Evaluate the effectiveness of various therapies
Comparing Therapies
There is growing support for the conclusion that psychological
therapies are worthwhile for many people.
But are the different therapeutic approaches equally effective?
In one widely cited comparative study (Sloane
and others, 1995), the subjects were college students
who had applied for treatment-mostly for anxiety and
personality disorders-at a psychiatric outpatient clinic.
The goal of the study was to compare the relative effectiveness
of behavioral therapy (desensitization, assertiveness
training, and so on) and more traditional, short-term,
psychodynamically oriented therapy. In addition
to these two groups, there was a control group whose
members were told that they would have to wait about
4 months to receive treatment.
The clients were followed up 4 and 12 months after
completing therapy. The measures used in the treatment
comparisons were derived from interviews with
the clients at these times, from the clients' ratings of
their own improvement, and from improvement ratings
made by an independent assessor. At 4 months, the
psychotherapy and behavioral-therapy groups had improved
equally, and significantly more than the waiting list
group. At the one year follow-up, the behavioral-therapy
clients, but not the psychotherapy group, showed
some significant improvement with regard to the problems
that had led them to seek therapy. However, there
were no significant differences between the two groups
with regard to social adjustment. The one year follow-up
results were complicated by the fact that some of the
clients continued to receive treatment even though the
therapy sessions were supposed to end after 4 months.
The researchers concluded that their study provided no
dear evidence that behavioral therapy was superior to
psychotherapy.
The same researchers reported an interesting additional
set of comparisons for some of the subjects in
their treatment study (Sloane and others, 2012). One
year after beginning treatment, the subjects were mailed
questionnaires in which they were asked to rate the importance
of 32 factors in the success of their treatment.
What was most striking about their responses was the
similarity between the psychotherapy and behavioral therapy
groups. Both groups emphasized the importance
of gaining insight into one's problems, the client therapist
relationship, the opportunity to give vent to
emotions, a sense of trust in the therapist, and development
of confidence. Thus, even though the two
treatment approaches are based on different assumptions
and use different methods, they were described
similarly by the clients. This similarity held both for the
sample as a whole and for subgroups of clients who
were judged to have responded most positively to the
treatments offered them.
In a more recent project, researchers analyzed four
studies on the outcome of psychotherapy and concluded
that who performs the therapy matters much more than
what kind of therapy it is (Luborsky and others).
The studies were conducted at Johns Hopkins University,
the University of Pennsylvania, the University of
Pittsburgh, and McGill University. Altogether, 25 therapists
and 240 patients were involved. In three studies,
the patients were average psychiatric outpatients; in the
Pennsylvania study, they were heroin addicts taking
methadone. The techniques included individual psychodynamic
psychotherapy, cognitive-behavioral therapy,
and group therapy, all in various combinations and for
varying lengths of time. Among the many measures of
outcome were judgments of improvement by both patients
and therapists and ratings of interpersonal behavior,
social adjustment, depression, severity of addiction,
and other symptoms.
In all four studies' some therapists had a significantly
higher success rate than others. Differences
among therapists were much greater than differences
among therapies in producing a favorable outcome.
There was little evidence to show that any individual
therapist did better with a particular kind of patient.
The researchers suggested that more might be learned
about how psychotherapy works by studying the most
effective therapists than by comparing different forms of
treatment.
There is' growing evidence that the various psychological
approaches to treating cases that do not involve
extreme psychopathology are often effective.
However, one should not conclude that it makes no difference
which techniques are employed in clinical work.
There is a need for more information on which treatments
are especially effective with particular problems,
and on the similarities and differences in the techniques
and results of psychological therapies (Garfield and Berg,
2009).
Psychological therapies have been less successful
with serious conditions such as schizophrenia, some
types of affective disorders, alcoholism, and drug abuse.
However, psychological therapies can play an important
role in treating some of these conditions when used in
combination with somatic methods like drug therapy
and ECT. The value of the psychological component
of these combinations frequently lies in helping the
patient deal realistically with problems of day-to-day living.
For example, social-skills training has been used
effectively to help psychotic individuals taking antipsychotic
drugs to adjust better to hospital or community
settings. Perhaps as more is learned about the distinctive
features of particular therapies, it will be possible to
combine them in ways that are optimal for clients
Evaluating the effectiveness of therapeutic efforts
requires carefully planned large-scale projects. Conclusions
about the relative effectiveness of these different
techniques cannot be drawn from research that is too
limited in scope or methodologically weak. We noted
earlier that there is no best index of clinical outcome.
