Psychosocial impact of COVID-19 pandemic on patients with solid malignancies and bone metastases (PsyCO-B): a multicentre prospective observational study

Graphical abstract

Highlights

• •This study shows the psychosocial impact of COVID-19 on bone metastatic patients.
• •Clinical severity directly predicted depressive symptoms.
• •Clinical severity was not associated with anxiety or psychosocial variables.
• •Pandemic’s traumatic impact was linked to anxiety symptoms.
• •Pandemic stress, but not clinical severity, shaped coping and support needs.

Abstract

Background

COVID-19 pandemic dramatically changed our daily habits. Patients with cancer, particularly those with bone metastases (BM), were among the most concerned individuals, due to both their clinical condition and social distancing.

Objectives

This study aimed to explore psychosocial effects of the pandemic on patients with BM, investigating the role of clinical severity, psychological symptoms, coping strategies, and perceived traumatic impact.

Methods

Six questionnaires were administered to patients with BM from three Institutions: Hospital Anxiety-Depression Scale, WHO Quality of Life-BREF, Attitude Toward Seeking Professional Psychological Help, Brief Illness Perception Questionnaire, Brief COPE, and Impact Event Scale-Revised (IES-R). Clinical, pathological, and socio-demographic data were collected and analyzed using descriptive statistics. Clinical severity was estimated by applying a specific prognostic model. Statistical analyses included correlations, linear regressions and mediation analyses.

Results

Clinical severity was significantly associated with depressive symptoms whereas the perceived traumatic impact of the COVID-19 pandemic was strongly associated with anxiety, the use of specific coping strategies (social support, avoidance and religiosity) and greater openness to psychological support. Mediation analysis showed that the IES-R mediates the relationship between subjective perception of illness and anxiety.

Conclusions

During the pandemic, depressive symptoms in patients with BM were mainly related to their specific clinical condition, whereas those associated with anxiety were mostly due to the experience of pandemic-associated trauma. The psychological impact of the pandemic influenced coping strategies and the request for support, regardless of the disease severity. These findings highlight the importance of psycho-oncological support focused on the patient’s experience and clinical condition.

1. Introduction

While the debilitating physical symptoms of cancer have long been recognized as a major clinical issue, attention to the psychological and social dimensions of the disease has emerged only relatively recently [1]. Originally, the emphasis was predominantly on adverse emotional consequences, including anxiety and depression. However, recent studies have begun to explore positive emotional constructs as well, including post-traumatic growth and cognitive adaptations associated with the cancer experience [2].

Following a diagnosis of cancer, patients frequently experience an initial psychological shock reaction, often accompanied by feelings of disbelief and denial [3], [4]. Such a reaction is commonly driven by the immediate association of cancer with intense physical suffering, social stigma, and death [5]. The diagnosis is thus perceived as a profoundly distressing event, frequently described as a psychological “nightmare”.

A significant proportion of patients exhibit intricate emotional patterns, characterized by fluctuations between denial and partial acceptance of the diagnosis [6]. As time progresses and based on the psychosocial resources of each individual patient, a proportion of patients begin to face the reality of their condition more comprehensively. This coping phase may involve psychological resistance, manifesting as conflictual interactions with healthcare providers, family members, or existential beliefs, as long as patients attempt to assert control over an uncontrollable situation [7].

It is evident that, when the limits of such strategies are perceived by individuals, this can lead to profound feelings of helplessness and hopelessness [8], [9]. In some cases, these feelings can evolve into proper depressive states, since not all patients achieve a state of psychological acceptance of their malignant illness, as an integrated aspect of their life experience.

In March 2020, the first official lockdown was implemented as a reaction to the outbreak of the SARS CoV-2 pandemic. Adopted as emergency provisions, measures including the use of medical and non-medical facemasks, social distancing, and targeted quarantine protocols had a profound impact on multiple aspects of oncology care across different cancer types [10], [11], [12], [13]. The pandemic significantly affected the lives of patients with cancer, raising concerns about accessibility and continuity of care, while weakening informal support networks due to social distancing [14], [15]. In patients receiving oncological treatments, this might have worsened mental health and emotional functioning.

A particularly vulnerable category of patients with cancer was represented by those with bone metastases (BM) [16].

In this regard, it is estimated that up to 80% of patients with advanced prostate cancer, approximately 65–75% of women with metastatic breast cancer and about 40% of subjects diagnosed with advanced lung cancer (among the most common neoplasms) develop secondary skeletal lesions [17]. Maintaining an adequate quality of life (QoL) in these patients depends on the adoption of suitable therapeutic strategies, based on the integration of different specialist skills [18]. However, a non-negligible percentage of patients affected by skeletal metastases still develop serious complications, collectively called “Skeletal Related Events (SREs)” [19], which lead to a progressive loss of autonomy in carrying out daily activities, whose psychological, socio-economic and relational impact is well documented [20], [21].

