Evaluation of outcomes for military mental health partial hospitalization program

ABSTRACT

In the military health system, there has been a growing demand for mental health services over the last two decades. Partial hospitalization programs fill a critical niche between outpatient and inpatient services. The present study evaluated immediate and long-term outcomes of a military mental health partial hospitalization program at a large military treatment facility. This study collected retrospective data of active duty patients who completed a 6-day partial hospitalization program within a 2-year period. Results showed that the majority of participants were young, male, and junior enlisted service members endorsing suicidal ideation as well as adjustment/stressor- and depressive-related psychiatric symptoms. Immediately after treatment, participants showed a significant reduction in psychiatric symptoms and dysfunction after treatment. In the long-term, most participants engaged in mental health services post-discharge, though engagement with services decreased over time. In addition, career-impacting medical recommendations were made for over half of participants with almost three-quarters of these recommendations made before or during enrollment in the program. This study expanded the limited evidence base for military mental health partial hospitalization programs. In addition, this study offered data on the frequency of career-impacting medical recommendations made for patients engaged in care at this level of acuity.

What is the public significance of this article?—This study suggests that mental health partial hospitalization programs may be helpful in reducing psychiatric symptoms and improving functioning for active duty military service members. In addition, it shows that most service members continue to engage in mental health services afterwards and that career-impacting recommendations may be common in this clinical setting.

Introduction

Over the last 50 years, there has been a steady rise in the availability of partial hospitalization programs (PHPs) for mental health services in civilian and military medical systems (Lande & Pourzand, 2016; Luber, 2013). PHPs offer hospital systems a viable option for delivering psychiatric services for patients who require acute stabilization but do not meet the threshold for psychiatric admission (Houvenagle, 2015). Thus, PHPs fit a unique niche between the scope of inpatient and outpatient services. Though research has been limited, mounting evidence demonstrates that PHPs can reduce psychiatric symptoms for patients with a variety of high severity mental health concerns, and they are at a lower cost than inpatient hospitalization services (Horvitz-Lennon, Normand, Gaccione, & Frank, 2001; Marshall, Crowther, Sledge, Rathbone, & Soares‐Weiser, 2011; Sledge et al., 1996; Tacchi, Joseph, & Scott, 2004).

In the military health system (MHS), there has been a marked increase in demand for mental health services over the last two decades. From 2005 to 2016, direct outpatient and inpatient services provided in the MHS increased by 151% and 132%, respectively (Psychological Health Center of Excellence, 2021a). PHPs have emerged across the MHS as specific demand for higher acuity services grows and budgetary constraints weigh on the health care system. Though the literature is sparse, researchers at Naval Medical Center Portsmouth (NMCP) investigated the cost savings and treatment response of a PHP for acute mental health services as compared to inpatient hospitalization (Manos et al., 2002). Results of this study showed that the PHP was both cost effective and had superior clinical results for patients in need of crisis intervention when compared to inpatient services. In another study, researchers at Walter Reed National Military Medical Center described characteristics of service members who completed a military-only PHP over a 5-year period (Lande & Pourzand, 2016). Findings revealed that a diverse range of patients were referred for treatment, cutting across age groups, ranks, diagnoses, time in service and combat exposure.

In the military, mental health providers routinely assess the impact of psychiatric conditions on occupational functioning, often known as fitness for duty evaluations. These evaluations determine whether a service member is unfit or unsuitable to perform the duties of his or her grade, rank, or assigned position (Monahan & Keener, 2012, p. 25). Thus, a mental health diagnosis may be accompanied with career-impacting recommendations such as being temporarily relieved of some or all of one’s assigned duties, medical retirement via referral to the physical evaluation board, or administrative separation (Monahan & Keener, 2012, p. 29). In outpatient settings, career-impacting recommendations are common, though it appears to vary across the branches, referral sources (e.g., command directed versus self-referred), and demographic and clinical factors (Ghahramanlou-Holloway et al., 2018, 2019; Rowan, Varga, Clayton, & Martin Zona, 2014). Given its position between in- and outpatient clinical settings, PHPs are likely to enroll patients with diagnoses that commonly lead to career-impacting recommendations. However, previous research on military PHPs has not reported the frequency of these recommendations.

