ORIGINAL ARTICLE

A Shortened Version of the Suicide Cognitions

Scale for Identifying Chronic Pain Patients at
Risk for Suicide

Craig J. Bryan, PsyD, ABPP*,†; Kathryn E. Kanzler, PsyD, ABPP‡; Emily Grieser,
PhD§; Annette Martinez, MA‡; Sybil Allison, MA‡; Donald McGeary, PhD, ABPP‡
*National Center for Veterans Studies; †Department of Psychology, The University of Utah, Salt
Lake City, Utah; ‡Department of Psychiatry, The University of Texas Health Science Center at
San Antonio; §Clinical Health Psychology, Wilford Hall Ambulatory Surgical Center, San
Antonio, Texas, U.S.A.

& Abstract analyses. Correlation and multivariate analyses supported the

construct and incremental validity of the SCS-S.
Objective: Research in psychiatric outpatient and inpatient
Conclusions: Results support the reliability and validity of
populations supports the utility of the Suicide Cognitions
the SCS-S among chronic pain patients, and suggest the scale
Scale (SCS) as an indicator of current and future risk for
may be a useful method for identifying high-risk patients in
suicidal thoughts and behaviors. Designed to assess suicide-
medical settings. &
specific thoughts and beliefs, the SCS has yet to be evaluated
among chronic pain patients, a group with elevated risk for
Key Words: suicide, chronic pain, screening, military
suicide. The purpose of the present study was to develop and
test a shortened version of the SCS (the SCS-S).
Study Design: A total of 228 chronic pain patients com-
pleted a battery of self-report surveys before or after a INTRODUCTION
scheduled appointment.
Setting: Three outpatient medical clinics (pain medicine, The notion of chronic vulnerability for suicidal thoughts
orofacial pain, and clinical health psychology). and behaviors as a distinct dimension of suicide risk is a
Methods: Confirmatory factor analysis, multivariate regres- core tenet of the fluid vulnerability theory of suicide.1,2
sion, and graded item response theory model analyses. Specifically, the fluid vulnerability theory conceptualizes
Results: Results of the CFAs suggested that a 3-factor suicide risk on 2 dimensions: baseline and acute.
solution was optimal. A shortened 9-item scale was identified
Baseline risk entails the persistent dimension of risk
based on the results of graded item response theory model
that remains reasonably constant over time, whereas
acute risk entails the fluctuations in risk that often occur
Address correspondence and reprint requests to: Craig J. Bryan, PsyD, in response to life stressors as well as moment-to-
ABPP, National Center for Veterans Studies, The University of Utah, 260 S.
moment change in various transient or state-dependent
Central Campus Dr., Room 205, Salt Lake City, UT 84112, U.S.A. E-mail:
Craig.bryan@utah.edu risk factors (eg, mood).2 The temporal process of suicide
Submitted: January 3, 2016; Revised: February 18, 2016; risk is therefore best conceptualized as the ebb and flow
Revision accepted: March 14, 2016
DOI. 10.1111/papr.12464 of risk factors around a relatively fixed “set point.”
These fluid dynamics provide a unique challenge for
© 2016 World Institute of Pain, 1530-7085/16/$15.00
identifying and intervening with patients: if assessed
Pain Practice, Volume , Issue , 2016 – during a period of relative “calm,” a high-risk individual
2  BRYAN ET AL.
might be missed. Indeed, research suggests that the
majority of individuals who die by suicide screened
negative for suicidal thoughts or intent prior to their
deaths across a wide range of medical settings.3–6 Newer
methods that assess the baseline (or chronic) dimension
of risk instead of the acute (or dynamic) dimension of
risk may therefore prove to be a useful strategy for
identifying higher risk patients who might otherwise
be missed by traditional screening and assessment
methods.
According to the fluid vulnerability theory, an
important component of baseline risk is the individual’s
sense of self. In particular, negative and self-deprecatory
identity-based perceptions such as perceived burden-
someness, self-hatred, and trait hopelessness serve as
long-term vulnerabilities to suicide risk. Individuals who
endorse such perceptions are therefore hypothesized to
be at increased risk for future suicidal crises and
behaviors, even if they deny recent emotional distress
and/or other time-limited risk factors for suicide (eg,
situational hopelessness, relationship problems). This
hypothesis has gained preliminary support in two
samples of military personnel receiving outpatient men-
tal health treatment,7 in which researchers found that
the 18-item scale designed to assess self-perceptions
consistent with the chronic dimension of suicide risk, the
Suicide Cognitions Scale (SCS), predicted concurrent
and future suicidal thoughts and behaviors beyond the
effects of other risk factors. Results suggested the SCS
had two latent factors, which were named Unlovability
(ie, perceptions that one is a burden on others, is
unworthy of love and respect, and deserves to be
punished) and Unbearability (ie, perceptions that one
is incapable of tolerating one’s emotional pain). Con-
sistent with the hypotheses of the fluid vulnerability
theory, both SCS subscales differentiated between
patients with a history of suicide attempts and those
with a history of suicidal thoughts only, and predicted
future suicide attempts better than current suicide
ideation. Subsequent research8 among psychiatric inpa-
tients suggests the SCS may actually have 3 latent factors
as opposed to 2: Unlovability, Unbearability, and
Unsolvability (ie, perceptions that one is hopelessly
incapable of solving one’s problems, and that suicide is
the only solution). Each of the 3 factors was significantly
better predictors of suicide ideation than depression and
hopelessness, with the Unsolvability and Unlovability
factors being the relative best predictors. In sum, the
self-deprecatory perceptions assessed by the SCS appear
to be better indicators of current and future suicide risk
than traditional constructs that are hypothesized to be
more state-dependent (eg, depression, hopelessness).
Despite these promising findings, additional research
on the SCS across a diversity of populations and settings
is needed. In particular, studies that examine the feasi-
bility of reducing the scale’s length are needed in order to
enhance the practicality of using the SCS within clinical
and medical settings. Examination of the SCS among
chronic pain patients provides a unique opportunity to
evaluate the scale’s properties for several reasons. First,
chronic pain is an established risk factor for suicide
ideation, attempts, and death.9–14 Among individuals
with chronic pain, estimated prevalence rates across
studies range from 17% to 66%.14 Although the mech-
anisms by which some chronic pain patients develop
suicide ideation and subsequently transition to suicide
attempts remain unclear, comorbid depression has been
identified as an especially important risk factor.12
Second, recent research has pointed to the incremental
validity of certain self-perceptions and cognitions, espe-
cially perceived burdensomeness, as indicators of suicide
ideation among chronic pain patients.15 These findings
converge with a considerable body of research in nonpain
populations.16–20 According to Van Orden et al.,21
perceived burdensomeness entails self-hatred (ie, shame,
low self-esteem) combined with the perception that one’s
death is worth more to others than one’s life (ie, liability).
Perceived burdensomeness may function in unique ways
among individuals with chronic pain, who may have
physical and/or functional limitations that require them
to depend on family members and caretakers to a greater
degree than individuals without such limitations.22,23
Over time, this self-perceived burden that is relatively
common among chronic pain patients could lead to the
emergence of emotional distress, resignation and defeat,
negative self-regard, and the erosion of relationships with
supportive others.23 Chronic pain patients are therefore
characterized by a range of cognitions, beliefs, and
behaviors that can increase vulnerability to suicidal
thoughts and behaviors.
The primary aims of the present study were to
develop a shortened version of the SCS (SCS-S) and to
evaluate its psychometric properties in a clinical sample
of chronic pain patients. We specifically sought to
answer the following questions:
1. What is the best fitting factor solution for the
SCS?
2. Which items should be included in a shortened
SCS scale?

