Enhancing the Reach of Cognitive-Behavioral Therapy Targeting Posttraumatic Stress in Acute Care Medical Settings

The integration of mental health and substance use screening and intervention within general medical settings has the potential to overcome barriers to accessing these services and constitutes a long-term public health objective (1,2). Collaborative care intervention models may be optimal approaches for achieving this objective in primary and acute care medical settings (3–7). Nationwide, over 30 million patients present to trauma care systems each year after sustaining a traumatic injury, and 1.5 to 2.5 million individuals are so severely injured that they require inpatient hospitalization (2,8). Psychiatric and substance-related comorbidities, including alcohol use disorders and posttraumatic stress disorder (PTSD), are common among injured patients in acute care medical settings (7,9).

A body of literature now supports the integrated delivery of depression and anxiety treatments, including cognitive-behavioral therapy (CBT), in primary care settings (10,11). However, less is known about how to effectively integrate CBT approaches into acute care medical settings. Session-based CBT is among the evidence-based treatment modalities for PTSD most consistently recommended by best-practice guidelines (12) and may be an effective early preventive intervention for PTSD for the subpopulation of trauma-exposed patients recruited into efficacy trials.

Emerging literature on early collaborative care and other interventions now informs preliminary American College of Surgeons’ guidelines supporting PTSD screening and intervention at U.S. trauma centers (6,13,14). However, ongoing dialogue in the dissemination research field suggests that multisession CBT can be challenging to implement for trauma survivors in real-world contexts, including acute care medical settings (15–17). For instance, in an acute care study of an early CBT intervention, less than 10% of trauma survivors symptomatic for PTSD received multisession CBT offered by the trial (18). Commentary from the dissemination and implementation research field has encouraged hybrid clinical trial approaches that aim to simultaneously test questions related to treatment effectiveness, feasibility, and acceptability of real-world intervention implementation (19,20).

The disassembling of traditional CBT approaches into implementable therapeutic elements (for example, psychoeducation, relaxation strategies, and activity scheduling) with accompanying homework assignments embedded in routine care management may advance the integration of PTSD treatment in acute care medical settings. Informed by the current evidence base for alcohol brief interventions in acute medical care and by public health and dissemination and implementation frameworks prioritizing population reach (21) and intervention feasibility and sustainability, our clinical research team developed CBT elements targeting PTSD and related comorbidities over a series of early collaborative care intervention trials (6,16).

A central feature of CBT is therapeutic assignment and completion of homework that provides opportunities for individuals to practice learned skills and for operant or classical conditioning learning to occur (22). Completion of homework is associated with symptom reduction. A recent meta-analysis demonstrated a small to medium association between homework completion and target symptom reduction (r=.26; 95% confidence interval=.19–.33) (23,24). Homework assignments play an equally important role in the CBT elements designed and integrated by the study team into care management; patients are taught behavioral skills and assigned activities to reduce symptoms of anxiety and depression that are common after traumatic injury. However, unlike patients in more traditional session-based approaches, patients of the care management team more frequently work with their care manager by telephone or at postinjury hospital and aftercare appointments, rather than in the therapist’s office. Whether homework is readily assigned and completed, and whether homework assignment and completion are associated with reduction of symptoms (that is, posttraumatic stress) within this therapeutic context, has not previously been empirically examined.

This study aimed to evaluate the feasibility and acceptability of embedding CBT elements within routine postinjury care management by using data from a clinical trial of stepped collaborative care targeting PTSD (25). The investigation also examined the association between completion of CBT elements homework and PTSD symptom reduction among patients receiving the stepped collaborative care intervention.

Methods

Data for this study were derived from a randomized controlled trial of a stepped collaborative care intervention targeting PTSD and related comorbidities among injured inpatients (Clinicaltrials.gov identifier NCT01625416) (25).

Participants and Procedures

This study included 115 of the original 121 patients enrolled in the larger trial; six patients were excluded because they did not complete any follow-up assessments (25). The University of Washington Institutional Review Board approved all study procedures prior to protocol initiation, and written informed consent was obtained from each participant. Study recruitment occurred from July 1, 2012, through February 28, 2013.

