Essential Reseach Findings in Counselling and Plsyh oth ecayy – Mihk Cooyec
C aythec 2: T e outhhomes of hounselling and ysyh oth ecayy
This chapter discusses:
- The evidence that counselling and psychotherapy can have a benefcial eeect
- The amount of change that therapy can bring about
- The percentage of clients who improve in therapy
- Whether or not some people deteriorate in therapy
- Whether or not therapy brings about change in ‘real world’ setngs
- The relatonship between how much people change and how much therapy they have
- The long-term eeects of therapy
- The relatve eeectveness of therapy versus medicaton
- The cost-eeectveness of therapy
- How researchers study, and report, therapeutc outcomes
When comparing client’s levels of wellbeing before and afer therapy on some indicator of psychological
distress it is found (again and again) that clients are beter of afer therapy than before it. You can either
assess changes in external characteristcs or assess changes in subjectve experiences (for instance with
psychological questonnaires). Studies that use questonnaires that measure global and specifc psychological
distress consistently show that clients rate themselves on average as less psychologically distressed afer
therapy as compared with before it.
T e ‘efhahy’ of hounselling and ysyh oth ecayy
On pre- and post-therapy measures alone, we cannot be sure that the counselling or psychotherapy was
actually responsible for the changes that came about (spontaneous recovery). So a control group, ofen in the
form of waitng for therapy or receiving treatment as usual is necessary to be fairly certain that the therapy,
and no other factor, is responsible for the changes.
When controlled for other factors the same fndings are there, but to a slightly lesser degree. Doing research
with a control group does seem to suggest that the psychological therapy is responsible for bringing about
positve change. But how do we know that this is really due to the ‘aactve ingredients’ of the therapy, and not
simply the fact that someone expects to be changed or improved by it? To take also a possible placebo efect in
consideraton, researchers have not only compared changes in therapy with changes in non-therapeutc control
groups, but also with changes in placebo conditons: generally some kind of supportve, listening-based
befriending, which clients are encouraged to perceive as a genuine therapeutc treatment.
Research shows us that such minimal, placebo interventons do actually bring about quite positve results as
compared with an entrely no-treatment control; but at the same tme, an actve therapeutc interventon stll
tends to do beter. This means, assuming that the clients do perceive the placebo as an equally credible
treatment, is that the efcacy of therapy cannot just be put down to clients believing that therapy will be
helpful to them or expectng to change.
Box 2.1 Randomized controlled trials
A study which uses a randomizaton procedure to allocate partcipants to the diferent conditons is called a
randomized controlled trial (RCT). This is considered by many as the ‘agold standard’ of research.
How muh efehth does hounselling and ysyh oth ecayy aveᝐ
What is essental to ask, is not just whether counselling and psychotherapy have a positve efect, but how
much positve efect it has. To report this, researchers use the efect size which is an expression of the strength
of the relatonship between two variables.
In a meta-analyses researchers combine fndings from lots of studies using a standardized measure of the size
of a relatonship like Cohen’s d. Meta-analyses draw conclusions from an extensive body of data and can be
considered one of the most reliable sources of informaton.
Summaries of these meta-analyses suggest that the average efect size for counselling and psychotherapeutc
practces, compared with a no-treatment control, is somewhere around 0.75 to 0.e5 (so counselling and
psychotherapy has a large efect).
,Another way of conceptualizing what a Cohen’s d of 0.e0 means is by convertng it into percentages. More
specifcally, a Cohen’s d of 0.e means that around 79 per cent of clients who have had therapy do beter than
the average person who has not had therapy.
Box 2.3 Eeect size statstcs
Cohen’s d is the most common used efect size.
- 0.2 = small efect
- 0.5 = medium efect
- 0.e = large efect
Clinical change
Improvements associated with counselling and psychotherapy can also be expressed in terms of ‘clinically
signifcant improvement’. This is the number of clients who have moved from levels of abnormally high
psychological distress (as defned by diagnostc criteria or by heightened scores on a measure of psychological
disorder) to levels that are within a normal range. It is important to look at this because, even if efect sizes are
very large, if people are coming into therapy with extremely high levels of psychological distress, then they may
stll be leaving it with higher than desired levels of distress.
Summarizing clinical trials, it is found that around 60 percent of clients in psychotherapy improved to an extent
that was clinically signifcant.
Change in the real world: the eeectveness of therapy
The RCTs are done in controlled circumstances. But in the real-world this is diferent, practtoners carry out an
idiosyncratc mixture of therapeutc strategies, with clients who generally experience a diverse range of
psychological difcultes. As well as knowing what diference therapy can make (its efcacy), it is also important
to establish what actual diference it makes (its eeectveness).
Studies of therapy in ‘aclinically representatve’ (i.e. real-world) conditons generally indicate that it is highly
efectve. In an extensive meta-analysis it is found that therapy in real-world conditons was no less efectve
than it is under more controlled conditons.
Clients’ self-reports of how helpful they fnd therapy also point towards its real-world efectveness. Most
studies show that the majority of clients report they experienced the therapy benefcial.
Do some yeoyle geth wocse in th ecayyᝐ
While most people improve as a consequence of counselling and psychotherapy, there is a signifcant minority
who do not, and the evidence indicates that therapists tend to underestmate this possibility (5-10% up to 10-
15% in abuse work). Also 20% of clients indicated there was something in their therapy that was harmful of
problematc; and ultmately half of patents will drop out of therapy.
