Ten-Year Use of Mental Health Services by Patients With Borderline Personality Disorder and With Other Axis II Disorders

The methodology of this study has been described in detail elsewhere ( 3 ). The study was approved by the McLean Hospital Institutional Review Board. Written informed consent was obtained from each participant after careful explanation of what the study involved and before any study procedures were administered.

All participants were inpatients at McLean Hospital in Belmont, Massachusetts, who were admitted and interviewed between 1992 and 1995. Each patient was first screened to determine that he or she fulfilled several inclusion criteria. Participants had to be between the ages of 18 and 35; have a known or estimated IQ of 71 or higher; have no history or current symptoms of schizophrenia, schizoaffective disorder, bipolar I disorder, or an organic condition that could cause psychiatric symptoms; and be fluent in English.

Each patient met with a master's-level interviewer blind to the patient's clinical diagnoses for a thorough diagnostic assessment. Three semistructured diagnostic interviews were administered—the Structured Clinical Interview for DSM-III-R Axis I Disorders ( 4 ), the Revised Diagnostic Interview for Borderlines (DIB-R) ( 5 ), and the Diagnostic Interview for DSM-III-R Personality Disorders (DIPD-R) ( 6 ). The interrater and test-retest reliability of all three measures have been found to be good to excellent ( 7, 8 ).

Treatment history was assessed with the Background Information Schedule (BIS), a semistructured interview designed to assess the psychosocial functioning and treatment history of patients with borderline personality disorder and patients with other axis II disorders. The interrater and test-retest reliability and the concurrent validity of the BIS have been found to be good to excellent ( 9 ).

At each of five follow-up waves, separated by 24 months, use of psychiatric treatment was reassessed by staff members blind to baseline diagnoses using interview methods similar to the baseline procedures. After informed consent was obtained, the follow-up analog to the BIS was administered—the Revised Borderline Follow-Up Interview. Follow-up interrater reliability (within one generation of follow-up raters) and follow-up longitudinal reliability (from one generation of raters to the next) were also found to be good to excellent for this interview ( 1 ).

Generalized estimating equations, with diagnosis and time as main effects, were used in longitudinal analyses of prevalence data. Tests of diagnosis-by-time interactions were conducted. These analyses modeled the log prevalence, yielding an adjusted relative risk ratio (RRR) and 95% confidence interval for diagnosis and time. Gender was also included in these analyses as a covariate because patients in the group with borderline personality disorder were significantly more likely to be female than those in the comparison group. Alpha was set at .05, two-tailed.

The Kaplan-Meier product-limit estimator (of the survival function) was used to assess the time to cessation of treatment and time to resumption of treatment for the three treatments studied: individual therapy, standing medication, and psychiatric hospitalization. For each modality, we defined time to cessation of treatment as the time until the interview at which cessation of the treatment was first reported. Possible values for this outcome were two, four, six, eight, or ten years. For example, time to cessation for a person who was in individual therapy at baseline and who first reported stopping therapy at the two-year interview was two years (even though the patient may have resumed and then stopped therapy during subsequent intervals). We defined time to resumption of treatment in a somewhat different manner—the number of years after a person had stopped a treatment that he or she resumed that treatment. Thus times for this outcome were two, four, six, or eight years after the first report of cessation of a given treatment.

A total of 290 patients met both DIB-R and DSM-III-R criteria for borderline personality disorder; 72 met DSM-III-R criteria for at least one axis II disorder other than borderline personality disorder (and neither set of criteria for borderline personality disorder). Of these 72 comparison participants, three (4%) met DSM-III-R criteria for an odd cluster personality disorder, 24 (33%) met DSM-III-R criteria for an anxious cluster personality disorder, 13 (18%) met DSM-III-R criteria for a dramatic cluster personality disorder, and 38 (53%) met DSM-III-R criteria for personality disorder not otherwise specified (operationally defined in the DIPD-R as meeting all but one of the required number of criteria for at least two of the 13 axis II disorders described in DSM-III-R ).

Baseline demographic data have been reported elsewhere (3). Briefly, 279 participants (77%) were female, and 315 (87%) were white, 20 (6%) were African American, nine (3%) were Hispanic, eight (2%) were Asian, and ten (3%) were biracial or of other racial or ethnic backgrounds. The mean±SD age of the participants was 27.0±6.3 years, the mean socioeconomic status was 3.3±1.5) (range of 1 to 5, with 5 indicating the lowest status) (10), and the mean Global Assessment of Functioning score was 39.8±7.8 (indicating major impairment in several areas, such as work or school, family relations, judgment, thinking, or mood).

A total of 309 of the surviving patients (90%) were reinterviewed at all five follow-up waves. More specifically, 249 (92%) of the 271 surviving patients with borderline personality disorder and 60 (85%) of the 71 surviving patients in the comparison group were evaluated six times (baseline and five follow-ups).

Table 1 presents prevalence data on use of the three treatment modalities by the two patient groups over ten years. A significantly higher percentage of patients with borderline personality disorder reported taking standing medications over the ten-year period (relative risk ratio [RRR]=1.29, or 29% higher). When all patients were considered together, the proportion of patients taking psychotropic medications declined significantly over time, with a relative decrease from baseline of 21% ([1–.79]×100=21).

However, significant interactions between diagnosis and time were found for individual therapy and psychiatric hospitalization, and the explanation of these results is more complicated. At baseline 96% of patients with borderline personality disorder and 86% of those in the comparison group reported that they were in individual therapy. At the ten-year follow-up, rates had declined to 73% and 45%, respectively. For patients in the comparison group, the likelihood of participating in individual therapy decreased by 51% from baseline to ten years (RRR for diagnosis=.49; [1–.49]×100=51). The significant interaction between diagnosis and time indicates that the decline over ten years for patients with borderline personality disorder was 29% ([1–.49×1.44]×100=29). Thus the rate of decline was significantly smaller for patients with borderline personality disorder than for those in the comparison group (29% compared with 51%).

Among patients with borderline personality disorder, 79% reported having been hospitalized before the baseline interview, compared with 50% of those in the comparison group. At the ten-year follow-up, 29% and 3%, respectively, of the groups reported hospitalization in the prior two years. At baseline patients with borderline personality disorder were 65% more likely to report prior hospitalization (RRR=1.65). For patients in the comparison group, the likelihood of being hospitalized over the ten-year period decreased by 91% (RRR=.09). The significant interaction between diagnosis and time indicates that the ten-year decline for patients with borderline personality disorder was 71% ([1–.09×3.21)×100%=71). Thus the rate of decline was significantly smaller for patients with borderline personality disorder than for those in the comparison group (71% compared with 91%).

Figure 1 shows the cumulative percentage at each follow-up interview of patients with borderline personality disorder who reported a first cessation of treatment in each modality. Of those who reported receipt of individual therapy in the two years before baseline, 52% (N=84) had reported ceasing therapy at least once by ten years. Of those who were taking standing medication at baseline, 44% (N=89) had reported ceasing psychopharmacology at least once by ten years. All 290 patients with borderline personality disorder were hospitalized at baseline. At ten years, 88% (N=178) had experienced at least one two-year period without a psychiatric hospitalization. (Rates of treatment cessation cannot be determined using the numbers in this paragraph because of censoring—that is, participants lost to follow-up because of death or because they declined to participate.)

Figure 2 shows the time to resumption for each treatment among patients with borderline personality disorder after they reported a first cessation of that treatment. The ten-year data indicate that of those who reported cessation of individual therapy, 85% (N=62) resumed therapy during a subsequent follow-up period. Of those who reported having stopped taking standing medication, 67% (N=44) resumed during a subsequent follow-up period. In addition, 49% (N=59) of patients who had not been hospitalized for at least one of the two-year follow-up periods had been rehospitalized by the ten-year follow-up. (Rates of treatment resumption cannot be determined using the numbers in this paragraph because of censoring—that is, participants lost to follow-up because of death or because they declined to participate.)