Anticipating mental health needs after chemotherapy in early-stage breast cancer using patient-reported symptom screening

Abstract

Purpose

Many patients with breast cancer experience depression and anxiety for years after completing systemic chemotherapy, which negatively impact overall symptom burden, quality of life, and treatment outcomes. The objective of this study was to examine the utility of patient-reported outcome (PRO) measures to predict mental health needs in patients with breast cancer during post-chemotherapy follow-up care.

Methods

In a sample of women with non-metastatic breast cancer, associations between patient-reported depression and anxiety at end of chemotherapy and post-chemotherapy mental health needs were evaluated using log-binomial regression adjusted for functional status, social activity limitations, and time from chemotherapy.

Results

In a sample of 149 women, 40% reported at least mild depressive symptoms and 52% reported at least mild anxiety at the end of chemotherapy. Over an average 3.2 years post-chemotherapy (range: 0.7–5.6 years), 23% received new psychiatric diagnoses, 21% engaged in mental health specialty care, and 62% were prescribed psychotropic medications. End of chemotherapy depression and anxiety were associated with future prescription of psychotropic medications (RR 1.52; 95% CI 1.14–2.03), as well as greater number of psychotropics. Associations were strongest with serotonin-norepinephrine reuptake inhibitors [(depression: RR 4.75; 95% CI 2.06–10.95); (anxiety: RR 3.68; 95% CI 1.62–8.36); (depression and anxiety: RR 2.98; 95% CI 1.65–5.36)].

Conclusion

Diagnosis of and treatment for depression and anxiety are common among women with breast cancer after completing chemotherapy. Prescriptions for psychotropic medications during the initial years after systemic chemotherapy can be anticipated by depression and anxiety screening at end of chemotherapy.

Introduction

At least 1 in 4 breast cancer patients experiences depression and anxiety [1–3]. While these symptoms may be most severe during active treatment [4–6], depression and anxiety may persist for years after the transition to post-treatment follow-up care [7, 8]. It is important to identify and treat persisting or new mental health symptoms, particularly during this transition, as they are associated with greater overall symptom burden [9, 10] and worse overall quality of life (QOL) [9, 11]. Additionally, depression is associated with decreased survival [12–15].

Existing guidelines emphasize the importance of mental health screening to identify psychological needs early, provide appropriate care, and ultimately minimize deleterious outcomes. Specifically, the Breast Cancer Survivorship Care guideline from the American Cancer Society (ACS) and American Society of Clinical Oncology (ASCO) recommends that clinicians screen patients for distress, depression, and/or anxiety and conduct a more probing assessment for patients perceived to be at greater risk for these symptoms [16]. ASCO has also published a guideline focused on depression and anxiety screening, assessment, and management, recommending that all cancer patients be evaluated for these mental health symptoms across the trajectory of care [17]. Certain transition periods during cancer care (e.g., end of treatment, post-treatment or at transition to survivorship, disease progression) are emphasized as the most critical times to screen for mental health symptoms and to direct patients along care pathways based on the severity of their symptoms [18].

Despite these existing guidelines, depression and anxiety screening remains far from routine in clinical practice. One potential barrier to implementation of mental health screening may be the lack of a standard way to measure depression and anxiety in patients with cancer. For example, the ASCO guidelines outline 11 different measures as potential options to evaluate depression and anxiety, which range in length from 7 (Generalized Anxiety Disorder-7 Item Scale, GAD-7 [19]) to 30 items (Geriatric Depression Scale [20]). In recognition of the need for a brief and clinically meaningful way to measure mental health symptoms, we have previously examined the utility of single-item patient-reported outcome (PRO) measures to screen for depression and anxiety during chemotherapy for breast cancer [21]. We found that compared to a longer, validated instrument [22], these single-item PROs were highly sensitive for depression and anxiety and, additionally, were associated in cross-sectional analysis with multiple objective indicators of psychiatric need. In the current study, we aimed to extend our previous work by using single-item PROs of depression and anxiety to measure symptoms at the end of chemotherapy and examine their relationship with future mental health needs in breast cancer patients. Specifically, we hypothesized that patients reporting at least mild symptoms of depression and anxiety at the end of chemotherapy were at greater risk for new psychiatric diagnoses, needing psychotropic medications, and seeking care from a mental health professional during the first several years post-treatment.

