ARTICLE IN PRESS
Social Science & Medicine 63 (2006) 3067 3079
www.elsevier.com/locate/socscimed
The interpersonal experience of health care through the eyes of
patients with diabetes$
Paul Ciechanowski , Wayne J. Katon
Department of Psychiatry and Behavioral Sciences, University of Washington School of Medicine, Box 356560, 1959 NE Pacific,
Seattle, WA 98195-6560, USA
Available online 25 September 2006
Abstract
Patients with chronic illness often face challenges navigating the US health care system because of the system s lack of
coordination and continuity. Patients with more difficulty relying on others and with reluctance in engaging frequently or
in-depth with providers, face even greater challenges obtaining optimal health care in this system. Using a self-report
measure of attachment style, we selected patients with varying degrees of comfort and trust in relationships. We conducted
qualitative semi-structured interviews with a purposive sample of 27 patients with type 2 diabetes attending the University
of Washington Diabetes Care Center in Seattle to explore issues of trust and collaboration in the health care setting. We
used a constant comparative approach in which contemporaneous data collection and analysis took place. A subset of
patients with fearful and dismissing attachment style reported having low levels of trust and an inability to collaborate with
others of longstanding duration. Many aspects of the current health care system, such as its rushed, impersonal nature and
a perceived   wall  between providers and patients were frustrating for most study patients. Patients with fearful and
dismissing attachment style reported that these aspects of the health care system often interfered with their ability to
partner with providers but also reported that patient-centered attitudes and behaviors by providers could improve their
trust and ability to engage in the health care system. Implications of using a conceptual model of attachment theory to
improve patient-centered care and customer service are discussed.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Diabetes; Attachment style; Trust; Patient provider relationship; Patient centered; Customer service; USA
Introduction become knowledgeable about their conditions,
share in decision making, receive education and
Chronic illness care is optimally carried out in a disease self-management support, and are provided
collaborative process with active follow-up and with optimal medication management in a sus-
tracking of outcomes and adherence by providers tained, consistent and timely fashion (Katon et al.,
and the health care system. Patients, in turn, 1997; Von Korff, Gruman, Schaefer, Curry, &
Wagner, 1997). Such a partnership between patients
and providers facilitates adoption of guideline and
$
Supported by grant K23 DK60652-01 (National Institute of
evidence-based treatments, increases patient-cen-
Digestive and Diabetes and Kidney Diseases).
tered interactions (Ciechanowski, Wagner et al.,
Corresponding author. Tel.: [phone].
2004; Neumeyer-Gromen, Lampert, Stark, &
E-mail addresses: [e-mail]
(P. Ciechanowski), [e-mail] (W.J. Katon). Kallischnigg, 2004; Stewart, 1995), and contributes
0277-9536/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.socscimed.2006.08.002
ARTICLE IN PRESS
3068 P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079
to optimal management of complex illnesses with particularly at vulnerable times (e.g. managing one s
complicated self-care and treatment regimens such disease or symptomatic and functional challenges of
as diabetes. illness). These models influence whether individuals
Despite promising changes in chronic disease deem themselves worthy of care (model of self) and
management in today s health system, the prevailing whether others are perceived as trustworthy to
delivery model may not support such a collabora- provide care (model of other). Studies demonstrate
tive process because of system-wide fragmentation. high stability and continuity of attachment models
As the recent Institute of Medicine Crossing the between infancy and adolescence (Hamilton, 2000)
Quality Chasm report describes, contemporary and infancy and adulthood (Waters, Merrick,
health care delivery is characterized by frequent Treboux, Crowell, & Albersheim, 2000). Based on
handoffs between providers, infrequent clinical empirical research in infants, children and adults
follow-up, reliance on in-person physician visits over the past three decades, social psychologists
with limited web-based or phone contact, lack of (Griffin & Bartholomew, 1994) have identified four
support for behavioral change, and lack of time and patterns of attachment behaviors in adults: secure,
resources for patient self-management training preoccupied, dismissing and fearful. These four
(2001). While the current health system is frustrat- attachment styles can be considered conceptually
ing for many patients with chronic conditions distinct dimensions and individuals may be char-
(1997), the capacity for a patient to successfully acterized interpersonally by varying degrees of each.
navigate this system and experience patient-centered Clinically and descriptively, however, it is often
care may be especially challenging for patients with more useful to conceptualize individuals in terms of
specific interpersonal characteristics. their predominant attachment style so as to better
The   Chasm  report recognizes the importance of understand developmental and behavioral charac-
individual differences in preferences and approaches teristics of each style.
in working with health care professionals (2001). Adults who have predominantly secure attach-
Patients with more reluctance depending on physi- ment style are generally believed to have experi-
cians and health care teams may have greater enced consistently responsive (Ainsworth, Blehar,
difficulty receiving high-quality health care and Waters, & Wall, 1978) early caregiving (in the
achieving optimal outcomes in a fragmented health process developing a positive model of self
system. Such patients may be more sensitive to the and other; Fig. 1) and are generally comfortable
lack of shared decision making and pervasive lack depending on and being readily comforted by
of coordination and continuity of routine health others. Adults with predominantly preoccupied
care compared to patients who are more comfor- attachment style are posited to have experienced
table in the traditional patient physician role. There inconsistently responsive caregiving (Bartholomew,
has not been a well-established theoretical approach 1990) and in an effort to ensure proximity to
to understanding and working with individual caregivers, they use strategies in which the attach-
differences in patients preferences for interacting ment behavioral system is   hyperactivated  through
with providers within the health care system. exaggeration of behaviors attracting support (Mi-
Attachment theory a model that recognizes that kulincer, Shaver, & Pereg, 2003). They are generally
all individuals have underlying cognitive-emotional emotionally dependent on others approval (positive
schemas guiding their perceptions and behaviors in model of other), often to the point of being
interpersonal relationships provides a promising   clingy,  but generally have poor self-worth (nega-
conceptual framework with which to practically tive model of self).
