درمان سوگ و داغداری و اختلالات پیچیده استفاده از مواد مخدر: یک مطالعه مقدماتی

Abstract Empirically supported treatments for co-occurring substance use disorders (SUDs) and grief problems are lacking, despite the salience of grief pathology in substance abusers. Identification of a syndrome of complicated grief, distinct from bereavement-related depression and anxiety, led to the development of a targeted treatment, but this treatment has not been tried with persons with SUDs. We recruited 16 adults with complicated grief and substance dependence or abuse into an open pilot study of a manualized 24-session treatment, incorporating motivational interviewing and emotion coping and communication skills into our efficacious complicated grief treatment. Completer and intent-to-treat analyses showed significant reductions in Inventory of Complicated Grief and Beck Depression Inventory scores, with large effect sizes. Timeline Followback percent days abstinent increased significantly in both analyses, with medium to large effect sizes, and cravings declined significantly. Study limitations notwithstanding, complicated grief and substance use treatment appears to be a promising intervention that merits further research.

1. Introduction Grief has long been recognized as salient in treating persons with substance use disorders (SUDs). Bellwood (1975) described addressing “unresolved grief” as a key to successful alcoholism treatment, and Blankfield (1982/1983) found intense grief or significant bereavement in 20% of consecutive inpatient admissions to a substance abuse treatment center. Yet, despite numerous published clinical accounts of grief treatment in those who abuse or are dependent on substances, no controlled study in which grief-specific symptoms were defined or in which both grief and substance abuse outcomes were assessed has yet been reported. A number of terms have been used in the literature to designate grief that is persistent and impairing. However, until the past decade, this work was not empirically based and there was no reliable way to identify such a condition. In contrast, several research groups have now identified a grief-specific condition characterized by prominent separation distress and causing chronic and clinically significant impairment (Horowitz et al., 1997 and Prigerson et al., 1999). Sufferers display persistent yearning or longing for the deceased, loneliness, preoccupation with thoughts of the deceased, intrusive images or memories, avoidance behaviors, anger and bitterness, survivor guilt, and inability to accept the death. This postloss stress syndrome is called complicated grief. A self-report instrument, the Inventory of Complicated Grief (ICG; Prigerson, Maciejewski, et al., 1995), was developed to assess grief-specific symptoms; a score of ™ 25 identifies the syndrome when the instrument is administered ≥6 months after a death. Factor analysis showed the ICG to measure a single underlying construct. The measure demonstrated excellent internal consistency (α = .94) and high 6-month retest reliability (r = .80). It showed good convergence (all r = .70–.87) with other measures designed to assess grief-related distress while also differentiating complicated grievers from normal grievers based on negative health consequences of bereavement. Several investigators have replicated the finding that complicated grief symptoms can be distinguished from depression and anxiety symptomatology (Boelen & van den Bout, 2005, Boelen et al., 2003, Ogrodniczuk et al., 2003, Prigerson et al., 1996 and Prigerson et al., 1995). Complicated grief is a postloss stress syndrome that bears some resemblance to posttraumatic stress disorder (PTSD). However, traumatic stress results from exposure to a life-threatening event, whereas complicated grief results from the loss of a life-sustaining person. As a result, sadness and loneliness are prominent in complicated grief, whereas fear and arousal are more pronounced in PTSD. Furthermore, symptoms of longing and yearning, as well as pleasurable reveries, are characteristic of complicated grief and clearly distinct from traumatic stress symptoms. Studies have shown moderate rates of comorbidity among complicated grief, major depressive disorder (MDD), and PTSD—similar to rates of comorbidity for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association [APA], 1994) mood and anxiety disorders ( Melhem et al., 2001, Melhem et al., 2004 and Silverman et al., 2000). In studies that controlled for the presence of both MDD and PTSD, complicated grief was associated with greater mental health and psychosocial impairments ( Ott, 2003, Prigerson et al., 1997, Prigerson et al., 2000 and Silverman et al., 2000), higher risk of suicidality ( Latham & Prigerson, 2004), and physical health problems in the aftermath of a loss ( Chen et al., 1999). A summary of the evidence for the specificity of complicated grief can be found in the work of Lichtenthal, Cruess, and Prigerson (2004). Emergent evidence suggests a link between intense grief and worsening of substance use (Prigerson et al., 1997). Parents who lost a child were found to be at significantly higher risk for hospitalization for substance abuse than parents who had not lost a child (Li, Laurson, Precht, Olsen, & Mortensen, 2005); the effect was especially strong on bereaved mothers, whose relative risk of hospitalization was more than double that of mothers who were not bereaved. In a survey study (Shear, Zuckoff, et al., 2005), we found a high rate of complicated grief among patients in a methadone maintenance program. Psychiatric severity, generally (Kranzler et al., 1996 and McLellan et al., 1983), and co-occurring mood or anxiety disorder, specifically (Charney et al., 1998, Dodge et al., 2005, Hasin et al., 2002 and Ouimette et al., 1998), are associated with poor SUD treatment outcomes. Treatment of co-occurring PTSD has been shown to be a positive predictor of 5-year substance use remission rates (Ouimette, Moos, & Finney, 2003), and successful treatment of depression diminishes the quantity of substance use (Nunes & Levin, 2004). Persons with SUDs who have co-occurring complicated grief would likewise seem likely to benefit from effective treatment of the syndrome. We developed and pilot tested (Shear, Frank, et al., 2001) a novel complicated grief treatment (CGT). Results of a randomized controlled trial of 16 sessions of CGT showed this treatment to be superior to a 16-session standard psychotherapy control (Shear, Frank, Houck, & Reynolds, 2005). However, persons with SUDs have been excluded from these studies, in the belief that special adaptations would be required to make treating them safe and feasible. We therefore undertook a treatment development project to adapt CGT for persons who abuse or are dependent on substances. The results of an open prospective pilot study are reported here.