Methods

Study design and participants

Participants were women recruited into one of three multisite intervention studies that encouraged moderate walking during (neo)adjuvant chemotherapy for early-stage breast cancer (NCT02167932, NCT02328313, and NCT03761706) between April 2014 and October 2018. Trial eligibility criteria were age 21 years or older, English-speaking, histologically confirmed stage I–III breast cancer, and scheduled for adjuvant or neoadjuvant chemotherapy. Additional criteria for the present analysis were (1) received cancer care at the University of North Carolina (UNC)-Chapel Hill; (2) completed depression and anxiety PROs within 1 month of the last dose of chemotherapy; and (3) monitored for mental health outcomes for at least 6 months and no more than 6 years after completion of chemotherapy. All data available through February 4, 2020, were included. All studies were approved by the UNC-Chapel Hill Institutional Review Board and Lineberger Comprehensive Cancer Center Protocol Review Committee. Written informed consent meeting university and federal guidelines was obtained from each participant.

Measures

Patient-reported depression and anxiety

A Patient-Reported Symptom Monitoring (PRSM) instrument [23], which inquired about depression and anxiety symptoms “in the past 7 days,” was administered within 1 month of receipt of the last dose of chemotherapy. Patient reports were not shared with their clinicians. The PRSM is very similar to the PRO-CTCAE [24, 25] which was not available to the general community when the first two intervention trials were started but became available and was used in the third trial. Participants rated their symptom severity as follows: 0 = none, 1 = mild, 2 = moderate, 3 = severe, or 4 = very severe. Individual PRSM and PRO-CTCAE items are shown in Supplemental Table 1.

Patient-reported physical function and social well-being

At the end of chemotherapy, participants also completed the patient-reported Karnofsky Performance Status (Patient-KPS) [26] and Medical Outcomes Study (MOS) Social Activity Limitations and Social Support Surveys [27]. The MOS Social Activities Limitations Survey is a four-item measure inquiring about interference with social activities due to physical or emotional health. The scale ranges from 0 to 100 with scores ≥ 50 indicating greater social activity limitations. The MOS Social Support Survey is a 12-item measure inquiring about the availability of support for certain tangible (e.g., if you are confined to a bed) and emotional needs (e.g., someone who will listen to you when you need to talk), resulting in Tangible and Emotional subscales. The subscale scores range from 0 to 100 with scores < 50 indicating lower support. These functional and psychosocial measures were included given their potential impact on relationships between end of chemotherapy depression and anxiety and future mental health outcomes.

Electronic medical records

Research staff abstracted information regarding breast cancer stage (American Joint Committee on Cancer, version 7), phenotype, and treatment from the electronic medical record (EMR). Engagement in mental health specialty care, medications for depression and anxiety, and new psychiatric diagnoses received during follow-up were also abstracted from the EMR. To best capture mental health needs during this post-chemotherapy period, we examined engagement with a mental health provider and psychiatric medication prescriptions exclusively after the last dose of chemotherapy. Because most of the psychiatric disorders evaluated in this study are life-long diagnoses and there is inadequate information in the EMR to reliably identify exacerbation of established diagnoses, we focused on new psychiatric diagnoses (those that were listed in the active problem list and past medical history section of any note in the EMR after, but not prior to, completion of chemotherapy).

Participants were deemed to be engaged in mental health care if they had at least one documented visit with a mental health professional (psychiatrist, psychologist, or other mental health therapist). Antidepressants and anxiolytics were classified as selective serotonin reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs), benzodiazepines, or other, and hereafter will be collectively referred to as “psychotropic medications.” A detailed list of queried medications is provided in Supplemental Table 2. To avoid including benzodiazepines used sparingly for alternative indications (e.g., insomnia), benzodiazepines were only included if prescribed on a scheduled basis or with the as needed indication of “anxiety.” Diagnoses were categorized as depressive disorders, anxiety disorders, adjustment disorder, or other (e.g., bipolar disorder, substance use disorders, post-traumatic stress disorder, psychotic disorders).

