|
Does Psychotherapy REALLY Work?
Effectiveness Studies / Methods
Predicting Treatment Outcome
As pointed out numerous times on this website, the treatment model a therapist uses predicts very little if any of the variance in treatment outcome. So what does predict success? As we argued in The Heroic Client, the response of the client in the first handful of visits with a clinician is among the most robust predictors of eventual success. Now, another study shows that by week 4, investigators using simple outcome measures can predict responders and nonresponders with a whopping 80% accuracy regardless of the diagnosis, chronicity of the client, or type of therapy offered!
You can begin tracking the outcome and process of your clinical work using simple measures developed by the members of ISTC. Click here for more information.
Crits-Christolph, P. et al. Early improvement during manual guided cognitive and dynamic psychotherapies predicts 16 week remission status. Journal of Psychotherapy Practice and Research, 10, 145-54
Does Therapy Work?
In the last update of the ISTC website, we indicated that anyone who questioned whether psychotherapy worked should be told that the research says, "Shut Up!" In that same vein, consider that available data indicate that the outcomes from the treatment of substance abuse compare favorably to those of diabetes, hypertension, and asthma. You don't hear anyone saying that payment should be denied for the treatment of these conditions do you?
McCann, B. (July/August, 2001). Incorporating advances in substance abuse rsearch. EAPA Exchange, 15-17.
Therapy Doesn't Work? Nonsense!
Hey, if bean counter says to you, "Therapy doesn't work anyway. All you research was done in labs" just tell them that the research literature says, "Shut Up!" A recent meta-analytic study of psychotherapy practiced under real world conditions across a broad range of clinical settings found effect sizes similar to those reported in earlier studies (mean d = .41). Also like earlier studies, greater number of sessions (up to a point) were correlated with a greater positive clinical effect.
Shadish, W. (2000). The effects of psychological therapies under clinically representative conditions. Psychological Bulletin, 126, 512-29.
The Definitive Proof: Yeah it Works (and it saves a bundle of money too)!
Way back in the pre-managed care days of the 1980's, psychologist/entrepreneur Nicholas Cummings published results which showed that psychotherapy offset medical expenses. This large-scale, meta-analytic study provides the definitive proof that psychological services decrease medical expenses. Analyzing studies from 1967 through 1997, researchers Chiles et al. found that a variety of psychological services (e.g., therapy, behavioral medicine, and psychiatric consultation) decreased medical expenses in patients undergoing surgery and those with a history of over utilization. The average savings was 20%–even when the cost of providing the services was subtracted from the savings!
Chiles, J. et al. (1999). The impact of psychological interventions on medical cost offset: a meta-analytic review. Clinical Psychology, 6(2), 204-220.
Therapy shown "cost-effective" . . . again (and not just CBT either)!
Researchers from Britain have provided additional evidence for the cost-effectiveness of psychological intervention–oh yeah, and not just of the cognitive-behavior variety! Indeed, these researchers found that a mere eight sessions of psychodynamic-interpersonal therapy resulted in significant reductions in inpatient stays, consultations with primary care physicians, nursing visits, use of all medications, and care provided by relatives. This were no run of the mill, garden variety, walking wounded types of clients either. Rather, those included in the study had been in regular medical treatment for "nonpsychotic psychiatric illness" an average of 10 years! The additional cost of the talk therapy was offset by significant reductions in overall costs ($2115 versus $2600).
Guthrie, E. et al. (1999). Cost-effectiveness of brief psychodynamic-interpersonal therapy in high utilizers of psychiatric services. Archives of General Psychiatry, 56, 19-26.
Therapy for Schizophrenia . . . NOT an oxymoron!
For whatever reason, the talk therapy world has largely abandoned the treatment of folks diagnosed as schizophrenia. Still laboring under a collective fear of having someone point the finger and say, "Oh yeah, you guys were the ones that blamed mothers for schizophrenia!" the field has shrunk away from clear and convincing data that the best treatment for this devastating problem is a caring relationship with a therapist! Yup, that's right! A relationship with a talk-style therapist. Instead, talk therapists have turned over people and families suffering from this terrible condition to the pharmaceutical companies. Think about it for a minute. If neuroleptic drugs offer the best chances of someone not suffering from terrifying delusions and hallucinations, then why are those with the condition so unwilling to take the medication as prescribed?
