General Hospital Psychiatry 36 (2014) 310–317

Contents lists available at ScienceDirect

General Hospital Psychiatry journal homepage: http://www.ghpjournal.com

Primary care clinician responses to positive suicidal ideation risk assessments in veterans of Iraq and Afghanistan Steven K. Dobscha, M.D. a, b,⁎, Lauren M. Denneson, Ph.D. a, b, Anne E. Kovas, M.P.H. a, b, Kathryn Corson, Ph.D. a, b, Drew A. Helmer, M.D., M.S. c, d, Matthew J. Bair, M.D., M.S. e a b c d e

VA HSR&D Center to Improve Veteran Involvement in Care (CIVIC), Portland Veterans Affairs Medical Center, Portland, OR, USA Department of Psychiatry, Oregon Health & Science University, Portland, OR, USA War-Related Illness and Injury Study Center, VA New Jersey Health Care System, East Orange, NJ, USA Rutgers University, New Jersey Medical School, Newark, NJ, USA Richard L. Roudebush Veteran Affairs Medical Center, Indianapolis, IN, USA

a r t i c l e

i n f o

Article history: Received 15 August 2013 Revised 21 November 2013 Accepted 25 November 2013 Keywords: Depression Screening Suicide Veterans

a b s t r a c t Objective: To examine primary care clinician actions following positive suicide risk assessments administered to Operation Enduring Freedom/Operation Iraqi Freedom (OEF/OIF) veterans. Methods: We identified OEF/OIF veterans with positive templated suicide risk assessments administered in primary care settings of three Veterans Affairs (VA) Medical Centers. National VA datasets and manual record review were used to identify and code clinician discussions and actions following positive assessments. Bivariate analyses were used to examine relationships between patient characteristics and discussions of firearms access and alcohol/drug use. Results: Primary care clinicians documented awareness of suicide risk assessment results for 157 of 199 (79%) patients with positive assessments. Most patients were assessed for mental health conditions and referred for mental health follow-up. Clinicians documented discussions about firearms access for only 15% of patients. Among patients whose clinicians assessed for substance abuse, 34% received recommendations to reduce alcohol or drug use. Depression diagnoses and suicidal ideation/behavior severity were significantly associated with firearms access discussions, while patient sex, military service branch, and substance abuse diagnoses were significantly associated with recommendations to reduce substance use. Conclusion: Greater efforts are needed to understand barriers to clinicians' assessing, documenting and counseling once suicidal ideation is detected, and to develop training programs and systems changes to address these barriers. Published by Elsevier Inc.

1. Introduction Veterans utilizing Veterans Health Administration (VHA) services die by suicide at a higher rate than the general population [1], and while specific rates of suicide among veterans of Operation Enduring Freedom and Operation Iraqi Freedom (OEF/OIF) are unknown, OEF/OIF veterans have high rates of previously documented risk factors for suicide including depression, substance use disorder (SUD), posttraumatic stress disorder (PTSD), chronic pain and traumatic brain injury [2–8]. Furthermore, the rate of suicide among active duty OEF/OIF soldiers has increased in recent years, a trend which could follow this group into postdeployment [9]. The Department of Veterans Affairs (VA) has implemented a multimodal strategy to improve detection and response to suicide risk

⁎ Corresponding author. Portland VA Medical Center P.O. Box 1034 (R&D 66) Portland, OR 97207, USA. Tel.: +1-503-220-8262x52207. E-mail address: [email protected] (S.K. Dobscha). 0163-8343/$ – see front matter. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.genhosppsych.2013.11.007

[10]. As one part of this strategy, the VA designated assessment for possible suicidal ideation (SI) among veterans at higher risk for suicide as a national performance goal. Specifically, since 2007, brief templated, structured suicide risk assessment tools are administered across the VA as part of routine care following positive depression and PTSD screens. These tools, typically consisting of less than five items, are designed to be used in conjunction with clinical judgment to assess suicide risk. Depression and PTSD screening, and the use of templated suicide risk assessments to assess for SI, frequently occur in primary care settings [11]. Prior research indicates that up to half of individuals have contact with primary care clinicians in the month prior to suicide, while a smaller proportion has contact with mental health care clinicians during that month [12,13]. As such, primary care clinicians may play a critical role in addressing suicide risk of veterans by detecting and treating important mental and general medical conditions, and being prepared to identify and intervene when veterans are at high risk. However, despite these high contact rates, there are substantial gaps in our knowledge of how primary care clinicians address suicide risk.

S.K. Dobscha et al. / General Hospital Psychiatry 36 (2014) 310–317

A recent study by Smith et al. [14] showed that a minority (34%) of VA patients with a history of depression had a mental health diagnosis coded during final nonmental health visits within 30 days prior to suicide, and only 41% were receiving recommended dosages of antidepressants. Vannoy et al. [15,16] found that while primary care clinicians may use language and communication approaches that support patient disclosures of SI, primary care clinicians are frequently challenged to go beyond assessment to develop welldefined and structured treatment plans when they encounter SI. Aside from this small group of studies, little is known about the specific actions primary care clinicians take when they encounter SI in their patients. Two risk factors of particular importance for clinicians to address with veterans who disclose SI are firearms access and alcohol and drug use. Nearly three quarters of veteran suicides in Oregon between 2000 and 2005 were firearms deaths [13]. Other studies suggest that veterans are significantly more likely to use firearms as a means for suicide than nonveterans [17]. National survey data have shown that individuals with an SUD have a sixfold greater likelihood of a lifetime suicide attempt than those without SUDs [18]. Among veteran suicide decedents in Oregon who received care in the VA healthcare system during the year prior to death, 20% were given an SUD diagnosis during that year, the second most common mental health disorder diagnosed after mood disorders [13,19]. In addition, it is welldocumented that intoxication frequently precedes suicide attempts and death by suicide [20,21], especially among those who die by gunshot wound [21]. The main objective of the current study was to describe primary care clinician discussions and clinical actions following depression screening and positive suicide risk assessments (indicating SI) administered to OEF/OIF veterans. A secondary objective was to identify correlates of documented clinician–patient discussions and actions related to two key suicide risk factors: discussions of firearms access and counseling to reduce alcohol or drug use.

