Neuroticism, Cancer Mortality Salience, and Physician Avoidance in Cancer Survivors: Proximity of Treatment Matters

Abstract

Objective:

To examine if the relationship between neuroticism and physician avoidance/physician visit concerns are mediated by perceptions that cancer is associated with death (“cancer mortality salience”; CMS) for cancer survivors to inform public health interventions and tailored health communications.

Methods:

Cancer survivors comprised 42.3% of the total sample (n=525). Participants completed a 4-item neuroticism scale, 4-item cancer perceptions scale, and 4-item physician avoidance and concerns scale. Multiple linear regression models were used to assess relationships among variables for cancer survivors and separately for those without a history of cancer.

Results:

Neuroticism was positively associated with CMS for cancer survivors, b = 0.26, (p < .001), and those without cancer, b = 0.22, (p < .001). There was an association between neuroticism and physician avoidance among cancer survivors with temporally distant treatment courses after controlling for CMS, b = 0.56 (p = .006), but not for those currently or recently having had undergone treatment (p = .949). There was also an indirect relationship between neuroticism and physician visit concerns that was mediated by CMS for cancer survivors, b = 0.07, CI = [0.03, 0.13], but this relationship was again driven by cancer survivors with more distal treatment courses.

Conclusions:

High neuroticism in cancer survivors is associated with physician avoidance and physician visit concerns when treatment is temporally distant. Interventions aimed at decoupling the association between cancer and death can help increase the willingness of cancer survivors to attain cancer care follow-ups and healthcare more generally.

Background

An estimated 67% of people with a history of cancer in 2019 were diagnosed five or more years before. Due to advances in detection and treatment that increase length of survival after diagnosis, that percent is expected to increase, with the population of cancer survivors in the United States expected to reach 22.1 million by 2030.^1,2 For this growing population of survivors, continued regular physician visits, healthcare screening, and follow-up care are instrumental for general health and preventing negative outcomes related to cancer and other morbidities. Physician avoidance is a widespread phenomenon with approximately 29.4% of individuals over the age of 50 and 40.4% of people under the age of 50 reporting avoiding visiting their doctor, even when they suspect they should.³ However, cancer survivors may be particularly unmotivated to seek follow-up care—or, further, may avoid it entirely. Based on the population based Medical Expenditure Panel (2008–2015), cancer survivors’ reported rates of visiting a doctor within the previous year (13.6%−19.2%) were much lower than those without a history of cancer (20.2%−33.3%).⁴ Experts have called for more in-depth research into the cultural and psychological factors behind healthcare-avoidant behaviors in cancer survivors.^5,6 As such, a greater understanding of the roots of physician avoidance in cancer survivors may guide public health interventions to increase the uptake of crucial follow-up screening and care, and healthcare-seeking behaviors more generally.

One particular psychological factor associated with physician avoidance is the perception of cancer as a “death sentence”,⁷ also known as cancer mortality salience (CMS). Across three timepoints from 2008–2017, between 57.6% and 64.1% of respondents in the United States espoused CMS⁸ and these individuals had 44% higher odds of avoiding their physicians.⁷ Having a prior history of cancer was also an important factor determining whether an individual associated cancer with death, with those previously diagnosed with cancer exhibiting lower levels of CMS,⁸ presumably because living with cancer weakened its association with death. Yet it is unclear if there are other underlying traits connecting cancer survivorship, CMS, and physician avoidance.

Another psychological variable critical to examine in clinical contexts for use of cancer care and healthcare settings more generally is personality, and neuroticism specifically, given its relationship with mental stress and quality of life.⁹ Preliminary evidence suggests that neuroticism—also referred to as negative emotionality or affectivity¹⁰—may play a role in physician avoidance. For example, females with high neuroticism are less willing to undergo mammograms and more uncomfortable while undergoing mammograms in death salient contexts (i.e., after being reminded of their physical nature).¹¹ However, no research to date has examined the relationships among neuroticism, CMS, and physician avoidance. The goal of this cross-sectional exploratory study is to examine the relationship between neuroticism and physician avoidance while testing CMS as a mediating variable shown to predict physician avoidance, both within cancer survivors and among those without a history of cancer. Results from this study will inform public health efforts to reduce physician avoidance, potentially leading to enhanced participation in screening and healthcare visits, during cancer survivorship.