3. How does a shortened SCS scale perform with
respect to reliability (ie, internal consistency) and
validity (ie, convergent, divergent, and incremental)?
METHODS
Participants and Procedures
Participants included 228 patients (56.1% male, 39.5%
female, 4.4% unknown) receiving outpatient treatment
for pain-related conditions at 3 clinics (pain medicine,
orofacial pain, and clinical health psychology) located at
a large military medical center in the southern United
States. The majority (64.5%) of participants were active
duty military (92.1% Army, 7.2% Air Force, and 0.7%
Navy); the remainder were military retirees and/or
family members. Current marital status was 68%
married, 6.6% dating or engaged, 12.7% separated or
divorced, and 0.9% widowed. Approximately half of
participants’ pain conditions were secondary to an
injury (24.1% work-related, 19.2% not work-related);
the other half were related to a medical condition or
other cause. The length of time since pain onset was less
than 5 years for 60.3% of participants.
Participants were invited to complete a self-report
survey before or after routine clinic visits. Surveys were
administered anonymously via a computer kiosk located
in clinic waiting rooms, or via paper survey packet, if
preferred. Inclusion criteria included being a Depart-
ment of Defense beneficiary (ie, active duty, retiree, or
dependent) at least 18 years of age who was receiving
treatment for pain of any etiology with a duration of at
least 4 months. In order to maximize the representa-
tiveness and generalizability of the sample, there were
no exclusion criteria. The study was reviewed and
approved by the Wilford Hall Medical Center’s Institu-
tional Review Board.
Materials
Suicide Cognitions Scale. The SCS7 is an 18-item self-
report scale designed to assess cognitions and beliefs that
are commonly expressed by suicidal individuals (eg,
“The world would be better off without me”; “I can’t
tolerate being this upset any longer”). Items are scored
on a 5-point scale ranging from 1 (disagree strongly) to 5
(agree strongly). Previous research has supported a 2-
factor7 as well as a 3-factor8 latent structure. The scale’s
reliability, concurrent validity, and predictive validity
Suicide Screener for Chronic Pain  3
have been supported in outpatient as well as inpatient
psychiatric samples.7,8
Beck Scale for Suicide Ideation (BSSI). The BSSI24 is a
19-item self-report scale designed to assess the intensity
of thoughts, attitudes, and behaviors associated with the
desire for suicide. Items are scored on a 3-point scale,
with higher scores indicating more severe suicide
ideation. The BSSI can be administered with respect to
2 different time frames: current (ie, past week) and worst
point (ie, most severe episode during the specified time
frame of interest). In the present study, both methods
were administered in order to assess the severity of
current as well as past suicide risk. For the worst-point
assessment, participants were directed to consider the
period of time since the onset of their chronic pain
during which they possessed the most intense or severe
desire to attempt suicide, and to complete the scale
according to that worst point. Research has shown that
both methods of assessment are reliable and valid,
although the worst-point assessment has been shown to
be a relatively stronger predictor of current and future
risk for suicidal behavior.25 In addition to its 19 core
items, the BSSI also includes 2 additional items that ask
about past suicide attempts (none, one, or multiple
attempts) as well as the level of suicidal intent during the
most recent attempt (none, mild, or moderate to severe).
Beck Depression Inventory, Second Edition (BDI-
II). The BDI-II26 is a 21-item self-report scale that
assesses depression symptom severity. Items are summed
to provide a metric of overall depression severity. The
BDI-II is a widely used measure of depression with well-
established reliability and validity.
Post-Traumatic Stress Disorder Checklist (PCL). The
PCL27 is a 17-item self-report scale that assesses post-
traumatic stress disorder (PTSD) symptom severity.
Items correspond to the Diagnostic and Statistical
Manual of Mental Disorders (4th edition, text revi-
sion)28 diagnostic criteria for PTSD, and are summed to
provide a metric of overall symptom severity. The PCL is
a widely used measure of PTSD symptoms with well-
established reliability and validity.
Pain Catastrophizing Scale (PCS). The PCS29 is a 13-
item self-report scale that assesses rumination about
pain, magnification of pain-related problems, and per-
ceived helplessness over one’s pain. The scale is a reliable
