Early response is predictive of outcome in intensive behavioral treatment for obsessive compulsive disorder

https://doi.org/10.1016/j.jocrd.2017.10.001Get rights and content

Highlights

  • Among the first studies to evaluate early improvement in ERP in an intensive residential treatment setting with severe OCD.

  • Early improvement was highly predictive of treatment response in a sample of 98 patients.

  • ROC analyses yielded clinical cutoffs to denote the point at which an early improvement could reasonably predict response.

  • Such cutoffs may be used by clinicians to aid in decision making during early stages of intensive treatment.

Abstract

Recent evidence indicates that CBT delivered in the context of intensive residential treatment (IRT) is an effective approach for severe OCD cases. However, no study has evaluated the extent to which process variables, such as early improvements, may be related to outcome. This paper describes a study that evaluates a sample of 98 patients with severe OCD who underwent IRT. Patients completed the Y-BOCS, a self-report measure, both at admission and discharge. Reliable clinical change in the first two weeks of residential treatment in relation to the improvement shown for the remainder of the patient's stay denoted an ‘early improver.’ We found 36 early improvers, and these patients experienced greater overall symptom reduction compared to patients who did not experience early improvement. With an area under the curve of .817, ROC analysis revealed that early response score was a reasonable metric in predicting significant reduction in OC symptoms during IRT. Coordinate of the curve analysis revealed that an early response cut-off point of 4.5 was sensitive and specific to treatment responders at 75% accuracy, meaning that a two-week post-admission decrease of 4.5 on the Y-BOCS was particularly predictive of response status at discharge. These findings and the future implications are discussed.

Introduction

Obsessive compulsive disorder (OCD) is a highly disabling psychiatric condition that affects between one and three percent of the population (American Psychiatric Association, 2013, Karno et al., 1988). Although specific symptom manifestations of the disorder are varied and nuanced, a common thread that underlies all presentations of OCD is the presence of distressing obsessive thoughts and experiences that are maintained via negative reinforcement through a behavior or set of behaviors intended to eliminate that distress. Such behaviors, i.e., compulsions, result in significant functional impairment and often cannot be resisted or are resisted at the cost of great emotional turmoil by the sufferer. In more severe presentations of OCD, compulsions can be quite pervasive (e.g., near-constant mental rituals), physically disabling (e.g., inability to move fluidly due to not just right obsessions) and sometimes dangerous to the patient (e.g., constant cleaning using household chemicals resulting in bodily harm).

Cognitive behavioral therapy (CBT) with exposure and response prevention (E/RP; Foa & Kozak, 1996; Abramowitz, Brigidi, & Roche, 2001; reviewed by McKay et al., 2015) is considered the “gold standard” of behavioral treatment for OCD. Recent meta-analyses indicate substantial effect sizes for E/RP in comparison to both active and wait-list control groups (Rosa-Alcazar et al., 2008; Ougrin, 2011) and that such gains are durable at six-month follow ups. However, there remains much room for improvement. A subset of patients with OCD might be considered “treatment refractory” given attenuated response to available intervention (Jenike & Rauch, 1994). Recent studies indicate that outcomes may be dependent upon symptom presentation. Both Abramowitz et al., 2003 and Rufer, Fricke, Moritz, Kloss, & Hand (2006) found such differential effects. Although improvement rates were high for some subgroups (i.e., 76% for those with symmetry obsessions and 70% for those with contamination), they were unacceptably low in others (46% for unacceptable thoughts in Abramowitz et al., 2003 and 41% for sexual/religious obsessions in Rufer et al., 2006). Further, patient adherence is a significant predictor of successful treatment (e.g., Simpson et al., 2011, Simpson et al., 2012) but patient attrition is a major issue for E/RP (as well as other treatments for OCD; Ong et al., 2016). Other factors that impact outcome include frequency of exposure sessions and therapist training (see Abramowitz et al., 2001). Intensive residential treatment (IRT) for OCD may be an appropriate setting to address these variables and improve outcomes in cases considered treatment refractory (Boschen, Drummond, & Pillay, 2008).

