If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Correspondence to: Mandy J. Kumpula, PhD, Veterans Health Administration, 2851 University Avenue, Green Bay, WI 54311. Phone: (414) 573-3587; fax: (920) 431-2940.
Department of Veterans Affairs Office of Mental Health and Suicide Prevention, Washington, District of ColumbiaDepartment of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North Carolina
Department of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North CarolinaDepartment of Psychiatry and Behavioral Sciences, Duke University School of Medicine, Durham, North Carolina
Department of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North CarolinaMental Health Service Line, Durham Veterans Affairs Medical Center, Durham, North CarolinaDepartment of Psychiatry and Behavioral Sciences, Duke University School of Medicine, Durham, North Carolina
Department of Veterans Affairs Office of Mental Health and Suicide Prevention, Washington, District of ColumbiaDepartment of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North Carolina
Department of Veterans Affairs Office of Mental Health and Suicide Prevention, Washington, District of ColumbiaDepartment of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North Carolina
Department of Veterans Affairs Office of Mental Health and Suicide Prevention, Washington, District of ColumbiaDepartment of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North Carolina
Department of Veterans Affairs Office of Mental Health and Suicide Prevention, Washington, District of ColumbiaDepartment of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North Carolina
Department of Veterans Affairs Office of Mental Health and Suicide Prevention, Washington, District of ColumbiaDepartment of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North Carolina
Department of Veterans Affairs Mid-Atlantic Mental Illness Research, Education, and Clinical Center, Durham, North CarolinaMental Health Service Line, Durham Veterans Affairs Medical Center, Durham, North CarolinaDepartment of Psychiatry and Behavioral Sciences, Duke University School of Medicine, Durham, North CarolinaDepartment of Veterans Affairs, Durham Center for Health Services Research in Primary Care, Durham, North Carolina
Although most suicide-related deaths occur among male veterans, women veterans are dying by suicide in increasing numbers. Identifying and increasing access to effective treatments is imperative for Department of Veterans Affairs suicide prevention efforts. We examined the impact of evidence-based psychotherapies for depression on suicidal ideation and the role of gender and treatment type in patients’ responses to treatment.
Methods
Clinicians receiving case consultation in interpersonal psychotherapy, cognitive–behavioral therapy for depression, and acceptance and commitment therapy for depression submitted data on depressive symptoms and suicidal ideation while treating veterans with depression.
Results
Suicidal ideation was reduced across time in all three treatments. A main effect for wave was associated with statistically significant decreases in severity of suicidal ideation, χ2 (2) = 224.01, p = .0001, and a subsequent test of the Gender × Wave interaction was associated with differentially larger decreases in ideation among women veterans, χ2 (2) = 9.26, p = .001. Within gender-stratified subsamples, a statistically significant Treatment × Time interaction was found for male veterans, χ2 (4) = 16.82, p = .002, with levels of ideation significantly decreased at waves 2 and 3 in interpersonal psychotherapy and cognitive–behavioral therapy for depression relative to acceptance and commitment therapy for depression; the Treatment × Wave interaction within the female subsample was not statistically significant, χ2 (4) = 3.41, p = .492.
Conclusions
This analysis demonstrates the efficacy of each of the three tested evidence-based psychotherapies for depression as a means of decreasing suicidal ideation, especially in women veterans. For male veterans, decreases in suicidal ideation were significantly greater in interpersonal psychotherapy and cognitive–behavioral therapy for depression relative to acceptance and commitment therapy for depression.
Suicide prevention is a top clinical priority within the Department of Veterans Affairs (VA). In response, the Veterans Health Administration (VHA) has called for expanded access to and engagement in effective treatments to prevent veteran suicide (
). The number of veterans dying by suicide in the United States has been increasing, and although the majority of suicide-related decedents among veterans are men, women veterans are dying by suicide in increasing numbers (
U.S. Department of Veterans Affairs (VA), Office of Mental Health and Suicide Prevention (OMHSP) Facts About Suicide Among Women Veterans: August 2017.
). The risk of suicide among women veterans using VHA services has also increased in recent years, from 14.4 per 100,000 in 2001 to 17.3 per 100,000 in 2014 (
U.S. Department of Veterans Affairs (VA), Office of Mental Health and Suicide Prevention (OMHSP) Facts About Suicide Among Women Veterans: August 2017.
). The narrowing of the gender gap in suicide risk and the expanding concurrent risk of death by suicide among women veterans make it imperative that women veterans have access to effective treatments to address these risks.