3. Results 3.1. Preliminary analyses 3.1.1. Participant characteristics Nine women and seven men signed informed consent forms for the study and had at least one treatment session. The mean time since the death that was the focus of treatment was 9.8 years (SD = 9.7 years, range = 0.7–31.7 years, Mdn = 7.6 years). Seven participants were grieving over violent deaths, and nine participants were grieving over nonviolent deaths. The mean age of the participants was 42.3 years (SD = 9.8 years, range = 24–57 years). Eight participants were African American, seven were Caucasian, and one was Native American. One was married; six were never married; and nine were widowed, separated, or divorced. Four had lower than high school education, two were high school graduates or equivalent, six had some postsecondary education, and four had a postsecondary degree. Most (n = 12) were unemployed. SUDs at baseline included alcohol dependence (n = 3) or abuse (n = 4), cannabis dependence (n = 3) or abuse (n = 1), and cocaine dependence (n = 3). Three participants entered treatment with opiate dependence and were on agonist therapy (methadone). Participants reported use of any substance in their lifetime a median of 24 years (1,250 weeks, range = 484–2,056 weeks). In their lifetime, all 16 participants used alcohol (Mdn = 546 weeks, range = 130–2,056 weeks), 15 used cannabis (Mdn = 260 weeks, range = 0–1,430 weeks), 12 used cocaine (Mdn = 172 weeks, range = 0–1,496 weeks), 11 used hallucinogens (Mdn = 1 week, range = 0–156 weeks), 9 used amphetamines (Mdn = 1 week, range = 0–364 weeks), 9 used opiates (Mdn = 1 week, range = 0–1,673 weeks), 7 used benzodiazepines (Mdn = 1 week, range = 0–520 weeks), 5 used inhalants (Mdn = 0 week, range = 0–6 weeks), and 4 used hypnotics (Mdn = 0 week, range = 0–520 weeks). Excluding prescribed methadone, during the 90 days prior to baseline, participants used a mean of 1.6 (range = 0–3) types of substances and used substances on 58% (SD = 36.5) of days. During this period, one participant on methadone was otherwise abstinent; 12 participants drank alcohol (median drinks per drinking day = 3, range = 0–17), 9 used cannabis, 4 used cocaine, 1 used benzodiazepines, and 1 used opiates. All participants had at least one nonsubstance use DSM-IV Axis I diagnosis at baseline (M = 2.1, range = 1–5), including MDD (n = 12), PTSD (n = 11), panic disorder (n = 4), generalized anxiety disorder (n = 4), and specific phobia (n = 1). Eleven participants were on psychotropic medication during study participation: 10 on antidepressants, 3 on benzodiazepines, 3 on neuroleptics, 2 on mood stabilizers, and 1 on nonbenzodiazepine sleep medication. In addition to the three participants on methadone who were enrolled in public methadone maintenance programs, two participants recruited from our low-income community clinic continued to receive supportive counseling during study participation and one participant was enrolled in a residential program for mothers with addictions. All participants judged grief to be their primary problem, with the exception of their SUDs. 3.1.2. Comparison of treatment completers and noncompleters Eight participants (five men and three women) completed the treatment, whereas six women and two men were noncompleters. The mean number of sessions for noncompleters was 9.3 (SD = 5.5, range = 1–15). One participant dropped out because of unwillingness to continue grief-focused procedures, and three dropped out for unknown reasons. Two participants were withdrawn for medical reasons (abnormal electrocardiogram, gastrointestinal disorder), and one was withdrawn for failure to attend treatment sessions. One participant was withdrawn for worsening substance use and depression after nine sessions. This participant was one of two enrolled early in the study whose condition worsened after telling the story of the death during the first