The bodily threat monitoring scale: Development and preliminary validation in adult and childhood cancer survivors

Bodily threat monitoring is a core clinical feature of Fear of cancer recurrence (FCR) and is targeted in psycho‐oncology treatments, yet no comprehensive self‐report measure exists. The aim of this study was the theory‐informed development and initial validation of the Bodily Threat Monitoring Scale (BTMS).


| INTRODUCTION
Fear of cancer recurrence (FCR), defined as the "fear, worry, or concern that cancer may come back or progress", is a prevalent concern among adults and young people living with and beyond cancer (i.e., cancer survivors). 1 Over half of cancer survivors report moderate or greater FCR 2,3 which is associated with more healthcare utilization 4 and poorer quality of life. 5Managing FCR is also the most reported unmet supportive care need in cancer survivors. 6To address this unmet clinical need, several theoretical models of FCR have been proposed over the last decade. 7These theories have led to significant advances in our understanding of FCR and have guided the development of psychological treatments, several of which have demonstrated efficacy in reducing FCR and associated healthcare costs. 8,9One prominent psychological process that appears across theories [10][11][12] is the tendency to monitor and (mis)interpret uncertain bodily symptoms-hereafter termed 'bodily threat monitoring'.
Bodily threat monitoring is an understandable response to a cancer diagnosis.As cancer treatment and survivorship are contexts in which there is risk for disease progression, disease recurrence, and late effects of treatment, symptom self-monitoring is often clinically encouraged as part of a 'better safe than sorry' strategy. 13t, while some degree of bodily threat monitoring may be helpful to respond appropriately to emerging symptoms of concern, persistent and excessive monitoring of bodily sensations and misinterpretation of benign bodily sensations as indicating disease could drive excessive reassurance-seeking, negatively impact mental health, and reduce quality of life. 11,14A recent Delphi study reported that bodily threat monitoring is a core characteristic that differentiates clinical from nonclinical levels of FCR. 15 Accordingly, existing FCR treatments attempt to reduce excessive bodily threat monitoring. 8,16t, there is no comprehensive self-report measure of bodily threat monitoring.This has precluded high-quality empirical research adequately testing FCR theories and thus establishing the role of bodily threat monitoring in the development, maintenance, or remittance of FCR.According to FCR theory, greater bodily threat monitoring should not only be associated with FCR but also with the broader construct of (intolerance of) uncertainty. 11Bodily threat monitoring may emerge as a motivated uncertainty-reducing behaviour, particularly in a context wherein uncertainty lies within the body itself.Critically, without a well-validated measure, bodily threat monitoring has not been adequately assessed as an outcome or mediator of change in FCR treatment studies.We thus do not know whether existing FCR interventions adequately reduce excessive bodily threat monitoring or whether novel treatment approaches are needed to target this core clinical correlate of FCR.Examining whether bodily threat monitoring is associated with healthcareseeking behaviours and health-related quality of life will also be important to establish its utility as a treatment target.Bodily threat monitoring is a complex phenomenon.Clinically, the phenomenon appears to comprise both the tendency to persistently attend to the body to detect symptoms of concern as well as the tendency to appraise bodily sensations as symptoms of bodily threat (e.g., as indicating cancer recurrence).Studies to date have typically either conflated these two cognitive processes or have assessed the general cognitive style without the motivational context of detecting bodily threat.For example, the Body Vigilance Scale (BVS) 17 which includes items such as "I am the kind of person who pays close attention to the way my body feels" captures attentional focus on the body but does not capture the motivation of this attentional focus.These items could be maladaptive or adaptive depending on whether the motivation for attending is to detect something being wrong with the body, or an accepting awareness and being 'in tune' with the body for the purpose of emotion regulation.Other studies have captured a more threat-focused style of relating to bodily sensations through measures such as the Pain Catastrophizing Scale, 18 however these are limited in their focus on pain.Sensations such as fatigue may be more relevant signals of threat for many cancer survivors.Lastly, the Symptom Worry Scale 19 captures the degree of worry about individual symptoms (e.g., pain, fatigue, itch) as a sign of cancer recurrence, however total scores on this measure are difficult to interpret as they may indicate a high level of worry about an individual symptom or a low level of worry about diffuse symptoms.Thus, we herein define the concept of bodily threat monitoring as comprising both the attending to and appraisal of bodily sensations for the motivational goal of detecting something being wrong with the body.A bodily threat monitoring measure that comprises subscales for both monitoring and appraising could usefully inform treatment approaches, for example, whether strategies to increase attentional control, reduce habitual threat interpretations, or both would be most helpful for FCR management.
The aim of this study is the development and preliminary psychometric assessment of the Bodily Threat Monitoring Scale (BTMS) in two samples of adult and childhood cancer survivors.We aim to examine the internal consistency, preliminary factor structure, and test-retest reliability of the BTMS.We also aim to examine its construct and criterion validity; we hypothesise that greater bodily threat monitoring will be associated, but not entirely overlapping, with greater body vigilance and anxiety sensitivity in adults (construct validity) as well as greater FCR in adults and children, intolerance of uncertainty, help-seeking behaviours, and lower quality of life in children (criterion validity).

