Psychological variables associated with quality of life in patients with head and neck cancer: the role of body image distress

Study design

This is a cross-sectional study: participants completed self-report questionnaires before undergoing a medical treatment for HNC (no other assessments occurred).

Participants

Patients with HNC who were about to receive a medical treatment approximately in a month (surgery, radiotherapy, chemotherapy, or concomitant radiotherapy and chemotherapy) were asked to take part in the study. Eligible patients were identified during multidisciplinary HNC visits at the Radiotherapy Unit of the Veneto Institute of Oncology (IOV) in Padua (Italy). Eligible criteria included patients who (1) were older than 18 years, (2) were diagnosed with a HNC, (3) were about to receive a medical treatment for HNC (all tumor sites, all treatment modalities), (4) were not receiving medical treatments for other cancer diseases at the time of the research, and (5) were competent to provide informed consent. No restriction was placed on the type of treatment participants received/tumor sites for inclusion in the study. The only exclusion criterion was the presence of a benign neoplasm localized on the head and neck anatomical district (e.g., parotid). Based on such criteria, 51 patients (37 males and 14 females) were considered eligible and were enrolled for the study (Table 1). In terms of response rate, only one patient considered eligible did not agree to take part in the study, a 50-year-old female with a tumor localized in the oral cavity.1

Recruitment procedures

Patients with HNC were recruited between February 2019 and February 2020. A licensed clinical psychologist with extensive experience managing pyscho-oncologic concerns in patients with HNC recruited patients during multidisciplinary HNC visits where patients received the cancer diagnosis and were offered a treatment. The clinical psychologist explained the study aims and patients were asked to complete a survey pertaining to the relationship between body image and physical and psychological well-being. No other assessments took place. Interested and eligible patients gave their written informed consent and completed self-report questionnaires before receiving treatment for HNC. Patients with HNC who accepted to participate in the study were offered two different modalities to complete the survey: in-person, at the Radiotherapy Unit of the Veneto Institute of Oncology (IOV), or at home. Two modalities of compilation were offered since most patients were willing to participate in the study but were distress after the HNC visit. When patients preferred the in-person modality, they completed self-report questionnaires in a quiet room at the IOV, specifically arranged for the self-report questionnaires compilation; the clinical psychologist was available to answer any inquiry during the filling process. The time for the compilation was approximately 40 min. When patients preferred the at-home modality, the clinical psychologist provided self-report questionnaires and a sealing envelop and instructed patients to complete the questionnaires at home and to take them back in the sealing envelop the day of the next HNC visit (approximately 2 weeks after). Patients did not receive any compensation for their participation. The study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of the Psychological Research of the University of Padua.

Measures

Socio-demographic information form: employed to assess socio-demographic information of participants such as sex (biological, assigned at birth; “What is your sex assigned at birth?”), gender (the internal/psychological sense of self, regardless of what sex a person was assigned at birth; “What is your current gender identity?”), age, education, and self-reported psychological disorders.

Personal medical history: information about participants’ medical history was collected through the electronic medical record of each patient, including cancer history (presence of recurrence and previous treatments for HNC), time of diagnosis, stage and localization of disease, and medications.

Body Image Scale (BIS; [27, 28]): self-report questionnaire aimed at investigating body image distress in patients with cancer. The BIS measures emotional, cognitive, and behavioral components of body image through 10 items on a 4-point Likert scale (from 1 = “never” to 4 = “very”). Higher values correspond to greater body image distress. The Italian version of the BIS showed good psychometric properties (Cronbach’s α = 0.92). In the current study, the BIS showed good internal consistency (Cronbach’s α = 0.93).

