Psychosocial impact of prognostic genetic testing in the care of uveal melanoma patients: protocol of a controlled prospective clinical observational study

This study is conducted within a network of three collaborative investigations (see Fig. 1) funded by the German Cancer Aid. A corporate time table was developed, in which the three projects were embedded (see Table 1).

Patients are included in the study if they are reliably diagnosed with uveal melanoma and if removal of a tumor sample is possible (enucleation or biopsy). Furthermore, they have to be aged between 18 and 90 years, have sufficient knowledge of the German language and give informed consent to take part in the study. Patients are excluded from the study if they have a pre-existing diagnosis of mental disability, psychosis or dementia.

The sample of this study is a consecutive sample of patients with uveal melanoma undergoing cancer therapy in the Department of Ophthalmology of the University Hospital in Essen, Germany, which is a national center for treatment of patients with ophthalmic tumors. Therefore, more than 170 patients with a first diagnosis of uveal melanoma undergo prognostic genetic testing for monosomy 3 every year.

This study is a non-randomized controlled prospective clinical observational trial. Patients are allocated to the groups by their choice (non-randomized) of whether to utilize prognostic genetic testing (intervention group, IG) or not (observationsal group, OG). Blinding is not possible.

Data are assessed preoperatively (at initial admission in the clinic) and postoperatively (at discharge from hospital after surgery, 6–12 weeks, 6 and 12 months after initial admission); the detailed time flow of the study can be seen in Fig. 2.

Exclusively at baseline assessment (initial admission in the clinic): sociodemographic data, medical history and initial protective factors are recorded. For assessment of initial protective factors (resilience and perceived social support), participants complete the short form of the sense of coherence scale [17], the short form of the social support questionnaire [18] and the subscale “social strain” from the long form of the social support questionnaire [19, 20].

The following measures are assessed both at baseline and at all follow-ups: psychological distress, (general and specific) health-related quality of life, attitude towards genetic testing, estimation of the perceived risk of metastasis and utilization of psycho-oncological interventions.

Psychological distress is measured by the Fear of Progression Questionnaire [21, 22], the Hospital Anxiety and Depression Scale (subscale depression, HADS-D, [23]), and the distress thermometer [24]. Specific health-related quality of life is assessed by the EORTC Quality of Life Questionnaire (EORTC QLQ-C30, [25]) and the ophthalmic module of the EORTC QLQ-C30 (EORTC QLQ-OPT 30, [26]). General health-related quality of life is measured by the short version of the Short Form Health Survey SF-36 (SF-12; [27]). Attitudes towards genetic testing are assessed with a modified version of the modified Attitudes Scale [28]. The estimation of perceived risk of metastasis is evaluated with a visual analogue scale (VAS), from 0 to 10. The utilization of psycho-oncological (crisis) interventions (control variable) and patient satisfaction with these interventions is recorded with a documentation form.

Detailed information about several of these measures is provided in the Additional file 1: Appendix. Which measures are assessed at which measurement point can be seen in Table 2.

Patients are informed at T1 that they may utilize psycho-oncological interventions. These crisis interventions follow a resource-oriented approach and are conducted by a clinical psychologist specialized in psycho-oncology. Frequency and kind of crisis intervention, as well as patient satisfaction, are assessed as covariates.

Data will be analyzed using the software SPSS for Microsoft Windows®. For descriptive analysis, data will be expressed as mean values, standard deviations, and frequencies; distributional characteristics of all variables will be given. To analyze simple linear relations between variables, Pearson correlations will be calculated. For verification/falsification of our hypotheses (group differences), ANOVAs (post hoc: Scheffé tests), repeated measurement ANOVAs, ANCOVAs, parametric tests (for independent and dependent samples) and non-parametric tests (Chi² tests, Mann–Whitney U-tests, Wilcoxon tests, Kruskal-Wallis tests) will be calculated (according to the respective levels of measurement and sample sizes). For prediction of the outcome variables, multiple regression analyses will be performed. For all tests, a significance level of p < 0.05 is predetermined.

Calculation of the sample size (GPower) for two-tailed t-tests with dependent samples, effect size = 0.6, probability of error α = 0.05, and power = 0.8 resulted in a sample size of n = 24 per group. For two-tailed t-tests with independent samples, the same calculation resulted in n = 45 per group. Due to anticipation of dropouts in the study progress, we enlarged this sample size by one third, so the final calculation of the sample size is n = 60 per group (N = 120). For the calculation of patient attendance see Fig. 2.

Patient recruitment started in October 2014. At the moment, 62 patients are already taking part in the study. For detailed information about the number of patients in each group and at each assessment point, please see Fig. 3. Baseline sociodemographic and disease-related data of all IG- and OG-patients included are presented in Table 3.