IRT is structured in such a way as to address elements that contribute to attenuated treatment outcomes for OCD. To address patient adherence, IRT typically involves several hours of both staff-assisted and self-directed exposure exercises each day. Although patients are often carrying out exposures in 1:1 sessions with their primary clinicians, they are also engaging in exposures with support staff in between these clinician sessions. Such staff members also work to create a “culture of ritual prevention” such that patients are monitored carefully and offered support when willingness to resist rituals is low. To address the extent to which symptom heterogeneity affects outcome, IRT typically includes “symptom-specific” psychoeducation and support groups designed to reduce stigma, increase motivation and willingness, and create solidarity around plights shared by the specific patient group. For example, the program described in the current study includes a group specifically designed for the subgroup of patients with unacceptable/intrusive thoughts, a group previously known to have attenuated treatment outcomes (Abramowitz et al., 2003; Rufer et al., 2006). IRT also addresses issues of therapist knowledge/competence by not only employing OCD specialists well-versed in ERP but by taking a “team approach” to treatment such that each patient has a multidisciplinary clinical team (1:1 therapist, family therapist, psychopharmacologist) tasked with addressing that patient's particular struggles. IRT has been shown to be an effective treatment for severe forms of OCD, resulting in clinically significant improvements in YBOCS scores, symptoms of depression, and general functioning (Stewart, Stack, Farrell, Pauls, & Jenike, 2005). IRT serves as a non-invasive alternative to neurosurgical interventions for those with severe, treatment refractory OCD, and indeed participation in an adequate trial of IRT often serves as a prerequisite for surgical candidacy (Dougherty et al., 2002).

Although IRT has its clear benefits, participation in the treatment comes at a considerable cost. As of the time of this publication, there are only three residential treatment facilities in the United States, necessitating that many of those in need of the treatment travel great distances (Osgood-Hynes, Riemann, & Bjorgvinsson, 2003). Further, the financial burden can be significant. With the onset of managed care, many insurance companies are resistant to cover payments at the residential level of care for an adequate duration, leaving patients and their families in a position to bear the costs out of pocket. Finally, IRT represents a significant investment of time and energy on the part of the patient and those supporting them. Relatively higher-functioning patients may be putting jobs and school attendance on hold in order to complete the program, and most are spending considerable time far from their social supports.

Because of this investment, it is important to understand predictors of treatment response in an IRT sample in order to determine the extent to which IRT is effective and for whom. Elucidating a profile of indicators that can be detected early in treatment would enable treatment teams to make informed clinical decisions in a timely manner. For example, it would be useful for clinicians to know the amount of time needed to adequately assess the efficacy of a given exposure plan or related intervention before making a change. Although there are relative few reliable clinical predictors, a handful have emerged in recent years (cf. Knopp, Knowles, Bee, Lovell, & Bower, 2013). Stewart, Yen, Stack, and Jenike (2006) reported that lower overall initial OCD severity, better psychosocial functioning, and female gender were all significant predictors of greater reductions in symptoms during IRT. However, this was contradicted in a more recent study by our group (Brennan et al., 2014) that demonstrated significant associations between higher baseline OCD severity, less alcohol use, and fewer hoarding symptoms; and pre-post symptom change, respectively. More broadly, a review across both adult and pediatric studies of CBT efficacy indicated that symptom severity, the presence of a comorbid personality disorder and/or significant depressive symptoms, and notable family accommodation were all relatively reliable predictors of treatment outcome (Keeley, Storch, Merlo, & Geffken, 2008). Similarly, a recent study evaluating putative predictors of intensive treatment response in pediatric OCD found that increased initial symptom severity, family accommodation, and female gender all predicted post-treatment symptom severity but no variable emerged as a reliable predictor of treatment response (Rudy, Lewin, Geffken, Murphy, & Storch, 2014). A recent review across scores of studies indicated that unmarried status and co-occurrence of hoarding symptoms also predicted attenuated treatment response, but this review largely concluded that inconsistencies in assessment and analytical methods in the literature limit capacity to inform treatment (Knopp et al., 2013). Related, the definition of “treatment response” can vary between studies, which likely contributes to the heterogeneity of results and further limits interpretability of the findings in the extant literature.

On the whole, the equivocality of extant literature on many pre-treatment, cross-sectional factors necessitates increased attention of treatment variables that may play a role in predicting outcomes. In particular, both the timing and magnitude of symptom improvements as they occur during the process of treatment are shown to be important variables when considering treatment efficacy across multiple psychiatric conditions (cf. Lewis, Simons, & Kim, 2012). Although somewhat limited compared to the depression literature (c.f, Tang & Derubeis, 1999), there is some emerging evidence that suggests the occurrence of sudden gains, i.e, steep decreases in symptoms occurring between successive sessions, has strong implications for outcome in anxiety/OCD treatment. Iderka et al. (2012) found that sudden gains predicted greater overall reductions in symptoms when treating OCD with cognitive therapy or exposure therapy, both as monotherapies or in combination with fluvoxamine. A recent attempt at replication and extension of this study also showed that sudden gains were evident in the treatment of OCD, but that they were not necessarily predictive of responder status when “response” was defined as at least a 35% symptom reduction on the Y-BOCS and at least a two point improvement on the Clinical Global Impression – Improvement scale (Collins & Coles, 2017).

