# Statistical analysis plan of the study titled “A deprescribing programme aimed to optimize blood glucose-lowering medication in older people with type 2 diabetes mellitus — the OMED2 study: a randomized controlled trial”

## TL;DR

This paper outlines the statistical methods for a study testing a deprescribing program to optimize diabetes medication in older patients.

## Contribution

The paper provides a detailed statistical analysis plan for evaluating a deprescribing program's impact on diabetes complications.

## Key findings

- The study will use generalized linear mixed models to compare diabetes complications between intervention and control groups.
- Both intention-to-treat and per protocol analyses will be conducted to assess the program's effectiveness.
- The plan includes methods to evaluate the implementation and cost-effectiveness of the deprescribing program.

## Abstract

The OMED2 (Optimization of Medication in Elderly with Diabetes) study addresses the effect and implementation of integrating a deprescribing programme (DPP) in general practice. The aim of the DPP is to reduce glucose-lowering medication (SU/insulin) in overtreated older patients. The protocol for this study has been published previously. This statistical analysis plan (SAP) contains a more elaborate outline of the (statistical) methods we plan to use for data analysis.

The OMED2 study is a randomized mixed-methods study with a 2-year follow-up period that compares the effect of the implementation of a DPP in general practice to regular care (control). In this SAP, we report on the (statistical) approaches that we plan to use to address the study objectives. The main objective of the OMED2 study is to examine the effect of the implementation of the DPP on diabetes complications, whereby the total number of diabetes complications related to undertreatment and overtreatment will be summed. Generalized linear mixed models with a Poisson distribution and the DPP as the main determinant will be used to test whether the total number of diabetes complications occurring from the start of the 2-year follow-up until the end of follow-up differs between intervention and control. The incident rate of the number of diabetes complications will be calculated to correct for possible differences in follow-up duration. The model will also include a random effect variable to allow for possible clustering effects by general practice. We will perform intention-to-treat analyses, which include all patients eligible for deprescribing, as well as per protocol analyses, which omit patients who were not deprescribed in the intervention arm. Additionally, approaches to study the implementation of the DPP and the cost-effectiveness of the implementation are outlined in the SAP.

ISRCTN Registry ISRCTN50008265. Registered on 1 November 2024.