The results from this study indicate the following: (1) Compared to healthy subjects from the German general population, the current sample of inpatients with OCD suffered from a diminished global, physical, psychological, and social QoL with baseline scores being one to two standard deviations below norm values. Environmental QoL was not impaired. (2) QoL was significantly improved 1 year after inpatient CBT with EX/RP and subsequent outpatient treatment in all domains except social QoL. Effect sizes were medium to large for global, physical, and psychological QoL (Cohen’s d between 0.7 and 0.9). However, only a small subgroup of the sample showed clinically significant improvements. (3) Despite the improvements, participants’ global, physical, psychological, and social QoL remained significantly lower than population norms. (4) QoL improvements were predicted by a reduction of depressive symptoms. The amelioration of BDI scores was a good predictor of global, physical, and psychological QoL improvements with the proportion of the explained variance being greater than 50%. It is to note that the present study has relevant limitations, most prominently the lack of a severity measure of OCD and the lack of standardized diagnostic interviews. Thus, the presented results do not allow for directly investigating the impact of alterations in OCD symptoms on changes in QoL. The diagnoses were established by experienced clinicians during the inpatient stay. However, we cannot exclude that the lack of structured interviews might have biased the results.
Our results are in line with previous reports on significantly reduced QoL in persons with OCD when compared to norm values from the general population [9, 10, 28]. The level of impairment observed in our study corresponds well to data by Hollander et al.  and Albert et al.  who reported that baseline scores were at least one z score below norm values for each mental health domain of the SF-36.
With regard to distinct QoL domains, reduced QoL in the psychological domain and social functioning have been reported based on different QoL measures [7, 9, 10]. Skevington and McCrate  recently published means and standard deviations for the WHOQOL-BREF for different conditions. The score for psychological QoL of our sample was lower than the scores of persons with chronic schizophrenia, sleep disorders, diabetes mellitus, and neurodegenerative disorders. Only persons with depression reported a lower quality of life in the respective domain.
The results concerning physical QoL are controversial: Some studies found that physical QoL in persons with OCD was not impaired , whereas others have reported impairments . One possible explanation is that impaired physical QoL in persons with OCD might reflect comorbid depressive symptoms. Moritz and colleagues , who found that physical QoL was impaired, reported that the criteria for major depression or dysthymia were fulfilled by 37% of their participants. In contrast, Koran et al.  used the same QoL measure (SF-36) and observed that physical QoL was comparable to norm values. Of note, in their sample only 2% of the participants were diagnosed with major depression and 12% with dysthymia. The differences concerning physical QoL cannot be attributed to age, as the participants of the study by Moritz et al. were younger than the sample by Koran et al. In our sample, physical QoL was reduced at baseline, which might be driven by the relatively high severity of comorbid depression (mean baseline BDI 19). More specifically, it is to note that the physical domain of the WHOQoL-BREF includes questions about the ability to manage daily routines, working ability, sleep quality, and energy for daily living. It seems plausibe that persons with mood disorders according to ICD-10 criteria who suffer from energy loss, diminished interest, and sleep disorders score low on this domain.
Environmental QoL, which is covered by the WHOQOL-BREF, but not by the SF-36, was not reduced in our sample and was slightly higher than in most other samples of persons with OCD (e.g., Cordioli  who found a baseline score which was more than 10 points lower than ours). Interestingly, Stengler-Wenzke and colleagues , who analyzed a similar German OCD sample seeking help at a University Medical Center (Leipzig), found a very similar baseline score for environmental QoL. This might be due to a selection bias. Patients with a stable socioeconomic background may be more likely to seek help in a University Medical Center in an urban area. This line of arguments fits well to the high level of education reported in our sample. Only two participants did not finish school, and 73% had at least a secondary level of education.
In line with other results, our findings indicate that the QoL of persons with OCD can be improved in the long run. In the randomized trial by Cordioli , the QoL was found to be improved 3 months after cognitive behavioral group therapy in comparison to baseline. This observation comprised improvements on all four domains of the WHOQOL-BREF with effect sizes within the medium range (effect sizes were calculated by the authors of the present paper based on means and standard deviations reported by Cordioli et al.).
Furthermore, we observed that the different QoL domains improved to a different extent: Improvements were most prominent for psychological and physical QoL, whereas social QoL was not significantly changed. Other studies with 12-month follow-up periods also found that social QoL did not improve in persons with depression aged 65 years or older  and in persons with social phobia . Possibly, profound changes concerning social relationships and social support require prolonged periods of time. Insufficient social support and impaired social networks might be an etiological and maintaining factor of various mental disorders rather than a consequence of psychiatric symptoms.
Another important observation is that—despite significant improvements—the QoL of persons with OCD remained below population norms. This is in line with other results .
With regard to potential mediators, we found that QoL improvement was strongly linked to a reduction of depressive symptoms. Previous studies indicated that QoL and QoL improvement might be more strongly associated with depression than with OCD severity [9, 16]. This implies that a careful diagnosis and, if applicable, treatment of comorbid depression is an important step in treating persons with OCD.
To the best of our knowledge, the present study is the first that repeatedly administered a psychometrically sound measure of QoL in persons with OCD over a course of 12 months. Thus, our results extend previous research by providing data on long-term changes of QoL in persons with OCD. However, a number of significant limitations need to be discussed. First, our study lacks a control group. Second, we did not control for medication. Thus, conclusions about potential causal relationships between the administration of inpatient CBT and subsequent outpatient treatment, QoL improvements, and symptomatic changes cannot be drawn. We can nevertheless draw on solid data demonstrating that OCD is a chronic disorder which does not show considerable improvement over time without treatment [33, 34]. Third, the current analysis is based on quality management data from our clinic that, within the period of observation, did not routinely include measures of OCD severity. For this reason, the baseline scores of our sample were compared to data reported in other studies in order to better characterize our sample. The baseline SCL-90-R global severity indices and OC-subscale scores of our sample correspond well to the scores of n = 43 persons with OCD reported by Grabe et al. . The baseline GAF score of 41.6 in our sample indicates severe impairment and is considerably lower than scores reported by Ramirez et al.  for psychiatric outpatients belonging to several diagnostic groups, including mood, anxiety, and eating disorders. Another limitation is that structured interviews have not been used. Future studies are needed to further disentangle the impact of distinct changes in psychopathology on long-term changes in the QoL. Furthermore, it remains possible that we overestimated QoL improvements, as our study suffered from a considerable amount of missing data. Patients who were less satisfied with the treatment were possibly less likely to send back the follow-up questionnaires. However, we showed that the analyzed sample did not differ from the original dataset concerning psychological distress and sociodemographic data.