Despite an extensive body of literature aiming to identify risk factors for suicide, a recent meta-analytic review concluded that existing research does not establish accurate or reliable predictors of suicide and “cannot provide much useful information about treatment and prevention targets,” appealing for improved methodology in suicide research (
). Amidst an ongoing evolution of empirical knowledge, the VA is tasked with addressing increasing rates of suicide among veterans and taking into account the unique contributions of gender.
Depressive disorders are a frequently cited risk factor for suicide (
) and, hence, provide a potential target for suicide prevention among women veterans. Women are more frequently diagnosed with depressive disorders than are men (
), highlighting an explicit link between depression and suicide risk. Women have higher lifetime prevalence of suicide ideation (20.1% vs. 12.7%) and twice the rate of lifetime suicide attempts compared with men (5.1% vs. 2.5%), and much of the increased risk of suicide attempts for women may be accounted for by increased ideation (
Lifetime suicidal behaviors and career characteristics among U.S. Army soldiers: Results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS).
Suicide and Life-Threatening Behavior.2017; 48: 230-250
concluded that depression is among the strongest predictors of suicide attempts, owing in large part to depression predicting the onset of suicidal ideation. In contrast, a thorough meta-analysis conducted by
did not identify internalizing disorders or suicidal ideation as substantial predictors of later suicidal behavior or death by suicide, leaving conclusions about the nature of these relationships unsettled.
The presence of heightened risk of suicide in veterans with depression is recognized in the VA/Department of Defense Clinical Practice Guideline for the Assessment and Management of Patients at Risk of Suicide (
U.S. Department of Veterans Affairs (VA), Office of Quality Safety and Value VA/DoD Clinical Practice Guideline for Assessment and Management of Patients at Risk for Suicide.
), which states, “research suggests that the effective treatment of mental health conditions (particularly major depression) reduces the risk of suicide and may decrease suicide rates.” Further, the VA/Department of Defense Clinical Practice Guideline recommends that “patients receive optimal evidence-based treatment for any mental health and medical conditions that may be related to the risk of suicide.” Among the first-line treatments recommended for major depressive disorder are evidence-based psychotherapies (EBPs), including cognitive–behavioral therapy for depression (CBT-D), interpersonal psychotherapy (IPT), and acceptance and commitment therapy for depression (ACT-D). CBT-D is a time-limited, structured psychotherapy that encourages behavioral changes that reduce depression, such as increases in pleasurable activities, and guides the patient through restructuring cognitions and beliefs that maintain depression. IPT addresses the relationship between adverse life events and depression by facilitating focus on processing affect related to interpersonal relationships (
Acceptance and commitment therapy and contextual behavioral science: Examining the progress of a distinctive model of behavioral and cognitive therapy.
) in reducing depressive symptoms. Within veteran populations, large within-group reductions in depressive symptoms have been observed in VHA clinical effectiveness studies of each intervention (CBT-D:
National dissemination of cognitive behavioral therapy for depression in the Department of Veterans Affairs health care system: Therapist and patient-level outcomes.
Journal of Consulting and Clinical Psychology.2012; 80: 707-718
Training in and implementation of acceptance and commitment therapy for depression in the Veterans Health Administration: Therapist and patient outcomes.
Although the effects of EBPs for depression on suicidality have not been explored thoroughly, available evidence suggests that these treatments are effective in reducing suicidal ideation. CBT-D and ACT-D were reported to reduce suicidal ideation specifically among veterans (
Differences in the effectiveness of psychosocial interventions for suicidal ideation and behaviour in women and men: A systematic review of randomised controlled trials.
reported superior responses to treatment among women in 33% of the studies they reviewed in a systematic review of psychosocial interventions for suicidal ideation and behavior. Treatment efficacy in reducing suicidal ideation or suicidal behavior did not differ by gender in the remaining studies. Notably, few of the reviewed studies examined EBPs for depression. Last, previous research on IPT and CBT-D from a large nonveteran sample found no associations between gender and treatment-related reductions in suicidal ideation (
). Against this conflicting background, additional research is needed to determine whether gender plays a role in moderating the effectiveness of evidence-based treatments for depression among veterans experiencing suicidal ideation.
Objectives
The VHA has been actively implementing a national initiative to disseminate and implement EBPs including CBT-D, IPT, and ACT-D. Since the initiative began in 2008, approximately 2,700 clinicians have received competency-based training to provide EBPs for the treatment of depression (for more details about the VA’s EBP training initiative, see
From the laboratory to the therapy room: National dissemination and implementation of evidence-based psychotherapies in the US Department of Veterans Affairs Health Care System.
). The first objective of the current program evaluation analysis was to examine the overall impact of participation in EBP for depression on suicidal ideation. We expected that participation in EBP for depression would be associated with statistically significant decreases in suicidal ideation across treatments and gender. Our second objective was to explore whether gender moderated patients’ response to treatment over time. The third and final objective was to examine whether treatment response varied by type of EBP treatment received.