In this study, a clinical severity score (CSS) was developed and we hypothesized that greater clinical severity would have been associated with higher levels of psychological distress, specifically anxiety and depression. The selection of prognostic variables was theory-driven (a priori). Specifically, anxiety and depressive symptoms were chosen as the main outcomes based on prior literature showing that clinical severity is associated with psychological distress in patients with bone metastases [2]. The psychological impact of COVID-19 was included as an additional predictor to investigate whether pandemic-related distress contributes independently to higher levels of anxiety and depression. Additionally, we assumed that the perceived psychological impact of the COVID-19 pandemic would contribute to worsening of these symptoms, beyond the effects of clinical severity. The secondary aim of the study was to explore how the clinical severity and psychological distress related to the pandemic were associated with the adoption of coping strategies and attitudes towards psychological support. Thus, this study aimed to investigate the relationships between the clinical severity of the disease, psychological symptoms, coping strategies, and Quality of Life (QoL); to assess the psychological impact of the pandemic on patients with BM; and to explore whether the perceived psychological impact of the pandemic mediates the relationship between subjective illness perception and psychological distress.

2. Material and methods

2.1. Study design and participants

PsyCO-B was a multicentre, prospective, observational study designed to assess the psychological impact of the pandemic on patients with solid malignancies and BM and explore the potential mediating role of perceived trauma, related to the pandemic, in the relationship between disease severity and mental well-being.

Following approval from local Ethics Committee (Project identification code: 6908), a total of 119 patients were enrolled from three different Italian hospitals (University Hospital “Policlinico” of Bari, Bari; University Hospital “SS. Antonio e Biagio e C. Arrigo”, Alessandria; IRCCS “Istituto Ortopedico Rizzoli”, Bologna) during COVID-19 pandemic between January 2022 and June 2023. This period corresponded to the late phase of the COVID-19 pandemic in Italy. During most of the recruitment period, the national state of emergency had ended (April 1, 2022), vaccination coverage was widespread, and infection rates were relatively stable compared to earlier pandemic waves. While most public restrictions had been lifted, precautionary measures remained in place in healthcare settings, including mandatory mask use and controlled access to hospital facilities.

Inclusion criteria included a histological diagnosis of solid malignancies, associated with radiologically/histologically confirmed BM, in patients aged 18 years or more, able to understand and complete the provided questionnaires.

All patients accessing the Outpatient Oncology Clinics of the abovementioned hospitals were screened by medical oncologists based on the previously-established inclusion criteria; participation in the study was offered to all patients who met such requirements.

The privacy rights of human subjects were observed and all participants provided written informed consent. For each patient, clinical, pathological, and socio-demographic information was collected in anonymized form at the time of enrollment and entered into a computerized database protected by a double access code/password.

2.2. Psychological and psychosocial measures

Participants completed a battery of validated self-report questionnaires assessing psychological symptoms, coping strategies, illness perception, and attitudes toward psychological support:

• i.Hospital Anxiety and Depression Scale (HADS) [22]: used to assess symptoms of anxiety (HADS-A) and depression (HADS-D) over the last week;
• ii.Impact of Event Scale – Revised (IES-R) [23]: applied to measure the subjective traumatic impact of the COVID-19 pandemic. The total score (IES-R TOT) reflects the severity of intrusion, avoidance, and hyperarousal symptoms related to the traumatic event. The participants were instructed to complete the IES-R questionnaire based on their personal experiences and emotional reactions to the pandemic, ensuring that the assessment focused on specific psychological distress related to the pandemic, as opposed to general stress or stress related to cancer;
• iii.World Health Organization Quality of Life – BREF (WHOQOL-BREF) [24]: used to assess overall QoL across physical, psychological, social, and environmental domains;
• iv.Brief COPE Inventory [25]: used to assess coping strategies. The following subscales were analyzed in the present study: Social support, Avoidant coping, Religion coping, Positive attitude, Problem-focused coping, Self-blame;
• v.Brief Illness Perception Questionnaire (BIPQ) [26]: used to assess cognitive and emotional representations of the illness, providing a global index of perceived illness threat;
• vi.Attitudes Toward Seeking Professional Psychological Help − Short Form (ATSPPH-SF) [27]: the 10-item measure is employed to evaluate the general attitudes of participants towards seeking psychological assistance. The scale comprises two subscales: Openness to Seeking Psychological Treatment (ATT.OPEN), defined as the willingness to engage with a mental health professional, and Recognition of Need for Help (ATT.NEED), indicated by the perceived need for psychological support. The total score (ATT.TOT) is indicative of overall attitudes towards psychological help, with higher scores denoting more positive attitudes.

