www.nature.com/ejhg ARTICLE OPEN Assessment of psychosocial difficulties by genetic clinicians and distress in women at high risk of breast cancer: a prospective study 1,2 ✉ Anne Brédart , Jean-Luc Kop 3, Anja Tüchler 4, Antoine De Pauw 5, Alejandra Cano6, Julia Dick4, Kerstin Rhiem4, 7 Peter Devilee , Rita Schmutzler4, Dominique Stoppa-Lyonnet 5 and Sylvie Dolbeault1,8 © The Author(s) 2022 We examined how often genetic clinicians correctly identify psychosocial difficulties in women at high breast cancer risk and explored effects of this assessment and the genetic test result on counselees’ distress. A prospective observational study of counselee–clinician dyads was performed in three French, German and Spanish genetic clinics, involving 709 counselees (participation rate, 83.4%) and 31 clinicians (participation rate, 100%). Counselee–clinician agreement in perceived psychosocial difficulties was measured after the pre-test genetic consultation. Multivariate mixed linear models accounting for clinicians were tested. Predicted distress levels were assessed after the pre- (T1) and post-test result disclosure consultations (T2). Depending on 1234567890();,: the difficulty domain, clinicians adequately assessed the presence or absence of difficulties in 51% (“familial issues”) to 59% (“emotions”) of counselees. When counselees’ and clinicians’ perceptions disagreed, difficulties were generally underestimated by clinicians. Counselees’ distress levels remained stable from T1 to T2, irrespective of clinicians’ appraisal adequacy, and the genetic test result disclosure. Psychological referral need were found in 20–42% of counselees, more frequently observed for difficulties in the “emotions” domain. Our findings suggest that the genetic test result is a suboptimal indicator for psychological referral. Instead, clinicians should focus on emotions expressed by counselees to appraise their needs for psychological support. European Journal of Human Genetics (2022) 30:1067–1075; https://doi.org/10.1038/s41431-022-01096-9 INTRODUCTION increasingly offered gene panel testing [4]. Following gene panel A panel of nine susceptibility genes has been identified as testing, a pathogenic variant, a VUS or a negative uninformative clinically useful for breast cancer risk prediction [1]. Gene panel result, which does not explain the personal or familial cancer tests offer important information for making clinical recommen- history, may be obtained. Following targeted single gene testing, dations to women with a personal or familial history of breast result may reveal either the presence or absence of a pathogenic cancer. However, with a wider range of high/moderate or lower variant. risk genes being tested and the increased possibility to discover Women with a personal or familial history of breast cancer who gene variants of uncertain significance (VUS), the information ask for genetic testing (i.e., the “counselees”) often experience delivered during genetic counseling has become very complex psychosocial difficulties in relation to genetic testing relating to: and a source of confusion [2]. Next generation sequencing and the informed choice (i.e., to choose between screening or risk increasing capacity of gene testing also extends the number of reducing surgery), understanding the cancer risks associated with women eligible for testing [3], leading to time management the result (e.g., appraisal of risks, coping with a result of unknown challenges for genetic clinicians. clinical significance), and the personal and familial consequences In routine practice, gene panel testing is firstly proposed to a of the test [5]. These difficulties may demand particular attention woman in the family who developed cancer (index case) to and deserve specific support. identify a possible breast cancer genetic factor. If a pathogenic Limited adverse psychological outcomes have been observed variant is found, blood relatives are proposed targeted single gene following the disclosure of the gene test result [6]. Carriers of a testing to identify or rule out the presence of the known pathogenic variant experienced higher short-term distress after pathogenic variant identified in the family. They are also testing [7] which remits after 1 year [8]; higher distress was also 1 Institut Curie, Supportive Care Department, Psycho-oncology Unit, 26 rue d’Ulm, 75005 Paris Cedex 05, France. 2Psychopathology and Health Process Laboratory UR4057, Psychology Institute, Paris University, 71 Avenue Edouard Vaillant, 92774 Boulogne-Billancourt, France. 3Université de Lorraine, 2LPN, 3 Place Godefroy de Bouillon, 54015 Nancy Cedex, France. 4Center for Hereditary Breast and Ovarian Cancer, Faculty of Medicine and University Hospital Cologne, University Hospital of Cologne, University of Cologne, Kerpener Str. 62, 50937 Cologne, Germany. 5Institut Curie, Department of Genetics, INSERM U830, Paris University, 26 rue d’Ulm, 75005 Paris Cedex 05, France. 6University Autónoma of Barcelona, Clinical and Health Psychology Department, Barcelona, Spain. 7Leiden University Medical Centre, Department of Human Genetics, Department of Pathology, S4-P, P.O. Box 9600, 2300 RC Leiden, The Netherlands. 8CESP, University Paris-Sud, UVSQ, INSERM, University Paris-Saclay, 16 Avenue Paul Vaillant-Couturier, 94807 Villejuif Cedex, France. ✉email: anne.bredart@curie.fr Received: 10 November 2021 Revised: 3 March 2022 Accepted: 22 March 2022 Published online: 11 April 2022 A. Brédart et al. 1068 reported when a VUS had been identified [6, 9]. However, some Counselee–clinician dyads were obtained for four samples from the two counselees, particularly those who feel unprepared or lack of successive cohorts constituted within BRIDGES. The first and second support, may experience distress [10]. Adequate identification of samples comprised women with a personal (index cases) or familial history counselees who might require additional genetic counseling and of breast cancer. Index cases received diagnostic breast cancer gene panel support would help minimize distress [11]. testing; women free of cancer were offered predictive targeted testing. These women were approached at Curie Institute (France, sample 1) and The major part of the cancer genetic consultation is usually the Catalan Institute of Oncology (Spain, sample 2) [17]. The third and devoted to the provision of biomedical information, leaving less fourth samples involved women free of a personal breast cancer who time to enquire about counselees’ concerns associated with underwent either predictive targeted testing and the gene panel testing genetic testing [12]. Moreover, although emotional factors are implemented within BRIDGES [20] if a pathogenic variant had been strong predictors of distress after genetic testing [7], psychological identified in the family, or otherwise only gene panel testing. These referral is more often based on the genetic test result than on women were approached at Curie Institute (sample 3) and