2. Methods This study was approved by the institutional review boards of the participating medical centers. Methods for the overall research project have been described in prior publications [11,22].

311

United States. Each VAMC is closely affiliated with a local university, has active research and teaching activities and has both metropolitan and rural-based primary care clinics. Each utilizes VA's computerized patient record system (CPRS), an integrated electronic medical record that contains progress notes, pharmacy and laboratory data. Brief screens for a variety of conditions and disorders including depression are administered routinely in the VA, facilitated by a reminder system embedded within CPRS. Specific screen items and scoring algorithms often vary by VA site. For veterans not currently receiving depression care, the VA requires annual depression screening using the 2-item or 9-item Patient Health Questionnaire (PHQ-2 or PHQ-9—the first two items of the PHQ-9 are the PHQ-2) [23]. A positive depression screen triggers a reminder to clinicians to conduct a suicide risk assessment. A suicide risk assessment tool based on questions from the National Clinical Reminder (NCR) Suicide Assessment Questions [24] was disseminated in late 2007; this five-item tool includes questions about hopelessness, thoughts of taking one's life and prior suicide attempts [24]. Individual VA facilities could choose to use this tool, the PHQ-9 (which includes the ninth item, “Over the past 2 weeks, how often have you been bothered by thoughts that you would be better off dead or of hurting yourself in some way?”), or develop their own assessment tool. The NCR tool and PHQ-9 were used as the suicide risk assessment tools most of the time (98%) at the three sites during the study period (Table 1) [11]. 2.2. Subjects We first identified veterans of OEF/OIF who had a recorded contact at one of the three VAMCs between April 1, 2008 and September 30, 2009. OEF/OIF status was determined using the national OEF/OIF Roster [11]. OEF/OIF Roster data were then matched to data from the VA Decision Support System (DSS), a national database containing demographic, clinical, utilization and cost information [25]. To be included in the current study, a veteran had to (a) have a positive depression screen conducted in a primary care setting at one of the three VAMCs; (b) a suicide risk assessment conducted using the NCR Suicide Assessment tool or the PHQ-9th item, conducted within a month of the positive depression screen; and (c) a positive suicide risk assessment result (Fig. 1). 2.3. Measures

2.1. Settings This study was conducted at three VA Medical Centers (VAMCs) located in the northwest, southwest and northeast regions of the

2.3.1. Independent variables Most demographic information, including race/ethnicity data, was extracted from DSS. The OEF/OIF Roster contained education, military

Table 1 Most commonly used templated suicide risk assessment tools Construct

Measure

Response options

Positive result

Depression

PHQ-2 Over the past 2 weeks, how often have you been bothered by: 1. Little interest or pleasure doing things? 2. Feeling down, depressed or hopeless?

0 = Not at all 1 = Several days 2 = NHalf the days 3 = Nearly every day

Sum ≥ 3

PHQ-9, 9th item Over the past 2 weeks, how often have you been bothered by thoughts that you would be better off dead or of hurting yourself in some way?

0 = Not at all 1 = Several days 2 = NHalf the days 3 = Nearly every day

Score ≥ 1

Yes/No

Item 2 = yes

Suicidal Ideation

VA NCR Pocket Card Risk Assessment 1. Feeling hopeless about the present/future? 2. Thoughts of taking your life? 3. When did you have these thoughts? 4. Do you have a plan to take your life? 5. Have you ever had a suicide attempt?

312

S.K. Dobscha et al. / General Hospital Psychiatry 36 (2014) 310–317

Fig. 1. Participant flow diagram. Depression screening conducted with the PHQ-2, or the sum of the first two items of the PHQ-9; PHQ-2≥3 is positive depression screen, range 3–6. Suicide risk assessment=use of either the PHQ-9 9th item, or the VA Pocket Card Risk Assessment. Primary care clinician awareness of the positive risk assessment includes n=87 who conducted the assessment themselves.