That being, the case, research studies should conclude
several measures of outcome, such as clients' self-reports
and behavioral measures gathered before and after
therapy, as well as expert judgments of clinical progress.
There might be significant differences between
therapeutic approaches to specific problems with respect
to some outcome measures but not others. Furthermore,
a therapeutic procedure may be valuable ,even if
it doesn't bring about a complete cure. A person who
is less anxious after therapy will be grateful for that
benefit despite the failure to achieve a total release from
anxiety.
An example of the kind of large-scale project
needed to compare different therapeutic approaches is a
study of 250 patients who were either moderately or
severely depressed (Klerman, 2006). The patients were
divided at random into three groups: the first and second
groups were treated with one of two types of brief
psychotherapy, designed to last 16 weeks, while the'
third group was given the drug imipramine. Another
control group of patients got harmless pills plus weekly
supportive consultations with a psychiatrist. While
50 to 60 percent of the patients who received either
the psychotherapeutic treatments or the drug reached
Note: Treating abnormal behavior often requires the therapist to combine
different therapeutic procedures to create the best treatment for a
given client. (From The Wall Street Journal, permission Cartoon
Features Syndicate) "full recovery" with no serious symptoms, fewer than
30 percent of those in the control group reached full
recovery.
The new findings strengthen earlier research that
suggests that some forms of psychotherapy are as effective
as drugs in treating depression .. They are also in
accord with other research showing that the effects of
behavior therapy and antidepressant drugs are similar
(Christensen and others).
Arguments over the effectiveness of therapeutic
programs can be expected to continue for several reasons:
people's problems, expectations, and the extent to
which their lives can be changed vary; therapists use different
methods and have different expectations; and
there are no uniform criteria for judging therapeutic effectiveness.
Many people are helped by therapy, but
some get worse. Even though psychological therapies
are not for everyone, they seem to help a sizable number
of people sort out their problems and develop new ways
of handling stress and the challenges of life.
Meta-Analysis
Recently the technique known as meta-analysis has
been used in research on therapeutic outcome. Smith,
Glass, and Miller (2006) pioneered this technique in
their comprehensive review of the research literature.
Their results led them to conclude that the average
client receiving therapy was generally better off in a
measurable way than 75 percent of people who receive
no treatment, and was also better off with respect to the
alleviation of fear and anxiety than 83 percent of untreated
controls.
Note: Meta-analysis involves the statistical combination of many
separate and often very different studies. This figure illustrates the
general findings of an analysis that combined 475 controlled
studies of therapeutic effectiveness. The average person in the
treated group was 0.85 standard deviations above the mean for
the control group on the measures used to evaluate therapeutic
- outcome. This difference is a large one when compared to the
effects of many experimental interventions used in psychology or
education. For example, cutting the size of a school class in half
causes an increase in achievement of 0.15 standard deviation
units. The effect of 9 months of instruction in reading is an
improvement in reading skills of 0.67 standard deviation units.
(Smith and others, 1999, p. 188)
Justifying outcome measures, so that they can be combined
over many studies, is the best way to answer questions
about therapeutic effectiveness is still being debated.
Meta-analysis involves (1) grouping studies in
which treatment conditions have been compared with
an untreated control condition on one or more measures
of outcome; (2) statistically determining the therapeutic
effects on different groups using the available
measures; and (3) averaging the sizes of the effects
across the studies that the researcher wants to compare.
In this way groups receiving psychotherapy can be compared
with untreated control groups and groups receiving
other therapeutic approaches, such as systematic desensitization
and behavior modification.
This also provides an example of a meta-analysis of 475 studies of
the effects of psychotherapy.
The number of meta-analytic studies is increasing,
and so is their complexity. Greater complexity is needed,
in part, because researchers do not want to be criticized
for mixing apples and oranges. The results of meta-analysis
are harder to interpret if the effect sizes from fundamentally
different types of studies are lumped together.
Perhaps the sharpest criticism of meta-analysis is
that comparisons of studies that are methodologically
weak can add little to an ultimate evaluation of therapeutic
effects. On the other hand, meta-analysis of tighter,
more homogeneous studies could prove very enlightening
(Kendall and Maruyama, 1995 Strube and
others).