2.3. Clinical severity score (CSS)

In order to quantify the severity of the disease in an objective and measurable way, a score was assigned to each patient, based on the prognostic model proposed by Katagiri et al. [28]. This model includes five clinical predictors of survival in patients with BM. All five factors were adapted to the available data in our database and applied as follows:

• i.Primary tumor: classified according to Katagiri's prognostic categories in those having “rapid”, “moderate” or “slow” growth. Three CSS points were assigned to the “rapidly growing” primary tumors (i.e. lung, colorectal, esophageal, gastric, bilio-pancreatic, head&neck cancer, urological malignancies, melanoma, hepatocellular carcinoma, cervical cancer, and cancers of unknown origin); two CSS points were assigned if the primary tumor was considered having “moderate growth” (e.g. lung cancer undergoing targeted treatment, hormone-receptor negative breast cancer, castration resistant prostate cancer, renal cell carcinoma, endometrial and ovarian cancer, sarcoma, and others); no points were assigned if the primary lesion was considered at “slow growth” (hormone-dependent breast and prostate cancer, thyroid cancer).
• ii.ECOG performance status (PS): used as an indicator of functional impairment. One CSS point was added when the ECOG PS was > 2.
• iii.Visceral and/or cerebral metastases: inferred from oncological records. Two CSS points were added in the presence of such secondary lesions.
• iv.Previous chemotherapy: one CSS point was assigned if the patient had already received chemotherapy.
• v.Multiple skeletal metastases: the presence of more than 2 skeletal metastases, identified by radiological imaging, resulted in the assignment of one CSS point.

Each factor contributed to the composite CSS, which was used as a continuous variable in the regression and mediation models. Higher scores indicated greater clinical severity and poorer prognosis.

2.4. Analysis of data

Statistical analysis was conducted using the JASP software [29]. First, the relationships between CSS, anxiety and depression (HADS-A and HADS-D), coping strategies (Brief-COPE subscales), the traumatic impact of the pandemic (IES-R TOT), and QoL (WHOQOL-BREF) were explored through the implementation of Spearman’s correlation analyses. This was done after verifying that several outcome measures did not follow a normal distribution, as assessed using the Shapiro-Wilk test and visual inspection through Q-Q plots. The use of Spearman’s correlation was chosen as it is more suitable for non-normally distributed data.

In order to further investigate the role of clinical severity, two ANCOVA analyses were performed, classifying the sample into two groups based on CSS: i) Low (CSS from 1 to 3); ii) High (CSS from 4 to 6). This was made with the aim of comparing levels of depression and anxiety between the groups, covariating for the traumatic impact of COVID-19 (IES-R TOT).

To examine whether the psychological impact of COVID-19 (IES-R TOT) mediated the relationship between perception of illness (BIPQ) and anxiety symptoms (HADS-A), a mediation analysis was conducted, with perception of illness (BIPQ) as the independent variable, anxiety symptoms (HADS-A) as the dependent variable, and the psychological impact of the pandemic (IES-R TOT) as the mediating variable.

Assumptions of linear regression models were evaluated prior to analysis. Linearity and homoscedasticity were assessed through visual inspection of residuals versus fitted values and residuals versus predictors, while normality of residuals was examined using Q–Q plots. Although the Clinical Severity Score (CSS) is a discrete variable with a limited number of possible values (1–6), residual diagnostics did not reveal systematic deviations from model assumptions or evidence of heteroscedasticity. Therefore, CSS was retained as a continuous predictor in the regression analyses, and no data transformations were applied.

Next, a series of linear regression analyses was made to determine whether clinical severity (Severity Score) and perceived COVID-19 trauma (IES-R TOT) predicted: i) Symptoms of anxiety (HADS-A); ii) Symptoms of depression (HADS-D); iii) QoL (WHOQOL-BREF total score); iv) Adoption of specific coping strategies (social support, avoidance, religious coping, problem orientation, positive attitude, and self-blame, from the Brief-COPE).

Additional regression analyses were conducted to evaluate whether the CSS and IES-R TOT predicted openness to psychological help (ATT.OPEN); perceived need for psychological support (ATT.NEED); overall attitude toward psychological support (ATT.TOT).

For each model, the standardized beta coefficients, R² values, and p-values were documented. All statistical tests were two-tailed, with a significance level of p < 0.05. No corrections for multiple comparisons were applied, as the analyses were based on theoretically-driven hypotheses.

Accordingly, no time-stratified or longitudinal analyses were performed.

3. Results

The total sample consisted of 119 participants with a mean age of 61.77 years (SD = 14.05; range = 20–86). Of these, 70 (58.8%) were women and 49 (41.2%) were men (see Table 1). Additional clinical data are available in Supplementary Table 1. The severity score ranged from 1 to 6, with most participants falling within the moderate to high range (scores 3–5).

Table 1.

Descriptive statistics for demographic and psychological variables. The table reports mean, standard deviation, and range for age and psychological measures (HADS, IES-R, BIPQ, WHOQOL-BREF, Brief-COPE subscales), and frequency distribution for gender.