service branch and additional race/ethnicity data (which were used to supplement DSS data when race data were missing). Using the date of the first depression screen administered during the study period, manual medical record review was conducted to determine PHQ-2 and subsequent suicide risk assessment results [10]. A PHQ-2 score of 3 is the cutoff for a positive depression screen [23,26]. If the PHQ-9 had been administered instead, we calculated the PHQ-2 score using the first two PHQ-9 items. The NCR Suicide Assessment tool was considered positive if the second item, “thoughts of taking your life,” was endorsed (yes/no). For the PHQ-9th item, a score of 1 or greater (several days or more) was considered positive [26]. Prior lifetime suicide attempt was coded as present when a patient had endorsed the fifth item on the NCR Suicide Assessment tool regarding past attempts. Additional information about prior attempts, timing and severity of suicide ideation and behaviors was captured from documentation in any clinician progress note from the day of or the day following the suicide risk assessment. We specifically coded for indicators of greater SI or suicide-related behavior severity, namely recent (within 1 month) SI, suicide intent, preparatory behavior or suicide attempt. Preparatory behavior is defined as engaging in preparatory actions for a suicide attempt, beyond verbalizations or thoughts, but before potential for harm has begun [27]. Finally, from DSS, we extracted International Classification of Diseases (ICD)-9Clinical Modification diagnoses recorded in CPRS on the dates of depression screen administration. Specific diagnoses examined included depression, SUD, PTSD and other anxiety disorders. 2.3.2. Dependent variables Manual medical record review was used to determine primary care clinician actions, which included documented discussion, consideration or assessment of specific risk factors, as well as treatment planning actions such as prescribing psychiatric medications, or arranging mental health follow-up, following positive suicide risk assessments. We created a checklist of clinical actions based on known risk factors for suicide [4], guidelines pertinent to evaluating

suicide risk [5,28] and prior research conducted by our research team using manual record review to evaluate treatment provided in primary care settings [19,29,30] (Table 2). We identified clinical actions that occurred on the same or next day from the time the suicide risk assessment was documented. Checklist items were scored dichotomously (present or not) to reflect whether the primary care clinician documented discussion or consideration of risk factors or made treatment changes, and for risk factors, that the documentation indicated the risk factor was present within the prior month. The record review tool underwent several revisions, wherein pairs of reviewers rated groups of 30 randomly selected medical records to refine operational definitions. Agreement after several rounds of revision and review was excellent (kappaN=0.75). All cases were double-rated, and final definitions and coding were decided through an arbitration process that included consultation with two of the authors (SD, LD). 2.4. Analyses We first described demographic and clinical characteristics of the main study group. Demographic variables included race/ethnicity, age, sex, education, rural/urban residence and reserve or National Guard military branch status. Based on raw distributions, age was separated into three categories (≤ 24, 25–34, ≥35), and education was collapsed into two (high school diploma/equivalent or less; some college or higher). Because race/ethnicity categories in the OEF/OIF dataset are treated as mutually exclusive and DSS utilizes a “multiple categories” option, we collapsed race/ethnicity categories into White, nonHispanic versus other. Clinical variables included PHQ-2 score, mental health diagnoses made on the date of depression screening (depressive disorder, SUD, PTSD and anxiety) and suicide behavior/timing category. PHQ scores were dichotomized (3–4 vs. 5–6). We described the proportion of patients whose clinicians or care team staff assessed for specific risk factors, and within these groups, identified the proportion of patients who endorsed specific risk factors.

S.K. Dobscha et al. / General Hospital Psychiatry 36 (2014) 310–317

313

Table 2 Operational definitions of suicide behavior severity and risk factors used in manual medical record review Suicide behaviors severity SI Suicidal intent Preparatory behavior Previous suicide attempts (suicidal self-directed violence) Risk factor domainsa Access to firearm(s) Mental health disorder(s) Partner/Relationship difficulties Legal problems Housing insufficiency Occupational difficulties Financial burdens Pain Drug and/or alcohol abuse

Treatment planning Psychiatric medication Mental health follow-up discussed or arranged Recommendation to restrict firearm access Recommendation to reduce alcohol or other drug use

Thoughts about engaging in suicide-related behavior. Intending to kill oneself, with understanding consequences of these actions; includes a specific plan/means of killing oneself. Acts or preparation towards engaging in self-directed violence beyond a verbalization or thought, before potential for injury has begun (e.g., buying a gun, writing a note). A previous suicide attempt or attempts occurring in lifetime. Self-directed behavior that deliberately results in injury or the potential for injury to oneself.

Patient ownership of, or access to, firearms. Assessment, consideration, or discussion of mental health disorders or symptoms, including substance abuse. Problems with a spouse or intimate partner. Any involvement in active, ongoing or pending legal process. Homelessness, unstable housing, or housing constraints (e.g., crowding). Job dissatisfaction, employment status or lack of desired employment. Documentation addressing any concerns about financial situation. Pain impact on physical, social, or occupational functioning; or formal pain assessment (e.g., 0–10 rating). Evaluation of substance use or abuse. Any illicit drug use was considered abuse, except, marijuana use was counted as endorsed only if clinician documented that the use was considered problematic. For purposes of describing the sample, problematic drug or alcohol use was assumed if provider made a recommendation to reduce or restrict substance use. Did not include caffeine or nicotine.

Clinician started, stopped, changed, continued medication or increased or decreased dosage. Clinician discussed or arranged a new mental health appointment or acknowledged/reminded patient of an upcoming mental health appointment. Clinician discussed removing patient's access to firearms or firearm safety. Clinician discussed reducing or stopping alcohol or substance use. Did not include caffeine or nicotine.

a Risk factor items were coded as discussed (yes/no) if documentation of the topic was present in the clinic notes on the same day or the day following the positive suicide risk assessment; risk factor items were coded as endorsed (yes/no) if documentation on the same day or day following the positive suicide risk assessment indicated that the risk factor had been present within the prior month.

