Screening patients with metastatic breast cancer for psychiatric disease: a cross sectional study

Study design and procedure

We conducted a cross-sectional, single-center study. The convenience sample was designed with inclusion criteria consisting of female patients 18 years of age or older, diagnosed with metastatic breast carcinoma under palliative treatment with chemotherapy, hormone therapy, cyclin inhibitors, or targeted anti-HER2 therapy, between November 2020 and May 2021, at the Medical Oncology Department of the Instituto Português de Oncologia de Coimbra Francisco Gentil, E.P.E. Exclusion criteria for the study were known brain metastasis, baseline cognitive impairment, refusal to participate in the study, and incomplete clinical records. The final sample size after included 91 of the initial 128 participants evaluated.

Patients were recruited at the outpatient daycare unit when they were present for treatment. The application and filling-out of the various scales (Mini Mental State Examination (MMSE), HADS, BSI, and WHOQoL-Bref instruments) was conducted by the first author before treatment in order to minimize bias that could be developed in the post-treatment period. MMSE was the first instrument to be used, since a score compatible with cognitive impairment would represent and exclusion criteria. Afterwards, HADS, BSI and WHOQoL-Bref scales were applied through direct questioning of the patients and presentation of various possibilities of response, with repetition of each if solicited. The referral for the Psychiatry appointment was made whenever the HADS scale exceeded the cut-off score of 8 in each part of the instrument¹⁷ or when the BSI exceeded the cut-off of 1.7.^18,19 The referred patients were posteriorly evaluated by an element of the psychiatric team consisting of two specialists, ideally within a month after the application of scales and at times within a week if the screening determined an elevated psychopathological risk. Determinants of elevated risk or severity included suicidal ideation and depressive symptoms with significant impact on daily functioning. The results, including possible diagnoses and eventual necessity of psychiatric medication from this evaluation were determined at this observation. The codification of psychiatric diagnoses was carried out using the International Classification of Diseases, Tenth Revision (ICD-10).

The following complementary data were collected either by direct questioning of the patient or consultation of the electronic files: gender; age; Eastern Cooperative Oncology Group – Performance Status (ECOG-PS); gynecological history; obstetric history; psychiatric; social history and characteristics of oncological disease.

The primary endpoints were assessing prevalence and type of psychiatric disorder and determining its relationship with characteristics of the oncological disease and treatment. Regarding secondary endpoints, quality of life (QoL) was evaluated.

Instruments

Various scales were applied including MMSE which is the most widespread cognitive impairment/dementia screening tool.²⁰ It is a brief 30-item measure that assesses six cognitive domains: visuospatial construction, language function, attention and concentration, working memory, memory recall, and orientation to time and place.²¹ Each item on the test is scored as 0 or 1 with the total score varying between 0 and 30. The MMSE utilized in this study was adapted to the Portuguese population.²² In this study, the cut-off points were based on the educational level of the Portuguese population, where cognitive impairment was present if ≤1 points if illiterate, ≤22 if 1-11 years of formal education and ≤27 if more than 11 years of school.

The Hospital Anxiety and Depression Scale (HADS) helps the diagnosis of anxiety and depression in patients without previous psychiatric history. This instrument excludes symptoms of severe psychiatric disease and physical disease and it is divided into two parts, anxiety and depression, with each sub-section consisting of seven questions. These two parts are classified separately and cover several domains with control of possible confounding factors.²³ A total subscale score of eight or more points out of 21 denotes considerable symptoms of anxiety or depression. The scale applied in this study was validated for the Portuguese population.²⁴

The Brief Symptom Index (BSI) instrument consists of 53 items which the subject must rate how much a certain problem has affected them within the last week. It is a smaller version of the instrument Symptom Checklist-90-Revised (SCL-90-R) and the classification is based on a Likert-type scale (a psychometric scale). An approach to nine symptom dimensions is made (somatization, obsession-compulsion, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, and psychoticism) and three global indices of distress (Global Severity Index, Positive Symptom Distress Index, and Positive Symptom Total. The cutoff point for considering the existence of positive symptoms/pathology is above 1.7.¹⁹ An adapted and validated scale aimed at the Portuguese population was utilized.²⁵