So what do the data say? Rather than focusing on the disease–as does the medical model–emphasis should be placed on the potential for growth in the individual. How could this be? Take, for example, the research of Courtenay Harding, Ph.D. This researcher followed "back ward schizophrenics" (those deemed unable to love outside of the hospital) for some 30 years following their discharge from state hospitals in Vermont in the late 1950's and early 60's. In contrast to the accepted rate of 10% "recovery rate" generally reported by the psychiatric profession, Dr. Harding found that 2/3's (66%) of these people, "showed no signs of at all of schizophrenia" (p. 27). What accounted for this success? To find out, Dr. Harding studied a similar group released from state hospitals in Maine. In contrast to the state of Vermont where emphasis was on integration into the community, skill building, and social support, treatment of those people released in Maine focused on education and medication. In the latter state, 48% of those followed over 30 years showed no signs of the diagnosis–an astounding figure still when compared to the recovery rates touted by the medical profession.
Perhaps, as finish research Jaako Seikkula argues, the development of chronic, unremitting schizophrenia is a result of the type of treatment received rather than some mysterious genetic or environmental factor. In his own research in Finland, they have found that actively involving the client and family in every stage of treatment planning, de-emphasizing medication, and providing support in the community significantly decreased if not eliminates altogether chronic schizophrenia. Perhaps this explains why research on the disorder continues to languish, mired as it is at looking "inside" sufferers rather than at beliefs and practices which continue to dominate the field.
McGuire, P.A. (2000). New hope for people with schizophrenia. APA Monitor, 31(2), 24-28.
Seikkula, J. (1996). Treating psychosis in western Lapland. ISM Skriftserie. (For a copy of this research, click here to send an email to Dr. Seikkula).
Efficacy and Effectiveness Studies: Don't Get the Two Confused!
Recent publications have begun stressing a distinction between two different types of research on psychotherapy: Effectiveness and Efficacy Studies. Most therapists are familiar with the latter group. These are the studies that everyone learned about in graduate school. Generally, they are conducted in a laboratory setting with strict controls (e.g., treatment manuals, single diagnosis, short duration of treatment). For years, these studies have been used as evidence that psychotherapy works. More recently, however, several researchers have pointed out that such studies provide NO evidence that treatment actually works in the real world–hence the new designation, "effectiveness studies." For the practicing therapist, this group of studies is the most important. Several such studies have been reviewed previously on this website. None of these studies finds any evidence that specific methods of treatment produce superior results but continue to indicate that the average treated client is better off than those without therapy.
Seligman, M (May, 1998). The effectiveness of therapy. APA Monitor, 2.
Who should gather intake and assessment information (or better what)?
For most therapists, gathering personal information about clients and establishing a DSM diagnosis are routine activities usually completed in the initial session. Building on similar findings in the past, a new study finds that computer-administered assessments may be better than face-to-face interviews with a mental health expert. Researchers Kobak et al. found that people were more likely to report having mental and emotional difficulties to a computer than to a professional–in the case of alcohol abuse, twice as likely and with obsessive-compulsive problems, three times as likely. Not only this, but the program can be administered over the phone and, because there are no written forms, to virtually anyone regardless of reading ability.
Kobak, K. et al (September, 1997). Journal of the American Medical Association, 278(11), 905-910.
Does Treatment Work for Men who Batter? Absolutely!
The treatment of men who batter their female partners always seems to generate controversy. Some claim that such programs actually make men worse, others that there is no effect al all. However, results reported at the International Conference on Family Violence show overwhelmingly that treatment works! Researchers found that 60% of men who received treatment did not re-offend during a 30-month follow-up period. Women also gave the programs high marks with nearly 70% reporting their lives to be better after their partners went through treatment. In addition, 83% of these same women reported that they felt "very safe" following their partner's participation in the program. Perhaps it's time to de-politicize this important and use the results to make a difference in the lives of men and women.
Peterson, K. (July 27, 1988). Programs help men unlearn violence. USA Today, p. 1.
