Dispositional shame and guilt as predictors of depressive symptoms and anxiety among adults with lung cancer: The mediational role of internalized stigma

Abstract

Purpose:

The current study investigated whether dispositional tendencies to experience shame and guilt (i.e., shame- and guilt- proneness) were associated with higher levels of internalized stigma and, in turn, higher depressive symptoms and anxiety in adults with lung cancer.

Method:

Participants (N = 50, 56.0% female) were men and women who received a clinical consultation for lung cancer and completed validated questionnaires. Mediation modeling using bootstrapping was used to characterize relationships between shame- and guilt-proneness, lung cancer stigma, depressive symptoms, and anxiety.

Results:

Higher guilt-proneness was associated significantly with higher anxiety (b = 0.69, SE=0.28, 95% CI [0.13, 1.26]) and higher shame-proneness was associated significantly with higher depressive symptoms (b = 0.56, SE = 0.19, 95% CI [0.18, 0.93]), beyond sociodemographic, medical, and smoking-related characteristics. Higher lung cancer stigma also significantly mediated the relationship between guilt-proneness and anxiety (indirect effect = 0.43, SE = .20, 95% CI [0.08, 0.89]) but not between shame-proneness and depressive symptoms.

Conclusions:

Shame- and guilt-proneness are associated significantly with depressive symptoms and anxiety, respectively, and the relationship between guilt-proneness and anxiety is explained in part by internalized stigma in a sample of newly diagnosed lung cancer patients. Findings carry implications for the early identification of lung cancer patients in need of additional supportive care services and highlight internalized stigma as a target for psychosocial intervention.

Stigma—a characteristic that designates an individual as socially devalued (Goffman, 1963)—is conceptualized as a stressor that contributes to psychological morbidity (Hatzenbuehler, Phelan, & Link, 2013; Meyer, 2003; Nelson, 2002). Patients diagnosed with diseases such as lung cancer or HIV are often targets of stigma and face negative societal attitudes that these diseases are self-inflicted (Chambers et al., 2012; Surlis & Hyde, 2001), owing to associations between these diseases and specific risky health behaviors (e.g., smoking in the case of lung cancer, unprotected sex and/or drug use in the case of HIV). These perceptions of negative societal attitudes can be internalized by the individual and are often experienced as feelings of shame and guilt (Stuber, Meyer, & Link, 2008). Indeed, shame and guilt are common feelings among patients who live with stigmatized health conditions including lung cancer (Chambers et al., 2012), HIV (Kalichman et al., 2009), diabetes (Browne, Ventura, Mosely, & Speight, 2013), alcoholism (Potter-Efron, 2002), obesity (Tomiyama et al., 2018), and mental illness (Clement et al., 2015). Additionally, higher internalized stigma is associated with poor mental health outcomes such as depressive symptoms and anxiety across a variety of populations (Gilkes, Perich, & Meade, 2019; Livingston & Boyd, 2010; Mak, Poon, Pun, & Cheung, 2007).

To better understand the nature and impact of stigma, there have been calls for researchers to examine the psychosocial contributors to and correlates of internalized stigma broadly (Gilkes et al., 2019; Livingston & Boyd, 2010) and for stigmatized medical populations specifically (Earnshaw, Smith, Chaudoir, Amico, & Copenhaver, 2013). Recent research in this area has demonstrated that internalized stigma is associated significantly with psychosocial factors such as emotion dysregulation (Wong et al., 2019), a low sense of belonging (Hirsch et al., 2019), and causal beliefs about illness (Carter, Read, Pyle, & Morrison, 2019). Additionally, there is a growing body of empirical evidence suggesting that dispositional tendencies to experience feelings of shame and guilt (i.e., shame- and guilt- proneness) may contribute to internalized stigma (Bennett, Traub, Mace, Juarascio, & O’Hayer, 2016; Hasson-Ohayon et al., 2012).