treatment session. The other participant was able to restabilize and successfully complete the treatment. Nonetheless, after these events, we changed the protocol such that the story was not told until after the patient completed the initial treatment phase; no further case of worsening occurred. Two completers each had one positive breathalyzer test; no noncompleter had positive breathalyzers. The two sessions in question were rescheduled, and we ensured that the patient was seen home safely, with no further complication. Although higher proportions of women and those grieving over violent deaths were noncompleters, there was no statistically significant relationship between completion and sex (71% males vs. 37% females; h = .78, p = .31) or type of death (67% nonviolent vs. 29% violent; h = .78, p = .31). Among participants on antidepressants, six completed treatment and four did not, whereas two participants not on antidepressants completed treatment and four did not (p = .61). Although the difference was not significant, completers had a lower proportion of abstinent days at baseline (32% vs. 53%; p = .51). 3.2. Treatment outcomes 3.2.1. Symptom scores Grief, depression, and substance use symptom outcomes are summarized in Table 1. Table 1. Scores on measures of grief, depression, and substance use in CGSUTa participants Participant group ICG BDI TLFB percent days abstinent Pretreatment Posttreatment Δ Initial session Final session Δ Pretreatment In treatment Δ Completer (n = 8) M 49.0 18.1 −30.9⁎ 26.5 11.0 −15.5⁎ 32.0 58.5 26.5⁎⁎ SD 9.8 14.0 15.4 12.7 10.9 5.5 26.8 33.7 29.8 Noncompleter (n = 8) b M 46.1 46.4 0.3 27.0 26.6 −0.4 52.6 67.0 14.4 SD 9.7 8.2 6.5 10.1 13.8 7.3 43.5 31.0 55.4 Intent-to-treat (n = 16) M 47.6 32.3 −15.3⁎ 26.8 18.8 −7.9⁎ 42.3 62.8 20.4⁎⁎ SD 9.5 18.3 19.7 11.1 14.5 10.0 36.5 31.6 43.4 a Manual-guided individual outpatient treatment conducted in 24 sessions over approximately 6 months. b Last observation carried forward. ⁎ p = .01. ⁎⁎ p < .05. Table options Significant pretreatment-to-posttreatment reductions were found in ICG scores in completers (M = 30.9, SD = 15.4, S = 18, p = .01) and intent-to-treat analysis (M = 15.3, SD = 19.7, S = 48, p = .01), with effect sizes of 2.01 and 0.78, respectively. In a comparable CGT pilot study ( Shear, Frank, et al., 2001), mean reductions in ICG scores were 22.8 (SD = 13.14, z = −3.11, p = .002) among completers and 16.9 (SD = 19.99, z = −3.51, p < .001) in intent-to-treat analysis, with effect sizes of 2.19 and 1.45, respectively. BDI scores showed corresponding reductions for both completer (M = 15.5, SD = 5.5, S = 18, p = .01) and intent-to-treat (M = 7.9, SD = 10.0, S = 40, p = .01) groups, with effect sizes of 2.82 and 0.79, respectively. Again, this was similar to reductions in BDI scores in the CGT pilot study for the completer (M = 13.1, SD = 10.19, z = −2.98, p = .003) and intent-to-treat (M = 10.4, SD = 9.93, z = −3.44, p = .001) groups, with effect sizes of 1.80 and 1.16, respectively. TLFB percent days abstinent from all substances increased significantly among the completer (M = 26.5, SD = 29.8, S = 15, p = .04) and the intent-to-treat (M = 20.4, SD = 43.4, S = 39, p = .04) groups. Effect sizes were 0.89 and 0.47, respectively. Among completers, mean reduction in ICG score was 30.8 for patients on antidepressants (n = 6) and 31.0 for patients not on antidepressants (n = 2). In intent-to-treat analysis, mean reduction in ICG score was 19.1 (SD =20.3) for the 10 participants on antidepressants and 9.0 (SD = 18.7) for the 6 participants not on antidepressants (p = .31). 3.2.2. Cravings A significant negative slope was found in the mixed-model analysis of the intent-to-treat sample, such that average cravings decreased over time. Mean craving score was 2.2 at treatment initiation, whereas the predicted value at treatment completion was 1.6, F (1, 13) = 5.30, p = .04, d = 1.30.