Methods
Program Procedures
Data for the present program evaluation analysis were collected as part of the VA EBP Training Programs for CBT-D, IPT, and ACT-D between 2008 and 2013. During this timeframe, the CBT-D, IPT, and ACT-D training programs provided competency-based training consisting of participation in a 3-day workshop followed by 6 months of weekly case consultation via telephone. Training participants were VA mental health clinicians who provided psychotherapy services to veterans with depression. Training workshops included didactic and experiential training focused on developing the knowledge and skills essential for effective treatment delivery. Case consultation involved verbal discussion of cases, review of audio-recorded therapy sessions, and support of the development and implementation of therapy-specific skills. Weekly consultation calls were conducted by trained experts in CBT-D, IPT, or ACT-D, with an average of four clinicians per consultation group. Detailed descriptions of the CBT-D, IPT, and ACT-D training programs have been previously reported by
National dissemination of cognitive behavioral therapy for depression in the Department of Veterans Affairs health care system: Therapist and patient-level outcomes.
Journal of Consulting and Clinical Psychology.2012; 80: 707-718
Training in and implementation of acceptance and commitment therapy for depression in the Veterans Health Administration: Therapist and patient outcomes.
The CBT-D protocol was adapted specifically for veterans and military service members and is designed to be administered in 12–16 individual sessions (
National dissemination of cognitive behavioral therapy for depression in the Department of Veterans Affairs health care system: Therapist and patient-level outcomes.
Journal of Consulting and Clinical Psychology.2012; 80: 707-718
). The IPT protocol consists of up to 16 weekly individual psychotherapy sessions across three phases of treatment: initial (sessions 1–3), intermediate (sessions 4–13), and termination (sessions 14–16;
). Patients were veterans with a primary presenting problem of a depressive disorder. IPT had the additional requirement that patients were experiencing current problems involving a considerable life change, conflict with an important person, interpersonal deficits, and/or the death of an important person (
Clinicians taking part in EBP training agreed to participate in program evaluation, which included collecting and reporting treatment outcome measures administered to at least two veterans engaged in therapy while receiving case consultation. Outcome data were collected as a part of the routine program evaluation for the VA National EBP Training Program. Data collection and statistical analyses were reviewed and determined to be consistent with nonresearch quality improvement activities by the chief consultant of the VA Office of Mental Health and Suicide Prevention.
Measures
Demographics
Clinicians collected demographic information from each patient, including age, gender, level of education, race, and ethnicity.
) is a well-established 21-item self-report measure used to assess depression severity. The BDI-II was administered in three waves: at the outset of treatment (before or at session 1), midtreatment (approximately session 7), and during the final session. Each item is scored on a 0 to 3 scale. The total score is the sum of all items (range, 0–63), with higher scores indicating greater symptom severity. Suicidal ideation was assessed via patients’ score on item nine of the BDI-II, which is a four-level scale where 0 = “I don't have any thoughts of killing myself”; 1 = “I have thoughts of killing myself, but I would not carry them out”; 2 = “I would like to kill myself”; and 3 = “I would kill myself if I had the chance.” Previous research identified that successive BDI-II item nine ratings were associated with increased suicide risk, with a score of 1 being an optimal cutoff for predicting long-term risk for death by suicide, and a score of 2 being predictive of more proximal suicide attempts (
Data were collected from a total of 3,703 veterans who participated in CBT-D, IPT, or ACT-D for the treatment of depression. To focus this report on the effect of treatment on suicidal ideation, the sample was restricted to patients who endorsed a score of 1 or higher on item nine of the BDI-II, indicating the presence of suicidal ideation. The final analysis sample was comprised of 1,416 male veterans and 364 female veterans (21.1%), including 844 participants in CBT-D (21.7% female), 271 participants in IPT (22.8% female), and 665 patients in ACT-D (19.4% female). Notably, the prevalence of gender did not vary by treatment, χ2 (2) = 2.194, p = .34. Demographics for the sample are shown in Table 1.
Table 1Demographic Characteristics of Patients by Type of EBP Treatment Received
Treatment
N
Mean or %
SD
Total sample (N = 1,780)
Women
364
20.4%
Age
52.4
12.6
Caucasian
1364
77.8%
Education
3.26
1.13
IPT cohort (n = 271)
Women
58
21.4%
Age
55.7
13.2
Caucasian
210
79.8%
Education
3.17
1.13
CBT-D (n = 844)
Women
181
21.5%
Age
51.9
12.5
Caucasian
656
78.8%
Education
3.19
1.10
ACT-D (n = 723)
Women
125
18.8%
Age
51.8
12.4
Caucasian
498
75.8%
Education
3.38
1.15
Abbreviations: ACT-D, acceptance and commitment therapy for depression; CBT-D, cognitive–behavioral therapy for depression; EBP, evidence-based psychotherapies; IPT, interpersonal psychotherapy; SD, standard deviation.