Methods

Participants

We used a prescreen (n=911) on a crowdsourcing convenience sampling platform called Prolific to determine personal history of cancer (e.g., “Have you ever been diagnosed with having cancer [Y/N?]”).¹² Cancer survivors were oversampled so that comparisons to those without a history of cancer could be made. After constraining the number of participants without a history of cancer from the prescreen we invited participants (n=683) to the main survey. A final sample of 525 participants (222 with a cancer history and 303 without) completed the main survey. Across the prescreen and main survey, 21 incongruencies were found for participant responses to the personal history of cancer question included in each survey (4% of entire sample). We contacted these participants anonymously on the Prolific messaging platform and 13 clarified their incongruency. For the other eight, we used their personal history of cancer response within the main survey. Analyses were run with and without these eight participants and no meaningful differences were found, so all were retained for analyses.

Procedure

This study received ethical approval by the National Institutes of Health Institutional Review Board (Protocol #539031) as part of a larger study. After agreeing to the online consent form for the main survey within Qualtrics, participants completed measures of neuroticism, cancer perceptions and physician avoidance questions, health status, and demographics.

Materials

Neuroticism.

Four items (e.g., “is relaxed, handles stress well [reverse scored],” “worries a lot,” “remains calm in tense situations [reverse scored],” “gets nervous easily) were used to assess the anxiety facet of neuroticism from the Big-Five Inventory.¹³ We used the anxiety facet of neuroticism due to the relationship that anxiety has with physical well-being in cancer patients¹⁴ which may inform the negative relationship between well-being and physician avoidance more generally.¹⁵ Participants were instructed to rate the degree to which they felt each item represented them on a 1–5 scale (disagree strongly to agree strongly). This scale exhibited good internal consistency (α = 0.85).

Cancer Mortality Salience.

CMS was measured with one survey item from the Health Information National Trends Survey (HINTS; https://hints.cancer.gov/), “when I think of cancer I automatically think of death”.⁷ CMS was measured on a 1–4 scale (disagree strongly to strongly agree).

Perceptions: Cancer Fatalism, Prevention, and Confusion.

To assess the unique predictive value of CMS as a mediator between neuroticism and physician avoidance, we controlled for three related cancer perception items from HINTS. Specifically, we measured cancer fatalism (e.g., “it seems like everything causes cancer”), prevention (e.g., “there’s not much you can do to lower your chances of getting cancer”) and confusion (e.g., “there are so many recommendations about preventing cancer, it’s hard to know which ones to follow”) on a 1–4 scale (disagree strongly to strongly agree).¹⁶

Physician Avoidance and Physician Visit Concerns.

Four items from HINTS assessed physician avoidance.⁷ The first item used binary true or not true response options (“Some people avoid visiting their doctor, even when they suspect they should. Would you say this is true for you or not true for you?”). The other three items (α = 0.74) examined physician visit concerns and were measured on a 1–4 scale (disagree strongly to agree strongly) using the following stem: “If I avoid seeing my doctor, it is because…I feel uncomfortable when my body is being examined,” “I fear I may have a serious illness,” and “It makes me think about dying.”

General Health.

For general health, participants were asked “in general, how would you say your health is,” on a scale of 1–5 (excellent to poor), and scores were reverse coded.¹⁷

Demographics.

Participants were asked their age, birth assigned sex, education, income, race, relationship status, and cancer treatment history, if applicable (e.g., “are you currently undergoing treatment for your cancer” [current] or “have you undergone treatment for your cancer in the last year” [recent]).