Statistical analysis

Descriptive statistics were used to characterize the participants. Log-binomial regression was used to evaluate the relationship between end of chemotherapy depression and anxiety (none vs. at least mild severity) and categorical mental health outcomes—mental health subspecialty care, psychotropic medications, new psychiatric diagnoses—during follow-up. Zero-inflated Poisson (ZIP) regression was used to examine associations between at least mild depression and anxiety symptoms and the rate of prescription of psychotropic medications, accounting for the time each individual was followed. Associations in the ZIP regression were only considered in the Poisson portion of the model, not the zero inflation. Associations significant at the p < 0.1 level in univariate analyses were adjusted for time from chemotherapy (to account for variability in the duration of participant follow-up) and for variables associated with the outcomes of interest in a previous study involving an overlapping cohort [21], namely patient-KPS and social activity limitations. Except where stated otherwise, all associations presented in this manuscript are from adjusted models. All statistical analyses were performed with SAS software, version 9.4 (SAS Institute, Cary, NC).

Results

Participant characteristics

Mean age was 55 (range: 24–83), 19% were Black, and 88% received education beyond high school (Table 1). Chemotherapy was given as neoadjuvant treatment to 39% and as adjuvant treatment following surgery in 61%. Thirty-six percent of patients received anthracycline-based regimens. On average, the time since chemotherapy completion was 3.2 years (median: 3.3 years; range: 0.7–5.6).

Table 1.

Participant characteristics (N = 149)

Variable	N (%)
Age, mean (SD) [range], years	55.2 (12.3) [24–83]
Race
White	111 (75)
Black or African American	29 (19)
Asian	5 (3)
Other	4 (3)
Education - beyond high school	131 (88)
Married	85 (57)
Living alone	31 (21)
Employed more than 32 h a week	60 (40)
Breast cancer stage
I	31 (21)
II	78 (52)
III	40 (27)
HER2 +	43 (29)
HR +	87 (62)
Chemotherapy regimen
Anthracycline-based	54 (36)
Not anthracycline-based	95 (64)
Chemotherapy timing
Neoadjuvant	47 (39)
Adjuvant	91 (61)
Chemotherapy duration
< 3 months	59 (40)
≥ 3 months	90 (60)
Time since chemotherapy, mean (SD) [range], years	3.2 (1.3) [0.7–5.6]
Surgery
Mastectomy	80 (54)
Lumpectomy	67 (46)
Radiation	109 (74)
Patient-KPS
< 80	20 (13)
≥ 80	129 (87)
MOS Social Activity Limitations
< 50 (fewer limitations)	67 (53)
≥ 50 (greater limitations)	59 (47)
MOS Social Support - Emotional
< 50 (low support)	4 (3)
≥ 50 (high support)	123 (97)
MOS Social Support - Tangible
< 50 (low support)	7 (6)
≥ 50 (high support)	119 (94)

Abbreviations: KPS, Karnofsky Performance Status; MOS, Medical Outcomes Survey

End of chemotherapy depression and anxiety

At the end of chemotherapy, 40% of participants reported mild, moderate, or severe depressive symptoms; 52% reported mild, moderate, or severe anxiety symptoms; and 33% reported both depression and anxiety symptoms (Fig. 1).

Fig. 1.

Patient-reported depression and anxiety severity at end of chemotherapy

Mental health needs post-chemotherapy

During the post-chemotherapy observation period, 23% of participants received new psychiatric diagnoses, most of which were depressive and anxiety disorders. Twenty-one percent were engaged in mental health specialty care and 62% were prescribed psychotropic medications at some point during the observation period. Benzodiazepines were the most prescribed medication (34%), followed by SSRIs (24%) and SNRIs (23%).