Short Form-12 Health Survey (SF-12; [29; 30]): a brief version of the SF-36 health survey. The SF-12 evaluates eight dimensions related to an individual’s life that can be influenced by the presence of a disease. Answers can be provided through dichotomous yes/no answers, or through items on a 3- or a 5-point Likert scale. In addition to the eight dimensions, the SF-12 produces two summary scores evaluating physical and mental health. Higher scores are associated with higher QoL. The Italian version of the SF-12 showed good content, construct, and criterion validity [31]. For the purposes of the study, we considered only the Physical Component Score (PCS) and the Mental Component Score (MCS).

Rosenberg Self-Esteem Scale (RSES; [32; 33]): measure made up of 10 items rated on a 4-point Likert scale (from 1 = “strongly disagree” to 4 = “strongly agree”) assessing global self-esteem, with higher scores indicating greater self-esteem. The Italian version of the RSES showed good psychometric properties: its internal consistency was α = 0.84 [33]. In the current study, the RSES showed adequate internal consistency (Cronbach’s α = 0.72).

Intolerance of Uncertainty Scale-12 (IUS-12; [34; 35]): 12-item self-report measure evaluating the tendency to find uncertainty upsetting and distressing. Individuals are asked to rate the extent to which each statement applies to themselves on a 5-point Likert scale (from 1 = “not at all like me” to 5 = “entirely like me”), with higher scores indicating greater IU. The IUS-12 demonstrated excellent internal consistency (Cronbach’s α = 0.80), convergent, and discriminant validity in its Italian version [35]. In the current sample, internal consistency was good (α = 0.87).

Social Interaction Anxiety Scale (SIAS; [36; 37]): 19-item self-report questionnaire designed to assess social interaction anxiety on a 5-point Likert scale (from 0 = “Not at all” to 4 = “Extremely”), with higher scores indicating greater social anxiety symptoms. The Italian version of the SIAS proved to be highly reliable (Cronbach’s α = 0.86) [37]. In our samples, the SIAS showed good internal consistency (Cronbach’s α = 0.84).

Coping Response Inventory Adult Form (CRI-Adult; [38; 39]): self-report questionnaire made up of 48 items assessing approach (logical analysis, positive reappraisal, guidance/support seeking, and problem solving) and avoidance coping (cognitive avoidance, resigned acceptance, alternative rewards, and emotional discharge). Each subscale is composed of six items rated on a 4-point Likert-type scale (0 = “Not at all”; 3 = “Fairly often”). Participants had to think specifically about how they cope with the diagnosis of HNC when replying to the items of the questionnaire. The administration of the Italian version of the CRI-Adult showed internal consistency values ranging from α = 0.58 to α = 0.68 for the approach scales, and ranging from α = 0.57 to α = 0.66 for the avoidance scales [39]. In the present sample, α coefficients for the approach scales ranged from α = 0.60 to α = 0.72, while α values for the avoidance scales ranged from α = 0.60 to α = 0.65.

Depression Anxiety Stress Scale-21 (DASS-21; [40; 41]): 21-item self-report questionnaire assessing depression, anxiety, and stress on a 4-point Likert scale (from 0 = “did not apply to me at all” to 3 = “applied to me very much”), with higher scores indicating greater distress. The Italian version of the DASS-21 proved to be highly reliable (from α = 0.74 to α = 0.90 [41]. Findings of the Italian version suggested that the use of the total score, measuring a “general distress” factor, could be more appropriate than calculating the three subscales separately [41]. Therefore, we focused only on the total score of the scale (i.e., general distress). In the current study, the DASS-21 total score showed good internal consistency (Cronbach’s α = 0.94).

Brief pain inventory (BPI; [42; 43]): 15-item self-report questionnaire designed to measure pain intensity and impact of pain on daily functioning. The BPI allows to calculate a total score and two specific scores (pain severity and pain interference) on a 10-point Likert Scale (0 = “no pain”/ “does not interfere”; 10 = “pain as bad as you can imagine”/ “interferes completely”), with higher scores indicating higher pain severity and interference. The Italian version of the BPI showed good psychometric properties [43]. For the purposes of the study, we focused only on the total score of the questionnaire. In the current study, the BPI showed good internal consistency (Cronbach’s α = 0.90).