and valid predictor of pain-related outcomes, including
4  BRYAN ET AL.
pain severity and recovery following surgical or medical
procedures.
Chronic Pain Acceptance Questionnaire (CPAQ). The
CPAQ30 is a 34-item self-report scale that assesses
acceptance of one’s pain along 2 dimensions: activity
engagement (ie, pursuit of life activities regardless of
pain) and pain willingness (ie, recognition that avoid-
ance and control are unhelpful strategies for adapting to
pain). Scale scores are associated with a range of pain-
related outcomes, including functional limitations, dis-
ability, depression, and anxiety.
West Haven–Yale Multidimensional Pain Inventory,
Version 2 (MPI). The MPI31 is a 52-item self-report scale
that assesses 12 different dimensions of pain experience.
For the present study, only the pain severity scale and the
3 scales related to caretaker/significant other relation-
ships were included: punishing responses, which assess
perceptions that the caretaker is angry or frustrated with
the respondent; solicitous responses, which assess per-
ceptions that the caretaker provides support; and dis-
tracting responses, which assess perceptions that the
caretaker engages in activities to help redirect the
respondent’s focus away from his or her pain.
Oswestry Disability Index (ODI). The ODI32 is a 10-
item self-report scale that assesses perceived disability
and functional impairment across a range of daily
activities, with higher scores indicating greater perceived
disability. The ODI has been validated across a wide
range of populations as a reliable and useful indicator of
overall disability and represents the gold standard of
self-report disability measures.
Data Analytic Approach
Confirmatory factor analysis (CFA) with robust
weighted least-squares estimation was used in the Mplus
6.12 software (Muthen & Muthen, 1998–2001) to
examine the latent structure of the SCS. Three separate
models were compared: a 1-factor solution in which all
items loaded onto a global factor; a 2-factor solution
identified by Bryan et al.7; and a 3-factor solution
identified by Ellis & Rufino.8 Good model fit was
determined using the following criteria: nonsignificant
chi-square, root mean square error of approximation
(RMSEA) < 0.05, comparative fit index (CFI) > 0.95,
Tucker-Lewis Index (TLI) > 0.95, and weighted root
mean square residual (WRMR) < 1.00.
In order to develop a shortened version of the SCS,
the graded response item response theory (IRT) model33
was used to examine item characteristics for the full
SCS. Items with the highest information and demon-
strating good discriminative properties were selected as
candidates for a shortened scale. The psychometric
properties and validity of the SCS-S scale were subse-
quently evaluated using Cronbach’s alphas, Pearson
correlation coefficients, and multivariate regression
modeling. Due to skewed distributions in relevant
outcome variables (worst-point suicide ideation, recent
suicide ideation), generalized linear regression with a
robust estimator was used. Overall, 30.4% of cases had
missing data for at least 1 variable. Little’s test indicated
that data were missing completely at random
(v2(340) = 375.35, P = 0.091). Missingness was there-
fore handled with full information maximum likelihood
estimation.
RESULTS
What is the Best Fitting Factor Solution for the SCS?
Results of the CFAs indicated that all 3 factor structures
had reasonably good fit (Table 1), but the 3-factor
solution showed especially good fit to the data:
v2(132) = 187.05, P = 0.001; RMSEA = 0.05 (90%
confidence interval [CI]: 0.03, 0.07), CFI = 1.00;
TLI = 1.00; WRMR = 0.79. In addition, the 3-factor
solution showed a statistically significant better fit than
the 1-factor solution (v2diff(3) = 43.67, P < 0.001) and
the 2-factor solution (v2diff(1) = 28.31, P < 0.001).
Item-factor loadings for the 3-factor solution are
displayed in Table 2 along with item descriptive statis-
tics, and all items had very high factor loadings (> 0.79).
Which Items Should Be Included in a Shortened SCS
Scale?
Consistent with the graded response IRT model, item
information curves were derived to identify those items
Table 1. Fit Statistics for Three Confirmatory Factor
Analyses of the Suicide Cognitions Scale
Model v2 df RMSEA (90% CI) CFI TLI WRMR
1 factor 230.72 135 0.07 (0.05, 0.08) 0.99 0.99 0.95
2 factors 215.36 134 0.06 (0.05, 0.08) 1.00 1.00 0.90
3 factors 187.05 132 0.05 (0.03, 0.07) 1.00 1.00 0.79
RMSEA, root mean square error of approximation; CFI, comparative fit index; TLI,
Tucker-Lewis Index; WRMR, weighted root mean square residual.
 Suicide Screener for Chronic Pain  5

