Psychiatry Research 217 (2014) 147–153
Contents lists available at ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Mental health utilization of new-to-care Iraq and Afghanistan Veterans following suicidal ideation assessment Lauren M. Denneson a,b,n, Kathryn Corson a,b, Drew A. Helmer c,d, Matthew J. Bair e, Steven K. Dobscha a,b a
Portland Center to Improve Veteran Involvement in Care (CIVIC), Portland VA Medical Center, Portland, OR, United States Department of Psychiatry, Oregon Health & Science University, Portland, OR, United States c War-Related Illness and Injury Study Center, VA New Jersey Health Care System, East Orange, NJ, United States d Department of Medicine and Dentistry, New Jersey Medical School, Newark, NJ, United States e Richard L. Roudebush Veteran Affairs Medical Center, Indianapolis, IN, United States b
art ic l e i nf o
a b s t r a c t
Article history: Received 25 September 2013 Received in revised form 8 February 2014 Accepted 14 March 2014 Available online 22 March 2014
We evaluated the impact of brief structured suicidal ideation (SI) assessments on mental health care among new-to-care Operations Enduring Freedom and Iraqi Freedom (OEF/OIF) veterans. National datasets provided military, demographic, and clinical information. For all new-to-care OEF/OIF veterans administered depression screens (PHQ-2: Patient Health Questionnaire-2) and structured SI assessments in primary care or ambulatory mental health settings of three Veterans Affairs (VA) Medical Centers between April 2008 and September 2009 (N ¼465), generalized estimating equations were used to examine associations between SI and number of subsequent-year specialty mental health visits and antidepressant prescriptions. Approximately one-third of the veterans reported SI. In multivariate models, PTSD and anxiety diagnoses, severe depression symptoms, being married, and SI assessment by a mental health clinician were associated with more mental health visits in the subsequent year. Depression, PTSD, and anxiety diagnoses, and SI assessment by a mental health clinician were associated with receiving antidepressants. Presence of SI did not significantly affect subsequent year mental health utilization when adjusting for diagnostic and clinician variables, but inaugural visits involving mental health clinicians were consistently associated with subsequent mental health care. Published by Elsevier Ireland Ltd.
Keywords: Depression Risk assessment Suicide Delivery of health care
1. Introduction Suicidal ideation (SI), defined as, “thoughts of engaging in suiciderelated behavior,” (Brenner et al., 2011; Crosby et al., 2011) is among one of the best predictors of suicide attempts (Nielsen et al., 1990; Szanto et al., 2003; Mann et al., 2008; Britton et al., 2012). Population estimates of SI vary considerably due to methods and sample selection, yet there is evidence that rates are high among Veterans Affairs (VA) healthcare-seeking veterans with mental health symptoms; up to 46% of Operation Enduring Freedom and Operation Iraqi Freedom (OEF/OIF) veterans referred for VA mental health services (Jakupcak et al., 2009) and 32% of OEF/OIF veterans with depression symptoms in VA care report SI (Corson et al., 2013). The VA, which operates the largest integrated healthcare system in the United States (Department of Veterans Affairs, 2009), has designated assessment of SI among veterans with positive depression
n Correspondence to: Portland VA Medical Center, P.O. Box 1034 (R&D66), Portland, OR 97207, United States. Tel: þ 1 503 220-8262x57351. E-mail address: [email protected] (L.M. Denneson).
http://dx.doi.org/10.1016/j.psychres.2014.03.017 0165-1781/Published by Elsevier Ireland Ltd.
or PTSD screens as a national performance goal since 2008. The main goal of this initiative is to identify at-risk veterans to engage them in appropriate mental health treatment. VA recommendations for such care include pharmacological treatment of acute symptoms, such as anxiety, insomnia, and psychosis as well as careful evaluation, treatment, and monitoring of any underlying mental health disorders, especially depression (Department of Veterans Affairs, 2008). However, the relationship between suicide screening, or risk detection, programs and treatment utilization, is not well understood (Gaynes et al., 2004; O'Connor et al., 2013). In a community setting, one study detected small associations between SI and subsequent receipt of further assessment, specialty mental health referral, and psychotropic medication (Bauer et al., 2013). One prior study in VA primary care revealed those with SI attended more primary care visits than those without SI (Lish et al., 1996). Given the significant time and financial burden routinized screening or assessment programs place on a healthcare system, it is imperative we understand the relative contribution of these structured SI assessments in meeting the desired objectives. In this retrospective study, we sought to evaluate the impact of VA's structured SI assessment procedures on veteran engagement in
148
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
mental health treatment. Among new-to-care OEF/OIF veterans who screened positive for depression in VA outpatient settings, we identified the correlates of subsequent-year VA mental health care utilization, with particular interest in examining the relationship between SI and utilization. Based on recommended follow-up for SI, described above (Department of Veterans Affairs, 2008), we hypothesized that veterans with positive SI assessments would have an increased number of subsequent mental health specialty visits and/or psychiatric medication prescriptions, as compared to veterans with negative SI assessments. 2. Methods We conducted a multi-site study of the SI assessment process and 1-year outcomes among OEF/OIF veterans screened for depression between April 1, 2008 and September 30, 2009 at three VA Medical Centers (VAMCs), representing the Northwestern, Southwestern, and Northeastern census regions of the US. Methods have been reported in detail elsewhere (Corson et al., 2013; Dobscha et al., 2013). 2.1. Settings Each of the VAMCs has metropolitan and rural-based primary care clinics, is closely affiliated with a local university, and provides a full range of patient care services. All VAMCs utilize the VA's electronic Computerized Patient Record System (CPRS). A clinical reminder system embedded within CPRS facilitates the routine administration of screens and assessment tools for a variety of conditions and disorders (Department of Veterans Affairs, 2007), including depression and SI. Items are administered by clinicians or clinic staff, and responses, with results, are listed in the electronic progress note. Positive results generate new reminders to prompt clinician follow-up. 2.2. Data sources The OEF/OIF Roster File, maintained by the Department of Defense and the Defense Manpower Data Center, contains demographic and service-related data on all OEF/OIF veterans discharged after September 11, 2001 who enrolled in or accessed VHA services after discharge. We used social security numbers to match the OEF/OIF Roster to VA Decision Support System data. Methods related to collection of depression screen and SI assessment results and inter-rater reliability are described in recent manuscripts (Corson et al., 2013; Dobscha et al., 2013). 2.3. Sample We limited our sample to new-to-care patients (no VA healthcare visits during the 5 years prior) because depression and suicide risk assessment results may have less effect on continuity of care than on initiation of care, especially among established patients with diagnosed psychiatric disorders. Also, at some VAMCs, patients currently engaged in mental health may be exempt from mental health screens. At the three medical centers during the study period, 600 new-to-care veterans screened positive for depression in a primary care or mental health outpatient setting. Within this group, 465 (77.5%) had a same-day structured SI assessment using one of two widely-used structured SI assessment tools (used in 98.2% of all structured SI assessments completed during the study period (Dobscha et al., 2013)). This group constituted the sample for the current study, with the date of the positive depression screen and SI assessment representing the patient's index date. The majority of the sample was under 34 years of age (80.2%), nonHispanic white (60.4%), and male (87.5%). The study was approved by the Institutional Review Boards of each participating site. Patient data presented here were obtained as part of routine clinical practice and thus waivers of informed consent were granted.
dead or of hurting themselves in some way. Any response 40 (not at all) is scored as a positive result (Kroenke et al., 2001; Schulberg et al., 2005). The VA “Pocket Card” Risk Assessment (VA-PCRA) (Department of Veterans Affairs, 2008) is a fouritem assessment, and a positive result requires endorsing, “thoughts of taking your life.” Reliability and validity of these instruments for SI detection have not been established, though two recent studies support concurrent validity of the PHQ-9 9th item as an SI screening tool (Uebelacker et al., 2011; Bauer et al., 2013). Veterans were administered one or the other (and sometimes both) of these assessments, depending on the facility, or clinician or staff preference; some veterans received multiple assessments on one day. Veterans with any positive SI assessment result on the same day as the positive PHQ-2 were categorized into the positive SI group; veterans with only negative results were categorized into the negative SI group. Demographic characteristics: age, sex, race/ethnicity, marital status, zip code of residence, and service connection status were obtained from the VA Decision Support System. To designate veteran rurality, we used the VA's Office of Rural Health’s classifications (Urban, Rural, or Highly Rural), which are based on geocoding and utilize a zip-code-based crosswalk system (West et al., 2010). Race/ ethnicity data were frequently missing from Decision Support System databases, so race/ethnicity identification was supplemented by OEF/OIF roster data. Healthcare data: utilization, diagnosis, and prescription data were obtained from DSS for 1 year, beginning with the patient's index date, at the VAMC of depression screening. Depressive disorder, PTSD, substance use disorder (SUD), and anxiety disorder diagnoses made on the index date were identified using International Classification of Diseases, Clinical Modification (ICD-9-CM) codes (full list available from corresponding author). 2.5. Analyses All analyses were completed using PASW Statistics (SPSS version 18). Consistent with prior studies (Corson et al., 2013; Dobscha et al., 2013), we collapsed age into three groups ( o24, 25–34, and Z 35), marital status into three groups (single/ never married, married, and divorced/widowed), and education into three groups (less than high school diploma, high school/some college, and college degree or higher). We dichotomized race/ethnicity into non-Hispanic white vs. other and rurality was dichotomized by combining the highly rural and rural categories. To denote a high PHQ-2 score (severe depression symptoms), PHQ-2 scores were dichotomized at PHQ-2 Z 5 (vs. PHQ-2 o 5); this score cutoff requires the patient to report experiencing both anhedonia and depression, with at least one of the symptoms occurring every day during the previous 2 weeks. For this study, the number of specialty mental health visits were summed for each patient over the course of the year follow-up period, and were ordinally categorized into nine levels (0 to Z 8) based on the positive skew and platykurtosis of the distribution in this sample. Univariate analyses (χ2 and Mann–Whitney U) were used to test demographic, clinical, and utilization differences between the positive SI and negative SI groups. Generalized estimating equation (GEE) models were then constructed to identify multivariate correlates of SI, and to identify multivariate correlates of subsequent year utilization, while controlling for potential non-independence of observations with sites. For specialty mental health visits, GEE ordinal logistic models were used. As in linear regression, log-odds regression coefficients show the change expected in the dependent variable for a one unit increase in the predictor, wherein change is defined as moving from one category/level to the next; one equation is estimated across all outcome variable levels and odds ratios are calculated by exponentiating the regression coefficients (Tabachnick and Fidell, 2001). We included a priori in each model sex, age, race/ethnicity, rurality, marital status, SI assessment by a mental health clinician (yes/no; includes non-licensed care staff), and SI assessment positive (yes/no). We also included PHQ-2 scores, and depressive disorder and PTSD diagnoses made on the index date; we previously reported on the association of these variables with receipt of SI assessment and with SI assessment result (Corson et al., 2013; Dobscha et al., 2013). Additional diagnosis variables were included if associated with one or more outcome variables at p r 0.10.