approach individual differences in preferences for The remaining two styles, dismissing and fearful
receiving health care (2001; Dozier, Cue, & Barnett, attachment styles, are characterized by strategies in
1994). which the attachment behavioral system is   deacti-
In developing attachment theory, John Bowlby vated  (Mikulincer et al., 2003 ), i.e. there may be
proposed that all individuals psychologically incor- avoidance of support-seeking behaviors or denial or
porate prior experiences with caregivers, forming minimization of emotions and cognitions associated
enduring mental representations of caregiving that with attachment needs. Adults with predominantly
persist into adulthood called   internal working dismissing attachment style are believed to have
models  (Bowlby, 1977). Such models are learned experienced early caregiving that was largely emo-
ways of interacting in relationships throughout life, tionally unresponsive. As a result, they develop
ARTICLE IN PRESS
P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079 3069
Fig. 1. Attachment style categories and model of self and other. Adapted from Bartholomew & Hororwitz (1991).
strategies from an early age in which they become matters, they are emotionally reactive, but do not
  compulsively self-reliant  (Bowlby, 1977) (positive actively deal with their distress or seek support.
model of self) and uncomfortable trusting others They can acknowledge feeling bad but avoid self-
(negative model of other). Individuals with dismiss- disclosure or appearing upset in front of others
ing attachment style are described as lacking in because of fear of rejection. Individuals with fearful
emotional self-disclosure and as emotionally cool or attachment style may have a few close relationships
aloof (Bartholomew, 1990). They distract them- that typically take years to establish and have
selves from emotions at times of upset and they difficulty breaking off such relationships because of
actively avoid seeking out support. Because of their fear of ever finding another relationship.
high interpersonal self-reliance, they may also have Previous studies in community, college and
moderate to high self-confidence. While downplay- medical populations have explored how adult
ing the importance of relationships they often stress attachment processes may interpersonally influence
the importance of independence, freedom and stress, coping and health-related outcomes. For
achievement. example, in student samples it has been confirmed
Individuals with predominantly fearful attach- that in stressful situations, compared to individuals
ment style may initially desire social contact (i.e., with secure attachment styles, individuals with
not highly self-reliant), but this desire is inhibited by dismissing attachment style demonstrate less self-
fear of rejection. These individuals are proposed to disclosure and reciprocity (Mikulincer & Nachshon,
have had overly critical, harsh or rejecting caregiv- 1991) and individuals with fearful attachment style
ing (negative model of self and other) and as adults demonstrate less collaboration (Lopez et al., 1997).
demonstrate interpersonal approach-avoidance In a study of expectant parents, parents with secure
behavior stemming from a fear of intimacy attachment style were more willing to seek out
(Bartholomew, 1990). Interpersonally, they appear therapy for mental health problems and were more
as hesitant, vulnerable, shy, self-conscious or as satisfied with care compared to parents with
having a low self-confidence (Bartholomew, 1990). insecure attachment styles (Riggs, 2001). In another
When confronted with problems or upsetting study, patients with dismissing attachment style
ARTICLE IN PRESS
3070 P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079
were least likely to seek out support in the form of demographic data, clinical characteristics, self-care
psychotherapy (Riggs, Jacobovitz, & Hazen, 2002). adherence and attachment styles of patients attend-
In medical populations, similar coping ap- ing this clinic within the prior 2 years. Response
proaches have also been found. Among patients rate to this questionnaire was 58%. To guide
with breast cancer, chronic leg ulcers and alopecia selection of a patient subgroup for recruitment to
those with dismissing attachment style more the qualitative study, purposive sampling (Patton,
often used denial coping compared to patients with 1990) was used to achieve maximum variation in
secure attachment style (Schmidt, Nachtigall, Wue- gender, age, race/ethnicity and in attachment styles
thrich-Martone, & Strauss, 2002). In HIV positive directly or inversely associated with trust and ability
patients, secure attachment style was associated to rely on others. To optimally explore health
with less perceived global stress in the prior month care experiences among patients with low levels of
compared to patients with insecure attachment style trust, we intended to recruit approximately twice as
(Koopman et al., 2000). many patients within each of dismissing and
Previous studies have explored the potential fearful attachment style groups as within the secure
influence of attachment styles on treatment adher- attachment style group.
ence and outcomes in patients with diabetes. For We also aimed to ensure maximum variation in
example, studies have demonstrated poorer diabetes glucose control among study patient based on
self-care, insulin and hypoglycemic medication glycosylated hemoglobin (HbA1c). HbA1c is accepted
adherence and higher glucose levels in diabetic as the best measure of recent glycemic control (last
patients with fearful and dismissing attachment 120 days) and is used to guide clinical management
style as compared to patients with secure attach- (Goldstein et al., 1995). Lowering HbA1c has been
ment style (Ciechanowski, Russo et al., 2004; associated with a reduction of microvascular and
Ciechanowski, Hirsch, & Katon, 2002; Ciechanows- macrovascular diabetic complications and the Amer-
ki, Katon, Russo, & Walker, 2001; Turan, Osar, ican Diabetes Association recommends developing
Turan, Ilkova, & Damci, 2003). or adjusting the management plan to achieve normal
In this paper, we present the results of a (4 6%) or near-normal (o7%) HbA1c values (2005).
qualitative investigation of the experiences of To measure HbA1c, the University of Washington
patients with type 2 diabetes in their interactions Diabetes Care Center laboratory uses a Bayer
with the health care system in managing diabetes, DCA2000, which is certified by the National
while taking into account their general capacity to Glycohemoglobin Standardization Program as hav-
rely on others based on attachment theory. We ing documented traceability to the Diabetes Control
hypothesized that compared to patients with secure and Complications Trial reference method.