Statistical Analyses

All analyses were conducted using SPSS version 25. Chi-square tests compared whether categorical demographic characteristics and physician avoidance (binary outcome) differed based on cancer survivorship status. T-tests compared if continuous variables differed based on cancer survivorship status (e.g., neuroticism, CMS, cancer perceptions, and physician visit concerns). All statistical tests were 2-tailed. The PROCESS macro for SPSS was used to estimate if the relationship between of neuroticism (anxiety) and 1) physician avoidance 2) physician visit concerns were mediated by CMS (Model 4).¹⁸ Estimates were derived from 5,000 bias corrected bootstrap samples. Analyses for the binary physician avoidance outcome and the continuous physician visit concern outcome were conducted separately for those with a history of cancer and those without. Additional regression analyses were conducted within the cancer survivor group based on recency of treatment. Control variables in each regression model included age, cancer perceptions (e.g., cancer fatalism, prevention, confusion), education, sex, general health, income, race, and relationship status.

Results

Participant Characteristics (Cancer Survivors vs. No Cancer History)

Chi-square analyses indicated that, compared to those with no history of cancer, cancer survivors were significantly more likely to be older, educated, female, less healthy, more wealthy, White/Caucasian, and married (ps < .013) (See Table 1 for participant characteristics). Cancer survivors were less likely to avoid their physician compared to those without a history of cancer (41.0% vs. 55.4%), p = .001. Bivariate correlations for continuous variables are found in Table 2. T-test analyses indicated that cancer survivors, compared to those without a history of cancer, exhibited significantly lower levels of neuroticism (anxiety), t(523) = −3.19, p = .002, and CMS t(523) = −4.34, p < .001. No significant differences were found between these two groups for cancer perceptions or the physician visit concerns (ps > .148).

Table 1.

Participant characteristics (n=525)

	Hx of Cancera	No Hx of Cancer
Characteristic	Total No. (%)	Total No. (%)
Total	222	303
Age*
18–39	73 (33.0)	242 (80.4)
40–64	112 (50.7)	55 (18.3)
65+	36 (16.3)	4 (1.3)
Education*
High school/GED or lower	39 (17.6)	92 (30.5)
Post high school training (vocational/technical)	24 (10.8)	17 (5.6)
College graduate (Associate’s degree)	37 (16.7)	38 (12.6)
College graduate (Bachelor’s degree)	74 (33.3)	107 (35.4)
College graduate (Grad degree)	48 (21.6)	48 (15.9)
Sex*
Male	65 (29.3)	136 (44.9)
Female	157 (70.7)	167 (55.1)
General Health*
Excellent	8 (3.6)	22 (7.3)
Very good	45 (20.3)	81 (26.7)
Good	82 (36.9)	135 (44.6)
Fair	71 (32.0)	59 (19.5)
Poor	16 (7.2)	6 (2.0)
Household Income*
<$35k	50 (22.5)	76 (25.1)
$35k–$75k	73 (32.9)	133 (43.9)
>$75k	97 (43.7)	92 (30.4)
Don’t know	2 (0.9)	2 (0.7)
Race/Ethnicity *
Asian	11 (5.0)	49 (16.2)
Black/African American	4 (1.8)	25 (8.3)
Hispanic/Latino	3 (1.4)	30 (9.9)
White/Caucasian	197 (88.7)	185 (61.1)
Other/Multiple	7 (3.2)	14 (4.6)
Relationship Status*
Married	120 (54.1)	94 (31.0)
Single	37 (16.7)	142 (46.9)
Divorced	21 (9.5)	8 (2.6)
In a relationship	38 (17.1)	55 (18.2)
Separated	6 (2.7)	4 (1.3)
Physician Avoidance (binary outcome)*
No	131 (59.0)	135 (44.6)
Yes	91 (41.0)	168 (55.4)
Recency of treatment
Not within last year	168 (75.7)	NA
Within last year	54 (24.3)	NA

^{* =}p < .05 for chi-squared test of independence (hx of cancer v no hx of cancer)

^{a =}Cancer type prevalence including individuals with multiple cancers: breast (21.6%), melanoma and nonmelanoma skin cancer (14.4%), thyroid (10.8%), cervical (9.5%), uterine/endometrial (6.8%), prostate (5.4%), colorectal (5.4%), non-Hodgkins lymphoma (5.0%), lung (3.2%), “other” (25.2%)

Table 2.