3. Results 2.4. Measures
3.1. Sample characteristics and differences between groups Depression: annual depression screening is conducted using the 9-item or 2-item Patient Health Questionnaire (PHQ-9 and PHQ-2, respectively). Both the PHQ-9 and PHQ-2 are well-validated and widely-used instruments for detecting and diagnosing depression and measuring depression severity (Kroenke et al., 2003). The PHQ-2 consists of the depressed mood and anhedonia items of the PHQ-9, and response options range from 0 (not at all) to 3 (nearly every day). A PHQ-2 score of 3 is the cut-point for a positive depression screen, with a sensitivity of 83% and a specificity of 92% for major depression (Kroenke et al., 2003). Suicidal ideation: at the time of the study, two brief, structured tools were predominantly used to detect suicidal ideation at the three medical centers. The 9th item of the PHQ-9 asks patients if they have had thoughts of being better off
Of the 465 OEF/OIF veterans in the sample, 147 (32%) had positive SI assessment results. In univariate comparisons of the positive SI and negative SI groups, the two groups did not differ on demographic characteristics, but did differ on several clinical variables (Table 1). Veterans in the positive SI group were more likely to have high PHQ-2 scores, be assessed for SI by a mental health clinician, and be diagnosed with a depressive disorder or a substance use disorder. In multivariate analyses, only depressive
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
149
Table 1 Demographic characteristics and VA healthcare utilization of new-to-care OEF/OIF veterans with a positive depression screena and same-day structured SI assessment, by SI assessment result. Characteristic
All patients (n ¼465)
Negative SI (n ¼318)
Site VAMC 1 VAMC 2 VAMC 3
107 289 69
(23.0) (62.2) (14.8)
65 204 49
(20.4) (64.2) (15.4)
42 85 20
(28.6) (57.8) (13.6)
Age r 24 25–34 Z 35 Male
110 263 92 407
(23.7) (56.6) (19.8) (87.5)
75 174 69 279
(23.6) (54.7) (21.7) (87.7)
35 89 23 128
(23.8) (60.5) (15.6) (87.1)
0.84
Non-Hispanic whiteb (8 missing)
276
(60.4)
196
(62.8)
80
(55.2)
0.12
Marital status (5 missing) Single/never married Married Divorced/widowed
195 173 92
(42.4) (37.6) (20.0)
138 118 57
(44.1) (37.7) (18.2)
57 55 35
(38.8) (37.4) (23.8)
Education (2 missing) Less than high school diploma High school/some college College degree or higher
72 363 28
(15.6) (78.4) (6.0)
48 246 23
(15.1) (77.6) (7.3)
24 117 5
(16.4) (80.1) (3.4)
Rural designation (4 missing) Urban Rural
358 103
(77.7) (22.3)
241 74
(76.5) (23.5)
117 29
(80.1) (19.9)
Service connectedc
188
(40.4)
126
(39.6)
62
(42.2)
0.60
Positive SI (n¼ 147)
P 0.15
0.29
a
0.33
0.27
0.38
237
(51.0)
140
(44.0)
97
(66.0)
o 0.001
SI assessment by a mental health clinician
173
(37.2)
103
(32.4)
70
(47.6)
o 0.01
Clinical diagnosesd Depressive disorder Post-traumatic stress disorder Substance use disorder Anxiety
254 224 79 94
(54.6) (48.2) (17.0) (20.2)
151 144 45 59
(47.5) (45.3) (14.2) (18.6)
103 80 34 35
(70.1) (54.4) (23.1) (23.8)
o 0.001 0.07 0.02 0.19
Mental health visitse, median (IQR) Antidepressant prescription
3.00 248
(7.00) (53.3)
3.00 156
(6.00) (49.1)
4.00 92
(6.00) (62.6)
o 0.01 o 0.01 0.98
PHQ-2 Z5
Benzodiazepine prescription Mental health visit þ antidepressant prescription Psychiatric inpatient stay
44 229 24
(9.5) (49.2) (5.2)
30 142 9
(9.4)
14
(9.5)
(44.7)
87
(59.2)
o 0.01
(2.8)
15
(10.2)
0.001
All results are presented as n (%), unless otherwise noted. OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; SI¼ suicidal ideation; VAMC ¼Veterans Affairs Medical Center; PHQ-2 ¼Patient Health Questionnaire depression screen; SI ¼suicidal ideation; IQR ¼interquartile range. Mental health visit þ antidepressant¼ any specialty mental health visit completed subsequent to initial assessment date and any antidepressant prescription filled. a
PHQ-2 43 is positive screen, range 3–6. Non-Hispanic white vs. other. c Service connected veterans receive compensation in the form of disability payments and/or treatment for the service-related disability. d International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. e Mental health visits, range 0–8þ ; specialty mental health visits attended subsequent to initial assessment date. b
disorder diagnoses and high PHQ-2 scores remained associated with suicidal ideation (Table 2). 3.2. Subsequent year utilization In the year following SI assessment, 377 (81.1%) had at least one mental health visit (median¼3.00, IQR¼7.00), 248 (53.3%) filled an antidepressant prescription, and 24 (5.2%) had a psychiatric inpatient stay. In univariate analyses, the groups differed significantly on number of specialty mental health visits, antidepressant prescriptions filled, any mental health visit and any antidepressant prescription fill, and any psychiatric inpatient stays (Table 1). Benzodiazepine prescription receipt did not differ between the two groups, and the small number of veterans completing inpatient stays precluded using benzodiazepine prescriptions or psychiatric inpatient stays in further analyses. Multivariate correlates of utilization outcomes are presented in Table 3. The odds of attending a greater number of mental health
visits were higher for married veterans (AOR¼ 1.49, p ¼0.04), veterans diagnosed with PTSD (AOR ¼1.70, p o0.01) or anxiety (AOR¼ 1.64, p ¼0.03), for veterans with high PHQ-2 scores (AOR¼ 1.41, p ¼0.049), and for those who were assessed for SI by a mental health clinician (AOR ¼1.72, p ¼0.01). The odds of antidepressant prescription receipt were greater for veterans diagnosed with PTSD (AOR¼2.38, p o0.001), anxiety (AOR ¼1.73, p¼ 0.046) or depressive disorders (AOR ¼2.02, p o0.01), and for veterans who were assessed for SI by a mental health clinician (AOR¼ 1.85, p ¼0.01). The odds of both attending a mental health visit and receiving an antidepressant prescription were greater for veterans who were married (AOR¼1.94, p ¼0.01), assessed for SI by a mental health clinician (AOR¼ 2.39, p o0.001), and were diagnosed with a depressive disorder (AOR¼1.71, p¼ 0.01), PTSD (AOR¼ 2.37, po 0.001), or anxiety (AOR ¼1.81, p ¼0.03). In all adjusted models, SI was not predictive of mental health care utilization. In post hoc analyses, we removed PHQ-2 score from the models to investigate whether the relationships between SI
150
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
Table 2 Multivariate correlates of suicidal ideation in new-to-care OEF/OIF veterans with a positive depression screen and structured SI assessment. AOR (95% CI)
P
– 1.06 (0.65–1.74) 0.71 (0.37–1.38)
– 0.82 0.31
0.81 (0.44–1.50)
0.50
Non-Hispanic white
0.66 (0.43–1.03)
0.07
Marital status Never married (reference) Married Divorced/widowed
– 1.37 (0.85–2.19) 1.42 (0.82–2.46)
– 0.20 0.21
Rural
0.96 (0.58–1.60)
0.87
PHQ-2 scores Z 5b
2.17 (1.43–3.30)
o 0.001
Clinical diagnosesc Depressive Disorder Substance Use Disorder Post-traumatic Stress Anxiety
2.16 (1.40–3.34) 1.41 (0.82–2.42) 1.10 (0.72–1.69) 1.18 (0.70–1.98)
0.001 0.21 0.67 0.53
SI assessment by mental health cliniciand
1.46 (0.90–2.36)
0.12
Age 18–24 (reference) 25–34 35þ Sex, female a
Odds ratios adjusted for non-independence of observations within sites (AORs) with 95% confidence intervals (CI); analyses included only those with complete data (n ¼449). OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; PHQ-2¼ Patient Health Questionnaire depression screen; SI ¼suicidal ideation.
services following SI endorsement on the PHQ-9 9th item in a primary care setting. Results supported an association between SI endorsement and 12-week mental health follow-up patterns; patients with SI were more likely to receive referrals for specialty mental health, more likely to receive psychotropic medication, and received earlier assessment by a care manager (Bauer et al., 2013). In contrast, we did not find an independent relationship between detection of SI and mental health utilization, although comparison is complicated by several methodological differences. We discovered that the relationship between SI and subsequent year specialty mental health visits was apparent only when PHQ-2 scores were not included in the model. This is not surprising, given that severe depression symptoms bring greater distress and have been repeatedly linked with SI (Mann et al., 2005; Pietrzak et al., 2010; Lemaire and Graham, 2011; Corson et al., 2013). As noted in Bauer et al. (2013), where endorsing SI on the PHQ-9 frequently co-occurred with severe depression symptoms on the PHQ-2 (Bauer et al., 2013), depression screening itself may identify a significant portion of those with SI. However, in a previous study of correlates of receiving an SI assessment through the VA's SI assessment initiative (Dobscha et al., 2013) we found that veterans with high PHQ-2 scores were more likely to receive an SI assessment than those with slightly lower scores. This systematic bias may have artificially inflated the relationship detected between high PHQ-2 scores and SI; the multivariate models of mental health utilization that control for high PHQ-2 scores help account for this confounding relationship.
a
Non-Hispanic white vs. other. PHQ-2 Z5 denotes high PHQ score. International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. d As opposed to primary care. b c
and the three utilization outcomes changed when PHQ-2 score are not taken into account. Results indicated that the odds of completing more specialty mental health care visits were greater for veterans with SI (AOR¼ 1.48, p ¼0.04) when not adjusting for PHQ-2 score (Table 4).