We next examined bivariate associations among demographic and clinical variables and primary care clinician discussion of firearms access and, within the subgroup whose primary care clinician asked about alcohol or drug use, recommendations to reduce or restrict alcohol/drug use. Discussions exploring for the presence or absence of firearms were collapsed with discussions involving specific recommendations to restrict firearms access because there were few occurrences of specific recommendations to reduce firearms access. We explored for associations between one or more indicators of SI/behavior severity and firearms discussions. Due to low frequency, anxiety disorders were not included in bivariate analyses. Small cell sizes also precluded the use of multivariate analyses. All analyses were conducted using IBM SPSS Statistics 21.0 (Chicago, IL), with two-sided alpha=.05 unless otherwise stated. 3. Results Across the three study sites, 1017 veterans had positive depression screens administered in primary care over the study period. Seven hundred twenty-two (71%) were subsequently administered suicide risk assessments in primary care; 97% of these were conducted within 1 day of depression screens. Among these 722 veterans, 199 (28%) had positive suicide risk assessments. For 157 (79%) of the patients with positive assessments, either their primary care clinicians conducted the assessment themselves (n= 87, 55%) or documented awareness of the assessment result when the assessment had been administered by another member of the care team (n= 70, 45%); this group constitutes the study group for our main analysis (Fig. 1). 3.1. Patient characteristics Table 3 shows the demographic and clinical characteristics of our main study group. The majority of patients were non-Hispanic White, male, had up to a high school diploma and were living in

urban areas; just over half were between 25 and 34 years of age. Almost two-thirds had PHQ-2 scores ≥5, and 80% were diagnosed with at least one psychiatric disorder (depression, SUD, PTSD or anxiety) on the date of depression screening. Documented pain (82%), financial burdens (71%), concerns about alcohol or drug abuse (60%) and occupational difficulties (46%) were common. Approxi-

Table 3 Characteristics of veterans whose clinicians were aware of positive suicide risk assessment results, n=157 Characteristic Age ≤24 25–34 ≥35 Male White, non-Hispanica Education ≤High school diploma Some college or higher Urban residenceb Reserve or National Guardc PHQ2≥5d Diagnoses on index date Depressive disorder SUD PTSD Anxiety disorder Greater SI severity (within 1 month SI/intent/prep. behav./attempts)

n

%

36 87 34 136 106

(22.9) (55.4) (21.7) (86.6) (70.2)

143 13 119 49 99

(91.7) (8.3) (75.8) (31.2) (63.1)

91 23 75 13 55

(58.0) (14.6) (47.8) (8.3) (35.0)

Index date=date of the veteran's positive PHQ-2 screen; Documented SI=SI noted; prior month SI/intent/prep/attempts=one or more indicators of greater SI severity noted (SI, suicidal intent, preparatory behaviors or suicide attempt) noted as occurring within the past month. a Versus other. b Versus rural residence. c Versus active duty. d PHQ-2≥3 is positive screen for depression, range 3–6.

314

S.K. Dobscha et al. / General Hospital Psychiatry 36 (2014) 310–317

Table 4 Documented primary care clinician actions after positive suicide risk assessments, n=157

Risk factor discusseda Access to firearms Partner/Relationship difficulties Involved in legal process Housing insufficiency Occupational difficulties Financial burdens Pain MH disorders Drug and/or alcohol abuse Treatment planningb Psychiatric prescription continuation or changec Mental health follow-up discussed or arranged

n

%

23 111 12 61 110 12 146 154 138

(14.6) (70.7) (7.6) (38.9) (70.1) (7.6) (93.0) (98.1) (87.9)

69 142

(43.9) (90.4)

a Risk factors were recorded as discussed if notation on the same or next day from when the suicide risk assessment was documented reflected consideration of the risk factor. b Treatment planning items were recorded as present if noted on the same or next day from when the suicide risk assessment was documented. c Psychiatric medication prescription continuation or change=notation of a new or continuing prescription of any mental health related medication (e.g., antidepressants, benzodiazepines).

mately one third of the group had at least one indicator of greater SI/ behavior severity. 3.2. Primary care clinician actions Table 4 displays frequencies of the discussions and actions of primary care clinicians following positive suicide risk assessments. Primary care clinicians documented assessment or consideration of mental health conditions for all but one veteran and arranged or noted previously planned mental health follow-up for 90% of patients. Primary care clinicians frequently documented discussions about drug or alcohol use (n=138; 88%), and specific recommendations to

reduce or restrict use of alcohol or drugs were documented for 47 of the 138 (34%). Primary care clinicians frequently documented discussing occupational (70%) and partner/relationship issues (71%) and pain problems (93%). However, primary care clinicians documented discussions about firearms access for only 15% of patients, and specific recommendations to restrict firearms access were found for just three patients. Table 5 shows bivariate associations between patient characteristics and (a) any primary care clinician discussions related to firearms (n= 157) and, (b) within the subsample of 138 veterans whose primary care clinician had asked about drug or alcohol use, primary care clinician recommendations to reduce alcohol or drug use. Veterans with diagnosed depressive disorders (P= .03) and those with indicators of more severe SI/behaviors (Pb .01) were significantly more likely to be asked about firearms access. Men (P= .01), veterans who had served in the reserve or National Guard (vs. active duty) (P= .03) and veterans given SUD diagnoses (Pb.001) were significantly more likely to be counseled to reduce or restrict alcohol or drug use by their primary care clinicians.