attachment style, those with dismissing and fearful
attachment style would be: (1) less satisfied with Attachment style determination
interactions with health care providers; (2) less
trusting of health care providers and (3) less able to Participants completed the Relationship Ques-
collaborate in health care settings. tionnaire (RQ) (Griffin & Bartholomew, 1994),
created by Bartholomew and colleagues based on
Study design Bowlby s attachment theory (Bowlby, 1977), which
measures respondent attachment style. This mea-
Subjects and settings sure demonstrates convergent and discriminant
validity with other self-report and interview ratings
Twenty-seven patients with type 2 diabetes (Scharfe & Bartholomew, 1994). Since preoccupied
attending the University of Washington Diabetes attachment style has not been consistently asso-
Care Center in Seattle, Washington were recruited ciated with diabetes treatment adherence or adverse
to the study. This tertiary care clinic provides outcomes compared to secure, dismissing and
diabetic health care for 3000 patients and is staffed fearful attachment styles (Ciechanowski et al.,
by eight physicians, two nurse practitioners and two 2006), an a priori decision was made not to include
nutritionists. Patients were identified from clinic this attachment style group in this study.
rosters and had initially responded to a self-report Items assessing each attachment style consist of
questionnaire from the Diabetes Care Study paragraphs describing each style (Table 1). Respon-
(Nź395 type 2 diabetes patients) which assessed dents are asked to think of all past and current close
ARTICLE IN PRESS
P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079 3071
Table 1 tives of diabetes care and health care, including the
Descriptive paragraphs of attachment styles (Bartholomew &
patient provider relationship. Telephone interviews
Horowitz, 1991)
were shorter in duration and addressed recent
intervening clinic visits that patients attended at
Secure attachment It is easy for me to become emotionally
style close to others. I am comfortable the Diabetes Care Center.
depending on them and having them
Interview topics included: (1) assessment of the
depend on me. I don t worry about
patients understanding of diabetes and related
being alone or having others not accept
complications; (2) a review of all prior and current
me
health care relationships related to diabetes treat-
Fearful attachment I am uncomfortable getting close to
style others. I want emotionally close ment; (3) discussion of health visit frequency, modes
relationships, but I find it difficult to
of contact with providers and perceived quality of
trust others completely, or to depend
prior and current patient provider relationships;
on them. I worry that I will be hurt if I
(4) assessment of diabetes self-management; (5)
allow myself to become too close to
discussion of patients attachment style character-
others
Dismissing I am comfortable without close istics; (6) patients perceptions of how family may
attachment style emotional relationships. It is very
help with or hinder diabetes self-management.
important to me to feel independent
Patients prior interviews were reviewed and
and self-sufficient, and I prefer not to
additional questions were added to subsequent
depend on others or to have others
interviews to address issues raised by a patient in
depend on me
Preoccupied I want to be completely emotionally a previous interview, or in response to emerging
attachment stylea intimate with others, but I often find
themes that arose from the interview with the same
that others are reluctant to get as close
patient or with other patients.
as I would like. I am uncomfortable
being without close relationships, but I
Analysis
sometimes worry that others don t
value me as much as I value them
We used a constant comparative approach in
a
Patients with preoccupied attachment style were not included
which contemporaneous data collection and the-
in this current study.
matic analysis of interview data took place (Strauss
& Corbin, 1990). All interviews were audiotaped,
relationships when completing the questionnaire
transcribed and managed using QSR N6 Version 6.0
and to choose the style suiting them best.
(QSR International Pty Ltd.). Themes related to
health care experiences were derived from data
Semi-structured interviews rather than being imposed in the analysis, though
the interviewer was not blinded to patients attach-
Each patient was enrolled for 3 months. In- ment style categories. Emerging themes, issues and
person interviews took place at the University of hypotheses from earlier interviews informed subse-
Washington Medical Center. Interviews usually quent interviews in an iterative process (Strauss &
occurred on the day of a scheduled Diabetes Care Corbin, 1990). Data were organized into initial and
Center appointment and were tape-recorded with higher level codes and clustered across transcripts to
patients consent and transcribed in full. The study derive primary interpretative themes.
protocol was reviewed and approved by the Uni-
versity of Washington institutional review board. Results
All participants gave written informed consent.
Semi-structured interviews were conducted by Among 395 patients with type 2 diabetes from the
the first author (PC). A repeat interview design clinic-wide sample: mean age was 56.8711.8 years
facilitated the development of trust and rapport (range 19.9 85.3 years); 194 patients (49.1%) were
(Mathieson, 1999), particularly around issues pa- female; 72 patients (18.2%) belonged to a race/
tients might have initial reluctance to discuss. Two ethnic minority; 329 patients (83.3%) had at least
in-person interviews lasting 30 50 min were con- one year of college education; 157 patients (41.8%)
ducted 3 months apart, interspersed with an inter- had secure attachment style; 114 patients (30.3%)
view conducted by telephone. In-person interviews had dismissing attachment style; 72 patients
enquired about various aspects of patients perspec- (19.1%) had fearful attachment style; 258 patients
ARTICLE IN PRESS
3072 P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079
(65.3%) were taking oral hypoglycemic medica- Interviewer: You can get close to people as long
tions; 245 patients (62.0%) were taking insulin and as you re in control?
126 (32.1%) were taking both classes of medica- Patient: I would say 95% of the time if I lose that
tions. Mean glycosylated hemoglobin values was control of myself, then I ve gotten hurt.
7.471.2% (range 4.6 to 13.0%). Interviewer: I see. And what is it like if you ?
In the sample selected for the qualitative study, Patient: What is it like to be there?yLonely.
sixteen women (59.3%) and eleven men (40.7%) with Patient 2069 Dismissing attachment style
type 2 diabetes between the ages of 27 and 79 years
This reflection on the loneliness resulting from
(mean age 54.4711.8 years) participated. Nine patients
being repeatedly interpersonally hurt may also
(33.3%) belonged to a racial/ethnic minority and 25
speak to the need for patients with dismissing
(96.2%) had at least one year of college. Five patients
attachment style to generally deny the importance
(18.5%) had secure attachment style, eleven patients
of relationships and avoid closeness and intimacy.
