The mental health impacts of receiving a breast cancer diagnosis: A meta-analysis

Abstract

Background

Receiving a breast cancer diagnosis can be a turning point with negative impacts on mental health, treatment and prognosis. This meta-analysis sought to determine the nature and prevalence of clinically significant psychological distress-related symptoms in the wake of a breast cancer diagnosis.

Methods

Ten databases were searched between March and August 2020. Thirty-nine quantitative studies were meta-analysed.

Results

The prevalence of clinically significant symptoms was 39% for non-specific distress (n = 13), 34% for anxiety (n = 19), 31% for post-traumatic stress (n = 7) and 20% for depression (n = 25). No studies reporting breast cancer patients’ well-being in our specific time frame were found.

Conclusion

Mental health can be impacted in at least four domains following a diagnosis of breast cancer and such effects are commonplace. This study outlines a clear need for mitigating the impacts on mental health brought about by breast cancer diagnosis. CRD42020203990.

Background

Among women, breast cancer is the most common type of cancer, with more than 2 million cases globally in 2018 [1]. Breast cancer can change the course of one’s life due, among other things, to the immense uncertainties brought about by the course of the disease, the treatment, the prognosis, as well as the profound changes to body image [2, 3]. As a consequence, patients could experience difficulties in many important areas, such as their social relationships [2, 3].

This life-changing experience induces significant psychological distress for many patients [4]. Poor mental health can be caused or exacerbated by the diagnosis itself [5, 6]. For many survivors from primary or recurrent breast cancer, receiving the actual cancer diagnosis was the most difficult moment of the whole process [2, 7]. In recent years, the remission rates have increased up to 96% in North America for primary breast cancer [8]. This increase in breast cancer survival also means, however, that some survivors live with enduring poor quality of life and psychological distress [8].

In spite of several publications on this topic, the nature and frequency of psychological symptoms experienced by breast cancer patients remain unclear. In oncology, psychological distress is broadly defined as a negative experience of a cognitive, behavioural, emotional, social, spiritual or psychological nature interfering with the ability to cope with cancer [3, 9]. Distress experienced by cancer patients can range from mild (i.e., common feelings of vulnerability) to disabling and can manifest in various ways [4]. Some studies have shown the increased risk of stress-related disorders immediately after the diagnosis of breast cancer, highlighting that diagnosis and treatment are two distinct distressing periods that ought to be each studied [10]. Although diagnosis-specific distress is distinct from distress pertaining to the overall cancer experience, many existing studies do not make the distinction between reactions to the acute post-diagnosis phase and more global responses to cancer [4].

Diagnosis-specific distress in breast cancer patients persists and has implications for outcomes, including treatment adherence [4]. Three reviews suggest that symptoms of anxiety in newly diagnosed breast cancer patients are pervasive and persistent regardless of the course of cancer treatment [11–13]. Another review suggests that >22% of women develop depressive symptoms following their diagnosis of breast cancer [12]. Depressive symptoms assessed after a cancer diagnosis are associated with a poor prognosis and a higher risk of mortality [14]. Furthermore, there is evidence that the life-threatening nature of a breast cancer diagnosis can elicit peritraumatic distress (i.e., emotional, behavioural and physiological responses that occur during and immediately after a highly stressful event) [15–17] and symptoms congruent with post-traumatic stress disorder (PTSD) in up to 32% of the women afflicted [18, 19]. Although these results are striking and require attention, they nonetheless highlight that the majority of patients experience subclinical distress symptoms and cope with the diagnosis and treatments relatively well [20]. Specifically, one study showed that despite the presence of distress symptoms, many patients reported higher levels of well-being following diagnosis compared to pre-diagnosis [20, 21]. The severity of anxiety, depressive and trauma-related symptoms can lead to poorer quality of life, whereas perceived well-being could improve mental health outcomes [20] particularly as patients transition from the diagnostic to the treatment phase [22, 23]. Relatedly, it is important to further understand the risk and protective factors that lead to clinical or subclinical diagnosis-specific distress. However, studies so far have rarely focused on the specific and important turning point of the diagnosis.

