Acceptability, feasibility, and utility of a safety aid reduction treatment in underserved veterans: a pilot investigation

Abstract

Access to mental health services, particularly for veterans residing in underserved communities, remain scarce. One approach to addressing availability barriers is through the use of group-based transdiagnostic or unified treatment protocols. One such protocol, Safety Aid Reduction Treatment (START), previously termed False Safety Behavior Elimination Treatment (FSET), has received increasing empirical support. However, prior research has only examined this treatment among civilians with a primary anxiety diagnosis. Thus, the purpose of the current study was to replicate and extend prior research by examining the acceptability, feasibility, and utility of START among veterans, particularly those living in underserved communities, and across a wider array of diagnoses. Veterans (n = 22) were assessed prior to, immediately after, and one month following the 8-week treatment. The majority of veterans found START useful and acceptable. Additionally, recruitment and retention rates suggest that the treatment was feasible. Notably, results revealed reductions in overall anxiety, depression, and safety aid usage, which were maintained throughout the brief follow-up period. These findings add to a growing body of literature highlighting the utility of transdiagnostic approaches in the amelioration of various anxiety and related disorders. Limitations include the small sample size and uncontrolled design.

Anxiety and related disorders (e.g. anxiety disorders, trauma-and stressor-related disorders, obsessive-compulsive and related disorders, and depressive disorders) are a highly prevalent, debilitating, and costly category of mental illness (Greenberg et al., 1999, 2003; Kessler et al., 2005, 2012). They are also the most common mental health problems faced by individuals returning from war (Hoge et al., 2004; Seal et al., 2007). Thus, the treatment of these disorders is a top priority within the Department of Veterans Affairs (VA; Ruzek et al., 2012). Although there are a number of evidence-based psychotherapies (EBPs) available to manage these conditions (Butler et al., 2006), a significant proportion of veterans, particularly those residing in underserved areas throughout the United States, do not receive mental health care (Cully et al., 2010). One of the largest barriers to care in such communities is the lack of available services (Health Resources and Services Administration, 2020). Thus, there remains a need for innovative approaches to increase accessibility of high-quality mental health services in underserved communities.

One approach to addressing availability barriers is through the use of group-based transdiagnostic or unified treatment protocols. These approaches are based on the understanding that anxiety and related disorders share a common etiology (Allen et al., 2008). This is most apparent in the elevated rates of comorbidity found across these conditions. Indeed, when lifetime diagnoses are considered, anxiety and related disorder comorbidity rates have been found to be as high as 81% (Brown et al., 2001). This degree of overlap can complicate diagnostic distinction and subsequent treatment selection. Transdiagnostic treatments were established to have broad applicability by attending to the commonalities across these conditions (Craske, 2012). For instance, by simultaneously addressing problematic thoughts and behaviors shared among anxiety and related disorders, clinicians can treat co-occurring conditions both within and across individuals during the same clinical hour, thereby increasing access to care.

Consequently, researchers have begun to examine the effectiveness of one type of transdiagnostic treatment specifically focused on the reduction of safety aids. Safety aids are maladaptive cognitive or behavioral strategies designed to prevent, avoid, or alleviate anxiety and/or distress (Helbig-Lang & Petermann, 2010; Salkovskis, 1991). Empirical work suggests that safety aids, such as situational avoidance (e.g. avoiding crowded stores), checking behaviors (e.g. repeatedly checking doors, windows, or locks), reassurance seeking (e.g. repeatedly seeking medical advice), and use of alcohol and substances (e.g. drinking to cope), temporarily alleviate negative affectivity in the context of stressors (Carter et al., 1995). Although useful in the short-term, it is well documented that repeated use of these behaviors in the long-term contributes to the development and maintenance of various anxiety and related disorders (Deacon & Maack, 2008; Ehring et al., 2006; Olatunji et al., 2011; Salkovskis et al., 1999; Wells et al., 1996; Woody & Rachman, 1994). However, the function of safety aid use can vary according to the diagnosis. For example, an individual with panic disorder may engage in safety behaviors such as repeatedly checking for exits and restrooms due to fears of needing to escape during a panic attack. Similarly, an individual with posttraumatic stress disorder (PTSD) may engage in the same safety aid (i.e. repeatedly checking/scanning for exits) but due to safety concerns or fears of encountering a trauma-related cue. Further, an individual with major depressive disorder (MDD) may repeatedly seek reassurance from friends and family due to concerns that others are mad at them, whereas an individual with social anxiety may engage in the same safety aid but due to fears of negative evaluation. Lastly, an individual with obsessive compulsive disorder (OCD) may repeatedly check doors, windows, and locks, more often than necessary, due to concerns about being responsible for harm or injury, whereas an individual with PTSD may engage in the same behavior due to fears regarding one’s safety.