the Center for distress [13]. This may result in unfulfilled psychosocial care needs Hereditary Breast and Ovarian Cancer at the University Hospital Cologne during the course of genetic testing [14]. (Germany, sample 4). Counselees with a new, recurrent or advanced breast cancer (in the first Because of these concerns, attention is increasingly devoted and second samples), another type of cancer or a major psychiatric to clinician–counselee communication in genetic testing [15]. In disorder were not included in the study. the broader oncology field, initiative such as the completion of Counselees who agreed to participate were invited to complete the psychosocial questionnaires before routine consultation seems study questionnaires at home (online or on paper) within 2 weeks after the to facilitate discussion about patients’ psychosocial concerns pre-test consultation (T1) and again within 2 months after the post-test during the consultation, which positively affects physicians’ consultation (T2). When necessary, a reminder call was made to awareness and management of patients’ problems and health counselees. Questionnaires not completed within 1 month afterwards outcomes [16]. Similarly, in the cancer genetic consultation, were considered missing. Clinicians completed questionnaires within adequate appraisal of counselees’ psychosocial difficulties 1 week after the pre-test consultation. would also positively affect counselees by eliciting genetic clinicians’ action such as providing extra counseling or support, Questionnaires and data collection psychosocial information or referral to psychosocial services and Socio-demographic and clinical data were collected from counselees after so lowering their distress. the pre-test consultation and from medical records. To date, only one study performed in a single country has The extent of genetic-specific psychosocial difficulties was assessed at T1 using the 26-item “Psychosocial Aspects of Hereditary Cancer” (PAHC) documented the genetic clinicians’ little awareness of counselees’ questionnaire [5] available in French, German and Spanish [21]. Items are psychosocial difficulties in the context of familial cancer genetic grouped into six conceptually defined domains [5]: hereditary predisposi- testing [14]. Moreover, that study was not designed to clarify tion, practical, familial, emotional, living with cancer and children-related which domains of psychosocial problems needed improved issues (Supplementary Table S2 provides examples of items). Possible counselors’ awareness to affect counselees’ distress level. answers are: Not at all (1), A little (2), Quite a bit (3) and Very much (4). The We designed this prospective observational study to evaluate presence of a difficulty in a domain is determined according to possible the extent to which genetic clinicians correctly identify domains of cut-offs to an item response scale as shown in Fig. 1. A difficulty was psychosocial difficulties in counselees at the initial (pre-test) considered present if a counselee indicated “quite a bit”, or “very much” on consultation for breast and ovarian cancer genetic susceptibility at least one item within a given difficulty domain, or alternatively, a testing. The study also examined the effect of clinicians’ appraisals response “very much”. For each domain, counselees are also asked about their need for additional help (Yes vs. No). of these difficulty domains on counselees’ distress after the initial Counselees’ self-reported distress was measured by French [22], German consultation and after the genetic test result disclosure. Lastly, the [23] and Spanish [24] versions of the Hospital Anxiety and Depression Scale study assessed the effect of disclosure of the genetic test result on (HADS) at T1 and at T2. HADS total scores range from 0 to 42. counselees’ distress. The PAHC questionnaire filled out by the clinician independently from counselees examines the presence and severity of problem in the same domains as those of the PAHC counselees’ version on a 3-level rating scale METHODS [5]: No problem (1), Minor problem (2) and Important problem (3). The study protocol was approved in France by the Comité consultatif sur le traitement de l’information en matière de recherche dans le domaine de la Statistical analyses santé (CCTIRS: Consultative committee for information management in A sample size of at least 500 counselees was predicted to allow for health research—No. 16.314) and the Comité de Protection des Personnes multivariate analyses. Ile-de-France V (CPP—No. 18.12.28.38743 CAT2), in Germany by the Ethics Statistical analyses of dyad agreement were restricted to questions that Committee of the University Hospital of Cologne (No. 16-098) and in Spain were answered by both counselees and clinicians. by the Ethics Committee of the Instituto Catalán de Oncología of Barcelona We assessed rates of agreement for counselee–clinician dyads using (No. PR111/16). All recruited women provided written informed consent. different possible cut-offs of the counselee and clinician response scales as shown in Fig. 1. The thick line in Fig. 1 indicates the selected cut-offs; using Study design these cut-offs, a counselee was considered as presenting difficulty in a This prospective observational study was part of a multicenter psychoso- cial research [17] undertaken within the European research program Not at all A lile Quite a bit Very much BRIDGES (“Breast Cancer Risk after Diagnostic Gene Sequencing”) (https:// PAHC Counselee’s bridges-research.eu). It comprised two successive cohorts of consecutive reported difficules female counselees undergoing testing for breast cancer risk, one using current national breast cancer gene panel [18, 19] (cohort 1—November PAHC 2016–April 2018) and the other using the BRIDGES (TruRisk® v3.1.1 [20]) Clinician’s assessment of gene panel (cohort 2—November 2019–December 2020). counselee’s difficules Not at all Minor problem Major problem Counselee–clinician dyads PAHC Counselee’s reported Counselees comprised women aged 18 years or over, with a personal and/ need for help No Yes or familial history of breast cancer, who were consecutively approached on the day of the initial (pre-test) visit. All clinicians in the genetic clinics of Fig. 1 Agreement calculation. Thick line indicates cut-offs for the Curie Institute (France), University Hospital of Cologne (Germany) and counselees’ and clinicians’ PAHC response scale which were selected Catalan Institute of Oncology (Spain) participated in the study. as they were optimal in terms of positive predictive value. European Journal of Human Genetics (2022) 30:1067 – 1075 A. Brédart et al. 1069 domain if she answered “Quite a bit” or “Very much” to at least one item in assumptions on relationships between factors deemed to influence this that domain and the clinician was considered as judging the presence of a effect [6–8] (Fig. 2). To this end, multivariable mixed linear models were problem if he/she