Statistical analyses

To investigate the relationship between body image distress (BIS) and physical (PCS-SF-12) and mental (MCS-SF-12) QoL, preliminary Pearson’s correlation analyses were performed. Correlations between age, disease duration, and self-report measures (i.e., scores obtained at the SF-12, BIS, RSES, IUS, SIAS, CRI-Adult, BPI, and DASS-21) were performed on the whole sample to identify variables to include in the regression model (see Supplementary Materials). Based on results emerged from correlational findings (i.e., only variables significantly correlated with the dependent variables were included in regression models), two multiple hierarchical regression models were performed. Pertaining to physical QoL (as measured by the PCS of the SF-12), a four-step multiple hierarchical regression analysis was performed, wherein the physical QoL (PCS) was included as dependent variable. Both time from diagnosis and DASS-21 total score were included in the first block to control for disease duration and general distress. Appropriate coping strategies (i.e., positive reappraisal and alternative rewards subscales of the CRI-Adult) were included in the second block, whereas variables related to mental QoL and pain (i.e., MCS and BPI total score) were included in the third block. Finally, psychological variables (i.e., RSES and BIS) were included in the fourth block. Also pertaining to mental QoL (as measured by the MCS of the SF-12), a four-step multiple hierarchical regression analysis was performed, wherein the mental QoL (MCS) was included as a dependent variable. The DASS-21 total score was included in the first block to control for general distress; coping strategies (i.e., logical analysis, resigned acceptance, and emotional discharge subscales of the CRI-Adult) were included in the second block, whereas variables related to physical QoL and pain (i.e., PCS and BPI total score) were included the third block. Then, psychological variables (i.e., RSES and BIS) were entered in the fourth block. The sequence of the 4 blocks of variables was driven by our specific research questions in conjunction with a theory-driven approach: we were interested in investigating the relationship between body image distress and QoL above and beyond relevant variables in HNC (HNC-related variables, general distress, coping strategies, physical/mental QoL, presence of pain, and psychological variables). Therefore, we statistically “control” for certain variables, to see whether adding variables significantly improved the model’s ability to account for physical/mental QoL. In the final/full model, it makes no difference as to when a given independent variable was entered; the estimated regression coefficients are conditional based on all other independent variables and the relationships in our data set, and not order of entry. We have indeed changed the sequence of the 4 blocks and tested both regression models: no differences between regression models emerged (i.e., body image distress emerged as associated with physical QoL and the presence of pain emerged as associated with mental QoL). To evaluate the final model fit, we employed the adjusted R squared (% of variance explained). All assumptions for multiple regression analysis were met.

To assess gender differences on socio-demographics (i.e., age and education), HNC-related variables, psychological variables, Chi-squared analyses (χ2), one-way analyses of variance (ANOVAs), and one-way analysis of covariance (ANCOVA)2 were performed. Before conducting these statistical analyses, we checked the normality of the distribution of self-report measures. Normality was not met for scores obtained at the SIAS, BIS, BPI, CRI-Adult guidance/support seeking, and CRI-Adult cognitive avoidance. However, analyses with the non-parametric Mann–Whitney test and the Rho Spearman correlation coefficient gave similar results to the ANOVA, ANCOVA, and Pearson’s, so we report on the latter here. Bonferroni correction was applied to the CRI-Adult and the SF-12 scores, revealing a significant p value of 0.006 for the CRI-Adult questionnaire and of 0.025 for the SF-12. Partial Eta Squared (η_p²) values were reported to evaluate the magnitude of effects. Cohen [44] has provided benchmarks to define small (η² = 0.01), medium (η² = 0.06), and large (η² = 0.14) effects when partial eta squared are computed.