Means, standard deviations, and bivariate correlations between continuous variables

Measures	Hx of Cancer		No Hx of Cancer		1	2	3	4	5
	M	SD	M	SD
1. Neuroticism (Anxiety)	2.83	1.02	3.11	1.01	-	-	-	-	-
2. Cancer Mortality Salience	2.42	0.96	2.76	0.81	0.30*	-	-	-	-
3. Cancer Fatalism	2.68	0.92	2.79	0.87	0.15*	0.31*	-	-	-
4. Cancer Prevention	1.96	0.75	2.02	0.80	0.07	0.23*	0.20*	-	-
5. Cancer Confusion	2.79	0.85	2.87	0.82	0.11	0.20*	0.35*	0.34*	-
6. Physician Visit Concerns	1.88	0.84	1.97	0.75	0.17*	0.35*	0.28*	0.20*	0.25*

Note.

^{* =}p < .01

Neuroticism, Cancer Mortality Salience, and Physician Avoidance

Cancer survivors.

Neuroticism was positively associated with CMS, b = 0.26, SE = 0.06, p < .001, confidence interval (CI) = [0.13, 0.38]. There was also a significant association between neuroticism and physician avoidance, b = 0.40, SE = 0.16, p = .016, CI [0.08, 0.72]. As neuroticism increased for cancer survivors, so did physician avoidance after controlling for CMS and other control variables (Figure 1). There was no indirect relationship between neuroticism and physician avoidance through CMS for cancer survivors, CI [−0.04, 0.19].

Figure 1.

Path estimates for relationship between neuroticism and physician avoidance (0 = No, 1 = Yes) mediated by CMS controlling for age, cancer perceptions (fatalism, prevention, and confusion), education, sex, general health, income, race, and relationship status.

Note. Values above line reflect estimates for those with a history of cancer (n = 221) and those below reflect estimates for those without (n = 298). †= p < .10 *= p < .05. **= p < .01.

Indirect relationship for cancer survivors: mean estimate = 0.06, Boot SE = 0.06, CI [−0.04, 0.19]

Indirect relationship for no history of cancer: mean estimate = 0.07, Boot SE = 0.04, CI [−0.01, 0.17]

To examine if the relationship between neuroticism and physician avoidance could be attributed to more proximal treatment courses for cancer survivors, two additional regression analyses were conducted for cancer survivors: 1) those who were not currently or had not recently undergone treatment (n = 167), and 2) those who were currently or had recently undergone treatment (n = 54) (See Demographics). A significant relationship between neuroticism and physician avoidance was only seen among cancer survivors who had not currently or recently undergone treatment, b = 0.56, SE = 0.21, p = .006, CI [0.16, 0.97]. There was not a significant relationship for cancer survivors who were currently or had recently gone through treatment (p = .949), indicating that, for cancer survivors, the relationship between neuroticism and physician avoidance was driven by those with more distal treatment courses.

No history of cancer controls.

There was a positive relationship between neuroticism and CMS, b = 0.22, SE = 0.05, p < .001, CI [0.13, 0.32] and a trending relationship between CMS and physician avoidance, b = 0.31, SE = 0.18, p = .084, CI [−0.04, 0.66]. However, there was no significant relationship between neuroticism and physician avoidance for those without a history of cancer (p = .377), nor was there an indirect relationship, CI [−0.01, 0.17].

Neuroticism, Cancer Mortality Salience, and Physician Visit Concerns (Continuous Outcome)

Cancer survivors.

For cancer survivors there was a significant relationship between neuroticism and CMS (as before) and a significant relationship between CMS and physician visit concerns, b = 0.29, SE = 0.06, p < .001, CI [0.17, 0.41]. There was also trending direct relationship between neuroticism and physician visit concerns for cancer survivors, b = 0.10, SE = 0.06, p = .088, CI [−0.01, 0.21], and a significant indirect relationship between neuroticism and physician mediated by CMS, b = 0.07, SE = 0.02, CI [0.03, 0.13]. For cancer survivors, as neuroticism increased, CMS increased, and that was in turn associated with increased physician visit concerns.

