Content uploaded by Perry Hoffman
Author content
All content in this area was uploaded by Perry Hoffman on Oct 19, 2017
Content may be subject to copyright.
Family Connections: A Program for Relatives of
Persons With Borderline Personality Disorder
PERRYD.HOFFMAN,PH.D.w
ALAN E. FRUZZETTI, PH.D.z
ELLIE BUTEAU, PH.D.‰
EMILY R. N EIDITCH#
DIXIANNE PENNEY, DR. P.H.ww
MARTHA L. BRUCE, PH.D. , MPHzz
FREDERIC HELLMAN ‰‰
ELMER STRUENING, PH.D. ##
This study assessed changes in family members who participated in Family Con-
nections, a 12-week manualized education program for relatives of persons with bor-
derline personality disorder (BPD). Family Connections, led by trained family
members, is based on the strategies of standard Dialectical Behavior Therapy (DBT)
and DBT for families. The program provides (a) current information and research on
BPD, (b) coping skills, (c) family skills, and (d) opportunities to build a support net-
work for family members. Forty-four participants representing 34 families completed
the pre-, post-, and 6-month postbaseline self-report questionnaires. Analyses employ-
ing hierarchical linear modeling strategies showed significant reductions in grief and
burden, and a significant increase in mastery from pre- to post-group assessment.
Changes were maintained at 6 months post baseline.
Fam Proc 44:217–225, 2005
Family Process, Vol. 44, No. 2, 2005 rFPI, Inc.
217
Special thanks to the National Institute of Mental Health for support for this study (NIMH Grant K08
MH001602-05).
wAssistant Clinical Professor, The Mount Sinai School of Medicine, New York, NY; President, National
Education Alliance for Borderline Personality Disorder, Rye, NY.
zAssociate Professor of Psychology, University of Nevada, Reno.
‰Instructor, Tufts University.
#Research Assistant, Weill Medical College of Cornell University, White Plains, NY.
wwExecutive Vice President, National Education Alliance for Borderline Personality Disorder (NEA-BPD).
zzProfessor, Weill Medical College of Cornell University, White Plains, NY.
‰‰Program Evaluation Specialist II, Epidemiology for Mental Disorders Research Department, New York
State Psychiatric Institute.
##Director, Epidemiology of Mental Disorders Research Department, New York State Psychiatric
Institute.
INTRODUCTION
The diagnosis of borderline personality disorder (BPD) reveals a complex, confus-
ing, and challenging disorder not only for patients and mental health professionals,
but also for family members of those with BPD. Not surprisingly, the behaviors that
commonly accompany BPD, such as suicide attempts, intense anger, and self-injury,
create stressful situations for persons with BPD and their family members, who fre-
quently report being overwhelmed by the chaos that results. Unlike other psychiatric
disorders, for which services exist to support patient and family member well-being,
the families of BPD persons are frequently neglected (Glick & Loraas, 2001; Harman
& Walso, 2001; Hoffman, Struening, Buteau, Hellman, & Neiditch, 2005). Although
family members of those with BPD often experience burden, depression, loss, grief,
and other kinds of distress (Berkowitz & Gunderson, 2002; Hoffman & Hooley, 1998;
Hoffman et al., 2005), no standardized programs to alleviate the impact of this dis-
order on relatives have been evaluated.
In striking contrast, family programs for relatives of persons with psychiatric dis-
orders other than BPD, when available, have taken a valued role in treatment settings
(Dixon et al., 2001). Although not yet widely disseminated enough, several modalities
of services are available for relatives of persons with schizophrenia, bipolar disorder,
and major depression, for example. Specifically, two categories of family programs
exist: family psychoeducation and family education (Hoffman & Fruzzetti, 2005;
McFarlane et al., 1995). In the former, the professional family psychoeducation model,
mental health professionals lead educational support groups that include both the
family members and patients. Complementing these professional services is the family
education model, in which trained family members lead educational support groups
consisting only of family members. Each modality is supported by data that demon-
strate improvements in the patient, the participating relative, or both. In family
psychoeducation, programs have been shown, often in combination with medication,
to effect a reduction in patient relapse and rehospitalization, and an increase in pa-
tient well-being (McFarlane et al.). Family education groups conducted solely for the
relatives and not directly targeting patient change have demonstrated significant
improvements in family member well-being (Dixon et al., 2001; Falloon & Pederson,
1985).
