Admission Volume and Quality of Mental Health Care Among Danish Patients With Recently Diagnosed Schizophrenia

The organizational structure of health care systems, including the size of hospital units, is a topic of discussion worldwide. The key objective is to ensure high-quality of care, patient satisfaction, and efficiency. Increasing evidence indicates an association between high admission volume per unit and better clinical outcome, including lower mortality and reduced complications in surgical care and across a variety of medical conditions (1–3). These findings have stimulated a trend toward centralization of care across medical specialties and health care systems. However, little is known about how to organize psychiatric units optimally. It has not been established whether an analogous association exists between admission volume and the quality of mental health care, although such an association is of major interest—not only for patients and clinicians but also for policy makers and administrators looking for new ways to improve the efficiency of the health care system.

To our knowledge, only four published studies have explored the relationship between admission volume and quality of care in mental health care. The results were inconsistent, however, making it difficult to draw any conclusions (4–7). To further assess the possible role of admission volume in mental health care, we conducted a nationwide population-based cohort study to examine the association between admission volume per psychiatric hospital unit and the quality of mental health care as reflected by the use of specific processes of care among Danish patients admitted with recently diagnosed schizophrenia.

Methods

The Danish public health care system, which is funded mainly by taxes, provides universal health coverage that is free of charge (8). In particular, health equity is a stated priority, with equal reimbursement across all institutional levels. If psychiatric treatment is required, patients with schizophrenia are admitted to public psychiatric hospitals and their use of inpatient services is recorded in registers with the patient’s unique, ten-digit civil registration number (9,10).

The Danish Schizophrenia Registry

The Danish Schizophrenia Registry was established with the objective of monitoring, documenting, and improving the quality of care among patients with schizophrenia and has been operative since 2004. All Danish psychiatric hospitals treating patients with schizophrenia must report data on all treated patients to the registry. The registry contains information on delivery of specified processes of care in accordance with national clinical guidelines (11–14). The following areas are assessed by monitoring 16 processes of care: diagnosing schizophrenia, antipsychotic medical treatment, screening for metabolic syndrome, family intervention, psychoeducation, suicide risk assessment at discharge, and referral to postdischarge support (12,14). The processes of care were identified by a multidisciplinary expert panel representing national scientific societies and professional organizations (12,14). The registry also contains information on covariates, including gender, age, substance abuse (alcohol, illicit drugs, benzodiazepines, or cannabis), and Global Assessment of Functioning (GAF) score, which is a single rating scale used to evaluate the overall psychosocial functioning of a patient. The score ranges from 1 to 100, with 1 representing the poorest functioning and 100 representing the best functioning (15,16). All data are prospectively collected from documentation in medical records by using a registration form with detailed instructions. For inpatients with schizophrenia, the processes of care are registered at the time of discharge from the psychiatric hospital unit by the health care professionals responsible for the care of the individual patient. For long-term inpatients, the processes are registered once a year on the date of admission. In cases of missing documentation, delivery of the processes of care is registered as unknown.

Of the 16 total processes of care monitored by the Danish Schizophrenia Registry, ten are relevant to inpatients (12,17–24) (Table 1). Changes occurred in the set of monitored processes during the study period, because some processes were added and others were omitted. Thus the time periods for the different processes of care varied, which was taken into account in the analyses. Staff contact with relatives and psychoeducation were assessed only among patients for whom the health care professionals considered the specific processes of care to be relevant. The remaining processes were a priori considered relevant for all patients.

Study Population

The study population included all patients (age 18 and older) admitted with schizophrenia and registered in the Danish Schizophrenia Registry between January 1, 2004, and December 31, 2011. Schizophrenia was defined according to the ICD-10 (codes F20.0 to F20.99) (25). The study included “incident inpatients,” including long-term inpatients. “Incident patients” were defined as individuals who had been given a diagnosis of schizophrenia within the preceding 12 months. A total of 14,228 patients were identified. The patients were admitted to 270 different psychiatric hospital units with 61,076 admissions during the study period. However, we included only the first recorded date of discharge per patient at each psychiatric hospital unit per year because readmissions of the same patient are not assumed to be comparable with a similar number of one-time admissions of different patients. We excluded psychiatric hospital units with fewer than 20 recorded admissions during the entire study period and, therefore, the patients admitted to these specific units (N=400). We made this exclusion because units with only sporadic admissions were more likely to have inadequate routines for reporting data to the Danish Schizophrenia Registry. Furthermore, we excluded patients with registration errors, including those who were registered on several psychiatric hospital units with the same date of discharge (N=164) and patients with missing information on all included processes of care (N=20). Of the remaining 13,644 patients with 33,916 admissions, we identified 3,209 “incident patients” with 3,555 admissions to 126 different psychiatric hospital units.

Psychiatric Hospital Units

On the basis of the entire study population, including 13,644 patients with 33,916 admissions, the admission volume was defined as the average number of admissions to each psychiatric hospital unit per year from 2004 to 2011. The admission volume was divided into four quartiles and referred to in this study as low volume (quartile 1, ≤75 admissions per year), medium volume (quartile 2, 76–146 admissions per year), high volume (quartile 3, 147–256 admissions per year), and very high volume (quartile 4, >256 admissions per year).

