The Agency for Health Care Research and Quality has released a report identifying a sharp rise in eating disorder hospitalizations. Can psychotherapy help with these debilitating disorders?
The eating disorders are particularly scary forms of mental health distress. There are a range of clinical presentations of eating disorders, some of which include: poor body image, unhealthy approaches to weight loss including restrictive diets and excessive exercise, binge eating, purging through use of laxatives or vomiting, or consuming non-edible objects (pica). Even in ‘milder forms' of these disorders, the physical cardiac, gastrointestinal, esophageal, and nutritional consequences can be devastating. Medically complicated hospitalizations are frequently necessary (often in light of a strong resistance to receiving necessary medical treatment) and accidental death rates are high. Many clinicians are extremely hesitant to admit to their caseloads patients with eating disorders due to the physical risks involved, the time-intensive collaboration with medical treaters, and the persistent nature of these disturbances.
Can psychotherapy help patients with eating disorders?
First, we need to look at how researchers and clinicians try to find out if psychotherapy works or not. In a follow-up post, we'll explore some published treatment results for eating disorder treatment.
There are three common research approaches to examining psychotherapy's effectiveness (something I've briefly discussed in a previous post on treating depression):
Randomized Control Trial (RCT): In a randomized control trial, participants are randomly assigned to a treatment condition. Some patients are given only psychotherapy, others may receive only psychiatric medication, some may be given placebo treatment, and some may not receive any treatment at all and be placed on an extended wait-list. The point of a psychotherapy RCT is to answer two questions: 1. Does psychotherapy work? and 2. Does psychotherapy work better than no-treatment and/or other treatment options? In order to make a causal statement (i.e. the patient got better BECAUSE of the therapy, and not for some unrelated reason), the experiment needs to be as controlled as possible...a sort of therapeutic "clean room". That means that patients are carefully selected to have pure forms of the disorder being examined and that treatments are rigidly manualized and treatment protocols are faithfully delivered. Furthermore, the participants are supposed to be "blind" to the treatment being received. That means the patient isn't supposed to know if they are receiving a placebo or an active treatment, and experimenters interacting with the participants aren't supposed to know what treatment is being delivered either.
You can already see one inherent flaw in this model: how does a patient not know they are receiving psychotherapy, and how can clinician not know that they are practicing a particular psychotherapy? Another dilemma in this model is how to deal with the factor of time. It is said that time heals all wounds, but is that true? It is very hard to maintain a clean psychotherapy study for any long period of time. How many people have a single, well-defined psychological disorder with no-other co-existing problems like depression, anxiety, physical illness or substance use? Furthermore, it's difficult to stick to a manual for a long time when you are interacting with other human beings who have unique needs and desires. On top of that, one has to assign a group of people to a condition where they receive no treatment.
Imagine you have an incredibly distressing and physically damaging condition. Now imagine your doctor telling you have to agree to wait a long period of time before they'll give you treatment. What ends up happening is that the treatments delivered in RCT are often very short (8-16 sessions, usually) and mostly behavioral, skills training, or cognitive-behavioral in nature. While it is clear that a short period of treatment can yield significant symptomatic improvements, patients completing an RCT treatment may still have significant pathology and a high proportion of people don't get better at all.
RCTs have been declared the "gold standard" for defining "evidence-based treatment". As you can see, there are a whole host of problems in allowing this research methodology to dictate clinical practice. In that sense, the term "gold-standard" might be well used here given that the term relates to an economic policy which has no current real-world implementation and fell apart as a feasible practice in the U.S. way back in the 1930s.
Effectiveness Studies: Do parachutes work to save the lives of human beings jumping out of an airplane? How do you know? To my knowledge, there's never been a randomized control trial where jumpers are randomly assigned to a parachute or no-parachute condition. What about wearing a seat-belt to prevent death in a car accident? Same thing...no way to do an RCT (okay it's possible, I guess, but I imagine it would come with a significant amount of jail time). Effectiveness studies measure treatment outcomes in people receiving psychotherapy in actual community or private treatment settings. Most people know pretty definitively that mental illness doesn't just go away with time (in fact, many times it gets worse without intervention). In an effectiveness study, we look at how patients come into treatment and measure how they change over time. If their symptoms, behaviors, and well-being improve, we infer that their treatments worked. Of course, in an effectiveness study, we can't draw a definitive conclusion that our treatment was the reason for that change...maybe the illness just got better on its own, maybe the patient got a great new job and felt better, or maybe they got a new pet that makes them happy. But if we have a large enough group of people who show a consistent pattern of getting better, we can make some pretty common sense inferences about the usefulness of treatment.
Meta-analyses: A meta-analysis involves digging up as many studies as one can find in a specific field (say, all studies done treating eating disorders). Any one research study could happen to give some exaggerated or inconsistent finding. To get a better picture of the true effect, a meta-analysis compiles more information from as many studies as possible. Because I love sports metaphors for psychology: Let's say the Boston Red Sox play a baseball game against their contemptible competitors, the New York Yankees. The Red Sox deliver a crushing defeat: 12-0. Are the Red Sox a better baseball team? (Yes, of course they are...but let's support that with more evidence) To answer that question more accurately, we would have to look at all the many games the teams play against each other through the season. If Boston consistently wins more games (of course they do), then we have more evidence to say that they are the better baseball team.
Now that we've seen some ways researchers evaluate psychotherapy, in a follow-up post we'll look to answer the question: Can psychotherapy help patients with eating disorders?
.png)
