Two-Year Follow-Up Study of Inpatients

Two-Year Follow-Up of Inpatients
With Dissociative Identity Disorder

Joan W. Ellason, M.A., L.P.C., and Colin A. Ross, M.D.

Objective: A patient group of 135 inpatients with dissociative identity disorder was followed for two years to monitor treatment outcome. Method: Fifty four of these patients were located and re-assessed after a two year period, by using the same self-report measures and structured clinical interviews that had been initially administered. Results: The patients showed marked improvement on Schneiderian first rank symptoms, mood and anxiety disorders, dissociative symptoms, and somatization, with a significant decrease in the number of psychiatric medications prescribed. Patients who were treated to integration, were significantly more improved than those who had not yet reached integration. Conclusion: Our findings, though preliminary, provide empirical validation of previous clinical impressions that patients with dissociative identity disorder may respond well to treatment.
(American Journal of Psychiatry 1997; 154:832-839)

Dissociative identity disorder (DSM-IV) is linked to intense childhood trauma (1-5) and is the most severe form of dissociative disorder (6-9). Traumatic experiences, including incest and childhood abuse experiences, have been linked to symptoms of posttraumatic stress disorder (PTSD) (10-18) and axis I and II disorders (16, 19-23), including Schneiderian first rank symptoms (2-4, 6, 9, 20, 24), and behavioral dysfunction (25-44). Dissociative identity disorder, which is understood to be a disturbance resulting from severe forms of childhood abuse (1-3, 5, 45), presents with auditory hallucinations (46-49), severe depression and suicidality (50), phobic anxiety, somatization, substance abuse (51), and borderline features (4, 6, 14, 52-58). Studies of patients with dissociative identity disorder in several countries, including Canada, the United States, the Netherlands, (59) and Turkey (60), consistently document multiple symptoms.

Freud's early theoretical framework recognized the developmental and emotional impact of trauma on the psyche (61, 62). Trauma, according to both Freudian theory and later studies manifests itself symptomatically (62) through dissociation, posttraumatic stress symptoms, somatization, other axis I syndromes, a developmentally disorganized characterological structure (1, 3, 5, 6, 17, 19, 23-27, 29-33, 38, 40, 42-45, 59), and personality decompensation (62).

Individuals with histories of severe trauma may change details of their retrospective reports over a period of time (63). Consistency of the trauma histories should be tracked systematically in follow-up studies of populations with amnesia symptoms.

Clinical research has begun to track the treatment progress of patients with dissociative identity disorder. These patients have shown noteworthy improvements in depression and anxiety symptoms, impulsivity and addictive behaviors, psychotic symptoms, interpersonal activities and day to day functioning among small patient groups (64; E. Smith, A. Mittelstet, unpublished data, 1995). Coons (65) followed the treatment progress of 20 patients and found that 5 had achieved complete integration over a 39 month period. Integration, as defined by Kluft (66), involves three stable months of continuity of contemporary memory, with the absence of behaviorally evident separate identities. The integrated patient will have a self-perception of internal unity, which is substantiated by the absence of alter identities during hypnotic re-exploration. Transference phenomena become consistent with the unity of the parts, and there is clinical evidence that the patient has incorporated previously separated attitudes and awarenesses into one unified whole. On meeting these criteria, 60% of patients maintain stable integration (66). Patients who reach integration tend to show a moderation of personality disorder characteristics (65), improved stability (66), and approximately half as much emotional turmoil as non-integrated patients (65).

Despite many clinical observations and preliminary studies (64-66; E Smith, A. Mittelstek, unpublished data, 1995) there remains a paucity of systematic empirical investigation of treatment outcome in dissociative identity disorder. The purpose of the present investigation was to track the long-term treatment progress, through symptom measures, and following hospitalization, in a large inpatient group of patients with dissociative identity disorder. These subjects, who participated in an earlier comorbidity study, conducted in 1992 and 1993, were contacted and reassessed following a two-year interval. Another aspect of this investigation involves comparison of patients' initial reports of childhood abuse with those described after a two year interval.

The major expectation of this study was to observe significant overall improvement in axis I and II symptom levels among patients with dissociative identity disorder. A secondary expectation was that patients achieving integration through treatment would show significantly more improvement than non-integrated patients. The number of medications and extent of hospitalization treatment was also examined among the integrated and non-integrated patients. A tertiary expectation was that patient reports of childhood abuse would be consistent from baseline to follow-up testing.

METHOD

Subjects

The initial patient group consisted of 135 inpatients at Charter Behavioral Health System of Dallas who were selected on the basis of having a clinical diagnosis of dissociative identity disorder, as defined by DSM-III-R, proposed DSM-IV criteria, and structured interview, and unequivocally clear behavioral evidence of dissociative identity disorder. At initial assessment, permission for patient participation was obtained through the internal review board and the attending physician; following a thorough description of the study to the subjects, signed, written, informed consent was obtained regarding baseline and follow-up research. Not all of the 135 subjects completed all measures at baseline.

Procedure

Subjects were contacted by telephone and by mail, following a two year interval that involved primarily outpatient psychotherapy with medication. Follow-up interviews included the same assessments that were administered during the initial phase of the study. These were the Dissociative Experiences Scale (68, 69) and the Dissociative Disorders Interview Schedule (67). Both of these instruments measure the severity of dissociation; the Dissociative Disorders Interview Schedule also inquires about childhood trauma and obtains demographic information. The Dissociative Experiences Scale is a 28-item self-report measure with good validity, and with test-re-test reliability of 0.84 (69); its updated version had a Pearson correlation of 0.95 with the original version among 87 inpatients with dissociative identity disorder (68), indicating good convergent validity. The Dissociative Disorders Interview Schedule consists of 131 structured interview items that include demographic information and components to assess history of childhood trauma, consisting of allegations of physical and sexual abuse. The overall interrater reliability is 0.68, with a sensitivity of 95% and a specificity of 100% for the diagnosis of dissociative identity disorder (67). The structured and clinical interview diagnoses had a 99.1 % agreement rate in our initial patient group. One interview with the Dissociative Disorders Interview Schedule resulted in a false negative diagnosis, but the patient later confirmed previously denied symptoms of dissociative identity disorder thereby establishing a positive diagnosis of dissociative identity disorder through clinical interview. The PTSD section of the National Institute of Mental Health Diagnostic Interview Schedule (70) was also administered.

