Zanarini Rating Scale For Borderline Zanarini2003
Zanarini Rating Scale For Borderline Zanarini2003
Zanarini Rating Scale For Borderline Zanarini2003
233
234 ZANARINI
Characteristic
% Female 76.0
% Caucasian 84.0
Mean Age 33.6 (SD 11.1)
Mean SES 3.2 (SD 1.3)
Mean Years of Education 14.5 (SD 2.4)
Mean GAF 48.5 (SD 12.1)
% Individual Therapy 91.5
% Psychotropic Medication 72.5
% Psychiatric Hospitalization 47.0
Shea, and Paul Soloff. We also had a kitchen cabinet of borderline patients
review the ZAN-BPD to assess its face validity. Both groups of experts found
that the questions of the ZAN-BPD covered the DSM-IV criteria for BPD well.
They also found that the 5-point anchored rating scale for each criterion ad-
equately assessed the continuum of borderline psychopathology likely to
occur in a 1-week time frame.
METHOD
Patients were recruited by an ad in a local newspaper and posters placed
around the McLean Hospital campus. The ad and poster, which specified
that we were looking for men and women between the ages of 18 and 60
years, asked: Are you extremely moody? Do you often feel distrustful of oth-
ers? Do you frequently feel out of control? Are your relationships painful
and difficult?
The exclusion criteria for the study were: (a) patients who ever met the
DSM-IV criteria for schizophrenia, schizoaffective disorder, or bipolar I dis-
order; (b) patients who met criteria for a substance use disorder in the pre-
ceding month; and (c) patients who had no history of psychiatric treatment.
The first two exclusion criteria follow standard practice in borderline re-
search of excluding patients whose Axis I state (i.e., psychosis, mania, or in-
toxication/withdrawal) is likely to interfere with an assessment of their
more enduring personality traits or symptoms. The third exclusion criteria,
excluding patients without a psychiatric treatment history, was intended to
try to equalize severity of impairment between patients judged to have or not
to have met the DSM-IV criteria for BPD.
Potential patients were prescreened by telephone to determine if they met
any of our exclusion criteria. Those who were not excluded were invited to
participate in a face-to-face-interview. After written informed consent was
obtained, four semistructured interviews of demonstrated reliability were
administered to each patient: (a) the Background Information Schedule
(BIS), which assesses psychosocial functioning and treatment history
(Zanarini et al., 2001) (b) the Structured Clinical Interview for DSM-IV Axis I
Disorders (SCID I) (Williams et al., 1992); (c) the BPD module of the DIPD-IV;
ZANARINI RATING SCALE 237
Note. ZANBPD = Zanavini Rating Scale for Borderline Personality Disorder; SCLR90 = Symptom Check-
list 90. aBoth ZANBPD and SCL90 pertain to past week.
and (d) the Revised Diagnostic Interview for Borderlines (DIB-R) (Zanarini,
Gunderson, Frankenburg, & Chauncey, 1989).
A second rater, blind to all information concerning that patient, then ad-
ministered the ZAN-BPD, inquiring about symptom severity during the past
week. The Symptom Checklist 90 (SCL-90) (Derogatis, Lipman, & Covi,
1973), which is a well-known and widely used self-report measure of change
in general psychopathology, was also administered at this time.
In addition, three separate substudies were undertaken to assess the
interrater reliability of the ZAN-BPD, the same-day test-retest reliability of
the ZAN-BPD, and the 1-week sensitivity to change of the ZAN-BPD. In the
first of these substudies, a second blind rater observed and independently
scored the ZAN-BPD interview the first blind rater conducted. A total of 32
conjoint interviews of this type were conducted. In the second of these
substudies, the patient received a second administration of the ZAN-BPD a
number of hours after the first administration. A blind interviewer was also
used for these 40 interviews. In the third of these substudies, 41 patients re-
turned 7 to 10 days after their first interview for a second administration of
the ZAN-BPD. Again, a blind rater assessed the patient. The SCL-90 was
also readminstered at this time.
Correlations between the continuous scores of the ZAN-BPD and the con-
tinuous scores of the SCL-90 and the DIB-R were assessed using
Spearmans . Between-group differences on the various scales of the
ZAN-BPD were assessed using Students t-test. The internal consistency of
the ZAN-BPD was assessed using Cronbachs . The standard error esti-
mate for the statistic was obtained using bootstrapping methods. Both
interrater and test-retest reliability were assessed using intraclass correla-
tion coefficients (ICCs), which correct for chance levels of agreement.
RESULTS
In all, 283 patients were screened by telephone. A total of 41 patients were
excluded from further participation after reporting that they had a
long-standing clinical diagnosis of a psychotic/bipolar I disorder, whereas
238 ZANARINI
Note. ZANBPD = Zanavini Rating Scale for Borderline Personality Disorder; DIBR = Revised Diagnostice
Interview for Borderlines. aZANBPD pertains to past week and DIBR to past 2 years.
symptoms of BPD. In contrast, three scales of the SCL-90 (i.e., anger, anxi-
ety, and depression) were aggregated to correlate with the affective sector
score of the ZAN-BPD.