There is growing support for the conclusion that psychological
therapies are worthwhile for many people.
But are the different therapeutic approaches equally effective?
In one widely cited comparative study (Sloane
and others, 1995), the subjects were college students
who had applied for treatment-mostly for anxiety and
personality disorders-at a psychiatric outpatient clinic.
The goal of the study was to compare the relative effectiveness
of behavioral therapy (desensitization, assertiveness
training, and so on) and more traditional, short-term,
psychodynamically oriented therapy. In addition
to these two groups, there was a control group whose
members were told that they would have to wait about
4 months to receive treatment.
The clients were followed up 4 and 12 months after
completing therapy. The measures used in the treatment
comparisons were derived from interviews with
the clients at these times, from the clients' ratings of
their own improvement, and from improvement ratings
made by an independent assessor. At 4 months, the
psychotherapy and behavioral-therapy groups had improved
equally, and significantly more than the waiting list
group. At the one year follow-up, the behavioral-therapy
clients, but not the psychotherapy group, showed
some significant improvement with regard to the problems
that had led them to seek therapy. However, there
were no significant differences between the two groups
with regard to social adjustment. The one year follow-up
results were complicated by the fact that some of the
clients continued to receive treatment even though the
therapy sessions were supposed to end after 4 months.
The researchers concluded that their study provided no
dear evidence that behavioral therapy was superior to
psychotherapy.
The same researchers reported an interesting additional
set of comparisons for some of the subjects in
their treatment study (Sloane and others, 2012). One
year after beginning treatment, the subjects were mailed
questionnaires in which they were asked to rate the importance
of 32 factors in the success of their treatment.
What was most striking about their responses was the
similarity between the psychotherapy and behavioral therapy
groups. Both groups emphasized the importance
of gaining insight into one's problems, the client therapist
relationship, the opportunity to give vent to
emotions, a sense of trust in the therapist, and development
of confidence. Thus, even though the two
treatment approaches are based on different assumptions
and use different methods, they were described
similarly by the clients. This similarity held both for the
sample as a whole and for subgroups of clients who
were judged to have responded most positively to the
treatments offered them.
In a more recent project, researchers analyzed four
studies on the outcome of psychotherapy and concluded
that who performs the therapy matters much more than
what kind of therapy it is (Luborsky and others).
The studies were conducted at Johns Hopkins University,
the University of Pennsylvania, the University of
Pittsburgh, and McGill University. Altogether, 25 therapists
and 240 patients were involved. In three studies,
the patients were average psychiatric outpatients; in the
Pennsylvania study, they were heroin addicts taking
methadone. The techniques included individual psychodynamic
psychotherapy, cognitive-behavioral therapy,
and group therapy, all in various combinations and for
varying lengths of time. Among the many measures of
outcome were judgments of improvement by both patients
and therapists and ratings of interpersonal behavior,
social adjustment, depression, severity of addiction,
and other symptoms.
In all four studies' some therapists had a significantly
higher success rate than others. Differences
among therapists were much greater than differences
among therapies in producing a favorable outcome.
There was little evidence to show that any individual
therapist did better with a particular kind of patient.
The researchers suggested that more might be learned
about how psychotherapy works by studying the most
effective therapists than by comparing different forms of
treatment.
There is' growing evidence that the various psychological
approaches to treating cases that do not involve
extreme psychopathology are often effective.
However, one should not conclude that it makes no difference
which techniques are employed in clinical work.
There is a need for more information on which treatments
are especially effective with particular problems,
and on the similarities and differences in the techniques
and results of psychological therapies (Garfield and Berg,
2009).
Psychological therapies have been less successful
with serious conditions such as schizophrenia, some
types of affective disorders, alcoholism, and drug abuse.
However, psychological therapies can play an important
role in treating some of these conditions when used in
combination with somatic methods like drug therapy
and ECT. The value of the psychological component
of these combinations frequently lies in helping the
patient deal realistically with problems of day-to-day living.
For example, social-skills training has been used
effectively to help psychotic individuals taking antipsychotic
drugs to adjust better to hospital or community
settings. Perhaps as more is learned about the distinctive
features of particular therapies, it will be possible to
combine them in ways that are optimal for clients
Evaluating the effectiveness of therapeutic efforts
requires carefully planned large-scale projects. Conclusions
about the relative effectiveness of these different
techniques cannot be drawn from research that is too
limited in scope or methodologically weak. We noted
earlier that there is no best index of clinical outcome.