4. Discussion In this study of new-to-care OEF/OIF veterans who screened positive for depression and were assessed for SI using brief, structured assessments, we note two key observations. First, detection and presence of SI is not associated with subsequent mental health care utilization when accounting for severity of depression symptoms. Second, when a veteran's inaugural visit to VA healthcare includes a mental health clinician, the veteran is more likely to attend more specialty mental health visits and to receive an antidepressant medication than veterans who are seen by a primary care clinician only. 4.1. Relationships among SI, depression severity, and mental health care utilization While some studies support the use of routine screening for mental health disorders among veterans, finding, for example, that veterans are more likely to complete initial mental health care visits following a positive screen for depression or PTSD (Seal et al., 2008; Lu et al., 2012), few studies exist on the use of systematized SI assessment and its effects on treatment patterns and outcomes (Gaynes et al., 2004; O'Connor et al., 2013). One known study, of a community sample, examined mental health care treatment
4.2. SI assessment in mental health vs. primary care SI assessment by a mental health clinician on one's first visit to VA care was the only predictor of subsequent mental health utilization across all multivariate models, suggesting that many veterans in our sample self-selected into mental health care prior to SI assessment and any subsequent referrals. Previous work has shown that although OEF/OIF veterans seen in integrated primary care/mental health clinics may be more likely to accept referrals to specialty mental health care than veterans of other eras, they are not more likely to actually attend subsequent specialty mental health care appointments stemming from referrals (Lindley et al., 2010). This may mean that self-selection into specialty mental health care is a larger driver of continued treatment than referrals from non-mental health clinicians, emphasizing the importance of OEF/OIF veterans' initial decision to seek specialty mental health treatment, and a continued emphasis on reducing the stigma of mental health treatment. We also know that veterans with serious mental illness and other risk factors associated with suicide, such as depression and PTSD, have been identified as high risk for treatment discontinuation, gaps in care, and under-treatment (McCarthy et al., 2007; Lindley et al., 2010; Seal et al., 2010). In fact, a recent study found that only 29% of OEF/OIF veterans completed a minimally adequate course of specialty treatment for PTSD within a year of a positive PTSD screen (Lu et al., 2011). As we measured mental health utilization over the course of an entire year, our findings may be indicative of such treatment gaps or discontinuation. More emphasis on supporting the continued engagement in mental healthcare for veterans with SI and associated risk factors may be warranted. Additionally, our findings may signify clinician uncertainty of how to incorporate patient disclosure of suicidal thoughts into treatment plan considerations (Vannoy et al., 2011; Betz et al., 2013). This uncertainty might be a reflection of the lack of empirical evidence regarding to what extent, or for whom, treatment should be intensified when suicidal ideation is present (Gaynes et al., 2004; Mann et al., 2005; O'Neil et al., 2012). Resources may also be
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
151
Table 3 Multivariate correlates of mental health utilization by new-to-care OEF/OIF veterans during the year following a positive depression screen and structured SI assessment, accounting for high depression score. MH visitsa
Antidepressant prescription
MH visitþ Antidepressant prescription
AOR (95% CI)
P
AOR (95% CI)
P
AOR (95% CI)
P
Age 18–24 (reference) 25–34 35 þ
– 1.09 (0.73–1.64) 1.20 (0.71–2.02)
– 0.66 0.50
– 0.90 (0.55–1.47) 0.92 (0.49–1.73)
– 0.68 0.80
– 0.89 (0.54–1.47) 0.91 (0.48–1.72)
– 0.66 0.77
Sex, female
1.46 (0.87–2.44)
0.15
1.06 (0.57–1.97)
0.86
1.18 (0.63–2.22)
0.61
Non-Hispanic whiteb
1.05 (0.74–1.50)
0.78
0.97 (0.63–1.49)
0.90
0.80 (0.52–1.24)
0.32
Marital status Never married (reference) Married Divorced/widowed
– 1.49 (1.02–2.18) 1.52 (0.96–2.40)
– 0.04 0.07
– 1.44 (0.91–2.29) 1.25 (0.72–2.18)
– 0.12 0.43
– 1.94 (1.21–3.11) 1.68 (0.96–2.94)
– 0.01 0.07
Rural
0.80 (0.53–1.20)
0.28
1.02 (0.62–1.67)
0.95
1.10 (0.67–1.83)
0.70
PHQ-2 scores Z 5c
1.41 (1.00–1.99)
0.049
1.18 (0.78–1.79)
0.44
1.08 (0.71–1.65)
0.71
SI assessment by mental health clinician
1.72 (1.17–2.52)
0.01
1.85 (1.18–2.90)
0.01
2.39 (1.52–3.75)
o0.001
Clinical diagnoses Depressive Disorder Substance Use Disorder Post-traumatic Stress Anxiety
0.89 1.04 1.70 1.64
0.53 0.87 o0.01 0.03
2.02 0.67 2.38 1.73
o 0.01 0.16 o 0.001 0.046
1.71 0.74 2.37 1.81
0.01 0.30 o0.001 0.03
SI assessment positive
1.38 (0.95–2.01)
0.09
1.21 (0.77–1.90)
0.40
1.29 (0.82–2.03)
d
(0.63–1.27) (0.65–1.65) (1.20–2.42) (1.06–2.54)
(1.33–3.07) (0.38–1.17) (1.56–3.63) (1.01–2.97)
(1.12–2.62) (0.41–1.31) (1.54–3.62) (1.05–3.11)
0.28
Odds ratios adjusted for non-independence of observations within sites (AORs) with 95% confidence intervals (CI); analyses included only those with complete data (n¼ 449). OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; PHQ-2 ¼Patient Health Questionnaire depression screen; SI ¼suicidal ideation a
Mental health visits, range 0–8þ ; specialty mental health visits attended subsequent to initial assessment date. Non-Hispanic white vs. other. PHQ-2 Z 5 denotes high PHQ score. d International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. b c
Table 4 Multivariate correlates of mental health utilization by new-to-care OEF/OIF veterans during the year following a positive depression screen and structured SI assessment. MH visitsa
Antidepressant prescription
MH visitþ Antidepressant prescription
AOR (95% CI)
P
AOR (95% CI)
P
AOR (95% CI)
P
– 1.12 (0.75–1.67) 1.17 (0.70–1.98)
– 0.59 0.55
– 0.91 (0.56–1.48) 0.91 (0.49–1.72)
– 0.70 0.77
– 0.90 (0.55–1.48) 0.90 (0.48–1.71)
– 0.67 0.76
1.49 (0.89–2.48)
0.13
1.07 (0.58–2.00)
0.82
1.19 (0.63–2.23)
0.59
Non-Hispanic white
1.07 (0.75–1.53)
0.70
0.98 (0.68–1.50)
0.92
0.80 (0.52–1.24)
0.32
Marital status Never married (reference) Married Divorced/widowed
– 1.43 (0.98–2.10) 1.51 (0.96–2.38)
– 0.06 0.08
– 1.42 (0.90–2.26) 1.26 (0.73–2.19)
– 0.13 0.41
– 1.93 (1.21–3.09) 1.69 (0.96–2.95)
– o 0.01 0.07
Age 18–24 (reference) 25–34 35 þ Sex, female b
Rural
0.80 (0.53–1.20)
0.28
1.02 (0.62–1.67)
0.95
1.10 (0.67–1.83)
0.70
SI assessment by mental health clinician
1.80 (1.23–2.63)
o 0.01
1.89 (1.21–2.96)
o 0.01
2.42 (1.54–3.78)
o 0.001
Clinical diagnosesc Depressive Disorder Substance Use Disorder Post-traumatic Stress Anxiety
0.91 1.03 1.67 1.64
0.59 0.90 o 0.01 0.03
2.02 0.66 2.37 1.72
o 0.01 0.16 o 0.001 0.048
1.72 0.73 2.36 1.80
0.01 0.29 o 0.001 0.03
SI assessment positive
1.48 (1.03–2.14)
0.04
1.25 (0.80–1.95)
0.32
1.31 (0.84–2.04)
(0.64–1.29) (0.65–1.63) (1.18–2.37) (1.06–2.53)
(1.33–3.07) (0.38–1.17) (1.56–3.61) (1.00–2.95)
(1.13–2.62) (0.41–1.31) (1.54–3.62) (1.05–3.10)
0.24
OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; PHQ-2 ¼Patient Health Questionnaire depression screen; SI ¼suicidal ideation. Odds ratios adjusted for non-independence of observations within sites (AORs) with 95% confidence intervals (CI); analyses included only those with complete data (n ¼449). a
Mental health visits, range 0–8þ ; specialty mental health visits attended subsequent to initial assessment date. Non-Hispanic white vs. other. c International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. b
152
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
limited, forcing clinicians to choose between standard outpatient treatment regimens or psychiatric hospitalization. It should also be considered that our findings may be indicative of veterans more willing to endorse SI in a mental health setting. Veterans are more comfortable disclosing thoughts of suicide with a provider whom they trust, and they describe thoughts of suicide as particularly private (Ganzini et al., 2013). If the veterans in our study found it more appropriate or more comfortable to disclose SI to a mental health clinician, rather than their primary care provider, the association between SI and subsequent mental health utilization may have been misestimated. However, our results did not support this explanation; although bivariate comparisons showed veterans with SI were more likely to have been assessed for SI by a mental health clinician, this association did not hold when depression diagnoses and PHQ-2 scores were taken into account. Several limitations should be considered in the interpretation of our results. Current relationships between SI assessment and utilization follow-up may differ from what we observed during this time period; our results reflect the first 18 months of VA's establishment of the SI assessment performance standard, the implementation of which varied between VAMCs (Dobscha et al., 2013). The three participating VAMCs are geographically-dispersed and our sample is demographically reflective of the OEF/OIF population (Institute of Medicine, 2010), but generalizability to other VAMCs and non-VA samples may be limited, and our findings reflect mental health services access through, or within, the VA system. While we did not detect a direct effect of the SI assessment result on utilization, above what could be explained by depression symptoms, it is possible the SI assessment program had indirect and unmeasured effects on mental health care, perhaps through increased awareness of suicide risk or increased detection of mental health disorders. Finally, this study focused on brief SI assessment results per se and did not consider the validity and reliability of such measures for detecting SI, an important, but unrelated, issue (Gaynes et al., 2004; Mann et al., 2005; Razykov et al., 2012). To our knowledge, this is the first study among OEF/OIF veterans to specifically address whether SI detection using structured assessments following positive depression screens increases subsequent mental health utilization. As healthcare systems begin to adopt similar suicidal ideation detection or screening procedures, it is important to understand whether these procedures are successful in engaging at-risk patients in mental health care. While our results suggest that SI is not independently associated with increased mental health care utilization in the following year, our findings suggest that promoting veteran self-selection into mental health care, reducing its stigma, and finding ways to support continued engagement for veterans with SI may increase mental health utilization independent of risk detection efforts. In conjunction with SI assessments, developing clearer empirical guidance for treatment and increasing healthcare resources for individuals with suicidal ideation constitute valuable directions for future work.