| STUDY 1 2.1 | Overview
In Study 1 we examine the BTMS internal consistency, preliminary factor structure, and association with FCR in a sample of women living with and beyond breast or ovarian cancer.The definition of FCR includes both fear of disease recurrence and progression, and FCR is common and impactful in patients receiving active treatment and those who have completed treatment.Moreover, bodily threat monitoring is not limited to the post-treatment phase, as patients 1886 -HEATHCOTE ET AL. may be motivated to detect symptoms of disease progression and spreading within the context of active treatment.Thus, in this first study we employed the BTMS in a mixed sample of patients both onand off-treatment.

| Initial development of the Bodily Threat Monitoring Scale
An initial pool of 36 items was developed by compiling and amending items from existing questionnaires and through feedback from an expert panel (see supplemental file S3 for full details).

| Participants
Females (N = 354) with a previous or current diagnosis of breast or ovarian cancer accessed an online survey circulated by Ovarian Cancer Australia and Breast Cancer Network Australia between June and September 2020.Eligible participants were female, fluent in English, aged over 18 years, and diagnosed with breast or ovarian cancer of any stage.Three-hundred and six respondents completed the BTMS and were included in analyses (breast N = 147; ovarian N = 159).Data from subsamples have been reported elsewhere. 20,21e study received approval from the University of Sydney Human Research Ethics Committee.Participants provided informed consent online.See Table 1 for full demographics.

| Fear of cancer recurrence inventory-Short form
The Fear of cancer recurrence inventory-Short form (FCRI-SF) is a 9-item questionnaire that assesses the severity of FCR. 22Items are rated on a 5-point scale; a cut-off score of 13 or higher indicates clinically significant levels of FCR. 23,24Cronbach's alpha for this sample was α = 0.88.

| Data analysis
Data were analysed using SPSS v28.0. 25Exploratory maximum likelihood factor analyses (EFA) with oblimin rotation were conducted to examine the BTMS factor structure.Pearson correlations (age), Spearman correlations (cancer stage), and independent-samples t-tests (cancer type, history of recurrence, clinical levels of FCR) were used to inspect demographic and medical correlates of bodily threat monitoring.Pearson correlations were used to assess associations with FCR (<0.3 weak; 0.3-0.5 moderate, >0.5 strong 26 ).

| Bodily Threat Monitoring Scale development & factor analyses
There were no items for which >90% of the sample had responses at the extreme ends of the scale (0-1, or 3-4), indicating no issues with ceiling or floor effects.Seven items with skew or kurtosis >2.0 were removed to retain items that approximate a normal distribution.
Seventeen items considered as overlapping as identified by a datadriven method (correlations >0.8) and considered to have considerable conceptual overlap were removed.The remaining 19 items were submitted to an unrestricted EFA.Three factors were indicated (Eigenvalues >1); however, the third factor had no unique items.A forced two-factor solution (see Table 2) was a good fit to the data, explaining 65.5% of the variance and with factor loadings >0.5.The two factors corresponded to dimensions of bodily monitoring ( 6items, e.g., "I pay attention to my body to check if something is wrong") and bodily threat appraisals (13 items, e.g., "When I have a bodily sensation that wasn't there before, I immediately think that something is wrong with my body").The two subscales were strongly associated with each other (r = 0.63, p < 0.001).For parsimony, a one-factor solution was also examined and showed a good fit to the data, explaining 54.9% of the data, with no items loading >0.4 on their factor (see supplemental file S2).The final 19-item BTMS total score and subscales showed excellent internal consistency (see Table 3).The BTMS is freely available to use by research and clinical teams (see S3).