The present study aims to expand the literature on military PHPs. First, this study will report demographic and clinical characteristics for those who enrolled in the program. Second, this study will test the hypothesis that completion of the PHP is associated with a significant reduction in clinical symptoms and dysfunction after completing treatment. To test this hypothesis, comparisons of pre- and post-treatment measures of depression, anxiety, and functional impairment will be conducted. Third, this study will provide descriptive data for patient outcomes 90 and 180 days after completing the program. At these intervals, we evaluated treatment utilization and career-impacting recommendations, which is a key indicator of occupational functioning in military settings. We hope these findings will bolster the empirical evidence base for military PHPs.

Method

Participants

This study was conducted as a retrospective evaluation of a PHP for active-duty service members treated at a large Navy military treatment facility. This PHP was a 6-day program with daily rolling admissions open to active duty service members from all service components. Data were collected from participants who enrolled in the program between January 1, 2018 and December 31, 2019. Out of a total 794 enrollees during this time period, 150 participants were randomly selected with 144 participants included in the final sample. Records for six participants could not be found using the patient log due to incomplete patient demographic information (e.g., incorrectly spelled name or incorrect electronic medical record number).

Materials

In this study, only total scores were available in patients’ records; thus, item level analyses are not possible for the measures described below.

Patient health questionnaire – 9 (PHQ-9)

The PHQ-9 is a 9-item self-report questionnaire used to measure symptoms of depression and symptom severity within the previous 2 weeks (Kroenke, Spitzer, & Williams, 2001). Each item is rated using a 4-point Likert scale, assessing the frequency at which individuals have been bothered by each symptom. Total scores range from 0 to 27, with higher scores indicating greater symptom severity. The measure is routinely used in military treatment settings, and is recommended by Departments of Veterans Affairs and Defense as a monitoring tool for adherence and response to treatment, suicidality, and psychosocial stress levels (Department of Veterans Affairs and Department of Defense, 2016).

Generalized anxiety disorder – 7 (GAD-7)

The GAD-7 is a 7-item self-report questionnaire used to assess symptoms of generalized anxiety disorder and symptom severity within the last 2 weeks (Spitzer, Kroenke, Williams, & Löwe, 2006). Like the PHQ-9, individuals rate the frequency of being bothered by anxiety symptoms on a 4-point Likert scale. Total scores range from 0 to 21, with higher scores indicating greater symptom severity.

Sheehan disability scale (SDS)

The SDS is a 5-item self-report rating scale that measures functional impairment across three domains: work/school, social life, and family life (Leon, Olfson, Portera, Farber, & Sheehan, 1997). Each of the three domains is rated on a 10-point Likert scale, rating the degree to which the individual’s symptoms have disrupted perceived functioning in that domain. Total scores range from 0 to 30 with higher scores indicating higher functional impairment.

Procedures

Data collection

Participants completed self-report measures immediately preceding both the intake and discharge interviews. These questionnaires were completed on paper and then recorded in the provider’s clinical note. These measures were not collected 90 or 180 days post-treatment.

Treatment program and curriculum

The treatment program was a 6-day PHP, with open enrollment capped at 16 patients per day. The program offered individual and group therapy, though the primary modality of treatment was group intervention. At the start of 2018, the curriculum consisted of treatment groups as well as complementary rehabilitative activities, including art therapy, recreation therapy, and self-guided exercise time. Regarding these complementary interventions, patients completed art therapy once per week, recreational therapy twice per week, and self-guided exercise twice per week. The treatment groups included psychoeducation on thoughts and emotions extracted from Dialectical Behavioral Therapy (DBT; Linehan, 2014; emotion regulation skills module), relationship-oriented groups focused on conflict management and boundary setting, and cohesion-oriented groups facilitating relationship formation within the cohort.

