What Works for Whom?
A Critical Review of Psychotherapy Research
By Anthony Roth Peter Fonagy
The Guilford Press
Copyright © 2005
The Guilford Press
All right reserved.
ISBN: 1-57230-650-5
Chapter One
DEPRESSION
DEFINITIONS
DSM-IV-TR describes a number of subcategories of depression; those particularly
relevant to research studies are defined as follows (adapted from
Wells, 1985).
Major Depressive Disorder
Major depressive disorder (MDD) is characterized by one or more major
depressive episodes and the absence of manic episodes. A major depressive
episode is defined by depressive mood or loss of interest or pleasure in almost
all usual activities, accompanied by other depressive symptoms. These include
disturbances in appetite, weight, and sleep; psychomotor agitation or retardation;
decreased energy; feelings of worthlessness or guilt; difficulty concentrating
or thinking; and thoughts of death or suicide, or suicidal attempts.
DSM-IV-TR specifies that at least five of nine specific depressive symptoms
must be present nearly every day for at least 2 weeks to make a diagnosis of
MDD, and that the symptoms cause clinically significant distress or impairment
in social, occupational, or other important areas of functioning.
Depressive episodes are distinguished from normal bereavement reactions.
Dysthymic Disorder
This disorder is characterized by depressed mood or loss of interest in nearly
all usual activities, though symptom severity is not sufficient to meet the criteria
for MDD. The disorder is, by definition, chronic. Symptoms should be
present for at least 2 years, and a diagnosis cannot be made if patients are
symptom-free for more than 2 months in any 2-year period. It is characterized
by depressed mood for most of the day, together with at least two of the
following six symptoms: poor appetite, insomnia or hypersomnia, low
energy, low self-esteem, poor concentration, and feelings of hopelessness. For
diagnostic purposes, these symptoms should be severe enough to cause clinically
significant distress or impairment in social, occupational, or other areas
of functioning.
"Double Depression"
Patients with dysthymic disorder frequently present with a superimposed
MDD; this is usually referred to as "double depression."
PREVALENCE AND NATURAL HISTORY
Prevalence
Only a portion of individuals with mental health problems present to family
physicians or mental health professionals (e.g., Bebbington et al., 2000a;
Goldberg & Huxley, 1980). Because of this, estimating treatment need is
better done through community-based surveys rather than relying on data
from clinical services. Two large-scale community surveys provide data
on the prevalence of psychiatric disorders in the United States. The
National Institute of Mental Health (NIMH) Epidemiologic Catchment Area
(ECA) program was a five-site project sampling approximately 20,000 adults
(Robins & Regier, 1991). The National Comorbidity Survey (NCS; Blazer
et al., 1994) had a slightly more restricted age range, and interviewed approximately
8,000 adults between ages 15 and 54. The prevalence rates derived
from these surveys need to be interpreted cautiously; for example, there is a
risk that they are inflated by individuals whose distress is transient. Deriving a
"correct" prevalence rate that accounts for the clinical significance of symptoms
is difficult and controversial. Narrow et al. (2002) have recomputed
prevalence rates from the ECA and NCS surveys, taking into account the
degree to which symptoms resulted in help-seeking behavior and led to significant
levels of distress (see Table 4.1). In addition, they attempted to reconcile
differences in prevalence rates between the surveys, some of which
relate to methodological differences. Their approach has been criticized as
inappropriately robust (Wakefield & Spitzer, 2002), and it is clear, that while
presentation to services appears to be linked to the severity of symptoms
(Bebbington et al., 2000b), lack of help seeking cannot be assumed to indicate
that distress is unimportant. Nonetheless, the revised rates are cited here
(and in Appendix III), since they probably yield a more accurate indicator of
service need.
The ECA and NCS estimate 1-year prevalence for MDD at 5.4% and
8.9%; corrected for clinical significance, these figures lower to 4.6% and
5.4%, respectively. Narrow et al.'s (2002) estimate, which combines data
from both surveys, is 4.5%. For dysthymic disorder, the ECA and NCS estimates
are 5.7% and 2.5%, respectively; with correction for clinical significance,
these reduce to 1.7% and 1.8%, respectively; the combined estimate is
1.6%. Other reviews derive somewhat different (uncorrected) rates. Angst
(1992), reviewing 17 studies, suggests that 1-year prevalence rates for MDD
lie between 2.6 and 6.2%, and for dysthymic disorder, between 2.3 and 3.7%.