Given the ubiquitous nature of these behaviors across anxiety and related disorders, treatments targeting these cognitive/behavioral strategies may offer a simplified approach to treating a wide range of problems, particularly during the same clinical hour. To this end, Schmidt et al. (2012) compared the effectiveness of a group-based safety aid reduction treatment, titled False Safety Behavior Elimination Therapy (FSET), to a wait-list control using a large treatment-seeking community sample (N = 96) with primary anxiety diagnoses (i.e. panic disorder, generalized anxiety disorder, and social anxiety disorder). Relative to a wait-list control, participants receiving FSET demonstrated improvements in overall anxiety, avoidance, depression, and clinician-rated severity, which were maintained throughout the follow-up period. Notably, effect size indices were comparable to those found in randomized clinical trials (RCTs) for individual anxiety and depression diagnoses, suggesting that the treatment is capable of producing clinical benefits. Extending these findings, Riccardi et al. (2017) examined the efficacy of FSET delivered in a briefer, individual treatment format among community patients with a primary anxiety diagnosis (i.e. panic disorder, generalized anxiety disorder, and social anxiety disorder). Once again, the authors found participants in the active condition (compared to those in the wait-list control) demonstrated improvements in anxiety, avoidance, depression, functional impairment, and clinician-rated severity, which were maintained throughout the follow-up period.

Although promising, no research to date has systematically examined this protocol among veterans, particularly in the context of underserved treatment settings. This is especially important as an estimated 5.2 million veterans reside in rural or underserved communities across the country, making it difficult for them to access quality medical and mental health care (Veterans Health Administration Office of Rural Health, 2014). Whereas the VA established Community Based Outpatient Clinics (CBOCs) as a way to improve access in these areas, many of these clinics have only one psychologist on staff, limiting the breadth of mental health services available (Panangala & Mendez, 2010). In addition, care is often delivered by providers who lack the hours needed to implement time intensive monotherapies. Group-based protocols targeting a number of different diagnoses are likely to be more attractive in these settings and, therefore, more easily disseminated as they require fewer training resources. Against this backdrop, the current study was designed to replicate and extend prior research by examining the acceptability, feasibility, and utility of FSET delivered in a group-based format among veterans receiving care at one of two CBOCs. Given the ubiquitous nature of safety aids, veterans were also recruited across a wider array of anxiety and related diagnoses (i.e. inclusion of OCD, PTSD, and MDD). To reflect these changes, the FSET manual was updated and renamed Safety Aid Reduction Treatment (START). It was hypothesized that START would be both acceptable and feasible to veterans. Further, it was hypothesized that START would lead to reductions in safety aid usage as well as reductions in anxiety, depression, and overall impairment both immediately and over time.