Disclosures None for any author.
Acknowledgments This material is based upon work supported by the U.S. Department of Veterans Affairs, Veterans Health Administration, and Health Services Research and Development Service Project
DHI-08-096. Dr. Denneson is a core investigator in the Center to Improve Veteran Involvement in Care (CIVIC) at the Portland VA Medical Center. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
References Bauer, A.M., Chan, Y.F., Huang, H., Vannoy, S., Unutzer, J., 2013. Characteristics, management, and depression outcomes of primary care patients who endorse thoughts of death or suicide on the PHQ-9. Journal of General Internal Medicine 28, 363–369. Betz, M.E., Sullivan, A.F., Manton, A.P., Espinola, J.A., Miller, I., Camargo Jr, C.A., Boudreaux, E.D., 2013. on behalf of the ED-SAFE Investigators, 2013. Knowledge, Attitudes, and Practices of Emergency Department Providers in the Care of Suicidal Patients. Depression and Anxiety 30, 1005–1012. Brenner, L.A., Breshears, R.E., Betthauser, L.M., Bellon, K.K., Holman, E., Harwood, J. E., Silverman, M.M., Huggins, J., Nagamoto, H.T., 2011. Implementation of a suicide nomenclature within two VA healthcare settings. Journal of Clinical Psychology in Medical Settings 18, 116–128. Britton, P.C., Ilgen, M.A., Rudd, M.D., Conner, K.R., 2012. Warning signs for suicide within a week of healthcare contact in Veteran decedents. Psychiatry Research 200, 395–399. Corson, K., Denneson, L.M., Bair, M., Helmer, D.A., Goulet, J.L., Dobscha, S.K., 2013. Prevalence and correlates of suicidal ideation among OEF/OIF Veterans. Journal of Affective Disorders 149, 291–298. Crosby, A.E., Ortega, L., Melanson, C., 2011. Self-Directed Violence Surveillance: Uniform Definitions and Recommended Data Elements, Version 1.0. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control, Atlanta, GA. Department of Veterans Affairs, 2007. Cinical Reminders. 〈http://www.va.gov/ vdl/documents/Clinical/CPRS-Clinical_Reminders/pxrm_2_6_um.pdf〉 (accessed 25.10.11). Department of Veterans Affairs, 2008. Suicide Risk Assessment Guide. 〈www.mental health.va.gov/docs/suicide-risk-assessment-guide.pdf〉 (accessed 27.2.12). Department of Veterans Affairs, 2009. Analysis of Unique Veterans Utilization of VA Benefits & Services. 〈http://www.va.gov/VETDATA/docs/SpecialReports/unique veteransMay.pdf\〉 (accessed 27.2.12). Dobscha, S.K., Corson, K., Helmer, D.A., Bair, M.J., Denneson, L.M., Brandt, C., Beane, A., Ganzini, L., 2013. Brief assessment for suicidal ideation in OEF/OIF veterans with positive depression screens. General Hospital Psychiatry 35, 272–278. Ganzini, L., Denneson, L., Press, N., Bair, M.J., Helmer, D.A., Poat, J., Dobscha, S.K., 2013. Trust is the basis for effective suicide risk assessment in veterans. Journal of General Internal Medicine 28, 1215–1221. Gaynes, B.N., West, S.L., Ford, C.A., Frame, P., Klein, J., Lohr, K.N., 2004. Screening for suicide risk in adults: a summary of the evidence for the U.S. Preventive Services Task Force. Annals of Internal Medicine 140, 822–835. Institute of Medicine, 2010. Returning Home from Iraq and Afghanistan: Preliminary Assessment of Readjustment Needs of Veterans, Service Members, and their Families. The National Academies Press, Washington DC. Jakupcak, M., Cook, J., Imel, Z., Fontana, A., Rosenheck, R., McFall, M., 2009. Posttraumatic stress disorder as a risk factor for suicidal ideation in Iraq and Afghanistan War veterans. Journal of Traumatic Stress 22, 303–306. Kroenke, K., Spitzer, R.L., Williams, J.B., 2001. The PHQ-9: validity of a brief depression severity measure. Journal of General Internal Medicine 16, 606–613. Kroenke, K., Spitzer, R.L., Williams, J.B., 2003. The Patient Health Questionnaire-2: validity of a two-item depression screener. Medical Care 41, 1284–1292. Lemaire, C.M., Graham, D.P., 2011. Factors associated with suicidal ideation in OEF/ OIF veterans. Journal of Affective Disorders 130, 231–238. Lindley, S., Cacciapaglia, H., Noronha, D., Carlson, E., Schatzberg, A., 2010. Monitoring mental health treatment acceptance and initial treatment adherence in veterans: veterans of Operations Enduring Freedom and Iraqi Freedom versus other veterans of other eras. Annals of the New York Academy of Sciences 1208, 104–113. Lish, J.D., Zimmerman, M., Farber, N.J., Lush, D.T., Kuzma, M.A., Plescia, G., 1996. Suicide screening in a primary care setting at a Veterans Affairs Medical Center. Psychosomatics 37, 413–424. Lu, M.W., Carlson, K.F., Duckart, J.P., Dobscha, S.K., 2012. The effects of age on initiation of mental health treatment after positive PTSD screens among Veterans Affairs primary care patients. General Hospital Psychiatry 34, 654–659. Lu, M.W., Duckart, J.P., O'Malley, J.P., Dobscha, S.K., 2011. Correlates of utilization of PTSD specialty treatment among recently diagnosed Veterans at the VA. Psychiatric Services 62, 943–949. Mann, J.J., Apter, A., Bertolote, J., Beautrais, A., Currier, D., Haas, A., Hegerl, U., Lonnqvist, J., Malone, K., Marusic, A., Mehlum, L., Patton, G., Phillips, M., Rutz, W., Rihmer, Z., Schmidtke, A., Shaffer, D., Silverman, M., Takahashi, Y., Varnik, A., Wasserman, D., Yip, P., Hendin, H., 2005. Suicide prevention strategies: a systematic review. Journal of the American Medical Association 294, 2064–2074. Mann, J.J., Ellis, S.P., Waternaux, C.M., Liu, X., Oquendo, M.A., Malone, K.M., Brodsky, B.S., Haas, G.L., Currier, D., 2008. Classification trees distinguish suicide
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
attempters in major psychiatric disorders: a model of clinical decision making. Journal of Clinical Psychiatry 69, 23–31. McCarthy, J.F., Blow, F.C., Valenstein, M., Fischer, E.P., Owen, R.R., Barry, K.L., Hudson, T.J., Ignacio, R.V., 2007. Veterans Affairs Health System and mental health treatment retention among patients with serious mental illness: evaluating accessibility and availability barriers. Health Services Research 42, 1042–1060. Nielsen, B., Wang, A.G., Brille-Brahe, U., 1990. Attempted suicide in Denmark. IV. A five-year follow-up. Acta Psychiatrica Scandinavica 81, 250–254. O'Connor, E., Gaynes, B.N., Burda, B.U., Soh, C., Whitlock, E.P., 2013. Screening for and treatment of suicide risk relevant to primary care: a systematic review for the U.S. Preventive Services Task Force. Annals of Internal Medicine 158, 741–754. O'Neil, M.E., Peterson, K., Low, A., Carson, S., Denneson, L.M., Haney, E., Shiroma, P., Kansagara, D., 2012. Suicide Prevention Interventions and Referral/Follow-up Services. A Systematic Review. VA-ESP Project No. 05-225. Pietrzak, R.H., Goldstein, M.B., Malley, J.C., Rivers, A.J., Johnson, D.C., Southwick, S. M., 2010. Risk and protective factors associated with suicidal ideation in veterans of Operations Enduring Freedom and Iraqi Freedom. Journal of Affective Disorders 123, 102–107. Razykov, I., Ziegelstein, R.C., Whooley, M.A., Thombs, B.D., 2012. The PHQ-9 versus the PHQ-8–is item 9 useful for assessing suicide risk in coronary artery disease patients? Data from the Heart and Soul Study. Journal of Psychosomatic Research 73, 163–168. Schulberg, H.C., Lee, P.W., Bruce, M.L., Raue, P.J., Lefever, J.J., Williams, J.W., Dietrich, A.J., Nutting, P.A., 2005. Suicidal ideation and risk levels among primary care
153
patients with uncomplicated depression. Annals of Family Medicine 3, 523–528. Seal, K.H., Bertenthal, D., Maguen, S., Gima, K., Chu, A., Marmar, C.R., 2008. Getting beyond "Don't ask; don't tell": an evaluation of US Veterans Administration postdeployment mental health screening of veterans returning from Iraq and Afghanistan. American Journal of Public Health 98, 714–720. Seal, K.H., Maguen, S., Cohen, B., Gima, K.S., Metzler, T.J., Ren, L., Bertenthal, D., Marmar, C.R., 2010. VA mental health services utilization in Iraq and Afghanistan veterans in the first year of receiving new mental health diagnoses. Journal of Traumatic Stress 23, 5–16. Szanto, K., Mulsant, B.H., Houck, P., Dew, M.A., Reynolds 3rd, C.F., 2003. Occurrence and course of suicidality during short-term treatment of late-life depression. Archives of General Psychiatry 60, 610–617. Tabachnick, B.G., Fidell, L.S., 2001. Using Multivariate Statistics, 4 ed. Allyn and Bacon, Boston. Uebelacker, L.A., German, N.M., Gaudiano, B.A., Miller, I.W., 2011. Patient health questionnaire depression scale as a suicide screening instrument in depressed primary care patients: a cross-sectional study. Primary Care Companion for CNS Disorders 13. Vannoy, S.D., Tai-Seale, M., Duberstein, P., Eaton, L.J., Cook, M.A., 2011. Now what should I do? Primary care physicians' responses to older adults expressing thoughts of suicide. Journal of General Internal Medicine 26, 1005–1011. West, A.N., Lee, R.E., Shambaugh-Miller, M.D., Bair, B.D., Mueller, K.J., Lilly, R.S., Kaboli, P.J., Hawthorne, K., 2010. Defining "rural" for veterans' health care planning. Journal of Rural Health 26, 301–309.
