Introduction and Background:
Valproate and related substances (sodium valproate, valproic acid, valproate semisodium, valpromide, and valproate magnesium) have been licensed since 1967 to treat epilepsy and since 1995 to treat bipolar disorder in Europe. Due to the risk of malformations and developmental disorders in babies exposed to valproate in utero, in 2014 restrictions on the use of valproate containing substances were applied to the label and risk minimisation measures were implemented. The results of various utilisation studies in Europe showed that the pattern of use of valproate in women of child bearing potential (WCBP) had not changed significantly over 2014-2016. Therefore, the Pharmacovigilance Risk Assessment Committee (PRAC) requested the marketing authorisation holders (MAHs) to continue an ongoing drug utilisation study conducted in France, Germany, Spain, Sweden and in the UK and to amend the study design to include the updated conditions of use for valproate and related substances. Databases capturing variables that would be used to assess the compliance to Pregnancy Prevention Program (PPP) were considered for the extension study including France, Germany, Spain, Netherlands, UK and Sweden.
Research Objectives:
The research question is, “”What is the impact of the implementation of the Risk Minimisation Measures and Pregnancy Prevention Programme (PPP) on the use of valproate in women of child bearing potential (WCBP) in Europe?””
The aim of this study is to describe the prescribing practices before and after the dissemination of the new risk minimisation measures (Q2 2018 – Q4 2018) in Europe and to assess the effectiveness of these measures.
Primary objectives:
? To describe and compare the prescribing practices in WCBP receiving valproate during the pre- and/or post-implementation period with respect to elements of the PPP (where available in each of the data sources) separately:
o Use of contraceptives without interruption during treatment
o Laboratory pregnancy tests before treatment
o Treatment reviews by a specialist at least once per year (using a proxy of consultation by a specialist as a marker for treatment review)
o Specialty of prescribing physician at initiation
? To describe and compare proportion of patients for which all elements of the PPP measurable with this DUS are fulfilled, during the pre- and/or post-implementation period
? To describe and compare the incidence of valproate exposed pregnancies, and characteristics of exposed pregnancies during the pre and post implementation period
Secondary objective:
? To describe and compare the following prescribing patterns in WCBP during the pre- and post-implementation periods, overall and in subgroups of patients:
o Demographic characteristics of WCBP prescribed or dispensed valproate in oral form
o Indication for use (epilepsy, bipolar disorder and other)
o Use of prior medication for valproate indications
Study Design:
This is a non-interventional longitudinal retrospective cohort study of WCBP exposed to valproate, conducted with secondary data obtained from electronic medical records or administrative healthcare databases. The study design has been informed by logic models describing and defining the parameters to be measured in the study in the pre- and post-implementation periods
A cohort of patients initiating valproate will be defined with a pre- and post-study design. This DUS extension will cover the 3-year pre and post-implementation period after the implementation of new (2018) risk minimisation measures and will be based on pre-existing databases recording prescriptions, patients’ demographics and diagnosis, in the selected European countries (France, Germany, the Netherlands, Spain, Sweden and the UK).
Population:
The study population will include all WCBP prescribed valproate during the pre-defined periods identified from the healthcare databases selected. Users will be defined as:
? Prevalent users (includes recurrent and incident users): At least one valproate prescription issued (UK) or dispensed (all other countries).
o Recurrent users: At least one valproate prescription issued (UK) or dispensed (all other countries) in the 12 months prior to the index date.
o Incident users: No valproate prescription issued (UK) or dispensed (all other countries) in the database in the 12 months prior to the index date.
First ever users ( a sub-group of incident users): No valproate prescription issued (UK) or dispensed (all other countries) at any time in the database prior to the index date.
Among the study populations, sub-populations of interest will be: pregnant women, valproate indication (epilepsy, bipolar disorder and other).
Inclusion criteria
? Female gender,
? Age 13-49
? At least one prescription of valproate in the selected databases during the pre-defined periods
Exclusion criteria
No exclusion criteria will be applied.
Variables:
The exposure is defined as one or more prescriptions of valproate during one of the study periods. The following variables will be considered: patient age, use of medications related to valproate indication ever recorded before valproate initiation, prescriber specialty, as well as PPP-specific variables such as contraceptive use and – if available – pregnancy testing. In addition, information on exposed pregnancies and outcomes will be provided.
Data Sources:
The following established longitudinal data sources will be utilised for data extraction:
? France: Système National des Données de Santé (SNDS)
? Germany: German Pharmacoepidemiological Research Database (GePaRD)
? Netherlands: PHARMO database network (Out-Patient Pharmacy Database and Hospitalization database) and Perinatal Registry (PRN)
? Spain: SIDIAP, the Information System for the Development of Primary Care Research in Catalonia.
? Sweden: national drug-, patient-, and birth registers
? UK: Clinical Practice Research Datalink (CPRD)
Study size:
All WCBP receiving valproate available in pre-defined periods of the selected databases will be included in the analysis.
The number of WCBP receiving valproate in the SIDIAP database was 2,700 in Spain in the most recent data periods available.
Data analysis:
Given the study objectives the analyses will be mainly descriptive and will be conducted by country and study time periods (both pre-implementation period and post-implementation period).
When applicable, quantitative variables will be statistically compared with a Student’s t-test (parametric test) or Wilcoxon signed-rank sum test (non-parametric test, when necessary). Categorical variables will be statistically compared with a Pearson Chi2 or with Fisher’s exact test (if the expected frequency lower or equal to 5 for one or several cells). Each statistical test will be two tailed and will not be adjusted for multiple comparisons. As appropriate, 95% confidence intervals will also be calculated.
In addition, an interrupted time series (ITS) analysis will be considered in case the conditions for this analysis will be met. Results for key variables will be presented and visualised over time where patient numbers permit to provide insights on trends during the pre- and post-implementation periods.
