55-1 Does psychotherapy work? How can we know?

The question, though simply put, is not simple to answer. Measuring therapy’s effectiveness is not like taking your body’s temperature after a fever. So how can we assess psychotherapy’s effectiveness? By how we feel about our progress? By how our therapist feels about it? By how our friends and family feel about it? By how our behavior has changed?

If clients’ testimonials were the only measuring stick, we could strongly affirm psycho-therapy’s effectiveness. When 2900 Consumer Reports readers (1995; Kotkin et al., 1996; Seligman, 1995) related their experiences with mental health professionals, 89 percent said they were at least “fairly well satisfied.” Among those who recalled feeling fair or very poor when beginning therapy, 9 in 10 now were feeling very good, good, or at least so-so. We have their word for it—and who should know better?

We should not dismiss these testimonials. But for several reasons, client testimonials do not persuade psychotherapy’s skeptics:

As researchers continue to document, we are prone to selective and biased recall and to making judgments that confirm our beliefs. Consider the testimonials gathered in a massive experiment with over 500 Massachusetts boys, aged 5 to 13 years, many of whom seemed bound for delinquency. By the toss of a coin, half the boys were assigned to a 5-year treatment program. The treated boys were visited by counselors twice a month. They participated in community programs, and they received academic tutoring, medical attention, and family assistance as needed. Some 30 years later, Joan McCord (1978, 1979) located 485 participants, sent them questionnaires, and checked public records from courts, mental hospitals, and other sources. Was the treatment successful?

674

Client testimonials were glowing. Some men noted that, had it not been for their counselors, “I would probably be in jail,” “My life would have gone the other way,” or “I think I would have ended up in a life of crime.” Court records offered apparent support: Even among the “difficult” boys in the treatment group, 66 percent had no official juvenile crime record.

But recall psychology’s most powerful tool for sorting reality from wishful thinking: the control group. For every boy in the treatment group, there was a similar boy in a control group, receiving no counseling. Of these untreated men, 70 percent had no juvenile record. On several other measures, such as a record of having committed a second crime, alcohol use disorder, death rate, and job satisfaction, the untreated men exhibited slightly fewer problems. The glowing testimonials of those treated had been unintentionally deceiving.

Do clinicians’ perceptions give us any more reason to celebrate? Case studies of successful treatment abound. The problem is that clients justify entering psychotherapy by emphasizing their unhappiness and justify leaving by emphasizing their well-being. Therapists treasure compliments from clients as they say good-bye or later express their gratitude, but they hear little from clients who experience only temporary relief and seek out new therapists for their recurring problems. Thus, the same person—with the same recurring anxieties, depression, or marital difficulty—may be a “success” story in several therapists’ files.

Because people enter therapy when they are extremely unhappy, and usually leave when they are less unhappy, most therapists, like most clients, testify to therapy’s success—regardless of the treatment.

How, then, can we objectively measure the effectiveness of psychotherapy? How can we determine which people and problems are helped, and by what type of psychotherapy?

In search of answers, psychologists have turned to controlled research studies. Similar research in the 1800s transformed the field of medicine. Physicians, skeptical of many of the fashionable treatments (bleeding, purging, infusions of plant and metal substances), began to realize that many patients got better on their own, without these treatments, and that others died despite them. Sorting fact from superstition required observing patients with and without a particular treatment. Typhoid fever patients, for example, often improved after being bled, convincing most physicians that the treatment worked. Not until a control group was given mere bed rest—and 70 percent were observed to improve after five weeks of fever—did physicians learn, to their shock, that the bleeding was worthless (Thomas, 1992).

In psychology, the opening challenge to the effectiveness of psychotherapy was issued by British psychologist Hans Eysenck (1952). Launching a spirited debate, he summarized studies showing that two-thirds of those receiving psychotherapy for non-psychotic disorders improved markedly. To this day, no one disputes that optimistic estimate.

Why, then, are we still debating psychotherapy’s effectiveness? Because Eysenck also reported similar improvement among untreated persons, such as those who were on waiting lists. With or without psychotherapy, he said, roughly two-thirds improved noticeably. Time was a great healer.

Later research revealed shortcomings in Eysenck’s analyses; his sample was small (only 24 studies of psychotherapy outcomes in 1952). Today, hundreds of studies are available. The best are randomized clinical trials, in which researchers randomly assign people on a waiting list to therapy or to no therapy, and later evaluate everyone, using tests and assessments by others who don’t know whether therapy was given. The results of many such studies are then digested by means of meta-analysis, a statistical procedure that combines the conclusions of a large number of different studies. Simply said, meta-analyses give us the bottom-line result of lots of studies.

675

Therapists welcomed the first meta-analysis of some 475 psychotherapy outcome studies (Smith et al., 1980). It showed that the average therapy client ends up better off than 80 percent of the untreated individuals on waiting lists (FIGURE 55.1). The claim is modest—by definition, about 50 percent of untreated people also are better off than the average untreated person. Nevertheless, Mary Lee Smith and her colleagues exulted that “psychotherapy benefits people of all ages as reliably as schooling educates them, medicine cures them, or business turns a profit” (p. 183).

Dozens of subsequent summaries have now examined psychotherapy’s effectiveness. Their verdict echoes the results of the earlier outcome studies: Those not undergoing therapy often improve, but those undergoing therapy are more likely to improve, and to improve more quickly and with less risk of relapse. Moreover, between treatment sessions for depression and anxiety, many people experience sudden symptom reductions. Those “sudden gains” bode well for long-term improvement (Aderka et al., 2012).

Is psychotherapy also cost-effective? Again, the answer is Yes. Studies show that when people seek psychological treatment, their search for other medical treatment drops—by 16 percent in one digest of 91 studies (Chiles et al., 1999). Given the staggering annual cost of psychological disorders and substance abuse—including crime, accidents, lost work, and treatment—psychotherapy is a good investment, much like money spent on prenatal and well-baby care. Both reduce long-term costs. Boosting employees’ psychological well-being, for example, can lower medical costs, improve work efficiency, and diminish absenteeism.

But note that the claim—that psychotherapy, on average, is somewhat effective—refers to no one therapy in particular. It is like reassuring lung-cancer patients that “on average,” medical treatment of health problems is effective. What people want to know is the effectiveness of a particular treatment for their specific problems.