Defining recovery: What contributes to recovery and remission from an ED?
In the first two blog posts of this series, I’ve looked at how researchers define recovery, and how patients do so. Now, I want to look at what factors predict recovery in eating disorders.
But first, a note. When I talk about things that predict a negative outcome, it’s possible that you might recognize some things about yourself and get all discouraged. “Is Carrie saying I’m never going to get better?” That’s not what I’m saying, not at all. We all have factors that make it both easier and more difficult to recover from an eating disorder. Every last one of us. So having something that predicts a lower likelihood of recovery doesn’t mean you won’t get better, it just means that, on average, people who have that trait have a more difficult time reaching recovery. It’s not set in stone. And as much as I hate it when people say, “If I can recover, you can too!!!” (you’re not me, I’m not you, and it sounds a bit patronizing, honestly), good treatment, good support, and your own efforts can help mediate these things. I’ve known people who have had the deck stacked against them in pretty much every way I can think of who have gotten better. So don’t be discouraged and don’t make dire predictions about your own recovery based on what I write next. Use it as an opportunity to get more support if you’re struggling.
</ gets off soapbox>
Most of the literature on factors that influence recovery have been done on anorexia. I found 2 studies on bulimia, but they didn’t seem to identify any major criteria that predicted recovery (Keel & Mitchell, 1997). So for this particular part of the post, I’m limited in discussing anorexia. So here’s what we know about what helps recovery in anorexia nervosa:
Early Detection. In a variety of studies, having a shorter duration of illness and a higher BMI at diagnosis seems to improve outcome. Clearly, this is an important variable, but the data isn’t that exciting. This isn’t rocket science, really.
Adequate weight restoration in the hospital. In a study of 212 adolescents hospitalized for AN in Europe, researchers found that nearly half required rehospitalization (Steinhausen et al., 2008). When they looked at the data, they identified five factors that predicted rehospitalization in over 2/3 of the patients: paternal alcoholism, eating disorder in infancy, periodic overactivity, low weight increase during first admission, and low BMI at first discharge.
The importance of adequate weight restoration is seen in many other studies. In a 21-year follow-up of 84 women hospitalized for AN, 50% had a good outcome according to Morgan-Russell criteria, 20% had an intermediate outcome, and 26% had a poor outcome. This did not count the 16.7% of women from the original sample who had died. In the table below, you can see the factors that predicted outcome in the study.
Odds Ratios are used to describe how closely two things are related. In this case, researchers are comparing the likelihood of a poor outcome of AN to the different factors you see in the table below. An odds ratio of 1.0 means there is no relationship between the two variables- it has no effect. In the table, you can see this in the age of onset. How old you were when you first got sick didn’t seem to have any effect. A longer duration of illness meant that you were 1.34 times more likely to have a poor outcome. For other variables, such as amount of weight gained in the hospital and the current BMI, made you less likely to have a poor outcome. Thus, the odds ratio value is less than 1.0.
The 95% confidence intervals tell you how likely the numbers are reflective of the actual relationship. Does the data indicate that this is just a freaky weird sample, or is it likely that this is what’s going on. Having large amounts of data and/or a really strong relationship between the two variables increases our confidence in the data.
Finally, a US study of 22 women hospitalized for anorexia nervosa found that reaching a normal weight before discharge was significantly associated with lower rates of rehospitalization and fewer ED symptoms at 29 month follow-up (Baran, Weltzin, & Kaye, 1995).
Food variety. Researchers found that the energy density of foods consumed, as well as the variety of foods, at the end of treatment could predict later relapse (Schebendach et al, 2008). The 47 women they followed were all hospitalized on the EDU of the New York State Psychiatric Institute at Columbia University. After reaching a minimum of a BMI of 20, the women then completed 4 days of food logs, and were then followed for one year. At this point 41 of the women could be classified as treatment success (n=29) or treatment failure (n=12). Those women that had the greatest food variety and ate the most energy dense foods were the most likely to have a good outcome. Interestingly, the women seemed to eat the same overall number of calories during the 4 day food diaries- it was the composition of the foods that seemed to differ. In a nutshell, eating more fats significantly improved outcome.
Not only do factors specifically addressed in treatment make a difference in recovery, there are also individual differences in someone’s likelihood of recovery.
Perfectionism and Distrust. In a study of 26 women admitted consecutively to a French hospital for a refeeding program for severe anorexia, scientists administered the Eating Disorders Inventory on admission and followed the women for around 8-10 years (Bizeul, Sadowsky, & Rigaud, 2001). Half of the women recovered during the follow-up period, half did not. The researchers defined recovery as normal weight and menses, no major episodes of ED behaviors for 2 years, satisfaction with life, autonomy from the family (that variable wasn’t defined), and insight into one’s own life.
On the EDI, researchers found that higher scores for perfectionism, interpersonal distrust, ineffectiveness, interoceptive difficulties, and drive for thinness were all individually associated with poor outcome, along with overall EDI scores. When the researchers looked at all of these factors together, however, only perfectionism and interpersonal distrust were significant for predicting illness severity. You can see the distribution (the asterisks and dots) and overall averages (bars) in the figure below.
Social Support. In a qualitative follow-up study of 69 women who had been referred to a local eating disorder service in New Zealand and diagnosed with AN, on average, 12 years prior (the study was conducted, on average, 15.4 years after ED onset), researchers conducted lengthy interviews to gather a set of themes for both what patients thought was the cause of their ED and, in those that had recovered, what factors emerged as helping them get better. The women were around 32 years old, and 90% had recovered, as defined by no active eating disorder (AN, BN, EDNOS) according to DSM-III-R criteria.
Although the criteria that the women said was important in their recovery varied (as shown in the table below), the most important factor, according to the authors, was a supportive relationship. Exactly who that relationship was with–partner, parents, or therapist–seemed to be less important than the relationship itself. Another subset of the women said that, basically, they matured out of the disorder. This is consistent with other data that shows many people with EDs in adolescence tend to recover in their mid-20s. Researchers think that increasing maturity and cognitive development help to make recovery more likely in this age group.
The researchers point this out at the end of the study:
The strength and, at the same time, the limitation of this study is the focus on the patients’ subjective perspectives. Individuals often try to give meaning to experiences in their life and may grasp onto stressful experiences for their explanatory power in creating a plausible story for why they developed an eating disorder (‘‘effort after meaning,’’ Cohen & Cohen, 1984). The perspective of the patients, together with clinical data, can provide a novel and rich understanding of etiology and outcome factors in AN.