(40.7%) had dismissing attachment style and eleven
patients (40.7%) had fearful attachment style. Nine- Patient: I ve had troubles with intimacy in my
teen patients (63.2%) were taking oral hypoglycemic relationships. It s been borne out in expecting
medications, 19 patients (63.2%) were taking insulin others to really be committed to taking care of
and 12 (44.4%) were taking both. Mean glycosylated you or rising to whatever occasion and then
hemoglobin value was 7.971.2% (range 5.9 10.8%). being mistaken often enough that you learn that
being independent is very, very important.
Development of trust Interviewer: Because you ve been failed before
when you depended on others?
Patients with dismissing or fearful attachment Patient: Yes, including the medical situation.
style described a long-standing general distrust or Patient 1852 Fearful attachment style
inability to collaborate with others, often extending
When I have been close with others, it s a great,
to their health care providers. This is consistent with
wonderful feeling. But then all of a sudden
an attachment theoretical framework in which
something happens in the relationship to where
patients with these attachment styles develop a
I m like the animal that comes up and eats out of
  negative model of other  based on consistently
your hand. And then all of a sudden you put the
unresponsive caregiving. Similar experiences of
other hand over to capture meyI m gone.
long-standing distrust were not observed among
Patient 2069 Dismissing attachment style
patients with secure attachment style.
You don t want to stand up and start asking
Trusting blindly, saying   OK, I ll go with
questions and say   Yeah, but what about this? 
whatever you say  , I don t have that trust.
or anything. You just kind of sit back and take it
Patient 3774 Dismissing attachment style
because we ve all been raised to sit back and take
Interviewer: When you do go to ask someone for it, you know.
help, how does it feel? Patient 3752 Fearful attachment style
Patient: I feel like I have to justify it. I feel like I
This last passage may speak to the perceived
have to say,   OK, I ve done this, this and this
harshness of early caregiving that is typically more
and you re my last resort.  Rather than just
characteristic of patients with fearful attachment
feeling to the point where I can say   I need your
style compared to those with dismissing attachment
help  .
style.
Patient 3752 Fearful attachment style
When asked about their need to be independent Implications of lack of trust in the health care setting
or their inability to rely on others, patients with
dismissing or fearful attachment style frequently Fearful and dismissing attachment style
gave as reasons the ways they were responded to in Such pervasive lack of trust may have implications
the past or prior experiences in which they were for health care utilization patterns or the quality of
emotionally hurt. interactions between patient and provider.
Patient: I can still turn around and cut you off I would say never go for being completely
and walk away, because I don t become attached. trusting because you re the only one that can
ARTICLE IN PRESS
P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079 3073
really know what it s like to be in your skin. And Interviewer: What makes you have trust in your
if you re having some difficulties you have to provider?
address those, and if the caregiver s not respond-
Patient: My provider demonstrates knowledge in
ing, you go see someone else.
current research or at least stuff that seems like it
Patient 1852 Fearful attachment style
is current because she says   I was reading this  
it s just that knowledge base she has.
Patient: I always want to be able to do things
Patient 3629 Secure attachment style
myself, and have the ability to do things myself.
And so sometimes it s very difficult to ask people
Interviewer: Is an ongoing relationship with your
for help.
provider over time important for you?
Interviewer: Before you had diabetes, are you
Patient: I think it is. My experience has been that
someone that typically went to the doctor?
frequently doctors become so busy that they
Patient: No. I stay away from those peopley
really don t remember patients themselves per-
Interviewer: Does your current doctor ask you to
sonally so they need to look at their charts, which
help make decisions about your diabetes care?
is only appropriate.
Patient: My doctor sometimes tricks me into that.
Patient 3620 Secure attachment style
Patient 3438 Dismissing attachment style
In this last passage, as in prior passages of
Patient perceptions of a power differential between
patients with dismissing attachment style, there is a
health care providers and patients
high sensitivity to or fear of being controlled.
Perceptions in patients with dismissing and fearful
Ok, Dr. X, I don t trust because I honestly feel
attachment styles: A   wall  between patient and
that attention wasn t paid. Might have been that
provider
day, might have been something else, but
Patients often perceived providers and the health
regardless, the doctor did that in front of me so
care system negatively. Patients with fearful or
we are not going to be able to build that trust.
dismissing attachment style often described a
Patient 1206 Fearful attachment style
barrier between providers (particularly physicians)
Patients with fearful attachment style, on the
and patients. They used metaphors such as   wall 
other hand, seem less concerned about being
or   line,  or described attitudes or perceived body
controlled than about not being responded to which
language to characterize this divide. Often, char-
is consistent with the observation that they are
acteristics of the health care setting (e.g. rushed
highly sensitive to rejection.
environment) were viewed as contributing to their
negative experience.
Secure attachment style
Sometimes you feel like there s this line you re
Patients with secure attachment style were gen-
the patient, they re the provider and you can t
erally more likely to trust providers and value
cross over They have this attitude. You re
an ongoing relationship, even when circumstances
sitting there waiting in the waiting room and
were not ideal. They often cited their health care
they re late to begin with and there s nothing else
provider s knowledge as an important determinant
to do. And finally they come in and they ve got
of their ability to trust their provider.
just this whole air about them that you re
Interviewer: Did you have a lot of trust in your
imposing on them almost because they re behind
provider at today s appointment?
schedule and   Get it over with, I ve got better
Patient: Going back to that rushing, if I don t
things to do  , is how it comes across. You re
catch my provider I still have trust in what we
almost afraid to speak up and say, you know,
did, and I got everything asked that was on the
  I ve got these questionsy  .
top of my head but I would have liked a little
Patient 3752 Fearful attachment style
more time.
Interviewer: What makes you trust your provider?