Method

Participants

Veterans (N = 24) were recruited from June 2018 to January 2019 at two VA CBOCs serving the southeast Louisiana catchment area to participate in a single arm pilot trial (clinicaltrials.gov identifier: NCT03285516). Inclusion and exclusion criteria were kept to a minimum to facilitate rapid recruitment and permit generalization of results to a real-world setting of veterans presenting for treatment at CBOCs. To be eligible, veterans had to: 1) meet criteria for an anxiety or related disorder (i.e. anxiety disorder, obsessive-compulsive and related disorder, trauma- and stressor-related disorder, or depressive disorder); and 2) be primarily served at a CBOC. Veterans were excluded if they were: 1) dependent upon substances and in need of immediate detoxification or treatment; 2) actively psychotic; or 3) suffered from uncontrolled bipolar disorder (i.e. not stable on medications for 30 days from the date of referral). Although not explicitly stated as an exclusion criterion, no veterans who were acutely suicidal (that is, requiring hospitalization due to imminent risk) were referred to the study, per clinic policy.

Of the 24 veterans referred, two failed to meet diagnostic criteria for an anxiety and related disorder, bringing the final sample size to 22. Conversely, all other participants met Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) criteria for at least one anxiety and related disorder (100%), including a depressive disorder (86%), an anxiety disorder (82%), a trauma- and stressor-related disorder (82%), and an obsessive-compulsive and related disorder (18%), with 23% also meeting criteria for a substance-related and addictive disorder. Further, veterans were primarily male (63%) with ages ranging from 25 to 74 years (M_age = 48.6; SD = 14.2). In terms of racial composition, 42% identified as Caucasian, followed by African American (33%), Other (e.g. Biracial; 17%), and American Indian/Alaskan Native (8%). Marital status was as follows: 38% married; 38% divorced, 21% never married; and 3% widowed. A little over half of the sample completed at least some college (54%), followed by high school or the equivalent (25%), college (17%), and a post-graduate degree (4%). With regard to military characteristics, 46% served in the Army, followed by Navy (17%), Marines (13%), Air Force (8%), and National Guard (3%), with 13% serving in more than one branch. Finally, war zone service included Operation Enduring Freedom/Iraqi Freedom/New Dawn (OEF/OIF/OND; 38%), Operation Desert Storm/Shield (ODS; 17%), Vietnam (17%), and other (e.g. peace keeping missions; 4%), with 4% serving in multiple war zones, 17% never having deployed, and 3% failing to respond.

Measures

Consistent with other transdiagnostic treatment trials (Barlow et al., 2017), an assessment battery was constructed such that each measure would be relevant to veterans, regardless of their primary diagnosis. Measures included the following:

Mini International Neuropsychiatric Interview (MINI)

The MINI is a brief, structured diagnostic interview for the DSM-5 (Sheehan et al., 1998). The MINI consists of close-ended questions designed to assess former Axis I psychiatric conditions. Diagnoses are made based on the number of affirmative responses. The MINI was administered at baseline by the principal investigator (PI; who is a licensed clinical psychologist) to ensure participant eligibility. Previous research has demonstrated that the MINI is both a reliable and valid diagnostic tool (Tolin et al., 2016). Given the pilot nature of the current study, inter-rater reliability data was not gathered.

Overall Anxiety Severity and Impairment Scale (OASIS)

The OASIS is a brief, 5-item self-report questionnaire designed to assess anxiety severity and impairment across various anxiety disorders (Campbell-Sills et al., 2009). Items were rated on a 5-point Likert-type scale ranging from 0 (No/None) to 4 (Constant/Extreme) and summed to create a total score. The OASIS has strong psychometric properties (Campbell-Sills et al., 2009) and was administered at all three timepoints to assess for treatment related changes in anxiety. In the present study, the OASIS demonstrated good internal consistency (αs ranged from 0.86 to 0.89).

Overall Depression Severity and Impairment Scale (ODSIS)

The ODSIS is a brief, 5-item self-report questionnaire designed to assess depression severity and impairment across mood disorders (Bentley et al., 2014). Items were rated on a 5-point Likert-type scale ranging from 0 (No/None) to 4 (Constant/Extreme) and summed to create a total score. The ODSIS has strong psychometric properties (Bentley et al., 2014) and was administered at all three timepoints to assess for treatment related changes in depression. In the present study, the ODSIS demonstrated excellent internal consistency (αs ranged from 0.92 to 0.93).