The study reported here evaluates Family Connections (FC), a 12-week multiple-
family education program modeled structurally after the National Alliance for the
Mentally Ill’s (NAMI) highly valued Family-to-Family Program (FFP). Dixon and
colleagues’ (2001) FFP assessment timeline model with three time periodsFpre-,
post-, and 6 months post baselineFwas followed. Designed to meet the needs of rel-
atives of persons with borderline personality disorder and its related problems such as
emotion dysregulation, self-injury, mood lability, and relationship difficulties, Family
Connections is led by trained family members who, after completing the required
training course, then conduct groups in their own areas. Training has been stand-
ardized to follow the FC program manual closely (Fruzzetti & Hoffman, 2004), and
includes (a) initial participation in an FC program as group members; (b) approxi-
mately 20 hours of formal training to lead an FC group (e.g., lecture, discussion, role
playing); and (c) weekly consultation after the leaders start a new group on their own.
Leaders serve in a teaching and mentoringcapacityfortheirgroupmembers,pro-
viding the course curriculum in a nonjudgmental, supportive environment. They also
FAMILY PROCESS218 /
www.FamilyProcess.org
model a no-blame approach, providing examples of skill application from their own
real-life experiences.
To help determine the suitability of individuals as leaders, potential leaders par-
ticipate in an in-person interview (or a phone interview if necessary), followed by
attendance at a training weekend. Several factors contribute to selection as FC
leaders: (a) active participation in the formal leaders’ training; (b) ability to role play
being a group leader and to provide competent psychoeducational instruction;
(c) ability to role play competent group leadership and group management, using
a nonblaming and nonpunitive style; (d) ability to role play with a coleader and work
together to manage a group effectively; (e) willingness to receive consultation and
supervision (including either being observed or having an audiotape available for
review for all groups); and (f) interest and enthusiasm about leading an FC group. A
final decision concerning appropriateness to colead a series is ideally a mutual decision
made by the trainers, in discussion with the potential leaders, at the completion of the
training course.
The content of the FC program, developed in consultation with several
family members and individuals with BPD, was adapted largely from existing
curricula created by the first two authors. The content includes psychoeducational
materials reflecting current literature on BPD and on family functioning,
and some skills adapted from individual Dialectical Behavior Therapy (DBT; Linehan,
1993). It also includes relationship and family skills developed by the authors, based
on DBT theory (e.g., Hoffman, Fruzzetti, & Swenson, 1999; Fruzzetti & Fruzzetti,
2003).
The FC program (Fruzzetti & Hoffman, 2004) is divided into six modules: Module 1:
the most current information and research on BPD (e.g., symptoms, course of illness);
Module 2: psychoeducation regarding the development of BPD, available treatments,
comorbidity, and a primer on emotion reactivity and dysregulation; Module 3: indi-
vidual skills and relationship skills to promote participant emotional well-being (in-
cluding emotion self-management, mindfulness, letting go of judgments, decreasing
vulnerability to negative emotions, and skills to decrease emotional reactivity);
Module 4: family skills to improve the quality of family relationships and interactions
(letting go of blame and anger, acceptance skills in relationships); Module 5: accurate
and effective self-expression (how to validate); and Module 6: problem management
skills (e.g., defining problems effectively, collaborative problem solving, knowing
when to focus on acceptance and when to focus on change). All modules include spe-
cific practice exercises and homework. In addition, throughout the program, FC
provides a forum in which participants can build a support network.
Program fidelity was ensured by periodic visits to sites by the first author, and
weekly hour phone supervision with each pair of group leaders to review the curric-
ulum presented the previous week and the class homework practice exercises com-
pleted. The FC program manual allowed the authors to evaluate the ability of group
leaders to deliver the program effectively and in a manner consistent with content in
the manual.
The FC program targeted change in consistently problematic constructs for family
members: burden, depression, grief, and mastery (Family Perspectives on Borderline
Personality Disorder Conference, 2002; Hoffman, Buteau, Hooley, Fruzzetti, & Bruce,
2003). It was hypothesized that (a) there would be a decrease in burden, grief, and
depression from pre- to postparticipation in FC; (b) there would be an increase in
HOFFMAN ET AL. / 219
Fam. Proc., Vol. 44, June, 2005
mastery from pre- to postparticipation in FC; and (c) the changes would be maintained
at 6 months post baseline.