Lifetime prevalence rates vary between 4.4 and 19.5%. Angst also reports data
from a Swiss prospective community survey carried out (to date) over 10
years. This was based on multiple interviews and hence avoided problems of
estimating prevalence based on recall. Lifetime prevalence to age 30 of MDD
was 14.5%, with around half of affected individuals seeking treatment.
The prevalence of depression varies by gender and age; prevalence of
MDD in the ECA and NCS was almost twice as high in women as men, and
greater in younger adults. In part, this may reflect the greater willingness of
younger adults to admit to mental health problems (Taube & Barrett, 1985;
Weissman et al., 1988), or problems of recall when older respondents are
interviewed in cross-sectional surveys (Fombonne, 1994). However, there is
evidence that prevalence within younger age groups is increasing (Burke et
al., 1991), though the degree to which this is associated with comorbid substance
abuse is unclear. Furthermore, there is some agreement that overall
rates of depression are increasing (Fombonne, 1998b; Klerman & Weissman,
1989).
Natural History
Most studies of "natural" history monitor longitudinal outcomes for patients
offered "treatment as usual" (TAU). Over a 2-year period, Wells et al. (1992)
followed up 626 outpatients; the sample included patients diagnosed with
MDD, dysthymic disorder, and double depression, and also contained clients
with subthreshold depressive symptoms. Patients with MDD had a 42%
probability of remission in the first year, and a 60% probability of remission
in the second year, if none had occurred in the first year. Clients with double
depression had a rather different course, depending on the severity of their
symptoms. Those with more severe symptoms had a 37% likelihood of
remission in the first year; if no remission occurred by this point, there was
only a 16% probability of remission in the second year. Both dysthymic
patients and those with subthreshold symptoms of depression were at considerable
risk of suffering an episode of MDD over the study period. Half the
patients with an initial diagnosis of dysthymic disorder and 25% of patients
with subthreshold symptoms of depression (with or without a prior history of
depression) experienced an episode of MDD over the 2-year period. Data
from patient samples in field trials for DSM-IV-TR confirm this pattern;
79% of patients with dysthymic disorder eventually developed MDD
(McCullough et al., 1992). The poorest clinical outcomes were found in
patients with double depression; there was a particularly low rate of remission
in patients with both double depression and high initial symptom severity.
Patients with dysthymic disorder (even in the absence of MDD) had higher
levels of depressive symptoms over the 2-year period of the study than
patients with MDD alone, despite the fact that dysthymic disorder is defined
by the presence of less severe (if persistent) depressive symptoms. In addition,
patients with dysthymic disorder were rated as having poorer social and emotional
functioning than those with MDD.
Keller and Shapiro (1982) and Keller et al. (1983) suggest that patients
with double depression tend to have a shorter episode of MDD but are also
likely to relapse more quickly than those with MDD alone. Double depressives
appear to have a faster "cycle time"; over a 2-year period, 62% of them
had completed a cycle of recovery and relapse, compared to 33% of the
MDD group.
Long-term monitoring confirms a pattern of vulnerability to relapse for
people with MDD. Piccinelli and Wilkinson (1994) reviewed 50 naturalistic
follow-up studies of in- and outpatients with unipolar depression, carried out
between 1970 and 1993. Although recovery rates increase over time (on
average, 53% of patients will recover at least briefly by 6 months), one-fourth
of the patients will have suffered a recurrence of the index episode within 1
year. Seventy-five percent of patients followed up for 10 years suffered a further
episode of depression, and 10% of patients suffered persistent depression.
Mueller et al. (1999) followed up patients over 15 years; all were in receipt of
TAU. Of 380 patients who had recovered from an index episode of MDD, a
cumulative proportion of 85% relapsed over this period. Of a further 105
patients who had recovered and remained well over 5 years, a cumulative
proportion of 58% relapsed. Though there were indications that TAU
included suboptimal delivery of medication, there was little information
available regarding the use of psychosocial interventions. Demographic or
clinical characteristics did not predict relapse, though there were indications
that individuals who had recovered but continued to experience subthreshold
symptoms were particularly vulnerable, a pattern found in other studies (e.g.,
Judd et al., 1998b).