Research has now established that psychotherapy works–at least in well-controlled, research settings. More recently, researchers have sought to determine whether treatment is actually helpful in clinically representative conditions. In other words, we know it works in theory, but does it work in practice? One of the first studies to review research from actual treatment settings was not encouraging. Weisz et al. (1992) found, "research focusing on a more representative treatment of referred clients in clinics has shown . . . modest effects; in fact, most clinic studies have not shown significant effects" (p. 1578, 1992). Researchers Shadish et al. (1997) and found that: (1) relatively few studies of clinic therapy exist in the therapy research literature; and (2) that the best estimate of the overall effect size for therapy conducted in a clinical setting is .59. According to the authors, the results provide "tentative" evidence that "clients who receive therapy under clinical conditions do better than those who do not receive therapy, and that the effect sizes for clinic therapy may be comparable with those yielded from past meta-analyses" (p. 361).
Shadish, W.R. et al. (1997). Evidence that therapy works in clinically representative conditions. Journal of Consulting and Clinical Psychology, 65(7), 355-365.
Psychotherapy Not Only Works But Saves Money At The Same Time!
The data are clear: psychotherapy services save money by off-setting healthcare expenditures, reducing employee absenteeism, and increasing productivity. This is, of course, not exactly, BIG news. Researcher and former APA president Nick Cummings published several studies back in the mid-eighties demonstrating that psychotherapy resulted in significant reductions in healthcare expenditures. In this study, however, researchers Gabbard, Lazar, Hernberger, and Spiegel reviewed 18 studies of different therapeutic modalities treating a variety of presenting problems and found that 90% resulted in significant cost savings. Some of the most significant savings involved the use of family therapy with people suffering from affective and personality disorders (i.e., borderline personality disorder). Importantly, the cost savings were often not related in any obvious way to the client's presenting complaint but rather reflected in general, across-the-board, savings in healthcare expenditures.
Gabbard, G.O., Lazar, S.G., Hornberger, J., and Spiegel, D. (February, 1997). The economic impact of psychotherapy: A review. American Journal of Psychiatry, 154(2), 147-155.
When Does Psychotherapy Work, When Doesn't Psychotherapy Work
In this award winning study, researcher Marnie Rice (1997) reports on the outcome of an ambitious therapeutic community treatment program initiated at the maximum security Oak Ridge Division of the Mental Health Centre in Ontario, Canada. Hailed by a blue-ribbon panel as "an exciting program [where] the impossible is apparently happening–psychopaths are being treated with success . . . Results of the program indicate a very low recidivism rate," the program was written about and copied extensively as the model for treatment of the very challenging population of violent offenders. However, when the program was subjected to rigorous empirical evaluation, the rhetoric fell far short of reality. To begin, the overall violent recidivism rate was 40%. Worse yet, when the researchers examined the effects of the treatment program on the target population–violent psychopaths (e.g., antisocial personality disorder)–they found that those who underwent the treatment had significantly worse outcomes than the untreated psychopaths. Indeed, the research shows that the treatment program effectively increased the self-esteem and empathic ability of violent psychopathic offenders thereby making it easier for them to manipulate, deceive, and consequently abuse others. In the words of the researchers, "there may be little wrong with psychopaths for therapy to fix" (p. 421). The researcher has used this information to develop an instrument that may be used to predict which offenders are most likely to re-offend in violent ways. In the meantime, they point out that programs which enhance efficacy, self-esteem, empathic ability/understanding are strongly contra-indicated in the treatment of this population.
Rice, M. (1997). Violent offender research and implications for the criminal justice system. American Psychologist, 52(4), 414-423.
Psychotherapy Alone Helps Keep Elderly Depression Free
In a recent study, Reynolds et al. (1997) found that the majority (90%) of elderly clients who were initially treated for depression with a combination of psychotherapy and medication maintained the changes when they were removed from the medication and but seen by a therapist for one maintenance session of psychotherapy per month over a 12-month follow-up period. In contrast, only 31% of those participants initially treated with both psychotherapy and medication maintained the changes when they attended a medication clinic and received a placebo during the same follow-up period.
Reynolds, C.F. (1997). American Journal of Psychiatry, 154(7), 958-962.
How do we know that therapy works?