The prevalence of current axis I and II psychopathology was measured by the Structured Clinical Interview for DSM-III-R (SCID I and SCID II) (71). The SCID I and II are widely published instruments designed to assess axis I and II disorders. The Beck Depression Inventory (72) and the Hamilton Rating Scale (73) were also administered. In addition, we tabulated previous and current psychiatric medications and the estimated total number of inpatient days and number of hospitalizations both before and after 1993.

Of the original patient group (N = 135), 62 (45.9%) were located. Of those contacted (N = 62), 54 (87.1%) were re-interviewed. Seventy three (54.1%) either had no forwarding address or did not respond to phone calls and letters. Only one patient (1.6%) of those contacted refused to participate because of retraction of the diagnosis. Other reasons given for declining follow-up were willingness to participate at a later date when less preoccupied with the therapeutic process (reason given by a small number of patients), refusal due to continued trust issues (one patient), and initial willingness to be re-interviewed which was prevented by a subsequent job injury. Patients who could not be reached by telephone were mailed information; letters that were returned with no forwarding address prevented further efforts at follow-up contact.

Follow-up interviews were completed by a series of telephone calls for patients in other parts of the United States and Canada (N = 36, 66.7%), with the exception of one non-local patient (1.9%), who traveled to our office for the interview. All interviews for local patients (N = 15, 27.8%) were conducted in person, with the exception of two patients (3.7%), whose schedules would not permit travel to the office. All re-interviewed patients were outpatients, with the exception of one who was in day hospital treatment during part of the interview. Along with follow-up symptom measures, patients were administered identical questions about their abuse histories, with no cues or information given about their baseline responses.

Following additional informed consent, therapists were consulted on all six of the criteria for integration outlined by Kluft (66). Patients were considered to be integrated on the basis of two prerequisites: 1) fulfillment of all six of Kluft's criteria for integration, and 2) current failure to meet DSM-III-R and DSM-IV criteria for dissociative identity disorder on structured interview (67).

Data Analyses

Baseline data on demographic, Dissociative Disorders Interview Schedule abuse history, and all measures were compared among participants and non-participants using two-tailed t tests for continuous data and chi square tests for dichotomous data. All subjects completed the Dissociative Disorders Interview Schedule at baseline and at follow-up. At baseline all 54 subjects completed the Dissociative Experiences Scale, while 51 completed this instrument at follow-up. The Beck inventory was completed by 47 patients initially and by 49 at follow-up. The Hamilton depression scale had 47 and 48 baseline and follow-up subjects, respectively. SCID I had 48 and 49 subjects, at baseline and follow-up, while the SCID II had the same number at baseline and follow-up (N = 47). Follow-up participants who did and did not achieve integration were compared at baseline also on the previously described variables.

Gender, marital status, age, employment status, types sexual of abuse, and the number of perpetrators of physical and sexual abuse were tabulated. Mean values were calculated on continuous variables, and the prevalence of axis I and II disorders were tabulated as percentages of subjects with such disorders. Chi Square and paired t tests were used to compare baseline and follow-up scores. The time frames for the Beck inventory, and the Hamilton depression scale were symptom levels during the previous week. The time frames for the Dissociative Disorders Interview Schedule, SCID I and SCID II were the previous year. For medications, the time frame was the past two years both at baseline and at follow-up. An analysis of variance (ANOVA) for repeated measures was conducted by using the General Linear Model program of SAS, which is designed to analyze data, unbalanced in subject number; SAS defaults at a maximum p value of 0.0001.

TABLE 1. SCID Axis I and II Diagnoses and Dissociative, Mood Disorder, and Overall Symptoms Among Patients With Dissociative Identity Disorder at Baseline and 2-Year Follow-Up

		1993		1995			Analysis^a

Measure	N	Mean	SD	Mean	SD	t	df	p

Number of SCID-I, SCID-II, and total current diagnoses
SCID-I (axis I)	40	7.3	2.2	3.5	2.2	7.6	39	0.00001
SCID-II (axis II)	43	3.6	2.3	2.2	2.5	3.0	42	0.004
SCID I and SCID-II total	38	10.8	3.8	5.4	3.9	6.0	37	0.00001
Dissociative, mood, and overall symptom scores
Dissociative Experiences Scale	50	52.4	19.1	33.0	20.1	6.5	49	0.00001
Beck Depression Inventory	45	33.4	12.1	23.2	13.1	4.7	44	0.00001
Hamilton Depression Rating Scale	46	43.4	11.6	28.4	12.9	6.5	45	0.00001
Dissociative Disorders Interview Schedule scores
Somatic Symptoms	54	17.2	7.0	10.8	7.5	5.6	53	0.00001
Substance Abuse	54	1.1	1.2	0.3	0.6	5.0	53	0.00001
Depression	54	7.6	0.8	6.2	2.9	3.6	53	0.0007
Schneiderian first-rank symptoms	54	6.8	2.9	4.2	3.0	5.8	53	0.00001
Features of dissociative identity disorder	54	11.5	2.7	8.6	3.8	5.3	53	0.00001
Borderline features	54	5.5	1.8	3.7	2.3	5.4	53	0.00001
Extrasensory perception	54	6.5	3.5	3.2	2.8	7.7	53	0.00001
Amnesia	54	4.7	0.6	3.9	1.6	3.9	53	0.0003