Contents lists available at ScienceDirect
Psychiatry Research journal homepage: www.elsevier.com/locate/psychres
Mental health utilization of new-to-care Iraq and Afghanistan Veterans following suicidal ideation assessment Lauren M. Denneson a,b,n, Kathryn Corson a,b, Drew A. Helmer c,d, Matthew J. Bair e, Steven K. Dobscha a,b a
Portland Center to Improve Veteran Involvement in Care (CIVIC), Portland VA Medical Center, Portland, OR, United States Department of Psychiatry, Oregon Health & Science University, Portland, OR, United States c War-Related Illness and Injury Study Center, VA New Jersey Health Care System, East Orange, NJ, United States d Department of Medicine and Dentistry, New Jersey Medical School, Newark, NJ, United States e Richard L. Roudebush Veteran Affairs Medical Center, Indianapolis, IN, United States b
art ic l e i nf o
a b s t r a c t
Article history: Received 25 September 2013 Received in revised form 8 February 2014 Accepted 14 March 2014 Available online 22 March 2014
We evaluated the impact of brief structured suicidal ideation (SI) assessments on mental health care among new-to-care Operations Enduring Freedom and Iraqi Freedom (OEF/OIF) veterans. National datasets provided military, demographic, and clinical information. For all new-to-care OEF/OIF veterans administered depression screens (PHQ-2: Patient Health Questionnaire-2) and structured SI assessments in primary care or ambulatory mental health settings of three Veterans Affairs (VA) Medical Centers between April 2008 and September 2009 (N ¼465), generalized estimating equations were used to examine associations between SI and number of subsequent-year specialty mental health visits and antidepressant prescriptions. Approximately one-third of the veterans reported SI. In multivariate models, PTSD and anxiety diagnoses, severe depression symptoms, being married, and SI assessment by a mental health clinician were associated with more mental health visits in the subsequent year. Depression, PTSD, and anxiety diagnoses, and SI assessment by a mental health clinician were associated with receiving antidepressants. Presence of SI did not significantly affect subsequent year mental health utilization when adjusting for diagnostic and clinician variables, but inaugural visits involving mental health clinicians were consistently associated with subsequent mental health care. Published by Elsevier Ireland Ltd.
Keywords: Depression Risk assessment Suicide Delivery of health care
1. Introduction Suicidal ideation (SI), defined as, “thoughts of engaging in suiciderelated behavior,” (Brenner et al., 2011; Crosby et al., 2011) is among one of the best predictors of suicide attempts (Nielsen et al., 1990; Szanto et al., 2003; Mann et al., 2008; Britton et al., 2012). Population estimates of SI vary considerably due to methods and sample selection, yet there is evidence that rates are high among Veterans Affairs (VA) healthcare-seeking veterans with mental health symptoms; up to 46% of Operation Enduring Freedom and Operation Iraqi Freedom (OEF/OIF) veterans referred for VA mental health services (Jakupcak et al., 2009) and 32% of OEF/OIF veterans with depression symptoms in VA care report SI (Corson et al., 2013). The VA, which operates the largest integrated healthcare system in the United States (Department of Veterans Affairs, 2009), has designated assessment of SI among veterans with positive depression
n Correspondence to: Portland VA Medical Center, P.O. Box 1034 (R&D66), Portland, OR 97207, United States. Tel: þ 1 503 220-8262x57351. E-mail address: [email protected] (L.M. Denneson).
http://dx.doi.org/10.1016/j.psychres.2014.03.017 0165-1781/Published by Elsevier Ireland Ltd.
or PTSD screens as a national performance goal since 2008. The main goal of this initiative is to identify at-risk veterans to engage them in appropriate mental health treatment. VA recommendations for such care include pharmacological treatment of acute symptoms, such as anxiety, insomnia, and psychosis as well as careful evaluation, treatment, and monitoring of any underlying mental health disorders, especially depression (Department of Veterans Affairs, 2008). However, the relationship between suicide screening, or risk detection, programs and treatment utilization, is not well understood (Gaynes et al., 2004; O'Connor et al., 2013). In a community setting, one study detected small associations between SI and subsequent receipt of further assessment, specialty mental health referral, and psychotropic medication (Bauer et al., 2013). One prior study in VA primary care revealed those with SI attended more primary care visits than those without SI (Lish et al., 1996). Given the significant time and financial burden routinized screening or assessment programs place on a healthcare system, it is imperative we understand the relative contribution of these structured SI assessments in meeting the desired objectives. In this retrospective study, we sought to evaluate the impact of VA's structured SI assessment procedures on veteran engagement in
148
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
mental health treatment. Among new-to-care OEF/OIF veterans who screened positive for depression in VA outpatient settings, we identified the correlates of subsequent-year VA mental health care utilization, with particular interest in examining the relationship between SI and utilization. Based on recommended follow-up for SI, described above (Department of Veterans Affairs, 2008), we hypothesized that veterans with positive SI assessments would have an increased number of subsequent mental health specialty visits and/or psychiatric medication prescriptions, as compared to veterans with negative SI assessments. 2. Methods We conducted a multi-site study of the SI assessment process and 1-year outcomes among OEF/OIF veterans screened for depression between April 1, 2008 and September 30, 2009 at three VA Medical Centers (VAMCs), representing the Northwestern, Southwestern, and Northeastern census regions of the US. Methods have been reported in detail elsewhere (Corson et al., 2013; Dobscha et al., 2013). 2.1. Settings Each of the VAMCs has metropolitan and rural-based primary care clinics, is closely affiliated with a local university, and provides a full range of patient care services. All VAMCs utilize the VA's electronic Computerized Patient Record System (CPRS). A clinical reminder system embedded within CPRS facilitates the routine administration of screens and assessment tools for a variety of conditions and disorders (Department of Veterans Affairs, 2007), including depression and SI. Items are administered by clinicians or clinic staff, and responses, with results, are listed in the electronic progress note. Positive results generate new reminders to prompt clinician follow-up. 2.2. Data sources The OEF/OIF Roster File, maintained by the Department of Defense and the Defense Manpower Data Center, contains demographic and service-related data on all OEF/OIF veterans discharged after September 11, 2001 who enrolled in or accessed VHA services after discharge. We used social security numbers to match the OEF/OIF Roster to VA Decision Support System data. Methods related to collection of depression screen and SI assessment results and inter-rater reliability are described in recent manuscripts (Corson et al., 2013; Dobscha et al., 2013). 2.3. Sample We limited our sample to new-to-care patients (no VA healthcare visits during the 5 years prior) because depression and suicide risk assessment results may have less effect on continuity of care than on initiation of care, especially among established patients with diagnosed psychiatric disorders. Also, at some VAMCs, patients currently engaged in mental health may be exempt from mental health screens. At the three medical centers during the study period, 600 new-to-care veterans screened positive for depression in a primary care or mental health outpatient setting. Within this group, 465 (77.5%) had a same-day structured SI assessment using one of two widely-used structured SI assessment tools (used in 98.2% of all structured SI assessments completed during the study period (Dobscha et al., 2013)). This group constituted the sample for the current study, with the date of the positive depression screen and SI assessment representing the patient's index date. The majority of the sample was under 34 years of age (80.2%), nonHispanic white (60.4%), and male (87.5%). The study was approved by the Institutional Review Boards of each participating site. Patient data presented here were obtained as part of routine clinical practice and thus waivers of informed consent were granted.
dead or of hurting themselves in some way. Any response 40 (not at all) is scored as a positive result (Kroenke et al., 2001; Schulberg et al., 2005). The VA “Pocket Card” Risk Assessment (VA-PCRA) (Department of Veterans Affairs, 2008) is a fouritem assessment, and a positive result requires endorsing, “thoughts of taking your life.” Reliability and validity of these instruments for SI detection have not been established, though two recent studies support concurrent validity of the PHQ-9 9th item as an SI screening tool (Uebelacker et al., 2011; Bauer et al., 2013). Veterans were administered one or the other (and sometimes both) of these assessments, depending on the facility, or clinician or staff preference; some veterans received multiple assessments on one day. Veterans with any positive SI assessment result on the same day as the positive PHQ-2 were categorized into the positive SI group; veterans with only negative results were categorized into the negative SI group. Demographic characteristics: age, sex, race/ethnicity, marital status, zip code of residence, and service connection status were obtained from the VA Decision Support System. To designate veteran rurality, we used the VA's Office of Rural Health’s classifications (Urban, Rural, or Highly Rural), which are based on geocoding and utilize a zip-code-based crosswalk system (West et al., 2010). Race/ ethnicity data were frequently missing from Decision Support System databases, so race/ethnicity identification was supplemented by OEF/OIF roster data. Healthcare data: utilization, diagnosis, and prescription data were obtained from DSS for 1 year, beginning with the patient's index date, at the VAMC of depression screening. Depressive disorder, PTSD, substance use disorder (SUD), and anxiety disorder diagnoses made on the index date were identified using International Classification of Diseases, Clinical Modification (ICD-9-CM) codes (full list available from corresponding author). 2.5. Analyses All analyses were completed using PASW Statistics (SPSS version 18). Consistent with prior studies (Corson et al., 2013; Dobscha et al., 2013), we collapsed age into three groups ( o24, 25–34, and Z 35), marital status into three groups (single/ never married, married, and divorced/widowed), and education into three groups (less than high school diploma, high school/some college, and college degree or higher). We dichotomized race/ethnicity into non-Hispanic white vs. other and rurality was dichotomized by combining the highly rural and rural categories. To denote a high PHQ-2 score (severe depression symptoms), PHQ-2 scores were dichotomized at PHQ-2 Z 5 (vs. PHQ-2 o 5); this score cutoff requires the patient to report experiencing both anhedonia and depression, with at least one of the symptoms occurring every day during the previous 2 weeks. For this study, the number of specialty mental health visits were summed for each patient over the course of the year follow-up period, and were ordinally categorized into nine levels (0 to Z 8) based on the positive skew and platykurtosis of the distribution in this sample. Univariate analyses (χ2 and Mann–Whitney U) were used to test demographic, clinical, and utilization differences between the positive SI and negative SI groups. Generalized estimating equation (GEE) models were then constructed to identify multivariate correlates of SI, and to identify multivariate correlates of subsequent year utilization, while controlling for potential non-independence of observations with sites. For specialty mental health visits, GEE ordinal logistic models were used. As in linear regression, log-odds regression coefficients show the change expected in the dependent variable for a one unit increase in the predictor, wherein change is defined as moving from one category/level to the next; one equation is estimated across all outcome variable levels and odds ratios are calculated by exponentiating the regression coefficients (Tabachnick and Fidell, 2001). We included a priori in each model sex, age, race/ethnicity, rurality, marital status, SI assessment by a mental health clinician (yes/no; includes non-licensed care staff), and SI assessment positive (yes/no). We also included PHQ-2 scores, and depressive disorder and PTSD diagnoses made on the index date; we previously reported on the association of these variables with receipt of SI assessment and with SI assessment result (Corson et al., 2013; Dobscha et al., 2013). Additional diagnosis variables were included if associated with one or more outcome variables at p r 0.10.