Safety Behavior Assessment Form (SBAF)

The SBAF is a 41-item self-report questionnaire designed to assess the frequency of safety behavior usage (Goodson et al., 2016). Items are rated on a 4-point Likert-type scale ranging from 0 (Never) to 3 (Always) and summed to create a total score. The SBAF has strong psychometric properties (Goodson et al., 2016) and was administered at all three timepoints to assess for treatment related changes in safety aid usage. In the present study, the SBAF demonstrated good internal consistency (αs ranged from 0.86 to 0.92).

Work and Social Adjustment Scale (WSAS)

The WSAS is a brief, 5-item self-report questionnaire designed to measure subjective interference of psychiatric symptoms in various life domains (Mundt et al., 2002). Items are rated on a 9-point Likert-type scale ranging from 0 (Not at all) to 8 (Very severely) and summed to create a total score. The WSAS has strong psychometric properties (Mundt et al., 2002) and was administered at all three timepoints to assess for treatment related changes in functional impairment across various life domains. In the present study, the WSAS demonstrated good internal consistency (αs ranged from 0.83 to 0.88).

Treatment acceptability questionnaire

This brief, 4-item investigator-developed self-report measure, which was modeled after other treatment acceptability questionnaires (Short et al., 2017), was administered post-intervention to assess veterans’ thoughts and opinions of START. In particular, veterans answered questions regarding the perceived utility (“How useful was the information and skills learned in treatment?”, “How likely are you to recommend this treatment to another veteran?”) and acceptability (“How difficult was it for you to participate in the treatment?”, “How likely are you to use the skills after treatment?”) of the intervention using a 5-point Likert-type scale ranging from 0 (Not at all/Not at all likely) to 4 (A lot/Very likely).

Treatment feasibility

Treatment feasibility was determined by examining treatment recruitment, eligibility, and completion rates. Recruitment, eligibility, and retention estimates were calculated as follows: percent recruited = 100* (# referred/# reached); percent eligible = 100* (# screened for eligibility/# eligible); percent retained = 100* (# enrolled/# completing treatment).

Procedure

Veterans were given verbal information about the study by a mental health provider during an intake appointment or treatment planning session. Veterans who expressed interest were contacted by the research coordinator and scheduled for an in-person baseline appointment to determine eligibility. All procedures were approved by the facilities Institutional Review Board and Research and Development committee.

Baseline appointment

During the baseline appointment, veterans completed an informed consent process where they were provided with all the necessary information to make an informed decision regarding their participation. Those who provided written consent completed a brief diagnostic interview and battery of self-report questionnaires. Eligible veterans were then provided with the start date and time of group, whereas ineligible veterans were referred back to their referring treatment provider. The baseline appointment took approximately one hour to complete, and veterans were monetarily compensated $50 for their time and travel, regardless of eligibility.

Treatment

Eligible veterans were treated in one of four groups (two groups conducted at each CBOC), which ranged in size from 6 to 7 participants (M = 5.75, SD = 0.96). Treatment consisted of eight group sessions, delivered once weekly, each lasting approximately one hour in duration. All groups were led by the PI (a licensed clinical psychologist) and a licensed clinical social worker at the CBOC. START includes many of the key elements found in empirically supported treatments for anxiety and related disorders including 1) psychoeducation regarding the nature of anxiety/distress and 2) exposure to internal sensations and external situations that are connected to one’s distress response via identification and reduction of safety aids. Notably, the original manual was updated to include safety aids relevant to a wider array of anxiety and related disorders (e.g. image or thought replacement found in OCD and PTSD), as well as to include veteran specific examples. With regard to specific session content, sessions 1–3 focused on the fear response (i.e. physical, cognitive, and behavioral) and factors that maintain the fear response (i.e. safety aids). Veterans were taught how to identify and eliminate (i.e. delay and/or decrease frequency or degree of use) various safety aids. Sessions 4–7 were then spent identifying and discussing specific categories of safety aids including cognitive avoidance, situational avoidance, checking behaviors, reassurance seeking, avoidance of bodily sensations, companions, and use of alcohol and other substances. Veterans were instructed to select specific safety aids within each category to fade while identifying the specific technique that would be used to fade. Finally, session 8 focused on reviewing the skills learned throughout treatment and factors that contribute to relapse (for additional details regarding the treatment see Schmidt et al., 2012).