The constructs for the study, as defined by their respective measurements, were (a)
burden: family member reported stressors due to the ill relative’s symptomatology and
behavior, both on other relationships and interfering in daily activities; (b) depression:
levels of depressive symptomatology that the family member experienced during the
past week, such as sleep restlessness, being bothered by things, and feeling hopeful
about the future; (c) grief: the cognitive, emotional, and psychological experiences
such as sadness, pain, and loss associated with having a relative with a mental illness;
and (d) mastery: the identification of self-management skills to cope with having
a relative with a mental illness.
METHODS
Participants learned about the 12-week program from various sources: their rela-
tive’s therapist, regional NAMI offices, facilities treating patients with borderline
personality disorder, and the Web site of the National Education Alliance for Bor-
derline Personality Disorder (http://www.borderlinepersonalitydiso.com) (Family Per-
spectives on Borderline Personality Disorder, 2003). Fifty participants enrolled, and 44
participants representing 34 families completed the Family Connections program at
three sites. Recruitment was conducted with comparative ease. In fact, after the
completion of the first series in any given location, each site accrued a waiting list. All
potential participants were willing to participate in research and actually stated ap-
preciation both for the development and research of a program to assist families. Each
person provided written informed consent prior to the initial assessment.
Participants
Thirty-nine of the 44 participants were parents (27 mothers [61.4%] and 12 fathers
[27.3%]), 4 were spouses/partners (2 husbands, 1 wife, 1 partner [9.1%]), and 1 was a
sibling [2.3%]). Mean age of group members was 55.5 (SD ¼10.0); mean age of the
BPD family member patients was 25.1(SD ¼9.8); and the mean number of years that
the relative had suffered from BPD was 7.7 (SD ¼7.3). Ninety-one percent of partic-
ipants reported a yearly income level at $50,000 or above. Group attendance was high,
and on average, participants missed only two of the 12 meetings (present for 83%).
The dropout rate from the program was 12%. A variety of reasons, such as the un-
expected death of a parent or discomfort with the presence of an ex-spouse, were re-
ported as causes. No participants reported dropping out because of dissatisfaction
with the FC program. Overall, 80% of participants completed the follow-up assess-
ment. A few participants were unreachable at 6-month postbaseline follow-up (n¼3),
and the remaining (n¼3) nonresponders reported that they did not have time to
complete the assessment packet.
Procedure
Before beginning the FC program, 2 weeks after program completion, and 6 months
post baseline, family members completed a research packet consisting of several scales
and demographic questions. To provide an opportunity to compare our research with
research on families with relatives with other mental illnesses, the assessment in-
cluded measures of depression, burden, grief, and mastery scales, all used in two
FAMILY PROCESS220 /
www.FamilyProcess.org
previous major family studies: the Family Impact Study (Struening et al., 1995) and
the NAMI Family–to-Family Study (Dixon et al., 2001).
Specifically included were the Burden Assessment Scale (Reinhard, Gubman,
Horwitz, & Minksy, 1994), which assesses levels of objective and subjective burden,
and the Perceived Burden Scale (Struening et al., 1995), which assesses interpersonal
burden and role strain. Participants also completed the Revised Center for
Epidemiologic Studies Depression Scale (CES-D; Struening et al.), the Grief Scale
(Struening et al.), which focuses on the respondent’s current feelings of grief, and the
Mastery Scale (Dixon et al., 2001), which rates perceptions of coping. The Cronbach’s
alphas of the four scales described ranged from .75 to .92. All measures were repeated
at posttest and at 6 months post baseline.
RESULTS
Hierarchical linear modeling (HLM)
1
was used to assess change from pre- to
posttest, and from posttest to 6 months post baseline. In the context of this study,
HLM was a more appropriate analysis than traditional repeated measures because the
data structure includes two levels of nesting: time points within family members and
family members within families. In addition, the current analysis includes 20%
missing data at the 6-month postbaseline assessment, and additional data missing at
random at the first two time points. Repeated measures analysis of variance (ANOVA)
procedures delete participants who do not have data at all time points involved in the
analysis, while HLM includes these participants in the analysis, thereby increasing
the power of the analysis. Therefore, we havethreetimepointsnestedwithineachof
44 family members, who are in turn nested within 34 families; this constitutes a three-
level model with time (level 1) nested within family members (level 2) nested within
families (level 3). Twenty-four of the family members were the only person from their
family participating in this study, while the remaining 20 family members were
coupled with someone else from their family.