Table 4 details the convergent validity of the ZAN-BPD with the relevant
scores of the DIB-R, which has a 2-year time frame. Although all the correla-
tions are highly significant, they are substantially lower than those found
for the SCL-90. This is probably due to the vastly different time frames of the
ZAN-BPD and the DIB-R (1 week vs. 2 years).
Discriminant validity findings are presented in Table 5. All comparisons
between those meeting the DSM-IV criteria for BPD and those not meeting
this criteria set were found to be highly significant. This was so regardless of
which of the 14 ZAN-BPD scores were being compared; those pertaining to
the nine criteria, the four sectors, or the total BPD psychopathology score.
The internal consistency of the nine criteria scores of the ZAN-BPD was
found to be high (Cronbachs = 0.85, SE = 0.01, 95% CI = 0.82 to 0.89).
This means that scores on the nine scales were strongly and consistently re-
lated to one another. This also suggests that these scores appear to be as-
sessing the same construct, the DSM-IV BPD.
Table 6 details the interrater and test-retest reliability of the ZAN-BPD.
Using the criteria of Fleiss (1981), correlations below .40 are considered to
be poor, correlations between .40 and .75 are considered to be fair to good,
and correlations higher than .75 are considered to be excellent. Using these
standards, all but one of the interrater reliability figures were in the excel-
lent range. The criterion pertaining to self-destructive efforts, which were
extremely rare, was in the fair to good range. Again using the Fleiss guide-
lines, all but two of the test-retest figures were in the excellent range. Both of
these criteria (i.e., affective instability and frantic efforts to avoid abandon-
ment) were in the fair to good range.
240 ZANARINI
The ability to detect change reliably is one of the most important proper-
ties of a continuous measure of change of severity of psychopathology. Of
the 41 patients who were reinterviewed 7 to 10 days after their initial inter-
view, only 4 or 9.8% had exactly the same score on the ZAN-BPD at both ad-
ministrations. In contrast, 51.2% (N = 21) of these patients were less
symptomatic at the time of the second administration and 39.0% (N = 16)
were more symptomatic. Of the 90% (N = 37) of patients who reported a
change on the ZAN-BPD at the second administration, 35% (N = 13) were
judged to have experienced a 1- to 2-point difference, 30% (N = 11) a 3- to
4-point difference, and 35% (N = 13) a 5- to 15-point difference.
Table 7 shows data related to another way of assessing change. It exam-
ines the correlations between the difference scores of the interview-based
ZAN-BPD and the relevant scales of the self-report SCL-90, both of which
were readministered 7 to 10 days after their initial administration. As can be
seen, all of the correlations between the difference scores (baseline value mi-
nus value at second administration) of these measures were statistically
significant. This shows that the ratings on these two quite different mea-
sures of borderline psychopathology were changing in similar directions
and in somewhat similar amplitudes over time.
DISCUSSION
Three important issues need to be addressed in the development of a new
continuous measure of psychopathology that will be administered multiple
times during the course of a treatment study. These issues are validity, reli-
ability, and sensitivity to change. The ZAN-BPD has been shown to have both
convergent validity and discriminant validity. In terms of convergent validity,
ZANARINI RATING SCALE 241
Note. ZANBPD = Zanavini Rating Scale for Borderline Personality Disorder; SCL90 = Symptom Checklist
90.
the scores on the various scales of the ZAN-BPD were highly correlated with
those obtained from both a self-report measure of general psychopathology
and an interview developed specifically to assess a somewhat different con-
ceptualization of BPD from that contained in the DSM-IV. In terms of
discriminant validity, the scores of the ZAN-BPD significantly discriminated
borderline patients from patients with a variety of serious Axis I disorders.
Reliability is an important concept in the development of any measure of
psychopathology. The ZAN-BPD was found to have high internal consis-
tency, an important component of reliability. In terms of forms of reliability
involving multiple raters, interrater reliability is a test of whether different
raters understand and score the same patient material in a similar manner.
Test-retest reliability is more complicated and depends on consistency of
patient self-report and interviewer differences in eliciting, understanding,
and scoring clinical material. High levels of both interrater and test-retest
reliability were obtained in the current study for the ZAN-BPD.
One cautionary comment is worth noting. Zimmerman (1994) has found
that reliability levels for diagnostic interviews for Axis II disorders are typi-
cally higher when reliability studies are conducted by the developer of that
measure and his or her colleagues. This suggests that the reliability of the
ZAN-BPD might well be somewhat lower (but still acceptable) when used by
other investigators.