Table 2. Item-Factor Loadings and Item Descriptive Statistics for the Suicide Cognitions Scale

% Endorsing Each Response Option

Item-Factor
Item Loading Item M (SD) 1 2 3 4 5

Unlovability
1 The world would be better off without me. 0.942 1.33 (0.82) 82.6 6.8 5.6 3.1 1.8
4 I’ve never been successful at anything. 0.796 1.12 (0.47) 82.5 10.0 3.8 3.1 0.6
6 I can never be forgiven for the mistakes I’ve made. 0.934 1.99 (1.32) 80.5 9.4 8.2 1.3 0.6
9 I am completely unworthy of love. 0.958 1.29 (0.74) 84.4 8.1 4.4 3.1 0.0
14 There is nothing redeeming about me. 0.972 1.53 (0.99) 82.5 7.5 5.6 2.5 1.9
18 No one is as loathsome as me. 0.955 1.32 (0.73) 90.0 6.2 2.5 0.6 0.6
Unsolvability
2 Suicide is the only way to solve my problems. 0.964 1.49 (0.97) 92.5 4.4 1.9 1.3 0.0
7 No one can help me solve my problems. 0.958 1.52 (1.04) 73.6 11.9 9.4 1.9 3.1
10 Nothing can help solve my problems. 0.974 1.26 (0.69) 76.7 11.9 5.7 3.8 1.9
15 Suicide is the only way to end this pain. 0.967 1.42 (0.90) 93.1 3.1 1.9 1.3 0.6
16 I don’t deserve to live another moment. 0.991 1.65 (1.19) 92.5 3.8 2.5 1.3 0.0
17 I would rather die now than feel this unbearable pain. 0.885 1.33 (0.74) 92.5 1.3 3.1 2.5 0.6
Unbearability
3 I can’t stand this pain anymore. 0.864 1.61 (1.06) 57.9 9.4 14.5 12.6 5.7
5 I can’t tolerate being this upset any longer. 0.898 1.34 (0.85) 73.6 8.8 10.1 6.3 1.3
8 It is unbearable when I get this upset. 0.905 1.13 (0.55) 76.1 6.9 8.8 5.7 2.5
11 It is impossible to describe how badly I feel. 0.939 1.13 (0.49) 71.7 8.2 9.4 5.0 5.7
12 I can’t cope with my problems any longer. 0.973 1.18 (0.66) 81.1 7.5 8.8 2.5 0.0
13 I can’t imagine anyone being able to withstand this kind of pain. 0.928 1.16 (0.54) 68.8 13.1 8.7 6.9 2.5

from each factor that accounted for the greatest amount
of variance. Item information curves are displayed in
Figure 1. Three items were selected for each subscale
since this is the number of items that must load
significantly on each factor in a multidimensional scale
in order to be successfully identified.34 As can be seen,
the items that yielded the most information were (in
descending order):
 Unlovability: “There is nothing redeeming about
me” (item 14); “I am completely unworthy of
love” (item 9); and “No one is as loathsome as
me” (item 18).
 Unsolvability: “I don’t deserve to live another
moment” (item 16); “Nothing can help me solve
my problems” (item 10); and “Suicide is the only
way to end this pain” (item 15).
 Unbearability: “I can’t cope with my problems
any longer” (item 12); “It is impossible to describe
how badly I feel” (item 11); and “I can’t imagine
anyone being able to withstand this kind of pain”
(item 13).
These items were used to create the shortened version
of each scale, although for the Unsolvability scale, item
15 (“Suicide is the only way to end this pain”) was
replaced with item 7 (“No one can help me solve my
problems”) in order to eliminate the use of the word
suicide from the shortened scales while providing a
comparable amount of information. We eliminated
items with the word suicide to reduce the likelihood of
motivated responding specific to face-valid suicide risk
instruments (eg, deliberate underreporting) and to
eliminate explicit overlap of the SCS-S with suicide-
specific measures. Overall, the item information curves
for these selected items were predominantly located on
the positive side of the latent variables (ie, to the right of
0), suggesting that these items are most useful for
distinguishing between those with slightly higher than
average levels of unlovable, unsolvable, and unbearable
beliefs and those with high levels of these beliefs.
Statistical tests on IRT model parameters revealed
statistically significant slopes (P < 0.001) for all steps
on each items. Table 2 shows the proportion of
participants endorsing each response option. Response
options 4 (agree) and 5 (strongly agree) were endorsed
infrequently for all items. Item characteristic curves
were also calculated for the 9 items identified in the
previous step and are displayed in Figure 2. The items
showed excellent discriminative capacity, as indicated
by the fairly discrete item characteristic curves (ie,
nonoverlapping peaks). Taken together, these findings
suggest that each item retained for the SCS-S con-
tributes directly and meaningfully to their respective
total factor scores.
How does a Shortened SCS Scale Perform with Respect
to Reliability and Validity?
In terms of internal consistency, Cronbach’s alphas for
the SCS and the SCS-S subscales were, respectively, as
6  BRYAN ET AL.