Post-treatment appointment

After the last treatment session, veterans were asked to complete a brief battery of self-report questionnaires. This appointment took approximately 30-minutes to complete, and veterans were monetarily compensated $25 for their time.

One-month follow-up appointment

One-month post-treatment, veterans returned to the clinic to complete a brief battery of self-report questionnaires. This appointment took approximately 30-minutes to complete, and veterans were monetarily compensated $50 for their time and travel, thanked for their participation, and directed back to their referring provider for additional treatment planning as needed.

Power considerations

Given the pilot nature of this one-year proposal, a formal power analysis was not conducted. Rather, we sought to determine the feasibility, acceptability, and utility of delivering START at two CBOCs. To best meet the individual needs of patients, it was determined that five to six veterans per group would be optimal. Further, given the one-year timeline and brief follow-up window, it was determined that we could conduct two groups at each CBOC. Thus, we aimed to retain 20 veterans throughout the treatment trial. Based on prior transdiagnostic treatment research, rates of attrition have been found to be approximately 15% (Schmidt et al., 2012). As such, the proposed sample size was increased by four to account for dropouts, thereby arriving at a final sample size of 24 veterans (six per group).

Data analytic strategy

All analyses were conducted using IBM SPSS Statistics version 26. Data were first inspected for entry errors, outliers, missingness, and normality. Missing data were evaluated using Little (1988) Missing Completely at Random (MCAR) test wherein a statistically significant value indicates that data are not missing completely at random. The item level linear trend at point imputation function in SPSS was used to account for missingness. Zero-order correlations for all self-report measures at baseline were then examined. Next, baseline differences in basic demographic (i.e. age, sex, race) and psychological variables (i.e. OASIS, ODSIS, SBAF, and WSAS) between the two CBOCs were examined using Fisher’s exact test for dichotomous variables (given that two cells had an expected count less than five) and independent samples t-test for continuous variables. Treatment acceptability and feasibility data were then calculated. Specifically, treatment acceptability was examined using basic descriptive statistics. Treatment feasibility was determined by examining treatment referral, entry, and completion rates. Treatment completion was defined as completing at least 60% of the treatment (i.e. five out of eight sessions). This benchmark was selected based on previous treatment outcome research (Jacobson et al., 1996; Raines et al., 2020). Finally, repeated measures ANOVAs were conducted to compare mean scores on the OASIS, ODSIS, SBAF, and WSAS at baseline, post-treatment, and one-month follow-up. Partial eta squared was reported as a measure of effect size using the guidelines proposed by Cohen (1988) wherein values at or above .01 = small, .06 = moderate, and .14 = large.

Results

Preliminary analyses

At the item level, missing data ranged from 18% to 32% solely due to participant dropout. According to Little’s MCAR test, data were missing at random, indicating that it was appropriate to impute missing values, χ2 = 65.91 (55), p = .149. Further, an examination of skew and kurtosis values revealed that all variables were normally distributed. Zero-order correlations for all self-report variables over time can be found in Table 1. When examining baseline differences between the two CBOCs, results revealed that there was not a significant association between sex or race and CBOC location (ps > .386). Further, there were no significant differences in age (t (20) = −.44, p = .664, two-tailed), OASIS (t (20) = .75, p = .464, two-tailed), ODSIS (t (20) = .10, p = .925, two-tailed), SBAF (t (20) = .48, p = .643, two-tailed) or WSAS scores (t (20) = .63, p = .537, two-tailed) across the two CBOCs (see Table 1).

