What is a stationary bug

Problems transitioning from inpatient to outpatient treatment for depression

Background: So far, hardly any data is available on the characteristics of inpatient depression treatment in Germany and the subsequent outpatient treatment. The aims of this work were to characterize the inpatient treatment and the subsequent outpatient treatment, to determine the readmission and mortality rates and to identify risk factors.

Methods: Descriptive analysis of routine administrative data from the BARMER health insurance company. All insured persons aged 18 to 65 who were treated in clinics for psychiatry and psychotherapy as well as psychosomatics in 2015 with the main diagnosis depression were included. Resumption risk factors and mortality were assessed using mixed logistic regression.

Results: Of the 22,893 people affected, 78% were treated in clinics for psychiatry and psychotherapy and 22% in clinics for psychosomatic medicine and psychotherapy. The median length of stay in hospital was 42 days. Follow-up treatment was not in accordance with the guidelines in 92% of patients who were hospitalized with major depression and 50% of those who had moderate depression. Readmissions within one year occurred in 21% of the cases. The age- and gender-adjusted one-year mortality for the German population was 961/100,000, 3.4 times the average population. In the regression model, more therapy units during inpatient treatment and subsequent psychotherapeutic treatment were associated with a reduced probability of resumption, while longer inpatient treatment and subsequent drug or psychotherapeutic treatment were associated with a lower risk of mortality.

Conclusion: After inpatient depression treatment, guideline-compliant care according to the national S3 care guidelines in the sectoral structures of outpatient and inpatient care in the German health care system is very inadequate. There is a clear excess mortality.

Depressive illnesses are among the most common and most debilitating illnesses worldwide (1). They cause great individual suffering (2) and are associated with increased mortality, both from suicides and comorbid diseases (3). Currently, around three percent of patients with depressive illnesses in Germany are treated as inpatients each year (4). After discharge, there is an increased risk of suicide (5, 6) and relapse and, in the absence of remission, the risk of chronification (7, 8). In order to achieve a full remission or to prevent recurrent and chronic courses as well as suicides, further treatment or remission-stabilizing maintenance therapy appropriate to the severity of the disease and the comorbidities is indicated according to the S3 / National Care Guideline. In severe depression, it should consist of a combined drug and psychotherapeutic treatment. In moderate depression, depending on the patient's preference, either drug or psychotherapeutic further treatment is sufficient (9). The guideline recommendations have the highest level of evidence “A”, that is, they are based on several randomized controlled studies. Investigations of routine health insurance data and survey studies that examined outpatient care for patients with depression indicated deficits in care (4, 10, 11, 12, 13). So far, hardly any data is available on the characteristics of inpatient depression treatment in Germany and the current status of outpatient further treatment, which in view of the current health policy debates about the care of the mentally ill makes it difficult to assess the necessary health policy measures. This work therefore examined the following questions in a large routine data set from the BARMER health insurance company with over 9 million insured persons:

  • How and for how long are patients with a depressive episode in Germany treated in psychiatric-psychotherapeutic or psychosomatic-psychotherapeutic clinics?
  • Is further outpatient treatment carried out in accordance with the guidelines?
  • What are the resumption and mortality rates within a year and
  • what socio-demographic, disease-related or treatment-related variables are influencing them?


From the approximately 9.4 million BARMER insured persons in 2015, those between 18 and 65 years of age were selected from inpatient psychiatric-psychotherapeutic or psychosomatic-psychotherapeutic treatment with an ICD-10 diagnosis of F32.x (depressive episode) or F33.x (depressive episode in recurrent depression). If a discharge took place in 2015, the index stay was considered to be an index stay. Following the discharge, a 365-day observation interval was defined in which the guideline-based drug and psychotherapeutic further treatment as well as readmission and mortality rates were examined (short version, see eMethods).


Characteristics of the index population

22,893 of the approximately 9.4 million BARMER insured persons (corresponding to an annual prevalence for an inpatient stay of 0.25%) fulfilled the inclusion criteria (= index population). 66% (n = 15 059) were women. The median age was 47 (range 47) years. 39% (n = 8,991) of the index population were registered in a large city (> 100,000 inhabitants).

Characteristics of the index hospital treatment

The median lay time was 42 days (minimum 1, maximum 816). Upon discharge from the index stay, 78% (n = 17 799) of the index population were treated in a clinic for psychiatry and psychotherapy and 22% (n = 5 094) in a clinic for psychosomatic medicine and psychotherapy. Figure 1 and eTable 1 provide an overview of the distribution of the cases (Figure 1a). What was noticeable here was a low proportion of day-clinic treatment, especially in psychosomatics (Figure 1b), as well as a significantly higher density of therapy by doctors / psychologists in psychosomatic-psychotherapeutic clinics / departments than in clinics / departments for psychiatry and psychotherapy, (Figure 1c), in however, the vast majority of patients were treated with severe and psychotic depression (Figure 1d). With regard to the secondary diagnoses, there were no major differences between the two types of clinic (Figure 1a).