A Unloveability significant positive correlations with recent suicide

4 ideation and worst-point suicide ideation. The SCS-S
3.5
Item 1
Unbearability subscale was also significantly correlated
3
Information

2.5 Item 4
with depression severity. None of the SCS or SCS-S
2
Item 6 subscales were significantly correlated with PTSD
1.5
1 Item 9 symptoms or any other pain-related variable, however,
0.5
0
Item 14 suggesting the subscales were independent of pain-
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4 Item 18 related constructs and other clinical indicators of
Standard Deviations From Mean distress.
Incremental validity was assessed using multivariate
B Unsolvability regression analyses to determine the association of each
7
SCS-S subscale with worst-point suicide ideation,
6
5 Item 2 beyond the effects of gender, depression, PTSD symp-
Information

4 Item 7 tom severity, and history of suicide attempt. When

3 Item 10 adjusting for these covariates, worst-point suicide
2 Item 15
1 ideation continued to be significantly associated with
Item 16
0 all 3 SCS-S subscales: Unlovability (b = 0.16,
Item 17
P = 0.008), Unsolvability (b = 0.15, P = 0.002), and
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5 4
Standard Deviations From Mean
Unbearability (b = 0.18, P = 0.023). These results sug-
C Unbearability gest that the SCS-S subscales are positively associated
4
with severity of past suicidal crises beyond the effects of
3.5 past suicidal behavior and other indicators of clinical
3 Item 3
distress. Results were unchanged when adjusting for
Information

2.5
2 pain severity.
1.5
1 Item 11
0.5
0
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
Standard Deviations from Mean
Figure 1. Item information curves for the (A) Unlovability, (B)
Unsolvability, and (C) Unbearability factors of the Suicide Cog-
nitions Scale. Items designated with black lines were retained for
the shortened SCS.
follows: Unlovability, a = 0.93 and a = 0.90; Unsolv-
ability, a = 0.91 and a = 0.85; Unbearability, a = 0.92
and a = 0.89. The SCS-S subscales were very strongly
correlated with the SCS subscales: Unlovability,
r = 0.97; Unsolvability, r = 0.96; Unbearability,
r = 0.96. Taken together, these results suggest the
SCS-S subscales do not show dramatic decreases in
internal consistency relative to the full scales, and
account for more than 92% of the variance in the SCS
subscale scores.
With respect to convergent and divergent validity,
the pattern of correlations for the SCS and SCS-S
subscales were comparable to each other, suggesting
the SCS-S was associated with clinical variables to a
similar degree as the full SCS (Table 3). As expected,
SCS and SCS-S subscale scores had statistically
Item 5
Item 8
These analyses were repeated for recent suicide
Item 12 ideation. When adjusting for gender, depression, PTSD
3 3.5 4 Item 13 symptom severity, and history of suicide attempt, the
SCS-S Unlovability (b = 0.16, P = 0.004) and Unsolv-
ability (b = 0.17, P = 0.001) subscales were signifi-
cantly associated with recent suicide ideation. The
SCS-S Unbearability subscale also showed a strong
trend toward significance (b = 0.20, P = 0.061). When
adding worst-point suicide ideation to the models as
covariates, the Unlovability (b = 0.08, P = 0.081) and
Unbearability (b = 0.05, P = 0.227) subscales were not
significantly associated with recent suicide ideation, but
the Unsolvability subscale was (b = 0.10, P = 0.008).
Results were unchanged when adjusting for pain
severity.
Sensitivity Analyses. We next conducted sensitivity
analyses to determine if observed patterns from the
multivariate analyses remained consistent across patient
subgroups that were dichotomized according to (1)
origin of pain (ie, injury- or non-injury-related) and (2)
duration of pain (ie, < 5 years or 5 years or more).
There were no between-groups differences in depression
severity, PTSD symptom severity, worst-point suicide
ideation, recent suicide ideation, or any of the 3 SCS-S
subscales between groups according to origin of pain

Item 9
Item 7
Item 11
Figure 2. Item characteristic curves for
9 SCS items, by factor.
(t < 0.88, P > 0.378) or duration of pain (t < 1.53,
P > 0.129).
Specific to origin of pain, standardized effects with
worst-point suicide ideation as the outcome were com-
parable among origin of pain subgroups as compared to
the full sample: noninjury-related pain patients (Unlov-
ability, b = 0.28, P < 0.001; Unsolvability, b = 0.17,
P = 0.010; and Unbearability, b = 0.29, P = 0.002) and
injury-related pain patients (Unlovability, b = 0.21,
P < 0.001; Unsolvability, b = 0.29, P < 0.001; and
Unbearability, b = 0.48, P < 0.001). Standardized
effects with recent suicide ideation as the outcome were
also comparable among origin of pain subgroups as
compared to the full sample: noninjury-related pain
Suicide Screener for Chronic Pain  7
Unlovability
Item 14 Item 18
Unsolvability
Item 10 Item 16
Unbearability
Item 12 Item 13
(Unlovability, b = 0.16, P = 0.304; Unsolvability,
b = 0.22, P = 0.040; and Unbearability, b = 0.12,
P = 0.694) and injury-related pain (Unlovability,
b = 0.22, P = 0.024; Unsolvability, b = 0.17,
P = 0.001; and Unbearability, b = 0.20, P = 0.003).
Specific to recency of pain onset, standardized effects
with worst-point suicide ideation as the outcome were
comparable in the < 5 years subgroup as compared to
the full sample (Unlovability, b = 0.21, P = 0.009;
Unsolvability, b = 0.29, P = 0.001; and Unbearability,
b = 0.26, P = 0.014), but were reduced in the 5+ years
subgroup (Unlovability, b = 0.14, P = 0.713; Unsolv-
ability, b = 0.16, P = 0.385; and Unbearability,
b = 0.11, P = 0.903). A similar pattern was found with
8  BRYAN ET AL.