Owing to missing data on race, denominators for calculating race proportions are n = 1,752 for total sample, n = 263 for IPT, n = 832 for, and n = 657 for ACT-D.
Evaluation of the primary outcome, suicidal ideation severity, was based on a set of generalized linear models using the GENMOD procedure available through SAS 9.4 (SAS Institute, Cary, NC). Generalized linear models are based on link functions that allow exponentially distributed variables to be modeled as linear. They also accommodate longitudinal (correlated) data using generalized estimating equations capable of incorporating a wide variety of time-dependent covariance structures. For the current analysis, the primary outcome measure, suicidal ideation severity, was analyzed using an assumed negative binomial distribution with a log link. This distribution, based on non-negative integer values, is similar to a Poisson distribution, but without the restrictive equivalence of mean and variance required of the latter. GENMOD procedures use all available pairs in estimating covariances and standard errors; under assumptions that missing data occur completely at random, these procedures have been shown to provide valid estimates. Covariance structure in all instances was assumed to be first-order autoregressive (i.e., correlations between repeated items were assumed to diminish exponentially with temporal distance). The ordinal outcome measure (BDI-II item nine) was regressed on models including various combinations of program, wave, and gender entered as both main effects and as a variety of interaction terms depending on the question; the baseline level of depression severity (operationalized as the total BDI-II score at baseline after removing item nine) was included in all models to account for initial severity of depression at the outset of treatment. In most instances, analyses were designed to test bivariate differences between conditions based on combinations of factors including gender, treatment, and wave. Specific contrasts were estimated using model-derived least square mean (LSM) estimates. Based on the substantial gender discrepancy in sample size and to avoid complexities of interpretation inherent in three-way interactions, relative efficacies of the three EBPs over time were assessed in gender-stratified subsamples. Because these analyses were primarily descriptive and exploratory, results are presented in unadjusted format.
Primary analyses were supplemented with a secondary responder analysis, with a treatment response defined as a zero score on the BDI-II item nine suicide item as assessed at the final session; in instances when the latter was missing, the response was based on mid-session data. Proportion of treatment responses was assessed across gender and program using standard chi square analyses to assess the magnitude of proportional differences between factors.
Results
Overall Impact of EBP Treatment on Suicidal Ideation
Across waves, suicidal ideation decreased sharply, both across and within genders: Summing over gender and treatment, the association between suicidal ideation and wave modeled as a main effect was statistically significant, χ2 (2) = 224.01, p = .0001, with ideation decreasing markedly at each measurement interval. Decreases were most pronounced from wave 1 to wave 2, where ideation scores decreased by 40.5% (model based), followed by a continued 16.9% decrease between wave 2 and wave 3, relative to the wave 1 score. In total, ideation scores decreased by 57.4% over the course of therapy (Table 2).
Table 2Suicidal Ideation Scores by Wave, Gender, and Treatment
Variable
Wave 1
Wave 2
Wave 3
n
Mean
SE
n
Mean
SE
n
Mean
SE
Sample mean
Total sample (N = 1,780)
1,780
1.10
0.01
1,318
0.68
0.02
1,089
0.53
0.02
Gender
Men (n = 1416)
1,416
1.11
0.01
1,017
0.72
0.02
852
0.56
0.02
Women (n = 364)
364
1.10
0.02
250
0.62
0.04
197
0.39
0.04
EBP
IPT (n = 271)
271
1.08
0.02
223
0.64
0.04
185
0.43
0.04
CBT-D (n = 844)
844
1.11
0.01
609
0.64
0.03
481
0.50
0.03
ACT-D (n = 665)
665
1.11
0.01
486
0.77
0.03
423
0.62
0.03
Model-based least squares mean estimate
Total sample (N = 1,780)
1,780
1.09
0.01
1,267
0.65
0.03
1,049
0.46
0.026
Gender
Men (n = 1416)
1,416
1.07
0.02
1,017
0.71
0.02
852
0.55
0.023
Women (n = 364)
364
1.08
0.02
250
0.60
0.05
197
0.40
0.042
EBP
IPT (n = 271)
271
1.07
0.02
220
0.60
0.06
182
0.37
0.051
CBT-D (n = 844)
844
1.09
0.01
579
0.62
0.03
458
0.47
0.034
ACT-D (n = 665)
665
1.12
0.02
468
0.74
0.04
409
0.57
0.041
Abbreviations: ACT-D, acceptance and commitment therapy for depression; CBT-D, cognitive–behavioral therapy for depression; EBP, evidence-based psychotherapies; IPT, interpersonal psychotherapy; SE, standard error.