reported a “minor” or “major” problem. These cut-offs fitted taking the TN category as reference. The basic model comprised the were chosen as they maximized the positive predictive value of clinicians’ intercept, the random effect of clinicians on the intercept and the fixed assessment (i.e., the probability of agreement on the presence of effect of samples. For distress at T1, the basic model was compared to: (1) a difficulties) as this reflects an appropriate referral of counselee in a context first model including the effect of agreement, and (2) a second model in of limited time and resources in psychological support (see Supplementary which the interaction between agreement and samples was added to the Table S3 for detailed data by cut-off). first model. For distress at T2, the basic model was compared to: (1) a first For each counselee–clinician dyad and PAHC domain, we calculated model including the effect of agreement, (2) a second model based on the proportions of responses where counselees and clinicians agreed (true first model plus time lapse between the pre- and post-test consultations positives (TP) and true negatives (TN)), and where clinicians under- (false and its interaction with agreement, and (3) a third model based on the negatives (FN)) or overestimated (false positives (FP)) counselees’ second model plus genetic test result (pathogenic variant, negative vs. difficulties. The level of agreement between counselees and clinicians uninformative results) and its interaction with agreement. Model selection was statistically quantified using the Kappa statistics with confidence was based on the Bayesian information criterion (BIC) [26]. Beta coefficients intervals (CI) [25]. derived from the selected model were used for predicting HADS means The Kappa statistic provides a quantitative measure of the magnitude of and 95% CI for each TN, TP, FP, and FN agreement category. The effect of agreement between counselees and clinicians. Kappa statistic ranges from inter-clinicians’ variability on the outcome was assessed by intra-class −1 (total disagreement) to +1 (perfect agreement). Kappa values from 0.01 correlation coefficients (ICC). to 0.20, 0.21 to 0.40, and above 0.41 equate to slight, fair, and moderate to Another set of multivariable mixed linear models examined the effect of substantial agreement [25]. A Kappa of zero and a 95% confidence of the genetic test result on counselees’ distress at T2. Because of the marked Kappa statistics including zero indicates no agreement at all. difference in personal breast cancer history [7], separate models were The effect of dyad agreement (i.e., TP, FP or FN vs. TN) on distress fitted for samples 1 and 2, and for samples 3 and 4. Moreover, VUS were measured by the HADS at T1 and at T2 was tested according to theoretical communicated to counselees in sample 1 and 2, but not in samples 3 and Genec test result disclosure consultaon Samples Agreement clinician-counselee on Distress psychosocial difficules T2 following inial genec consultaon (T1) Distress T1 Time between pre- and post-test consultaons Fig. 2 Hypotheses on the effect of agreement between clinicians and counselees on distress at T1 and T2, and factors susceptible to influence this effect. The thick arrows represent the effect being tested. T1 = within 1 month after the initial genetic consultation; T2 = within 3 months after the genetic test disclosure consultation. The theoretical model presumes that a “sample” would influence “agreement” at T1 and “distress” at T1 and T2; b that the effect of “agreement” at T1 on distress at T2 would depend on time elapsed between T1 and T2; c the influence of distress at T1 on distress at T2 was not considered because distress at T1 is part of the link between “agreement” at T1 and distress at T2; d genetic test result at T2 is independent from “agreement” and distress at T1; the influence of this factor on distress at T2 is tested independently. Personal history Genec test result of breast cancer disclosure consultaon Samples Distress Distress T2 T1 Time between pre- and post-test consultaons Fig. 3 Hypotheses on the effect of genetic test results on distress at T2, and factors susceptible to influence this effect. The thick arrow represents the effect being tested by the study. T1 = within 1 month after the initial genetic consultation; T2 = within 3 months after the genetic test disclosure consultation. The theoretical model presumes that a distress at T2 would depend on distress at T1; b a personal history of breast cancer would influence distress associated with the disclosure of the genetic test result at T2; c “sample” would influence “distress” at T2 via the personal history of breast cancer; d the effect of “distress” at T1 on “distress” at T2 would depend on the time elapsed between T1 and T2. European Journal of Human Genetics (2022) 30:1067 – 1075 A. Brédart et al. 1070 Table 1. Counselees’ socio-demographic and clinical characteristics (N = 850). Sample 1—FR1 N Sample 2—SP N Sample 3—FR2 N Sample 4—GE N = 258 = 170 = 200 = 222 Age Mean (SD) 48.0 (12.1) 47.8 (12.8) 39.3 (13.3)a 41.0 (9.9) Age Median (range) 47.5 (21–82) 47.8 (19–81) 36 (21–80) 41 (21–71) Having children Yes n (%) 203 (78.7) 114 (77.0) 97 (48.5) 140 (63.1) Clinical data Personal history of breast cancer n (%) (Yes) 208 (80.6) 108 (63.9) NA NA Genetic test result in first person tested in the family n (%) Pathogenic variant 34 (13.2) 49 (28.8) 200 (100) 54 (25.0) Non-informative result NA NA NA 97 (44.9) Variant of uncertain significance NA NA NA 4 (1.9) First person tested in the family not accessible NA NA NA 37 (17.1) First person eligible for testing in the family not NA NA NA 24 (11.1) yet tested Respondents at T1 N = 213 N = 133 N = 157 N = 206 Education level n (%) Compulsory education or below 6 (2.8) 57 (33.3) 4 (2.6) 8 (3.9) Secondary or technical/vocational education 60 (28.4) 55 (32.6) 36 (23.4) 117 (56.8) Higher education or above 145 (68.7) 58 (34.1) 114 (74.0) 81 (39.3) Marital status n (%) Married/partnered 149 (70.3) 102 (77.3) 94 (60.3) 146 (70.9) Others (widowed, separated/ divorced, single/ 63 (29.7) 30 (22.7) 62 (39.7) 60 (29.1) never married) Loss of family member because of breast/ovarian 86 (42.8) 60 (46.9) 76 (49.0) 122 (59.2) cancer n (%) (Yes) Past psychological help n (%) (Yes) 113 (53.1) 51 (38.3) 73 (47.1) 87 (42.2) Respondents at T1 and T2 Sample 1—FR1 N Sample 2—SP N Sample 3—FR2 N Sample 4—GE N = 164 = 67 = 118 = 199 Time lapse between pre- and post-test 163 (17) 100 (57) 108 (17) 97 (33) consultations (days) Mean (SD) Number of counselees with BRCA1, BRCA2 vs. other 24/1 16/0 31/2 17/14 high and moderate-risk pathogenic variant Number of counselees with a negative result 14 14 85 24 Number of counselees with an uninformative vs. 113/12 23/14 0 143 VUSb FR France, GE Germany, SP Spain, T1 within 1 month after the pre-test consultation, T2 within 3 months after the post-test consultation, NA not applicable either because no index case tested yet (samples 1 and 2) or counselee’s eligibility only if family member carrier of a pathogenic variant (sample 3). a Comparisons between eligible counselees and respondents at T1 and at T2; participant in sample 3 at T1 are older (mean (SD) age = 40.3 (13.2); p value = <0.05). b VUS = variant of uncertain clinical significance; VUS not communicated to counselees in samples 3 and 4. 