3. Results 2.4. Measures
3.1. Sample characteristics and differences between groups Depression: annual depression screening is conducted using the 9-item or 2-item Patient Health Questionnaire (PHQ-9 and PHQ-2, respectively). Both the PHQ-9 and PHQ-2 are well-validated and widely-used instruments for detecting and diagnosing depression and measuring depression severity (Kroenke et al., 2003). The PHQ-2 consists of the depressed mood and anhedonia items of the PHQ-9, and response options range from 0 (not at all) to 3 (nearly every day). A PHQ-2 score of 3 is the cut-point for a positive depression screen, with a sensitivity of 83% and a specificity of 92% for major depression (Kroenke et al., 2003). Suicidal ideation: at the time of the study, two brief, structured tools were predominantly used to detect suicidal ideation at the three medical centers. The 9th item of the PHQ-9 asks patients if they have had thoughts of being better off
Of the 465 OEF/OIF veterans in the sample, 147 (32%) had positive SI assessment results. In univariate comparisons of the positive SI and negative SI groups, the two groups did not differ on demographic characteristics, but did differ on several clinical variables (Table 1). Veterans in the positive SI group were more likely to have high PHQ-2 scores, be assessed for SI by a mental health clinician, and be diagnosed with a depressive disorder or a substance use disorder. In multivariate analyses, only depressive
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
149
Table 1 Demographic characteristics and VA healthcare utilization of new-to-care OEF/OIF veterans with a positive depression screena and same-day structured SI assessment, by SI assessment result. Characteristic
All patients (n ¼465)
Negative SI (n ¼318)
Site VAMC 1 VAMC 2 VAMC 3
107 289 69
(23.0) (62.2) (14.8)
65 204 49
(20.4) (64.2) (15.4)
42 85 20
(28.6) (57.8) (13.6)
Age r 24 25–34 Z 35 Male
110 263 92 407
(23.7) (56.6) (19.8) (87.5)
75 174 69 279
(23.6) (54.7) (21.7) (87.7)
35 89 23 128
(23.8) (60.5) (15.6) (87.1)
0.84
Non-Hispanic whiteb (8 missing)
276
(60.4)
196
(62.8)
80
(55.2)
0.12
Marital status (5 missing) Single/never married Married Divorced/widowed
195 173 92
(42.4) (37.6) (20.0)
138 118 57
(44.1) (37.7) (18.2)
57 55 35
(38.8) (37.4) (23.8)
Education (2 missing) Less than high school diploma High school/some college College degree or higher
72 363 28
(15.6) (78.4) (6.0)
48 246 23
(15.1) (77.6) (7.3)
24 117 5
(16.4) (80.1) (3.4)
Rural designation (4 missing) Urban Rural
358 103
(77.7) (22.3)
241 74
(76.5) (23.5)
117 29
(80.1) (19.9)
Service connectedc
188
(40.4)
126
(39.6)
62
(42.2)
0.60
Positive SI (n¼ 147)
P 0.15
0.29
a
0.33
0.27
0.38
237
(51.0)
140
(44.0)
97
(66.0)
o 0.001
SI assessment by a mental health clinician
173
(37.2)
103
(32.4)
70
(47.6)
o 0.01
Clinical diagnosesd Depressive disorder Post-traumatic stress disorder Substance use disorder Anxiety
254 224 79 94
(54.6) (48.2) (17.0) (20.2)
151 144 45 59
(47.5) (45.3) (14.2) (18.6)
103 80 34 35
(70.1) (54.4) (23.1) (23.8)
o 0.001 0.07 0.02 0.19
Mental health visitse, median (IQR) Antidepressant prescription
3.00 248
(7.00) (53.3)
3.00 156
(6.00) (49.1)
4.00 92
(6.00) (62.6)
o 0.01 o 0.01 0.98
PHQ-2 Z5
Benzodiazepine prescription Mental health visit þ antidepressant prescription Psychiatric inpatient stay
44 229 24
(9.5) (49.2) (5.2)
30 142 9
(9.4)
14
(9.5)
(44.7)
87
(59.2)
o 0.01
(2.8)
15
(10.2)
0.001
All results are presented as n (%), unless otherwise noted. OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; SI¼ suicidal ideation; VAMC ¼Veterans Affairs Medical Center; PHQ-2 ¼Patient Health Questionnaire depression screen; SI ¼suicidal ideation; IQR ¼interquartile range. Mental health visit þ antidepressant¼ any specialty mental health visit completed subsequent to initial assessment date and any antidepressant prescription filled. a
PHQ-2 43 is positive screen, range 3–6. Non-Hispanic white vs. other. c Service connected veterans receive compensation in the form of disability payments and/or treatment for the service-related disability. d International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. e Mental health visits, range 0–8þ ; specialty mental health visits attended subsequent to initial assessment date. b
disorder diagnoses and high PHQ-2 scores remained associated with suicidal ideation (Table 2). 3.2. Subsequent year utilization In the year following SI assessment, 377 (81.1%) had at least one mental health visit (median¼3.00, IQR¼7.00), 248 (53.3%) filled an antidepressant prescription, and 24 (5.2%) had a psychiatric inpatient stay. In univariate analyses, the groups differed significantly on number of specialty mental health visits, antidepressant prescriptions filled, any mental health visit and any antidepressant prescription fill, and any psychiatric inpatient stays (Table 1). Benzodiazepine prescription receipt did not differ between the two groups, and the small number of veterans completing inpatient stays precluded using benzodiazepine prescriptions or psychiatric inpatient stays in further analyses. Multivariate correlates of utilization outcomes are presented in Table 3. The odds of attending a greater number of mental health
visits were higher for married veterans (AOR¼ 1.49, p ¼0.04), veterans diagnosed with PTSD (AOR ¼1.70, p o0.01) or anxiety (AOR¼ 1.64, p ¼0.03), for veterans with high PHQ-2 scores (AOR¼ 1.41, p ¼0.049), and for those who were assessed for SI by a mental health clinician (AOR ¼1.72, p ¼0.01). The odds of antidepressant prescription receipt were greater for veterans diagnosed with PTSD (AOR¼2.38, p o0.001), anxiety (AOR ¼1.73, p¼ 0.046) or depressive disorders (AOR ¼2.02, p o0.01), and for veterans who were assessed for SI by a mental health clinician (AOR¼ 1.85, p ¼0.01). The odds of both attending a mental health visit and receiving an antidepressant prescription were greater for veterans who were married (AOR¼1.94, p ¼0.01), assessed for SI by a mental health clinician (AOR¼ 2.39, p o0.001), and were diagnosed with a depressive disorder (AOR¼1.71, p¼ 0.01), PTSD (AOR¼ 2.37, po 0.001), or anxiety (AOR ¼1.81, p ¼0.03). In all adjusted models, SI was not predictive of mental health care utilization. In post hoc analyses, we removed PHQ-2 score from the models to investigate whether the relationships between SI
150
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
Table 2 Multivariate correlates of suicidal ideation in new-to-care OEF/OIF veterans with a positive depression screen and structured SI assessment. AOR (95% CI)
P
– 1.06 (0.65–1.74) 0.71 (0.37–1.38)
– 0.82 0.31
0.81 (0.44–1.50)
0.50
Non-Hispanic white
0.66 (0.43–1.03)
0.07
Marital status Never married (reference) Married Divorced/widowed
– 1.37 (0.85–2.19) 1.42 (0.82–2.46)
– 0.20 0.21
Rural
0.96 (0.58–1.60)
0.87
PHQ-2 scores Z 5b
2.17 (1.43–3.30)
o 0.001
Clinical diagnosesc Depressive Disorder Substance Use Disorder Post-traumatic Stress Anxiety
2.16 (1.40–3.34) 1.41 (0.82–2.42) 1.10 (0.72–1.69) 1.18 (0.70–1.98)
0.001 0.21 0.67 0.53
SI assessment by mental health cliniciand
1.46 (0.90–2.36)
0.12
Age 18–24 (reference) 25–34 35þ Sex, female a
Odds ratios adjusted for non-independence of observations within sites (AORs) with 95% confidence intervals (CI); analyses included only those with complete data (n ¼449). OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; PHQ-2¼ Patient Health Questionnaire depression screen; SI ¼suicidal ideation.
services following SI endorsement on the PHQ-9 9th item in a primary care setting. Results supported an association between SI endorsement and 12-week mental health follow-up patterns; patients with SI were more likely to receive referrals for specialty mental health, more likely to receive psychotropic medication, and received earlier assessment by a care manager (Bauer et al., 2013). In contrast, we did not find an independent relationship between detection of SI and mental health utilization, although comparison is complicated by several methodological differences. We discovered that the relationship between SI and subsequent year specialty mental health visits was apparent only when PHQ-2 scores were not included in the model. This is not surprising, given that severe depression symptoms bring greater distress and have been repeatedly linked with SI (Mann et al., 2005; Pietrzak et al., 2010; Lemaire and Graham, 2011; Corson et al., 2013). As noted in Bauer et al. (2013), where endorsing SI on the PHQ-9 frequently co-occurred with severe depression symptoms on the PHQ-2 (Bauer et al., 2013), depression screening itself may identify a significant portion of those with SI. However, in a previous study of correlates of receiving an SI assessment through the VA's SI assessment initiative (Dobscha et al., 2013) we found that veterans with high PHQ-2 scores were more likely to receive an SI assessment than those with slightly lower scores. This systematic bias may have artificially inflated the relationship detected between high PHQ-2 scores and SI; the multivariate models of mental health utilization that control for high PHQ-2 scores help account for this confounding relationship.
a
Non-Hispanic white vs. other. PHQ-2 Z5 denotes high PHQ score. International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. d As opposed to primary care. b c
and the three utilization outcomes changed when PHQ-2 score are not taken into account. Results indicated that the odds of completing more specialty mental health care visits were greater for veterans with SI (AOR¼ 1.48, p ¼0.04) when not adjusting for PHQ-2 score (Table 4).