We again conducted two additional regression analyses for cancer survivors based on recency of treatment. For those without a recent history of treatment, there was a significant indirect relationship between neuroticism and physician visit concerns which was mediated by CMS, b = 0.06, SE = 0.03, CI [0.02, 0.12], and there was also a significant direct relationship between neuroticism and physician visit concerns, b = 0.15, SE = 0.07, p = .033, CI [0.01, 0.28]. For cancer survivors who were currently or had recently gone through treatment, there was no significant direct relationship between neuroticism and physician visit concerns (p = .949), nor was there a significant indirect relationship between the two through CMS, CI [−0.01, 0.24].

No history of cancer controls.

For those without a history of cancer, there was a positive relationship between neuroticism and CMS. There was also a significant relationship between CMS and physician visit concerns, b = 0.23, SE = 0.05, p < .001, CI [0.12, 0.33], but no significant direct relationship between neuroticism and physician visit concerns (p = .248). However, there was a significant indirect relationship between neuroticism and physician visit concerns which was mediated by CMS, b = 0.05, SE = 0.02, CI [0.02, 0.08].

Discussion

In this study, cancer survivors exhibited significantly lower levels of CMS than their counterparts with no cancer history, congruent with previous research demonstrating that a history of cancer may be associated with decreased perceptions of cancer as a “death sentence”.⁷ Results also indicated that as neuroticism increased for cancer survivors, so did physician avoidance, even after controlling for CMS. However, cancer survivors higher in neuroticism exhibited increased physician avoidance only when their treatment course was more distal (i.e., they were not currently or had not recently undergone treatment). This finding may reflect a stronger motivation to attend healthcare appointments when faced with the immediate threat of cancer (e.g. during and directly following cancer treatment). However, the conclusion of primary treatment may decrease the immediacy of reminders to seek follow-up care for those high in neuroticism who may be avoidant of such health contexts.

Previous studies indicate that whereas neuroticism or high perceived cancer risk alone may motivate individuals to seek out screening, care, and health-related information, the combination of neuroticism plus high perceived cancer risk can lead to negative expectations and, ultimately, the avoidance of health-related information.^3,19,20 Indeed, anxiety and fear surrounding cancer recurrence has been shown to fluctuate, with levels of fear of recurrence often increasing as time passes after healthcare encounters.²¹ Accordingly, cancer survivors with higher levels of neuroticism and distal treatment courses may have been exhibiting increased physician avoidance due to the higher perceived cancer risk. Of note, it seems that associating cancer with death may contribute to cancer survivor avoidance of physicians through physician visit concerns. As neuroticism increased so too did CMS, which in turn, predicted stronger endorsement of physician visit concerns. Like the other findings to emerge, this effect was robust with respect to a number of relevant control variables (e.g., more general cancer perceptions). Further, the absence of an indirect relationship between neuroticism and physician avoidance mediated by CMS for cancer survivors may have been due to a lack of variability often seen in binary outcomes.

Avoidance of medical visits during cancer survivorship may produce a number of negative outcomes, including worse prognosis, treatment options, and response to treatment.²² The present results identify cancer survivors at higher risk of avoiding medical care, and highlight clinical approach needs of cancer survivors with more distal treatment courses. Identifying cancer survivors higher in neuroticism and targeting interventions to reduce anxiety around fear of cancer recurrence could attenuate susceptibility to physician avoidance and concerns about physician visits. As a personality trait, neuroticism is relatively easy to screen for in a clinical setting—for example, using a neuroticism subscale of the Big Five Inventory.¹³

Clinical Implications

Cognitive behavioral therapy techniques²³ and mind-body techniques²⁴ have been shown to be efficacious in reducing fear of cancer recurrence through teaching cancer survivors to identify, reframe, and cope with harmful patterns of thoughts, emotions, and behaviors surrounding cancer recurrence. These practices may present useful techniques for reducing the anxiety that could provoke physician avoidance, particularly in survivors high in neuroticism.