In addition to the occurrence of sudden gains, it is likely that the extent to which they occur early in treatment plays a significant role in dictating outcome. Early improvement is shown to be an important predictor of outcome in numerous studies over the past two decades. Stiles et al. (2003) showed that individuals in routine community outpatient treatment for depression experienced better outcomes when they had a sudden gain early in treatment; an effect also observed in a similar sample undergoing CBT by Kelly, Roberts, and Ciesla (2005). Another study of depression evaluating response to cognitive therapy showed that only sudden gains experienced during the first half of treatment predicted response. Grilo, Masheb, and Wilson (2006) showed rapid responses (i.e., significant symptom decrease within the first four weeks of treatment) in a sample of patients undergoing treatment for binge-eating, with steeper decreases for those receiving CBT vs. medication alone. In a sample of patients with generalized anxiety disorders, Bradford et al. (2011) found that extent of symptoms improvement in the first month of CBT treatment for geriatric patients was highly predictive of outcome.

More evidence that early improvement predicts treatment outcome in OCD/anxiety disorders comes from the pharmacological literature. Three recent studies evaluated the predictive utility of early improvement to pharmacological intervention in OCD, with all three finding some evidence that early response/improvement to selective serotonin-reuptake inhibitors (SSRIs) was predictive of eventual response (Bloch et al., 2013, Da Conceição Costa et al., 2013, Jakubovski et al., 2013). Similarly, two known studies of treatment course in generalized anxiety disorder (GAD) demonstrated the importance of early improvement. The first, by Rynn, Khalid‐Khan, Garcia‐Espana, Etemad, and Rickels (2006), found that symptom improvement by weeks 1 and 2 significantly predicted responder status at week 8 when GAD was treated with benzodiazepine, serotonin receptor agonists, or placebo. A study looking at pharmacological treatment of panic disorder found that changes in total scores on the clinician-rated Hamilton Anxiety Rating Scale (HAM-A) and Clinical Global Impression-Improvement Scale within the first three weeks of treatment were predictive of outcome over and above changes in panic symptoms themselves (Pollack et al., 2002). The extant anxiety/OCD literature, then, points to the importance of early improvement in pharmacological intervention, though more research is necessary to build off of groundwork laid by Aderka et al. (2012) and Collins and Coles (2017) to evaluate whether this treatment variable carries similar utility in CBT-based intervention.

In the current study, we sought to evaluate a) the occurrence of “early improvements” in a sample of individuals with OCD undergoing IRT, b) the extent that such improvements were predictive of post-treatment symptom severity, and c) the extent that early improvement might predict treatment response. We sought to operationally define an “early improvement” using criteria gleaned from the sudden gains literature (see below) and compare the symptom trajectories of patients with and without an early improvement across time spent in IRT. We sought to use a signal detection analytical approach in order to understand the extent that early symptom changes could reliably predict treatment responder status. We hypothesized that a subset of our patients would indeed evince an early improvement and that these patients would be less symptomatic post-treatment and would be more likely classified as having had a better treatment response.

Section snippets

Participants and setting

Participants in this study were 103 individuals with OCD admitted to an intensive residential program for OCD due to severe symptoms that resulted in a profound loss of function. This treatment facility provides comprehensive CBT-based OCD treatment to individuals aged 16 and older with an average length of stay of approximately 60 days. Each patient is assigned a three person treatment team: a case manager/family therapist, individual behavioral therapist, and psychiatrist. In addition to

Evidence for and characteristics of “early improver” participants

Across our entire sample, we found that participants experienced a statistically significant decrease in OCD symptoms as measured by the Y-BOCS: mean (SD) Y-BOCS at admission was 26.9 (5.6) and at discharge was 17.5 (6.7), p <.001. Using the RCI criteria established by Jacobson and Truax (1991), we utilized pre-treatment standard deviation (5.6) and test-retest reliability of the self-report Y-BOCS (.88, as indicated in Steketee et al., 1996) to determine that a drop of 5.3 points on the Y-BOCS

Discussion

Although IRT for OCD is shown to be an effective intervention, the topography of treatment trajectory, as well as its implication for eventual treatment response, is not yet well understood. The results of the current analyses reveal that many patients undergoing IRT experience an initial drop in symptoms within the first two weeks of treatment, which accounts for the majority of their overall improvement. Further, this early improvement appears a useful metric in predicting eventual responder

References (55)

  • A.I. Rosa-Alcazar et al.

    Psychological treatment of obsessive-compulsive disorder: a meta-analysis

    Clinical Psychology Review

    (2008)
  • B.M. Rudy et al.