RESULTS

Independent samples t test comparison of follow-up participants and nonparticipants on all measures used in this study showed no significant differences on demographic data, abuse history, or symptom variables at baseline, except for SCID I panic disorder. This disorder was present at baseline in 40 (83.3%) of 48 participants and was present in 34 (57.6%) of 59 non-participants (X². = 7.0, df = 1, p = 0.008). Twelve subjects (22.2%) were identified as having achieved integration, according to Kluft's criteria (66) and structured interview (67). At baseline, the integrated subjects differed slightly from nonintegrated subjects in the number of Dissociative Disorders Interview Schedule depression items (mean= 7.3, SD = 1.2, versus mean = 7.8, SD = 0.3; t = 2.1, df = 52, p = 0.04), and the integrated subjects had lower initial scores on the Hamilton depression scale (N = 12, mean = 36.5, SD = 11.9) than did the non integrated subjects (N = 37, mean = 46.3, SD = 10.4) (t = 2.7, df = 47, p = 0.009). Otherwise, there were no differences between integrated and non-integrated subjects at baseline.

Most of the patients who participated in follow-up were women (N = 48, 88.9%). The patients were fairly evenly distributed in marital status among single (N = 17, 31.5%), married (N = 18, 33.3%), and separated, widowed or divorced (N = 19, 35.2%). The mean age was 39.2 (SD = 9.2). No inquiry was conducted on socioeconomic status, however, slightly under half (N = 24, 44.4%) of the patients were employed at the time of the follow-up interview.

Medications and Hospitalizations

The reported average number of medications prescribed during the previous two years (mean = 3.8, SD = 2.4) was less in 1995 (mean = 2.5, SD = 1.4) (t = 4.1, df = 36, p = .0002). Among integrated subjects, the reduction in number of medications from 1993 to 1995 fell just short of significance, however, the non-integrated subjects reported significantly fewer medications in 1995 (mean = 2.8, SD = 1.4) than in 1993 (mean = 4.4, SD = 2.3) (t = 4.1, df = 28, p = .0003). At follow-up the patients reported (N = 32) an average of 2.7 hospitalizations (SD = 3.2) during the previous two years. For those who estimated their average number of inpatient days (N = 29), an average of 10.9 days (SD = 17.5) was indicated. All patients except for four (7.4%) reported consistently having outpatient psychotherapy during most of this two-year follow-up period.

Reports of Childhood Abuse

There were no significant differences in patients' 1993 and 1995 reports of childhood abuse. Fifty-three (98.1%) of the subjects reported having been physically or sexually abused during childhood. None of these subjects retracted their initial allegations.

Symptoms

Table 1 shows the follow-up results among all patients. In 1995 integrated and non-integrated patients differed significantly on all variables listed in table 1 except substance abuse: this was because of the excellent improvement in both groups on substance abuse scores. Paired samples t test showed that nonintegrated patients significantly improved in the number of current SCID I (N = 32) diagnoses (1993 mean = 7.2 SD = 2.4; 1995 mean = 4.1 SD = 2.0 t = 5.7, df = 31, p = .00001), Hamilton depression scale (N = 34) depression scores (1993 mean = 45.9, SD = 10.7, 1995 mean = 32.7 SD = 10.8, t = 5.1, df = 33, p = .00001), Dissociative Experiences Scale (N = 39) dissociative experiences scores (1993 mean = 52.9 SD = 20.8, 1995 mean = 38.0 SD = 19.4; t = 4.5, df = 38, p = .0001), Dissociative Disorders Interview Schedule (N = 42) extrasensory perception experiences (1993 mean = 6.6 SD = 3.7; 1995 mean = 3.5 SD = 3.0, t = 6.3, df = 41, p = .0001), Dissociative Disorders Interview Schedule (N = 42) Schneiderian first rank symptoms (1993 mean = 7.0 SD = 3.0; 1995 mean = 5.0 SD = 2.9; t = 4.0, df = 41, p = .0002), number of combined current SCID I and II (N = 30) diagnoses (1993 mean = 10.6 SD = 3.9; 1995 mean = 6.3 SD = 3.7; t = 4.3, df = 29, p = .0002), Dissociative Disorders Interview Schedule (N = 42) somatic symptoms (1993 mean =18.0 SD = 7.2; 1995 mean = 12.6 SD = 7.4; t = 4.0, df = 41, p = 0.0003), Dissociative Disorders Interview Schedule (N = 42) substance abuse items (1993 mean = 1.0, SD = 1.2; 1995 mean = 0.3 SD = 0.7; t = 3.6, df = 41, p = .0008), Beck inventory (N = 36) scores (1993 mean = 34.8 SD = 11.8, 1995 mean = 26.6 SD = 12.3); t = 8.2, df = 35, p = .001), Dissociative Disorders Interview Schedule (N = 42) borderline features (1993 mean = 5.4 SD = 1.8; 1995 mean = 4.3 SD = 2.3; t = 3.4, df = 41, p = .002), and Dissociative Disorders Interview Schedule (N = 42) features of dissociative identity disorder (1993 mean = 11.6 SD = 2.9; 1995 mean = 9.8 SD = 2.7; t = 3.4, df = 41, p = .002). Improvement in the number of SCID II (N = 34) diagnoses fell just short of significance (1993 mean = 3.6 SD = 2.2; 1995 mean = 2.6 SD = 2.7; t = 3.6, df = 33, p = 0.056) among non-integrated patients.