Although empirical studies and systematic reviews point to a variety of mental health problems and different levels of well-being following a diagnosis of breast cancer, a quantitative account of the nature and frequency of such mental health problems and levels of well-being is lacking. The previously cited studies do not distinguish the pre-treatment and treatment phases, mixing newly diagnosed women with those undergoing treatment. It is important to isolate the moment following the diagnosis and ending just before treatments in order to better understand its impact on breast cancer patient’s mental health and thus adapt the interventions according to the phases of the cancer experience. This meta-analysis fills that gap by determining the nature and prevalence of clinically significant diagnosis-specific distress symptoms before the treatment begins in breast cancer patients.

Methods

The meta-analysis protocol is registered on PROSPERO (CRD42020203990).

Data collection

Ten databases were searched for published and unpublished studies: CINAHL, Clinicaltrial.org, Cochrane Central, Google Scholar, PTSDpubs, ProQuest, PsycInfo, PubMed, Research Gate and Web of Science. Keywords were chosen with the help of an academic librarian. We searched for the keywords [distress* or stress* or anxiety or anxious* or depress* or PTSD or peritraumatic distress/reactions or well-being] AND [breast cancer or breast neoplasm* or mammary neoplasm*] AND [diagnosis* or announcement or pre-treatment*] in the databases with no year constraint.

Literature screening

Inclusion criteria were developed based on literature search and discussions with the two senior authors. In order to be included, a study needed to: (1) measure depressive, anxiety, non-specific distress, peritraumatic distress or post-traumatic stress symptoms or well-being; (2) measure the symptoms after the diagnosis but before the beginning of treatments; (3) report data specifically for breast cancer patients; (4) use of a valid self-report measure or a structured interview; (5) present the results as the proportion of study participants scoring above the clinical threshold or cut-off scores; and (6) be written in English or French.

Data extraction

Articles whose title and abstract matched the inclusion criteria were exported to the Rayyan website [24] and divided according to symptom dimensions. Studies measuring several dimensions were included in multiple categories. Two team members independently selected the articles for inclusion based on the abstract. The consensus was achieved in 93% of cases. Disagreements were resolved by a senior author weighing in. Information was extracted from the article using the PRISMA checklist [25], and the exclusion criteria were applied. If a study did not clearly report that the data were collected before the cancer treatment started (n = 5), the corresponding author was contacted and given 30 days to respond. Two papers were thus excluded for lack of such a response.

Quality assessment

To assess the risk of bias, we used the tool from Hoy et al. [26] which evaluates the external and internal validity of studies but also the analysis bias. For each of the ten items, the evaluator chooses between a high, moderate or low bias rating [26]. Each study was evaluated by two independent raters. Disagreements (15%) were resolved by a senior author. Only papers with a low/moderate risk of bias were included in the final sample (see Fig. 1).

Fig. 1. PRISMA flow diagram [63].

The figure shows the flow of information through the different phases our meta-analysis following PRISMA methodology [63].

Data analysis

A meta-analysis was performed using the RStudio statistical software. ‘Prevalence’ estimate of each symptom dimension was calculated by pooling the study-specific estimates using random-effect meta-analysis that accounted for heterogeneity among studies. The heterogeneity of the preliminary studies was evaluated and quantified with the I² Cochrane test. The inverse variance method with the logit transformation was performed to stabilise the variance of each study’s proportion in order to get a better estimate of the summary proportion and its confidence interval (CI). Forest plots for the random-effect models were constructed. Linear regression tests of the funnel plot asymmetry as bias tests were conducted. We performed a sensitivity analysis to confirm the robustness of the results by conducting each meta-analysis multiple times, each time removing a different result (effect size) to see if the overall results remained similar. This was done to make sure that no effect size had too much influence on the resulting combined effect size. Finally, we conducted an exploratory meta-regression analysis using a mixed-effect model to examine if the time since diagnosis had an effect on the overall psychological symptom severity for all studies combined.

Results

Our final sample included 39 studies (see Fig. 1). The descriptive information is summarised in Table 1. As 19 studies reported more than one symptom dimension, they appear in more than one category. The definition of diagnosis varied between studies, such that distress symptoms were measured following the initial diagnosis of breast cancer (n = 15), following the pathological/clinical diagnosis (confirmed after surgery) (n = 12), or at an unspecified point (n = 12; see column “Diagnosis Definition” in Table 1). Linden et al. [27] was the only study reporting the prevalence of anxiety and depressive symptoms for women and men patients in breast cancer. However, we kept the study in our analyses since gender did not obscure the findings. For each symptom domain, the asymmetry linear regression tests and sensitivity analyses were nonsignificant, confirming that the analysis assumptions were satisfied.