| Associations with demographic and medical factors
As seen in Table 3, there were no significant differences in the bodily threat monitoring total score or its subscales according to partici-

| Associations with fear of cancer recurrence
As seen in  more severe fear of recurrence. 29Cronbach's alpha for this sample was 0.90.

| Childhood anxiety sensitivity index (CASI- Autonomic/physical concerns) (U.S. Site only)
The CASI-autonomic/physical concerns subscale comprises 9 items assessing fear of sensations associated with the experience of anxiety and a misinterpretation of anxiety-related sensations as dangerous. 30tal scores range from 9 to 27 and higher scores indicating greater anxiety sensitivity.Cronbach's alpha for this sample was 0.88.

| Body Vigilance Scale
The BVS comprises 3 items assessing attentional focus on and perceived sensitivity to bodily changes. 31The BVS comprises total scores from 0 to 30 with higher scores indicating greater body vigilance.Cronbach's alpha for this sample was 0.79.

| Intolerance of uncertainty scale (IUS-Revised)
The IUS-R child version 32 comprises 12 items for assessing intolerance of uncertainty in adults.Items are summed with total scores ranging from 12 to 60, with higher scores indicating greater intolerance of uncertainty.Cronbach's alpha for this sample was 0.92.

| Childhood illness attitudes scale (CIAS-Help seeking)
The help-seeking subscale of the CIAS 33 comprises 9 items assessing help-seeking behaviours, including asking to go to the doctor and informing parents about feeling unwell.Scores range from 9 to 27, with higher scores indicating more help-seeking behaviours.Cronbach's alpha for this sample was 0.83.

| Health-related quality of life (PROMIS paediatric global health)
The paediatric global health (PGH-7) is a 7-item PROMIS paediatric measure of health-related quality of life. 34Raw scores are rescaled into standardized scores (t-scores) such that a score of 50 represents the average (mean) for the US paediatric population (SD = 10 points).
Cronbach's alpha for this sample was 0.77.

| Data analysis
Data were analysed using SPSS v28.0. 25Pearson correlations and paired-samples t-tests were used to examine test-retest reliability.
Pearson correlations and Spearman correlations (treatment intensity) were used to inspect demographic and medical correlates of bodily threat monitoring.Partial correlations were used to assess unique associations with FCR.

| Psychometric properties of the 19-item Bodily Threat Monitoring Scale in a paediatric sample
The BTMS total score and subscales yielded excellent internal consistency in childhood cancer survivors (see

| Associations with demographic and medical factors
As can be seen in Table 3, older participants, those diagnosed at an older age, and females engaged in significantly more bodily threat monitoring.The significant association with age at diagnosis and sex was strongest for the bodily monitoring subscale but did not hold for the bodily threat appraisals subscale.There were no significant differences in bodily threat monitoring according to participants' time since treatment completion or treatment intensity.

| Associations with Fear of cancer recurrence
As seen in Table 3, there was a strong association between FCR and bodily threat monitoring in childhood cancer survivors.The BTMS total score and subscales were significantly, positively correlated with anxiety sensitivity and body vigilance, indicating construct validity.Aligning with theory, body vigilance was more strongly associated with the bodily monitoring subscale, which was expected as these both capture styles of attending.Moreover, anxiety sensitivity was more strongly associated with the bodily threat appraisals subscale as anticipated as these both capture threat-focused appraisals of bodily sensations.Partial correlations revealed that the association between the BTMS total score and the FCRI-SF remained significant when controlling for anxiety sensitivity (r = 0.44, p < 0.001) and body vigilance (r = 0.43, p < 0.001), indicating a unique contribution of this theory-informed measure.

| Association with intolerance of uncertainty, help-seeking behaviours, and quality of life
As seen in Table 3, there were small to moderate positive correlations between the BTMS total score, intolerance of uncertainty, and help-seeking behaviours, and a negative association with healthrelated quality of life.Intolerance of uncertainty was more strongly associated with the bodily threat appraisals subscale, even more so than the BTMS total score.The bodily threat appraisals subscale was significantly, moderately associated with quality of life.