In this last passage describing a clinical encounter
Patient: I think my provider has a lot of practical in a patient with fearful attachment style, themes
knowledge. of perceived harshness and rejection arise. In
Patient 3576 Secure attachment style contrast, in the following passage from a patient
ARTICLE IN PRESS
3074 P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079
with dismissing attachment style, a theme of being When patients perceived that providers tried to get
controlled is apparent. to know them or that they were accepted with a
nonjudgmental attitude, they were more satisfied
It s like asking the difference between,   Have a
and reported being more likely to engage in the
seat and tell me what you need or what I can do
patient provider relationship.
to help you  and   Sit down and let s find out
what s wrong with you and I ll tell you how to He doesn t just come in, do your treatment and
get better or not  . The   doc thing  to me is a leave. He kind of talks, you know,   How things
body language that says   I have the information going? Tell me about yourself  , and he has a
you need and if you follow and take this fabulous memory. He remembers about those
information you re going to get well. I m the things that you tell himy I don t know if it s just
doctor, you re the patient, I have the answers a really good memory, or he puts notes in the
that you need and that s the end  . chart, or whatever, but it s just yyhe makes you
Patient 1979 Dismissing attachment style feel comfortable coming iny With my doctor
who I have now, I feel like I can be my weird,
If a point would come in holding on to that
twisted self and my doctor understands that.
  power doc  thing that some people do then I d
Patient 3752 Fearful attachment style
be gone.
Patient 1371 Fearful attachment style
Uh, Dr. Z, I don t know too well yet. But I feel
comfortable in the fact that if I come with a
My health care provider always goes on about
question, no matter what it might be, I can ask.
how wonderfully I m doing. I am reminded that
Patient 1743 Fearful attachment style
my provider sees people who are much, much
worse off than I am. Those people are not
The benefits of a relationship built on trust
relevant to me at the moment when I am sitting
particularly where the patient feels empowered
in the room during my appointment!
about diabetes self-care can be significant.
Patient 2478 Fearful attachment style
But somehow the way she approaches her
practice of working with diabeticsyat least for
Perceptions in patients with secure attachment style
me, she has been able to communicate a sense of
Patients with predominantly secure attachment
  I am bigger than this disease  and that together
style also acknowledged shortcomings of the health
as a team this can be managed. And it s not
care system and providers behavior. However, such
something that is so big and so powerful.
negative attributes did not appear to significantly
Patient 1371 Fearful attachment style
diminish trust or threaten the treatment relation-
ship.
She finds something good no matter what.
There s not much when you get a bad report.
Dr. Y is very intense, and if I m not thinking fast
She always finds some good in there. She works
enough, is out the door before I can get every-
on the positive and I like that instead of always
thing out that I might want to ask. Today, I felt
the negative, you know, like a lot of doctors do.
like the appointment ended before I was ready
Patient 1156 Fearful attachment style
for it to end. Sometimes I feel rushed and that my
doctor doesn t take as much time with me as I d
When providers were perceived as nonjudgmen-
like, focusing on one thing and ignoring other
tal, patients felt more comfortable acknowledging
things yit is kind of frustrating for mey but it
when they had inadequate insurance coverage for
is not a major thing.
components of their treatment and were more able
Patient 3576 Secure attachment style
to negotiate around their treatment plan.
Overcoming lack of trust in the clinical setting And I feel comfortable saying I don t have good
coverage for a specific treatmenty   Can we
Patient-centered actions by providers consider something else?  yWhereas with the
other doctor, I never would have thought about
The attitude, clinical approach or behaviors of
providers can potentially enhance the capacity for it, because he would be,   This is what I ordered
patients with dismissing or fearful attachment style and that s what you re going to be on. 
to trust or engage in the health care relationship. Patient 3752 Fearful attachment style
ARTICLE IN PRESS
P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079 3075
Even small patient-centered acts by providers   game  they thought would allow them to stay in
such as acknowledging being late and that the the relationship and ultimately get what they
patient has been left waiting can be experienced as needed.
validating in patients with difficulty trusting others.
And what I don t want to do is have to look for
And you can t control it obviously, but he was somebody else. I hate doing that. It s like dating
very good and you know, if he was running late almost. I mean, I don t want to go there. And so
he would pop his head in and   OK, I ll be in in I guess I m willing to sacrifice the   Dr. Welby 
mentality for the stress I d have to go through to
10 min  or whatever. And get you noticed. So at
least you knew what was going on. You re not find somebody else.
just sitting there, wondering. Patient 1610 Fearful attachment style
Patient 3752 Fearful attachment style
Interviewer: So, what if you don t play the role
of the   good patient  , what do you fear the
Patients with difficulty trusting others reported
consequence might be?
relying on body language as a way of gauging
Patient: Well, I think that they cut down the time
safety in the health care relationship. Certain
that they spend with you. And I felt that when I
examples of body language were seen as powerful
disagreed about how to deal with my medical
means of enhancing the quality of the health care
situation: my provider absolutely withdrew and I
relationship.
would have less time in the office.
Interviewer: You say you like face-to-face
Patient 1852 Fearful attachment style
visitsyWhat do you like about them?
This so-called game I have figured outy..I think
Patient: Because I can read what the doctor is
it s particularly prevalent and obvious in the case
thinking in his eyes.
of specialists. But I find it with my general
Patient 2069 Dismissing attachment style
practitioner as well. So over the years I ve figured
Of course, when she is sitting on the step that
this out and my task as a patient is to go along
pulls out on the bottom of examination table,
with whatever protocols are set. But at the same
you can t help to think of her as a person. Cause
time, be assertive enough to get what I think I
she s kind of sitting down there. And that s
need. Meaning, I have to play by the rules of this
important. You know body language is really
game that I think I ve figured out. But at the
important. She pulled that thing out, sat down.
same time remind the caregiver continuously that
This is two times in a row!
I m not just a statistic.