Table 3. Correlations of Full SCS Subscales and Shortened SCS Subscales With Other Clinical Variables

Full Scale Shortened Scales

Unlovability Unsolvability Unbearability Unlovability Unsolvability Unbearability

BSSI
Current 0.18 0.09 0.11 0.18 0.09 0.17
Worst 0.20 0.12 0.22 0.20 0.15 0.26
PCL 0.04 0.11 0.07 0.06 0.12 0.02
BDI 0.13 0.07 0.09 0.14 0.06 0.19
ODI 0.12 0.15 0.15 0.13 0.16 0.14
PCS
Rumination 0.05 0.00 0.04 0.03 0.04 0.07
Magnification 0.05 0.04 0.04 0.04 0.02 0.11
Helplessness 0.02 0.01 0.03 0.02 0.02 0.09
CPAQ
Activities engagement 0.12 0.01 0.05 0.10 0.00 0.11
Pain willingness 0.08 0.02 0.08 0.06 0.00 0.12
MPI
Pain severity 0.02 0.07 0.03 0.05 0.07 0.06
Social support 0.03 0.01 0.01 0.04 0.02 0.05
Punishing response 0.07 0.13 0.07 0.06 0.13 0.04
Solicitousness 0.00 0.01 0.02 0.01 0.04 0.03

SCS, Suicide Cognitions Scale; BSSI, Beck Scale for Suicide Ideation; PCL, PTSD Checklist; BDI, Beck Depression Inventory; ODI, Oswestry Disability Index; PCS, Pain Catastrophizing
Scale; CPAQ, Chronic Pain Acceptance Questionnaire; MPI, Multidimensional Pain Inventory. Values in bold are statistically significant at P < 0.05. Values at or above r = |0.17| are
statistically significant at P < 0.05; values at or above r = |0.23| are statistically significant at P < 0.01; values at or above r = |0.28| are statistically significant at P < 0.001.

recent suicide ideation as the outcome: standardized
effects in the < 5 years subgroup were comparable to the
full sample (Unlovability, b = 0.16, P = 0.042; Unsolv-
ability, b = 0.12, P = 0.016; and Unbearability,
b = 0.15, P = 0.037) but were reduced in the 5+ years
subgroup (Unlovability, b = 0.13, P = 0.794; Unsolv-
ability, b = 0.01, P = 0.400; and Unbearability,
b = 0.03, P = 0.665).
DISCUSSION
Although suicidal thoughts and behaviors are common
among chronic pain patients, accurate and early detec-
tion of high-risk patients is challenging due to the
dynamic nature of many suicide risk factors, suicide
ideation, and suicidal intent. Newer methods that can
efficiently and reliably capture elevated suicide risk
despite these fluctuations are therefore needed. The SCS
was developed to assess identity-based suicide-specific
beliefs and schemas that reflect the chronic dimension of
suicide risk that persists as a more stable construct
underneath the more dynamic and transient aspects of
suicide risk. Previous research with the SCS supports its
utility as a concurrent and prospective predictor of
suicidal thoughts and behaviors within psychiatric
samples,7,8 but to date no studies have evaluated the
scale’s use in nonpsychiatric medical settings, and no
studies have sought to reduce the full SCS from 18 items
to a briefer version that can be implemented practically
in busy clinical settings.
Similar to previous findings,8 results of the present
study indicate that the full SCS is best conceptualized as
a 3-factor scale composed of statements that assess the
patient’s perceived worthlessness, failure, and self-
hatred (ie, Unlovability); perceived hopelessness and
deserving of punishment (Unsolvability); and perceived
inability to tolerate or cope with emotional pain
(Unbearability). Graded IRT analyses indicated that
the full SCS could be reduced to a 9-item SCS-S with
minimal adverse impact on the scale’s psychometric
properties and construct validity. Consistent with the
intent of the scale, which is to assess perceptions and
belief that are highly specific to the suicidal state,
patterns of SCS item endorsement were highly skewed,
with greater than 75% of patients endorsing the lowest
value possible (1, “strongly disagree”) on 14 of the 18
items. One notable exception was SCS item 3 (“I can’t
stand this pain anymore”), which showed much less
skew in endorsement, which is not surprising in a
chronic pain population. Overall, these patterns suggest
that most chronic pain patients do not subjectively
identify the beliefs and perceptions assessed by the SCS
as indicative of their internal state despite the strong
validation of the SCS in the present sample. Lower
reported levels of distressing cognitions may be a
function of the complexity of chronic pain and a
tendency among chronic pain patients to eschew psy-
chiatric symptoms in favor of less stigmatizing medical
symptoms (ie, somatization) resulting in increased
utilization of medical care with increased emotional