| DISCUSSION
The BTMS is a psychometrically promising self-report measure for survivors of adult and childhood cancers aged 10 and above.As expected, the BTMS total score and its subscales yielded moderate-to-strong correlations with FCR in both adults and young survivors; these significant associations held while controlling for existing constructs of body vigilance and anxiety sensitivity.The effect sizes findings are particularly striking given that the BTMS makes no explicit reference to cancer.Moreover, the BTMS differentiated between those with clinical and subclinical levels of FCR.
6][37] They also align with cross-sectional studies showing that survivors who report more somatic symptoms, including pain and fatigue, are also more fearful of cancer recurrence. 5,19,38,39In addition, these findings align with recent studies showing that FCR is associated with a catastrophic appraisal of pain 18 and a tendency to interpret ambiguous health-related information derived from bodily sensations as threatening. 21Fear of recurrence may bias attention towards and motivate monitoring of bodily sensations, lowering the detection thresholds for these sensations, hence fuelling increased fear of recurrence.This indicates that bodily threat monitoring may act as a mechanism maintaining a bi-directional association between FCR and somatic symptoms over time.The current findings provide support that bodily threat monitoring is a core clinical correlate of FCR and further supports its targeting in FCR treatment.
In the childhood cancer survivor sample, we further probed the criterion validity of the BTMS.Greater bodily threat monitoring was associated with elevated intolerance of uncertainty, a dispositional characteristic involving emotional, cognitive, and behavioural responses to an unknown outcome. 402][43] Yet, intolerance of uncertainty has been inadequately studied in relation to the experience of somatic symptoms.Our findings align with that of Hall and colleagues who found that greater perception of cancerrelated uncertainty was associated with more fatigue and insomnia among breast cancer survivors. 44Bodily threat monitoring may emerge as an information-seeking style to reduce uncertainty-this warrants empirical investigation.Lastly, we found that bodily threat monitoring was associated with greater help-seeking behaviours and lower health-related quality of life in childhood cancer survivors.
These data provide initial evidence that bodily threat monitoring is related to salient clinical and behavioural outcomes, such as helpseeking, FCR, and HRQoL, that align with higher healthcare costs and morbidity in cancer survivors.

| Clinical implications
The current findings are relevant for guiding survivorship care and policy.Less than one third of paediatric oncologists believe that 1892 -HEATHCOTE ET AL.
their off-treatment patients worry unnecessarily about symptoms of cancer recurrence or late effects of treatment. 13Oncologists often favour a 'better safe than sorry' approach in counselling off-therapy patients to self-monitor for symptoms of disease recurrence and late effects. 13Yet, the current findings add to the growing body of evidence indicating that not only do cancer survivors commonly monitor and worry about their bodily sensations but also that this bodily threat monitoring matters-it is associated with increased healthcare-seeking behaviours and lower health-related quality of life in children.While some degree of self-monitoring is clinically warranted, more discussion is needed within clinical care and policy around striking a balance between mitigating medical risk (e.g., early detection of recurrence) while not unintentionally increasing fear and decreasing quality of life.There may be opportunities within existing clinical care pathways to actively stress the limited utility and mental health consequences of symptom self-monitoring when patients are receiving regular surveillance testing.This may be particularly significant for individuals living with the chronic high threat of cancer associated with cancer predisposition syndromes (e.g., hereditary breast and ovarian cancer).Offering anticipatory guidance as to which symptoms should not be a cause for concern, as well as accessible pathways for patients to report symptoms of concern early to the oncology team, could also be helpful to prevent an escalating cycle of bodily threat monitoring and fear of recurrence. 14

| CONCLUSIONS
In conclusion, the BTMS shows evidence of sound psychometric properties in two samples of adult and childhood cancer survivors.
Given that bodily threat monitoring is proposed as a core clinical correlate of FCR and is already targeted in FCR interventions, the BTMS would be useful to include in future studies examining the maintenance and management of FCR.

Following
initial pilot work, a large item pool was administered to two samples of adult and childhood cancer survivors across three countries (Australia, United States, Canada).Aligning with theory, exploratory factor analyses in the larger adult sample indicated a two-factor structure; subscales represented 'bodily monitoring' and 'bodily threat appraisals'.The BTMS total score and subscales demonstrated excellent internal consistency across both samples and showed adequate test-retest reliability across a brief 2 week interval in the childhood cancer sample.
Demographic and clinical characteristic the Study 1 sample (N = 306).
T A B L E 1

Table 3
Bodily Threat Monitoring Scale (BTMS) factor loadings obtained by exploratory factor analysis (two-factor solution).Psychometric properties of the Bodily Threat Monitoring Scale (BTMS) across studies.

Table 3
p = 0.38) were not different across the two recruitment sites.