RCOG 2002 Br J Obstet Gynaecol 109, pp. 386 – 394