To assess QoL and due to the need for an instrument that would be easily applicable in clinical trials or large-scale epidemiological studies, the World Health Organization (WHO) created the WHO-QoL Bref, which encompasses four domains related to quality of life (physical, psychological, social, and environmental). The score can be calculated in all or each domain, and QoL is better the higher the score.²⁶ An adapted and translated scale was applied to the sample.^27,28

Statistical analysis

Data were analyzed using SPSS statistics (IBM SPSS Data Editor Version 25.0 IBM Corp, Armonk, NY) software. Descriptive statistics (median and range for continuous variables, frequency and percentages for categorical variables) and univariate analysis (Mann-Whitney U Test, linear regression, one-way ANOVA and independent sample T-test for continuous data, chi-square and Fisher exact test for categorical data) were utilized, presenting data on the odds ratio (OR) and 95% confidence interval (CI), p≤0.05 was considered significant.

Ethics approval

The study was reviewed and approved by the Ethics Committee of the Instituto Português de Oncologia de Coimbra Francisco Gentil, E.P.E (02/2020), and performed following the ethical standards of the 1964 Declaration of Helsinki,²⁹ as well as its subsequent amendments. Upon inclusion, the study was carefully explained to all patients and an informed consent form was presented and signed by the patients who agreed to participate.

Demographic sample characterization

A total of 128 patients were evaluated and 91 included (37 patients refused to participate because they did not consider this evaluation necessary). All patients were female, median age was 59.79 years at study inclusion, 49.90 years at breast cancer diagnosis and ECOG-PS was 0 in 62.6% of patients.

Regarding gynecological and obstetric history, menarche was between 10 and 12 years old in 58.3%; 69.2% of patients had used oral contraceptives at some point in their lives; 96.7% of patients had children and 64.8% had their first child before age 30; 75.8% of patients were postmenopausal or under ovarian function suppression at the time of data collection.

A history of psychiatric disorder was present in 16.5%: anxiety was the most frequent disease (7.7%), 16.5% were using psychiatric medication and 6.6% were already followed in a Psychiatry consultation. Regarding habits, 73.6% were non-smokers and 65.9% did not have alcoholic habits.

The most frequent molecular subtype was Luminal B HER2 negative (47.3%). At diagnosis 60.4% of patients had non-metastatic tumors, 19.8% underwent neoadjuvant chemotherapy, 51.6% underwent surgery with curative intent – more frequently mastectomy with axillary dissection (25.3%), and 34.1% underwent adjuvant chemotherapy. At the time of data collection 26.3% of patients were under first-line of palliative treatment, 51.8% under second-line, and 21.9% under third-line. Treatment consisted in hormonal therapy in 69.19% of patients, chemotherapy in 36.2%, chemotherapy with anti-HER2 therapy in 31.8%, hormonal therapy with cyclin inhibitors 4/6 in 20.8%, hormonal with anti-HER2 therapy in 9.89% and anti-HER2 therapy in 5.49%. Grade 3-4 toxicity had occurred in 32.9% of patients. Patient characteristics can be consulted in Table 1.

Prevalence of psychiatric disorder

MMSE scale application showed that none of the patients had cognitive impairment. HADS scale results showed that 18.7% of the patients scored positively for anxiety and 17.6% for depression. BSI scale application showed that 2.2% of patients surpassed the cut-off of 1.7, presenting positive symptoms. A total of 17 (18.6%) patients fulfilled HADS (n=17) or BSI criteria (n=2) for Psychiatric referral, of these, only 52.9% were consulted by a psychiatrist since the remaining patients refused observation.

There was a statistically significant relationship between an anxiety subscale score of >8 (HADS) and ovarian function suppression (median 12.40, p<0.001), use of neoadjuvant therapy regardless of the type of treatment (median 7.47, p<0.001), and use of hormonal and anti-HER2 therapy in tsecond-line of palliative treatment (median 7.44, p=0.024).