Table 1.

Means, standard deviations, and zero-order correlations between all self-report measures at baseline, post-treatment, and one-month follow-up.

Variable	1	2	3	4	5	6	7	8	9	10	11	12
1. BL OASIS	-
2. BL ODSIS	.55**	-
3. BL SBAF	.42*	.47*	-
4. BL WSAS	.43*	.56**	.36	-
5. PO OASIS	.69***	.50*	.29	.35	-
6. PO ODSIS	.54**	.48*	.28	.24	.84***	-
7. PO SBAF	.44*	.28	.48*	.15	.61**	.56**	-
8. PO WSAS	.73***	.61**	.45*	.54**	.79***	.63**	.55**	-
9. 1 M OASIS	.62**	.30	.18	.45*	.77***	.67***	.51**	.68***	-
10. 1 M ODSIS	.42*	.28	.11	.22	.59**	.51**	.48*	.48*	.76***	-
11. 1 M SBAF	.16	.07	.20	.13	.34	.29	.70***	.27	.48*	.34	-
12. 1 M WSAS	.45*	.37	.28	.44*	.65**	.54**	.51**	.75***	.68***	.71***	.45*	-
CBOC 1 M	14.36	13.27	72.08	27.82	12.19	11.46	56.54	24.81	11.74	11.16	64.54	28.50
CBOC 1 SD	3.29	5.18	6.66	11.63	2.29	5.36	12.11	9.07	3.30	4.71	9.67	7.19
CBOC 1 Ranges	9-20	4-20	63-85	5-40	7-20	0-19	37-79	9-39	7-17	0-18	53-84	13-37
CBOC 2 M	13.27	13.09	69.18	25.00	9.36	8.65	44.22	16.19	8.67	8.14	43.45	14.91
CBOC 2 SD	3.55	3.59	19.09	9.30	2.71	3.34	10.85	8.90	2.76	3.30	11.85	6.39
CBOC 2 Ranges	8-18	9-20	40-91	7-39	6-15	4-17	29-66	6-33	4-14	3-14	27-70	4-27

n = 22; BL = Baseline; PO = Post-treatment; 1 M = One-month follow-up; OASIS = Overall Anxiety Severity and Impairment Scale; ODSIS = Overall Depression Severity and Impairment Scale; SBAF = Safety Behavior Assessment Form; WSAS = Work and Social Adjustment Scale; CBOC = Community Based Outpatient Clinic; M = Mean; SD = standard deviation.

^**p < .010

^*p < .05.

Treatment acceptability

Treatment acceptability data can be found in Table 2. All veterans found the information and skills learned in treatment at least “somewhat” useful with the majority indicating that the information and skills were “very” useful. Further, the majority of veterans reported that it was “a little” or “not at all difficult” for them to participate in treatment. Most veterans also reported that they were at least “somewhat likely” to use the skills beyond treatment with the majority indicating that they were “very likely” to continue implementing lessons learned. Finally, all veterans reported that they were at least “somewhat likely” to recommend the treatment to another veteran with the majority indicating that they were “very likely” to refer a fellow service member.

Table 2.

Frequencies of Responses to Acceptability Items.

Responses % (n)
How useful were the information and skills learned in treatment?
Not at all	0% (0)
A little	0% (0)
Somewhat	17.6% (3)
Quite	29.5% (5)
Very	52.9% (9)
How difficult was it for you to participate in the treatment program?
Not at all	58.8% (10)
A little	23.5% (4)
Somewhat	5.9% (1)
Quite	5.9% (1)
Very	5.9% (1)
How likely are you to use the skills after treatment?
Not at all likely	6.7% (1)
A little likely	0% (0)
Somewhat likely	13.3% (2)
Quite likely	20.0% (3)
Very likely	60.0% (9)
How likely are you to recommend this treatment to another veteran?
Not at all likely	0% (0)
A little likely	0% (0)
Somewhat likely	6.3% (1)
Quite likely	12.5% (2)
Very likely	81.2% (13)