Across treatments, a Gender × Time interaction was associated with reduced ideation among women veterans, χ2 (2) = 9.26, p = .001. LSM estimates between men and women favoring women were not statistically significant at wave 1 or wave 2 and emerged as statistically significant at wave 3 (LSM difference = -0.32, p = .005; Table 2; Figure 1).
Figure 1Plot of the Gender × Wave interaction (N = 1,780) showing reductions over time in suicidal ideation by gender.
Impact of Type of EBP Treatment Received on Suicidal Ideation
Across gender, a Treatment × Time interaction did not reach statistical significance, χ2 (4) = 9.21, p = .056. Within all three treatments, suicidal ideation decreased from wave 1 to wave 3 (Table 2). Comparisons between treatments at each wave were not statistically significant.
Gender by Treatment Effects on Suicidal Ideation
To preclude interpreting three-way interactions, analyses addressing gender-related treatment differences were conducted on gender-stratified subsets of the study cohort. Within the male cohort, the association between suicidal ideation and the difference in Wave x Treatment interaction was statistically significant, (χ2 [4] = 16.82, p = .002). Using LSM contrasts to explore wave-to-wave differences, suicidal ideation decreased over the course of therapy irrespective of treatment modality, that is, summing across treatments (Table 3), with the greatest decreases in ideation taking place moving from wave 1 to 2; changes at the treatment level were similar (Table 3). Exploring the magnitude of symptom levels between treatments by wave revealed programmatic differences. Decreases in the LSM-estimated ideation symptom levels were decreased in male patients receiving treatment in either the IPT or CBT-D modalities relative to levels in ACT-treated patients (Table 3; Figure 2) at both waves 2 and 3. Presented as percent change, based on the change in the LSM estimates for suicidal ideation at wave 1 and 3, male patients in IPT and CBT-D experienced decreases in ideation severity over the course of therapy of 61.1% and 52.6%, respectively, relative to a decrease of 41.1% for ACT-treated veterans.
Table 3Mean Differences by Wave and Treatment and Gender
Test
Women
Men
Diff
SE
z-Val
Pr > |z|
Diff
SE
z-Val
Pr > |z|
Differences in waves by treatment
All
W1 vs W2
0.59
0.08
7.50
<.001
0.49
0.06
7.71
<.001
W1 vs W3
1.01
0.11
9.52
<.001
0.70
0.04
17.24
<.001
W2 vs W3
0.42
0.11
3.94
<.001
0.26
0.04
7.04
<.001
IPT
W1 vs W2
0.66
0.19
3.41
<.001
0.49
0.06
7.71
<.001
W1 vs W3
1.29
0.26
4.99
<.001
0.83
0.09
8.79
<.001
W2 vs W3
0.63
0.27
2.30
0.02
0.34
0.08
4.02
<.001
CBT-D
W1 vs W2
0.61
0.09
6.52
<.001
0.52
0.04
11.95
<.001
W1 vs W3
0.94
0.13
7.13
<.001
0.75
0.06
12.73
<.001
W2 vs W3
0.33
0.12
2.67
0.01
0.23
0.05
4.12
<.001
ACT
W1 vs W2
0.50
0.10
5.16
<.001
0.32
0.04
8.07
<.001
W1 vs W3
0.81
0.13
6.10
<.001
0.53
0.05
10.49
<.001
W2 vs W3
0.31
0.12
2.62
0.01
0.21
0.04
4.69
<.001
Differences in treatments by wave
Wave 1
IPT vs CBT
0.00
0.04
0.06
0.95
−0.03
0.02
−1.38
0.17
IPT vs ACT
−0.08
0.05
−1.47
0.14
−0.01
0.02
−0.54
0.59
CBT vs ACT
−0.08
0.04
−1.90
0.06
0.02
0.02
1.07
0.29
Wave 2
IPT vs CBT
−0.05
0.23
−0.22
0.82
0.00
0.08
−0.06
0.95
IPT vs ACT
−0.24
0.22
−1.06
0.29
−0.18
0.08
−2.40
0.02
CBT vs ACT
−0.19
0.13
−1.41
0.16
−0.18
0.06
−2.93
0.00
Wave 3
IPT vs CBT
−0.34
0.29
−1.18
0.21
−0.12
0.11
−1.03
0.30
IPT vs ACT
−0.56
0.29
−1.92
0.06
−0.32
0.11
−2.86
0.00
CBT vs ACT
−0.21
0.19
−1.14
0.26
−0.20
0.08
−2.50
0.01
Abbreviations: ACT-D, acceptance and commitment therapy for depression; CBT-D, cognitive–behavioral therapy for depression; IPT, interpersonal psychotherapy; SE, standard error.