4. The effect of the genetic test result on distress was tested according to and genetic test result of the first person tested in the family, theoretical assumptions on relationships between factors deemed to except in sample 3 where participants at T1 were slightly older. influence this effect (Fig. 3). The basic model comprised the intercept, the Counselee–clinician dyads were constituted for 213 counselee random effect of clinicians on the intercept and the fixed effect of samples. respondents in sample 1 (France), 133 in sample 2 (Spain), 157 in The basic model was compared to: (1) a first model including HADS at T1, sample 3 (France), and 206 in sample 4 (Germany). (2) a second model based on the first model plus time lapse between the pre- and post-test consultations, (3) a full model based on the second Among eligible counselees, 208 (80.6%) in sample 1 and 108 model plus genetic test results. For samples 1 and 2, addition of the (63.9%) in sample 2 had a personal breast cancer history. In presence or absence of personal breast cancer was added to the first, sample 3 and 4, 200 (100%) and 54 (25%) counselees had a family second and third models. Model selection was based on the BIC. member carrier of a breast cancer pathogenic variant. Statistical analyses were performed with R software (R Core Team, 2020). Among counselees’ respondents, 25 (sample 1), 16 (sample 2), 33 (sample 3) and 31 (sample 4) were informed they were carrying a pathogenic variant. Supplementary Table S4 details comparisons RESULTS of the four samples’ characteristics. A total of 850 eligible counselees were consecutively approached, All 31 clinicians (6 of them being the same in samples 1 and 3) 709 of which (83.4%) agreed to participate (Table 1). Eligible involved in genetic counseling in the three clinics participated in the counselees and non-participants at T1 or at T2 did not significantly study (Supplementary Table S1). Clinicians were mostly female differ by age, having children or, where applicable, disease status (86%), with variable experience in genetic counseling. Among them, European Journal of Human Genetics (2022) 30:1067 – 1075 Table 2. Assessment of genetic-specific psychosocial difficulties by counselees and by clinicians. PAHC domainsa Counselees’ responses Clinicians’ responses Agreement rate b (true Underestimationb (false Overestimationb (false Agreement level “Quite a bit”/ “Minor”/“Important” positives and negatives) positives) Kappa (95% CI) “Very much” negatives) Hereditary 388 (55.0) 335 (50.0) 377 (56.5) 164 (24.6) 126 (18.9) 0.13 (0.05–0.21) predisposition Sample 1 (FR1) 113 (53.1) 82 (41.6) 112 (56.9) 55 (27.9) 30 (15.2) 0.15 (0.02–0.28) Sample 2 (SP) 86 (65.2) 51 (38.4) 78 (59.1) 45 (34.1) 9 (6.8) 0.24 (0.10–0.38) Sample 3 (FR2) 89 (57.1) 66 (42.3) 91 (58.7) 43 (27.7) 21 (13.5) 0.19 (0.05–0.34) Sample 4 (GE) 100 (48.8) 136 (73.9) 96 (52.5) 21 (11.5) 66 (36.1) 0.06 (−0.07–0.18) Practical issues 218 (30.9) 228 (34.0) 373 (56.0) 137 (20.6) 156 (23.4) 0.001 (−0.08–0.08) Sample 1 (FR1) 79 (37.1) 32 (16.2) 118 (59.6) 60 (30.3) 20 (10.1) 0.01 (−0.11–0.13) Sample 2 (SP) 59 (45.4) 36 (27.1) 78 (60.0) 38 (29.2) 14 (10.8) 0.16 (0.01–0.32) Sample 3 (FR2) 45 (28.8) 54 (34.8) 92 (59.7) 26 (16.9) 36 (23.4) 0.08 (−0.08–0.24) Sample 4 (GE) 35 (17.0) 106 (57.6) 85 (46.2) 13 (7.1) 86 (46.7) 0.02 (−0.08–0.12) Familial issues 399 (56.7) 304(45.4) 340 (51.1) 202 (30.4) 123 (18.5) 0.04 (−0.04–0.11) European Journal of Human Genetics (2022) 30:1067 – 1075 Sample 1 (FR1) 145 (68.7) 73 (37.1) 101 (51.8) 78 (40.0) 16 (8.2) 0.12 (0.01–0.23) Sample 2 (SP) 79 (60.3) 58 (43.6) 70 (53.4) 42 (32.1) 19 (14.5) 0.10 −0.06–0.26 Sample 3 (FR2) 103 (66.0) 54 (34.6) 83 (53.6) 60 (38.7) 12 (7.7) 0.15 (0.03–0.28) Sample 4 (GE) 72 (35.0) 119 (64.7) 86 (46.7) 22 (12.0) 76 (41.3) 0.02 (−0.10–0.14) Emotions 288 (40.9) 220 (47.7) 393 (59.0) 116 (17.4) 157 (23.6) 0.17 (0.09–0.25) Sample 1 (FR1) 96 (45.5) 69 (34.8) 127 (64.8) 46 (23.5) 23 (11.7) 0.28 (0.15–0.41) Sample 2 (SP) 52 (39.7) 62 (46.6) 78 (59.5) 22 (16.8) 31 (23.7) 0.18 (0.01–0.35) Sample 3 (FR2) 75 (48.1) 75 (48.1) 100 (64.5) 27 (17.4) 28 (18.1) 0.29 (0.14–0.44) Sample 4 (GE) 65 (31.6) 114 (62.0) 88 (47.8) 21 (11.4) 75 (40.8) 0.04 (−0.08–0.16) A. Brédart et al. Living with cancer 643 (91.5) 355 (53.1) 349 (52.7) 284 (42.9) 29 (4.4) 0.002 (−0.04–0.05) Sample 1 (FR1) 197 (93.4) 83 (42.3) 85 (43.8) 105 (54.1) 4 (2.1) 0.02 (−0.04–0.08) Sample 2 (SP) 129 (98.5) 63 (47.4) 62 (47.3) 68 (51.9) 1 (0.8) −0.002 (−0.04–0.04) Sample 3 (FR2) 133 (85.8) 67 (43.2) 73 (47.7) 72 (47.1) 8 (5.2) 0.04 (−0.06–0.14) Sample 4 (GE) 184 (89.3) 142 (77.2) 129 (70.1) 39 (21.2) 16 (8.7) −0.05 (−0.17–0.06) Children- 369 (70.6) 285 (47.1) 245 (51.5) 172 (36.1) 59 (12.4) 0.04 (−0.04–0.12) related issues Sample 1 (FR1) 142 (78.9) 81 (46.6) 83 (52.2) 63 (39.6) 13 (8.2) 0.06 (−0.06–0.18) Sample 2 (SP) 92 (92.0) 49 (37.7) 45 (45.0) 53 (53.0) 2 (2.0) 0.05 (−0.04–0.13) Sample 3 (FR2) 63 (63.6) 35 (29.4) 55 (61.1) 31 (34.4) 4 (4.4) 0.29 (−0.13–0.44) Sample 4 (GE) 72 (50.0) 120 (65.9) 62 (48.8) 25 (19.7) 40 (31.5) −0.04 (−0.21–0.13) Table entries are number (%) of counselees, except column of concordance level. CI confidence interval. a Genetic-specific difficulties assessed with the Psychosocial Aspects in Hereditary Cancer (PAHC) questionnaire covering six domains of difficulties. b Agreement, under- and overestimation based on data available for both counselees and clinicians. 1071 A. Brédart et al. 1072 Table 3. Counselees’ HADS means (95% CI) and number (%) of counselees with HADS scores >12 at T1 and at T2 by sample. Psychosocial scale or parameter (range of values) Sample 1—FR1 Sample 2—SP Sample 3—FR2 Sample 4—GE Number of counselees at T1 213 133 157 206 Hospital Anxiety and Depression Scale at T1 (HADS) [0–42]— 11.6 (10.8–12.4) 10.0 (8.8–11.2) 10.7 (9.8–11.7) 8.5 (7.6–9.5) Distressa,b Number (%) of counselees with HADS score >12 at T1 89 (42.2) 43 (32.8) 52 (32.5) 41 (19.9) Number of counselees at T2 164 67 118 199 Hospital Anxiety and Depression Scale at T2 (HADS) [0–42]— 11.4 (10.5–12.3) 9.1 (7.3–10.9) 10.5 (9.4–11.6) 7.5 (6.5–8.6) Distressa,c Number (%) of counselees with HADS score >12 at T2 64 (37.4) 19 (28.4) 44 (34.4) 43 (21.5) Across samples for the HADS and assessment times (T1 and T2), <5 missing data were observed and internal consistencies (Cronbach’s alpha coefficients) are very good all above 0.80. Comparisons between samples: a(sample 1 vs. sample 4), p < 0.0001; b(sample 3 vs. sample 4): p value = <0.05, <0.01; c(sample 3 vs. sample 4): p value = <0.01 5 of 10, and 8 of 12 were clinical geneticists with