Relationship between end of chemotherapy depression and anxiety and mental health needs post-chemotherapy

Depression and anxiety symptoms of at least mild severity at end of chemotherapy treatment were associated with 50% greater risk of being prescribed any psychotropic medication during the post-chemotherapy period (RR 1.52, p = 0.004) (Table 2). End of treatment anxiety was associated with 3.68-fold risk of future prescription for SNRIs (p = 0.002). Participants reporting depressive symptoms at end of chemotherapy were almost 5 times as likely to receive SNRIs (RR 4.75, p = 0.0003) and almost 3 times as likely to receive other classes of psychotropic medications (RR 2.89, p = 0.03). End of chemotherapy depression and anxiety were not associated with future prescription of benzodiazepines. Participants who reported both depression and anxiety at end of chemotherapy were prescribed a greater number of psychotropic medications compared to those who reported only depression, only anxiety, or neither (p = 0.02, Fig. 2).

Table 2.

Association between patient-reported depression and anxiety at end of chemotherapy and post-chemotherapy mental health outcomes

		Unadjusted RR [95% CI]			Adjusted RR [95% CI]a
Variable	N (%)	Depression	Anxiety	Depression and Anxiety	Depression	Anxiety	Depression and Anxiety
Psychotropic medications
Any	92 (62)	1.25 [0.97–1.60]	1.29 [0.99–1.68]	1.37 [1.08–1.74] **	--	1.33 [0.98–1.82]	1.52 [1.14–2.03] **
SSRI	36 (24)	1.33 [0.75–2.34]	1.82 [0.99–3.36]	1.83 [1.05–3.19] *	--	1.53 [0.73–3.20]	1.65 [0.85–3.20]
SNRI	35 (23)	3.71 [1.92–7.15] ***	3.07 [1.49–6.31] **	3.45 [1.91–6.26] ***	4.75 [2.06–10.96]b **	3.68 [1.62–8.36] **	2.98 [1.65–5.36] **
Benzodiazepine	51 (34)	1.22 [0.78–1.90]	1.41 [0.89–2.24]	1.43 [0.92–2.22]	--	--	--
Other	23 (15)	2.31 [1.07–4.99] *	3.28 [1.28–8.37] *	2.23 [1.06–4.68] *	2.89 [1.08–7.69] *	2.87 [0.98–8.45]	2.39 [0.95–6.00]
Mental Health Specialty Care	31 (21)	1.39 [0.75–2.59]	1.91 [0.97–3.78]	1.68 [0.90–3.12]	--	1.27 [0.57–2.81]	--
New psychiatric diagnoses
Any	34 (23)	1.32 [0.73–2.37]	1.30 [0.71–2.38]	1.26 [0.69–2.31]
Depression	17 (11)	1.32 [0.54–3.22]	1.30 [0.52–3.23]	1.11 [0.44–2.83]	--	--	--
Anxiety	24 (16)	1.06 [0.50–2.23]	1.27 [0.60–2.68]	1.02 [0.47–2.22]
Adjustment disorder	7 (5)	1.98 [0.46–8.52]	1.21 [0.28–5.24]	1.53 [0.36–6.57]
Other	1	--	--	--

Abbreviations: SSRI, selective serotonin reuptake inhibitor; SNRI, serotonin-norepinephrine reuptake inhibitor

^aAll univariate associations significant at the p < 0.1 level were examined in multivariable analyses adjusting for functional status, social activity limitations, and time since chemotherapy

^bThe adjusted RR between Depression and SNRI could not be estimated controlling for all three variables (due to small event rate), so the RR shown has been adjusted for social activity limitations and time since chemotherapy only

^*p < 0.05

^**p < 0.01

^***p < 0.0001

Fig. 2.

Number of psychotropic medications prescribed after chemotherapy by patient-reported depression and anxiety symptom status at end of chemotherapy

The univariate association between anxiety symptoms and future care with a mental health specialist bordered on statistical significance (RR 1.91, p = 0.06), but was not significant after adjustment. Neither depression nor anxiety at end of treatment was associated with new psychiatric diagnoses.

