The psychosocial impact of prostate cancer on patients and their partners

Our multisite study was designed to elicit data from two groups of couples facing PCA: (i) couples where the patient has just learnt he has localised (potentially curable) PCA; and (ii) couples where the patient has been told that he has metastatic (incurable) PCA. Consecutive clinic attendees who met study criteria (Box 1) were recruited between March 2002 and March 2005 by oncologists and urologists working in public hospitals and private practices in metropolitan Melbourne. (The patients were also recruits for a larger, patient only, 18-month longitudinal study.) Interviews took place before the launch of definitive treatments, such as prostatectomy or radiotherapy, or soon after.

All patients had a living female partner, and both patient and partner gave written consent to participate. If patients died, their partners became ineligible. Ethical approval was obtained from human research ethics committees at each of the five public hospitals involved in the study and from the University of Melbourne.

Patients and partners simultaneously and confidentially completed self-report questionnaires at the time of diagnosis of localised or metastatic PCA (Time 1), and 6 months later (Time 2). Demographic data were collected from all patients and partners. Most Time 1 interviews took place in the couples’ homes; Time 2 questionnaires were sent by mail with telephone follow-up. Patients and partners were interviewed and completed their questionnaires separately, so that their responses could be kept confidential from their spouse. A computer-based psychiatric diagnostic interview (see below) was conducted with the partners at Time 1, and with the patients 3 months later (both will be referred to as part of Time 1 data hereafter).

The anxiety and depression modules of the Composite International Diagnostic Interview (CIDI)7 were administered to patients and partners. The CIDI, which derives psychiatric diagnoses according to the Diagnostic and statistical manual of mental disorders 4th edition (DSM-IV),8 has confirmed reliability and validity.7 The CIDI computer software takes account of symptoms arising from physical illness (including cancer and its treatments) to ensure they are not incorrectly ascribed to depression or anxiety diagnoses.7 The CIDI has recency codes that separate past psychiatric disorders from current disorders. For our study, only DSM-IV diagnoses over the 12 months before the interview are reported (12-month prevalence rates).

The Brief Symptom Inventory (BSI)9 is a 53-item self-report psychological symptom scale. It yields a global rating of psychological status called the Global Severity Index (GSI), comprising a grand total of scores divided by the number of valid items. The BSI also provides subscale scores for depression and for anxiety. In each case, a higher score denotes greater psychological distress. The BSI’s ability to discriminate clinically distressed from non-distressed samples has been well established, using structured psychiatric interviews as benchmarks.9 Extensively used in psycho-oncology research, the BSI is sensitive to changes in psychological adjustment and to psychopathology.10 Specific threshold scores of psychiatric “caseness” (where an individual’s levels of symptoms are deemed to correspond to a diagnosable psychiatric disorder) have been established for both men and women.11

The Dyadic Adjustment Scale (DAS)12 is a six-item self-report scale widely used in the context of families experiencing cancer.13 The DAS is designed to measure the quality and level of satisfaction of romantic relationships in cohabiting couples.14 The DAS yields the DAS Full Scale score, comprising the sum of the recoded item scores; a higher score represents greater marital satisfaction (maximum 30).

In the data presented below, we cite the number of participants at Time 1 and Time 2 who had usable scores for each analysis. SPSS, version 11.5 (SPSS Inc, Chicago, Ill, USA) was used throughout.

The independent sample t test was used to compare psychological distress and marital satisfaction in couples who participated at Time 2 and those who did not, and χ2 tests were used to compare couples facing localised PCA with those facing metastatic PCA. Rates of DSM-IV diagnoses (given as percentages) were compared directly with 12-month community prevalence rates. Paired samples t tests were used to compare psychological distress and marital satisfaction of patient and partner at each time point, and a repeated-measures analysis of variance (ANOVA) was used to examine the interaction of psychological distress and marital satisfaction at Time 1 and Time 2 in patients and partners.

Referring clinicians identified 141 eligible patients from consecutive clinic attendees: 28 patients refused permission for their partners to be approached. Of the partners approached, 10 were ineligible, could not be contacted, or declined to participate.