Summary
Studies of the prevalence and natural history of depression have a number of
implications for research. Although precise estimation is complicated, it is
clear that depression is a relatively common syndrome affecting at least 4.5%
of the population, with prevalence among women about double that among
men. The course of depression appears to differ according to subtype (MDD,
dysthymic disorder, or double depression). It is likely that 80% of patients
with dysthymic disorder will eventually develop an MDD, suggesting that
dysthymic disorder and acute depression are variants of a similar condition.
Relapse is a serious challenge: 85% of patients followed up over 15 years, and
75% of patients followed up over 10 years will have suffered a further episode
of MDD, and 10% of these will have endured persistent depression. The
probability of relapse is increased in patients with more than three previous
episodes of MDD but is greatest in patients with a diagnosis of dysthymic disorder;
these patients show a faster cycle of recovery and relapse than those
with MDD alone. Even among those patients who have "recovered,"
subthreshold symptoms are common and are associated with an increased
likelihood of relapse.
The risk-indeed, the probability-of relapse has obvious implications
for treatment trials. The effectiveness of a treatment needs to be judged not
only by its capacity to manage an index episode but also by its ability to
maintain remission. This poses a challenge, in part, because on the basis of
figures given above, long-term follow-up of at least 2 years would be necessary
to provide a conclusive result that is not confounded with the natural
history of this disorder. It is also likely that outcome in clinical trials will be
influenced by case mix, and particularly by the presence of patients with double
depression or a history of recurrent MDD. Because of the exclusion criteria
applied in at least some research trials, it is possible that the clinical population
will contain comparatively more patients with chronic depression and
dysthymic disorder. This may lead to overestimation of treatment effects;
poorer outcomes might be expected in clinical practice than in trials. However,
as an increasing number of studies concern themselves with "treatment
resistant" patients, this may be a less pertinent issue than before.
LANDMARK STUDIES OF EFFICACY
Subsequent chapters review individual studies in the context of meta-analyses
and qualitative reviews. This chapter adopts a different strategy, describing in
some detail a small number of high-quality individual studies that help to
contextualize the broader body of evidence. These trials give indications of
the acute and longer term efficacy of the major treatment approaches in this
area (cognitive-behavioral therapy, interpersonal psychotherapy, short-term
psychodynamic therapy, and medication), and of the challenge posed by
relapse.
Cognitive-Behavioral Therapy and Interpersonal Psychotherapy:
NIMH Treatment of Depression Collaborative Research Program
This major and widely cited research program (summarized in Elkin, 1994)
set a standard against which other studies can be judged. Patients were randomized
to receive one of four interventions: cognitive-behavioral therapy
(CBT; Beck et al., 1979), interpersonal psychotherapy (IPT; Klerman et al.,
1984), imipramine plus clinical management (IMI-CM), or placebo plus clinical
management (PLA-CM). Clinical management consisted of a weekly
meeting of 20-30 minutes to discuss medication, side effects, and the
patient's clinical status. In addition, and where necessary, support, encouragement,
and direct advice were also offered. On this basis, it is worth noting
that both medication conditions contained psychotherapeutic elements. This
research design has been misinterpreted as a test of therapy against medication
(Elkin, 1994); more accurately, the intention was to use medication condition
as a "benchmark" against which to compare the psychological therapies.
The study was carried out at three research sites in the United States.
Five hundred sixty outpatients were initially screened, essentially ensuring
that patients met criteria for a DSM-III-R diagnosis of unipolar depression.
Two hundred fifty patients, all moderately to severely depressed, were
selected for the trial; 239 actually entered it. Of these, 60% had been
depressed for more than 6 months; for only 36% was this a first episode of
depression.
Treatments were carried out by experienced therapists (10 each in IPT
and pharmacotherapy, eight in CBT) chosen for their expertise in applying
their respective therapy and supervised regularly throughout the clinical trial.
To ensure that therapies were conducted as intended, sessions were taped,
and the process of therapy was checked against measures of therapy adherence.
Though there were some differences in attrition from each condition,
these were not statistically significant. Rates of dropout across treatment
modalities were as follows: 23% (n = 14) for IPT, 32% (n = 19) for CBT,
33% (n = 19) for IMI-CM, and 40% (n = 25) for PLA-CM.
Continues...
Excerpted from What Works for Whom?
by Anthony Roth Peter Fonagy
Copyright © 2005 by The Guilford Press.
Excerpted by permission.
All rights reserved. No part of this excerpt may be reproduced or reprinted without permission in writing from the publisher.
Excerpts are provided by Dial-A-Book Inc. solely for the personal use of visitors to this web site.