Included fee schedule items (GOP) from the Uniform Assessment Standard (EBM) of the National Association of Statutory Health Insurance Physicians 2015/2016

Severity-adapted, guideline-based follow-up treatment

With regard to the severity-adapted guideline recommendations, 92% of the patients with severe depression (n = 12,395) did not receive a guideline-compliant combined drug and psychotherapeutic follow-up treatment (Figure 2a). Of the patients with moderate depression, 50% (n = 4,605) did not receive any follow-up treatment that was either medicinal or psychotherapeutic in accordance with the guidelines (Figure 2b).

The investigation of outpatient drug treatment revealed that 84% of the 13,427 patients with severe depression (F3X.2 or F3X.3) and 70% of the patients with moderate depression (F3X.1, n = 9,270) had at least one antidepressant - Redeem the prescription in the entire follow-up year. But only 57% (n = 7,651) of the severely and 42% (n = 3,908) of the moderately severely depressed patients received - as recommended in the guideline - a prescription in the first quarter after discharge and, if necessary, follow-up prescriptions with a sufficiently defined definition for at least four months Daily Doses (DDDs). For an overview of the DDDs of the prescribed substances, see the e-graphic.

With regard to further outpatient psychotherapeutic treatment, it was found that of the patients with severe depression only 33% (n = 4,428) and of the patients with moderate depression 37% (n = 3,474) received any one hour of psychotherapy within the one-year observation interval. According to the guidelines, only 12% (n = 1,676) of patients with severe depression and only 15% (n = 1,376) of patients with moderate depression received the first hour of therapy within the first quarter of discharge and at least eight hours of therapy within the one-year observation interval.

The interval between discharge from hospital and the first hour of therapy was examined in the 4,311 patients who had not received any psychotherapy in the year before hospital admission. The median interval between discharge and start of psychotherapy was 111 days (95% confidence interval: 106; 115). The psychotherapies were mainly individual therapies with behavioral therapy 55% and depth psychology 43%, whereas psychoanalysis and group therapies only played a marginal role (Table 1).

ETable 3 shows whether a specialist or family doctor / internist was consulted in the first quarter of the treatment of the observation interval.


Within the observation interval, 21% (n = 4 798) of the index population were re-admitted to inpatient or semi-inpatient psychiatric-psychotherapeutic or psychosomatic treatment, 5% (n = 1 103) twice or more. In order to check whether there are any indications of the preventive effects of a guideline-compliant follow-up treatment, we used a multi-level regression model to check indications of the risk factors that can be mapped in the data of a readmission in the second half of the year after discharge. In the "Random Intercept" null model, the hospital and region explained a substantial proportion of the variation in terms of readmission. In the three-level regression model, with regard to sociodemographic factors, older age increased the likelihood of resumption. With regard to disease-related factors, the main diagnosis of major depression and secondary diagnoses of personality disorder, post-traumatic stress disorder, obsessive-compulsive disorder, addiction, or alcoholism in the model significantly increased the probability of readmission. With regard to treatment-related factors, the likelihood of resumption was higher in the case of treatment in a psychiatric and psychotherapy clinic and guideline-compliant antidepressant follow-up treatment. More therapy units during the index treatment as well as guideline-based psychotherapeutic follow-up treatment reduced the likelihood of resumption. McFadden's pseudo-R2 was 0.03; the predictors thus explained a moderate proportion of the variation beyond the clinical and regional variation (14) (Table 2).


1.1% (n = 256) of the index population died within the observation period. The result was adjusted for age and gender for the 18- to 65-year-old German population (15). The one-year mortality of 961/100,000 was 3.4 times higher than that of the age- and gender-equivalent average population (282/100,000). The cause of death is not recorded in the available data. In the “Random Intercept” null model, “hospital” but not “region” explained variation in mortality. In the 3-level regression model, age and male gender were the fixed effects as sociodemographic risk factors, disease-related factors were the main diagnosis of severe depression, psychiatric secondary diagnoses and severe somatic comorbidities (represented by the Charlson index) and treatment-related factors were treatment in a clinic for psychiatry and psychotherapy was associated with an increased likelihood of mortality. In contrast, in the model, a longer inpatient stay and (at least minimal) antidepressive and psychotherapeutic follow-up treatment reduced the probability of mortality. McFadden's pseudo-R2 was 0.16; the predictors thus explained a significant proportion of the variation beyond the clinical and regional variation (14) (Table 3 a, b).

3-level regression model mortality - Random effects of the null model and the predictor model

In order to be able to gain indications of the preventive effects of a therapy according to the guidelines, we planned to repeat the regression with the further indicators "guideline-compliant medication" and "guideline-compliant psychotherapy" for deaths in the second half of the year, but none of those who died in the second half of the year received in the first Half a year a guideline-based therapy.