Table 1.

Studies characteristics.

Studies (n = 39)	Country	Target population sex (n total; mean age)	Symptom categories	Diagnosis definitiona	Instrument (cut-off scores or clinical threshold (CT))	Time (mean days) of testing (post-diagnosis; pre-treatment)	Design	Risk of bias
Andreu et al. [64]	Spain	Women (102; 50.5)	Distress	Pathological/clinical	BSI-18 (>63)	Not specified	Longitudinal	Moderate
Arnaboldi et al. [65]	Italy	Women (150; 47.0)	Anxiety PTSD	Announcement	STAI (CT) IES (CT)	30; not specified	Longitudinal	Low
Burgess et al. [66]	UK	Women (222; 48.4)	Anxiety Depression	Pathological	SCID-anxiety (CT) CES-D (CT)	Not specified	Cross-sectional	Low
Chang et al. [67]	USA	Women (110; 60.3)	Depression	Pathological/ clinical	BSI‐18 (> 60)	Not specified	Cross-sectional	Low
Cimprich [68]	USA	Women (74; 56.0)	Distress	Announcement	SDS (CT)	Not specified; 11	Cross-sectional	Low
Cimprich & Ronis [69]	USA	Women (47; 64.0)	Distress	Pathological/ clinical	SDS (CT)	16; 12	Cross-sectional	Low
Cimprich et al. [70]	USA	Women (184; 54.6)	Distress	Announcement	POMS (CT)	Not specified; 18	Cross-sectional	Low
Civilotti et al. [58]	Italy	Women (436; 57.2)	Anxiety Depression Distress	Announcement	HADS-A (>8) HADS-D (>8) DT (CT)	28; 14	Cross-sectional	Low
Denieffe et al. [71]	Ireland	Women (94; 54.9)	Depression	Announcement	HADS-D (≥8)	14; 10	Prospective	Low
Farragher [72]	Ireland	Women (33; 51)	Anxiety Depression	Pathological/ clinical	LSAA (>7) LSAD (>7)	Not specified	Prospective	Low
Faye‐Schjøll & Schou-Bredal [73]	Norway	Women (367; 54.9)	Anxiety Depression	Pathological/ clinical	HADS-A (≥8) HADS-D (≥8)	Not specified	Longitudinal	Low
Galloway et al. [74]	USA	Women (60; 57.7b)	Anxiety Depression	Not specified	STAI (CT) CES-D (CT)	Not specified	Cross-sectional	Low
Gibbons et al. [75]	Ireland	Women (94; 57.2)	Anxiety Depression Distress	Announcement	HADS-A (CT) HADS-D (CT) Cancer-related distress (CT)	14; not specified	Cross-sectional	Low
Glinder & Compas [76]	USA	Women (76; 54.8)	Depression Distress	Pathological/ clinical	SCL-90R-Depression (CT) SCL-90-R (CT)	10; not specified	Prospective	Low
Hegel et al. [77]	USA	Women (236; 57.4)	Anxiety Depression Distress PTSD	Announcement	PHQ-Anxiety (≥8) PHQ-9- Depression (>9) SCL-90R-Distress (>5) PC-PTSD (≥8)	Not specified	Cross-sectional	Low
Hegel et al. [78]	USA	Women (345; 57.8)	Depression	Announcement	PHQ-9- Depression (≥7)	Not specified	Cross-sectional	Moderate
Henselmans et al. [79]	Netherlands	Women (171; 54.8)	Distress	Pathological/ clinical	GHQ-12 (≥2)	12; not specified	Longitudinal	Low
Jones et al. [80]	USA	Women (6949; 62.7)	Depression	Not specified	CES-D (≥1.61)	Not specified	Cross-sectional	Low