Using HLM, burden, depression, grief, and mastery were modeled separately as a
function of change between pre and post, and between post and the 6 months post
baseline. The passage of time during each period is represented in the model by a
dummy variable at level 1. Therefore, the model includes two dummy variables at
level 1, the first representing the difference between pre- and posttest, and the second
representing the difference between posttest and the 6-month postbaseline follow-up.
Posttest is the category chosen as the reference group (represented by the intercept in
the model), so that the level 1 slope coefficients would represent the relevant mean
levels of change over time.
The HLM models indicate that levels of burden and grief decreased, and mastery
increased from pre-Family Connections to post-Family Connections. Results indicate
that from pre- to post-Family Connections, the mean score of all family units on the
Burden Assessment Scale
2
decreased by an average of 4.14 points
3
(d¼.28)
4
. Overall,
1
The software HLM 5.04 (Raudenbush, Bryk, & Congdon, 2000) was used; the method of esti-
mation used was full maximum likelihood.
2
Because of a near singularity occurring in the BAS data, the intercept variance at level 2 was
set to zero.
3
This average is the regression coefficient at level 2 of the HLM model.
4
Effect sizes were calculated from raw means.
HOFFMAN ET AL. / 221
Fam. Proc., Vol. 44, June, 2005
the mean grief score of all family units decreased by an average of 4.99 points (d¼.45),
and the mean mastery score increased by an average of 4.43 points (d¼.58), Burden
Assessment Scale, t(32) ¼2.15, po.05; Grief Scale, t(32) ¼2.78 po.01; Mastery Scale,
t(33) ¼3.31, po.01. Results showed no change in the mean family unit depression
from pre- to post-FC, t(28) ¼.1.42, p>.05, or in the mean family unit Perceived Burden
Scale score, though a trend did appear for this change, Perceived Burden Scale,
t(27) ¼2.00, p¼.06. See Table 1 for level 1 means at the three time points.
During the 3 months after FC ended, results show that the average family’s score
on the Burden Assessment Scale continued to decrease by an average of 5.78 points
(d¼.65), t(32) ¼2.53, po.05. Changes in mastery and grief were maintained during
those following 3 months, with no significant increase or decrease: mastery,
t(33) ¼.17, p>.05; grief, t(32) ¼1.34, p>.05. No change in depression or perceived
burden occurred during the 3 months post-FC, t(28) ¼1.66, p>.05; t(27) ¼.37,
p>.05.
The level 3 variance components were tested for significance; this significance level
indicates whether individual families significantly vary from the mean intercept and
slopes for all families. If significant variation exists, future research should consider
family-level variables to be entered that may explain some of the variation in burden,
grief, and mastery across families.
For perceived burden and mastery, a significant amount of variation does exist
around the mean intercept for all families (i.e., posttest scores on the scales). For grief,
no variation existed around the intercept, and for the Burden Assessment Scale, the
variance for the intercept had been fixed to zero (see Note 2). For both burden
measures, grief and mastery, significant variation exists around the mean rate of
change for all families from pre to post, and the mean rate of change between post and
6 months post.
DISCUSSION
Families of persons with borderline personality disorder are an underserved pop-
ulation. No standardized information has been available to educate them, support
them, and help them understand the disorder, the chaos that often exists, or the
impact of the illness on their ill relative or themselves (Hoffman et al., 2003). The
Family Connections program is one opportunity for these families to receive much-
needed information, skills, and support. Analyses from this pilot study indicate that
TABLE 1
Descriptive Statistics
Variable Range
Pre Post Follow-Up
Mean (SD)Mean(SD)Mean(SD)
BAS 20–80 51.41 (10.98) 48.35 (11.27) 40.76 (12.06)
PBS 7–28 20.47 (4.13) 18.06 (4.38) 17.04 (4.07)
CESD 14–56 26.58 (9.03) 25.53 (7.32) 24.71 (8.78)
Grief 15–75 52.41 (10.49) 47.62 (10.60) 44.46 (10.64)
Mastery 15–60 39.36 (6.95) 43.28 (6.48) 44.40 (5.51)
Note. Means were run separately for samples at each time point.