Descriptive Statistics
Variables	Mean (DS)	Minimum	Maximum	N (%)
Age	61.77 (14.05)	20	86	—
Sex	—	—	—	F: 70 (58.8%)
				M: 49 (41.2%)
HADS A	13.202 (4.986)	0.000	28.000
HADS D	16.782 (4.375)	2.000	26.000
IES-R TOT	3.709 (2.375)	0.000	11.500
BIPQ_TOT	6.060 (1.846)	1.750	10.000
COP_SupSoc	9.034 (4.367)	0.000	18.000
COP_Avoid	6.790 (4.069)	0.000	21.000
COP_Pos	7.882 (2.401)	1.000	12.000
COP_Prob	7.983 (3.479)	0.000	12.000
COP_Rel	2.958 (2.226)	0.000	6.000
COP_Autoacc	1.933 (1.867)	0.000	6.000
WHO.TOT	81.849 (12.893)	17.000	107.000
ATT.OPEN	15.017 (3.798)	5.000	20.000
ATT.NEED	14.303 (3.366)	6.000	20.000
ATT.TOT	29.319 (6.037)	11.000	39.000

3.1. Correlations between COVID-19 psychological impact, anxiety, coping, and illness perception

The Spearman’s correlation analyses showed that a higher psychological impact of Covid-19 (IES-R TOT) was strongly and significantly associated with increased anxiety symptoms (HADS-A; rho = 0.421; p < 0.001), as well as with various coping strategies, including social support (Cop_SupSoc; rho = 0.281; p = 0.002), avoidance (Cop_Avoid; rho = 0.255; p = 0.005), and religiosity (Cop_Rel; rho = 0.216; p = 0.018); it was also associated with a more negative perception of the disease (BIPQ; rho = 0.411; p < 0.001). No significant correlations were found between the CSS and psychological or QoL measures (all p values > 0.05).

4. Effects of clinical severity and COVID-19 impact on depression and anxiety

From the ANCOVA analysis, performed to examine whether depression levels (HADS-D) differed based on clinical severity (CSS, treated as a continuous variable), covariating for the effect of the psychological impact of the COVID-19 pandemic (IES-R TOT), a significant main effect of clinical severity (F_1,115 = 8.252, p = 0.005) emerged, indicating that patients with higher CSS reported significantly higher levels of depression than those with low severity scores. This result was also confirmed by the post-hoc test, which showed a significant mean difference in HADS-D scores between the two groups (mean difference = 2.221, p = 0.009).

The main effect of the IES-R TOT, on the other hand, was not significant (p = 0.205), suggesting that the perceived impact of the pandemic alone did not explain a significant portion of the variability in depression levels. However, the interaction between clinical severity and psychological impact of COVID-19 showed a trend towards significance (F_1,115 = 3.245, p = 0.074), indicating that the link between the traumatic impact of the pandemic and depression may vary depending on clinical severity, although it did not reach robust statistical significance (Table 2).

Table 2.

ANCOVA results for depressive symptoms (HADS-D) according to clinical severity (CSS) and psychological impact of COVID-19 (IES-R TOT). The table reports Type III sum of squares, degrees of freedom (df), mean squares, F values, and significance levels (p). Post hoc comparisons for clinical severity (high vs. low) are also shown, with Bonferroni-adjusted p-values.

ANCOVA − HADS D
Cases					Sum of Squares		df		Mean Square				F		p
SEVERITY					148.726		1		148.726				8.252		0.005
IES-R TOT					29.338		1		29.338				1.628		0.205
SEVERITY * IES-R TOT					58.484		1		58.484				3.245		0.074
Residuals
					2.072.727		115		18.024
Note. Type III Sum of Squares
Post Hoc Tests
Standard
Post Hoc Comparisons − SEVERITY
				Mean Difference		SE		df		t		pbonf
high		low		2.221		0.835		115		2.660		0.009

Further ANCOVA analyses were conducted to examine whether anxiety levels (HADS-A) differed according to clinical severity (high vs low), controlling for the psychological impact of the pandemic (IES-R TOT) as a covariate. The results showed that the main effect of clinical severity was not significant (F_1,115 = 0.455, p = 0.501), indicating that patients with greater clinical severity did not report significantly different anxiety levels compared to those with lower severity. The post-hoc test also found no significant difference in anxiety scores between the two groups (mean difference = 1.455, p = 0.094). However, the main effect of the IES-R TOT was statistically significant (F_1,115 = 24.628, p < 0.001), suggesting that patients who perceived the pandemic as having a greater impact reported higher levels of anxiety (Table 3).

Table 3.

ANCOVA results for anxiety symptoms (HADS-A) according to clinical severity (CSS) and psychological impact of COVID-19 (IES-R TOT). The table reports Type III sum of squares, degrees of freedom (df), mean squares, F values, and significance levels (p). Post hoc comparisons for clinical severity (high vs. low) are also shown, with Bonferroni-adjusted p-values.

ANCOVA − HADS A
Cases		Sum of Squares		df		Mean Square		F		p
SEVERITY		8.736		1		8.736		0.455		0.501
IES-R TOT		472.870		1		472.870		24.628		<.001
SEVERITY * IES-R TOT		0.720		1		0.720		0.037		0.847
Residuals		2.208.023		115		19.200
Note. Type III Sum of Squares

Post Hoc Tests
Standard
Post Hoc Comparisons − severity
				Mean Difference		SE		df		t		pbonf
high		low		1.455		0.862		115		1.689		0.094

4.1. Mediation between illness perception, COVID-19 impact, and psychological symptoms

Given the significant correlation between subjective perception of illness and anxiety symptoms, we further explored whether the relationship between perception of illness (BIPQ) and anxiety symptoms (HADS-A) was mediated by the psychological impact of COVID-19, measured by the IES-R TOT. The findings indicated that the direct impact of BIPQ_TOT on HADS-A is substantial (β = 0.858, SE = 0.230, z = 3.734, p < 0.001, 95% CI [0.408, 1.309]). The indirect effect of BIPQ_TOT on HADS-A through IES-R_TOT was also significant (β = 0.405, SE = 0.219, z = 5.770, p < 0.001, 95% CI [0.834, 1.692]), indicating that the psychological impact caused by the pandemic mediates the association between perceived illness and anxiety symptoms. The total effect of BIPQ_TOT on HADS-A is also significant (β = 1.263, SE = 0.219, z = 5.770, p < 0.001, 95% CI [0.834, 1.692]) (Table 4).