Patients, who achieved integration, however, improved significantly on all variables. Paired samples t tests revealed the greatest improvement in SCID I (N = 8), (1993 mean =7.8 SD = 1.5; 1995 mean = 1.3 SD = 1.6; t = 10.9, df = 7, p = .00001), Dissociative Experiences Scale (N = 11) dissociative experiences scores (1993 mean = 50.5 SD = 12.2; 1995 mean = 15.4 SD = 10.2; t = 8.0, df = 10, p = .00001), and Dissociative Disorders Interview Schedule (N = 12) borderline features (1993 mean = 5.9 SD = 1.7; 1995 mean = 1.7. SD = 1.2; t = 7.0, df = 11, p = .00001), followed by Dissociative Disorders Interview Schedule (N = 12) Schneiderian first rank symptoms (1993 mean = 6.2 SD = 2.4; 1995 mean = 1.4 SD = 1.7; t = 5.7, df = 11, p = .0001), Dissociative Disorders Interview Schedule (N = 12) features of dissociative identity disorder (1993 mean = 11.2 SD = 1.7; 1995 mean = 4.1 SD = 3.9; t = 6.4, df = 11, p = .0001), Dissociative Disorders Interview Schedule (N = 12) somatic symptoms [1993 mean = 14.3 SD = 5.3; 1995 mean = 4.3 SD = 2.8; t = 5.8, df = 11, p = .0001), total number of SCID I and II (N = 8) diagnoses [1993 mean = 11.5 SD =3.7; 1995 mean = 1.9 SD = 2.0; t = 7.6, df = 7, p = .0001), Dissociative Disorders Interview Schedule (N = 12) amnesia items (1993 mean = 4.8 SD = 0.8; 1995 mean = 2.3 SD = 2.0; t = 4.6, df = 11, p = .0008), Dissociative Disorders Interview Schedule (N = 12) extrasensory perception (1993 mean = 5.8 SD = 3.2; 1995 mean = 1.9 SD = 1.5; t = 4.6, df = 11, p = .000), Beck inventory (N = 9) depression scores (1993 mean = 27.9 SD = 12.2; 1995 mean = 9.7 SD = 4.8; t = 4.0, df = 8, p = .004), Hamilton depression scale (N = 12) depression scores (1993 mean = 36.5 SD = 11.9; 1995 mean = 16.0 SD = 10.0; t = 4.2, df = 11, p = 0.002), Dissociative Disorders Interview Schedule (N = 12) substance abuse items (1993 mean = 1.3 SD = 1.2; 1995 mean = 0.0, SD = 0.0; t = 4.0, df = 11, p = .002), Dissociative Disorders Interview Schedule depression items (1993 mean = 7.3 SD = 1.2; 1995 mean = 4.3, SD = 3.6; t = 2.7, df = 11, p = .02), and number of SCID II (N = 9) diagnoses (1993 mean = 3.6 SD = 2.5; 1995 mean = 0.7 SD = 1.0; t = 3.0, df = 8, p = .02). Results of ANOVA for main effects of baseline/follow-up, integration, and their interaction effects are presented in Table 2.

DISCUSSION

At follow-up, the overall group had improved considerably on dissociative symptoms; the average Dissociative Experiences Scale score was typical for dissociative identity disorder in 1993 but had dropped out of this range in 1995 (74). The patients showed significant improvement on substance abuse, depression and symptoms that mimic psychosis, while simultaneously reducing their number of antidepressant and antipsychotic drugs. Although dosage was not inquired about routinely, those who reported dosage described a reduction in dosage over the two year period. In 1995, all cases of psychotic disorder not otherwise specified, except for one, were the result of a more severe baseline psychotic diagnosis decreasing in symptom level; as a result, those patients no longer met full criteria for a specified psychotic disorder. The one exception was a case of a consistently maintained SCID diagnosis of psychotic disorder not otherwise specified.

A considerable amount of improvement among both integrated and non-integrated patients occurred in several areas. A similar decrease in the number of SCID I & II (axis I & II) and total SCID diagnoses occurred among these two groups, with the most significant improvement occurring in the number of axis I symptoms, regardless of integration status. The SCID I diagnostic categories that showed the most consistent improvement, among patients in both the preintegration and postintegration stages of progress, were the anxiety, somatic, and psychotic diagnoses. Patients with two years of treatment who integrated showed more substantial improvement on Dissociative Disorders Interview Schedule secondary features of dissociative identity disorder and amnesia symptoms, with slightly more significant improvement on suicidality and on Dissociative Experiences Scale and Beck inventory scores than those who did not integrate.

Although at baseline, the integrated patients had lower Hamilton depression scale scores and fewer Dissociative Disorder Interview Schedule depression symptoms than the nonintegrated patients, the interaction effect indicates a substantially greater improvement on Dissociative Disorders Interview Schedule depression symptoms among integrated patients than among non-integrated patients. The Hamilton depression scale measures a broad range of mood, somatic, sleep disorder, and other comorbid depression symptoms, while the Dissociative Disorders Interview Schedule depression items are congruent with APA diagnostic criteria for major depression (DSM-II-R and DSM-IV). The combined effect of integration and treatment on depressive symptoms is a major treatment benefit: depression is the most common comorbid feature in dissociative identity disorder (75).