BAS is Burden Assessment Scale; PBS is Perceived Burden Scale.
FAMILY PROCESS222 /
www.FamilyProcess.org
family members experienced decreases in their levels of burden and grief, and in-
creases in their level of mastery over the duration of the FC program. Analyses also
indicate that these changes were maintained 3 months after the program had ended.
Although often feeling deskilled by BPD, with its associated problematic behaviors
that are difficult to comprehend, family members report that they are frequently
isolated and alone in dealing with their loved one’s struggles. Equally difficult is ex-
periencing friends and family members as judgmental about the often traumatizing
crises that occur and what they endure (Hoffman et al., 1999). Family members often
report a sense of ‘‘surplus stigma,’’ even from those more enlightened about mental
illness. Surplus stigma is stigma that is over and above what is typically experienced
by family members of persons with other serious mental illnesses. The phenomenon is
perhaps best illustrated by the example of a family member being told by some mental
health professionals and occasional hospital administrators, ‘‘I don’t accept BPDs in
my practice,’’ or ‘‘We don’t want those patients in our hospital.’’
Family Connections is a program specifically designed to addresses these issues.
Data indicate that family members who participated in FC reported significant
changes in several key dimensions that are consistently reported to play a central role
in their lives: burden, grief, and mastery. These constructsFburden, grief, and
masteryFencompass phenomena that often plague family members with issues such
as financial concerns, worries that one’s own behavior may exacerbate BPD symp-
tomatology, mourning lost expectations, and guilt around etiology of the disorder. By
acquiring the most current information on the disorder, learning new emotion and
coping skills, and having a support network, several important aspects of family
member well-being improved significantly from pre- to post-FC group participation.
Importantly, the FC program itself was conducted by several pairs of leaders in dif-
ferent localities, suggesting the transportability of the program.
It is important to note the lack of change in depression on the CES-D vis-a
`-vis
improvements on other dimensions, but this lack of statistical significance for change
in depression should not necessarily be interpreted as an inability of FC to decrease
depression. This finding could simply reflect low statistical power due to the small
sample size used. That change over time was detected in other variables shows the
strength of the relationship between FC and these other well-being variables.
As hypothesized, well-being gains were maintained beyond program completion,
and levels of burden were further reduced during the post-3-month period. Although
it is difficult to explain definitively the additional decrease in burden at follow-up, one
hypothesis consistent with the data is that skills learned in the FC program had a
salutary effect either on relationships in the family (decreasing actual burden), or on
acceptance of the situation such that the actual ‘‘burden’’ of caregiving was experi-
enced as less problematic.
These results provide initial support for the FC program. However, much further
research is needed. Given that there was no control group in this study, it is with
caution that we attribute changes in participants’ well-being solely to participation in
the FC program. However, that pre- to post-FC changes were maintained from post-
FC to the 6-month postbaseline assessment does offer support for FC being respon-
sible for these improvements. But the effectiveness of the FC program must be
evaluated in comparison with one or more control conditions in the future. Although
the link between the program and improvement in burden, grief, and mastery from
pre- to post-Family Connections was supported by the results obtained at the 6-month
HOFFMAN ET AL. / 223
Fam. Proc., Vol. 44, June, 2005
postbaseline follow-up, further research is needed to understand these ongoing im-
provements, and longer follow-ups must be collected and reported to understand
whether the impact of FC is long lasting.
Although family members reported experiencing the value of the three components
of FC (information, skills, and support network) in their efforts to remain centered
amid confusion, we need to identify and understand the mechanisms of change
demonstrated in this initial evaluation. Given that results of the HLM analyses in-
dicated that there is a significant amount of variation in change over time to explain at
the family level, one way to examine important explanatory variables is to enter
relevant variables that vary by family member into level 2 of the HLM models,
and variables that vary by family into level 3. Using these predictors, we may be able
to explain some of the variation that currently exists between and within families
over time.
This study is the first critical step in these lines of research, and equally important,
serves as a major step toward filling the void that exists for family members of persons
with BPD.
REFERENCES
Berkowitz, C.B., & Gunderson, J.G. (2002). Multifamily psychoeducational treatment of bor-
derline personality disorder. In W.R. McFarlane (Ed.), Multifamily groups in the treatment of
severe psychiatric disorders (pp. 268–290). New York: Guilford Press.