4. Discussion This is the first study to examine primary care clinician actions taken in response to routinely administered brief, templated assessments for SI. There were a number of encouraging findings. Primary care clinicians usually documented awareness of positive suicide risk assessment results and almost always assessed for mental health conditions, including SUDs. Mental health referrals or acknowledgement of upcoming appointments were almost universal. Clinicians usually discussed some psychosocial stressors with their patients. However, documented primary care clinician recommendations to reduce or restrict use of alcohol or drugs took place only a third of the time, and documented discussions about firearms access were infrequent. Patients with more recent or severe SI or other suicide-related behaviors were more likely to discuss firearms access

Table 5 Demographic and clinical factors associated with primary care clinician discussions of firearm access and recommendations made to reduce alcohol use Firearm access discussed (n=157)a

Male Age ≤24 25–34 ≥35 White, non-Hispanicc Education ≤High school diploma Some college or higher Urban residenced Reserve or National Guarde PHQ≥5f Diagnoses Depressive disorder SUD PTSD Prior month SI/intent/prep/ attemptsg a

Recommendation to reduce alcohol/drug use (n=138)b

No (n=134)

Yes (n=23)

P

No (n=92)

Yes (n=46)

P

115 (85.8)

21 (91.3)

.48 .07

76 (82.6)

45 (97.8)

.01 .74

27 (20.1) 79 (59.0) 28 (20.9) 87 (68.0)

9 (39.1) 8 (34.8) 6 (26.1) 19 (82.6)

22 53 17 61

11 24 11 31

123 (92.5) 10 (7.5) 104 (77.6) 41 (30.6) 82 (61.2)

20 (87.0) 3 (13.0) 15 (65.2) 8 (34.8) 17 (73.9)

73 (54.5) 20 (14.9) 60 (44.8) 40 (29.9)

18 (78.3) 3 (13.0) 15 (65.2) 15 (65.2)

.16 .38

(23.9) (57.6) (18.5) (68.5)

(23.9) (52.2) (23.9) (70.5)

.20 .69 .24

84 (92.3) 7 (7.7) 74 (80.4) 23 (25.0) 59 (64.1)

43 (93.5) 3 (6.5) 35 (76.1) 20 (43.5) 26 (56.5)

.03 .81 .07 b.01

54 (58.7) 6 (6.5) 45 (48.9) 33 (35.9)

27 17 20 16

(58.7) (37.0) (43.5) (34.8)

.82 .80

.55 .03 .39 1.00 b.001 .55 .90

Firearms discussed: n=157; six missing race/ethnicity data, one missing education data. Recommendation to reduce alcohol or drug use: analyses conducted within the group whose primary care provider had asked about current alcohol/drug use, n=138; five missing race/ethnicity data, one missing education data. c Versus other. d Versus rural residence. e Versus active duty. f PHQ-2≥3 is positive depression screen, range 3–6. g Prior month SI/intent/prep/attempts=One or more indicators of greater SI severity noted (SI, suicidal intent, preparatory behaviors, or suicide attempt) noted as occurring within the past month. b

S.K. Dobscha et al. / General Hospital Psychiatry 36 (2014) 310–317

with their primary care clinicians. Women and members of the reserves or National Guard were less likely to have medical record documentation of counseling to reduce or restrict substance use. That clinicians frequently considered mental health conditions when evaluating these patients is not surprising, given that a positive depression screen was required for inclusion in the sample. The majority of patients also reported pain; many had substance abuse problems, occupational, financial or legal difficulties; and one third reported relationship problems. These results are consistent with prior studies showing that the rate of psychosocial stressors in the OEF/OIF population, especially in the postdeployment period, is high [31,32]. Our findings further suggest that patients with SI frequently experience multiple psychosocial stressors as well as clinical conditions such as major depression, SUD or pain. The finding that clinicians frequently conducted or acknowledged suicide risk assessment results and assessed for mental health conditions may reflect VA efforts to enhance suicide prevention efforts system wide. Over the past few years, the VA has developed and promoted multiple approaches to improve detection and response to suicide risk [10], including enhancing access to mental health services, implementing a wide range of education programs for VA staff, veterans and caregivers, and establishing national toll-free crisis hotline and Web-based chat programs [33]. In addition, VA's electronic medical record utilizes a number of automated clinical reminders to help clinicians attend to high-priority clinical issues; clinicians are accustomed to responding to clinical reminder results, and mental health referrals are easily made using the CPRS consult package. On the other hand, as compared to more biomedical or traditional clinical problems such as symptoms of depression, substance abuse or pain, primary care clinicians infrequently documented discussions regarding important psychosocial issues, principally legal involvement, partner/relationship problems and housing. Yet, these factors, in addition to occupational and financial factors, have been associated with suicide risk in multiple studies [4]. While documentation regarding the presence or absence of alcohol or drug abuse was common, specific recommendations to reduce alcohol or drug use were given to just 34% of patients with positive suicide risk assessments whose primary care clinicians discussed drug or alcohol use. We found that 17 of 21 women in the sample had an assessment for drug or alcohol abuse documented in their records, yet only one was counseled to reduce drug or alcohol use. These findings are consistent with prior work showing that women are less likely to receive counseling about alcohol use [34–36], and may reflect gender stereotypes regarding problematic alcohol or substance use, or that women may endorse such abuse less frequently than men. However, in our sample, males and females endorsed alcohol or drug abuse at statistically similar rates (63% vs. 41%, χ 2= 2.91, P=.09). Our results suggest that increasing SI severity is associated with firearms discussions; however even among those with indicators of more severe ideation, less than a third had these discussions. This finding is consistent with several prior studies suggesting that physicians frequently do not discuss firearms with their patients [37,38]. In one of these studies (Walters et al., 2012), qualitative interviews were conducted with VA clinicians and leaders, patients and family members to explore perceptions of gun safety and of interventions to delay gun access during high-risk periods. These stakeholders felt that discussing and enhancing gun safety was important. However, the clinicians expressed concerns that discussing gun access would negatively impact the therapeutic alliance and that some patients might not be honest about gun access. In addition, due to their own lack of knowledge or personal experience with guns, some clinicians did not feel comfortable discussing firearms. This is especially unfortunate, as veterans frequently own firearms and are more likely to use firearms as a means for suicide compared to