a medical which psychological referral is recommended (>12) [27], ranged background in sample 1 and 3, respectively; all 4 clinicians were from 19.9 to 42.2% across the four samples. genetic counselors in sample 2, and 10 of 11 clinicians were Of the various statistical models that were fitted with distress gynecologists with a genetic training in sample 4. Clinicians did not as outcome, the lowest BIC at both T1 and T2 was obtained for differ by country on age, gender and length of clinical experience. models that included dyad agreement (i.e., TP, FP, FN vs. TN) and samples (1 to 4) in five of the six PAHC domains: for Counselee–clinician agreement example, to predict distress at T1 and testing agreement in the Less than 10% and 11% missing data were observed for “Hereditary predisposition” domain, the statistical models that counselees’ and clinicians’ individual items respectively, except included either samples (basic model), or sample and agree- for children-related issues, which had up to 36% and 25% missing ment (first model), or the interaction between samples and data for counselees’ and clinicians’ items, respectively (Supple- agreement (second model) resulted in BIC estimates of mentary Table S3b). For all domains, answers from both 4417.451, 4382.435 (i.e., lower BIC, first model) and 4432.558, counselees and clinicians were available for 94% of dyads. An respectively; to predict distress at T2, the statistical models that exception was the domain of children-related issues with 67% included either samples (basic model), or samples and dyads. For the different PAHC scales, the overall number of TN, TP, agreement (first model), or samples, agreement and the FP and FN assessments by clinicians were 170, 207, 126, 164 interaction between agreement and time between T1 and T2 (Hereditary predisposition), 302, 71, 156, 137 (Practical issues), 162, (second model), or sample, agreement and the interaction 178, 123, 202 (Familial issues), 233, 160, 157, 116 (Emotions), 26, between the genetic test result and agreement (third model) 323, 29, 284 (Living with cancer), and 72, 173, 59, 172 (Children- resulted in BIC estimates of 3417.414, 3399.314 (i.e., lower BIC, related-issues) (Supplementary Table S3). first model), 3419.991, 3436.978, respectively (see Supplemen- Overall, comparable proportions of difficulties were reported by tary Tables S5 and S6 showing similar results for other PAHC counselees and clinicians for the first four domains (Table 2). domains except “Living with Cancer” for which samples (basic Counselee–clinician agreement per domain ranged from 51.1% model) best explained data.). (familial issues) to 59.0% (emotions). Disagreement between counse- Length of time between the pre- (T1) and post-test (T2) lees and clinicians varied across samples and domains. In samples 1, 2 consultations did not improve the statistical model fit according to and 3, underestimation was more pronounced than overestimation, the BIC parameter. whereas the reverse was observed for sample 4. Underestimation was Table 4 provides predicted mean of HADS and 95% CI at T1 particularly marked in all four samples for the domain of “living with (second column) and at T2 (third column) for counselees for cancer”. In this domain, 91.5% of counselees reported difficulties, but which clinicians’ assessments were TN, TP, overestimation (FP) 53.1% of clinicians did; most differences originated from the or underestimation (FN) for the different PAHC domains. It systematic underestimation of difficulties by clinicians, which resulted shows that mean HADS scores and 95% CI in the four in the low negative predictive value of 7% (i.e., TN/(TN + FN)) (see agreement categories remained comparable from T1 to T2. Supplementary Table S3 for details). Agreement between clinicians For all PAHC domains, clinicians’ adequate perception of low and counselees for the domain of “emotions” ranged from 59.5 to amount of difficulties (TN) as well as overestimation of 64.8% in samples 1, 2 and 3. Underestimation by clinicians was difficulties (FP) at both T1 and at T2 were associated with generally less marked in this domain. HADS scores ranging from 6.6 to 9.4. In contrast, clinicians’ The kappa statistic showed no or only slight agreement adequate perception (TP) of greater amount of difficulties as between counselees’ and clinicians’ perception of difficulties well as underestimation (FN) were associated with significantly across all PAHC domains with Kappa statistics (95% CI) ranging higher HADS scores (p < 0.001), ranging from 9.6 to 14.8. from 0.001 (−0.08–0.08) (practical issues) to 0.17 (0.09–0.25) Differences in HADS scores for the domain “living with cancer” (emotions). No agreement (95% CI including 0) was found for were less consistent owing to the small number of counselees “practical issues”, “familial issues”, “living with cancer”, and in the TN and FP categories. “children-related issues”. Kappa was fair (>0.20) for “hereditary The largest difference between categories in HADS scores were predisposition” (sample 2), “emotions” (samples 1 and 3) and observed for the PAHC “emotions” domain. Five to seven point- “children-related issues” (sample 3). differences in HADS scores were noticeable between the TP or FP categories, and the TN or FN categories. Meanwhile in other Effect of counselee–clinician agreement on distress after pre- domains, differences between these pairs of categories did not and post-test consultations exceed five-point scores. Moreover, the lower CI limits of mean HADS showed fairly stable distress levels from T1 to T2 (Table 3). HADS scores in the TP of the “emotions” domain exceeded 12 at Proportions of women with HADS scores above the threshold at T1 and T2. The picture was nearly equivalent for underestimation. European Journal of Human Genetics (2022) 30:1067 – 1075 A. Brédart et al. 1073 (pathogenic variant) to 10.0 (8.4–10.9) (uninformative result) at T2 Table 4. HADS predicted means (95% confidence interval) for distress (Supplementary Tables S7 and S8). after the pre-test (T1) and after the post-test (T2) consultations according to counselee–clinician agreement in perceived genetic- specific psychosocial difficulties in statistical models best fitting DISCUSSION the data. This large prospective observational study in the context of breast Distress at T1 Distress at T2 cancer genetic testing in three European clinical settings is the PREDICTOR (number of counselees by HADS predicted HADS predicted first to address the ability of genetic clinicians to identify categories of counselee–clinician agreement) means (95% CI) means (95% CI) PAHC hereditary predisposition counselees with psychosocial difficulties. True negative (n = 170) 7.7 (6.7–8.7) 7.0 (5.7–8.2) Agreement rates between psychosocial difficulties reported by True positive (n = 207) 12.1 (11.2–13.0)** 11.2 (10.1–12.4)** counselees and clinicians’ ratings of these difficulties were Overestimation (n = 126) 9.0 (7.8–10.2) 8.8 (7.4–10.3)* generally