Repeating analyses within the female stratum, in contradistinction with results reported above for male patients, the Treatment × Wave interaction was not statistically significant, χ2 (4) = 3.41, p = .492. As with male veterans, wave-to-wave ideation severity based on LSM estimates decreased markedly over the course of therapy both across (Table 3) and within treatments (Table 3; Figure 2). Again mirroring effects observed among men, decreases in symptomology were most pronounced between the outset of therapy and its midpoint. Comparable percentage-wise decreases in suicidal ideation severity over the course of therapy for IPT, CBT-D, and ACT were 72.4%, 61.0%, and 55.4%, respectively.
Responder Analyses
Results of responder analyses were generally consistent with results of primary analyses. Rates of response among men differed across programs (Table 4). In subsequent bivariate comparisons between programs, the proportions of male responders in both IPT, χ2 = 5.63; p = .018, and CBT-D, χ2 = 7.82; p = .005, exceeded rates observed for veterans treated in ACT-D (Table 4); proportional differences in response levels between CBT-D and ACT-D programs did not differ significantly, χ2 = 0.07; p = .796. Among women, the rate of response did not differ significantly between programs. The proportions of female responders across programs were slightly higher relative to men but similar in pattern; however, as indicated by the omnibus test, above, χ2 = 3.798; p = .150, bivariate tests of proportional program differences between female responders failed to attain statistical significance. The lack of statistical significance between programs in part may reflect a type II error related to limited sample size. The proportion of responders among female participants across all programs exceeded the proportion of male responders across programs, although this difference was not statistically significant (54.7%–48.0%; χ2 = 3.73; p = .053). Within programs, the response rates for women treated within IPT exceeded levels observed in male participants by more than 14%, although the difference was not statistically significant (χ2 = 2.91; p = .088). Gender differences between veterans in the remaining two programs also did not differ significantly (Table 4; CBT-D: χ2 = 0.36; p = .548; ACT-D: χ2 = 1.39; p = .238).
Table 4Treatment Responders
Program
Men
Women
No Response
Response
Totals
No Response
Response
Totals
Program by gender
IPT
85 (47.5%)
94 (52.5%)
179
15 (33.3%)
30 (66.7%)
45
CBT-D
229 (48.6%)
242 (51.4%)
471
57 (45.6%)
68 (54.4%)
125
ACT-D
233 (58.1%)
168 (41.9%)
401
44 (51.2%)
42 (48.5%)
86
Totals
547
504
1051
116
140
256
Contrast
χ2
p
χ2
p
Overall
9.6043
.0082
3.7979
.1497
IPT vs CBT
0.07
.7961
2.04
.1533
IPT vs ACT
5.63
.0176
3.79
.0514
CBT vs ACT
7.82
.0052
0.63
.4267
IPT
CBT-D
ACT-D
No Response
Response
Totals
No Response
Response
Totals
No Response
Response
Totals
Gender by program
Men
85 (47.5%)
94 (52.5%)
179
229 (48.6%)
242 (51.4%)
471
233 (58.1%)
168 (41.9%)
401
Women
15 (33.3%)
30 (66.7%)
45
57 (45.6%)
68 (54.4%)
125
44 (51.2%)
42 (48.8%)
86
Totals
100
124
224
286
310
596
277
210
487
Contrast
χ2
p
χ2
p
χ2
p
Men vs women
2.91
.0878
0.36
.548
1.39
.2382
Abbreviations: ACT-D, acceptance and commitment therapy for depression; CBT-D, cognitive–behavioral therapy for depression; IPT, interpersonal psychotherapy.
Associations between attrition at both waves 2 and 3 and veterans’ BDI-II item nine scores at the immediately preceding wave were nonsignificant, as were tests of wave 3 attrition based on wave 1–wave 2 difference scores. The tests were repeated using the full BDI-II measure as well as a set of available demographic covariates. In all instances, results were not statistically significant. Owing to the absence of any apparent systematic associations within variables present in the data, missing data were assumed to occur at random.