A simple, cost-effective measure for proving and improving the effectiveness of clinical work
Traditionally, clinicians have relied on client self-report and their own judgement to evaluate the effectiveness of therapy. Historically, more formal assessment of treatment outcome using standardized measures has not taken place as such instruments have not "fit" with outpatient care. The tests have been too long, too costly, or irrelevant to treatment. Enter the OQ-45, a simple, brief (2 to 5 minutes to score and interpret), cheap (.03 cents per administration), and clinically relevant outcome measure. The 45-item measure provides normative scores on client level of symptomatic distress, social and interpersonal functioning. In a study by Lambert et al. (1996), the OQ was shown to have excellent psychometric properties–test-retest in the .80's, internal consistency in the .90's, and good concurrent and construct validity. Software supported versions of the measure provide the clinician with feedback regarding the progress of cases–including warnings when treatment may need to be altered to optimize client change.
Ongoing research being conducted by the Founders of ISTC finds that the measure can reliably predict the outcome of treatment by the 3rd or 4th visit! The basic idea behind such predictions is based on the finding that the majority of people in treatment experience change earlier rather than later in treatment regardless of the model or technique employed (see figure 1). This same data shows that the longer a person goes without change, the greater the likelihood of a negative outcome.
Source: Howard, et al (1986). The dose effect response in psychotherapy. American Psychologist, 41(2), 159-164.
Importantly, such findings should not be taken to mean that all treatment should be brief in duration. Indeed, this research goes a long way toward proving that the distinction between brief and long term treatment obscures rather than clarifies the important issues in treatment outcome. The only time treatment should be brief is when little or no progress is being made in the early sessions–then it should be as short in duration as possible. As long a clients are making documentable progress and are interesting in continuing, however, treatment should be extended.
Lambert, M., Burlingame, G., Umphress, V., Vermeersch, D., Clouse, G., and Yanchar, S. (1996). The reliability and validity of the Outcome Questionnaire. Clinical Psychology and Psychotherapy, 3(4), 249-258.
Are Clients who are Satisfied also Clients who've Benefited from Treatment?
Developers of therapeutic approaches like to report high rates of client satisfaction as evidence that people treated with their approach are also clinically improved. Many researchers have cautioned that such measures are typically inflated as well as unrelated to either empirical or client report of change in the presenting complaint. Researchers Lunnen and Ogles (1998) have driven this point home once again by finding that clients "tend to report high levels of satisfaction regardless of outcome" (p. 406). What is a better predictor? Well, for one, the therapeutic alliance–both client and therapist ratings of the alliance matched categorical groupings of improvers versus no-changers and deteriorators.
There are several simple, easy to use measures that can be employed to assess the alliance. One measure that assesses the client's ratings of the four common factors is Johnsons (1994) Session Rating Scale. Ongoing research at ISTC is employing a new measure that assesses: client satisfaction, therapeutic alliance, and degree of fit with the client theory of change (click here page for more information on this important construct). The measure has good internal consistency and is currently under study for ability to predict change in treatment. For copies of these instruments see the table above.
Other research being conducted at the center employs the Stages of Change Scale (SOCRATES). Briefly, this measure helps determine which stage of change the client is in so that interventions may be tailored to the client's motivational status. There are two versions (long and short). Available research say that the short version is as reliable as the longer version. All are available on Does it Work page.
Lunnen, K.M., and Ogles, B.M. (1998). A multivariable evaluation of reliable change. Journal of Consulting and Clinical Psychology, 66(2), 400-410.
So, how do you feel about this? The ineffectiveness of changes in mood as a predictor of outcome
The question, "How do you feel about that?" has become–thanks to therapists–a cliche' in modern culture. For years, therapists have used clients' reports of changes in affect to determine whether treatment is progressing. Researchers at Beth Israel Medical Center in New York City have found, however, that changes in feelings are a poor predictor of treatment outcome. Researchers Gorman, Safran, Twining, Wallner, and Winston used five intermediate (sub-outcome) measures to determine which had the most predictive power in terms of ultimate outcome (Anxiety shift, depression shift, cognitive shift, optimism shift, and therapeutic alliance). Not surprisingly, these researchers found that shift in cognition and the strength of the therapeutic alliance during therapy were the strongest predictors of ultimate outcome.
Gorman, B., Safran, J., Twining, L., Wallner, L. & Winston, A. (1995). Linking in-session change to overall outcome in short-term cognitive therapy. Journal of Consulting and Clinical Psychology, 63(4), 651-657.
|