distress.15 If true, then ongoing surveillance of cogni-
tions and emotional states related to suicide risk in
medical environments is highly indicated, and a pain-
validated, abbreviated version of the SCS (as established
in the present study) offers unparalleled capacity for this
screening.
Despite the SCS items’ skewed endorsement pat-
terns, the 5-point Likert response option format for the
9 items retained for the SCS-S nonetheless showed
excellent discriminative capacity, such that the endorse-
ment of higher scores on each item contributed useful
incremental information. In other words, patients who
endorsed a 2 (“disagree”) on any given item were
meaningfully different from patients who endorsed a 1
(“strongly disagree”); those who endorsed a 3 (“neu-
tral”) were meaningfully different from patients who
endorsed a 2; and so on. This suggests that even very
low levels of endorsement of SCS-S items are clinically
relevant.
With respect to convergent and discriminant validity,
comparison of the SCS and SCS-S subscale scores
indicated that both versions were significantly correlated
with severity of current and worst-point suicide idea-
tion, but were not correlated with emotional distress,
disability, pain-related cognitions, pain severity, or
caretaker support. These patterns suggest the SCS and
SCS-S are similarly tapping into constructs that are
directly related to suicidal thoughts and desires but are
independent of emotional distress and pain-related
variables. This provides support for the specificity of
the SCS-S to suicide risk, making the SCS-S valuable
beyond more unidimensional measures of risk like
depression screening measures (eg, BDI, Patient Health
Questionnaire-9) commonly used in medical environ-
ments. Results of the multivariate regression analyses
further support the construct validity of the SCS-S and
also demonstrate incremental validity beyond the effects
of other relevant risk factors like depression and PTSD
symptom severity regardless of source of pain (ie, injury
or noninjury).
Results of our post-hoc sensitivity analyses further
suggest the SCS-S’s incremental validity may be most
pronounced among chronic pain patients with relatively
recent pain onset. Among patients with 5 or more years
of pain, however, the strength of association between
SCS-S scores and suicide ideation was reduced. It is
possible that this is attributable to reduced statistical
power associated with the small subgroup sizes, but the
observed reductions in effect sizes suggest otherwise. An
alternative possibility is that the nature of one’s beliefs
Suicide Screener for Chronic Pain  9
about unlovability, unsolvability, and unbearability
change over time as one’s pain continues to persist.
Additional research using larger samples and longitudi-
nal methods are needed to further understand this
preliminary finding.
Consistent with the conceptualization of the SCS-S
as a measure of chronic and persistent suicide risk,
the SCS-S subscales (especially the Unbearability
scale) had somewhat stronger correlations with
worst-point suicide ideation as compared to current
suicide ideation, although the present sample size was
too small to detect a statistically significant difference
between the correlation values. This aligns with the
fluid vulnerability theory’s notion of chronic suicide
risk, which is distinguished from the notion of acute
or dynamic risk. Specifically, chronic suicide risk
denotes that dimension of suicide risk that remains
somewhat stable and persists over time, whereas
acute suicide risk denotes that dimension of suicide
risk that fluctuates concurrent with the individual’s
context and psychological state. Worst-point suicidal
crises are more strongly associated with risk for
suicidal behaviors than current or recent suicidal
crises,25 suggesting that the predictive utility of worst-
point suicidal crises endures over time, even in the
presence of more recent crises. The SCS-S might
therefore be a useful measure because it taps into
beliefs that are temporally stable and highly specific
to suicide risk. These characteristics are noteworthy
when one considers that none of the SCS-S’s items
include the word suicide, which may enhance identi-
fication and detection of at-risk patients who are
otherwise reluctant to disclose suicidal thoughts.
Additional research is needed to confirm this possi-
bility.
Limitations of the present study include restriction
to a single sample composed of military beneficiaries;
results should therefore be considered preliminary
until replicated in additional samples. Second, our
cross-sectional design restricts our ability to evaluate
the utility of the SCS-S as a prospective predictor of
future suicidal thoughts and behaviors. Longitudinal
research is needed to confirm the predictive validity of
the SCS-S. Third, we were unable to assess the age of
participants, which limits our ability to understand
the generalizability of results to the wider population
of chronic pain patients. Despite these limitations, the
results of the present study support the psychometric
properties of the SCS-S in a nonpsychiatric population
with elevated risk for suicide, and suggest the scale
10  BRYAN ET AL.
may be a useful and practical tool for identifying and
monitoring suicide risk.
ACKNOWLEDGEMENTS
Funding provided by U.S. Department of Defense, U.S.
Army Medical Research and Materiel Command
(D61_I_10_J5_148).
REFERENCES
1. Rudd MD. Fluid vulnerability theory: a cognitive