Kaiser et al. [81]	Germany	Women (27; 53.2)	Anxiety Depression	Not specified	HADS-A (CT) HADS-D (CT)	Not specified	Cross-sectional	Low
Kant et al. [82]	Germany	Women (181; 53.0)	Distress	Pathological/clinical	GHQ-12 (≥11)	Not specified	Prospective	Low
Kennedy et al. [83]	UK	Women (42; 60.2)	Anxiety Depression	Not specified	HADS-A (≥11) HADS-D (≥11)	44.7; not specified	Prospective	Low
Kyranou et al. [28]	USA	Women (390; 54.9)	Anxiety Depression	Announcement	STAI (Trait ≥31.8; State ≥32.2) CES-D (≥16)	Not specified	Cross-sectional	Low
Lally et al. [84]	USA	Women (100; 54.2)	Depression Distress PTSD	Not specified	CES-D (≥16) DT (≥4) IES (CT)	Not specified	Experimental	Low
Lee et al. [85]	Korea	Women (111; 44.2)	Anxiety Depression	Not specified	HADS-A (≥8) HADS-D (≥8)	Not specified	Prospective	Low
Linden et al. [27]	Canada	Women and men (2250; 58.9)	Anxiety Depression	Not specified	21-item Psychosocial Screen for Cancer assesses anxiety and depressive symptoms (≥11)	Not specified	Cross-sectional	Moderate
Mertz et al. [3]	Denmark	Women (41; 60.0)	Distress	Announcement	DT (≥7)	Not specified	Cross-sectional	Low
Miranda et al. [86]	Brazil	Women (20; 49.8)	Depression	Pathological/ clinical	BDI (CT)	Not specified	Cross-sectional	Moderate
Ng et al. [87]	Malaysia	Women (211; 55.0)	Distress	Pathological/ clinical	DT (CT)	Not specified	Cross-sectional	Low
Oh & Cho [88]	Korea	Women (50; 48.9)	Anxiety Depression	Not specified	HADS-A (≥11) HADS-D (≥11)	Not specified; 1–3	Longitudinal	Low
Oliveri et al. [89]	Italy	Women (150; 49.0)	PTSD	Pathological/ clinical	PTSD (CT)	Not specified	Prospective	Low
Park et al. [90]	USA	Women (54; 35.2)	Anxiety Depression	Not specified	HADS-A (≥11) HADS-D (≥11)	33; not specified	Cross-sectional	Low
Primo et al. [31]	USA	Women (85; 54.8)	PTSD	Not specified	IES (CT)	11.7; not specified	Prospective	Low
Ramalho et al. [91]	Portugal	Women (418; 56.7b)	Anxiety Depression	Announcement	HADS-A (≥11) HADS-D (≥11)	Not specified	Prospective	Low
Romeo et al. [92]	Italy	Women (147; 54.0)	Depression	Announcement	HADS-D (≥8)	Not specified	Longitudinal	Moderate
Stafford et al. [93]	Australia	Women (58.9)	Anxiety Depression	Announcement	HADS-A (≥16) CES-D (≥11)	Not specified	Cross-sectional	Low
Tjemsland et al. [32]	Norway	Women (106; 50.9b)	PTSD	Announcement	IES (≥19)	14; not specified	Cross-sectional	Low
Villar et al. [29]	Spain	Women (339; 58.9)	Anxiety	Not specified	STAI (CT)	Not specified	Prospective	Low
Voigt et al. [94]	Germany	Women (166; 50.4)	PTSD	Announcement	SCID (CT)	Not specified	Prospective	Low
Watson et al. [95]	UK	Women (359; 55.8)	Anxiety Depression	Not specified	HADS-A (≥10) HADS-D (≥8)	28–84; not specified	Cross-sectional	Moderate