The primary distinction made between shame and guilt is that shame is characterized by a negative evaluation of the self, whereas guilt is a negative appraisal of one’s own specific behavior rather than the self as a whole (Tangney, Yourman, & Stuewig, 2009). Relatedly, shame- and guilt-proneness are conceptualized as trait-motivational characteristics of individuals who have a tendency to experience feelings of shame and guilt (Giner-Sorolla, Piazza, & Espinosa, 2011; Tangney, 1996; Tangney et al., 2009). These dispositional tendencies are associated with higher depressive symptoms and anxiety in the general population (Candea & Szentagotai-Tata, 2018; Fergus, Valentiner, McGrath, & Jencius, 2010; Young, Neighbors, DiBello, Traylor, & Tomkins, 2016) and in chronic illness samples (Grynberg, de Timary, Van Heuverswijn, & Maurage, 2017; Wertheim et al., 2018). Given that prior research has demonstrated significant relationships between shame- and guilt-proneness and internalized stigma (Bennett et al., 2016; Hasson-Ohayon et al., 2012) and between internalized stigma and depressive symptoms (Livingston & Boyd, 2010; Ostroff et al., 2019), it is possible that a higher dispositional tendency to experience shame and guilt may predict stronger feelings of internalized stigma, which in turn may predict poorer psychological outcomes such as depressive symptoms and anxiety (See Figure 1 for conceptual model). The current study tested this model within a sample of adults recently diagnosed with lung cancer—a disease stigmatized primarily due to its robust association with smoking and the perception of the disease as self-inflicted (Chambers et al., 2012; Hamann et al., 2014).

Figure 1.

Hypothesized model of shame- and guilt-proneness predicting higher internalized stigma and, in turn, higher psychological distress in the context of living with a stigmatized chronic disease such as lung cancer.

Much like other stigmatized patient populations, lung cancer patients commonly report feelings of internalized stigma (Cataldo, Slaughter, Jahan, & Hwang, 2011; Hamann et al., 2014; Hamann, Shen, Thomas, Lee, & Ostroff, 2018) and evidence higher depressive symptoms and anxiety compared to the general population and other cancer samples (Brintzenhofe-Szoc et al., 2009; Zabora, Brintzenhofeszoc, Curbow, Hooker, & Piantadosi, 2001). Additionally, higher levels of lung cancer stigma are associated with poorer psychological, physical, and functional outcomes (Criswell, Owen, Thornton, & Stanton, 2016; Steffen, Vowles, Smith, Gan, & Edelman, 2018; Williamson et al., 2018), including higher depressive symptoms (Chambers et al., 2012; Gonzalez & Jacobsen, 2012; Maguire, Lewis, Kotronoulas, McPhelim, & Cataldo, 2019; Ostroff et al., 2019) and anxiety (Brown Johnson, Brodsky, & Cataldo, 2014; Cataldo & Brodsky, 2013). As such, it is crucial to identify characteristics associated with higher internalized stigma and poor mental health outcomes in lung cancer patients to promote early identification of patients who could benefit from psychosocial support services that aim to reduce stigma and improve health.

No study to our knowledge has examined shame- and guilt-proneness in a sample of lung cancer patients specifically. One study, however, measured shame- and guilt-proneness in a general sample of cancer patients in which participants with any type of cancer were eligible (Wertheim et al., 2018). This study demonstrated that shame-proneness (but not guilt-proneness) was associated with higher depressive symptoms and anxiety. Additionally, general and cancer- specific concealment (i.e., actively trying to conceal information related to the one’s cancer) partially mediated the relationships between shame-proneness and depressive symptoms as well as between shame-proneness and anxiety, which prompted the authors to highlight the need for investigating additional mediators of these relationships among people living with cancer (Wertheim et al., 2018).

Research is needed to understand relationships between dispositional tendencies (e.g., shame- and guilt-proneness) and contextual factors (e.g., being diagnosed with lung cancer, facing negative societal attitudes) that may contribute to higher feelings of internalized stigma, depressive symptoms, and anxiety (McQueary Flynn, Eisenlohr-Moul, Segerstrom, Logue, & Studts, 2017; Wertheim et al., 2018). Such an investigation would advance psychological theory and yield clinical implications for supportive care. As such, the primary aim of this study is to test the hypotheses that higher shame- and guilt-proneness would be associated significantly with higher depressive symptoms and anxiety and that those relationships would be mediated, in part, by higher internalized lung cancer stigma.