The remaining 103 couples (73% of identified couples) were recruited. The mean age of the patients was 66 years (SD, 9.5; range, 43–92), and that of their partners was 62 years (SD, 9.3; range, 41–83) (Box 2). The patients’ ages were in the expected range.15

Forty-seven patients (46%) were diagnosed with localised PCA and 56 (54%) with metastatic disease. Twenty-four of the 103 patients (23%) received no active treatment during the study period, and 49 (48%) had hormone treatment. Of the 47 patients with localised PCA, 13 (28%) had a radical prostatectomy and five (11%) had some form of radiotherapy. Of the 56 patients with metastatic PCA, eight (14%) received chemotherapy. The range of treatments was generally consistent with Victorian Cancer Registry data.16 Seventy-one of the patients (69%) were recruited from public hospital clinics, the remainder from private practice.

The longest time between first PCA diagnosis and recruitment into the study as a patient with metastatic PCA was 14 years, the shortest was within a year (five patients), with a mean duration of 4.6 years. Twenty eligible patients and 11 partners reported having previously had other types of cancer.

Twenty patients (19%) were no longer participating at Time 2: nine had died and 11 withdrew. Thirteen of these patients (seven dead, six withdrawn) were lost to the study at Time 1 before the CIDI was administered, so that DSM-IV diagnoses were not obtained for these patients. Fourteen partners (13.5%) were no longer participating at Time 2: 10 were no longer eligible to continue, nine because the patient they partnered had died, and one had suffered a disabling stroke. Four other partners withdrew because of lack of time or feeling overwhelmed by their situation.

Patients and partners not participating at Time 2 were significantly more likely to have been facing advanced PCA at Time 1 (if a patient died, his partner was not eligible to continue, but if he survived but withdrew she remained eligible) (Box 3). The age of patients (but not partners) was significantly higher among non-participants at Time 2. There was no significant difference between participants and non-participants at Time 2 in GSI or DAS scores (data not shown).

Contrary to our expectations, differences in psychological distress or marital satisfaction between patients (and between partners) facing localised and metastatic PCA at both Time 1 and Time 2 did not reach statistical significance. In subsequent analyses, patients with localised and metastatic disease (and their partners) were combined into one group.

The rates of DSM-IV anxiety and depression for 90 patients and 103 partners at Time 1 are given in Box 4. The partners had rates of major depression more than twice the normal rate for the Australian (Victorian) community;17 patient rates were also higher than the community rate but proportionally not as elevated. A similar pattern emerged for rates of generalised anxiety disorder and other anxiety disorders in patients and partners.

Overall mean psychological distress (GSI) in partners at Time 1 was almost twice as high as in patients (Box 5). Similarly, depressive symptoms and anxiety symptoms at Time 1 (BSI Depression and Anxiety Subscales) were substantially higher in partners compared with patients.

Patients experienced an increase in distress at Time 2 compared with Time 1 (F1,81, 7.26; P < 0.05; n = 82), whereas partners reported a reduced level of distress at Time 2 compared with Time 1 (F1,87, 3.99; P < 0.05; n = 88) (Box 6A). At Time 2, distress in partners did not differ significantly from that of patients, unlike the situation 6 months earlier.

A mixed-model repeated-measures ANOVA (n = 76), with the within-subject factor of time and the between-subject factor of patient versus partner, revealed a significant convergent interaction overall in mean GSI scores between patients and partners over time (F1,75, 14.256; P < 0.001).

Patients’ and partners’ marital satisfaction at Time 1 was not significantly different. At Time 2, however, partners’ satisfaction was lower than that of patients (t, 2.35; df, 70; P < 0.05; n = 71). Moreover, partners’ satisfaction fell from Time 1 to Time 2 (F1,88, 5.71; P < 0.05, n = 89), whereas that of patients did not change significantly (Box 6B).

A mixed-model repeated-measures ANOVA (n = 77), with the within-subject factor of time and the between-subject factor of patient versus partner, revealed a borderline significant divergent interaction overall in mean DAS Full Scale score between patients and partners over time (F1,76, 3.917; P < 0.05). The effect of time alone was significant (F1,76, 6.565; P < 0.01), but the effect of patient versus partner alone was not (F1,76, 2.775; P = 0.1).

At Time 1, 15% of partners (n = 102) exceeded the female GSI threshold for psychiatric caseness;11 at Time 2, 14% of partners (n = 89) met criteria for psychiatric caseness. Ten per cent of patients (n = 103) met criteria for psychiatric caseness at Time 1, but this went up to 15% of patients (n = 82) at Time 2 (Box 7).