315

nonveterans [17]. A review of suicide prevention approaches has suggested that restricting access to lethal means such as firearms is one of two strategies shown to be effective in reducing suicide deaths [39]. Indeed, the VA Office of the Medical Inspector and the Geriatrics and Extended Care Group has advised routine screening of older veterans for access to guns and provision of gun safety information, and the VA Office of Mental Health Services—Suicide Prevention has developed a safety planning process that incorporates discussion of firearms access. Of special concern may be interactions of risk factors. Our data show that over half of the veterans in the sample were considered to have some type of alcohol/drug problem by their clinicians. At the same time, several data sources suggest that veterans are more likely to own firearms compared to individuals in the general population [17,40]. The combination of alcohol use and firearm access may generate special risks [21]. Individuals intoxicated at the time of suicide death are more likely to choose more fatal means of suicide than those not intoxicated; intoxication has been found to increase the likelihood of using a firearm by over 75% [20]. Finally, in one study, over half of veteran suicide decedents with an SUD diagnosis received VA healthcare during the 30 days prior to death [41]. Due to the prevalence and potential impact of substance use and firearms ownership, clinicians should assess and address active SUDs as well as firearm access among veterans who present with increased risk for suicide, especially younger veterans who may be at increased risk for violent deaths [42,43]. Taken together, our findings suggest that primary care clinicians would benefit from additional training and/or support regarding how to respond and document when they encounter increased suicide risk in their patients. In other words, training primary care clinicians and other front-line providers in how to detect suicidality is not sufficient. Prior work has suggested that clinicians are often uncertain about how to respond to patient disclosures of suicidal thoughts [15,44]. While this uncertainty may reflect a lack of clarity regarding the extent to which, or for whom, treatment should be intensified when SI is present [39,45,46], our results suggest that certain types of important information may be especially challenging for clinicians to discuss. To this extent, the Suicide Prevention Resource Center, in collaboration with the American Foundation for Suicide Prevention has developed the Counseling on Access to Lethal Means training program [47], which is designed to train clinicians and other providers to help young patients and their families reduce access to lethal means. Adapting this training for clinicians of other patient populations who may be at risk for suicide, perhaps adding a component related to counseling about substance use, deserves consideration. There are a number of important limitations to this study. The study was conducted at three large, VAMCs with close academic affiliations, and we only examined OEF/OIF veterans; the findings may not be generalizable to smaller VAs, to veterans of other eras or to other patient and primary care clinician populations. However, the VAMCs included were located in different regions of the US and in post hoc analyses (not shown), rates of discussions of firearms use and alcohol use were not substantially different when comparing the three sites, which may enhance generalizability of our findings. Unfortunately our relatively small sample size limited our bivariate analyses and did not allow us to conduct multivariate analyses; some of our significant findings may reflect confounding by other variables. For example, we suspect that military branch status may not be independently related to primary care clinician recommendations to reduce substance use; post hoc analyses indicate that reserve/ National Guard individuals are also more likely to be older (≥ 35; 42% vs. 10%, Pb.001) and more highly educated (Nhigh school; 17% vs. 3%, Pb .01) than active duty status veterans. Thus, age and education status may confound the relationship we observed between military branch status and counseling to reduce or restrict alcohol/drug use.