In 2019, authors JM and MS revised the curriculum so that it was more directly connected to empirically-supported cognitive-behavioral interventions. The new curriculum included daily mindfulness practice and two treatment groups covering one new topic each day. The topics were drawn from cognitive-behavioral therapy manuals, and broadly included psychoeducation and experiential sessions covering thoughts, emotions, emotion regulation, interpersonal skills, values, and sleep hygiene. Groups addressing the relationship between thoughts and emotions drew their content from manuals available through the Mental Illness Research, Education, and Clinical Center, specifically manuals for depressed mood (Wenzel, Brown, & Karlin, 2011) and anxiety (Shrestha & Stanley, 2011). Groups that focused on emotion, emotion regulation, and interpersonal skills were drawn from the second edition of the DBT manual (Linehan, 2014). To facilitate behavioral activation, values clarification and goal setting activities were drawn from a group therapy focused Acceptance and Commitment Therapy (ACT) manual (Westrup & Wright, 2017). Psychoeducational materials for sleep hygiene were drawn from a Cognitive Behavioral Therapy for Insomnia manual (Edinger & Carney, 2014). The mindfulness activities were drawn from the DBT and ACT manuals (Linehan, 2014; Westrup & Wright, 2017). The complementary rehabilitative activities were retained after the curriculum revision.

Given the higher acuity level of enrolled patients, patients completed suicide risk screenings with staff at check in and check out every day. Patients who endorsed suicidal ideation were required to complete an evaluation with a privileged provider before participating program activities or departing for the day. If a patient was deemed unsafe, he or she would be referred to the emergency department in the hospital for psychiatric admission consideration.

Data analytic plan

Participant characteristics

Descriptive statistics were used to describe participant characteristics and treatment histories prior to treatment.

Immediate treatment effects

To evaluate treatment response, measures of depression, anxiety, and functional impairment were used to determine if enrollment in the program led to a significant reduction in symptoms. All variables met normality criteria for paired samples t-test, and Bonferroni corrections were made to adjust for multiple comparisons (adjusted α = .02). In addition, descriptive data were provided on utilization of mental health service post-discharge, along with fitness and suitability decisions at discharge.

Long-term outcomes

Descriptive analyses were used to analyze patient outcomes 90 and 180 days after completing the program.

Results

Participant characteristics

The mean age of participants was 24.37 years (SD = 5.86). The majority of participants were male (64.6%, n = 93), and serving in the US Navy (95.1%, n = 137). Pay grade ranged from E-1 to O-3 with the modal ranks being E-3 and E-4 (25%, n = 36, and 30%, n = 43, respectively). Participants were nearly all enlisted (98.6%, n = 142). Most participants described their ethno-racial background as White (50%, n = 75) with Black (30%, n = 45), Hispanic (7.3%, n = 11), Other (6%, n = 9), and Asian or Pacific Islander (5.3%, n = 8) backgrounds also represented. Two participants did not identify with a specific ethno-racial background (1.3%). The majority of participants were serving on operational platforms (e.g., ships, submarines, or active flying wings; 66.2%, n = 95) with fewer serving on shore duty assignments (63.8%, n = 49).

Prior to enrollment in the program, 58.3% of participants had engaged in at least one session of psychotherapy (Md = 5). Fewer participants had engaged in at least one medication consultation appointment, 23.6% (Md = 4). Nearly all participants had at least one visit to the emergency department for mental health concerns (88.2%; n = 127) with only 12.5% having had two or more visits to the emergency department (n = 18). A third of participants had at least one hospitalization prior to enrollment (33.3%; n = 48).

Immediately prior to enrollment, the majority of participants were referred directly from the emergency department (51.5%, n = 74), with inpatient and outpatient services also referring participants (31.2% [n = 45] and 7.3% [n = 25], respectively). The most common acute concern precipitating a participant’s referral was recent suicidal ideation (46.5%, n = 66).

Given these referral pathways, there was limited diagnostic clarity at intake for many patients. Amidst the wide variety of diagnoses at discharge from the program, Adjustment Disorders (50%, n = 72) and Depressive Disorders (19.4%, n = 28) were most common. In addition, mental health providers had made career-impacting recommendations for 20.1% (n = 29) of participants prior to enrollment (e.g., unfit for duty or unsuitable for military service; n = 16 & n = 13, respectively).

Immediate treatment effects

Symptom questionnaires were not recorded for any participants before March 2018, resulting in 15 cases without documented PHQ-9 or GAD-7 questionnaire data. In addition, seven participants did not have recorded discharge data and one participant did not complete any intake or discharge questionnaires.