Patients randomly assigned to the intervention were identified as at risk of developing PTSD on the basis of a ten-item electronic medical record screen (26) and as symptomatic for PTSD on the basis of a score of ≥35 on the PTSD Checklist–Civilian DSM-IV Version (PCL-C) (6,27). Patients were randomly assigned to either the stepped care intervention or usual care control condition. Blinded to the study condition, research associates assessed patients at baseline and at one, three, and six months after baseline.

Intervention Condition

Patients randomly assigned to the intervention condition (N=56) received stepped measurement-based care from a trauma center–based mental health team over the course of six months postinjury (25), which included care coordination, pharmacotherapy targeting PTSD and related comorbidity, motivational interviewing (MI) targeting alcohol use and other risk behaviors, and CBT elements with associated homework targeting PTSD and comorbidity (for example, depression and suicidal ideation).

Rationale and approach to CBT element development and implementation.

As with “broad reach” MI interventions that have been successfully implemented in trauma center–based alcohol screening and brief intervention (2), the CBT elements were designed to be brief and readily deliverable within routine care management (that is, at bedside in the hospital or on the emergency department gurney, at outpatient medical follow-up appointments, or via telephone). The elements included principles and skills similar to those taught in session-based CBT targeting anxiety and depression, such as psychoeducation about symptoms and posttraumatic recovery; anxiety and stress reduction techniques, such as training in progressive muscle relaxation and breathing retraining; problem-solving; and activity scheduling consistent with a behavioral activation paradigm (28). The elements were designed to help patients overcome behavioral avoidance patterns consequent to anxious avoidance, withdrawal, or functional impairments related to the injury. Initial investigation has suggested that behavioral activation is efficacious for reducing symptoms of depression and PTSD (18,29,30). Therefore, the elements used emphasized working with patients to identify reinforcing activities that are pleasurable or meaningful or that help patients to build self-efficacy and mastery.

Common relaxation exercises targeting anxiety and hyperarousal included practicing slow breathing and progressive muscle relaxation for a few minutes each day. Examples of activity scheduling included engaging in pleasant events throughout the week, such as spending time outdoors or with friends. Patients were also assigned activities to face feared situations with the aim of increasing contact with enjoyable or meaningful life activities. The care manager documented the amount of time spent during each encounter with a patient, which included the time spent completing intervention activities for the patient (for example, care coordination). On average, the care manager spent 155.5 minutes, or 2.6±1.3 hours, per patient.

The care manager (SO) has a doctoral degree in psychology and completed graduate and fellowship training in CBT and MI. The care manager collaboratively identified appropriate CBT homework assignments with patients on the basis of the elements to target avoidance-withdrawal and anxiety-stress symptoms, according to the patient’s report and standardized measures. The nature and frequency of embedded CBT elements and homework assigned varied by patient depending on the patient’s level of engagement, time available, and treatment preferences and targets. As in previous trials by the study team, outpatient appointments for traditional cognitive-behavioral psychotherapy with the care manager were made available but were rarely utilized (1% of all encounters) (16,18).

Usual care control condition.

Patients in the control group (N=59) received postinjury care as usual, which prior investigation suggests includes routine outpatient surgical, primary care, and emergency department visits, as well as the occasional use of specialty mental health services (6,13,25).

Measures and Variables

Demographic and injury variables.

Demographic variables, including gender, race, ethnicity, and age, were assessed through patient self-report. The investigation determined injury severity at baseline during the index admission from the medical record ICD-9 codes by using the Abbreviated Injury Scale and Injury Severity Score (31).

PTSD symptoms.

PTSD symptoms were assessed with the PCL-C (27), which has established reliability and validity across trauma-exposed populations (32,33). All patients were asked to report their PTSD symptoms since the injury event at baseline and during the previous month at follow-up assessments. A 10-point difference on the PTSD checklist has been identified as clinically meaningful change (34).

Mental health service utilization.