4. Discussion In this study of new-to-care OEF/OIF veterans who screened positive for depression and were assessed for SI using brief, structured assessments, we note two key observations. First, detection and presence of SI is not associated with subsequent mental health care utilization when accounting for severity of depression symptoms. Second, when a veteran's inaugural visit to VA healthcare includes a mental health clinician, the veteran is more likely to attend more specialty mental health visits and to receive an antidepressant medication than veterans who are seen by a primary care clinician only. 4.1. Relationships among SI, depression severity, and mental health care utilization While some studies support the use of routine screening for mental health disorders among veterans, finding, for example, that veterans are more likely to complete initial mental health care visits following a positive screen for depression or PTSD (Seal et al., 2008; Lu et al., 2012), few studies exist on the use of systematized SI assessment and its effects on treatment patterns and outcomes (Gaynes et al., 2004; O'Connor et al., 2013). One known study, of a community sample, examined mental health care treatment
4.2. SI assessment in mental health vs. primary care SI assessment by a mental health clinician on one's first visit to VA care was the only predictor of subsequent mental health utilization across all multivariate models, suggesting that many veterans in our sample self-selected into mental health care prior to SI assessment and any subsequent referrals. Previous work has shown that although OEF/OIF veterans seen in integrated primary care/mental health clinics may be more likely to accept referrals to specialty mental health care than veterans of other eras, they are not more likely to actually attend subsequent specialty mental health care appointments stemming from referrals (Lindley et al., 2010). This may mean that self-selection into specialty mental health care is a larger driver of continued treatment than referrals from non-mental health clinicians, emphasizing the importance of OEF/OIF veterans' initial decision to seek specialty mental health treatment, and a continued emphasis on reducing the stigma of mental health treatment. We also know that veterans with serious mental illness and other risk factors associated with suicide, such as depression and PTSD, have been identified as high risk for treatment discontinuation, gaps in care, and under-treatment (McCarthy et al., 2007; Lindley et al., 2010; Seal et al., 2010). In fact, a recent study found that only 29% of OEF/OIF veterans completed a minimally adequate course of specialty treatment for PTSD within a year of a positive PTSD screen (Lu et al., 2011). As we measured mental health utilization over the course of an entire year, our findings may be indicative of such treatment gaps or discontinuation. More emphasis on supporting the continued engagement in mental healthcare for veterans with SI and associated risk factors may be warranted. Additionally, our findings may signify clinician uncertainty of how to incorporate patient disclosure of suicidal thoughts into treatment plan considerations (Vannoy et al., 2011; Betz et al., 2013). This uncertainty might be a reflection of the lack of empirical evidence regarding to what extent, or for whom, treatment should be intensified when suicidal ideation is present (Gaynes et al., 2004; Mann et al., 2005; O'Neil et al., 2012). Resources may also be
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
151
Table 3 Multivariate correlates of mental health utilization by new-to-care OEF/OIF veterans during the year following a positive depression screen and structured SI assessment, accounting for high depression score. MH visitsa
Antidepressant prescription
MH visitþ Antidepressant prescription
AOR (95% CI)
P
AOR (95% CI)
P
AOR (95% CI)
P
Age 18–24 (reference) 25–34 35 þ
– 1.09 (0.73–1.64) 1.20 (0.71–2.02)
– 0.66 0.50
– 0.90 (0.55–1.47) 0.92 (0.49–1.73)
– 0.68 0.80
– 0.89 (0.54–1.47) 0.91 (0.48–1.72)
– 0.66 0.77
Sex, female
1.46 (0.87–2.44)
0.15
1.06 (0.57–1.97)
0.86
1.18 (0.63–2.22)
0.61
Non-Hispanic whiteb
1.05 (0.74–1.50)
0.78
0.97 (0.63–1.49)
0.90
0.80 (0.52–1.24)
0.32
Marital status Never married (reference) Married Divorced/widowed
– 1.49 (1.02–2.18) 1.52 (0.96–2.40)
– 0.04 0.07
– 1.44 (0.91–2.29) 1.25 (0.72–2.18)
– 0.12 0.43
– 1.94 (1.21–3.11) 1.68 (0.96–2.94)
– 0.01 0.07
Rural
0.80 (0.53–1.20)
0.28
1.02 (0.62–1.67)
0.95
1.10 (0.67–1.83)
0.70
PHQ-2 scores Z 5c
1.41 (1.00–1.99)
0.049
1.18 (0.78–1.79)
0.44
1.08 (0.71–1.65)
0.71
SI assessment by mental health clinician
1.72 (1.17–2.52)
0.01
1.85 (1.18–2.90)
0.01
2.39 (1.52–3.75)
o0.001
Clinical diagnoses Depressive Disorder Substance Use Disorder Post-traumatic Stress Anxiety
0.89 1.04 1.70 1.64
0.53 0.87 o0.01 0.03
2.02 0.67 2.38 1.73
o 0.01 0.16 o 0.001 0.046
1.71 0.74 2.37 1.81
0.01 0.30 o0.001 0.03
SI assessment positive
1.38 (0.95–2.01)
0.09
1.21 (0.77–1.90)
0.40
1.29 (0.82–2.03)
d
(0.63–1.27) (0.65–1.65) (1.20–2.42) (1.06–2.54)
(1.33–3.07) (0.38–1.17) (1.56–3.63) (1.01–2.97)
(1.12–2.62) (0.41–1.31) (1.54–3.62) (1.05–3.11)
0.28
Odds ratios adjusted for non-independence of observations within sites (AORs) with 95% confidence intervals (CI); analyses included only those with complete data (n¼ 449). OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; PHQ-2 ¼Patient Health Questionnaire depression screen; SI ¼suicidal ideation a
Mental health visits, range 0–8þ ; specialty mental health visits attended subsequent to initial assessment date. Non-Hispanic white vs. other. PHQ-2 Z 5 denotes high PHQ score. d International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. b c
Table 4 Multivariate correlates of mental health utilization by new-to-care OEF/OIF veterans during the year following a positive depression screen and structured SI assessment. MH visitsa
Antidepressant prescription
MH visitþ Antidepressant prescription
AOR (95% CI)
P
AOR (95% CI)
P
AOR (95% CI)
P
– 1.12 (0.75–1.67) 1.17 (0.70–1.98)
– 0.59 0.55
– 0.91 (0.56–1.48) 0.91 (0.49–1.72)
– 0.70 0.77
– 0.90 (0.55–1.48) 0.90 (0.48–1.71)
– 0.67 0.76
1.49 (0.89–2.48)
0.13
1.07 (0.58–2.00)
0.82
1.19 (0.63–2.23)
0.59
Non-Hispanic white
1.07 (0.75–1.53)
0.70
0.98 (0.68–1.50)
0.92
0.80 (0.52–1.24)
0.32
Marital status Never married (reference) Married Divorced/widowed
– 1.43 (0.98–2.10) 1.51 (0.96–2.38)
– 0.06 0.08
– 1.42 (0.90–2.26) 1.26 (0.73–2.19)
– 0.13 0.41
– 1.93 (1.21–3.09) 1.69 (0.96–2.95)
– o 0.01 0.07
Age 18–24 (reference) 25–34 35 þ Sex, female b
Rural
0.80 (0.53–1.20)
0.28
1.02 (0.62–1.67)
0.95
1.10 (0.67–1.83)
0.70
SI assessment by mental health clinician
1.80 (1.23–2.63)
o 0.01
1.89 (1.21–2.96)
o 0.01
2.42 (1.54–3.78)
o 0.001
Clinical diagnosesc Depressive Disorder Substance Use Disorder Post-traumatic Stress Anxiety
0.91 1.03 1.67 1.64
0.59 0.90 o 0.01 0.03
2.02 0.66 2.37 1.72
o 0.01 0.16 o 0.001 0.048
1.72 0.73 2.36 1.80
0.01 0.29 o 0.001 0.03
SI assessment positive
1.48 (1.03–2.14)
0.04
1.25 (0.80–1.95)
0.32
1.31 (0.84–2.04)
(0.64–1.29) (0.65–1.63) (1.18–2.37) (1.06–2.53)
(1.33–3.07) (0.38–1.17) (1.56–3.61) (1.00–2.95)
(1.13–2.62) (0.41–1.31) (1.54–3.62) (1.05–3.10)
0.24
OEF/OIF ¼ Operation Enduring Freedom/Operation Iraqi Freedom; PHQ-2 ¼Patient Health Questionnaire depression screen; SI ¼suicidal ideation. Odds ratios adjusted for non-independence of observations within sites (AORs) with 95% confidence intervals (CI); analyses included only those with complete data (n ¼449). a
Mental health visits, range 0–8þ ; specialty mental health visits attended subsequent to initial assessment date. Non-Hispanic white vs. other. c International Classification of Diseases, Clinical Modification (ICD-9-CM) codes recorded on date of positive PHQ-2. Substance Use Disorder includes abuse of cannabis, alcohol, opiates, cocaine, amphetamines, polysubstance, and “other,” excluding tobacco. b
152
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
limited, forcing clinicians to choose between standard outpatient treatment regimens or psychiatric hospitalization. It should also be considered that our findings may be indicative of veterans more willing to endorse SI in a mental health setting. Veterans are more comfortable disclosing thoughts of suicide with a provider whom they trust, and they describe thoughts of suicide as particularly private (Ganzini et al., 2013). If the veterans in our study found it more appropriate or more comfortable to disclose SI to a mental health clinician, rather than their primary care provider, the association between SI and subsequent mental health utilization may have been misestimated. However, our results did not support this explanation; although bivariate comparisons showed veterans with SI were more likely to have been assessed for SI by a mental health clinician, this association did not hold when depression diagnoses and PHQ-2 scores were taken into account. Several limitations should be considered in the interpretation of our results. Current relationships between SI assessment and utilization follow-up may differ from what we observed during this time period; our results reflect the first 18 months of VA's establishment of the SI assessment performance standard, the implementation of which varied between VAMCs (Dobscha et al., 2013). The three participating VAMCs are geographically-dispersed and our sample is demographically reflective of the OEF/OIF population (Institute of Medicine, 2010), but generalizability to other VAMCs and non-VA samples may be limited, and our findings reflect mental health services access through, or within, the VA system. While we did not detect a direct effect of the SI assessment result on utilization, above what could be explained by depression symptoms, it is possible the SI assessment program had indirect and unmeasured effects on mental health care, perhaps through increased awareness of suicide risk or increased detection of mental health disorders. Finally, this study focused on brief SI assessment results per se and did not consider the validity and reliability of such measures for detecting SI, an important, but unrelated, issue (Gaynes et al., 2004; Mann et al., 2005; Razykov et al., 2012). To our knowledge, this is the first study among OEF/OIF veterans to specifically address whether SI detection using structured assessments following positive depression screens increases subsequent mental health utilization. As healthcare systems begin to adopt similar suicidal ideation detection or screening procedures, it is important to understand whether these procedures are successful in engaging at-risk patients in mental health care. While our results suggest that SI is not independently associated with increased mental health care utilization in the following year, our findings suggest that promoting veteran self-selection into mental health care, reducing its stigma, and finding ways to support continued engagement for veterans with SI may increase mental health utilization independent of risk detection efforts. In conjunction with SI assessments, developing clearer empirical guidance for treatment and increasing healthcare resources for individuals with suicidal ideation constitute valuable directions for future work.