low, although better than rates reported by previous Underestimation (n = 164) 11.3 (10.3–12.3)** 11.0 (9.7–12.2)** oncology (non-genetic) studies (i.e., 26–38% (health information) PAHC practical issues and 28–40% (psychological difficulties) [28, 29] compared to True negative (n = 302) 8.9 (8.1–9.8) 8.4 (7.4–9.5) 51–59% in this study). However, clinicians did under-estimate True positive (n = 71) 12.6 (11.1–14.1)** 12.6 (10.7–14.6)** difficulties in up to 42.9% of counselees. Overestimation (n = 156) 9.4 (8.3–10.6) 8.5 (7.2–9.9) As the role endorsed by genetic clinicians places greater Underestimation (n = 137) 12.7 (11.6–13.9)** 11.7 (10.3–13.0)** emphasis on biomedical information [30], we expected that their PAHC familial issues ability to identify counselees’ difficulties would be noticeable in True negative (n = 162) 7.7 (6.7–8.7) 7.3 (6.0–8.6) the domain of hereditary predisposition (e.g., counselees worrying True positive (n = 178) 12.7 (11.7–13.6)** 12.0 (10.7–13.2)** about the chance of being a carrier of a genetic mutation). Overestimation (n = 123) 7.4 (6.1–8.6) 6.6 (5.1–810) However, this was only found in the Spanish sample. Underestimation (n = 202) 11.7 (10.8–12.7)** 11.4 (10.1–12.6)** Difficulties in the “emotions” domain were better identified, PAHC emotions especially by clinicians from the French setting (65%). Emotions True negative (n = 233) 7.4 (6.6–8.2) 7.1 (6.0–8.1) may be more easily recognizable [31, 32] than difficulties in other True positive (n = 160) 14.8 (13.9–15.7)** 13.6 (12.5–14.8)** domains which may require more in-depth evaluation of the Overestimation (n = 157) 7.4 (6.5–8.4) 7.4 (6.2–8.7) personal and familial context. Underestimation (n = 116) 13.3 (12.2–14.3)** 11.9 (10.5–13.2)** Distress scores were low on average across samples. However, PAHC living with cancer around 20–42% of counselees displayed distress levels eligible for True negative (n = 26) 6.6 (4.01–9.09) 7.1 (4.0–10.2) psychological referral. Regardless of samples, time between pre- True positive (n = 323) 11.0 (10.2–11.8)** 10.1 (9.0–11.1) and post-test consultations and genetic test results, Overestimation (n = 29) 7.4 (5.0–9.4) 7.1 (4.3–10.0) counselee–clinician agreement on psychosocial difficulties best Underestimation (n = 284) 10.0 (9.1–10.8)* 9.6 (8.4–10.7) explained distress levels after consultations. However, counselees’ PAHC children-related issues distress remained stable between the pre- and post-test True negative (n = 72) 7.8 (6.3–9.4) 7.3 (5.3–9.3) consultations and this appeared independent from clinicians’ True positive (n = 173) 11.7 (10.7–12.8)** 11.9 (10.5–13.3)** adequate or inadequate perception of the presence or absence of Overestimation (n = 59) 6.7 (4.9–8.5) 6.7 (4.5–8.8) difficulties. Counselees experienced higher and persisting distress Underestimation (n = 172) 10.9 (9.8–12.0)** 10.4 (8.9–11.8)** levels when their difficulties were underestimated or when they were correctly identified, suggesting that both improving T1 = within 1 month after the pre-test consultation; T2 = within 3 months clinicians’ appraisals and their response to counselees’ difficulties after the post-test consultation. Table entries refer to predicted mean values of HADS and 95% confidence interval at T1 (second column) and at in terms of psychological referral should be promoted. Accurate T2 (third column) for counselees for which clinicians’ assessments were appraisal of the presence of difficulties by clinicians was not true negative, true positive, overestimation (false positive) or under- associated to lower distress levels, suggesting a lack of clear estimation (false negative) for the different PAHC domains. Statistical clinical pathways between detection of emotional difficulties and significance tests taking “True negative” in agreement as the reference intervention in these cancer genetic clinical setting [11]. category. Best models selected based on Bayesian information criterion Furthermore, all counselees considered together, distress (BIC) estimates includes the random effect of clinicians, samples and seemed to remain unaffected by the disclosure of the genetic agreement, except for the PAHC domain “living with cancer” where the test result. This counterintuitive finding has already been observed basic model is best. in other cancer genetics studies [6, 33, 34]. It suggests that FR France, GE Germany, SP Spain. *,**p values <0.05; <0.001. psychological referral should not rely on the genetic test result criterion only. Genetic counseling is meant to minimize distress experienced Hence for the “emotions” domain, most counselees falling in the by counselees [35]. However, addressing psychological concerns is TP and FN categories experienced anxious or mood symptoms rarely on the genetic consultations agenda [36]. In our study, a associated with a HADS scores above 12. psychological consultation was proposed to counselees depend- The modest inter-clinicians’ variability on HADS scores (ICC of ing on clinician’s personal judgment. 0.04 or below in retained models) indicates similar distress levels As mentioned in the introduction, a randomized trial in the between counselees across clinicians. Netherlands tested the usefulness of the PAHC questionnaire in routine practice [14]. This questionnaire was completed by Effect of genetic test results counselees allocated to the intervention group before the The communication of the genetic test result, i.e., a pathogenic consultation but not by the control group. After evaluating the variant on a high breast cancer susceptibility gene such BRCA1 or PAHC responses, clinicians undertook initiatives to discuss overall BRCA2, a negative or an uninformative test result, or a VUS psychosocial difficulties with counselees. However, this interven- (samples 1 and 2) had no significant influence on distress tion reduced counselees’ distress levels, by −1.4 (95% CI: 0.2–2.6) measured after the post-test consultation: HADS predicted means point of mean HADS, almost negligible compared to the 4 to 7 (95% CI) ranged from 8.8 (6.9–10.8) (negative result) to 10.8 average difference in mean HADS between counselees in the TN (10.0–12.0) (uninformative result) at T1 and from 8.3 (7.2–9.4) vs. FN categories found in our study. Of note, a two-point score European Journal of Human Genetics (2022) 30:1067 – 1075 A. Brédart et al. 1074 difference is the minimal clinically important difference in HADS 3. Hiraki S, Rinella ES, Schnabel F, Oratz R, Ostrer H. Cancer risk assessment using scores [37]. genetic panel testing: considerations for clinical application. J Genet Couns. In contrast, this study found large differences (>5-point HADS) 2014;23:604–17. in distress levels between counselees with more and less 4. Culver JO, Ricker CN, Bonner J, Kidd J, Sturgeon D, Hodan R, et al. Psychosocial outcomes following germline multigene panel testing in an ethnically and eco- difficulties for the PAHC “emotions” domain. Feasible approaches nomically diverse cohort of patients. Cancer. 