Method

Participants

Participants (N = 50) were men and women who received a clinical consultation for lung cancer. Participants were eligible if they were: 1) a new patient diagnosed within the past six months or within the first three visits at the institution; 2) being seen for lung cancer or a mass suspicious of lung cancer; 3) formerly or currently smoking cigarettes (i.e., smoked >100 packs in their lifetime); 4) fluent in English; 5) able to provide informed consent; and 6) able and willing to complete all study procedures. Participants who never smoked were excluded from participation in an effort to target patients at high risk for experiencing internalized stigma (Williamson et al., 2018). A power analysis for mediation (Schoemann, Boulton, & Short, 2017), which used 50,000 bootstrap replications and 20,000 Monte Carlo draws per repetition, revealed that a sample size of 54 participants would provide 75% power to detect as significant (p < .05) a moderate-to-large effect size, which is equivalent to effect sizes from previous research on shame- and guilt-proneness and mental health outcomes (Cândea & Szentagotai-Tăta, 2018; Grynberg et al., 2017). All participants provided written informed consent, and all procedures were approved by the Institutional Review Board at Memorial Sloan Kettering Cancer Center.

Procedure

Eligible participants for the study were identified through physicians’ clinics at Memorial Sloan Kettering Cancer Center and were approached by a member of the research team during a regularly scheduled medical appointment. Eligible participants who provided informed consent were asked to complete validated questionnaires on sociodemographic information, medical characteristics, shame- and guilt-proneness, lung cancer stigma, depressive symptoms, and anxiety within three days of their consultation. Participants could complete the questionnaires remotely through a secure web-based portal, over the phone with a member of the research team, or onsite via a paper-and-pencil survey.

Measures

Shame- and guilt-proneness.

The 10-item Personal Feelings Questionnaire was used to assess dispositional traits of shame- and guilt-proneness (Harder & Zalrna, 1990). Using this well-validated global adjective checklist, participants endorsed the degree to which they experience frequent feelings that reflect dispositional guilt (e.g., “regret, remorse”) or shame (e.g., “humiliated, stupid”) on a scale from 0 (“never experience this feeling”) to 4 (“continuously experience this feeling”). Of note, patients responded to these items without reference to lung cancer. Subscale scores for shame- and guilt-proneness were computed and current estimates of internal consistency reliability were adequate (α = .68, .71).

Lung cancer stigma.

The Cataldo Lung Cancer Stigma Scale was used to assess patients’ experience of stigma. The 31-item scale has demonstrated good reliability and validity (Cataldo et al., 2011) and is comprised of four subscales (stigma and shame, social isolation, discrimination, and smoking). Participants rated items on a 4-point Likert scale (“strongly disagree” to “strongly agree”) with higher scores indicating higher stigma. To focus specifically on internalized lung cancer stigma, the 11-item stigma and shame subscale (e.g., “I feel guilty because I have lung cancer”) was used¹. Internal consistency reliability for the stigma and shame subscale was adequate (α = .76).

Depressive symptoms and anxiety.

The 14-item Hospital Anxiety and Depression Scale was used to assess depressive symptoms and anxiety (Zigmond & Snaith, 1983). Participants rated items on a 4-point Likert scale (“not at all” to “nearly all the time”). The 7-item anxiety and depression subscales were used as the dependent variables in this study. Internal consistency reliability estimates for the anxiety (α = .86) and depression (α = .69) subscales were adequate. The scale has been extensively validated for use in oncology samples (Bjelland, Dahl, Haug, & Neckelmann, 2002).