Box 2.4 The process of therapy: drop-out
A client can be defned as dropping out when they fail to atend a last scheduled visit, or when they withdraw
from therapy before the therapist thinks advisable. Research indicates about half of all clients drop out, with
20-57 per cent of outpatents in mental health agencies dropping out afer a frst session.
Clients who drop out partcularly in the frst few sessions do tend to have poorer outcomes and are less
satsfed with therapy. Research indicates that one-third drop out because they are dissatsfed with their
therapist or with the progress they are making; with a another third dropping out because they feel that their
problems are sufciently ameliorated, and a fnal third dropping out because of environmental constraints.
How muh th ecayy do hlienths needᝐ T e dose-efehth celations iy
Although researchers cannot predict exactly how much therapy each individual client will need to improve,
what they can do is to calculate how much therapy is needed for a certain percentage of clients (normally used
is 50, half of the client group). This is known as the median efectve dose of ED50. Studies of this type suggest
that the number of therapy sessions that is required before 50 of clients show clinical improvement is between
10 and 20. For 75 per cent of clients to clinically improve: 5e. But of course this depends on the kinds of
problems and symptoms being looked ad.
Research into the dose-eeect relatonship makes it clear that the more therapy clients have, the beter they
tend to get, and this has been shown in both efcacy and the efectveness. However, this relatonship is not
linear but negatvely acceleratng (law of diminishing return); as clients have more and more sessions, the
added beneft of each session becomes less and less.
1
,Research on the dose-efect relatonship has important implicatons for the debate around short-term versus
long-term therapy. Clients who get long-term treatment will tend to do beter. At the same tme, the law of
diminishing returns means that short-term therapies can be viewed as relatvely more efectve than long-term
ones: for instance, clients who have eight sessions of therapy will do more than half as well as clients who have
sixteen sessions. So when resources are scare, providing more clients with fewer sessions is likely to produce
greater overall benefts than providing fewer clients with more sessions. Also, there is some evidence to
suggest that change occurs more rapidly when tghter tme limits are imposed.
Box 2.5 The process of counselling and psychotherapy: sudden gains
An interestng discovery is that around 40 percent of clients have more than half of their total symptom
improvement concentrated in one between-session interval: sudden gains. It is not clear what brings sudden
gains about, although in CBT they seem to be preceded by sessions in which there are substantal cognitve
changes, and in psychodynamic therapy by a greater number of accurate interpretatons and a beter
therapeutc relatonship.
W ath ayyens afec th ecayyᝐ
On average, clients do not tend to improve one the therapy is over (a hypothesized ‘asleeper efect’), but equally
tend do not to deteriorate rapidly. Rather, therapy gains tend to be maintained at follow-up, though there is
some evidence of a decrease in improvement as tme goes on. Across diferent therapies a drop in efect size
from 0.6e to 0.55 has been found. In general, how well clients are doing at follow-up is strongly predicted by
how well they are doing at the end of therapy.
A limitaton is that the tmespan of the follow-up measurement is limited due to methodological and ethical
constraints.
Box 2.6 The process of therapy: improvement predicts improvement, deterioraton predicts deterioraton
Research on the long-term outcomes of counselling and psychotherapy suggest that clients who do well at the
end of therapy tend to do well at follow-up, while those who make few gains during therapy tend to show litle
improvement further down the line. This is consistent with the research which indicates that clients who do
well at the start of therapy tend to keep on doing well, while those who show an inital poor response to
therapy are more likely to deteriorate. If a clients’ symptoms are getng worse, this is a strong predictor that
they will contnue to get worse or drop out of therapy.
T ecayy vecsus medihation
While some studies suggest that drugs produce a faster inital response, by the end of treatment psychological
therapies are usually as efcacious as pharmacological therapies and have a more enduring efect.
The willingness of individuals to partcipate in verbal therapies is bigger compared to pharmacological
treatments.
Findings on psychological therapy compared to psychological therapy plus medicaton are mixed. However, it
should be noted that there is no strong evidence to suggest that pharmacological treatments reduce the
efcacy of psychological therapies.
Costh-efehtiveness
Researchers have tended to compare the cost of the talking therapies with the cost of those interventons
(such as medicaton or hospitalizaton) that are necessary if clients do not use counselling or psychotherapy
services. It has been found that psychotherapy can make a substantal reducton to the utlizaton of medical
care. A meta-analysis has shown psychotherapy to be economically benefcial compared to a no-therapy
control.
Given that the greatest costs for those who do not receive psychotherapy tend to be in the form of increased
hospitalizaton, cost-savings in inpatent setngs tends to be greater than in outpatent setngs. Along these
lines, cost-savings are also partcularly high when behavioural and psycho-educatonal procedures are used in
medical setngs: for instance relaxaton training for patents awaitng surgery. Because of the savings from
reduced hospitalizaton, counselling and psychotherapy seem to be most efectve with older clients, and also
those who experience the highest levels of psychological distress. For those psychological difcultes as
afectve disorder (where hospitalizaton is a less likely outcome), psychotherapy and counselling appear to be
about cost-neutral.
Most fndings take only into account the direct costs of psychological illness, and do not consider the wider,
less tangible costs of mental distress on a country’s social and economic wellbeing. Furthermore, such
2
, calculatons do not take into account the beneft for the individual or society in terms of increased happiness or
social wellbeing.
3