    Predictors of treatment response to intensive cognitive-behavioral therapy for pediatric obsessive-compulsive disorder

    Psychiatry Research

    (2014)
  • G. Steketee et al.

    The Yale-Brown Obsessive Compulsive Scale: Interview versus self-report

    Behaviour Research and Therapy

    (1996)
  • S.E. Stewart et al.

    Effectiveness of intensive residential treatment (IRT) for severe, refractory obsessive-compulsive disorder

    Journal of Psychiatric Research

    (2005)
  • S.E. Stewart et al.

    Outcome predictors for severe obsessive–compulsive patients in intensive residential treatment

    Journal of Psychiatric Research

    (2006)
  • M.P. Twohig et al.

    Increasing willingness to experience obsessions: Acceptance and commitment therapy as a treatment for obsessive-compulsive disorder

    Behavior Therapy

    (2006)
  • J.S. Abramowitz et al.

    Cognitive-behavioral therapy for obsessive-compulsive disorder: a review of the treatment literature

    Research on Social Work Practice

    (2001)
  • J.S. Abramowitz et al.

    Treatment compliance and outcome in obsessive-compulsive disorder

    Behavior Modification

    (2002)
  • J.S. Abramowitz et al.

    Symptom presentation and outcome of cognitive-behavioral therapy for obsessive-compulsive disorder

    Journal of Consulting and Clinical Psychology

    (2003)
  • I., M. Aderka et al.

    Sudden gains in the treatment of obsessive-compulsive disorder

    Psychotherapy and Psychosomatic

    (2012)
  • American Psychiatric Association

    Diagnostic and statistical manual of mental disorders

    (2013)
  • R.C. Arnau et al.

    Psychometric evaluation of the Beck Depression Inventory-II with primary care medical patients

    Health Psychology

    (2001)
  • A.T. Beck et al.

    Comparison of Beck Depression Inventories–IA and –II in psychiatric outpatients

    Journal of Personality Assessment

    (1996)
  • M. Bloch et al.

    Long-term outcome in adults with obsessive-compulsive disorder

    Depression and Anxiety

    (2013)
  • M.J. Boschen et al.

    Treatment of severe, treatment-refractory obsessive-compulsive disorder: A study of inpatient and community treatment

    CNS spectrums

    (2008)
  • L. Collins et al.

    Sudden gains in exposure therapy for obsessive-compulsive disorder

    Behaviour Research and Therapy

    (2017)
  • D.D. Dougherty et al.

    Prospective long-term follow-up of 44 patients who received cingulotomy for treatment-refractory obsessive-compulsive disorder

    The American Journal of Psychiatry

    (2002)
  • Cited by (10)

    • Examination of outcomes among sexual minorities in treatment for obsessive-compulsive and related disorders

      2022, Journal of Obsessive-Compulsive and Related Disorders
      Citation Excerpt :

      All participants engaged in up to 4 h of either coached or self-directed ERP per day for a mean of 62.2 (SD = 22.7) days in the program. For more information about the IRT program, refer to Krompinger et al. (2017) and Stewart et al. (2005). For each construct of interest (YBOCS, DTS, HAMD), we submitted scores to separate 2 (Group: sexual minorities, heterosexual) X 2 (Time: baseline, discharge) analysis of variance (ANOVAs) with repeated measures on time.

    • Feasibility and Acceptability of Cognitive Bias Modification for Interpretation as an Adjunctive Treatment for OCD and Related Disorders: A Pilot Randomized Controlled Trial

      2022, Behavior Therapy
      Citation Excerpt :

      Further examination revealed that the effects were generally largest in magnitude between pre-intervention and mid-intervention as opposed to mid-intervention and post-intervention, suggesting that the intervention has its largest impact after the first half of the eight training sessions. This is consistent with several recent studies supporting early gains in treatment in IRT for OCD (Brennan et al., 2014; Falkenstein et al., 2019; Krompinger et al., 2017). Possible reasons for these gains during the first four sessions could include that participants’ levels of motivation and willingness to engage in the training may have been highest during this time, along with having recently heard the rationale for CBM-I, and the training being more novel and less repetitive to the participants at this stage.

    • Empirically-derived response trajectories of intensive residential treatment in obsessive-compulsive disorder: A growth mixture modeling approach

      2019, Journal of Affective Disorders
      Citation Excerpt :

      This required us to make a partially theory-driven selection of the quadratic 2-class model. Examination of the trajectories within this quadratic 2-class model was also consistent with clinical impressions and prior research in other samples in our IRT program (Krompinger et al., 2017), showing that there is a meaningful difference between early and later response. Fourth, the participants in this study were receiving different types of medications which limits our ability to fully evaluate the effects of medication on trajectories of response.

    View all citing articles on Scopus
    View full text