Patient 1371 Fearful attachment style
Patient 2478 Fearful attachment style
And the nonverbal communication: The sound
These last passages again point to the difficulty
of his voice, the pitch of his body, the way he tries
that patients with fearful attachment style have with
to maintain his schedule, what he has to say and
rejection and the extent to which they may go to
how he says it when he can t just being human.
avoid having to start again once relationships are
Someone that is just open if you meet them
established.
they ll shake your hand and say hello and look at
you, you know, and take a little bit of time to see
Discussion
what you re about before they just look at your
chart and say and judge, and just put you in
As has been previously reported in a large
this block   Okay, you re this  and not talk to
American Hospital Association survey of 37,000
you or get your point of view about it.
patients across the US, it is highly instructive and
Patient 1979 Dismissing attachment style
revealing to see the health care system   through the
patient s eyes  (1997). The prevailing health care
Tolerating the health care relationship: playing system was described as a   nightmare to navigate 
games and roles and was characterized as complex, cold and
Patients with low levels of trust, particularly impersonal with fragmented information- or deci-
patients with fearful attachment style, were often sion-sharing, and a perception that the system
reluctant to change providers who they did not blocked access, reduced quality and limited spend-
perceive as ideal or they played a   role  or a ing for care at expense of patients. The influential
ARTICLE IN PRESS
3076 P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079
Institute of Medicine Crossing the Quality Chasm tion and that potentially facilitated engagement
report several years later acknowledged positive within the health care relationship. Many patients
changes had occurred to rectify some of these issues, with fearful attachment styles also learned that
but that many of these perceptions were still well- changing providers took an emotional toll and
founded (2001). Such reports have served as learned instead to play   games  or   roles  such as
important focal and anchor points from which to   the good patient  as a way of tolerating a
improve the experience of health care from the potentially difficult health care interaction (Fig. 2),
patient s perspective. which is consistent with descriptions of individuals
In the current study, we gained valuable insight with fearful attachment style in the literature.
into patients past and current experiences with In a recent large epidemiological study of over
diabetes-related health care while taking into 4000 primary care patients with diabetes (Ciecha-
account their cognitive-emotional schemas related nowski, Russo et al., 2004) and in the initial
to trust based on attachment theory. Patients with screening component of this study, 48 49% of
dismissing and fearful attachment style may be like patients had dismissing or fearful attachment style.
  canaries in the coalmine  who more acutely see From a population-based and public health per-
and feel the inherent problems with chronic illness spective, there are many opportunities for interac-
care in the prevailing health care system. On the tions between patients who have difficulty trusting
other hand, having secure attachment style may others and a health care system that is often
mitigate against experiencing the frustrations of the fragmented, impersonal, and often felt by patients
health care experience as a reason to change to recapitulate earlier, sometimes emotionally pain-
providers or quit the system. ful experiences. Given characteristics of the current
We found that most patients in our sample health care system, clinicians are left with the main
perceived the health care system as rushed, im- responsibility to recognize when patients have
personal and fragmented, regardless of their attach- difficulty collaborating in the health care setting
ment style. Patients with dismissing and fearful and to change their own behaviors and attitudes in
attachment styles (negative model of other) also order to facilitate and nurture trust, particularly
reported perceiving a distinct division, or power with patients who are characteristically less trusting.
differential, between providers and patients that There may be limitations however, in how
threatened their ability to engage in the health care accurately providers can assess who is at risk for
system. Patients with fearful attachment style were difficulty with collaboration since patients, when
highly attuned to indications of rejection and asked about satisfaction with providers or their
patients with dismissing attachment style were health care, typically soften any criticism because
highly sensitive to being controlled. While these of social desirability biases. Furthermore, patients
patients could identify provider behaviors or system who attempt to truthfully answer questions in a
characteristics that jeopardized their relationship clinical setting such as   What degree of contact do
with the health care system, they could also describe you prefer with your provider?  or   Why do you
behavioral and health care system attributes they delay seeking care for your condition?  may have
felt mitigated against health care system fragmenta- difficulty responding because of shame, social
Prevailing
health care system:
Patient health
care utilization
" Characterized by
patterns:
fragmentation
Patient
" Engagement
attachment style
" Reluctance to seek
and
care
capacity to trust
" Leaving care
Health care
" Frequently changing
provider:
providers
" Playing a  role or
" Ä… patient-centered
 game to tolerate
" Ä… potential power
care
differential
Fig. 2. Diagrammatic summary of findings.
ARTICLE IN PRESS
P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079 3077
desirability or because they truly do not know. By that the primary locus of power and control in the
using an indirect approach to assess patients trust health care relationship lies in the provider (Mayer &
and ability to rely on others in the health care Cates, 1999). On the other hand, when patients are
setting (i.e. by asking about past and current less acutely ill assuming a more   vertical  posi-
relationships in general and not specific to the tion they may expect to be treated like customers
health care setting), we were able to avoid some of where power and control in the relationship is
these limitations and potentially derive better shared. It is instructive to review what business
determinants of satisfaction and health care beha- literature has long identified as the most important
viors than might be derived through direct ques- elements of exemplary customer service resulting in
tioning. This is a significant advantage of using customer satisfaction and loyalty to a service,
attachment theory-based questionnaires and is product or company (Mooney, 2002): empowerment
consistent with recent health services (Hohmann, of customers, provision of choice, clear communica-
1999) and customer service (Christensen & Olson, tion, customer relevant education and information,
2002; Zaltman, 2003) literature that recognizes that continuous access, multiple modes of access, capa-
indirect methodologies that tap underlying cognitive city to receive a timely response, product/service
and emotional schemas may have benefits over expertise, attentive listening to customers, and
direct assessment, and may even be essential to development of a shared action plan.
understanding and working with   irrational  or less Key components of recent chronic care models
understood patient behaviors (Hohmann, 1999). As (Glasgow et al., 2005; Katon et al., 1997; Von Korff
clinicians better understand patients and their inner et al., 1997) where patients are educated about
emotional world, patient behaviors are experienced their condition(s), activated and share in decision
as less and less   irrational  . making, are provided a choice in treatment, and set
Prior studies have found that patients with goals to work on are almost identical to the valued
dismissing and fearful attachment style had fewer elements that have been cornerstones in customer
social supports, more adverse childhood experiences service for decades. Given greater dissatisfaction
(e.g. loss, neglect, abuse), and were more dissatisfied with health care among patients who live in a world
with their providers and health care systems than where they can expect and demand tailored,
patients with secure attachment style (Ciechanowski customer service in their daily pursuit of other
et al., 2001; Ciechanowski et al., 2006). For such services and products, it is not surprising, perhaps,
patients, the health care system can impose signifi- that medicine is finally starting to shift toward a
cant distress and a sense of invalidation. For patients customer service model.