Table 2. Symptoms at Baseline and 2-Year Follow-Up Among 54 Integrated and Nonintegrated Patients With Dissociative Identity Disorder, With Interaction Effects Between Integration and Symptom Improvement

			Analysis^a

Main and Interaction Effects	Mean Square	F (df=1, 52)	p


Number of current SCID-I symptoms
All patients' 1993 versus 1995 symptoms	110.00	36.81	0.0001
Integrated versus nonintegrated	22.00	4.55	0.04
Integration and 1993/1995 scores^b	6.67	2.23	n.s.
Between-group error^c	4.83
Within group error^c	2.99
Number of Current SCID-II symptoms
All patients' 1993 versus 1995 symptoms	36.64	9.75	0.003
Integrated versus nonintegrated	13.14	2.20	n.s.
Integration and 1993/1995 scores^b	10.04	2.67	n.s.
Between-group error^c	5.99
Within group error^c	3.76
Total number of current SCID-I and SCID-II symptoms
All patients' 1993 versus 1995 symptoms	266.19	29.64	0.0001
Integrated versus nonintegrated	62.73	4.11	0.05
Integration and 1993/1995 scores^b	34.00	3.79	0.06
Between-group error^c	15.26
Within group error^c	8.98
Dissociative Experience Scale scores
All patients' 1993 versus 1995 symptoms	11,028.82	54.80	0.0001
Integrated versus nonintegrated	2,963.17	6.23	0.02
Integration and 1993/1995 scores^b	1,423.17	7.07	0.01
Between-group error^c	475.32
Within group error^c	201.25
Beck Depression Inventory scores
All patients' 1993 versus 1995 symptoms	3,130.31	31.75	0.0001
Integrated versus nonintegrated	2,119.01	12.33	0.001
Integration and 1993/1995 scores^b	490.31	4.97	0.03
Between-group error^c	171.82
Within group error^c	98.61
Hamilton Depression Rating Scale scores
All patients' 1993 versus 1995 symptoms	4,896.17	41.86	0.0001
Integrated versus nonintegrated	2,240.93	21.54	0.0001
Integration and 1993/1995 scores^b	131.22	1.12	n.s.
Between-group error^c	104.02
Within group error^c	116.98
Dissociative Disorders Interview Schedule
Somatic symptoms
All patients' 1993 versus 1995 symptoms	1,107.43	32.12	0.0001
Integrated versus nonintegrated	757.21	12.33	0.0009
Integration and 1993/1995 scores^b	84.76	2.46	n.s.
Between-group error^c	61.41
Within group error^c	34.48
Substance Abuse items
All patients' 1993 versus 1995 symptoms	17.79	27.79	0.0001
Integrated versus nonintegrated	0.01	0.01	n.s.
Integration and 1993/1995 scores^b	2.38	3.72	0.06
Between-group error^c	1.17
Within group error^c	0.64
Depression
All patients' 1993 versus 1995 symptoms	85.24	25.92	0.0001
Integrated versus nonintegrated	35.29	9.31	0.004
Integration and 1993/1995 scores^b	30.57	9.30	0.004
Between-group error^c	3.79
Within group error^c	3.29
Schneiderian first-rank symptoms
All patients' 1993 versus 1995 symptoms	211.13	41.65	0.0001
Integrated versus nonintegrated	88.11	8.46	0.005
Integration and 1993/1995 scores^b	35.91	7.08	0.01
Between-group error^c	10.42
Within group error^c	5.07
Secondary features of dissociative identity disorder
All patients' 1993 versus 1995 symptoms	374.01	75.27	0.0001
Integrated versus nonintegrated	177.46	17.56	0.0001
Integration and 1993/1995 scores^b	136.12	27.19	0.0001
Between-group error^c	10.11
Within group error^c	4.97
Borderline symptoms
All patients' 1993 versus 1995 symptoms	145.48	58.47	0.0001
Integrated versus nonintegrated	24.13	4.67	0.04
Integration and 1993/1995 scores^b	49.29	19.81	0.0001
Between-group error^c	5.17
Within group error^c	2.49
Extrasensory perception experiences
All patients' 1993 versus 1995 symptoms	226.34	45.37	0.0001
Integrated versus nonintegrated	26.72	1.73	n.s.
Integration and 1993/1995 scores^b	3.52	0.71	n.s.
Between-group error^c	15.46
Within group error^c	4.99
Number of amnesia symptoms
All patients' 1993 versus 1995 symptoms	110.64	37.02	0.0001
Integrated versus nonintegrated	57.56	9.21	0.004
Integration and 1993/1995 scores^b	41.01	13.72	0.0005
Between-group error^c	6.25
Within group error^c	2.99
Suicidality symptoms
All patients' 1993 versus 1995 symptoms	57.95	43.94	0.0001
Integrated versus nonintegrated	7.43	1.99	n.s.
Integration and 1993/1995 scores^b	6.10	4.62	0.04
Between-group error^c	3.74
Within group error^c	1.32

Consistent with clinical evidence that internal voices cease following integration (66), our patients showed a significant decrease in the number of Schneiderian first rank symptoms following integration. Apparently, there is an overall decrease in Schneiderian symptoms with treatment, and integration is a contributing factor to that decrease. The average of 1.4 Schneiderian symptoms among the integrated group, however, includes pre-integration symptoms experienced during the past year, because the integrated patients consistently reported that internal voices became totally absent after integration. Therefore, further follow-up may show a more highly significant interaction and main effect due to integration.

An intriguing finding was the significant reduction in borderline symptoms, which occurred following both treatment and integration. This finding substantiates prior clinical and research observations that DID patients exhibit increased stability after integration (65, 66, 76).

We do not have information on the date of the initial diagnosis of dissociative identity disorder or on the total length of psychiatric treatment for each of the patients, therefore the full treatment response profile of dissociative identity disorder cannot be determined within the scope of this study. Because of selection criteria at baseline, 100% of the patients met criteria for DID. Our data indicate that the diagnosis of dissociative identity disorder appear to be stable and correct, except among patients who reach integration, at which point the diagnosis goes into remission but is not revised retrospectively. Future research will track our patients prospectively from the time they are diagnosed with dissociative identity disorder. Generally patients with dissociative identity disorder are treated with psychodynamic therapy, hypnosis, and adjunctive treatment (77), including medication and group therapy; further studies will involve systematic assessment of the specific outpatient treatment techniques used in prospectively followed cases.