The anxiety subscale general score was higher in postmenopausal patients (median 6.53, p=0.002), particularly in patients with ovarian function suppression (median 12.40, p<0.001). It was also higher in patients with stage IV disease at diagnosis (median 6.86, p=0.045), treated with hormonal therapy as first-line (median 6.64, p=0.041), treated with anti-HER2 therapy and chemotherapy as second-line (median 11, p=0.044), and treated with chemotherapy as third-line (median 7.82, p=0.020).

Regarding the depression subscale score >8 of HADS, there was a statistically significant relationship with previous mastectomy with axillary dissection (median 7.96, p=0.022), HER2-negative Luminal B tumors (median 6.52, p=0.020), and grade 3-4 toxicities in the first (median 6.44, p=0.018) and third-line treatments regardless of the type of treatment (median 9.43, p=0.031). The depression subscale general score of HADS was also higher in older patients (median 6.43, p=0.006), postmenopausal patients (median 6.37, p=0.007), and in those treated with chemotherapy as third-line (median 10, p< 0.001).

Regarding BSI score >1.7, there was a statistically significant relationship with moderate alcohol habits (p=0.013). Details concerning the HADS and BSI scores can be consulted in Table 2.

Psychiatric diagnosis

Of the 17 patients with referral criteria for a Psychiatry appointment (17 scored positively on the HADS scale and 2 on the BSI scale), only 9 patients accepted the referral and of these all were diagnosed with some type of psychiatric disorder.

Of the nine patients who accepted referral for psychiatric consultation, 55.55% were diagnosed with dysthymic disorder (F34.1), 33.33% with depression (F33), and 11.11% with adjustment disorder (F43.2). All patients were considered to have indication from treatment with an antidepressant, 44.44% required an association with an anxiolytic, and 22.22% with an antipsychotic.

Quality of life assessment

The median total score on the WHOQoL-Bref was 92.33 points, the environment domain was the highest with a median score of 80.53, followed by the social relationships domain with 77.89, psychological domain with 77.05, and physical domain with 69.79.

The better the ECOG-PS the higher the WHOQoL-Bref score, particularly in the physical and social relationships domains (p<0.001). Menopause also seemed to influence physical, psychological and social relationships domains, particularly the physical domain, in which premenopausal women had a higher median scored postmenopausal women (median 79.14 versus 66.81, p<0.001). On the other hand, women who underwent ovarian function suppression scored lower in the psychological domain (median 57.83 versus 74.80 for natural menopause vs 80.94 for surgical menopause, p=0.002). The absence of previous psychiatric history/medication seemed to influence the environmental domain (median 81.58, p=0.002), which was also related to alcoholic habits with abstinent patients scoring higher (median 80.13, p=0.018). Patients with initial stage IV tumors scored lower in all domains, except social relationships (median 78.31, p=0.007). Patients under hormone therapy and cyclin inhibitors 4/6 as second-line treatment scored higher in all domains, particularly the environment domain (median 89.29, p<0.001). Occurrence of grade 3-4 toxicities also seemed to affect scores, particularly the physical domain in patients under first-line treatment (median 61.75, p=0.016), social relationships domain in second-line (median 88,60, p<0.001) and third-line (p=0.009), and environment domain in third-line (median 83.71, p=0.013). Details concerning the scores of WHO-QOL Bref are shown in Table 3.

Study limitations

Our study has some limitations: the fact that it is a single-center study limits the generalization of our findings, and our sample of patients is heterogenous, since we included patients under different treatments and at different stages of advanced disease. We were also subjected to selection bias by selecting our patients at the outpatient daycare unit where most patients are present for treatment with subcutaneous or intravenous treatments such as ovarian suppression medications and chemotherapy, which translates into a younger sample of patients than expected for patients with breast cancer.

Clinical implications

Screening for psychological and psychiatric disorders requires a specific time allocated to the oncology consultation or more time available for patient observation, which may not be possible in all cancer centers and departments. This highlights the importance of screening tools that are simple and quick to apply, such as the HADS scale. On the other hand, the existence of a Psychiatry department with a quick response capacity and patient orientation is necessary in order to guarantee timely and adequate evaluation as well as detection and intervention of relevant psychopathology which may influence patient orientation and response.

Underlying data

Figshare: Database.xlsx, https://doi.org/10.6084/m9.figshare.18364970.v1.

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).