Since 1969, the Danish Psychiatric Central Research Registry (PCRR) has maintained a systematic collection of all psychiatric contacts for the entire population of Denmark. The registry contains dates of admission and discharge, start and end of outpatient treatment, and place of treatment (the specific psychiatric hospital unit is identified) (26,27). It is mandatory for all Danish psychiatric hospital units to report data to the PCRR. To characterize the clinical infrastructure, each psychiatric hospital unit assigned to the admission volume quartiles was linked with data from the PCRR. In 2011, the median length of stay was 8.0 hospital days in low-volume units and 5.5 hospital days in very-high-volume units (28).

Statistical Analyses

The use of processes of care by each patient was assessed both overall and individually for each process. The overall quality of care was calculated by dividing the number of received processes of care by the number of relevant processes for the individual patient. Missing and irrelevant data were excluded from the calculation of the use of the individual processes of care.

The association between admission volume and the quality of care—that is, the overall quality and the individual processes of care—was examined by using binary regression. In this case, the association between admission volume and the overall quality of care was examined by setting a pragmatic cut point of 80%; that is, high overall quality of care was defined as a patient’s receipt of 80% or more of all relevant recommended processes of care. The analyses were also repeated with alternative cut points varying from 60% to 90%. All 95% confidence intervals [CIs] were corrected for clustering of patients within psychiatric hospital units by using robust estimates of the variance. The analyses were stratified according to patient characteristics: gender, age, substance abuse (alcohol, illicit drugs, benzodiazepines, or cannabis), and GAF score. A two-sided p value of ≤.05 was considered significant.

Results

Characteristics of patients in the four admission quartiles did not vary considerably, but the proportion of men was greater than the proportion of women in all quartiles (Table 2). Most patients were between the ages of 18 and 29, and about a third had a GAF score <40, meaning that the patients were dysfunctional in several areas. In addition, 39% of the patients had some form of substance abuse (alcohol, illicit drugs, benzodiazepines, or cannabis).

Table 2 also presents data on the overall quality of care by admission quartile, shown as the proportion of relevant processes of care received by the patients. The largest proportions of patients were in the group receiving 60% to 79% of the relevant recommended care processes, and 26%−36% of the patients received ≥80% of the recommended processes. Table 3 shows the association between admission volume and the overall quality of care. Patients admitted to very-high-volume psychiatric hospital units were 1.40 times more likely than patients admitted to low-volume units to receive high overall quality of care (≥80% of the relevant recommended processes). When alternative cut points between 60% and 90% were used, the association was likewise confirmed, with risk ratios for high overall quality of care ranging between 1.22 (CI=1.06–1.41) and 1.69 (CI=1.04–2.76) when very-high-volume units were compared with low-volume units.

Table 4 presents data on the use of the ten individual processes of care by admission volume quartile. The proportion of patients receiving the individual processes varied from 25% to 93%. Table 5 shows the association between admission volume and the individual processes of care. Patients admitted to very-high-volume psychiatric hospital units were more likely to receive several of the individual processes. In particular, patients admitted to very-high-volume units were 1.05 times more likely than those admitted to low-volume units to be assessed for psychopathology by a specialist in psychiatry and 1.16 times more likely to receive psychoeducation. Moreover, patients admitted to high-volume units were 1.14 times more likely than those admitted to low-volume units to receive suicide risk assessment at discharge. In contrast, staff contact with relatives was 1.10 times more likely for patients admitted to medium-volume units than for those admitted to low-volume units. No clear associations were found for the remaining processes of care. We found no evidence of systematic interaction when the analyses stratified patients according to gender, age, substance abuse (alcohol, illicit drugs, benzodiazepines, or cannabis) and GAF score (data not shown).

Discussion

Our results show that patients with recent diagnoses of schizophrenia admitted to very-high-volume psychiatric hospital units had the best chance of receiving guideline-recommended processes of care. When these patients were admitted to very-high-volume units, they were more likely to receive high overall quality of care and to receive several of the individual processes, including assessment of psychopathology by a specialist in psychiatry and psychoeducation. In high-volume units, these patients were more likely to receive suicide risk assessment at discharge. In contrast, staff contact with relatives was more likely for patients admitted to medium-volume psychiatric hospital units than to low-volume units. No clear associations were found for the remaining processes of care. These results may imply that very-high-volume and high-volume psychiatric hospital units provide better quality of care—at least in some areas—for admitted patients with recently diagnosed schizophrenia, compared with low-volume psychiatric hospital units.

Strengths and Limitations

The strengths of this study include the nationwide population-based design, with prospectively collected data and a relatively large study population. The Danish Schizophrenia Registry has high coverage—for example, it was estimated in 2011 to include records for 93% of all inpatients with schizophrenia in the Danish psychiatric health care system (12). Confounding is considered to be of minor importance for the study findings because the included processes of care in principle are relevant for admitted patients independent of admission volume.