The attrition rate is an important limitation of the study. Although the differences in demographic features and symptoms among participants and non-participants are minimal, our findings are preliminary and give no indication of treatment response in the non-follow-up group. Many patients could not be located despite three or more telephone calls and mail correspondence. Nonresponders could refuse participation for several reasons, ranging from lack of improvement to priority placed on therapy over participation in research. Since participants and non-participants did not differ statistically on baseline measures, with the exception of the higher prevalence of panic disorder among participants, it is possible that inclusion of non-participants might have resulted in lower symptom scores at follow-up.

Baseline characteristics such as compliance, reliability, suggestibility, and positive or negative therapeutic alliance at discharge were not directly measured in the patients, so we cannot comment on them. In future studies we will attempt to minimize the drop-out rate by more frequent contact with patients and we will monitor a broader number of variables including medications and dose. We also plan to attempt to validate the abuse allegations of patients with documented or second party corroboration. In the present study no effort was made to investigate the accuracy of reported childhood trauma. Despite the methodological limitations of the study, the findings indicate that many patients with dissociative identity disorder improve on 2-year follow-up, while those achieving integration do extremely well.

Overall, our results appear to be consistent with recent findings on other dissociative patients (64, 65; E. Smith, A. Mittelstet, unpublished data, 1995). The Dissociative Experiences Scale scores, which include items on memory problems, showed significant improvement among both integrated and non-integrated patients in 1995, with consistent recollections of childhood abuse history from 1993 to 1995. Although recollections of abuse can be prone to error (78) and inconsistency (63), our patients did not change their recollections of childhood abuse. None of our subjects recanted their earlier reports of abuse. In addition, patients who achieved integration by 1995 reported no less severe physical or sexual trauma during childhood than patients who had not yet reached integration. This suggests that a severe abuse history may not preclude a positive prognosis.

Methodological strengths of this investigation include the fact that our patient group encompasses a broad range of age and geographical areas, including many parts of the United States and Canada. The follow-up subjects are representative of the entire original patient group (N = 135), because their baseline (1993) profile did not differ on most items from those of the non-participants.

CONCLUSIONS

Our findings provide information about the treatment response of inpatients with dissociative identity disorder who were willing and able to be reassessed. Despite the limitations of the study, our findings add substance to the claims of Loewenstein (79) regarding the positive course of many correctly diagnosed dissociative identity disorder patients.

The reader should not infer from these data that integration alone directly causes symptom improvement. It is our position that the treatment process leads to recovery, with integration being a marker of more complete recovery. According to research interviews with our patients, integration occurred following substantial inpatient and outpatient therapeutic work. Much of the symptom improvement can be attributed statistically to the combined effects of treatment and integration, but integration is not a separate phenomenon clinically. Although dissociative identity disorder is a severe dissociative disorder requiring long-term treatment, one can conclude from our findings that patients with this disorder often respond very well to treatment within a reasonable period of time, given their degree of baseline comorbidity.

REFERENCES

1. Kluft RP (Ed): Childhood Antecedents of Multiple Personality Disorder. Washington,DC, American Psychiatric Press, 1985

2. Ross CA Multiple Personality Disorder: Diagnosis, Clinical Features, and Treatment. New York, John Wiley & Sons, 1989.

3. Ross CA, Norton GR, Wozney K: Multiple personality disorder: An analysis of 236 cases. Can J Psychiatry 1989; 34:413-418

4. Ross CA, Joshi S, Currie R: Dissociative experiences in the general population Am J Psychiatry 1990; 147:1547-1552

5. Putnam FW, Guroff JJ, Silberman EK, Barban L, Post RM: The clinical phenomenology of multiple personality disorder: Review of 100 recent cases. Am J Psychiatry 47:285-293, 1986.

6. Boon S, Draijer N: Multiple Personality in the Netherlands. Amsterdam, Swets & Zeitlinger, 1993.

7. Coons PM: Dissociative identity disorder not otherwise specified: A clinical investigation of 50 cases with suggestions for typology and treatment. Dissociation 1992; 5:187-195

8. Ross CA: DSM-III problems in diagnosing partial forms of multiple personality disorder: Discussion paper. Journal of the Royal Society of Medicine 1985; 78:933-936

9. Ross CA, Anderson G, Fraser G, Reagor, P, Bjornson, L, Miller, S: Differentiating multiple personality disorder and dissociative disorder not otherwise specified. Dissociation 1992; 5:87-90

10. Kuch K, Cox BJ: Symptoms of PTSD in 124 survivors of the holocaust. Am J Psychiatry 1992; 149:337-340

11. Southwick SM, Yehuda R, Giller EL: Personality disorders in treatment seeking combat veterans with posttraumatic stress disorder. Am J Psychiatry; 1993 150: 1020-1023

12. Mellman TA, Randolph CA, Brawman-Mintzer O, Flores LP, Milanes FJ: Phenomenology and course of psychiatric disorders associated with combat-related posttraumatic stress disorder. Am J Psychiatry 1992; 149:1568-1574

13. Carlson EB, Rosser-Hogan R: Trauma experiences, posttraumatic stress, dissociation, and depression in Cambodian refugees. Am J Psychiatry 1991; 148:1548-1551

14. Roszell DK, McFall ME, Malas KL: Frequency of symptoms and concurrent psychiatric disorder in Vietnam veterans with chronic PTSD. Hosp Community Psychiatry 1991; 42:293-296

15. Shalev AY, Schreiber S, Galai T: Early psychological responses to traumatic injury. Journal of Traumatic Stress 1993; 6:441-450