It may also be possible to interrupt the negative cycle of anxiety and physician avoidance through a broader social agenda of demystifying and destigmatizing cancer, death, and dying. At present, Western social norms dictate relatively rigid taboos around death and dying, and cancer itself is still routinely associated with death. Public health messaging may benefit from taking a cue from the death positivity movement to diminish taboos surrounding cancer, with a goal of reframing cancer from a “death sentence” with a proactive approach including regular and consistent follow-up medical care—even as time passes after treatment. Such approaches may be especially promising given the patterns observed in the present study, wherein CMS played a role in the reasons survivors with more distal treatment courses may avoid physicians and have concerns about physician visits.

Finally, cancer survivors more temporally distal to cancer treatment may not be followed as closely by the healthcare system. Left to their own devices, cancer survivors—particularly those higher in neuroticism—may be stranded in a vicious cycle of anxiety and healthcare avoidance, increasingly intractable as time passes and uncertainty grows. The National Academy of Medicine recommends that all cancer survivors have a Survivorship Care Plan (SCP)—a document comprising a treatment summary and recommendations for future follow-up care and surveillance.²⁵ The SCP has the potential to decrease physician avoidance in cancer survivors by increasing awareness of the need for follow-up care and decreasing fear of recurrence through education about the efficacy of continued screening. However, SCPs alone are likely not enough to mitigate cancer-related anxiety²⁶—or prevent physician avoidance—in cancer survivors who are more distal to treatment. Additional strategies, such as the use of lay health workers, may be warranted to 1) decrease fear of recurrence and 2) educate about the importance of life-long follow-up care for cancer survivors.

Limitations

This study used cross-sectional data and thus, it was not possible to assess cause and effect or directionality. We also used a convenience sample from an online crowdsourcing platform (Prolific), and self-report survey measures. A longitudinal analysis of how neuroticism relates to changes in CMS and physician avoidance over time for cancer survivors would have allowed us to make a stronger case for the temporal nature of these effects. Our sample was also predominantly White/Caucasian, limiting generalizability. Given the sample we recruited, it was difficult to ascertain the number of cancer survivors with temporally proximal vs. distal treatment courses, which resulted in relatively small sample sizes for those in the proximal treatment group. Further, the type of treatment cancer survivors underwent was not assessed in nor was the time since treatment for those with more distal treatment courses. Survivors who were treated further in the past may have undergone more harsh or aversive treatment regimens that have since been replaced by newer treatment advances. Such measurement could have provided further explanatory power to target when to intervene to decrease physician avoidance. However, the statistical robustness of the effects for cancer survivors with more distal treatment courses were substantial.

Conclusion

This work represents a step toward better understanding psychological and affective correlates of physician avoidance and physician visit concerns in cancer survivors, suggesting that survivors high in trait neuroticism may be particularly at risk for avoiding healthcare as their treatment becomes more temporally distal. Further research is necessary to probe the relationship between personality traits and healthcare avoidance in individuals with a history of cancer, to distinguish the most salient targets for public health messaging and to increase uptake of follow-up care recommendations within this population.

Figure 2.

Path estimates for relationship between neuroticism and physician visit concerns (3-item composite) mediated by CMS controlling for age, cancer perceptions (fatalism, prevention, and confusion), education, sex, general health, income, race, and relationship status.

Note. Values above line reflect estimates for those with a history of cancer (n = 220) and those below reflect estimates for those without (n = 298). †= p < .10 *= p < .05. **= p < .01.

Indirect relationship for cancer survivors: mean estimate = 0.07, Boot SE = 0.02, CI [0.03, 0.13]

Indirect relationship for no history of cancer: mean estimate = 0.05, Boot SE = 0.02, CI [0.02, 0.08]

Data Availability Statement

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.