Data validity is always a relevant concern in registry-based studies. The data in the Danish Schizophrenia Registry are collected by several clinicians during routine clinical work, and registration errors and variation in registration practice may occur. Registering processes of care once a year for long-term inpatients constitutes a particular challenge in regard to accurate reporting of data. However, extensive efforts are made to ensure data validity by designating key persons in each psychiatric hospital unit with the responsibility of securing correct data collection and reporting. In addition, uniformity and validity are ensured by detailed instructions, with explicit data definitions included in standardized registration forms and regular structured audit processes carried out on a local, regional and national basis. The audit processes critically assess the quality of the data and provide continuous feedback to the psychiatric hospital units (12). However, it should be noted that inconsistencies in the documentation practices in Danish psychiatric hospitals have been reported. Insufficient medical records for patients with schizophrenia may prevent the use of psychiatric medical records as the gold standard when validating registry data. This makes it difficult to draw firm conclusions about the validity of data recorded in the Danish Schizophrenia Registry (29). It must be recognized that the quality of care was assessed by using recommended processes of care, whereas no information on clinical outcomes was provided. Thus only changes in delivery of the processes were documented. Furthermore, the quality of care was simplified because of the dichotomous data. In a clinical setting, variations may occur in delivering processes to patients; however, detailed instructions included in the registration forms are intended to reduce such disparities.

Comparison With Other Studies

The relationship between admission volume and quality in mental health care has been explored in only a limited number of studies, and differences between the studies make direct comparisons difficult. Nevertheless, a U.S. study examined the relationship between admission volume per unit and five mental health care quality measures (5). The quality of care was measured by adequate seven-day and 30-day follow-up after admission for mental illness, the appropriateness of antidepressant medication management by prescriptions filled during a 12-week period and for at least six months, and at least three follow-up visits in the 12 weeks after diagnosis of a new episode of depression. According to all five measures, patients admitted to units with a low admission volume received poorer quality of care than those admitted to units with high admission volume. Other studies have assessed the quality of mental health care by length of hospital stay and readmission (4,6,7). The studies found that high admission volume per psychiatrist was associated with both a shorter stay and a higher readmission rate. However, increased length of stay and readmissions may not necessarily indicate poor-quality mental health care. On the contrary, an extended stay may be required if a patient is severely ill or psychotic, and readmissions may be essential for the stabilization of psychiatric patients.

In this study, we examined admission volume defined as the average number of admissions in each psychiatric hospital unit per year from 2004 to 2011. This measure reflects the experience and capacity of an organization and not of individual health care professionals. In contrast, other studies have examined provider-level volume, defined as the total number of mental health admissions by a given psychiatrist (4,6,7). It should be noted that this “psychiatrist volume” reflects something other than the departmental volume used in our study, because it primarily reflects the experience of the individual psychiatrist rather than the organization within which the psychiatrist works.

Several underlying mechanisms in high-volume psychiatric hospital units, including specialization, greater clinical experience, and better resources, might explain the observed association between very-high and high-volume psychiatric hospital units and the highest overall quality of care for patients with recently diagnosed schizophrenia. Moreover, a greater number of beds and shorter stays may characterize high-volume psychiatric hospital units. In this study, the median length of stay did not vary substantially between low-volume and very-high-volume units. It must nonetheless be recognized that unmeasured variables may have affected the quality of care.

A contributing factor, at least in the Danish context, may also be the Danish specialized assertive intervention program (OPUS). OPUS focuses on early detection and assertive community treatment for patients with first-episode schizophrenia spectrum disorders and comprises a two-year treatment, including intensive psychosocial assertive community treatment, family treatment, social skills training, multifamily groups, and psychoeducation. Studies show that OPUS improved patients’ adherence to treatment and clinical outcome, including improved effects on negative and psychotic symptoms, greater treatment satisfaction, reduced secondary substance abuse, and a lower dosage of antipsychotic medication after two years of treatment (30,31). However, most of the positive clinical effects were not sustained three years after the end of OPUS treatment (32), which indicates that intensified and structured organization of interventions is likely to improve care. Increased focus on and more intensive treatment of patients with newly diagnosed schizophrenia as a result of OPUS may correlate with the results of this study because OPUS treatment is primarily offered on larger psychiatric hospital units. However, this factor must be interpreted as intermediary because of the greater capacity and resources of those units.

Of note, our study showed that staff contact with relatives was more likely for patients admitted to medium-volume psychiatric hospital units than to low-volume units. In contrast, contact with relatives was less likely on very-high and high-volume units, although these findings were not significant. The explanation for this difference may be a greater awareness of relatives in low-volume units and thus greater involvement of relatives.

We encourage further studies of the association between admission volume and quality of care to confirm the generalizability of our findings for specific mental disorders, including schizophrenia. A final question remains about whether qualified diagnosis, treatment, and care in very-high-volume and high-volume psychiatric hospital units are cost-effective. In this case, the cost of providing higher-quality care, health consequences, and specific short- and long-term costs need to be clarified.

Conclusions

This nationwide population-based cohort study demonstrated that patients with newly diagnosed schizophrenia who were admitted to very-high-volume psychiatric hospital units were more likely to receive care in accordance with clinical guidelines. However, the absolute differences were modest for most of the examined processes of care, and further studies are warranted to determine the clinical implications of the possible differences in care.