USA   United States of America, UK   United Kingdom, PTSD   post-traumatic stress disorder, STAI   State-Trait Anxiety Inventory, SCID   shortened version of the structured clinical interview, HADS   Hospital Anxiety and Depression Scale-Anxiety/Depression subscale, LSAA/LSAD   Leeds general Scales for the self-Assessment of Anxiety/Depression, BSI-18   The Brief Symptom Inventory 18 item, CES-D   Center for Epidemiologic Studies Depression Scale, SCL-90-R Symptom Checklist-90—Revised, PHQ-9   Patient Health Questionnaire 9‐Item, BDI   Beck Depression Inventory, SDS   The Symptom Distress Scale, POMS   Profile of Mood States, DT   Distress Thermometer, GHQ-12   General Health Questionnaire-12 items, PC-PTSD   4-item Primary Care PTSD Screen, IES   Impact of Event Scale.

^aAnnouncement is the diagnosis provided before surgery and pathological/clinical is the definitive diagnosis provided after surgery. All before the medical treatments began.

^bMean age calculated by authors since the only percentage by age decade was available.

Anxiety symptoms

Nineteen studies were included in the anxiety category. The random-effect model showed that 34.0% (95% CI = [26.0, 43.1]; I² = 97.1%) of breast cancer patients present clinically significant anxiety symptoms after their diagnosis but before beginning medical treatments (see Fig. 2). Two studies [28, 29] using the State-Trait Anxiety Inventory provided the trait and state anxiety scores separately, whereas the majority of studies reported anxiety symptoms using the total score for this tool. To be consistent with the literature, we computed the mean number of clinical cases for the study and used this number in the meta-analysis.

Fig. 2. Forest plot for anxiety symptoms.

Instruments. STAI State-Trait Anxiety Inventory, SCID shortened version of the Structured Clinical Interview, HADS-A Hospital Anxiety and Depression Scale- Anxiety subscale, LSAA Leeds General Scales for the Self-Assessment of Anxiety. Other terms. Events: n total of participants with severe anxiety symptoms; Total: n total.

Depressive symptoms

We included 25 studies that reported the prevalence of depressive symptoms in our specified time frame. The random-effect model showed that 19.7% (95% CI = [15.9, 24.3]; I² = 95.2%) of pre-treatment breast cancer patients present clinically significant depressive symptoms (see Fig. 3).

Fig. 3. Forest plot for depressive symptoms.

Instruments. SCID shortened version of the Structured Clinical Interview, BSI-18 The Brief Symptom Inventory 18 item (depression subscale), HADS-D Hospital Anxiety and Depression Scale- Depression subscale, LSAD Leeds general scales for the Self-Assessment of Depression, CES-D Center for Epidemiologic Studies Depression Scale, SCL-90-R Symptom Checklist-90—Revised (depression subscale), PHQ-9 Patient Health Questionnaire 9‐Item Depression Module, BDI Beck Depression Inventory. Other terms. Events: n total of participants with severe depressive symptoms; Total: n total.

Non-specific distress symptoms

Thirteen studies were included in the non-specific distress category. The random-effect model showed that the prevalence of distress symptoms is 39.0% (95% CI = [28.1, 51.2]; I² = 95.7%) for breast cancer patients who just received their diagnosis (see Fig. 4).

Fig. 4. Forest plot for non-specific distress symptoms.

Instruments. BSI-18 The Brief Symptom Inventory 18 items (distress), SDS The Symptom Distress Scale, POMS Profile of Mood States, DT distress thermometer, SCL-90-R Symptom Checklist-90—Revised (distress), GHQ-12 General Health Questionnaire-12 items. Other terms. Events: n total of participants with severe non-specific distress symptoms; Total: n total.

Trauma-related symptoms

Seven studies were included in the PTSD symptom category. Three studies using the Impact of Event Scale to measure PTSD symptoms provided Intrusion and Avoidance scores separately [30–32], so we computed the mean number of clinical cases for the study and used that number in the meta-analysis. Results from the random-effect model showed that 31.4% (95% CI = [14.6, 55.0]; I² = 97.3%) of pre-treatment breast cancer patients present clinically significant PTSD symptoms (see Fig. 5). We did not find any study that measured peritraumatic distress symptoms with respect to our inclusion criteria.

Fig. 5. Forest plot for PTSD symptoms.

Instruments. IES-I/A Impact of Event Scale Intrusion and Avoidance scores summed, PC-PTSD 4-item Primary Care PTSD Screen, IES Impact of Event Scale, SCID shortened version of the structured clinical interview. Other terms. Events: n total of participants with PTSD symptoms; Total: n total.

Well-being

Through our article search, seven papers met the inclusion criteria for well-being. However, following data extraction, most of them were either excluded or divided into the symptom categories “anxiety” or “depression,” based on the instrument used. Finally, no studies reporting breast cancer patients’ well-being (e.g., emotional, physical, social and functional well-being) in our specific time frame were found.

Meta-regression analysis

Thirteen studies (excluding the studies that did not disclose the time elapsed since diagnosis) were included in the latter analysis, the meta-regression. We did not find a significant association between time since breast cancer diagnosis and prevalence (P = 0.69).

Discussion

This meta-analysis indicates that high diagnosis-specific distress is manifested in a significant minority of breast cancer patients specifically during the window of time between the diagnosis and the onset of medical treatments and clearly delineates the nature of such symptoms.