Treatment feasibility

Out of the 38 veterans referred to START, 24 (63%) responded to contact attempts, indicated an interest in participating, and were subsequently scheduled for a baseline appointment to determine eligibility. Of the 24 veterans who completed an eligibility assessment, 22 (92%) were eligible and enrolled in the treatment trial. Four veterans (18%) discontinued the intervention with another two veterans (9%) failing to complete the one-month follow-up. The mean number of sessions completed was 5.95 (SD = 2.21). In terms of frequency of attendance, seven veterans (32%) completed all eight appointments, three (14%) completed seven appointments, six (27%) completed six appointments, two (9%) completed five appointments, one (5%) completed four appointments, one (5%) completed two appointments, and two (9%) completed only one appointment.

Treatment utility

Means and standard deviations by timepoint can be found in Table 3. The mean OASIS and ODSIS scores at baseline were well above the clinical cut scores (Bentley et al., 2014; Campbell-Sills et al., 2009). The mean SBAF was higher than that found in a non-clinical sample (M = 60.05, SD = 16.10) but slightly lower than that found in a clinical sample (M = 76.07, SD = 24.19). Further, the mean WSAS score at baseline suggested moderately severe or worse psychopathology (Mundt et al., 2002). When examining self-report scores over time, results revealed a significant effect for OASIS (Wilks’ Lambda =.37, F (2, 20) = 17.2, p < .001) and ODSIS scores (Wilks’ Lambda =.62, F (2, 20) = 6.14, p = .008), which were associated with partial eta squared statistics of .63 and .38, respectively. Further, there was a significant effect of time on SBAF (Wilks’ Lambda =.28, F (2, 20) = 25.41, p < .001) and WSAS scores (Wilks’ Lambda =.72, F (2, 20) = 3.88, p = .038), which were associated with partial eta squared statistics of .72 and .28, respectively.

Table 3.

Means and standard deviations for all self-report variables over time.

Timepoint	OASIS M (SD)	ODSIS M (SD)	SBAF M (SD)	WSAS M (SD)
Baseline	13.82	13.18	70.63	26.41
	(3.39)	(4.35)	(14.03)	(10.37)
Post-Intervention	10.77	10.06	50.38	20.50
	(3.79)	(4.59)	(12.87)	(9.82)
One-month follow-up	10.20	9.65	54.00	21.71
	(3.36)	(4.26)	(15.10)	(9.61)

OASIS = Overall Anxiety Severity and Impairment Scale; ODSIS = Overall Depression Severity and Impairment Scale; SBAF = Safety Behavior Assessment Form; WSAS = Work and Social Adjustment Scale; M = mean; SD = standard deviation.

Discussion

The current study was designed to examine the acceptability, feasibility, and utility of a group-based transdiagnostic treatment specifically focused on the identification and elimination of safety aids using a small sample of veterans receiving care at one of two CBOCs. Although we did not have an a priori benchmark regarding acceptability, the majority of veterans found START useful and acceptable. Additionally, recruitment and retention rates suggest that the treatment was feasible. Indeed, dropout rates in other transdiagnostic treatment trials range from 13%-15% (Barlow et al., 2017; Schmidt et al., 2012) and dropout rates among veterans receiving group-based PTSD treatment reach as high as 52% (Sloan et al., 2013). Thus, the 18% dropout rate in the current study supports treatment feasibility. Moreover, results revealed reductions in overall anxiety, depression, and safety aid usage which were maintained throughout the brief follow-up period. These findings are consistent with hypotheses and prior safety aid reduction treatment trials (Arai et al., in press; Riccardi et al., 2017; Schmidt et al., 2012; Short & Schmidt, 2020). Importantly, the current study expanded upon previous work by demonstrating the utility of such interventions among veterans, in particular those receiving care in underserved areas, and across a wider array of anxiety and related disorders (i.e. inclusion of OCD, PTSD, and MDD).