Table 4.

Mediation analysis examining the indirect effect of illness perception (BIPQ_TOT) on anxiety symptoms (HADS-A) through the psychological impact of COVID-19 (IES-R TOT). The table reports direct, indirect, and total effects, as well as path coefficients, with estimates, standard errors, z-values, p-values, and 95% confidence intervals.

Mediation Analysis BIPQ X IES X HADS A
Parameter estimates
Direct effects
							95% Confidence Interval
			Estimate	Std. error	z-value	p	Lower	Upper
BIPQ_TOT	→	HADS A	0.858	0.230	3.734	<.001	0.408	1.309
Note. Estimator is ML.

Indirect effects
								95% Confidence Interval
					Estimate	Std. error	z-value	p	Lower	Upper
BIPQ_TOT	→	IES-R TOT	→	HADS A	0.405	0.127	3.194	0.001	0.156	0.653
Note. Estimator is ML.

Total effects
							95% Confidence Interval
			Estimate	Std. error	z-value	p	Lower	Upper
BIPQ_TOT	→	HADS A	1.263	0.219	5.770	<.001	0.834	1.692
Note. Estimator is ML.

Path coefficients
							95% Confidence Interval
			Estimate	Std. error	z-value	p	Lower	Upper
IES-R TOT	→	HADS A	0.711	0.180	3.949	<.001	0.358	1.064
BIPQ_TOT	→	HADS A	0.858	0.230	3.734	<.001	0.408	1.309
BIPQ_TOT	→	IES-R TOT	0.569	0.105	5.430	<.001	0.364	0.774
Note. Estimator is ML.

A further mediation analysis was conducted to examine whether post-traumatic distress (IES-R_TOT) mediated the relationship between illness perception (BIPQ_TOT) and depressive symptoms (HADS-D). The results revealed that the direct effect of BIPQ_TOT on HADS-D was not statistically significant (β = 0.110, SE = 0.242, z = 0.456, p = 0.648, 95% CI [-0.364, 0.585]). The indirect effect of BIPQ_TOT on HADS-D through IES-R_TOT was also not significant (β = 0.012, SE = 0.108, z = 0.107, p = 0.915, 95% CI [-0.200, 0.223]), indicating that IES-R_TOT did not mediate the association between illness perception and depressive symptoms. Similarly, the total effect of BIPQ_TOT on HADS-D (i.e., the sum of direct and indirect effects) did not reach statistical significance (β = 0.122, SE = 0.217, z = 0.574, p = 0.574, 95% CI [-0.303, 0.547]). Among the path coefficients, only the association between BIPQ_TOT and IES-R_TOT was statistically significant (β = 0.569, SE = 0.105, z = 5.430, p < 0.001, 95% CI [0.364, 0.774]), suggesting that higher illness perception was associated with increased post-traumatic distress. However, the path from IES-R_TOT to HADS-D was not significant (β = 0.020, SE = 0.190, z = 0.107, p = 0.915, 95% CI [-0.352, 0.392]) (Table 5).

Table 5.

Mediation analysis examining the indirect effect of illness perception (BIPQ_TOT) on depressive symptoms (HADS-D) through the psychological impact of COVID-19 (IES-R TOT). The table reports direct, indirect, and total effects, as well as path coefficients, with estimates, standard errors, z-values, p-values, and 95% confidence intervals. Analyses were conducted using maximum likelihood (ML) estimation.

Mediation Analysis BIPQ TOT X IES X HADS D
Parameter estimates
Direct effects
							95% Confidence Interval
			Estimate	Std. error	z-value	p	Lower	Upper
BIPQ_TOT	→	HADS D	0.110	0.242	0.456	0.648	−0.364	0.585
Note. Estimator is ML.

Indirect effects
										95% Confidence Interval
					Estimate	Std. error	z-value	p	Lower		Upper
BIPQ_TOT	→	IES-R TOT	→	HADS D	0.012	0.108	0.107	0.915	−0.200		0.223
Note. Estimator is ML.

Total effects
							95% Confidence Interval
			Estimate	Std. error	z-value	p	Lower	Upper
BIPQ_TOT	→	HADS D	0.122	0.217	0.563	0.574	−0.303	0.547
Note. Estimator is ML.