Diagnosis-specific distress

The usual stress response to a cancer diagnosis is characterised by shock and denial [33]. Despite efforts to remain focused, patients may have difficulty processing information when faced with this unexpected event. The fact that we found a higher prevalence of non-specific distress (39%) after the diagnosis announcement could indicate that there is a phase during which patients are experiencing shock, distress, uncertainty, and existential and anticipatory anxiety in the face of this new information and its perceived threat. This finding is in line with a literature review that reported a 32.8% prevalence of distress symptoms in breast cancer patients, but some studies included patients who had already started treatment at the time of data collection [34]. Receiving a breast cancer diagnosis changes patients’ views of their future, as they suddenly must consider the possibility of losses, ‘unfinished business’, and death while managing the reactions of their loved ones [35]. These changes to life plans can be accompanied by high levels of distress for many patients [35]. A study that reported the trajectory of distress symptoms in women with breast cancer showed high levels of distress after diagnosis [36]. Psychological distress severity can either decrease over time during the treatment and remission phases [36] or increase continuously after diagnosis, negatively altering the immune system and treatment efficacy [37].

Following the initial shock, anxiety symptoms may develop [33], which can explain the high prevalence of both distress (39%) and anxiety (34%) symptoms in this study. Anxiety is expected in response to receiving a life-threatening diagnosis of cancer. Within 3 weeks following the diagnosis, it becomes easier to discriminate between patients who are having a transient, as opposed to enduring, anxious reaction to the diagnosis [38, 39]. Indeed, research suggests that some breast cancer patients see a decrease of anxiety symptoms as treatment unfolds, but for several others, the symptomatology endures or worsen [40]. Patients with severe anxiety experience many physical symptoms (e.g., vomiting, nausea) that compound cancer-related physical symptoms [39].

In the category of PTSD symptoms, a prevalence of 31% within our specific time frame was found. A recent meta-analysis reported a 9.6% prevalence of PTSD symptoms in breast cancer patients; in that study, participants were mixed, some undergoing treatment and others not [41]. The literature on the prevalence of PTSD symptoms in different cancer phases is rich but lacks consistency since such symptoms are often confused with adjustment reactions [42, 43]. One possible reason for this confusion is the fact that a cancer diagnosis is not always construed as a qualifying event for receiving a PTSD diagnosis according to DSM-5 criteria [43]. Further, the short amount of time between receiving the diagnosis and starting treatment may contribute to the difficulty of diagnosing PTSD during this initial phase, leading clinicians to overlook this diagnosis in favour of other entities, such as a more generic anxiety disorder or an adjustment disorder [43]. There is also a conceptual difficulty in understanding PTSD symptoms during the diagnostic phase in the illness—patients have received their cancer diagnosis, but technically are still undergoing the “trauma” [42]. The Peritraumatic Distress Inventory could be the more appropriate measure to administer while patients are still in the midst of their stressors. Unfortunately, we did not find any study examining the prevalence of peritraumatic distress that met our inclusion criteria. The existing literature suggests that peritraumatic distress is present before undergoing surgery, but this kind of study is clearly missing in oncology [44–46].

According to the Learned Helplessness model, depression often follows anxiety reactions [47]. This model could explain why in our specified time frame, the prevalence of depressive symptoms (19%) was the lowest compared to every other diagnostic-specific distress symptom, especially anxiety. Supporting this theory, trajectory studies show that depression symptoms increase over time and that their prevalence can reach 66.1% during the remission phase [11, 48]. This suggests that depressive symptoms take more time to develop following a breast cancer diagnosis than do distress and anxiety symptoms. Another meta-analysis found a prevalence (32%) seemingly at odds with ours, but that study included patients in the treatment phase [49]. Still, the fact that nearly one in five patients experience clinically severe depressive symptoms makes it a common mental condition among breast cancer patients [12, 13].

We did not find any study falling in the well-being category. First, by using the search term “well-being,” most studies found were actually relying on a broad definition of well-being and measured (the absence of) anxiety or depression. Some studies that examined more specific dimensions of well-being (e.g., optimism, life satisfaction, personal growth) could have been left undetected by the “well-being” search term. Second, because one of the inclusion criteria was to “present the results as the proportion of study participants scoring above the clinical threshold or cut-off scores”, scarce well-being studies could have been included since the vast majority of well-being scales do not have a clinical threshold. Moreover, well-being is assessed more often in the post-treatment/survivorship phase than in the time frame between diagnosis and treatments [50]. However, two promising studies were found using The Emotional Well-Being (EWB) Subscale of the Functional Assessment of Cancer Therapy—General (FACT-G [51]) to measure patient well-being after receiving a cancer diagnosis but prior to treatments [52]. Sohl et al. [53] showed that well-being was lower prior to radiotherapy than after treatments in breast cancer patients. Positive aspects of well-being during the post-diagnosis phase should be investigated in future research.