Discussion
Our findings in a large clinical sample of veterans participating in EBPs for depression found that CBT-D, IPT, and ACT-D each accomplished reductions in suicidal ideation, with decreases evident by treatment midpoint. These findings add to the emerging literature examining the impact of depression treatments on suicidal ideation. Although published findings on putative gender differences in treatment response for interventions targeting suicidal ideation are conflicting, these results describe greater decreases in the severity of suicidal ideation among women relative to men. These findings demonstrate that interventions with a strong evidence base for treating depression are concurrently effective in decreasing suicidal ideation among veterans and, as suggested, may be particularly effective in decreasing suicidal ideation in women veterans. Although women veterans showed sharper declines in suicidal ideation severity with treatment, responder analyses revealed that women were not more likely to evidence elimination of ideation relative to men.
Owing to the limited available data comparing the effectiveness of ACT-D to other EBPs for depression in decreasing suicidal ideation (
), our finding that ACT-D was less effective in this analysis is novel. Furthermore, gender-stratified analyses demonstrated that this effect was largely driven by the poorer response of male veterans to ACT-D relative to IPT and CBT-D. It may follow that IPT and CBT-D are preferable treatments for male veterans reporting suicidal ideation; however, as previously suggested in a review (
), unequivocal evaluation of gender differences in treatment of depression and suicidality would require rigorous controlled trials with appropriate randomization of men and women across treatments (including ACT-D). Supporting this, a meta-analysis of ACT-D studies noted that studies with greater proportions of men were associated with larger effect sizes (
), thus indicating that design artifacts in outcomes related to depression warrant subgroup analysis by gender.
Limitations
The generalizability of these program evaluation findings is increased by the provision of the therapies by clinicians receiving training, the relatively few exclusions, and the use of routine clinical referral pathways to identify patients. Conversely, results are limited by the lack of experimental assignment to treatments and the unavailability of data concerning concurrent treatments and clinician factors that might have influenced changes in clinical outcomes. It must also be noted that clinicians received expert consultation during treatment provision, which does not reflect routine clinical practice. Additionally, although ACT-D was designed as a 12-session treatment, IPT and CBT-D protocols allowed up to 16 sessions of treatment, a difference in dose that may account for the greater decreases in symptoms experienced by men in IPT and CBT-D relative to ACT-D by the wave 3 measurement. Arguing against this supposition, it is noteworthy that differences were observed between ACT-D and both IPT and CBT-D among men by wave 2, indicating that differences were evident before differences in treatment dose. Veterans lost to follow-up and resulting missing data also present limitations, and conclusions cannot be drawn about the status of veterans without data at all waves.
These results are further limited by the single-item measure of suicidal ideation used in the present analyses. Although BDI-II item nine has been shown to be a valid measure for assessing suicidal ideation (
), there are more optimal measures, and future work in this area would benefit from the use of a longer, more reliable assessment of suicidal ideation. The present results also cannot specify that any reduction in suicidal ideation evidenced through participation in EBPs for depression results in subsequent reduction in risk of suicidal behavior; thus, future research would benefit from longitudinal assessment of suicidal ideation and attempts. Finally, these conclusions are limited by a lack of data to characterize the factors associated with suicide risk beyond suicidal ideation, most notably any history of suicide attempts or psychiatric hospitalizations.
Implications for Practice and/or Policy
The results of this report support the effectiveness of EBPs for depression for decreasing suicidal ideation in veterans. They also support the inclusion of veterans experiencing suicidal ideation as a part of standard VHA clinician training aimed at promoting the generalization of clinician skills and increasing veterans’ access to EBPs for depression. We also note that these results provide particularly strong support for the effectiveness of EBPs for depression for reducing suicidal ideation in women veterans. Owing to the rising rate of suicide in women veterans (
U.S. Department of Veterans Affairs (VA), Office of Mental Health and Suicide Prevention (OMHSP) Facts About Suicide Among Women Veterans: August 2017.
Lifetime suicidal behaviors and career characteristics among U.S. Army soldiers: Results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS).
Suicide and Life-Threatening Behavior.2017; 48: 230-250
Lifetime suicidal behaviors and career characteristics among U.S. Army soldiers: Results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS).
Suicide and Life-Threatening Behavior.2017; 48: 230-250
) with suicide among women, these findings provide a rationale for facilitating access to EBPs for depression in women veterans with depression. In addition, the relative advantages of IPT and CBT-D for reducing suicidal ideation in male veterans would perhaps argue for the initiation of strategies designed to tailor treatments for male veterans based on levels of pretreatment suicidal ideation and/or further analysis of gender differences based on more rigorous clinical trial formats as an entrée to matching the strategies suggested elsewhere in this article. Clinical guidance should also inform the role of EBPs for depression in reducing risk of suicide in the context of suicide-focused psychotherapies and other interventions employed in efforts to prevent suicide.