approach to understanding the process of acute and chronic
suicide risk. In: Ellis TE, ed. Cognition and Suicide: Theory,
Research, and Therapy. Washington, DC: American Psycho-
logical Association; 2006:355–368.
2. Bryan CJ, Rudd MD. The importance of temporal
dynamics in the transition from suicidal thought to behavior.
Clin Psychol Sci Pract. 2016:23:21–25.
3. Kovacs M, Beck AT, Weissman A. The communication
of suicidal intent: a reexamination. Arch Gen Psychiatry
1976;33:198–201.
4. Hall RC, Platt DE, Hall RC. Suicide risk assessment: a
review of risk factors for suicide in 100 patients who made
severe suicide attempts: evaluation of suicide risk in a time of
managed care. Psychosomatics. 1999;40:18–27.
5. Coombs DW, Miller HL, Alarcon R, Herlihy C, Lee
JM, Morrison DP. Presuicide attempt communications
between parasuicides and consulted caregivers. Suicide Life
Threat Behav 1992;22:289–302.
6. Busch KA, Fawcett J, Jacobs DG. Clinical correlates of
inpatient suicide. J Clin Psychiatry 2003;64:14–19.
7. Bryan CJ, Rudd MD, Wertenberger E, et al. Improving
the detection and prediction of suicidal behavior among
military personnel by measuring suicidal beliefs: an evaluation
of the Suicide Cognitions Scale. J Affect Disord 2014;159:15–
22.
8. Ellis TE, Rufino KA. A psychometric study of the
Suicide Cognitions Scale with psychiatric inpatients. Psychol
Assess 2015;27:82–89.
9. Treharne GJ, Lyons AC, Kitas GD. Suicidal ideation in
patients with rheumatoid arthritis: research may help identify
patients at high risk. BMJ 2000;321:1290.
10. Juurlink DN, Herrmann N, Szalai JP, Kopp A,
Redelmeier DA. Medical illness and the risk of suicide in the
elderly. Arch Intern Med 2004;164:1179–1184.
11. Ratcliffe GE, Enns MW, Belik SL, Sareen J. Chronic
pain conditions and suicidal ideation and suicide attempts: an
epidemiologic perspective. Clin J Pain 2008;24:204–210.
12. Fishbain DA, Cutler R, Rosomoff HL, Rosomoff RS.
Chronic pain-associated depression: antecedent or conse-
quence of chronic pain? A review Clin J Pain 1997;13:116–137.
13. Timonen M, Viilo K, Hakko H, et al. Suicides in
persons suffering from rheumatoid arthritis. Rheumatology
(Oxford) 2003;42:287–291.
14. Fishbain DA. The association of chronic pain and
suicide. Semin Clin Neuropsychiatry. 1999;4:221–227.
15. Kanzler KE, Bryan CJ, McGeary DD, Morrow CE.
Suicidal ideation and perceived burdensomeness in patients
with chronic pain. Pain Pract. 2012;12:602–609.
16. Van Orden KA, Lynam ME, Hollar D, Joiner TE.
Perceived burdensomeness as an indicator of suicidal symp-
toms. Cognit Ther Res. 2006;30:457–467.
17. Joiner TE Jr, Van Orden KA, Witte TK, et al. Main
predictions of the interpersonal-psychological theory of suici-
dal behavior: empirical tests in two samples of young adults. J
Abnorm Psychol 2009;118:634–646.
18. Bryan CJ, Morrow CE, Anestis MD, Joiner TE. A
preliminary test of the interpersonal-psychological theory of
suicidal behavior in a military sample. Personality Individ
Differ 2010;48:347–350.
19. Bryan CJ, Clemans TA, Hernandez AM. Perceived
burdensomeness, fearlessness of death, and suicidality among
deployed military personnel. Personality Individ Differ
2012;52:374–379.
20. Bryan CJ, Hernandez AM, Allison S, Clemans T.
Combat exposure and suicide risk in two samples of military
personnel. J Clin Psychol 2013;69:64–77.
21. Van Orden KA, Witte TK, Cukrowicz KC, Braithwaite
SR, Selby EA, Joiner TE Jr. The interpersonal theory of suicide.
Psychol Rev 2010;117:575–600.
22. Cousineau N, McDowell I, Hotz S, Hebert P. Measur-
ing chronic patients’ feelings of being a burden to their
caregivers: development and preliminary validation of a scale.
Med Care 2003;41:110–118.
23. McPherson CJ, Wilson KG, Lobchuk MM, Brajtman S.
Self-perceived burden to others: patient and family caregiver
correlates. J Palliat Care 2007;23:135–142.
24. Beck AT, Steer RA. Manual for the Beck Scale for
Suicide Ideation. San Antonio, TX: Psychological Corpora-
tion; 1991.
25. Joiner TE Jr, Steer RA, Brown G, Beck AT, Pettit JW,
Rudd MD. Worst-point suicidal plans: a dimension of suici-
dality predictive of past suicide attempts and eventual death by
suicide. Behav Res Ther 2003;41:1469–1480.
26. Beck AT, Steer RA, Brown GK. Manual for the Beck
Depression Inventory-II. San Antonio, TX: Psychological
Corporation; 1996.
27. Weathers F, Litz B, Herman D, Huska J, Keane T. The
PTSD Checklist (PCL): Reliability, Validity, and Diagnostic
Utility. San Antonio, TX: Paper presented at the International
Society for Traumatic Stress Studies; 1993.
28. American Psychiatric Association. Diagnostic and
Statistical Manual Of Mental Disorders, Text Revision
(DSM-IV-TR). Washington, DC: American Psychiatric Asso-
ciation; 2000.

29. Sullivan MJ, Bishop SR, Pivik J. The pain catastro-
phizing scale: development and validation. Psychol Assess
1995;7:524.
30. McCracken LM, Vowles KE, Eccleston C. Acceptance
of chronic pain: component analysis and a revised assessment
method. Pain 2004;107:159–166.
31. Kerns RD, Turk DC, Rudy TE. The West Haven-Yale
Multidimensional Pain Inventory (WHYMPI). Pain 1985;23:
345–356.
Suicide Screener for Chronic Pain  11
32. Fairbank JC, Pynsent PB. The Oswestry disability
index. Spine. 2000;25:2940–2953.
33. Samejima F. Estimation of Latent Ability Using a
Response Pattern of Graded Scores. Richmond, VA: Psycho-
metric Society; 1969.
34. Raubenheimer J. An item selection procedure to
maximise scale reliability and validity. S Afr J Indust Psychol.
2004;30:59–64.