The ZAN-BPD was developed for use in treatment studies involving bor-
derline patients. It hopefully will be equally useful whether the treatment in
question is psychosocial or pharmacologic in nature. However, its useful-
ness will depend on whether it can accurately assess symptomatic change
over time. The results of this study suggest that it has this capability. More
specifically, difference scores on the ZAN-BPD were significantly correlated
with difference scores on the BPD-related scales of a self-report measure of
general psychopathology (the SCL-90), which is often used in treatment
studies of borderline patients because of its validity as a measure of change.
A limitation of the current study is that we did not assess the full array of
Axis II disorders due to time constraints. Our clinical impression was that
the nonborderline patients did not have a substantial amount of Axis II
comorbidity. However, because we did not complete a full DIPD-IV assess-
ment, we cannot be sure of this. Another limitation is that all of the patients
in this study had a treatment history. Whether the ZAN-BPD would perform
as well in a community-based sample is an open question.
242 ZANARINI
REFERENCES
American Psychiatric Association. (1994). Di- Young, R. C., Biggs, J. T., Ziegler, V. E., &
agnostic and statistical manual of men- Meyer, D. A. (1978). A rating scale for
tal disorders (4th ed.). Washington, mania: Reliability, validity, and sensi-
DC: American Psychiatric Association. tivity. British Journal of Psychiatry,
Bender, D. S., Dolan, R. T., Skodal, A. E., 133, 429-435.
S a n i s l o w , C . A . , Dyc k , I. R . , Zanarini, M. C., & Frankenburg, F. R. (2001).
McGlashan, T. H., Shea, M. T., Attainment and maintenance of reli-
Zanarini, M. C., Oldham, J. M., & ability of axis I and II disorders over the
Gunderson, J. G. (2001). Treatment course of a longitudinal study. Compre-
utilization by patients with personality hensive Psychiatry, 42, 369-374.
disorders. American Journal of Psychia- Zanarini, M. C., Frankenburg, F. R.,
try, 158, 295-302. Chauncey, D. L., & Gunderson, J. G.
Derogatis, L, R., Lipman, R. S., & Covi, L. (1987). The diagnostic interview for
(1973). SCL-90: An outpatient psychiat- personality disorders: Interrater and
ric rating scale: Preliminary report. test-retest reliability. Comprehensive
Psychopharmacology Bulletin, 9, 13-28. Psychiatry, 28, 467-480.
Fleiss, J. L. (1981). Statistical methods for Zanarini, M. C., Frankenburg, F. R., Khera,
rates and proportions (2nd ed., p. 218). G. S., & Bleichmar, J. (2001). Treat-
New York: Wiley. ment histories of borderline inpatients.
Hurt, S. W., Hyler, S. E., Frances, A., Clarkin, C o m p r e h e n s i v e P s y c h i a tr y , 4 2 ,
J. F., & Brent, R. (1984). Assessing bor- 144-150.
derline personality disorder with Z a n a r i n i , M. C . , G u n d e r s o n , J . G . ,
self-report, clinical interview, or Frankenburg, F. R., & Chauncey, D. L.
semistructured interview. American (1989). The revised diagnostic inter-
Journal of Psychiatry, 141, 1228-1231. view for borderlines: Discriminating
Skodol, A. E., Gunderson, J. G., McGlashan, BPD from other axis II disorders. Jour-
T. H., Dyck, I. R., Stout, R. L., Bender, nal of Personality Disorders, 3, 10-18.
D. S., Grilo, C. M., Shea, M. T., Zanarini, M. C., Gunderson, J. G.,
Zanarini, M. C., Morey, L. C., et al. Frankenburg, F. R., & Chauncey, D. L.
(2002). Functional impairment in (1990). Discriminating borderline per-
schizotypal, borderline, avoidant, and sonality disorder from other Axis II dis-
obsessive-compulsive personality dis- orders. American Journal of Psychiatry,
orders. American Journal of Psychia- 147, 161-167.
try,159, 276-283. Zanarini, M. C., Skodol, A. E., Bender, D.,
Swartz, M., Blazer, D., George, L., & Winfield, Dolan, R., Sanislow, C., Schaefer, E.,
I. (1990). Estimating the prevalence of Morey, L. C., Grilo, C. M., Shea, M. T.,
borderline personality disorder in the McGlashan, T. H., & Gunderson, J. G.
community. Journal of Personality Dis- (2000). The collaborative longitudinal
orders, 4, 257-272. personality disorders study: II. Reliabil-
Williams, J. B. W., Gibbon, M., First, M. B., ity of axis I and II diagnoses. Journal of
Spitzer, R. L., Davies, M., Borus, J., Personality Disorders, 14, 291-299.
Howes, M., Kane, J., Pope, H. G., Zimmerman, M. (1994). Diagnosing person-
Rounsaville, B., & Wittchen, H. (1992). ality disorders: A review of issues and
The structured clinical interview for research methods. Archives of General
DSM-III-R (SCID). II. Multi-site test-re- Psychiatry, 51, 225-245.
test reliability. Archives of General Psy-
chiatry, 49, 624-629.