Conclusions
This program evaluation project provides the first comparison of the clinical effectiveness of three EBPs for depression for reducing suicidal ideation across male and female veterans. This analysis revealed that 1) all three EBPs for depression that were examined resulted in statistically significant reductions in suicidal ideation and reductions were evident by the midpoint of treatment; 2) women veterans reported greater decreases in suicidal ideation severity across all three EBP treatments; and 3) decreases in suicidal ideation were greater among male veterans treated in IPT or CBT-D relative to ACT-D.
Acknowledgments
The authors express their gratitude to the project staff and clinicians who collected data as part of the clinical care they provided to their patients. The authors thank the veterans who sought treatment for depression and provided data for this evaluation project.
Acceptance and commitment therapy and contextual behavioral science: Examining the progress of a distinctive model of behavioral and cognitive therapy.
From the laboratory to the therapy room: National dissemination and implementation of evidence-based psychotherapies in the US Department of Veterans Affairs Health Care System.
National dissemination of cognitive behavioral therapy for depression in the Department of Veterans Affairs health care system: Therapist and patient-level outcomes.
Journal of Consulting and Clinical Psychology.2012; 80: 707-718
Differences in the effectiveness of psychosocial interventions for suicidal ideation and behaviour in women and men: A systematic review of randomised controlled trials.
Lifetime suicidal behaviors and career characteristics among U.S. Army soldiers: Results from the Army Study to Assess Risk and Resilience in Servicemembers (Army STARRS).
Suicide and Life-Threatening Behavior.2017; 48: 230-250
Training in and implementation of acceptance and commitment therapy for depression in the Veterans Health Administration: Therapist and patient outcomes.
Mandy J. Kumpula, PhD, is Program Evaluator for the National Evidence-Based Psychotherapy Program within VA Office of Mental Health and Suicide Prevention and VA Mid-Atlantic Mental Illness Research, Education, and Clinical Center. Interests include mechanisms, effectiveness, and implementation of evidence-based psychotherapies.
H. Ryan Wagner, PhD, is a Statistician and Data Scientist at the VA Mid-Atlantic Mental Illness Research, Education, and Clinical Center and Associate Professor in the Department of Psychiatry and Behavioral Sciences at Duke University School of Medicine.
Eric A. Dedert, PhD, is a Psychologist with the Durham VA Health Care System and VA Mid-Atlantic Mental Illness Research, Education, and Clinical Center. He is an Assistant Professor at Duke University. His interests include the clinical effectiveness of evidence-based treatments.
Chris M. Crowe, PhD, serves as National Mental Health Director of Psychotherapy for the Veteran Healthcare Administration Office of Mental Health and Suicide Prevention.
Kristine T. Day, PhD, is the National Evidence-Based Psychotherapy Program Manager within the Department of Veterans Affairs. Within this role, she oversees implementation, maintenance, monitoring, and evaluation of evidence based psychotherapy training programs within the VA.
Kristin Powell, PhD, serves as National EBP Implementation Program Manager, VA Office of Mental Health and Suicide Prevention and VA Mid-Atlantic Mental Illness Research, Education, and Clinical Center. Professional interests include EBP training, implementation, and evaluation.
Wendy H. Batdorf, PhD, is a clinical psychologist serving as the VA Evidence-Based Psychotherapy (EBP) Training Coordinator for the Cognitive Behavioral Therapy for Depression and Acceptance and Commitment Therapy for Depression Programs. Her professional interests include EBP training, depression, and geropsychology.
Hani Shabana, PhD, is a clinical psychologist serving as the VA Evidence-Based Psychotherapy (EBP) Training Coordinator for the Interpersonal Psychotherapy for Depression and Cognitive Behavioral Therapy for Chronic Pain Programs.
Ellie Kim, BA, is a psychology technician for the National Evidence-Based Psychotherapy Program within the Department of Veterans Affairs and the VA Mid-Atlantic Mental Illness Research, Education, and Clinical Center.
Nathan A. Kimbrel, PhD, is Research Psychologist and Assistant Director for Implementation Science and Program Evaluation at the Durham VA Medical Center and VA Mid-Atlantic Mental Illness Research, Education, and Clinical Center. He is an Associate Professor at Duke University.
Article info
Publication history
Accepted:
April 19,
2019
Received in revised form:
April 15,
2019
Received:
August 27,
2018
Footnotes
Supported by the National Evidence-Based Psychotherapy Training Program, Department of Veterans Affairs Office of Mental Health and Suicide Prevention.
Conflicts of Interest: The authors have no conflicts of interest to declare. The views expressed in this paper are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs or the United States government.
30503-6&id=gr1.jpg){kind=link}
30503-6&id=gr2.jpg){kind=link}