316

S.K. Dobscha et al. / General Hospital Psychiatry 36 (2014) 310–317

Importantly, our results are highly dependent on the accuracy and extent of providers' documentation in medical records. It is likely that more topics were discussed between patients and clinicians than were captured in progress notes, and clinicians may have documented more frequently about risk factors when they were present than when they were not present. On the other hand, due to medical–legal concerns, clinicians facing time constraints may be more likely to document suicide risk factors than other types of clinical issues. The VA designated assessment for possible SI among veterans with positive depression or PTSD screens as a national performance goal in 2007, which encourages documentation regarding detection of SI. It is also likely that the clinicians who documented severity of SI were more likely to document more carefully other suicide risk factors, which may have contributed to the association we found between SI severity and documentation of key risk factors. Finally, suicide prevention efforts have continued to evolve in the VA since the study period, including changes in training and support for clinicians; this should be kept in mind when considering next steps. 5. Conclusion Our results suggest that barriers remain to clinicians querying and counseling veterans at risk for suicide, including discussion of firearms access and psychosocial stressors relevant to suicide risk. Women with SI may be less likely to receive counseling to reduce use of alcohol and drugs. Our findings call upon investigators and other stakeholders to further study clinician, patient and system barriers to exploring and documenting key risk factors for patients at risk for suicide and to develop effective education and support to help clinicians respond appropriately when risk factors are detected. Acknowledgements We gratefully acknowledge Kathryn Dickinson, MPH; Megan Crutchfield, MPH; Anna Beane, BA; Joseph Warren, MA; and April Wilson, MPH, for assistance with reviewing medical records and coding and organizing data. The research reported here was supported by the Department of VA, VHA, Health Services Research and Development Service projects DHI-08-096 and IIR 10–331. Dr. Dobscha is Director of the Center to Improve Veteran Involvement in Care at the Portland VAMC. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of VA. References [1] Blow FC, Bohnert ASB, Ilgen MA, Rosalinda I, McCarthy JF, Valenstein MM, et al. Suicide mortality among patients treated by the Veterans Health Administration from 2000 to 2007. Am J Public Health 2012;102:S98–S104. [2] Seal KH, Maguen S, Cohen B, Gima KS, Metzler TJ, Ren L, et al. VA mental health services utilization in Iraq and Afghanistan veterans in the first year of receiving new mental health diagnoses. J Trauma Stress 2010;23:5–16. [3] Schlenger WE, Kulka RA, Fairbank JA, Hough RL, Jordan BK, Marmar CR, et al. The psychological risks of Vietnam: the NVVRS perspective. J Trauma Stress 2007;20: 467–79. [4] Martin, J., Ghahramanlou-Holloway, M., Lou, K., Tucciarone, P. A comparative review of U.S. Military and civilian suicide behavior: Implications for OEF/OIF suicide prevention efforts. J Ment Health 2009;31:101,118; 101. [5] Haney EM, O'Neil ME, Carson S, Low A, Peterson K, Denneson LM, et al. Suicide risk factors and risk assessment tools. Washington, DC: Department of Veterans Affairs; 2012. [6] Kaplan MS, McFarland BH, Huguet N, Valenstein M. Suicide risk and precipitating circumstances among young, middle-aged, and older male veterans. Am J Public Health 2012;102(Suppl 1):S131–7. [7] Kline A, Ciccone DS, Falca-Dodson M, Black CM, Losonczy M. Suicidal ideation among National Guard troops deployed to Iraq: the association with postdeployment readjustment problems. J Nerv Ment Dis 2011;199:914–20. [8] Pietrzak RH, Russo AR, Ling Q, Southwick SM. Suicidal ideation in treatmentseeking Veterans of Operations Enduring Freedom and Iraqi Freedom: the role of coping strategies, resilience, and social support. J Psychiatr Res 2011;45:720–6.

[9] Department of the Army. Army health promotion risk reduction suicide prevention report 2010. Available at http://usarmy.vo.llnwd.net/e1/HPRRSP/ HP-RR-SPReport2010_v00.pdf. [10] Department of Veterans Affairs Office of Inspector General. Healthcare Inspection: Implementing VHA’s Mental Health Strategic Plan Initiatives for Suicide Prevention. Report # 06–03706–126. Available at: http://www.va.gov/oig/54/ reports/VAOIG-06-03706-126.pdf. Accessed November 13, 2012. Washington, DC: U.S. Department of Veterans Affairs, 2007. [11] Dobscha SK, Corson K, Helmer DA, Bair MJ, Denneson LM, Brandt C, et al. Brief assessment for suicidal ideation in OEF/OIF veterans with positive depression screens. Gen Hosp Psychiatry 2013;35:272–8. [12] Luoma JB, Martin CE, Pearson JL. Contact with mental health and primary care providers before suicide: a review of the evidence. Am J Psychiatry 2002;159: 909–16. [13] Basham C, Denneson LM, Millet L, Shen X, Duckart J, Dobscha SK. Characteristics and VA health care utilization of U.S. Veterans who completed suicide in Oregon between 2000 and 2005. Suicide Life Threat Behav 2011;41:287–96. [14] Smith EG, Craig TJ, Ganoczy D, Walters HM, Valenstein M. Treatment of Veterans with depression who died by suicide: timing and quality of care at last Veterans Health Administration visit. J Clin Psychiatry 2011;72:622–9. [15] Vannoy SD, Tai-Seale M, Duberstein P, Eaton LJ, Cook MA. Now what should I do? Primary care physicians' responses to older adults expressing thoughts of suicide. J Gen Intern Med 2011;26:1005–11. [16] Vannoy SD, Fancher T, Meltvedt C, Unutzer J, Duberstein P, Kravitz RL. Suicide inquiry in primary care: creating context, inquiring, and following up. Ann Fam Med 2010;8:33–9. [17] Kaplan MS, Huguet N, McFarland BH, Newsom JT. Suicide among male veterans: a prospective population-based study. J Epidemiol Community Health 2007;61:619–24. [18] Kessler RC, Borges G, Walters EE. Prevalence of and risk factors for lifetime suicide attempts in the National Comorbidity Survey. Arch Gen Psychiatry 1999;56: 617–26. [19] Denneson LM, Basham C, Dickinson KC, Crutchfield MC, Millet L, Shen X, et al. Suicide risk assessment and content of VA health care contacts before suicide completion by veterans in Oregon. Psychiatr Serv 2010;61:1192–7. [20] Kaplan MS, McFarland BH, Huguet N, Conner K, Caetano R, Giesbrecht N, et al. Acute alcohol intoxication and suicide: a gender-stratified analysis of the National Violent Death Reporting System. Inj Prev 2013;19:38–43. [21] Cherpitel CJ, Borges GL, Wilcox HC. Acute alcohol use and suicidal behavior: a review of the literature. Alcohol Clin Exp Res 2004;28:18S–28S. [22] Corson K, Denneson LM, Bair M, Helmer DA, Goulet JL, Dobscha SK. Prevalence and correlates of suicidal ideation among OEF/OIF Veterans. J Affect Disord 2013;149: 291–8. [23] Kroenke K, Spitzer RL, Williams JB. The Patient Health Questionnaire-2: validity of a two-item depression screener. Med Care 2003;41:1284–92. [24] Department of Veterans Affairs. Suicide risk assessment guide. www.mentalhealth. va.gov/docs/suicide-risk-assessment-guide.pdf; 2008. Accessed 2/27/12. [25] Department of Veterans Affairs. VHA Decision Support System (DSS). Available at: http://www.virec.research.va.gov/DataSourcesName/DSS/DSSintro.htm. Accessed April 11, 2012. Washington, DC: U.S. Department of Veterans Affairs, 2007. [26] Kroenke K, Unutzer J, Callahan C, Perkins A, Lowe B. Monitoring depression with a brief self-report scale (PHQ-9). J Gen Intern Med 2004;19(Supplement 1):181. [27] Crosby AE, Ortega L, Melanson C. Self-directed violence surveillance: uniform definitions and recommended data elements, version 1.0. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control; 2011. [28] Centers for Disease Control and Prevention (CDC). Self-directed violence surveillance: uniform definitions and recommended data elements, version 1.0. http://www.cdc.gov/violenceprevention/pdf/Self-Directed-Violence-a.pdf. Atlanta, GA: Centers for Disease Control and Prevention, National Center for Injury Prevention and Control, 2011. [29] Dobscha SK, Gerrity MS, Corson K, Bahr A, Cuilwik NM. Measuring adherence to depression treatment guidelines in a VA primary care clinic. Gen Hosp Psychiatry 2003;25:230–7. [30] Corson K, Doak MN, Denneson L, Crutchfield M, Soleck G, Dickinson KC, et al. Primary care clinician adherence to guidelines for the management of chronic musculoskeletal pain: results from the study of the effectiveness of a collaborative approach to pain. Pain Med 2011;12:1490–501. [31] Elbogen EB, Johnson SC, Wagner HR, Newton VM, Timko C, Vasterling JJ, et al. Protective factors and risk modification of violence in Iraq and Afghanistan War veterans. J Clin Psychiatry 2012;73:e767–73. [32] Sayers SL, Farrow VA, Ross J, Oslin DW. Family problems among recently returned military veterans referred for a mental health evaluation. J Clin Psychiatry 2009; 70:163–70. [33] Kemp JE, McKeon R. VA Suicide Prevention Hotline and Strategies. SAMHSA's Suicide Prevention Initiatives and Collaborations; 2009 [Presentation]. [34] Buchsbaum DG, Buchanan RG, Poses RM, Schnoll SH, Lawton MJ. Physician detection of drinking problems in patients attending a general medicine practice. J Gen Intern Med 1992;7:517–21. [35] Denny CH, Serdula MK, Holtzman D, Nelson DE. Physician advice about smoking and drinking: are U.S. adults being informed? Am J Prev Med 2003; 24:71–4. [36] Dobscha SK, Dickinson KC, Lasarev MR, Lee ES. Associations between race and ethnicity and receipt of advice about alcohol use in the Department of Veterans Affairs. Psychiatr Serv 2009;60:663–70. [37] Kaplan MS, Adamek ME, Rhoades JA. Prevention of elderly suicide. Physicians' assessment of firearm availability. Am J Prev Med 1998;15:60–4.