Analytic Strategy

Shame- and guilt-proneness were entered as independent variables in separate models, internalized lung cancer stigma was entered as a single mediator in all models, and depressive symptoms and anxiety were entered as dependent variables in separate models. Mediation analyses were conducted following well-established guidelines (Preacher & Hayes, 2008). Four mediation models were conducted using PROCESS, a statistical program which uses an ordinary least squares framework to test total, direct, and indirect effects. Bootstrap estimates using 20,000 repetitions were used to construct 95% bias-corrected confidence intervals for indirect effects, which is an analytic approach to mediation that has been used previously with small samples (Aldred, Green, Emsley, & McConachie, 2012; Killgore, Sonis, Rosso, & Rauch, 2016). Indirect effects were determined to be statistically significant if the bootstrap confidence interval did not include zero. Notably, research indicates that indirect effects can exist in the absence of a significant direct effect or a non-significant total effect (MacKinnon, Krull, & Lockwood, 2000; Preacher & Hayes, 2008; Shrout & Bolger, 2002). As such, the traditional “four-step” process (Baron & Kenny, 1986) of testing mediation was not used. Based on research, age, sex, disease stage, and smoking status were selected as a priori covariates (Hinz et al., 2010; Linden, Vodermaier, MacKenzie, & Greig, 2012; Lo et al., 2010; Williamson et al., 2018). Per an a priori decision, any other sociodemographic variable (i.e., education, marital status, race/ethnicity) associated with the predictor, mediator, or outcome at p < .05 was also added as a covariate. Two-tailed tests were used for all analyses, and p < .05 was considered statistically significant. Variance inflation factor and tolerance statistics were computed to test for multicollinearity.

Results

Descriptive Statistics

Sample characteristics.

Of the 99 patients approached to participate, three were ineligible, 26 declined to participate, and 14 did not complete the study materials. Additionally, six participants who had a mass suspicious of lung cancer that was later not diagnosed as lung cancer were excluded from the analyses². As shown in Table 1, on average, participants were 67.78 years old. The majority was married or living as married, non-Hispanic white, formerly smoked, and had relatively early stage disease (i.e., stages I-IIIA non-small cell lung cancer or limited stage small cell lung cancer). Participants with less than a bachelor’s degree level of education were combined into one group for analyses and compared to participants who obtained a bachelor’s degree or beyond. Zero-order bivariate correlations between shame-proneness, guilt- proneness, lung cancer stigma, anxiety, and depressive symptoms are displayed in Table 2.

Table 1.

Sample characteristics and demographics (N = 50)

	n	Mean	Standard Deviation	Possible Range
Age (in years)	50	67.78	9.08	--
PFQ Shame-Proneness	49	2.01	2.56	0-20
PFQ Guilt-Proneness	49	2.22	2.28	0-20
CLCSS Shame and Stigma	49	17.94	4.95	11-55
HADS Anxiety	50	6.11	4.57	0-21
HADS Depressive Symptoms	50	4.31	3.09	0-21
	n	%
Sex
Male	22	44.0
Female	28	56.0
Race/Ethnicity
Non-Hispanic white	39	78.0
Other race/ethnicity	9	18.0
Did not report	2	4.0
Marital status
Married/Living as married	31	62.0
Not married	17	34.0
Did not report	2	4.0
Education
Partial high school	3	6.0
High school graduate	7	14.0
Partial college	10	20.0
Bachelor’s degree	12	24.0
Graduate degree	18	36.0
Smoking status
Currently smoked	9	18.0
Formerly smoked	41	82.0
Stage of disease
Early Stage	27	54.0
Late Stage	21	42.0
Missing	2	4.0

Note: PFQ = Personal Feelings Questionnaire. CLCSS = Cataldo Lung Cancer Stigma Scale. HADS = Hospital Anxiety and Depression Scale. Early Stage = Stages I-IIIA non-small cell lung cancer and limited stage small cell lung cancer. Late stage = Stages IIIB-IV non-small cell lung cancer and extensive stage small cell lung cancer.

Table 2.

Zero-order correlations between shame-proneness, guilt-proneness, lung cancer stigma, anxiety, and depressive symptoms (N = 50)

		1	2	3	4
1	PFQ Shame-proneness	--	--	--	--
2	PFQ Guilt-Proneness	.54*	--	--	--
3	CLCSS Shame and Stigma	.17	.49*	--	--
4	HADS Anxiety	.24	.35*	.45*	--
5	HADS Depressive Symptoms	.43*	.34*	.22	.48*

Note: PFQ = Personal Feelings Questionnaire. CLCSS = Cataldo Lung Cancer Stigma Scale. HADS = Hospital Anxiety and Depression Scale.

^*= p < .05.

Evaluating covariates.