Inpatient treatment

The routine data from BARMER show that in Germany, inpatient treatment of depression was carried out in three out of four patients in clinics for psychiatry and psychotherapy. The number of beds in 2015 was 50,972 (psychiatry without addiction) to 10,439 (psychosomatics) (16). In clinics for psychiatry and psychotherapy, particularly those affected with severe and psychotic depression were treated, in psychosomatic clinics mainly with moderate depression. The density of therapy in psychiatry and psychotherapy clinics was, however, significantly lower than in psychosomatic medicine. This is due to the requirements of the 30 year old Psychiatry Personnel Ordinance (PsychPV), which limits the possibilities of intensive and guideline-compliant psychotherapy (17, 18). Since the staffing in psychosomatic clinics is generally not subject to the PsychPV, a full medical / psychological staff member is only entrusted with treating half as many patients on average (19, 20). At 42 days, the hospital stays were shorter than 15 years ago, for example (21). The fact that day-clinic treatment, especially in psychosomatic clinics, is only the exception is astonishing in view of the need to promote the integration of patients into their living environment and the lower costs of this form of treatment.

Outpatient follow-up treatment

After discharge, only 8% of the patients with severe depression and 50% of the patients with moderate depression received further treatment in accordance with the guidelines. Of the patients with severe depression, only 12% received the further psychotherapeutic treatment recommended for this group in the guideline. The reasons for this could not be clearly determined on the basis of the study data. The fact that supply structure deficits in Germany play an important role in the long interval of 16.7 weeks until the start of treatment is evident from the information provided by the Federal Chamber of Psychotherapists, according to which nationwide the waiting time for a psychotherapy place is on average 19.9 weeks, with a higher one Patient preference for this treatment method (22). The long intervals are medically questionable, as data from the English health system show that waiting times of more than four weeks significantly reduce the chance of a positive effect from outpatient psychotherapy (23). Structural support could be provided through improved coordination between clinics, specialists in psychiatry and psychotherapy / general practitioners and guideline psychotherapists (24, 25, 26). The use of group therapies as an extension of the offer to compensate for the lack of resources also rarely takes place in Germany, according to the chambers of psychotherapists because of bureaucratic hurdles (27).

There were also clear deficits with regard to further drug treatment. Only 57% of the seriously or psychotically ill received medication of sufficient duration or dose - although the indicator chosen in the study represents a very conservative estimate. The extent to which this was due to resentment on the part of patients against long-term use of medication or a non-guideline-compliant approach by doctors could not be determined from the data.


The literature agrees that comorbidities and older age are associated with worse outcomes (28). The fact that patients treated in a clinic / department for psychiatry and psychotherapy are more likely to be re-admitted can be explained by differences in the patient mix and the more acute treatment setting. The fact that the regression model indicates a connection between inpatient therapy intensity and readmission rates raises the question of the appropriateness of staffing levels in psychiatric-psychotherapeutic clinics. After discharge, patients with follow-up psychotherapeutic treatment in the model have a lower probability of resumption, which indicates the importance of implementing this guideline recommendation. The finding that guideline-based further drug treatment in the model is associated with a higher probability of resumption appears counterintuitive, but could be explained by the fact that there is a harder indication for consistent medication in severe and therefore more at risk of relapse patients. Alternatively, there could also be a connection with the rebound phenomena that have only been described in the more recent literature when antidepressants are not properly discontinued (29).


Deaths represent a central, clinically highly relevant outcome measure that is available in the routine data. The present study showed an unexpectedly high mortality rate. The mortality in the year following the depression hospital discharge was 3.4 times higher than that of the normal population.This means that instead of the expected 65 deaths in 22,893 people of the normal population in the course of a year in the present (adjusted) index sample, 151 more people died. The recording of mortality in the present data set can be assumed to be reliable. While factors such as older age, male gender, severity of depression, combined addictions, the Charlson index, and treatment in a clinic / department of psychiatry and psychotherapy can be used to identify populations at risk, the treatment-related factors indicate the risk that Failure to continue psychotherapeutic and drug treatment as well as shorter hospital stays could contribute to increased mortality. It was particularly worrying that none of those who died in the second half of the year had received treatment in line with the guidelines in the first half of the year.


The use of administrative health insurance data offers the opportunity to examine large populations and the treatment reality outside of studies over the course of the year. A disadvantage of routine data is the limited validity of the diagnoses, especially the severity classification used in this work. The focus of this work on patients, who were mostly treated for several weeks in a specialist hospital or department, suggests a better validity of at least the diagnoses of mental disorders, on the other hand the outpatient treatment reality of some of the severely affected is in focus. By summarizing stays, this work gives a more realistic picture of hospital stays and readmission rates than previous studies. Since health insurance companies are not allowed to have further clinical variables at their disposal, these must be opened up using indicators. The reality of treatment can therefore differ in individual cases. In addition, routine data analyzes can only describe the supply situation. The regressions carried out can only provide indications for plausible relationships; causal interpretations are not possible. In addition, the relationships shown in the regression models are only valid, interpreted in a correlative way, if there are no disregarded confounding variables with substantial effects on the examined outcomes. In addition, given the relatively small number of events compared to the complexity of the statistical models, some results could be biased. Furthermore, it should be noted that "guideline fairness" could only be operationalized very roughly in the present study, because guidelines make general recommendations, but explicitly eliminate the possibility of deviating from these recommendations and taking into account factors beyond the severity of the disease in individual cases Treatment decisions.