Disclosures None for any author.
Acknowledgments This material is based upon work supported by the U.S. Department of Veterans Affairs, Veterans Health Administration, and Health Services Research and Development Service Project
DHI-08-096. Dr. Denneson is a core investigator in the Center to Improve Veteran Involvement in Care (CIVIC) at the Portland VA Medical Center. The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
References Bauer, A.M., Chan, Y.F., Huang, H., Vannoy, S., Unutzer, J., 2013. Characteristics, management, and depression outcomes of primary care patients who endorse thoughts of death or suicide on the PHQ-9. Journal of General Internal Medicine 28, 363–369. Betz, M.E., Sullivan, A.F., Manton, A.P., Espinola, J.A., Miller, I., Camargo Jr, C.A., Boudreaux, E.D., 2013. on behalf of the ED-SAFE Investigators, 2013. Knowledge, Attitudes, and Practices of Emergency Department Providers in the Care of Suicidal Patients. Depression and Anxiety 30, 1005–1012. Brenner, L.A., Breshears, R.E., Betthauser, L.M., Bellon, K.K., Holman, E., Harwood, J. E., Silverman, M.M., Huggins, J., Nagamoto, H.T., 2011. Implementation of a suicide nomenclature within two VA healthcare settings. Journal of Clinical Psychology in Medical Settings 18, 116–128. Britton, P.C., Ilgen, M.A., Rudd, M.D., Conner, K.R., 2012. Warning signs for suicide within a week of healthcare contact in Veteran decedents. Psychiatry Research 200, 395–399. Corson, K., Denneson, L.M., Bair, M., Helmer, D.A., Goulet, J.L., Dobscha, S.K., 2013. Prevalence and correlates of suicidal ideation among OEF/OIF Veterans. Journal of Affective Disorders 149, 291–298. Crosby, A.E., Ortega, L., Melanson, C., 2011. Self-Directed Violence Surveillance: Uniform Definitions and Recommended Data Elements, Version 1.0. Centers for Disease Control and Prevention, National Center for Injury Prevention and Control, Atlanta, GA. Department of Veterans Affairs, 2007. Cinical Reminders. 〈http://www.va.gov/ vdl/documents/Clinical/CPRS-Clinical_Reminders/pxrm_2_6_um.pdf〉 (accessed 25.10.11). Department of Veterans Affairs, 2008. Suicide Risk Assessment Guide. 〈www.mental health.va.gov/docs/suicide-risk-assessment-guide.pdf〉 (accessed 27.2.12). Department of Veterans Affairs, 2009. Analysis of Unique Veterans Utilization of VA Benefits & Services. 〈http://www.va.gov/VETDATA/docs/SpecialReports/unique veteransMay.pdf\〉 (accessed 27.2.12). Dobscha, S.K., Corson, K., Helmer, D.A., Bair, M.J., Denneson, L.M., Brandt, C., Beane, A., Ganzini, L., 2013. Brief assessment for suicidal ideation in OEF/OIF veterans with positive depression screens. General Hospital Psychiatry 35, 272–278. Ganzini, L., Denneson, L., Press, N., Bair, M.J., Helmer, D.A., Poat, J., Dobscha, S.K., 2013. Trust is the basis for effective suicide risk assessment in veterans. Journal of General Internal Medicine 28, 1215–1221. Gaynes, B.N., West, S.L., Ford, C.A., Frame, P., Klein, J., Lohr, K.N., 2004. Screening for suicide risk in adults: a summary of the evidence for the U.S. Preventive Services Task Force. Annals of Internal Medicine 140, 822–835. Institute of Medicine, 2010. Returning Home from Iraq and Afghanistan: Preliminary Assessment of Readjustment Needs of Veterans, Service Members, and their Families. The National Academies Press, Washington DC. Jakupcak, M., Cook, J., Imel, Z., Fontana, A., Rosenheck, R., McFall, M., 2009. Posttraumatic stress disorder as a risk factor for suicidal ideation in Iraq and Afghanistan War veterans. Journal of Traumatic Stress 22, 303–306. Kroenke, K., Spitzer, R.L., Williams, J.B., 2001. The PHQ-9: validity of a brief depression severity measure. Journal of General Internal Medicine 16, 606–613. Kroenke, K., Spitzer, R.L., Williams, J.B., 2003. The Patient Health Questionnaire-2: validity of a two-item depression screener. Medical Care 41, 1284–1292. Lemaire, C.M., Graham, D.P., 2011. Factors associated with suicidal ideation in OEF/ OIF veterans. Journal of Affective Disorders 130, 231–238. Lindley, S., Cacciapaglia, H., Noronha, D., Carlson, E., Schatzberg, A., 2010. Monitoring mental health treatment acceptance and initial treatment adherence in veterans: veterans of Operations Enduring Freedom and Iraqi Freedom versus other veterans of other eras. Annals of the New York Academy of Sciences 1208, 104–113. Lish, J.D., Zimmerman, M., Farber, N.J., Lush, D.T., Kuzma, M.A., Plescia, G., 1996. Suicide screening in a primary care setting at a Veterans Affairs Medical Center. Psychosomatics 37, 413–424. Lu, M.W., Carlson, K.F., Duckart, J.P., Dobscha, S.K., 2012. The effects of age on initiation of mental health treatment after positive PTSD screens among Veterans Affairs primary care patients. General Hospital Psychiatry 34, 654–659. Lu, M.W., Duckart, J.P., O'Malley, J.P., Dobscha, S.K., 2011. Correlates of utilization of PTSD specialty treatment among recently diagnosed Veterans at the VA. Psychiatric Services 62, 943–949. Mann, J.J., Apter, A., Bertolote, J., Beautrais, A., Currier, D., Haas, A., Hegerl, U., Lonnqvist, J., Malone, K., Marusic, A., Mehlum, L., Patton, G., Phillips, M., Rutz, W., Rihmer, Z., Schmidtke, A., Shaffer, D., Silverman, M., Takahashi, Y., Varnik, A., Wasserman, D., Yip, P., Hendin, H., 2005. Suicide prevention strategies: a systematic review. Journal of the American Medical Association 294, 2064–2074. Mann, J.J., Ellis, S.P., Waternaux, C.M., Liu, X., Oquendo, M.A., Malone, K.M., Brodsky, B.S., Haas, G.L., Currier, D., 2008. Classification trees distinguish suicide
L.M. Denneson et al. / Psychiatry Research 217 (2014) 147–153
attempters in major psychiatric disorders: a model of clinical decision making. Journal of Clinical Psychiatry 69, 23–31. McCarthy, J.F., Blow, F.C., Valenstein, M., Fischer, E.P., Owen, R.R., Barry, K.L., Hudson, T.J., Ignacio, R.V., 2007. Veterans Affairs Health System and mental health treatment retention among patients with serious mental illness: evaluating accessibility and availability barriers. Health Services Research 42, 1042–1060. Nielsen, B., Wang, A.G., Brille-Brahe, U., 1990. Attempted suicide in Denmark. IV. A five-year follow-up. Acta Psychiatrica Scandinavica 81, 250–254. O'Connor, E., Gaynes, B.N., Burda, B.U., Soh, C., Whitlock, E.P., 2013. Screening for and treatment of suicide risk relevant to primary care: a systematic review for the U.S. Preventive Services Task Force. Annals of Internal Medicine 158, 741–754. O'Neil, M.E., Peterson, K., Low, A., Carson, S., Denneson, L.M., Haney, E., Shiroma, P., Kansagara, D., 2012. Suicide Prevention Interventions and Referral/Follow-up Services. A Systematic Review. VA-ESP Project No. 05-225. Pietrzak, R.H., Goldstein, M.B., Malley, J.C., Rivers, A.J., Johnson, D.C., Southwick, S. M., 2010. Risk and protective factors associated with suicidal ideation in veterans of Operations Enduring Freedom and Iraqi Freedom. Journal of Affective Disorders 123, 102–107. Razykov, I., Ziegelstein, R.C., Whooley, M.A., Thombs, B.D., 2012. The PHQ-9 versus the PHQ-8–is item 9 useful for assessing suicide risk in coronary artery disease patients? Data from the Heart and Soul Study. Journal of Psychosomatic Research 73, 163–168. Schulberg, H.C., Lee, P.W., Bruce, M.L., Raue, P.J., Lefever, J.J., Williams, J.W., Dietrich, A.J., Nutting, P.A., 2005. Suicidal ideation and risk levels among primary care
153
patients with uncomplicated depression. Annals of Family Medicine 3, 523–528. Seal, K.H., Bertenthal, D., Maguen, S., Gima, K., Chu, A., Marmar, C.R., 2008. Getting beyond "Don't ask; don't tell": an evaluation of US Veterans Administration postdeployment mental health screening of veterans returning from Iraq and Afghanistan. American Journal of Public Health 98, 714–720. Seal, K.H., Maguen, S., Cohen, B., Gima, K.S., Metzler, T.J., Ren, L., Bertenthal, D., Marmar, C.R., 2010. VA mental health services utilization in Iraq and Afghanistan veterans in the first year of receiving new mental health diagnoses. Journal of Traumatic Stress 23, 5–16. Szanto, K., Mulsant, B.H., Houck, P., Dew, M.A., Reynolds 3rd, C.F., 2003. Occurrence and course of suicidality during short-term treatment of late-life depression. Archives of General Psychiatry 60, 610–617. Tabachnick, B.G., Fidell, L.S., 2001. Using Multivariate Statistics, 4 ed. Allyn and Bacon, Boston. Uebelacker, L.A., German, N.M., Gaudiano, B.A., Miller, I.W., 2011. Patient health questionnaire depression scale as a suicide screening instrument in depressed primary care patients: a cross-sectional study. Primary Care Companion for CNS Disorders 13. Vannoy, S.D., Tai-Seale, M., Duberstein, P., Eaton, L.J., Cook, M.A., 2011. Now what should I do? Primary care physicians' responses to older adults expressing thoughts of suicide. Journal of General Internal Medicine 26, 1005–1011. West, A.N., Lee, R.E., Shambaugh-Miller, M.D., Bair, B.D., Mueller, K.J., Lilly, R.S., Kaboli, P.J., Hawthorne, K., 2010. Defining "rural" for veterans' health care planning. Journal of Rural Health 26, 301–309.