Dixon, L., Stewart, B., Burland, J., Delahanty, J., Lucksted, A., & Hoffman, M. (2001). Pilot
study of the effectiveness of the Family-to-Family Education Program. Psychiatric Services,
52, 965–967.
Falloon, I., & Pederson, J. (1985). Family management in the prevention of morbidity
of schizophrenia: The adjustment of the family unit. British Journal of Psychiatry,147,
156–163.
Family Perspectives on Borderline Personality Disorder. (2002). From Research to Community.
Breakout Proceedings Conference of the National Education Alliance for Borderline Per-
sonality Disorder and Columbia University College of Physicians and Surgeons, New York,
NY.
Family Perspectives on Borderline Personality Disorder. (2003). BPD and Co-occurring Dis-
orders. Conference of the National Education Alliance for Borderline Personality Disorder
and Weill Medical College of Cornell University, White Plains, NY.
Fruzzetti, A.E., & Fruzzetti, A.R. (2003). Borderline personality disorder. In D.K. Snyder &
M.A. Whisman (Eds.), Treating difficult couples: Helping clients with coexisting mental and
relationship disorders (pp. 235–260). New York: Guilford Press.
Fruzzetti, A.E., & Hoffman, P.D. (2004). Family Connections workbook and training manual.
Rye, NY: National Education Alliance for Borderline Personality Disorder.
Glick,I.D.,&Loraas,E.L.(2001). Family treatment of borderline personality disorder. In M.M.
MacFarlane (Ed.), Family therapy and mental health: Innovations in theory and practice
(pp. 135–154). New York: Haworth Clinical Practice Press.
Harman, M.J., & Walso, M. (2001). Family treatment of borderline personality disorder through
relationship enhancement therapy. In M.M. MacFarlane (Ed.), Family therapy and mental
health: Innovations in theory and practice (pp. 215–235). New York: Haworth Clinical Practice
Press.
Hoffman, P.D., Buteau, E., Hooley, J.M., Fruzzetti, A.E., & Bruce, M.L. (2003). Family mem-
bers’ knowledge about borderline personality disorder: Correspondence with levels of de-
pression, burden, distress, and expressed emotion. Family Process,42, 469–478.
FAMILY PROCESS224 /
www.FamilyProcess.org
Hoffman, P.D., & Fruzzetti, A.E. (2005). Psychoeducation. In J.M. Oldham, A. Skodol, &
D. Bender (Eds.), Textbook of personality disorders (pp. 375–385). Washington, DC: American
Psychiatric Publishing.
Hoffman, P.D., Fruzzetti, A.E., & Swenson, C.R. (1999). Dialectical behavior therapy: Family
skills training. Family Process,38, 399–414.
Hoffman, P.D., & Hooley, J.M. (1998). Expressed emotion and treatment of borderline per-
sonality disorder. In Session: Psychotherapy in Practice,4(3), 39–54.
Hoffman, P.D., Struening, E., Buteau, E.,Hellman,F.,&Neiditch,E.(2005).Family Perspec-
tives Survey. Manuscript in preparation.
Linehan, M.M. (1993). Skills training manual for treating borderline personality disorder.New
York: Guilford Press.
McFarlane,W.R.,Lukens,E.,Link,B.,Dushay,R.,Deakins,S.A.,&Newmark,M.etal.(1995).
Multiple-family groups and psychoeducation in the treatment of schizophrenia. Archives of
General Psychiatry,52, 679–687.
Raudenbush, S., Bryk, A., & Congdon, R. (2001). HLM for Windows (Version 5.04). Lincoln-
wood, IL: Scientific Software International, Inc.
Reinhard, S., Gubman, G., Horwitz, A., & Minksy, S. (1994). Burden Assessment Scale for
families of the seriously mentally ill. Evaluation and Program Planning,17, 261–269.
Struening, E., Stueve, A., Vine, P., Kreisman, D., Link, B., & Herman, D. (1995). Factors as-
sociated with grief and depressive symptoms in caregivers of people with mental illness.
Research in Community and Mental Health,8, 91–124.
HOFFMAN ET AL. / 225
Fam. Proc., Vol. 44, June, 2005