Path coefficients
							95% Confidence Interval
			Estimate	Std. error	z-value	p	Lower	Upper
IES-R TOT	→	HADS D	0.020	0.190	0.107	0.915	−0.352	0.392
BIPQ_TOT	→	HADS D	0.110	0.242	0.456	0.648	−0.364	0.585
BIPQ_TOT	→	IES-R TOT	0.569	0.105	5.430	<.001	0.364	0.774

4.2. Predictors of anxiety symptoms

Additionally, a linear regression model with significant results (R² = 0.231; Adjusted R² = 0.218; F_{2, 116} = 17.437; p < 0.001), which examined predictors of anxiety symptoms (HADS-A) as the dependent variable, showed that IES-R TOT was a significant positive predictor (B = 1.010; β = 0.477; p < 0.001) (Fig. 1), in contrast to CSS (p = 0.509) (Fig. 2). This indicates that anxiety symptoms were primarily associated with the subjective impact of the pandemic, rather than with clinical status.

Fig. 1.

Residual plot for IES-R total and HADS-A. The plot displays the relationship between the residuals of IES-R total and HADS-A; the solid line indicates the linear trend, while the dashed lines represent the confidence bounds, used to assess linearity and homoscedasticity of the model.

Fig. 2.

Residual plot for severity score and HADS-A. The plot shows the relationship between the residuals of the severity score and HADS-A; the solid line represents the linear trend and the dashed lines indicate the confidence bounds, used to evaluate linearity and homoscedasticity of the regression model.

4.3. Predictors of coping strategies

Linear regression models were conducted to examine whether CSS and IES-R TOT predicted the application of specific coping strategies. IES-R TOT significantly predicted a greater need for social support, as well as avoidance and religiosity strategy use. In particular, the regression model was significant for the purpose of social support (R² = 0.108, Adjusted R² = 0.092, F_2,116 = 6.988, p = 0.001), primarily due to the significant impact of IES-R (B = 0.561, β = 0.303, p < 0.001) rather than that of the CSS (Fig. 3).

Fig. 3.

Residual plots of the psychological impact of COVID-19 (IES-R TOT) and clinical severity (CSS) in relation to the Social Support coping strategy (Cop_SupSoc).

The regression model was also significant for the purpose of avoidance (R² = 0.098, Adjusted R² = 0.082, F_2,116 = 6.282, p = 0.003), again primarily due to the significant impact of IES-R (B = 0.539, β = 0.312, p < 0.001), rather than CSS one (Fig. 4).

Fig. 4.

Residual plots of the psychological impact of COVID-19 (IES-R TOT) and clinical severity (CSS) in relation to the Avoidance coping strategy (Cop_Avoid).

Moreover, the regression model was significant for the purpose of religiosity (R² = 0.071, Adjusted R² = 0.051, F_2,116 = 4.459, p = 0.014). This was primarily due to the significant impact of IES-R (B = 0.222, β = 0.235, p = 0.010), rather than that of the CSS (Fig. 5).

Fig. 5.

Residual plots of the psychological impact of COVID-19 (IES-R TOT) and clinical severity (CSS) in relation to the religiosity coping strategy (Cop_Rel).

Positive attitude, problem-focused coping, and self-blame strategies were not significantly predicted by either the Severity Score or IES-R TOT. Furthermore, the linear regression that examined the relationship between IES-R TOT, CSS, and the use of the Positive Attitude coping strategy showed an overall significant model (R² = 0.061; Adjusted R² = 0.045; F_2,116 = 3.758; p = 0.026). However, in this case, the Severity Score had a significant negative impact on the adoption of this strategy (β = –.223; p = 0.015), while the IES-R TOT was not a significant predictor (p = 0.253) (Fig. 6).

Fig. 6.

Residual plots of the psychological impact of COVID-19 (IES-R TOT) and clinical severity (CSS) in relation to the Positive Attitude coping strategy (Cop_PosAtt).

4.4. Attitude and openness towards psychological support

The results of further regressions, conducted to verify whether CSS and IES-R TOT could predict attitude and openness towards psychological help showed that only IES-R TOT predicted openness to psychological support, perceived need for psychological support, and general attitude towards psychological help. However, the association between these latter variables and the CSS was not significant.

As for the openness to psychological support (ATT.OPEN), the regression model was significant (R² = 0.126, Adjusted R² = 0.110, F_2,116 = 8.326, p < 0.001). This was primarily due to the significant impact of IES-R (B = 0.550, β = 0.341, p < 0.001), rather than that of the CSS (Table 6).

Table 6.

Results of the linear regression model predicting openness toward psychological support (ATT.OPEN) from the psychological impact of COVID-19 (IES-R TOT) and clinical severity (Severity Score). Model summary, ANOVA statistics, and unstandardized and standardized regression coefficients are reported.