16. Briere J, Evan D, Runtz M, Wall T: Symptomatology in men who were molested as children. Am J Orthopsychiatr 1988; 58:457-461

17. Goodwin J: Mary Reynolds: A post-traumatic reinterpretation of a classic case of multiple personality disorder. Hillside J of Clinical Psychiatry 1987; 9:89-99

18. Jaschke VA, Spiegel D: A case of probable dissociative disorder. Bulletin of the Menninger Clinic 1992; 56:246-260

19. Pribor EF, Dinwiddie SH: Psychiatric correlates of incest in childhood. Am J Psychiatry 1992; 149:52-56

20. Anderson G, Yasenik L, Ross CA: Dissociative experiences and disorders among women who identity themselves as sexual abuse survivors. Child Abuse & Negl 1993; 17:677-686

21. Goldfarb LA: Sexual abuse antecedents to anorexia nervosa, bulimia, and compulsive overeating: Three case reports. International J of Eating Disorders 1987; 6:675-680

22. Walker EA, Katon WJ., Roy-Byrne PP, Jemelka RP, Russo J: Histories of sexual victimization in patients with irritable bowel syndrome or inflammatory bowel disease. Am J Psychiatry 1993; 150:1502-1506

23. Elliot DM, Briere J: Sexual abuse trauma among professional women: Validating the Trauma Symptom Check List-40 (TSC-40). Child Abuse & Negl 1992; 16:391-398

24. Ross CA, Joshi S: Schneiderian symptoms and childhood trauma in the general population. Comprehensive Psychiatry 1992; 33:269-273

25. Goldman SJ, D'Angelo EJ, DeMaso DR, Mezzacappa E: Physical and sexual abuse histories among children with borderline personality disorder. Am J Psychiatry 1992;149:1723-1726

26. Briere JN: Child Abuse Trauma. London, Sage Publications, 1992.

27. van der Kolk, BA, Perry JC, Herman JL: Childhood origins of self-destructive behavior. Am J Psychiatry 1991; 148:1665-1671

28. Cavaiola AA, Schiff M: Behavior sequelae of physical and/or sexual abuse in adolescents. Child Abuse & Neglect 1988; 12:181-188

29. van der Kolk BA: Psychological Trauma. Washington, DC, American Psychiatric Press, 1987.

30. Zanarini MC, Gunderson JG, Marino MF, Schwartz EO, Frankenburg FR: Childhood experiences of borderline patients. Comprehensive Psychiatry 1989; 30:18-25

31. Braver M, Bumberry J, Green K, Rawson R: Childhood abuse and current psychological functioning in a university counseling center population. J Counseling Psychology 1992; 39:252-257

32. Nigg JT, Silk KR, Westen D, Lohr NE, Gold LJ, Goodrich S, Ogata, S: Object representations in the early memories of sexually abused borderline patients. Am J Psychiatry 1991; 148:864-869

33. Ogata SN, Silk KR, Goodrich S: Childhood sexual and physical abuse in adult patients with borderline personality disorder. Am J Psychiatry 1990;147:1008-1013

34. Westen D, Ludolph P, Misle B, Russins S, Block J: Physical and sexual abuse in adolescent girls with borderline personality disorder. Am J Orthopsychiatr 1990; 60:55-66

35. Hurlbert DF, Apt C, White LC: An empirical examination into the sexuality of women with borderline personality disorder. J of Sex and Marital Therapy 1992; 18:231-242

36. Frank H, Paris J: Recollections of family experience in borderline patients. Arch Gen Psychiatr 1981; 38:1031-1034

37. Zweig-Frank H, Paris J: Parents' emotional neglect and overprotection according to the recollections of patients with borderline personality disorder. Am J Psychiatry 1991; 148:648-651

38. Paris J, Nowlis D, Brown R: Developmental factors in the outcome of borderline personality disorder. Comprehensive Psychiatry 1988; 29:147-150

39. Torgergen S, Alnaes R: Differential perception of parental bonding in schizotypal and borderline personality disorder patients. Comprehensive Psychiatry 1992; 33:34-38

40. Gartner AF, Gartner J: Borderline pathology in post-incest female adolescents. Bulletin of the Menninger Clinic 1988; 52:101-113

41. Ludolph PS, Westen D, Misle B, Jackson A, Wixom J, Wiss FC: The borderline diagnosis in adolescence: Symptoms and developmental history. Am J Psychiatry 1990; 147:470-476

42. Pollock VE, Briere J, Schneider L, Knop J. Mednick SA, Goodwin DW: Childhood antecedents of antisocial behavior: Parental alcoholism and physical abusiveness. Am J Psychiatry 1990; 147:1290-1293

43. Herman JL, Perry JC, van der Kolk BA: Childhood trauma in borderline personality disorder. Am J Psychiatry 1989; 146:490-495

44. Perry JC, Herman JL, van der Kolk BA: Psychotherapy and psychological trauma in borderline personality disorder. Psychiatric Annals 1990; 20:33-43

45. Ross CA, Miller SD, Bjornson L, Reagor P, Fraser G, Anderson G: Abuse histories in 102 cases of multiple personality disorder. Can J Psychiatry 1990; 36:97-101

46. Kluft RP: Unsuspected multiple personality disorder: An uncommon source of protracted resistance, interruption, and failure in psychoanalysis. The Hillside Journal of Clinical Psychiatry 1987; 100-115

47. Ross CA, Miller SD, Reagor P, Bjornson L, Fraser G, Anderson G: Structured interview data on 102 cases of multiple personality disorder from four centers. Am J Psychiatry 1990; 147:596-601

48. Ross CA, Miller SD, Reagor P, Bjornson L, Fraser G, Anderson G: Schneiderian symptoms in multiple personality disorder and schizophrenia. Comprehensive Psychiatry 1990; 31: 111-118

