Studies using the Psychiatric Case Register Middle Netherlands
Routine databases like the PCR-MN are attractive for research because of their large size, longitudinal perspective, and practice-based information. Several studies have already been conducted using data from the PCR-MN case register. From an epidemiological perspective, etiological, intervention and prognostic studies are feasible with and without using links to other sources for additional information. Below we will give a number of examples using the linked PCR-MN database.
One study examined whether being diagnosed with both an anxiety disorder and a depression increases the hazard-ratio for death during follow-up compared to only one of these disorders. All cases were linked to the causes of death register of Statistics Netherlands in a 10 year follow-up cohort study, together with a random control group, after the initial diagnosis. An increased risk was demonstrated in patients with an anxiety disorder and in patients with a depression. However, the hazard ratio among people with both disorders was similar to those with only a depression (Laan W, Termorshuizen F, Smeets HM, Boks M, Wit NJ, Geerlings MI. A comorbid anxiety disorder does not result in an excess risk of death above the risk of death associated with a depressive disorder alone. Submitted
). A second study was performed on the risk of several psychiatric disorders among minorities in Utrecht compared to Dutch natives. All cases with a relevant psychiatric diagnosis were subsequently linked to their birth records in the database of Statistics Netherlands. The findings in this cohort study showed significantly increased relative risks for treatment because of a depression or a non-affective psychosis among all immigrant groups [7
In another case-control study, we investigated whether the use of corticosteroids was associated with a decreased risk of subsequent psychosis. All glucocorticosteroid prescriptions prior to the first registered psychotic episode in patients from the PCR-MN database were extracted from the AHD. Then the number of defined daily doses (DDDs [8
]) was compared to a control group and the results showed that the risk of psychosis was lower in those who have used these drugs [9
In a prognostic study at last the effect of non-compliance to anti-psychotic drugs on the risk of a psychotic relapse was measured. From all cases of the PCR-MN database that were discharged from an inpatient clinic of Altrecht, the reimbursed antipsychotic drugs were extracted from the AHD. The results of this study indicated a significantly increased risk for relapse with increasing non-compliance using a continuous compliance scale [10
Strengths and weaknesses
A major strength of the PCR-MN database is that it enables researchers to do longitudinally studies without interfering with regular care, thus allowing cohort studies nested within the PCR-MN main cohort, but also nested case-control studies. In a nested cohort, patients are selected based upon a determinant in the cohort; in a nested case-control study, patients are sampled based on the outcome.
Another strength is that the PCR-MN covers all psychiatric healthcare provided in the region. Patients referred to another psychiatric hospital in the region can still be identified in the case-register.
Also, because the registration of diagnosis and psychiatric contacts is based upon the financial databases of the psychiatric centers, the registration is very accurate. If the diagnoses are not registered well the costs will not be reimbursed by the insurance company.
Finally, because the PCR-MN case register is one of four Dutch psychiatric case-registers it is also possible to compare results from the Utrecht region to other regions covered by one of the other case-registers or even to analyze aggregated data from the four PCRs together.
A disadvantage of the registration in the PCR-MN is that some patient characteristics are subject to changes during follow-up. For example, the postal-code can result in loss to follow-up, because a patient may move to another part of the region and will not longer be treated at the same psychiatric hospital. The particular patient would then be identified as a new patient increasing the incidence and prevalence of disorders in the region. Also, patients who die or move from the region are not reported as such to the PCR-MN. It is difficult to determine to what extent this loss to follow-up occurs
A second limitation is that diagnoses in the PCR-MN are clinical diagnoses and not the result of a standardized structured diagnostic questionnaire. Also the reason to terminate psychiatric healthcare is not registered i.e. it is unknown if the patient is cured.
From 2010 onward we aim to also include the Axis 5 (Global Assessment of Functioning (GAF) score) of the DSM-IV in the PCR-MN database. The GAF score is a numeric scale with a range of 0 to 100 that aims to describe the social, occupational, and psychological functioning of a person.
In the coming years, the identification of patients in the database based upon the gender, postal code and date of birth will be replaced by a registration based upon the Dutch social security number (BSN). This requires the use of a 'third trusted party' to protect the privacy of patients in the case-register, which is currently being explored.