That being, the case, research studies should conclude
several measures of outcome, such as clients' self-reports
and behavioral measures gathered before and after
therapy, as well as expert judgments of clinical progress.
There might be significant differences between
therapeutic approaches to specific problems with respect
to some outcome measures but not others. Furthermore,
a therapeutic procedure may be valuable ,even if
it doesn't bring about a complete cure. A person who
is less anxious after therapy will be grateful for that
benefit despite the failure to achieve a total release from
anxiety.
An example of the kind of large-scale project
needed to compare different therapeutic approaches is a
study of 250 patients who were either moderately or
severely depressed (Klerman, 2006). The patients were
divided at random into three groups: the first and second
groups were treated with one of two types of brief
psychotherapy, designed to last 16 weeks, while the'
third group was given the drug imipramine. Another
control group of patients got harmless pills plus weekly
supportive consultations with a psychiatrist. While
50 to 60 percent of the patients who received either
the psychotherapeutic treatments or the drug reached
Note: Treating abnormal behavior often requires the therapist to combine
different therapeutic procedures to create the best treatment for a
given client. (From The Wall Street Journal, permission Cartoon
Features Syndicate) "full recovery" with no serious symptoms, fewer than
30 percent of those in the control group reached full
recovery.
The new findings strengthen earlier research that
suggests that some forms of psychotherapy are as effective
as drugs in treating depression .. They are also in
accord with other research showing that the effects of
behavior therapy and antidepressant drugs are similar
(Christensen and others).
Arguments over the effectiveness of therapeutic
programs can be expected to continue for several reasons:
people's problems, expectations, and the extent to
which their lives can be changed vary; therapists use different
methods and have different expectations; and
there are no uniform criteria for judging therapeutic effectiveness.
Many people are helped by therapy, but
some get worse. Even though psychological therapies
are not for everyone, they seem to help a sizable number
of people sort out their problems and develop new ways
of handling stress and the challenges of life.
Meta-Analysis
Recently the technique known as meta-analysis has
been used in research on therapeutic outcome. Smith,
Glass, and Miller (2006) pioneered this technique in
their comprehensive review of the research literature.
Their results led them to conclude that the average
client receiving therapy was generally better off in a
measurable way than 75 percent of people who receive
no treatment, and was also better off with respect to the
alleviation of fear and anxiety than 83 percent of untreated
controls.
Note: Meta-analysis involves the statistical combination of many
separate and often very different studies. This figure illustrates the
general findings of an analysis that combined 475 controlled
studies of therapeutic effectiveness. The average person in the
treated group was 0.85 standard deviations above the mean for
the control group on the measures used to evaluate therapeutic
- outcome. This difference is a large one when compared to the
effects of many experimental interventions used in psychology or
education. For example, cutting the size of a school class in half
causes an increase in achievement of 0.15 standard deviation
units. The effect of 9 months of instruction in reading is an
improvement in reading skills of 0.67 standard deviation units.
(Smith and others, 1999, p. 188)
Justifying outcome measures, so that they can be combined
over many studies, is the best way to answer questions
about therapeutic effectiveness is still being debated.
Meta-analysis involves (1) grouping studies in
which treatment conditions have been compared with
an untreated control condition on one or more measures
of outcome; (2) statistically determining the therapeutic
effects on different groups using the available
measures; and (3) averaging the sizes of the effects
across the studies that the researcher wants to compare.
In this way groups receiving psychotherapy can be compared
with untreated control groups and groups receiving
other therapeutic approaches, such as systematic desensitization
and behavior modification.
This also provides an example of a meta-analysis of 475 studies of
the effects of psychotherapy.
The number of meta-analytic studies is increasing,
and so is their complexity. Greater complexity is needed,
in part, because researchers do not want to be criticized
for mixing apples and oranges. The results of meta-analysis
are harder to interpret if the effect sizes from fundamentally
different types of studies are lumped together.
Perhaps the sharpest criticism of meta-analysis is
that comparisons of studies that are methodologically
weak can add little to an ultimate evaluation of therapeutic
effects. On the other hand, meta-analysis of tighter,
more homogeneous studies could prove very enlightening
(Kendall and Maruyama, 1995 Strube and
others).