2021;127:1275–85. may be suggested to help clinicians appraise the severe difficulties 5. Eijzenga W, Bleiker EM, Hahn DE, Kluijt I, Sidharta GN, Gundy C, et al. Psychosocial requiring psychological referral, such as questions focusing on Aspects of Hereditary Cancer (PAHC) questionnaire: development and testing of emotions to be completed by counselees before the pre-test a screening questionnaire for use in clinical cancer genetics. Psychooncology. consultation, and multidisciplinary team discussions involving 2014;23:862–9. psychological experts to ensure appropriate psychological care. 6. Yanes T, Willis AM, Meiser B, Tucker KM, Best M. Psychosocial and behavioral outcomes of genomic testing in cancer: a systematic review. Eur J Hum Genet. Strengths and limitations 2019;27:28–35. The present study has limitations. Generalization is limited as 7. Meiser B. Psychological impact of genetic testing for cancer susceptibility: an update of the literature. Psychooncology. 2005;14:1060–74. samples were constituted of counselees’ willing to be tested, from 8. Beran TM, Stanton AL, Kwan L, Straatsma BR, Burgess BL. The trajectory of psy- only one genetic clinic per country, at specific time points of the chological impact in BRCA1/2 genetic testing: does time heal? Ann Behav Med. cancer genetic testing trajectory (after the pre- and post-test 2008;36:107–16. consultations). Samples 1 and 3 comprised mostly highly 9. Bredart A, Kop JL, Depauw A, Caron O, Sultan S, Leblond D, et al. Short-term educated women. However, sample biases were accounted for psychological impact of the BRCA1/2 test result in women with breast cancer in statistical models. Moreover, in contrast to other studies [29, 38], according to their perceived probability of genetic predisposition to cancer. Br J levels of agreement were similar across clinicians regardless of Cancer. 2013;108:1012–20. samples and their cultural and health care delivery differences. 10. Jacobs C, Patch C, Michie S. Communication about genetic testing with breast Collection of distress data before the initial consultation was not and ovarian cancer patients: a scoping review. Eur J Hum Genet. 2019;27:511–24. 11. Butow P, Price MA, Shaw JM, Turner J, Clayton JM, Grimison P, et al. Clinical possible as women could have been sensitized by the cancer pathway for the screening, assessment and management of anxiety and genetic context at any time before the study took place. Because depression in adult cancer patients: Australian guidelines. Psychooncology. clinicians had variable encounters with counselees with different 2015;24:987–1001. personal and familial history, we could not explore in depth the 12. Meiser B, Irle J, Lobb E, Barlow-Stewart K. Assessment of the content and process relationships between clinicians’ characteristics and agreement. of genetic counseling: a critical review of empirical studies. J Genet Couns. Counselee–clinician verbal and non-verbal interactions during the 2008;17:434–51. consultation should be investigated to understand communica- 13. Maheu C, Bouhnik AD, Nogues C, Mouret-Fourme E, Stoppa-Lyonnet D, Lasset C, tion factors underlying agreement or disagreement in perceived et al. Which factors predict proposal and uptake of psychological counselling psychosocial difficulties [39]. after BRCA1/2 test result disclosure? Psychooncology. 2014;23:420–7. 14. Eijzenga W, Aaronson NK, Hahn DE, Sidharta GN, van der Kolk LE, Velthuizen ME, This study rests on a large sample of participants, including all et al. Effect of routine assessment of specific psychosocial problems on perso- clinicians of each genetic clinic, a high rate of respondents among nalized communication, counselors’ awareness, and distress levels in cancer counselees, and little differences between respondents and non- genetic counseling practice: a randomized controlled trial. J Clin Oncol. respondents. It was performed in three European country settings, 2014;32:2998–3004. reflecting different genetic testing practices. The study included 15. Biesecker B, Lewis MA. Advancing genomic translation: Investigations in com- counselees routinely met in cancer genetic clinics, either for munication. A special series on communication research in the context of predictive or diagnostic genetic testing for breast cancer, and is genomic medicine. Patient Educ Couns. 2021;104:933–4. particularly relevant considering the rapidly evolving cancer 16. Street RL Jr. How clinician-patient communication contributes to health genetic testing field. improvement: modeling pathways from talk to outcome. Patient Educ Couns. 2013;92:286–91. To conclude, these findings suggest that referral to psycholo- 17. Bredart A, Kop JL, Dick J, Cano A, De Pauw A, Anota A, et al. Psychosocial gical assessment and support should not be based on the genetic problems in women attending French, German and Spanish genetics clinics test result criterion. Enhancing assessment of emotional difficulties before and after targeted or multigene testing results: an observational pro- particularly would provide more relevant guidance for appraising spective study. BMJ Open. 2019;9:e029926. psychological needs. This could be achieved through the 18. Moretta J, Berthet P, Bonadona V, Caron O, Cohen-Haguenauer O, Colas C, et al. systematic implementation of questions focusing on emotions [The French Genetic and Cancer Consortium guidelines for multigene panel to be answered by counselees before the pre-test consultation, analysis in hereditary breast and ovarian cancer predisposition]. Bull Cancer. and by multidisciplinary team discussions involving psychological 2018;105:907–17. experts. 19. Llort G, Chirivella I, Morales R, Ales-Martinez JE, Andres R, Balmana J, et al. SEOM clinical guidelines in Hereditary Breast and ovarian cancer. Clin Transl Oncol. 2015;17:956–61. 20. Bick U, Engel C, Krug B, Heindel W, Fallenberg EM, Rhiem K, et al. High-risk breast DATA AVAILABILITY cancer surveillance with MRI: 10-year experience from the German consortium The study database is hosted at Institut Curie (France). It may be available after main for hereditary breast and ovarian cancer. Breast Cancer Res Treat. publication for BRIDGES research program performed (in 2023). 2019;175:217–28. 21. Bredart A, Anota A, Dick J, Cano A, De Pauw A, Kop JL, et al. The “Psychosocial Aspects in Hereditary Cancer” questionnaire in women attending breast cancer CODE AVAILABILITY genetic clinics: Psychometric validation across French-, German- and Spanish- The data generated and analyzed during this study can found within the article and language versions. Eur J Cancer Care. 2020;29:e13173. its Supplementary files. Additional data are available from the corresponding author 22. Razavi D, Delvaux N, Farvacques C, Robaye E. Screening for adjustment disorders on reasonable request. and major depressive disorders in cancer in-patients. Br J Psychiatry. 1990;156:79–83. 