In two meta-analyses, a significant impact of genetic counseling for familial cancer on anxiety (reduction), and on accuracy of perceived risk and knowledge (both improved) was found, suggesting genetic counseling had no adverse psychological effects [29, 30]. Comparably, beneficial effects of multidisciplinary genetic risk counseling for familial colorectal cancer on psychosocial outcome were found: General anxiety, familial cancer-specific distress and general cancer worries were significantly reduced after genetic counseling [31]. Moreover, distress and depression levels were reduced within the first 6 months after counseling and testing [32], and intrusion and avoidance significantly reduced over time [33]. Furthermore, the authors found passive and palliative coping styles, excessive breast self-examination, and overestimation of breast cancer risk to be predictive of increased long-term distress. In contrast, decreased long-term distress could be predicted by promoting reassuring thoughts.

When comparing carriers and non-carriers, carriers were found to be significantly more distressed from testing and to report increased risk perception levels and surveillance (up to four years) [34]. Furthermore, Meiser (2005) [35] differentiated in their review between individuals having never been affected by cancer and individuals affected by cancer. In the former condition, non-carriers psychologically benefitted significantly, whereas carriers experienced no adverse effects. In individuals affected with cancer, the individual former cancer experience seemed to influence the effects of genetic testing.

While waiting for disclosure of the test result, acute anxiety may emerge [36]. Phelps et al. (2013) [37] investigated the effectiveness of a self-help coping intervention in patients awaiting their genetic result. Intrusive thoughts during the waiting period could be reduced significantly in patients with moderate baseline levels of intrusion. Furthermore, distress levels were decreased in patients with low or moderate intrusive worries at baseline. Yet, no intervention effect on the sample as a whole could be demonstrated, suggesting that patients with clinically high levels of psychological distress might need more intensive psychological treatments. The authors conclude that their intervention could provide help while waiting for test results and feeling uncertain, in various oncological patient groups. Moreover, there seem to be no harmful effects on those patients that are likely to not benefit from the intervention.

Our study was designed to investigate if these findings could also account for uveal melanoma patients (who have already been diagnosed with cancer).

Affected persons have to weigh the pros and cons of genetic testing when making their decision. Pros are, for example, relief from uncertainty, adjustment of important life decisions and life planning, as well as making decisions about intensified prophylaxis and screening programs. Cons are, for example, risks of persistent psychological distress, such as depressive rumination about one’s individual risk of disease.

It is an important ethical issue, as to how decision-making concerning genetic testing can be improved, such as through educative interventions both for patients and for medical staff. Wakefield et al. (2008) [38] developed a decision aid intervention (specifically for informed decision making concerning genetic testing for hereditary non-polyposis colorectal cancer risk) and tested its efficacy in a randomized controlled trial. Participants of the intervention group experienced less decisional conflict and were better informed and therefore had a better chance of making an informed decision with regard to genetic testing. Furthermore, Ilic et al. (2015) [39] (review, prostate cancer) found improvements in knowledge, reduction in decisional conflict and an increase in decisional satisfaction after utilization of a decision aid intervention. Moreover, Dieng et al. (2014) [40] (review, mainly breast cancer studies) found changes in risk perception level and accuracy after educational interventions in prospective observational studies (n = 28), but not in randomized controlled trials (n = 12). We expect our study to shed further light on questions concerning the influence of decision for or against genetic testing.

Another important research question is the appropriate and correct ethical conduct with information. The right of self-determination and respect for the personal autonomy of the patient form the ethical and legal basis for the demand for information and participation in medical decisions. It is important to note, that the right of self-determination includes also the right of not being informed, because negatively appraised health-related information can cause severe psychological distress. So, apart from the ethical principle of autonomy, the principles of non-maleficence and beneficence have to be taken into account for medico-ethical analyses [41]. To minimize psychosocial distress due to prognostic genetic testing, genetic counseling before and directly after communication of the results is recommended by the guidelines of the German Medical Association [42]. This recommendation was taken into account in the elaboration of the German Genetic Diagnostics Act. In a systematic review (93 studies), it was found that most studies about communication of prognosis to cancer patients were conducted in early stages of disease [43]. It remained uncertain which approach of communication would be the best. We expect our study to provide further information about patient autonomy and disclosure of test results.

There are two limitations of our study. First, randomization and blinding is not possible. Second, patients receive different cancer treatments (see Table 3), which might cause differences concerning the psychosocial impact of genetic testing. On the other hand, no differences concerning quality of life, psychological distress, depression and anxiety were found between enucleation and radiotherapy patients or between patients receiving different methods of radiotherapy [44, 45].

We created optimal conditions for the investigation of the psychosocial impact of prognostic genetic testing and the course of disease, in uveal melanoma patients. The naturalistic design we chose is very suitable for research in the clinical setting, but takes time and effort. We therefore thank the German Cancer Aid for promoting this network of three parallelized collaborative studies from different scientific disciplines, giving the opportunity to investigate uveal melanoma patients from different points of view.

We expect that our findings will be transferable to other oncological entities and clinical syndromes. Furthermore, we expect that empirical data from our study will contribute to the continuing medico-ethical debate on prognostic genetic testing, in the light of patient autonomy and decision-making processes.