Paired samples t-tests were conducted to evaluate the mean differences between pre- and post-treatment questionnaire data. Results showed that self-reported depressive symptomatology on the PHQ-9 reduced from pre- (M = 16.36, SD = 6.04) to post-treatment (M = 10.69, SD = 6.57) assessment (t = 9.57 [df = 120], p < .001, 95% CI [4.31, 6.55]). Cohen’s d was calculated, yielding an effect size of .89 (large, per Cohen, 1988). Results similarly showed that self-reported anxiety-related symptomatology on the GAD-7 reduced from pre- (M = 13.45, SD = 5.60) to post-treatment (M = 9.28, SD = 6.13) assessment (t = 7.43, [df = 120], p < .001, 95% CI [2.84, 4.90]). Cohen’s d was .71 (medium-to-large effect size per Cohen, 1988). Analysis of self-reported dysfunction on the SDS revealed a significant reduction from pre- (M = 18.13, SD = 7.64) to post-treatment (M = 11.54, SD = 7.34) assessment (t = 9.59, [df = 135], p < .001, 95% CI [5.03, 7.65]), Cohen’s d = .87 (large effect per Cohen, 1988).

Given that three separate tests were being conducted, Bonferoni corrections were made for all t-test results. With these corrections, all three tests showed a significant reduction in symptomatology or dysfunction from pre- and post-treatment (i.e., p < .001 for all three t values).

After completing the program, 84.7% of patients met at least once with a mental health provider, with next available follow up at approximately one week for most participants (Md = 6 days). At discharge, 28.7% of participants who were considered fit and suitable at intake (n = 115) were deemed unfit (n = 16) or unsuitable (n = 17; 13.9% and 14.8%, respectively) while enrolled in the program.

Long-term outcomes

From discharge to 90 days, 81.9% of participants who completed the PHP engaged in some form of mental health service (n = 118). Of this group of post-PHP treatment-engagers, 19.5% of participants returned to the emergency department for mental health concerns at least once (n = 23), and 10.2% of participants experienced at least one inpatient psychiatric hospitalization (n = 12). Regarding outpatient services, 88.1% of participants kept at least one psychotherapy session (n = 104; Md = 4) and 44.1% of participants kept at least one medication consultation session (n = 52, Md = 2). Regarding career-impacting recommendations, 69% of treatment engagers at 90 days were considered fit and suitable at discharge (n = 81). Of these participants, 25.9% had been deemed unfit (n = 14) or unsuitable (n = 7) 90 days post-discharge.

Between 90 and 180 days, 51.4% of participants were engaged in some form of mental health services (n = 74). Of participants still engaged in services, 13.5% of participants returned to the emergency department at least once (n = 10), 6.8% of participants experienced at least one inpatient psychiatric hospitalization (n = 5), 90.5% of participants kept at least one psychotherapy session (Md = 4), and 56.8% of patients kept at least one medication consultation session (Md = 2). Regarding career-impacting recommendations, 59% of treatment engagers at 180 days were considered fit and suitable at discharge and at 90 days (n = 44). Of these participants, 6.8% had been deemed unfit (n = 2) or unsuitable (n = 1) 180 days post-discharge.

Overall, 59.7% of participants (n = 86) were found unfit (n = 48) or unsuitable (n = 38) for military service before, during, or after the program. Of those found unfit, 60.1% of participants were considered to have a disabling mental health condition (n = 23) or other medical condition (n = 6), and were referred to the disability evaluation system for possible medical separation or retirement. 84.2% of participants who were deemed unsuitable were recommended for administrative separation due to difficulties adjusting to the military (n = 27) or personality disorder (n = 5).

Discussion

With demand for mental health services growing in the military, partial hospitalization programs fill a critical niche between traditional inpatient and outpatient services. This niche offers medical systems a viable cost-saving and effective treatment option in civilian and military settings (Horvitz-Lennon et al., 2001; Manos et al., 2002; Tacchi et al., 2004). Despite their demonstrated value, there is little empirical research on the effectiveness and outcomes of these programs in military medical settings. This study bolsters the limited research on military PHPs.

The first aim of this study was to report demographic and clinical characteristics of participants who enrolled in the PHP. Similar to previous research on military PHPs, the majority of participants were young, male, junior enlisted service members (Lande & Pourzand, 2016; Manos et al., 2002). These sample characteristics are consistent with the distribution of age, gender, and rank in the DoD overall. Although the sample was ethnically and racially diverse, there were some notable demographic differences between this sample and the DoD overall. Black participants were overrepresented (this study = 30%; DoD, 18% for enlisted and 8% for officer ranks) while Hispanic participants were underrepresented (this study = 7.3%; DoD, 19% for enlisted and 8% for officer ranks; Department of Defense, 2020). In outpatient mental health settings, nonwhite service members have been shown to be less likely to utilize mental health services (McKibben et al., 2013). Thus, PHPs may fill a critical niche that may be in higher demand for black military service members.