Women (compared to men) and participants with earlier stage disease (compared to late stage) reported higher anxiety (all p < .01). Participants with less than a bachelor’s degree level of education (compared to those with bachelor’s or beyond) and those who currently (compared to formerly) smoked reported higher levels of lung cancer stigma (all p < .05). Finally, non-married participants reported higher shame- and guilt-proneness, as compared to married participants (p = .009). Shame-proneness, guilt-proneness, lung cancer stigma, anxiety, and depressive symptoms were not associated significantly with any other potential covariate (all p > .05). Therefore, in addition to a priori covariates (i.e., age, sex, disease stage, smoking status), education and marital status were included as covariates.

Mediation Analyses

In fully adjusted models, the total effect of guilt-proneness on anxiety was statistically significant (See Figure 2a), but the total effect of guilt-proneness on depressive symptoms was non-significant (See Figure 2b). Higher guilt-proneness (but not shame-proneness) was associated significantly with higher internalized lung cancer stigma. In turn, higher lung cancer stigma was associated significantly with higher anxiety (but not depressive symptoms). There was a significant indirect effect of lung cancer stigma in mediating the relationship between guilt-proneness and anxiety (but not depressive symptoms), beyond sociodemographic, medical, and smoking-related factors. After controlling for the indirect effect, the direct effect between guilt-proneness and anxiety was no longer statistically significant, indicating that internalized lung cancer stigma fully mediated this relationship (Baron & Kenny, 1986). A post-hoc analysis demonstrated that this mediated effect remained statistically significant even when adding shame-proneness as an additional covariate to the model (indirect effect = .50, SE = .24, 95% CI [0.11, 1.06]). Due to the cross-sectional nature of the data, we also conducted a post-hoc analysis to test whether guilt-proneness mediated the relationship between lung cancer stigma and anxiety. The indirect effect, however, was statistically non-significant (indirect effect = .06, SE = .208 95% CI −0.08, 0.23]), further supporting the hypothesized model that a dispositional tendency to experience guilt is predictive of greater disease-specific feelings of internalized stigma, which in turn are associated with higher anxiety.

Figure 2.

Results for analyses of internalized lung cancer stigma mediating the relationship between guilt-proneness and anxiety and depressive symptoms. All results are adjusted for age, sex, education, marital status, disease stage, and smoking status. LC = lung cancer. SE = standard error. CI = confidence interval. * = p < .05.

As shown in Figure 3, the total effect of shame-proneness on depressive symptoms was statistically significant but the total effect of shame-proneness on anxiety was non-significant, controlling for covariates. Higher shame-proneness was not associated significantly with internalized lung cancer stigma. Higher lung cancer stigma was associated significantly with higher anxiety (but not depressive symptoms). There was no significant indirect effect of lung cancer stigma in mediating the relationship between shame-proneness and either outcome.

Figure 3.

Results for analyses of internalized lung cancer stigma mediating the relationship between shame-proneness and anxiety and depressive symptoms. All results are adjusted for age, sex, education, marital status, disease stage, and smoking status. LC = lung cancer. SE = standard error. CI = confidence interval. * = p < .05.

Regarding covariates, smoking status was associated significantly with lung cancer stigma in all models (all p < .05). Specifically, participants who currently smoked reported higher levels of stigma than those who formerly smoked. Additionally, women (vs. men) and those with early stage (vs. late stage) disease reported higher levels of anxiety (all p < .02). No other covariate was associated significantly with the lung cancer stigma, depressive symptoms, or anxiety in fully-adjusted models (all p > .05). None of the variables included in the analyses were found to have variation inflation scores above 2.50 or tolerance statistics below .40, which would indicate prominent multicollinearity (Allison, 1999)³.

Discussion

In this cross-sectional study of adults newly diagnosed with lung cancer, higher guilt-proneness was associated significantly with higher anxiety and higher shame-proneness was associated significantly with higher depressive symptoms, beyond sociodemographic, medical, and smoking-related characteristics. Additionally, higher internalized lung cancer stigma significantly mediated the relationship between guilt-proneness and anxiety. Results are consistent with research demonstrating that higher levels of shame- and guilt-proneness are associated with poor mental health outcomes (Cândea & Szentagotai-Tăta, 2018; Gray, Fitch, Phillips, Labrecque, & Fergus, 2000; Grynberg et al., 2017; Wertheim et al., 2018; Young et al., 2016) and that higher lung cancer stigma is correlated with higher anxiety (Brown Johnson et al., 2014; Cataldo & Brodsky, 2013). The current findings highlight shame- and guilt-proneness and internalized lung cancer stigma as important psychosocial constructs related to depressive symptoms and anxiety in lung cancer patients.