Discussion

At the end of chemotherapy, approximately half of breast cancer patients reported at least mild depressive or anxiety symptoms. Over an average 3-year post-chemotherapy follow-up period, 23% received new psychiatric diagnoses, 21% were engaged in mental health care, and over 60% were prescribed antidepressants or anxiolytics, representing an even higher proportion of patients receiving medications and engaged with a mental health specialist compared to what we have previously described in an overlapping cohort during chemotherapy [21]. Those who reported depression and/or anxiety at end of chemotherapy were more likely to receive prescriptions for psychotropic medications, particularly SNRIs.

Compared to previous studies, we observed a high prevalence of psychiatric diagnoses, psychotropic medication prescriptions, and utilization of mental health specialty care. For example, in a claims database study of 1111 cancer survivors at least 5 years from diagnosis, 33% had mental health diagnoses [28]. Although only 23% of patients in our study received new diagnoses in the years following chemotherapy, our previous work with this cohort suggests that the overall prevalence of psychiatric diagnoses during these initial post-chemotherapy years was over 50% [21]. As it relates to mental health treatment, an analysis of close to 100,000 US adults from the National Health Interview Survey estimated that 7% of cancer patients receive mental health services [29]. Participants in our study were three times as likely to receive care with a mental health provider.

We have previously described that the higher utilization of mental health care at our institution may be related to the availability of a comprehensive cancer support program, which provides mental health care that is highly integrated into the patients’ oncology care, offered in the same location where patients are receiving cancer treatment, and available regardless of payor status [21]. The higher prevalence reported in our study may also reflect our study design which monitored these mental health needs cumulatively over several years after chemotherapy instead of limiting examination to a single timepoint. Finally, our participants were younger than the overall cancer population, female, and had breast cancer, all possible risk factors for greater mental health need [29–32].

The relationships we observed between end of treatment depression and anxiety, identified using single-item PROs, and prescription of specific classes of psychiatric medications suggest several areas for future research with potential clinical implications. For example, relative to other medication classes, we observed strongest associations between patient-reported depression and anxiety and future prescription of SNRIs. The strong association with SNRIs may reflect that SNRIs are generally recommended for more severe mood symptoms that are resistant to initial treatment with SSRIs. Additionally, due to antagonism of CYP2D6, an enzyme essential to the metabolism and availability of the biologically active form of Tamoxifen, several SSRIs, including paroxetine and fluoxetine, are often avoided in patients with breast cancer [33]. The strength of the association between our single-item PROs and SNRIs may also reflect their utility not only to treat depression and anxiety, but also to improve vasomotor symptoms and neuropathic pain, which are common symptoms in these patients [34]. To that end, while the current study was focused on mental health symptoms, future research examining how multiple symptoms experienced by patients with breast cancer change over time and what interventions are used to address them would be of great interest.

One concerning finding in our study was the high prevalence of benzodiazepine prescriptions. Benzodiazepines have an adverse side effect profile that includes abuse potential, risk for falls, cognitive impairment, and, particularly when combined with opioids, respiratory depression and death. Many benzodiazepine prescriptions may have been initiated during treatment to address chemotherapy-induced nausea and vomiting or as time-limited interventions for insomnia and/or anxiety precipitated by corticosteroids. However, the continued prescription after completion of chemotherapy is concerning [18], particularly given the lack of association with patient reports of end of treatment anxiety. Future research is needed to clarify the overall and comparative effectiveness of pharmacotherapies for depression and anxiety for patients with breast cancer. However, our data already reveal that there is a subset of patients who require relatively complex psychiatric medication management and, importantly, that patients with these more complex future needs can be identified at the critical transition following chemotherapy by screening for depression and anxiety using single-item PROs.