S.K. Dobscha et al. / General Hospital Psychiatry 36 (2014) 310–317 [38] Walters H, Kulkarni M, Forman J, Roeder K, Travis J, Valenstein M. Feasibility and acceptability of interventions to delay gun access in VA mental health settings. Gen Hosp Psychiatry 2012;34:692–8. [39] Mann JJ, Apter A, Bertolote J, Beautrais A, Currier D, Haas A, et al. Suicide prevention strategies: a systematic review. JAMA 2005;294:2064–74. [40] Smith TW, Marsden P, Hout M, Kim J. General Social Surveys, 1972-2012[MachineReadable Data File]. Chicago: National Opinion Research Center [producer]; Storrs, CT: The Roper Center for Public Opinion Research, University of Connecticut [distributor], 2013. [41] Ilgen MA, Conner KR, Roeder KM, Blow FC, Austin K, Valenstein M. Patterns of treatment utilization before suicide among male veterans with substance use disorders. Am J Public Health 2012;102(Suppl 1):S88–92. [42] Kaplan MS, McFarland BH, Huguet N. Firearm suicide among veterans in the general population: findings from the National Violent Death Reporting System. J Trauama 2009;67:503–7.

317

[43] Kaplan MS, McFarland BH, Huguet N. Characteristics of adult male and female firearm suicide decedents: findings from the National Violent Death Reporting System. Inj Prev 2009;15:322–7. [44] Betz ME, Sullivan AF, Manton AP, Espinola JA, Miller I, Camargo Jr CA, et al. Knowledge, attitudes, and practices of emergency department providers in the care of suicidal patients. Depress Anxiety 2013;30:1005–12. [45] Gaynes BN, West SL, Ford CA, Frame P, Klein J, Lohr KN. Screening for suicide risk in adults: a summary of the evidence for the U.S. Preventive Services Task Force. Ann Intern Med 2004;140:822–35. [46] O'Neil ME, Peterson K, Low A, Carson S, Denneson LM, Haney E, et al. Suicide prevention interventions and referral/follow-up services. Washington, DC: U.S. Department of Veterans Affairs, Evidence-based Synthesis Program; 2012. [47] Frank E, Ciocca M. Counseling on access to lethal means. Available at: http://www. sprc.org/bpr/section-III/calm-counseling-access-lethal-means. Accessed July 30, 2013.