49. Ellason JW, Ross CA: Positive and negative symptoms in dissociative identity disorder and schizophrenia: a comparative analysis. Journal of Nervous and Mental Disease 1995; 183: 236-241

50. Coons PM, Bowman ES, Milstein V: Multiple personality disorder: A Clinical investigation of 50 cases. J Nerv Ment Dis 1988; 176:519-527

51. Norton GR, Ross CA, Novotny MF: Factors that predict scores on the Dissociative Experiences Scale. J Clinical Psychology 1990; 46:273-277

52. Ross CA, Heber S, Norton R, Anderson G: Differences between multiple personality disorder and other diagnostic groups on structured interview. J Nerv Ment Dis 1989;177:487-491

53. Solomon RS, Solomon V: Differential diagnosis of multiple personality. Psychological Reports 1982; 51:1187-1194

54. Horevitz RP, Braun BG: Are multiple personalities borderline? Psychiatr Clin North Am 1984; 7:69-87

55. Buck OD: Single case study: Multiple personality as a borderline state. J Nerv Ment Dis 1983; 171:62-65

56. Benner DG, Joscelyne B: Multiple personality as a borderline disorder. J Nerv Ment Dis 1984; 172:98-104

57. Clary WF, Burstin KJ, Carpenter JS: Multiple personality and borderline personality disorder. Psychiatr Clin North Am 1984; 12:325-335

58. Ellason JW, Ross CA, Fuchs DL: Assessment of dissociative identity disorder with the Millon Clinical Multiaxial Inventory-II. Psychological Reports 1995; 76:895-905

59. Boon S, Draijer N: Multiple personality disorder in the Netherlands: a clinical investigation of 71 patients. Am J Psy 1993; 150:489-494

60. Sar V, Yargic LI, Tutkun H. Structured interview data on 35 cases of dissociative identity disorder in Turkey, Am J Psychiatry 1996; 153:1329-1333

61. Marmer SS: Multiple personality disorder: a psychoanalytic perspective. Psychiatr Clin North Am 1991; 14:677-693

62. Ulman RB., Brothers D: The Shattered Self: A psychoanalytic Study of Trauma. New Jersey, The Analytic Press, 1988.

63. Mollica RF, Caspi-Yavin Y: Measuring torture and torture-related symptoms. Psychological Assessment 1991; 3:581-587.

64. Roundtree T: The Effects of Shortterm Inpatient Treatment of Dissociative Identity Disorder. (master's thesis). Yellow Springs, Ohio, Antioch University, Clinical Psychology Department, 1997

65. Coons PM: Treatment progress in 20 patients with multiple personality disorder. J Nerv Ment Dis 1986; 174:715-721

66. Kluft RP: Treatment of multiple personality disorder. Psy Clin North Am 1984; 7:9-29

67. Ross CA, Heber S, Norton GR, Anderson G, Barchet P: The Dissociative DisordersInterview Schedule: A structured interview. Dissociation 1989; 2:169-189

68. Ellason JW, Ross CA, Mayran LW, Sainton K: Convergent validity of the new form of the DES. Dissociation 1994; 7:101-103

69. Bernstein EM, Putnam FW: Development, reliability, and validity of a dissociation scale. J Nerv Ment Dis 1986; 174:727-733

70. Wittchen HU, Semler G, von Zerssen D: A comparison of two diagnostic methods. Arch Gen Psychiatry 1985; 42:677-684

71. Spitzer RL, Williams JBW, Gibbon M, First MB: Users Guide for the Structured Clinical Interview for DSM-III-R.. Washington, DC, American Psychiatric Press, 1990.

72. Gallagher D, Breckenridge J, Steinmetz J, Thompson L: The Beck Depression Inventory and research diagnostic criteria: Congruence in an older population. J of Consulting and Clinical Psychology 1983; 51:945-946

73. Rehm LP, O'Hara MW: Item characteristics of the Hamilton Rating Scale for Depression. J Psychiatr Res 1985; 19:31-39

74. Carlson EB, Putnam FW, Ross CA, Torem M, Coons P, Dill DL, Loewenstein, RJ, & Braun, BG: Validity of the Dissociative Experiences Scale in screening for multiple personality disorder: A multicenter study. Am J Psychiatry 1993; 150:1030-1036

75. Ellason JW, Ross CA, & Fuchs DL: Lifetime Axis I and II Comorbidity and Childhood Trauma History in Dissociative Identity Disorder. Psychiatry 1996; 59:255-266.

76. Ellason JW, Ross CA: Millon Clinical Multiaxial Inventory-II follow-up of patients with dissociative identity disorder. Psychological Reports 1996; 78:707-716.

77. Putnam FW, Loewenstein RJ: Treatment of multiple personality disorder: a survey of current practices. Am J Psychiatry 1993; 150:1048-1052

78. Ganaway GK: Historical truth versus narrative truth: calrifying the role of exogenious trauma in the etiology of multiple personality and its variants. Dissociation 1989; 2:205-220

79. Loewenstein RJ: Diagnosis, epidemiology, clinical course, treatment, and cost effectiveness of treatment for dissociative disorders and multiple personality disorder: report submitted to the Clinton Administration Task Force on Health Care Financing Reform. Dissociation 1994; 7:3-11

Received Dec. 19, 1995; revisions received June 13 and October 28, 1996; accepted Jan. 30, 1997. From the University of Texas at Arlington and the Colin A. Ross Institute for Psychological Trauma, Richardson, Tx. Address reprint requests to Dr. Ross, 1701 Gateway, Suite 349, Richardson, Texas 75080.

The authors thank Stephanie Allred, M.S., for her assistance with statistical analysis.