Trajectory studies suggest that different symptoms emerge at different time points among cancer survivors according to the phase of the illness they are undergoing: screening, diagnosis, treatments, and remission. Even in a short time frame between diagnosis and treatments, our results support that the prevalence of diagnosis-specific distress symptoms varies according to the distress dimension. After diagnosis, distress symptoms are the most prevalent, closely followed by severe anxiety symptoms; severe depressive symptoms are relatively less common after the diagnosis. Further research is needed to understand the prevalence of PTSD symptoms in the post-diagnosis phase, as well as whether such symptoms are better characterised as peritraumatic distress [54].

Although our results target the psychological consequences associated with receiving the diagnosis of breast cancer per se, it is important to note that some patients experience stress prior to receiving their diagnosis [21]. The screening phase and the anxious anticipation related to possibly receiving a cancer diagnosis are not to be neglected and could also have an impact on the prevalence that we observed. It is also important to emphasise that our results show that the majority of breast cancer patients do not display clinically significant symptoms. The coping strategies and resilience of this group of patients should be recognised and further studied, as this could help develop helpful clinical interventions.

Clinical implications

Currently, in North America, screening for distress symptoms in cancer patients at all stages of the disease is part of the standard of care routine [4]. The high prevalence we found for diagnosis-specific distress symptoms suggests that clinicians should pay more attention to these symptoms as soon as patients receive their breast cancer diagnosis, even if their cancer treatments have not begun. Early screening and treatments for distress could facilitate adherence to breast cancer treatment protocols and increase the quality of life and the chances of survival [4]. This is aligned with our results and with the notion that psychosocial interventions are needed to prevent symptom worsening. A recent meta-analysis showed that cognitive-behavioural stress management interventions (e.g., relaxation training, stress reappraisal, training in interpersonal skills) can reduce stress, serum cortisol, anxiety, depression, avoidance, intrusive thoughts, and negative mood in breast cancer patients undergoing treatments [55]. Such care could be offered to breast cancer patients as an early intervention once the diagnosis is confirmed to reduce distress and prevent escalating symptoms.

Limitations

High heterogeneity

Our results have limitations related to the high heterogeneity found in each symptom category. The heterogeneity varied from 95 to 97%, and it may be accounted for by clinical and methodological factors [56]. Clinical heterogeneity occurs when there are significant differences between participants’ characteristics [56].

In that respect, sociodemographic variables, except for age, were heterogeneous from one study to another. Although we recognise that it is difficult to have a homogenous sample in oncology studies [57], it is important to address these differences to better interpret our results. Some included studies were conducted only on women who received a primary breast cancer diagnosis (e.g., refs. [3, 58]), and others on women diagnosed with either a primary or recurrent breast cancer (e.g., ref. [36]). However, according to Cohen [59], given their poorer prognosis, women with recurrent cancer present more distress symptoms than women with primary cancer, which could lead to greater variability in psychological distress outcomes following diagnosis. In our results, we did not find clear evidence that studies with mixed primary/recurrent samples had higher distress symptom prevalence. In addition, another influential variable was the different stages of breast cancer at the time of diagnosis in our sample: if the cancer is diagnosed at a more advanced stage, the prognosis will be proportionally poorer and the distress may be more severe [60]. In our meta-analysis, not all included studies reported these elements and those that did pool the prevalence for every type of breast cancer diagnosis together.

Second, methodological heterogeneity occurs when the design and protocol differ across studies [57]. The time since receiving a cancer diagnosis in our sample varied from 10 days to a few months. This could impact the severity of symptoms measured as, generally speaking, diagnosis-specific distress symptoms could fade or worsen with elapsed time since the diagnosis was announced [36]. However, we did not find an interaction between time since breast cancer diagnosis and any diagnostic-specific distress symptoms.