Linear Regression ATT OPEN X IES E SEV SCORE
Model Summary − ATT.OPEN
Model		R		R2		Adjusted R2		RMSE
M0		0.000		0.000		0.000		3.798
M1		0.354		0.126		0.110		3.582
Note. M1 includes IES-R TOT, SEVERITY SCORE
ANOVA
Model			Sum of Squares		df		Mean Square		F	p
M1	Regression		213.657		2		106.829		8.326	<.001
	Residual		1.488.309		116		12.830
	Total		1.701.966		118
Note. M1 includes IES-R TOT, SEVERITY SCORE
Note. The intercept model is omitted, as no meaningful information can be shown.
Coefficients
Model			Unstandardized		Standard Error		Standardized		t	p
M0	(Intercept)		15.017		0.348				43.134	<.001
M1	(Intercept)		11.767		1.280				9.191	<.001
	IES-R TOT		0.550		0.140		0.341		3.930	<.001
	SEVERITY SCORE		0.303		0.282		0.093		1.075	0.285

For the perceived need for psychological support (ATT.NEED), the regression model was significant (R² = 0.062, Adjusted R² = 0.046, F_2,116 = 3.817, p = 0.025), due to the significant impact of IES-R (B = 0.333, β = 0.233, p = 0.011), rather than that of the CSS (Table 7).

Table 7.

Results of the linear regression model predicting perceived need for psychological support (ATT.NEED) from the psychological impact of COVID-19 (IES-R TOT) and clinical severity (Severity Score). Model summary statistics, ANOVA results, and unstandardized and standardized coefficients are reported.

Model Summary − ATT.NEED
Model	R	R2	Adjusted R2	RMSE
M0	0.000	0.000	0.000	3.366
M1	0.248	0.062	0.046	3.289
Note. M1 includes IES-R TOT, SEVERITY SCORE

ANOVA
Model		Sum of Squares	df	Mean Square	F	p
M1	Regression	82.568	2	41.284	3.817	0.025
	Residual	1.254.541	116	10.815
	Total	1.337.109	118
Note. M1 includes IES-R TOT, SEVERITY SCORE
Note. The intercept model is omitted, as no meaningful information can be shown.

Coefficients
Model		Unstandardized	Standard Error	Standardized	t	p
M0	(Intercept)	14.303	0.309		46.349	<.001
M1	(Intercept)	12.088	1.175		10.283	<.001
	IES-R TOT	0.333	0.128	0.233	2.589	0.011
	SEVERITY SCORE	0.246	0.259	0.085	0.949	0.345

For the general attitude towards psychological help (ATT.TOT), the regression model was also significant (R² = 0.130, Adjusted R² = 0.115, F_2,116 = 8.701, p < 0.001) as a consequence of the significant impact of IES-R (B = 0.882, β = 0.344, p < 0.001), rather than CSS (Table 8).

Table 8.

Results of the linear regression model predicting overall attitude toward psychological support (ATT.TOT) from the psychological impact of COVID-19 (IES-R TOT) and clinical severity (Severity Score). Model summary statistics, ANOVA results, and regression coefficients are presented.

Model Summary − ATT.TOT
Model	R	R2		Adjusted R2	RMSE
M0	0.000	0.000		0.000	6.037
M1	0.361	0.130		0.115	5.677
Note. M1 includes IES-R TOT, SEVERITY SCORE
ANOVA
Model		Sum of Squares	df	Mean Square	F	p
M1	Regression	560.928	2	280.464	8.701	<.001
	Residual	3.738.938	116	32.232
	Total	4.299.866	118
Note. M1 includes IES-R TOT, SEVERITY SCORE
Note. The intercept model is omitted, as no meaningful information can be shown.
Coefficients
Model		Unstandardized	Standard Error	Standardized	t	p
M0	(Intercept)	29.319	0.553		52.984	<.001
M1	(Intercept)	23.855	2.029		11.755	<.001
	IES-R TOT	0.882	0.222	0.344	3.979	<.001
	SEVERITY SCORE	0.549	0.447	0.106	1.228	0.222

4.5. Quality of life

The residual regression analyses, including those that examined overall QoL (WHOQOL-BREF TOT) in relation to the psychological impact of COVID-19 (IES-R TOT) and clinical severity (CSS), did not show any statistically significant results.

5. Discussion

The findings of this study suggest a relevant psychosocial burden associated with the COVID-19 pandemic in patients with BM, emphasizing the intricate interplay between clinical severity of the disease, psychological distress and coping strategies. The results describe significant associations between perceived pandemic-related distress and psychological outcomes. The data show that clinical severity is significantly associated with depressive symptoms, but not with anxiety symptoms or other psychological or psychosocial variables. This suggests that the perception of a serious clinical condition may contribute to emotional distress linked to feelings of resignation, loss of hope or existential reflection on prognosis, components frequently observed in depressive conditions in oncology [30], [31].

Depression and anxiety are the predominant symptoms of psychological distress in patients with advanced cancer, often associated with the fear of death, loss of autonomy in daily activities, and worsening symptoms as the disease progresses [32]. However, in the present study, anxiety was not associated with clinical severity, but it was significantly associated with the subjective psychological impact of the pandemic. This finding is consistent with previous studies reporting a highly variable prevalence of anxiety in patients with cancer, ranging from 29% to 52% [33], [34], [35].

The expression of psychological distress can be influenced by multiple factors, including the type of cancer, patient age, symptom control, family support, and physical, emotional, and cognitive functioning [32], [35], [36].