Pre- and postinjury outpatient mental health service utilization among all patients was assessed by using items from the National Study on the Costs and Outcomes of Trauma (35), which were administered at baseline and at the one-, three-, and six-month follow-up interviews. At baseline, patients reported whether in their lifetime prior to their injury they had at least one outpatient visit with a provider about problems with emotions or nerves or about use of alcohol. At the one-month follow-up, patients were asked to respond to the same question in reference to the year prior to their injury, and at all follow-ups, they were asked to respond in reference to the last study interview; they also indicated the type of provider seen (psychiatrist, psychologist, or mental health counselor) since the last study interview. At the three- and six-month follow-ups, patients were asked to indicate whether they received psychotherapy or counseling during the study period targeting any of the following issues: anxiety, depression, alcohol use, drug use, traumatic brain injury, or suicidal ideation.

Patient willingness for treatment.

At baseline, all patients were asked to respond to the questions, “On a scale from 0 to 10, with 0 being not at all willing and 10 being totally willing, how willing would you be to talk to a trauma support counselor in the hospital, at follow-up medical appointments, or over the phone about your physical condition and emotional distress?” On the same 0-to-10 scale they were asked about willingness “to take medications for emotional distress.”

Homework completion.

For intervention patients only, the total number of homework assignments completed over the course of the six-month intervention was determined by reviewing care manager treatment notes, which included whether homework was assigned and whether the patient reported completing prior assignments since the last encounter.

Intervention acceptability.

At the six-month follow-up, intervention patients were asked to respond to three questions to assess intervention acceptability (on a scale from 1, not at all, to 9, very): “How helpful was the trauma support counseling to you?” “How successful do you think the trauma support counseling was at reducing your symptoms?” “How confident would you be in recommending the trauma support counseling to a friend or acquaintance that experiences similar symptoms?”

Data Analyses

We first described and compared baseline demographic, injury, and clinical variables, including PTSD symptom levels, willingness to receive treatments, and preinjury mental health service utilization across patients in the intervention and control group conditions. Next we compared PTSD symptom levels and patterns of health service utilization. Chi-square and independent-sample t test statistics were used to examine statistical significance of differences between groups.

To examine the feasibility and acceptability of CBT element delivery, we descriptively examined intervention patients’ self-report of trauma support counseling helpfulness, the success of the counseling at reducing symptoms, and confidence in recommending the counseling to others.

To examine the potential effectiveness of CBT element homework completion, we examined the association among intervention patients between the total number of homework assignments completed and symptoms of PTSD at the six-month assessment by using multiple linear regression. As in previous analyses of intervention effects, we controlled for age, gender, race-ethnicity, Injury Severity Score, total care manager time spent with the patient, and baseline PTSD symptoms (25). We used SPSS, version 19.0, for all analyses.

Results

On average, at baseline, patients in both the intervention and control groups were highly willing to talk to a trauma support counselor about their physical condition and emotional distress and somewhat willing to take medications for emotional distress (Table 1). Preinjury outpatient mental health service utilization was common, with nearly all patients having sought outpatient services for mental health purposes at some point in their lifetime. There were no significant differences in baseline PTSD symptoms between intervention and control patients.

About half of all patients reported accessing outpatient mental health services during the six-month study period. However, few patients in the control group reported receiving psychotherapy or counseling, whereas nearly all intervention patients reported receiving this specific type of service (Table 2). On average, intervention patients had high scores on the measure of intervention acceptability (Table 2). Intervention patients had significantly lower scores on the PCL-C than the control group at the six-month assessment (Table 2).

The care manager reported delivering CBT elements to 54 (96%) intervention patients and assigning homework to 52 (92%) patients, with a mean±SD number of assignments of 5.04±2.74; of the 52 patients assigned homework, 81% completed at least one assignment. At follow-up encounters with the care manager over the six-month intervention period, intervention patients reported successful completion of homework assignments on average 2.8±2.5 times.

Adjusted linear regression analysis revealed a statistically significant and clinically meaningful association between CBT element homework completion and PTSD symptom reduction for intervention patients (Table 3). The regression coefficient for homework completion in the final model indicated that for every report patients made of having completed assigned homework at an encounter with the care manager, PCL scores dropped by 2.45 points.