Internalized lung cancer stigma significantly mediated the relationship between guilt-proneness and anxiety but not between shame-proneness and depressive symptoms. A potential explanation for this finding is that internalized stigma might explain the relationship between guilt-proneness and anxiety, whereas a distinct process—such as self-concealment—may explain the relationship between shame-proneness and depressive symptoms. Indeed, findings from Wertheim and colleagues (2018) demonstrated that trait-level and disease-specific self-concealment significantly mediated the relationship between shame-proneness and depressive symptoms but not between guilt-proneness and anxiety. Future research should examine whether internalized stigma is more closely related to indicators of anxiety (e.g., rumination) and whether self-concealment is more closely associated with indicators of depression (e.g., hopelessness). Taken together, the pattern of findings from Wertheim et al. (2018) and the current study suggest that both internalized stigma and self-concealment may confer risk for adverse mental health in cancer patients, consistent with theoretical perspectives of lung cancer stigma (Hamann et al., 2014). Additionally, lung cancer patients high in shame- and guilt-proneness may be at particular risk to experience high levels of self-concealment and internalized stigma, respectively.

Surprisingly, internalized stigma was not associated significantly with depressive symptoms, which is inconsistent with past research (Brown Johnson et al., 2014; Gonzalez & Jacobsen, 2012; Maguire et al., 2019; Ostroff et al., 2019). Notably, only 4% (2 out of 50) of participants in the current study reported depressive symptoms above the suggested clinical cut-off of 11 (Crawford, Henry, Crombie, & Taylor, 2001), whereas 18% (9 out of 50) reported clinical levels of anxiety. In another study of more than 600 lung cancer patients, 18% experienced clinical levels of depression and 26% experienced clinical levels of anxiety (Linden et al., 2012), suggesting that participants in the current study evidenced relatively low levels of depressive symptoms. Therefore, the non-significant relationship between internalized stigma and depressive symptoms may be explained by participants’ overall low levels of depressive symptoms.

The present study also showed that guilt-proneness is associated with anxiety and that shame-proneness is associated with depressive symptoms, indicating differential relationships between guilt- and shame-proneness and specific indicators of psychological distress. Some studies have shown that shame- and guilt-proneness are associated with both higher depressive symptoms and anxiety (Cândea & Szentagotai-Tăta, 2018; Wertheim et al., 2018; Young et al., 2016). Others, however, have found that shame-proneness is more strongly related to depressive symptoms and anxiety than guilt-proneness (Fergus et al., 2010; Pineles, Street, & Koenen, 2006), perhaps because a negative evaluation of the entire self (shame) is more indicative of a depressogenic attribution style than a negative evaluation of one’s own specific behavior (guilt) (Pineles et al., 2006). Also, some researchers have suggested that relationships between guilt- proneness and psychological symptoms are accounted for by shame-proneness (Pineles et al., 2006; Tangney, Burggraf, & Wagner, 1995). However, our results showed that the indirect effect of lung cancer stigma mediating the relationship between guilt-proneness and anxiety remained significant (and even yielded a larger effect size) when shame-proneness was added as a covariate in the model. We suggest that the contextual factors of a lung cancer diagnosis and the timing of recently completing a clinical consultation may contribute to the current findings. Specifically, guilt experienced by lung cancer patients is often tied closely to one’s smoking history (LoConte, Else-Quest, Eickhoff, Hyde, & Schiller, 2008). Clinical consultations for lung cancer include gathering information about patients’ smoking history (Novello et al., 2016), which may prompt feelings of guilt and anxiety for some patients. Given that all participants in the current study currently or formerly smoked, feelings of guilt may be more salient than shame and may relate more strongly to experiencing anxiety in a clinical consultation with a physician.