Notably, the current findings also add to a growing body of literature highlighting the utility of transdiagnostic approaches in the amelioration of anxiety and related disorders (Farchione et al., 2012; Gros, 2014). In contrast to other transdiagnostic treatments, namely the Unified Protocol for the Transdiagnostic Treatment of Emotional Disorders (Barlow et al., 2017), START does not explicitly focus on thoughts or correcting thinking errors. Rather patients are taught to identify and reduce the use of various cognitive (e.g. distraction) and behavioral (e.g. situational avoidance) strategies, leading to the development and maintenance of anxiety and related conditions. This relatively simplified approach has numerous benefits for both the patient and provider. Indeed, cognitive treatments can be mentally challenging (Dimidjian et al., 2006), especially for patients who lack adequate verbal skills and the ability to articulate thoughts and feelings (Berardis et al., 2008). Further, the purely behavioral nature of START may offer additional benefits to the provider in that less time and training are needed to master the treatment protocol. Such benefits are of particular importance in underserved communities where access to effective mental health care and resources remain scarce.

There are a number of study specific factors including the sample, assessments, and procedures used that make it difficult to directly compare outcomes across transdiagnostic treatment trials. All things being equal, it is still useful to evaluate findings from the current investigation in light of previous research. Effect size indices in the current study were large and comparable to those reported in other transdiagnostic treatment trials (Bullis et al., 2015; Farchione et al., 2012) and in RCTs for individual anxiety diagnoses (Butler et al., 2006). Compared to other transdiagnostic treatments (Farchione et al., 2012; Gros, 2014), START is considerably briefer (i.e. 8 sessions compared to 12–18) and specifically designed to be delivered in a group-based format. One of the largest barriers to care within the VA is an acute shortage of mental health providers (US Department of Veterans Affairs, 2014). Group-based transdiagnostic treatments, which allow multiple patients to be seen within the same clinical hour, may offer a cost-effective and efficient solution to solving access issues. Such protocols are also likely to be more attractive to providers than other more time intensive therapies and thus easier to disseminate.

Findings from the current study should be considered in light of limitations and opportunities for future research. First, given the pilot nature, the current study lacked a control or comparison group. Although reductions in overall anxiety, depression, impairment, and safety aid usage were found, we cannot conclusively rule out that veterans merely improved due to the passage of time or because of nonspecific aspects of the treatment/study procedure (i.e. therapist contact). Second, given the one-arm design, we were unable to test potentially relevant mediators and/or moderators of treatment. Future investigations should utilize wait-list controls or active comparators to further determine the efficacy of START and specific mechanisms of action, as the current findings suggest that reductions in safety aid usage may be the avenue through which START reduces anxiety and depression. Third, all outcomes of interest were measured using self-report assessments. Despite the use of well-validated instruments, future research would benefit from multimethod assessments. Fourth, despite the emphasis on substance use as a safety aid, the SBAF does not measure substance related behaviors (e.g. consuming alcohol to cope). As such, it will be important for future studies to assess changes in alcohol and or substance use over time. Fifth and finally, the brief follow-up window did not allow for an assessment of treatment effects on remission rates of anxiety and related conditions. A larger sample followed over a longer period of time would allow for gold-standard diagnostic interviews to confirm symptom remission.

Despite these limitations, the current study adds to a growing body of literature highlighting the utility of transdiagnostic treatments. Expanding upon prior safety aid reduction treatment trials, we now have support for the acceptability and feasibility of START among veterans, particularly those seeking care in underserved areas. Further, we now have support for the utility of safety aid reduction treatments across a wider array of diagnoses, namely OCD, PTSD, and MDD. Pending replication, future studies should examine the utility of this transdiagnostic treatment in a virtual format as an alternate method of delivery. Future studies should also illicit feedback from clinicians on the relative training and burden of this approach particularly compared to monotherapies.

Figure 1.

CONSORT chart of participants detailing patient flow, assignment, and dropout. CONSORT = consolidated standards of reporting trials.