23. Herrmann-Lingen CBU, Snaith RP. HADS-D: Hospital Anxiety and Depression Scale REFERENCES —German; Manual. Bern: Hans Huber; 2011. 1. Dorling L, Carvalho S, Allen J, González-Neira C, Luccarini C, Wahlström KA, et al. 24. Herrero MJ, Blanch J, Peri JM, De Pablo J, Pintor L, Bulbena A. A validation study Breast cancer risk genes—association analysis in more than 113,000 women. N of the hospital anxiety and depression scale (HADS) in a Spanish population. Gen Engl J Med. 2021;384:428–39. Hospital Psychiatry. 2003;25:277–83. 2. Esteban I, Vilaro M, Adrover E, Angulo A, Carrasco E, Gadea N, et al. Psychological 25. Viera AJ, Garrett JM. Understanding interobserver agreement: the kappa statistic. impact of multigene cancer panel testing in patients with a clinical suspicion of Fam Med. 2005;37:360–3. hereditary cancer across Spain. Psychooncology. 2018;27:1530–7. 26. Gideon S. Estimating the dimension of a model. Ann Stat. 1978;6:461–4. European Journal of Human Genetics (2022) 30:1067 – 1075 A. Brédart et al. 1075 27. Singer S, Kuhnt S, Götze H, Hauss J, Hinz A, Liebm A, et al. Hospital anxiety and research site (SiRIC) under agreement INCa-DGOS-Inserm_12554. The funder of the depression scale cutoff scores for cancer patients in acute care. Br J Cancer. study had no role in study design, data collection, data analysis, data interpretation, 2009;100:908–12. or writing of the report. 28. Newell S, Sanson-Fisher RW, Girgis A, Bonaventura A. How well do medical oncologists’ perceptions reflect their patients’ reported physical and psychosocial problems? Data from a survey of five oncologists. Cancer. 1998;83:1640–51. COMPETING INTERESTS 29. Shin DW, Kim SY, Cho J, Sanson-Fisher RW, Guallar E, Chai GY, et al. Discordance DSL declares the following conflict of interest: AstraZeneca is funding the French in perceived needs between patients and physicians in oncology practice: a COVAR study coordinated by the Institut Curie and aiming to the classification of nationwide survey in Korea. J Clin Oncol. 2011;29:4424–9. BRCA1/2 Variants of Unknown Significance. The Department of Genetics of the 30. Butow PN, Lobb EA. Analyzing the process and content of genetic counseling in Institut Curie is performing genotyping (PRS) of MammoRisk®, Predilife. DSL and the familial breast cancer consultations. J Genet Couns. 2004;13:403–24. Institut Curie have received honorarium for her participation in education meetings 31. Gouveia L, Lelorain S, Bredart A, Dolbeault S, Bonnaud-Antignac A, Cousson-Gelie organized by AstraZeneca or Tesaro®. RS declares the following conflict of interest: F, et al. Oncologists’ perception of depressive symptoms in patients with grants or contracts received from AstraZeneca, Amgen. Payment or honoraria advanced cancer: accuracy and relational correlates. BMC Psychol. 2015;3:6. received from AstraZeneca, JanssenCilag. Participation on an Advisory Board for 32. Passik SD, Dugan W, McDonald MV, Rosenfeld B, Theobald D, Edgerton S. AstraZeneca; MSD, GSK, Clovis Oncology. PD reports grants from EU Horizon 2020 Oncologists’ recognition of depression in their patients with cancer. J Clin Oncol. program during the conduct of the study. All other authors declare no conflict of 1998;16:1594–1600. interest. 33. Vos J, Oosterwijk JC, Gomez-Garcia E, Menko FH, Collee MJ, van Asperen CJ, et al. Exploring the short-term impact of DNA-testing in breast cancer patients: the counselees’ perception matters, but the actual BRCA1/2 result does not. Patient ETHICS APPROVAL AND CONSENT TO PARTICIPATE Educ Couns. 2012;86:239–51. The study protocol was approved in France by the Comité consultatif sur le 34. Lombardi L, Bramanti SM, Babore A, Stuppia L, Trumello C, Antonucci I, et al. traitement de l’information en matière de recherche dans le domaine de la santé Psychological aspects, risk and protective factors related to BRCA genetic testing: (CCTIRS: Consultative committee for information management in health research— a review of the literature. Support Care Cancer. 2019;27:3647–56. No. 16.314) and the Comité de Protection des Personnes Ile-de-France V (CPP—No. 35. Biesecker BB. Goals of genetic counseling. Clin Genet. 2001;60:323–30. 18.12.28.38743 CAT2), in Germany by the Ethics Committee of the University Hospital 36. Biesecker BB, Lillie SE, Amendola LM, Donohue KE, East KM, Foreman AKM, et al. of Cologne (No. 16-098) and in Spain by the Ethics Committee of the Instituto Catalán A review and definition of ‘usual care’ in genetic counseling trials to standardize de Oncología of Barcelona (No. PR111/16). All recruited women provided written use in research. J Genet Couns. 2021;30:42–50. informed consent. 37. Wynne SC, Patel S, Barker RE, Jones SE, Walsh JA, Kon SS, et al. Anxiety and depression in bronchiectasis: Response to pulmonary rehabilitation and minimal clinically important difference of the Hospital Anxiety and Depression Scale. ADDITIONAL INFORMATION Chronic Respiratory Dis. 2020;17:1479973120933292. Supplementary information The online version contains supplementary material 38. Söllner W, DeVries A, Steixner E, Lukas P, Sprinzl G, Rumpold G, et al. How available at https://doi.org/10.1038/s41431-022-01096-9. successful are oncologists in identifying patient distress, perceived social support, and need for psychosocial counselling? Br J Cancer. 2001;84:179–85. Correspondence and requests for materials should be addressed to Anne Brédart. 39. Medendorp NM, Hillen MA, Visser LNC, Aalfs CM, Duijkers FAM, van Engelen K, et al. A randomized experimental study to test the effects of discussing uncer- Reprints and permission information is available at http://www.nature.com/ tainty during cancer genetic counseling: different strategies, different outcomes? reprints Eur J Hum Genet. 2021;29:789–99. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ACKNOWLEDGEMENTS We are grateful to all participants, clinicians and counselees, who provided their time to complete the questionnaires. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, AUTHOR CONTRIBUTIONS adaptation, distribution and reproduction in any medium or format, as long as you give AB, JLK, PD, DSL, RS and SD designed the project. AB, AT, AC, ADP, JD, KR, DSL, RS and appropriate credit to the original author(s) and the source, provide a link to the Creative SD collected data. AB, JLK, AT, KR, DSL, PD, RS and SD analyzed, interpret data and Commons licence, and indicate if changes were made. The images or other third party draft the work. All authors revised and provided final approval of the version to be material in this article are included in the article’s Creative Commons licence, unless published. indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. FUNDING org/licenses/by/4.0/. This project has received funding from the European Union Horizon 2020 research and innovation program under grant agreement No. 634935 (BRIDGES). In France, at Institut Curie, this work was partly financed, within the designated integrated cancer © The Author(s) 2022 European Journal of Human Genetics (2022) 30:1067 – 1075