A post-hoc analysis of the sample size was established using the G*Power 3.1 software [45]. The variables of the multiple hierarchical regression were considered: disease duration, DASS-21 total score, CRI-Adult logical analysis, CRI-Adult alternative rewards, CRI-Adult emotional discharge, CRI-Adult resigned acceptance, MCS, PCS, BPI total score, BIS total score, and RSES. A total sample of 51 patients allows to achieve a power of 0.81 in reliably detecting a one-tailed effect (f²) of 0.13, with a type I error of 0.05.

Less than 0.5% of the total dataset was missing. Given the minimal missing data, these were replaced using the mean replacement method (i.e., replacement using the mean of valid surrounding values).

All statistical analyses were conducted using IBM SPSS statistics [46], version 26, and the G*Power 3.1 software [45].

Psychological variables associated with physical QoL

The overall model explained 37.4% (adjusted R square) of the variance in physical QoL (PCS). Disease duration and general distress (DASS-21) were entered in the first step, but were not significantly associated with physical QoL (PCS) (F_(2,40) = 1.81, p = 0.18). The inclusion of coping strategies in the second step of the model did not explain an additional variation (13.6%) in physical QoL (PCS) (F change = 3.14; p = 0.06), despite alternative reward coping strategy (CRI-Adult) emerged as significant (p = 0.04) (i.e., alternative reward strategy emerged as significant, but the general 2nd step was not). The inclusion of mental QoL (MCS) and pain (BPI) in the third step of the model did not explain an additional variation (10.3%) in physical QoL (PCS) (F change = 2.60; p = 0.09), despite the step emerged as significant (p = 0.03). Finally, the inclusion of self-esteem (RSES) and body image distress (BIS) explained an additional 17.3% of the variation in QoL (PCS) (F change = 5.53; p = 0.01). Results showed that body image distress (BIS) was the only variable significantly associated (negatively) with physical health (PCS), whereas all the other variables were not (Table 2).

Psychological variables associated with mental QoL

The overall model explained 42.4% (adjusted R square) of the variance in mental QoL (MCS). General distress (DASS-21), entered in the first step of the regression model, emerged as significantly associated (negatively) with mental QoL (MCS) (F_(1,39) = 18.95, p < 0.001). The inclusion of coping strategies in the second step did not significantly explain an additional variation (4.1%) in mental QoL (MCS) (F change = 0.77; p = 0.52). Physical QoL (PCS) and pain (BPI), entered in the third step of the regression model, significantly accounted for mental QoL (MCS), F change = 5.48; p = 0.01, explaining an additional 15.6% of the variation. Finally, the inclusion of self-esteem (RSES) and body image distress (BIS) did not significantly explain an additional variation (1.2%) in mental QoL (MCS) (F change = 0.39; p = 0.68). Results showed that pain (BPI) was the only variable significantly associated (negatively) with mental QoL (MCS), whereas all the other variables were not (Table 3).

Gender differences in age, years of education, HNC-related variables, QoL, and psychological variables relevant in HNC

No significant differences between genders with respect to years of education emerged (p = 0.40), while age differed: males were older than females (p = 0.02). In terms of HNC-related variables, no gender differences emerged (all ps > 0.05).

With respect to psychological variables, significant differences between genders emerged: females scored higher than males on body image distress (BIS), social anxiety symptoms (SIAS), general distress (DASS-21), CRI-Adult emotional discharge, and pain (BPI) and scored lower than males on mental QoL (MCS) and self-esteem (RSES). No significant differences emerged with respect to physical QoL (PCS), intolerance of uncertainty (IUS-12), and other CRI-Adult subscales (all ps > 0.05). As a covariate, age was significant for both CRI-Adult positive reappraisal (F_(1,49) = 8.46, p = 0.01, ηp² = 0.15) and alternative rewards (F_(1,49) = 5.73, p = 0.02, ηp² = 0.11) (Table 4).

Below is the link to the electronic supplementary material.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee of University of Padova (Date: 07/11/2018; No. E402E1706176FBFC03E79ECD4E8E1BED).

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.