with difficulty interacting or trusting others at times In their article on health service excellence, Mayer
of need, interpersonal validation and a sense of safety and Cates (1999) recommend that clinicians make a
must be attained before adequate treatment adher- customer service diagnosis in addition to a clinical
ence, optimal health care utilization patterns and diagnosis when assessing a patient. That is, clin-
effective disease self-management can be expected. icians should ask if they are meeting the unique
However, patients less satisfied with providers or the needs of a patient in a caring, competent, compas-
health care system, who miss appointments or who sionate and skillful way. A primary step in
are less adherent with self-care regimens may be delivering tailored, attuned customer service (i.e.
perceived as   difficult patients  (MacDonald, 2003). patient-centered care) to patients with chronic
While such designations are often made in earnest in illness may necessarily include recognizing and
order to improve clinical responses (Hahn et al., conceptually understanding characteristics of a
1996), these labels may absolve a health care system patient s attachment style especially with patients
of the need to consider whether it offers an adequate less inclined to collaborate with the health care
range of responses to patients with varying inter- system. By analogy, depression treatment in pri-
personal preferences. mary care did not improve significantly until sound
In our study, many patients cited the power methods of recognizing and understanding the
differential between patient and provider as a needs of depressed patients were well delineated
significant deterrent to forming a collaborative (Katon, 2003). Once patients with problems trusting
health care relationship. Literature on health service health care providers are identified through clinical
excellence suggests that acutely ill,   horizontal  (i.e. assessment using measures of attachment style,
bed-ridden) patients may traditionally anticipate changes in the health care system to address
ARTICLE IN PRESS
3078 P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079
patients unique needs can take place. In addition Bartholomew, K. (1990). Avoidance of intimacy: An attachment
perspective. Journal Social and Personal Relationships, 7,
to provider education about attachment styles,
147 178.
effective interventions may make use of ancillary
Bartholomew, K., & Horowitz, L. M. (1991). Attachment styles
providers who either facilitate interactions between
among young adults: A test of a four-category model. Journal
providers and patients or who actually take on a
of Personality and Social Psychology, 61, 226 244.
significant care management role in helping patients Bowlby, J. (1977). The making and breaking of affectional bonds.
I. Aetiology and psychopathology in the light of attachment
with their interpersonal patterns and fears and
theory. An expanded version of the Fiftieth Maudsley Lecture,
in getting their emotional needs met clinically.
delivered before the Royal College of Psychiatrists, 19
Such approaches would likely be well received by
November 1976. British Journal of Psychiatry, 130, 201 210.
providers and health care systems, but perhaps most
Christensen, G. L., & Olson, J. C. (2002). Mapping consumers
by patients with difficulty trusting who currently mental models with ZMET. Psychology and Marketing, 19(6),
477 502.
have to resort to repeatedly changing providers,
Ciechanowski, P., Russo, J., Katon, W., Von Korff, M.,
  playing games  or choosing not to seek medical
Ludman, E., Lin, E., et al. (2004). Influence of patient
care when they most need it.
attachment style on self-care and outcomes in diabetes.
There are several limitations in this study.
Psychosomatic Medicine, 66(5), 720 728.
Based on prior observational studies that did not
Ciechanowski, P., Russo, J., Katon, W., Von Korff, M., Simon,
G., Lin, E., et al. (2006). The association of patient relation-
consistently demonstrate an association between
ship style and outcomes in collaborative care treatment for
preoccupied attachment style and diabetes self-
depression in patients with diabetes. Medical Care, 44,
care behaviors and outcomes, we did not include
283 291.
patients with this style in the current study. Future
Ciechanowski, P., Wagner, E., Schmaling, K., Schwartz, S.,
qualitative studies with similar or larger sample sizes
Williams, B., Diehr, P., et al. (2004). Community-integrated
home-based depression treatment in older adults: A rando-
might benefit from inclusion of patients with
mized controlled trial. Journal of the American Medical
preoccupied attachment style since this style is
Association, 291(13), 1569 1577.
characterized by high medical utilization, often for
Ciechanowski, P., Walker, E. A., Katon, W. J., & Russo, J.
medically unexplained physical symptoms (Ciecha-
(2002). Attachment theory: A model for health care
nowski, Walker, Katon, & Russo, 2002). Another
utilization and somatization. Psychosomatic Medicine, 64(4),
660 667.
potential limitation is our use of a self-report
Ciechanowski, P. S., Katon, W. J., Russo, J. E., & Walker, E. A.
measure that has only single-item responses. Future
(2001). The patient provider relationship: Attachment theory
studies might benefit from the use of one of a
and adherence to treatment in diabetes. American Journal of
number of other attachment style questionnaires
Psychiatry, 158(1), 29 35.
(Stein, Jacobs, Ferguson, Allen, & Fonagy, 1998)
Ciechanowski, P. S., Hirsch, I. B., & Katon, W. J. (2002).
Interpersonal predictors of HbA(1c) in patients with type 1
some of which may have stronger reliability and
diabetes. Diabetes Care, 25(4), 731 736.
validity. Future qualitative studies may also benefit
Dozier, M., Cue, K. L., & Barnett, L. (1994). Clinicians as
from use of interview assessments of attachment
caregivers: Role of attachment organization in treatment.
categories such as the Adult Attachment Interview
Journal of Consulting and Clinical Psychology, 62(4), 793 800.
(George, Kaplan, & Main, 1985).
George, C., Kaplan, N., & Main, M. (1985). The Berkeley adult
attachment interview. Unpublished protocol, University of
California, Berkeley.