Participants who currently smoked reported higher stigma than those who formerly smoked, which is consistent with research (Williamson et al., 2018) and theory (Hamann et al., 2014). Women reported higher levels of anxiety than men, also consistent with research (Linden et al., 2012; LoConte et al., 2008). Surprisingly, participants with early stage (vs. later stage) disease reported higher anxiety in this study, whereas other studies have found opposite (Hinz et al., 2010) or null findings (Kurtz, Kurtz, Stommel, Given, & Given, 2002). Disease stage was not associated significantly with any other sociodemographic, medical, or psychosocial variable. However, disease stage may be a proxy for a more proximally related psychological variable not measured in the current study (e.g., threat appraisals), which may relate more closely to anxiety (Britton, Lissek, Grillon, Norcross, & Pine, 2011).

Study limitations and recommendations for future research should be noted. First, the cross-sectional nature of the data precludes any interpretation of causality. Although shame- and guilt-proneness are posited to be stable, trait-like characteristics (Tangney et al., 2009), direct conclusions about directionality are also precluded. However, a post-hoc analysis supported the hypothesized direction of findings in that lung cancer stigma mediated the relationship between guilt-proneness and anxiety, whereas guilt-proneness was not a significant mediator of the relationship between stigma and anxiety. Longitudinal research is needed to test whether shame- and guilt-proneness predict increases in lung cancer stigma and subsequent increases in depressive symptoms and anxiety, which would bolster causal inference. Second, variables not included in this study may be associated with the independent and/or dependent variables. Specifically, future research should test internalized stigma and self-concealment as simultaneous mediators of the relationships between shame- and guilt-proneness and adverse mental health outcomes to test their unique roles in explaining these relationships. Third, the statistical power to detect significant mediation was restricted by the sample size, particularly if the true effect sizes of the hypothesized indirect effects are small-to-medium. As such, additional research with larger samples are needed to replicate the current findings and examine whether small-to-medium effects are produced from statistical models that were nonsignificant in the current study. Relatedly, an additional limitation of this study is the relatively low number of participants who currently smoked. Finally, the majority of participants was non-Hispanic white and attained a bachelor’s degree level of education or beyond; thus, caution is warranted in generalizing the findings more broadly. Additional research is needed to test whether the current findings generalize to the broader population of newly diagnosed lung cancer patients, particularly to those from diverse racial, ethnic, and sociodemographic backgrounds and/or who are treated in community oncology practices or rural settings. Finally, researchers have critiqued the use of the Hospital Anxiety and Depression Scale to measure depressive symptoms and anxiety in cancer populations, due to the heterogeneity of the latent structure of the measure across studies (Cosco, Doyle, Ward, & McGee, 2012); however, the internal consistency estimates in the current study were adequate and indicate it is likely an appropriate measure.

This study has several theoretical and clinical implications. These findings contribute to theoretical models of lung cancer stigma specifically (Hamann et al., 2014) and health-related stigma broadly (Weiss, Ramakrishna, & Somma, 2006) through identifying dispositional traits that may contribute to internalized stigma. Theoretical models would benefit from the incorporation of upstream dispositional psychosocial traits such as shame- and guilt-proneness, which may precede the internalization of stigma and its downstream effects on negative health outcomes. Regarding implications for practice, the current findings can be used to identify patients at risk for developing high levels of internalized stigma, depressive symptoms, and anxiety, who could benefit from psychosocial support. Although data from the current study are cross-sectional and directionality cannot be inferred, patients’ tendency to experience feelings of shame and guilt are theoretically posited to be stable traits that exist prior to the diagnosis of lung cancer and the subsequent internalization of lung cancer stigma. Thus, lung cancer patients high in shame- and guilt-proneness may be able to be identified early in the cancer care trajectory (e.g., at diagnosis) for proactive, multi-level psychosocial interventions (e.g., health care provider communication skills training, acceptance and commitment therapy) that may prevent internalized stigma and mitigate psychological maladjustment.

This study adds to the growing body of evidence that lung cancer stigma is associated with adverse consequences for patients’ health and well-being (Chambers et al., 2012; Maguire et al., 2019; Ostroff et al., 2019; Williamson et al., 2018). Additionally, our findings may promote early identification of patients who need psychosocial support and point to additional factors that can be targeted for prevention and intervention. Finally, this study supports the recommendation that psychological processes (e.g., shame- and guilt-proneness) should be considered when developing interventions to reduce stigma (Hamann, Ver Hoeve, Carter-Harris, Studts, & Ostroff, 2018).