In a review by Paredes et al. [37], anxiety was described in 25% of patients with cancer, while depression was present in 11.1%. In a study of patients with limb sarcoma, Tang et al. [38] found that approximately 30% of subjects exhibited anxiety symptoms associated with pain, functional impairment and a worsened QoL.

These findings highlight that the emotional burden introduced by the pandemic context, characterized by uncertainty, fear of worsening clinical conditions and difficulties in accessing care [14], may have contributed to higher levels of anxiety. Mediation analysis provided further support for these mechanisms: the subjective psychological impact of the COVID-19 mediated the relationship between perception of illness and anxiety, suggesting that a more threatening perception of one's clinical condition makes patients more vulnerable to the activation of an anxious response in the presence of events perceived as stressful.

Conversely, clinical severity directly predicted depressive symptoms; however, the contribution of the psychological impact of the pandemic in this relationship appeared not significant. This was also indicated by the ANCOVA analysis, which showed that the effects of the pandemic did not explain a significant proportion of the variance in depression levels.

In accordance with the study by Lee et al. [39], the perception of the disease played a pivotal role in mediating the relationship between contextual factors and psychological distress. When the disease was perceived as serious, chronic or uncontrollable, anxiety and depression levels increased, regardless of the objective clinical severity. Consequently, anxiety would not be directly determined by clinical severity, but rather by the subjective perception of threat, probably amplified by external events such as the pandemic.

Focusing on coping mechanisms, the study findings indicate that clinical severity did not correlate directly with the adoption of specific coping strategies. This suggests that behavioural responses are more closely associated with subjective experiences perceived as stressful than with the objective medical condition. This aligns with the observations made in depressive disorders, where attitudes of withdrawal, passivity and reduced behavioural activation are prevalent and often associated with feelings of helplessness and a perceived lack of control [40], [41].

In contrast, the perceived traumatic impact of the pandemic was found to be significantly associated with anxiety symptoms and increased use of coping strategies, particularly social support, avoidance, and religiosity.

These results suggest that in conditions of elevated perceived stress, patients adopt an array of coping strategies, some oriented towards seeking help, others towards emotional management or distancing. As evidenced by the existing literature, patients with cancer use a diverse range of coping strategies, including both adaptive and maladaptive mechanisms. These strategies include denial, disengagement, acceptance, emotional expression, and self-blame [42], [43].

In addition, the psychological impact of the pandemic was interestingly associated with an increased openness of individuals to seek psychological support, regardless of the severity of the cancer [44], [45]. This finding suggests that individuals sought for psychological assistance not due to the cancer itself, but as a result of the emotional distress arising from the challenging pandemic experience.

Finally, the issue of providing psychological and social support during the pandemic must be considered. While it is widely recognized that physical distancing can diminish interpersonal connections [46], some studies have reported limited effects on social relationships and loneliness [47], [48]. The pandemic may have strengthened the perception of support for patients with cancer, either by activating existing networks or by enhancing available resources [49], [50]. Many patients with cancer had pre-existing support networks consisting of family members, healthcare professionals, and caregivers, which may have mitigated the psychological impact of isolation [51]. Furthermore, patients undergoing active treatment often take hygiene and health precautions, even outside of pandemic times, thus reducing the impact of distancing measures [52]. Once again, the quality and quantity of social ties are confirmed as a protective factor for psychological well-being during pandemic [53].

It is important to acknowledge that the cross-sectional design of the study, in combination with the absence of pre-pandemic data or a control group, limits the possibility of making causal inferences or conclusions regarding the extent of the psychosocial impact of the pandemic. In addition, data were collected during a relatively extended period of the late COVID-19 pandemic, during which public attitudes and perceived threat may have evolved over time. As the study was not powered to examine temporal effects, potential variations related to the timing of enrollment could not be formally tested. Instead, the focus is on the associations between perceived pandemic-related distress and psychological outcomes.

In conclusion, the results of this study suggest that, in the management of patients, it is crucial to assess not only the clinical severity of the disease itself, but also the role of external stressors, such as pandemic-related psychological distress, which have the potential to amplify the overall psychological impact. In daily clinical practice, it is essential to offer early psycho-oncological support, both through standard pathways and, if necessary, through ad hoc interventions during periods of collective major stressful events, such as pandemics, in order to support patients' psychological resources and encourage the adoption of effective coping strategies.

CRediT authorship contribution statement

Paolo Taurisano: Writing – original draft, Methodology, Formal analysis, Data curation. Ester Cornacchia: Methodology, Investigation. Maria Fara De Caro: Methodology, Data curation. Vittorio Fusco: Writing – review & editing, Investigation, Data curation. Toni Ibrahim: Writing – review & editing, Investigation. Anna Ragno: Data curation. Francesco Salonne: Investigation, Data curation. Giulia Paparella: Methodology, Data curation. Marta Betti: Investigation, Data curation. Serena Penpa: Investigation, Data curation. Maura Rossi: Investigation, Data curation. Rossella Hakim: Investigation, Data curation. Michela Pierini: Investigation, Data curation. Camillo Porta: Writing – review & editing, Supervision. Stella D’Oronzo: Writing – review & editing, Writing – original draft, Supervision, Methodology, Formal analysis, Data curation, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: