Treatment Rates for Patients With Borderline Personality Disorder and Other Personality Disorders: A 16-Year Study

As noted above, the study is part of the McLean Study of Adult Development, a multifaceted longitudinal study of the course of borderline personality disorder. The methodology, which was reviewed and approved by the McLean Hospital Institutional Review Board, has been described in detail elsewhere (9). Briefly, all participants were initially inpatients at McLean Hospital in Belmont, Massachusetts. Each patient was screened to determine that he or she was between the ages of 18 and 35; had a known or estimated IQ of ≥71; had no history or current symptoms of schizophrenia, schizoaffective disorder, or bipolar I disorder or an organic condition that could cause serious psychiatric symptoms (for example, lupus and multiple sclerosis); and was fluent in English.

After the study procedures were explained, written informed consent was obtained. Each patient then met with a masters-level interviewer who was blind to the patient’s clinical diagnoses for a thorough treatment history and a diagnostic assessment. Four semistructured interviews were administered: the Background Information Schedule, which assesses lifetime psychiatric treatment history (5); the Structured Clinical Interview for DSM-III-R Axis I Disorders (10); the Revised Diagnostic Interview for Borderlines (11); and the Diagnostic Interview for Personality Disorders (12). The interrater reliability of the Background Information Schedule was carefully assessed in a sample of 45 patients with personality disorders and was found to be good to excellent (5). As a measure of validity, we compared self-report of treatment history according to this interview with the medical records of 15 patients who had received all of their psychiatric care at McLean Hospital. Convergent validity was also found to be good to excellent (5). In addition, the interrater and test-retest reliability of all three diagnostic measures have been found to be good to excellent (13,14).

The psychiatric treatments used by study participants over the years of follow-up were assessed with the treatment section of the Revised Borderline Follow-up Interview (15)—the follow-up analog to the Background Information Schedule. This measure, as well as our diagnostic battery, was readministered every two years over 16 years of prospective follow-up by raters blind to previously collected information. The follow-up interrater reliability (within one generation of follow-up raters) and follow-up longitudinal reliability (from one generation of raters to the next) of these four interviews have also been found to be good to excellent (7,13,14).

A vast majority of our follow-up interviews were conducted within several months of the date of each participant’s scheduled interview. However, two participants who were unavailable for interview at the 12- and 14-year waves of data collection provided six years of data at the 16-year follow-up. A third participant who was unavailable for interview at the eight-, ten-, 12-, and 14-year waves of data collection provided ten years of data at the 16-year follow-up. Overall, eight of 2,881 interviews (or .3%) assessed a longer time period than our typical two years.

Data on psychiatric treatment were assembled in panel format (that is, multiple records per patient, with one record for each follow-up period for which data were available). Generalized estimating equations, appropriately accounting for repeated measures for the same patients, were used to fit log-linear regression models assessing the association between diagnostic group and the prevalence of treatment use over time. Specifically, these analyses modeled the log prevalence as a piecewise-linear function of time, with separate slopes for the change from baseline to eight-year follow-up and for the corresponding change from eight- to 16-year follow-up; the models also included the effect of diagnostic group. Preliminary tests of diagnostic group × time interactions were also conducted to assess whether the pattern of change in prevalence differed by diagnostic group. Because there was no evidence of interaction, main effects of diagnostic group and time are reported; results of these analyses yielded an adjusted relative risk ratio and 95% confidence interval for diagnostic group and the two time trends. Given the large number of comparisons for the 16 treatment modalities, we applied the Bonferroni correction for multiple comparisons to the analysis of each treatment modality, resulting in a corrected alpha level of .0031 (.05/16).

The sample and its diagnostic characteristics have been described before (9). A total of 290 patients met DSM-III-R criteria for borderline personality disorder as well as the criteria in the Revised Diagnostic Interview for Borderlines. A total of 72 patients met DSM-III-R criteria for at least one nonborderline axis II disorder (and met neither criteria set for borderline personality disorder). Of these 72 comparison participants, three (4%) met DSM-III-R criteria for a personality disorder in the odd cluster, 24 (33%) met DSM-III-R criteria for a personality disorder in the anxious cluster, 13 (18%) met DSM-III-R criteria for a personality disorder in the nonborderline dramatic cluster, and 38 (53%) met DSM-III-R criteria for a personality disorder not otherwise specified (which was operationally defined in the Revised Diagnostic Interview for Personality Disorders 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 also been reported before (9). Briefly, 279 (77%) of the 362 participants were female. A total of 315 (87%) were white, 20 (6%) were African American, nine (3%) were Hispanic, eight (2%) were Asian, and ten (3%) were biracial. The mean±SD age of the participants was 27.0±6.3. The mean score on a measure of socioeconomic status was 3.3±1.5 (possible scores range from 1 to 5, with higher scores indicating lower status). 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.

In terms of continuing participation in the study, which has also been described before (16), 231 (88%) of 264 of surviving patients with borderline personality disorder were reinterviewed at all eight follow-up waves (13 died by suicide and 13 died of other causes). A similar rate of participation was found for the comparison participants; 58 (83%) of 70 surviving patients in this group were reassessed at all eight follow-up waves (one died by suicide and one died of other causes).

Table 1 summarizes data on rates of outpatient treatments and more intensive treatments, such as psychiatric hospitalizations, reported by participants in the two groups over 16 years of prospective follow-up. Patients with borderline personality disorder were not significantly more likely than those with other axis II disorders to report being in individual therapy, intensive psychotherapy, or couples or family therapy. However, they were significantly more likely to report being in group therapy and self-help groups. They were also significantly more likely to report taking any standing medication, all forms of polypharmacy (from taking two to five standing medications), and all forms of more intensive treatment studied except electroconvulsive therapy (ECT) (day treatment, residential treatment, any psychiatric hospitalizations, multiple hospitalizations, and hospitalizations of ≥30 days).

During the first eight years of follow-up, reported rates of all forms of treatment were found to decline significantly for those in both study groups, except for taking four or more or five or more medications concurrently and ECT. Conversely, during the most recent eight years of follow-up (year 8 to year 16), reported rates of almost all forms of treatment remained relatively flat or stable over time for both groups; the only exception was ≥30 days of psychiatric hospitalization; this rate continued to decline significantly for both groups.

Because these results are complicated, we present two examples from Table 1 so that they can be better understood. As noted above, the rates of reported individual therapy were not significantly different between the groups over time. However, a significant decline in the rates during the first eight years of follow-up was noted, which was the same for both groups; specifically, there was a 29% decline ([1–.71] × 100%) in the rates from baseline to year 8 in both groups. Thereafter, during years 8–16, no significant change was noted in rates for either group. As also noted above, patients with borderline personality disorder reported significantly higher rates (2.3 times higher) of ≥30 days of psychiatric hospitalization than participants in the comparison group. However, the significant declines in the reported rates were the same for those in both groups ([1–.14] × 100%=86% decline in years 0–8, followed by [1–.39] × 100%=61% decline in years 8–16).

Three main findings emerged from this study. The first is that compared with patients with other axis II disorders, patients with borderline personality disorder reported significantly higher rates of use of 12 of the 16 treatment modalities studied—all but individual therapy, intensive psychotherapy, couples or family therapy, and ECT. Looked at another way, all forms of pharmacotherapy and more intensive treatment studied (for example, psychiatric hospitalizations and day treatment) were reported by a significantly larger proportion of patients with borderline personality disorder over time, compared with patients who had other axis II disorders. This finding is consistent with participants’ treatment history at study entry (5) and over six years of prospective follow-up (7). It also highlights the consistent severity of the psychopathology of borderline personality disorder compared with that of other personality disorders (17). In addition, it is consistent with the higher rates of co-occurring disorders reported over time by patients with borderline personality disorder compared with participants with other axis II disorders (18,19).

The second main finding is that rates of participation in 13 treatment modalities declined significantly over the first eight years of follow-up for participants in both groups. Only the rates of four or five concurrent medications and ECT remained stable over these eight years. This finding too is consistent with the results we found at the six-year follow-up (7).

The third main finding is that the rates of participation in 15 of 16 treatment modalities did not continue to decline significantly over the second eight years of follow-up for those in either study group. Only the rate of psychiatric hospitalizations of ≥30 days declined significantly from year 8 to year 16 of follow-up for both groups. This is a new finding and an important one with public health significance. It suggests that the cost of treating patients with borderline personality disorder declines in the short and medium terms, but is relatively stable in the longer term. For example, 97% of all participants reported participating in individual therapy at study entry, and this rate declined significantly to 73% at the eight-year follow-up. However, this 73% declined only to 65% at the 16-year follow-up. The same pattern was found, for example, for any standing medication. More specifically, 84% of patients with borderline personality disorder reported taking a standing medication at study entry, and this rate declined significantly to 71% over the first eight years of follow-up. However, the rate remained a steady 71% over the second eight years of follow-up. As a third example, 79% of patients with borderline personality disorder had a history of prior hospitalizations at baseline, and the rate of hospitalizations declined significantly to 28% at the eight-year follow-up. However, this 28% rate declined only to 24% by the time of the 16-year follow-up.

Although rates for these three major treatment modalities declined substantially over time for patients with borderline personality disorder, the fact that they barely declined in the second eight years of follow-up suggests that these may be chronic rates of treatment going forward. The good news in terms of cost is the relatively low rate of psychiatric hospitalizations at the 16-year follow-up among patients with borderline personality disorder (24%). In a similar vein, only 15% of patients with borderline personality disorder had multiple hospitalizations, while only 4% spent ≥30 days in a psychiatric inpatient unit.

However, the question remains whether these rates will remain stable going forward or whether there is anything clinicians or the health care system can do or would want to do to lower these rates further. Among patients with borderline personality disorder at the 16-year follow-up, 65% were receiving individual therapy and 71% were receiving standing medication (71%); these interventions may be helping these patients stay out of the hospital. The rate of more costly forms of these outpatient modalities had also dropped. More specifically, the rate of intensive individual therapy dropped from 36% to 13%, and the rate of aggressive polypharmacy (three or more concurrent medications) dropped from 46% to 36%.

However, the relatively large percentages of patients who reported that they were using individual therapy and standing medications over the second eight years of follow-up is striking, particularly given the high rates of sustained remissions found over the 16 years (16). For example, 99% of the patients with borderline personality disorder reported a two-year remission, 95% reported a four-year remission, and 90% reported a six-year remission (16). It may be that these patients were dealing with residual symptoms of borderline personality disorder, particularly the less dramatic or temperamental symptoms (for example, anxiety, abandonment concerns, and undue dependency) that have been found to resolve more slowly than the acute symptoms of this disorder (for example, self-harm, suicide attempts, and quasi-psychotic thought) (17). It may also be that these patients were being treated for axis I disorders that had never remitted or that had recurred (18).

The main limitation of this study was that all the patients with borderline personality disorder were severely ill inpatients at the time of study entry. Rates of treatment use might be substantially lower for individuals with borderline personality disorder who have never been hospitalized or in psychiatric treatment. This limitation also applies to the axis II comparison participants. In addition, reported rates of many forms of more intensive treatments (for example, residential treatment and ≥30 days of psychiatric hospitalization) were very low for axis II comparison participants during the most recent eight years of follow-up; consequently, because of the sparseness of data, extra caution is required when interpreting results from our regression analyses for these particular treatments.

It is worth noting that most study participants were treated in the community. Psychotherapy was primarily provided by hundreds of community-based psychologists and social workers located throughout the United States. In addition, these therapies were mostly supportive in nature, and almost none were empirically based (20–24). Medications were prescribed by psychiatrists in general practice or, increasingly, by primary care physicians.

The study was supported by grants MH47588 and MH62169 from the National Institute of Mental Health.

The authors report no competing interests.