An important question raised, but not fully answered, by the findings from our study is whether patients who would benefit from mental health care are in fact receiving it. There may be system-level barriers that impede access to care (e.g., fragmented delivery system, capacity and sites of care, cost), but another barrier may be under-recognition of symptoms by providers and suboptimal communication between providers and patients [21, 35, 36]. Providers may not have time to inquire about their patients’ emotional needs [37] or may assume that patients would make their mental health needs known. However, research suggests that cancer patients seek help for only a subset of their concerns [38], emphasizing the importance of systematically screening for potentially unmet needs [39] and doing so using methods compatible with providers’ time constraints (i.e., very brief screening instruments). Even though a relatively high proportion of participants in our study were engaged with a mental health provider, it remains possible that more could have benefited from mental health services. Furthermore, among the patients who had at least one visit with a mental health provider, the intensity or quality of care may still not have been optimal. It should be noted, however, that direct care with a mental health provider may not be the best use of resources or the most effective way for patients, especially as they transition into survivorship, to receive comprehensive care for their combined mental and physical health needs. Instead, cancer survivors may benefit most from care directed by a primary care provider [38, 40, 41], who can leverage expertise of mental health specialists, tailored to the needs of their patients, through collaborative care models [42].

Strengths and limitations

Our study has several strengths. First, our focus on depression and anxiety, as opposed to more vaguely defined distress or overall emotional well-being, offers clear targets for intervention. Second, we assessed multiple mental health outcomes, including psychiatric diagnoses, psychotropic medication prescriptions, and mental health specialty care. Third, our study included a relatively high proportion of Black women (19%), comparable to the proportion of Black women in the state of North Carolina where our study was conducted. Finally, our examination of single-item PROs for mental health symptom reporting facilitates contextualization of our findings within the larger, burgeoning field of patient-reported symptom monitoring [24].

There are also limitations to acknowledge. The first is related to the reliance on the EMR for the outcomes of interest. For example, it is possible that prescribed medications listed in the EMR did not accurately reflect what patients were actually taking. Especially for patients who transitioned aspects of care to providers outside of our EMR system, there may have been other mental health information that was not fully up to date. This limitation of the quality of more detailed or nuanced data also prevented us from examining changes in the stability of existing psychiatric diagnoses. Of note, the limitations inherent in the EMR data most likely led to an underestimation of our study outcomes, further emphasizing the magnitude of mental health needs in these patients. It is also important to consider limitations related to generalizability. Participants were from one academic medical center, generally highly educated, and not only agreed to participate in supportive care clinical trials, but were among the more compliant participants, as this analysis required for them to have provided symptom reports at the end of treatment. The variability in the time of follow-up, especially given that patients may be more likely to receive mental health diagnoses and treatment over time, is also important to acknowledge. We have mitigated the effects of accumulating needs over time by adjusting for time since chemotherapy in our analyses.

Conclusion

Taken together, our findings support a growing literature that depression and anxiety symptoms are prevalent in patients with breast cancer and that use of psychotropic medications and mental health specialty care are common in the initial years after completion of primary treatment. Our findings that single-item PROs of depression and anxiety, assessed at the end of chemotherapy, were associated with these future mental health needs highlight an opportunity to identify patients early on who might benefit from timely referral to mental health resources [16]. The importance of early identification of depression and anxiety is emphasized not only by the fact that these symptoms in themselves are highly treatable [31], but because worse mental health symptoms are often associated with other unmet psychosocial and physical care needs [43]. As previously discussed, ASCO guidelines recommend that depression and anxiety be screened throughout the trajectory of care, beginning at diagnosis, at regular intervals during treatment, at end of treatment, and beyond [18]. Depression and anxiety have also been recommended as two of 12 core PROs to be included in all population-based studies of cancer survivors [44]. In order for these symptoms to be routinely evaluated, it is important to have screening instruments that are feasible for use in busy oncology clinics. The single-item PROs for depression and anxiety used in this study may not only be feasible for routine screening but may also enable early recognition of symptoms that are predictive of future clinical need.

Availability of data and material

The data that support the findings of this study are available from the corresponding author upon reasonable request.