The presence of clinical and methodological heterogeneity in our sample shows that there is considerable variability between studies that measure the occurrence of psychological distress-related symptoms during the post-diagnosis and pre-treatment period. We argue that it is crucial to increase standardisation between studies, in conjunction with providing more information about the participants and methods.

Pre-diagnosis background

Another limitation was the studies’ specific methodological limitations that cannot be accounted for by the high heterogeneity. For example, it has been previously shown that not having a family member diagnosed with breast cancer can result in poorer psychological adjustment than having a family history of cancer [35]. The latter case brings about apprehension and mental preparation for that patient, unlike patients who receive a positive diagnosis without a family history. Having a prior mental health condition is also a risk factor for poorer adjustment to cancer [60]. We suggest that studies report these participants’ characteristics in their studies, to better document the symptoms they are measuring, which was not always the case in the studies we reviewed.

Methodology

We limited our data collection to papers written in French or in English. However, ten studies were not included on the basis that they were written in another language. There is a possibility that the inclusion of these studies could have changed our results.

Recommendation for future studies

The methodology is an important parameter impacting the capacity to replicate and compare results across studies. In our sample of studies, the lack of information regarding the time of measurement is a common omission. Future studies should report the time at which symptoms were measured along with the stage of cancer diagnosis and scheduled start of treatment. The methodology implemented by Civilotti et al. [58] serves as a good example.

Many studies did not differentiate between those newly diagnosed breast cancer patients who had and had not started medical treatments. Future studies should more clearly differentiate between those two groups since their symptom profile can be very different.

Psychological distress in cancer research appears to be measured using instruments that are derived from the DSM-5 diagnostic entities. However, these diagnostic entities present a certain degree of overlap. For example, items pertaining to insomnia and cognitive impairments (e.g., difficulty concentrating) can be found in instruments measuring anxiety, depression, distress, and PTSD [61]. Applying the framework suggested by the National Institute of Mental Health’s Research Domain Criteria [RDoC], which proposes to study domains of functioning cutting across diagnostic entities, could help identify the domains that are the most affected in breast cancer patients [61].

Finally, receiving a cancer diagnosis—a life-threatening illness—can be traumatic [17]. Oddly, we did not find any study that measured the presence of trauma-related symptoms in our specific time frame. Many patients develop PTSD symptoms upon learning that they have a cancer diagnosis, but there is much controversy about the proper way to measure these symptoms in the context of cancer [17]. One possible reason for the lack of clarity around cancer-related PTSD could be that previous studies attempted to measure PTSD symptoms too soon after the cancer diagnosis. An alternative would be to measure peritraumatic distress reactions first and PTSD later, as is typically the case in trauma research. Peritraumatic reactions index the subjective response to an event and can help determine whether the event was traumatic or not [16]. Peritraumatic distress symptoms is a robust predictor of who will develop PTSD and could be used for prognosis purposes in cancer research [62].

Receiving a cancer diagnosis can represent a turning point of high magnitude and can elicit psychological symptoms such as anxiety, depression, post-traumatic stress and non-specific distress. This meta-analysis aimed to shed light on the nature and occurrence of psychological symptoms in newly diagnosed breast cancer patients who have not yet begun treatment. Our results revealed that such symptoms are clinically severe for approximately one-third of our sample, and develop soon after the diagnosis is announced. Given the implications of diagnosis-specific distress for treatment adherence and prognosis, it should be carefully assessed and addressed before breast cancer treatments begin.

Author contributions

JF coordinated the study, carried out all the search and wrote most of the paper. ML assisted JF in all stages of the project, she also reviewed the table of results and wrote some sections of the paper. GE performed the statistical analyses and created the figures and helped with the revisions of the paper. Drs. M-FM and AB provided mentorship at all stages of the study and as the paper was being written and reviewed. Dr. MJC also provided mentorship and precious feedback, especially in the field of breast cancer and mental health. All authors have reviewed this version of the article and agreed to its publication.

Funding information

The authors did not receive funding to conduct the meta-analysis.

Data availability

Data are available upon request to the corresponding author.

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

This meta-analysis study did not thus involve human participants, human data or human tissue. No ethics approval or contentment to participate were not required.

Consent to publish

This manuscript does not contain any individual person’s data in any form, thus, consent for publication was not required.

Supplementary information

The online version contains supplementary material available at 10.1038/s41416-021-01542-3.