Discussion

This is the first study to simultaneously assess the feasibility, acceptability, and potential effectiveness of CBT disassembled into elements and integrated in a stepped collaborative care intervention for traumatically injured inpatients, a population for whom engagement in session-based CBT has been consistently challenging (16,18). The CBT elements were feasibly delivered with homework as embedded treatments within routine trauma-focused care management. The care manager was able to engage in CBT elements with nearly all intervention patients, and most patients were able to successfully complete a homework assignment. Patients receiving the intervention generally reported that encounters with the care manager were helpful, that the counseling contributed to reduced symptoms, and that they would recommend the counseling to others.

Consistent with findings from the CBT psychotherapy process literature and the meta-analysis by Mausbach and colleagues (23), completing CBT element homework assignments was associated with a modest (β=–.38) improvement in PTSD symptoms (23,24). Specifically, our results indicate that it would take two completed homework assignments to see a 5-point difference on the PCL-C and four completed homework assignments to see a more substantial 10-point PCL-C difference (34). We acknowledge, however, that additional research is required to better understand clinically meaningful change parameters for commonly used PTSD measures, such as the PCL-C. This finding suggests that embedding elements may be an effective, initial broad-reach CBT implementation approach, with great potential public health–population impact in acute care medical settings (36). However, we acknowledge that a major shortcoming of a stand-alone approach to CBT element delivery is that many patients in the trial remained symptomatic for PTSD at follow-up. A potentially effective strategy to address patients with recalcitrant PTSD symptoms that could be tested in future randomized trials is the delivery of more intensive session-based CBT through stepped-care procedures (6).

Our findings suggest that embedding CBT elements into care management improves patients’ perceptions of having accessed counseling services after traumatic injury. Patients in the control and intervention groups reported an equally high degree of willingness to talk to a trauma support counselor about emotional distress and physical conditions at baseline; however, considerably more patients in the intervention group reported receiving psychotherapy or counseling that targeted noted comorbidities during the study period (93% of intervention patients versus 10% of patients in the control group). Of note, although patients in both groups reported obtaining mental health services during the study period, it appears that only those in the intervention group perceived themselves to have received psychotherapy or counseling for one of the comorbidities targeted by the intervention. However, the disparity and the nature of services patients receive outside the care management intervention team warrants closer examination in future studies.

Prior to enrollment in the trial, patients in our sample appeared to access mental health services at a rate higher than national estimates (37), which may be a reflection of how participants were recruited into the larger trial (that is, patients with prior mental health conditions were more heavily recruited because they were at greater risk of postinjury PTSD). Such prior exposure may increase patients’ willingness to engage in the care management intervention and CBT elements. Populations of patients with less prior mental health service use may respond differently to the intervention as a whole and to CBT elements in particular.

A key limitation is the exclusive use of a Ph.D.-level care manager and the associated limits to the generalizability of the investigative findings to providers with less training, such as frontline trauma center master’s-level social workers and registered nurse practitioners. Another limitation is that our model for homework completion among intervention patients was not causal. It is possible that the association between homework completion and symptom improvement reflected the fact that patients who experienced improvement in symptoms were better able to complete homework or that a third variable, such as engagement with the care manager, accounted for the observed association. Also, the intervention arm of the trial involved multiple components, including information technology enhancements and pharmacotherapy that may also be associated with observed treatment effects. It is not uncommon for studies of CBT service delivery that examine both effectiveness and implementation to employ nonrandomized designs (19,20). Future experimental research could test whether homework completion causes symptom reduction by randomly assigning patients to homework or no homework groups. Careful examination of the nature of homework assignments and association between homework completion and symptom reductions by using DSM-5 PTSD criteria is warranted.

Conclusions

The American College of Surgeons has demonstrated willingness to integrate mental health and substance use service delivery in acute care medical settings through best-practice guidelines and policy requirements (14). The development of readily implementable, MI-informed, brief interventions has catalyzed the college to develop requirements for universal screening and intervention for patients with alcohol use disorders. Similar development of readily deliverable CBT elements targeting PTSD and comorbidity could help facilitate the movement of current college guideline recommendations for PTSD screening and intervention toward nationwide requirements.