Compliance to weight reduction programs is usually poor in obese individuals and therefore they often regain weight [20]. Thus improved compliance might contribute to the therapeutic effect in those individuals with secure attachment style. This assumption is supported by the finding that the patient-therapist relationship was assessed to be more positive in secure compared to insecure participants. Indeed, previous publications confirmed that securely attached adults are considerably more cooperative in a patient-therapist situation [22]. This corresponds to observations indicating that securely attached individuals have better access to their emotions and eating habits, can therefore better perceive, understand and name them, and thus also have better prerequisites for successful weight reduction [23]. Individuals with an insecure attachment style are more likely to have problems with these issues; in the case of emotional entanglement, for example, preoccupied attached individuals do not perceive eating as an attempt to compensate. Dismissively attached individuals, on the other hand, tend to have difficulties in experiencing emotions and thus compensate by eating; in the case of therapeutic failure, they tend to blame others - in this setting the therapist.
It may be assumed that insecurely attached patients have difficulties in coping with the disease and also with time-limited therapies [34]. Thus the limited number of group therapy sessions offered in our setting might have negatively affected especially preoccupied attached patients, while securely attached patients were more stable in this respect. In addition, different reactions of the therapists, which are well known to occur and may have been induced by the patients' attachment style [11], may also have affected differences in weight loss.
Some limitations of our study should be mentioned: For estimating the distribution of attachment styles in representative, non-therapy seeking obese individuals compared to normal weight individuals, our study results are not representative and further studies are required. This trial was a clinical outcome study, it was not randomized and there was no control group. Predominantly women were studied; thus, the results can not be transferred to male subjects. Due to the small number of cases, we did not analyze the subgroups of binge eating, night or emotional eaters, and the effect of additional psychiatric co-morbidities in this study. To estimate the relevance of attachment styles according gender, age, and different subgroups, future studies are desirable. But there is a strong support for the assumption that the attachment style and the helping alliance between patient and physician are important in the therapy of obesity and probably other diseases.