Acknowledgements
Glasgow, R. E., Wagner, E. H., Schaefer, J., Mahoney, L. D.,
Reid, R. J., & Greene, S. M. (2005). Development and
Validation of the Patient Assessment of Chronic Illness Care
The authors would also like to thank Lorna
(PACIC). Medical Care, 43(5), 436 444.
Rhodes, Ph.D. for her input in the design and
Goldstein, D. E., Little, R. R., Lorenz, R. A., Malone, J. I.,
interpretation of the results of this study. The
Nathan, D., & Peterson, C. M. (1995). Tests of glycemia in
authors would also like to acknowledge the support
diabetes. Diabetes Care, 18(6), 896 909.
of Ms. Natalie Brown who served as a research Griffin, D., & Bartholomew, K. (1994). The metaphysics of
measurement: The case of adult attachment. Advances in
assistant on this study.
Personal Relationships, 5, 17 52.
Hahn, S. R., Kroenke, K., Spitzer, R. L., Brody, D., Williams, J.
B., Linzer, M., et al. (1996). The difficult patient: Prevalence,
References
psychopathology, and functional impairment. Journal of
General Internal Medicine, 11(1), 1 8.
Ainsworth, M. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Hamilton, C. E. (2000). Continuity and discontinuity of
Patterns of attachment: A psychological study of the strange attachment from infancy through adolescence. Child Devel-
situation. Hillsdale, NJ: Erlbaum. opment, 71(3), 690 694.
ARTICLE IN PRESS
P. Ciechanowski, W.J. Katon / Social Science & Medicine 63 (2006) 3067 3079 3079
Hohmann, A. A. (1999). A contextual model for clinical mental systematic review and meta-analysis of randomized controlled
health effect effectiveness research. Mental Health Services trials. Medical Care, 42(12), 1211 1221.
Research, 1(2), 83 91. Patton, M. (1990). Qualitative evaluation and research methods.
Katon, W., Von Korff, M., Lin, E., Unutzer, J., Simon, G., Newbury Park, CA: Sage Publications.
¨
Walker, E., et al. (1997). Population-based care of depression: Riggs, S. A. (2001). Adult discourse and internal working models
Effective disease management strategies to decrease preva- of attachment: Relation to history of therapy and mental
lence. General Hospital Psychiatry, 19(3), 169 178. health. Dissertation Abstracts International: Section B: The
Katon, W. J. (2003). The institute of medicine   Chasm  report: Sciences and Engineering, 61(11-B), 6167.
Implications for depression collaborative care models. Gen- Riggs, S. A., Jacobovitz, D., & Hazen, N. (2002). Adult
eral Hospital Psychiatry, 25(4), 222 229. attachment and history of psychotherapy in a normative
Koopman, C., Gore-Felton, C., Marouf, F., Butler, L. D., Field, sample. Psychotherapy: Theory, research, practice, and train-
N., Gill, M., et al. (2000). Relationships of perceived stress to ing, 39(4), 344 353.
coping, attachment and social support among HIV-positive Scharfe, J., & Bartholomew, K. (1994). Reliability and stability of
persons. AIDS Care, 12(5), 663 672. adult attachment patterns. Personal Relationships, 1, 23 43.
Lopez, F. G., Gover, M. R., Leskela, J., Sauer, E. M., Schirmer, Schmidt, S., Nachtigall, C., Wuethrich-Martone, O., & Strauss,
L., & Wyssmann, J. (1997). Attachment styles, shame, guilt, B. (2002). Attachment and coping with chronic disease.
and collaborative problem-solving orientations. Personal Journal of Psychosomatic Research, 53(3), 763 773.
Relationships, 4(2), 187 199. Stein, H., Jacobs, N. J., Ferguson, K. S., Allen, J. G., & Fonagy,
MacDonald, M. (2003). Seeing the cage: Stigma and its potential P. (1998). What do adult attachment scales measure? Bulletin
to inform the concept of the difficult patient. Clinical Nurse of the Menninger Clinic, 62, 33 82.
Specialist, 17(6), 305 310 quiz 311 302. Stewart, M. A. (1995). Effective physician patient communica-
Mathieson, C. (1999). Interviewing the ill and the healthy: tion and health outcomes: A review. Canadian Medical
Paradigm or process? In M. Murray, & K. Chamberlain Association Journal, 152(9), 1423 1433.
(Eds.), Qualitative health psychology: Theories and methods Strauss, A., & Corbin, J. (1990). Basics of qualitative research:
(pp. 117 132). London: Sage. Grounded theory procedures and techniques. Newbury Park,
Mayer, T., & Cates, R. J. (1999). Service excellence in health care. CA: Sage Publications.
Journal of the American Medical Association, 282(13), Turan, B., Osar, Z., Turan, J. M., Ilkova, H., & Damci, T. (2003).
1281 1283. Dismissing attachment and outcome in diabetes: The mediat-
Mikulincer, M., & Nachshon, O. (1991). Attachment styles and ing role of coping. Journal of Social and Clinical Psychology,
patterns of self-disclosure. Journal of Personality and Social 22, 607 626.
Psychology, 61(2), 321 331. Von Korff, M., Gruman, J., Schaefer, J., Curry, S. J., & Wagner,
Mikulincer, M., Shaver, P. R., & Pereg, D. (2003). Attachment E. H. (1997). Collaborative management of chronic illness.
theory and affect regulation: The dynamics, development, and Annals of Internal Medicine, 127(12), 1097 1102.
cognitive consequences of attachment-related strategies. Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheim,
Motivation and Emotion, 27(2), 77 102. L. (2000). Attachment security in infancy and early adult-
Mooney, K. (2002). The ten demandments: Rules to live by in the hood: A twenty-year longitudinal study. Child Development,
age of the demanding consumer. New York: McGraw-Hill. 71(3), 684 689.
Neumeyer-Gromen, A., Lampert, T., Stark, K., & Kallischnigg, Zaltman, G. (2003). How customers think: Essential insights into the
G. (2004). Disease management programs for depression: A mind of the market. Boston, MA: Harvard Business School Press.