Clinically, the most common acute presenting concern was suicidal ideation with the majority of participants being referred from the emergency department or the inpatient ward. The most common diagnoses made were Adjustment and Depressive Disorders, which have high prevalence rates in the DoD overall (Psychological Health Center of Excellence, 2021b). Military hospitals, particularly those with emergency departments, may benefit by adding PHPs to their menu of clinical services given that suicidality can be a feature of both Adjustment and Depressive Disorders (Casey, Jabbar, O’Leary, & Doherty, 2015).

The second aim of this study was to test the hypothesis that completion of the PHP would be associated with a significant reduction in clinical symptomatology and dysfunction after completing treatment. Results showed significant reductions in depressive and anxiety symptoms, with large and medium-to-large effects, respectively. In addition, results showed a significant, large reduction in functional impairment. These results demonstrate clinical efficacy, notably for symptoms that are common features of Adjustment and Depressive Disorders, which were the most commonly diagnosed conditions. Though research on PHPs is limited, these results are consistent with findings in military intensive outpatient programs (Hoyt et al., 2018).

The study’s last aim was to describe outcomes after completing the program, as measured by treatment utilization and career-impacting decisions. By 90 days, most participants had engaged in some form of mental health services post-discharge. By 180 days, about half of the participants were still utilizing mental health services. Utilization of psychotherapy was more common than psychotropic medication consultation at 90 and 180 days. Some participants also utilized high acuity services after completing the PHP, including returning to the emergency department or being admitted for inpatient services. Though research on hospitalization rates after PHP is scant, one study found shared and unique risk factors based on an individual’s referral source (Beard et al., 2016). Shared risk factors included histories of inpatient admission, suicidal ideation, and substance abuse for patients referred from inpatient and outpatient referral settings (Beard et al., 2016). For individuals referred directly from an inpatient setting, suicidal ideation and psychotic symptoms were associated with increased likelihood of re-hospitalization. For individuals referred from outpatient settings, poor relational functioning was associated with increased risk of first-time hospitalization. In the present study, the number of hospitalizations after 90 or 180 days was too small for meaningful analysis (n = 17). Future research on military PHPs should explore the risk factors for hospitalization or re-hospitalization after completion of a PHP.

For some participants, continued utilization of mental health services was accompanied by career-impacting recommendations after 90 and 180 days (27%). However, the majority of career-impacting recommendations were made before and during enrollment in the PHP (73%). Notably, over half of the participants in this study sustained a career-impacting recommendation (e.g., found unfit for duty or unsuitable for military service). Though research on the frequencies of career-impacting recommendations is limited, these recommendations have been shown to vary across service branches, referral sources, and demographic and clinical factors (Ghahramanlou-Holloway et al., 2018, 2019; Rowan et al., 2014). The frequency of career-impacting recommendations in this study was higher than those reported in outpatient samples (Ghahramanlou-Holloway et al., 2018, 2019; Rowan et al., 2014), likely due to the higher acuity of the treatment setting.

These results need to be considered in light of a few limitations. First, this study was conducted as a retrospective study. Future studies could aim to collect data directly from interviews or web-based platforms rather than collecting only retrospective data from an electronic medical record. Second, the data collected in this study was extracted from one particular PHP that served mostly US Navy personnel. Thus, these results may not generalize to PHPs in different regions or PHPs that serve other component services. Third, this study does not have a control group. Therefore, it is possible that the treatment effect could be accounted for by the passage of time outside of one’s work environment. Given the clinical severity of patients referred to PHPs, having a waitlist control group may not viable. However, future researchers could explore strategies for parsing out these effects. Similarly, there is no direct comparison group against which to compare the long-term outcomes.

Despite these limitations, this study makes a meaningful contribution to the clinical literature on military PHPs. These findings provide support for the efficacy of PHPs to meet the ever-growing demand for mental health services.

Disclosure statement

No potential conflict of interest was reported by the author(s).