Applied Statistics Mini-Project: The Effectiveness of a
Peer-Supporter Programme to Manage Opiate Misuse in
England
Introduction to Statistical Thinking and Data Analysis
MSc in Epidemiology and MSc in Health Data Analytics, Imperial College London 23 December 2019 to 10 January 2020
Background
Dependence and abuse of prescription opiate drugs is on the cusp of a public health emergency in England. Deaths from opiate misuse have increased from 1100 in 2012 to over 2200 in 2018. Opioid prescriptions for chronic pain more than doubled during this period.
In addition to risk of death resulting from drug overdose, opiate misuse is linked to a number of other negative health and social wellbeing consequences, such as job loss, family breakdown, and homelessness.
Approximately 145,000 clients sought help from the NHS alcohol and drug treatment services for opiate substance use in 2016-17. A major challenge is sustaining positive rehabilitation treatment outcomes following discharge, including avoiding relapse of drug misuse and recovering from the broader social harms wrought by substance use. Identifying and expanding successful interventions to achieve this is a major priority for the NHS given the expanding epidemic of opiate misuse in England.
Data
The data set consists of the results of a randomised controlled trial undertaken by NHS England to study the effectiveness of a peer mentoring programme following completion of rehabilitation therapy. The trial was undertaken in four regions of England: North West, South East, West Midlands, North East. The trial enrolled 864 patients presenting for drug rehabilitation treatment for opiate addiction between April and August 2017.
Upon entry to rehabilitation therapy, patients were randomised to receive the peer mentor support programme or normal standard of care following completion and discharge from therapy. Randomisation was stratified by region of England. The peer mentor support intervention involved assignment of a rehabilitated addict residing in the same local neighborhood. Following discharge from therapy, patients were introduced to their peer mentor and met or spoke to their peer mentor on a weekly basis and participated in monthly social events.
Patients were followed until study closure in July 2019. There were two outcomes of interest from the trial: (1) whether the peer mentoring programme reduced drug relapse, and (2) whether the peer mentoring programme facilitated integration back into the local community and mitigated the longer social and economic impacts of drug addiction. Relapse was tracked on a quarterly basis from treatment discharge until the study closure. When relapse occurred, it was recorded as the number of days from discharge to relapse.
Social wellbeing was measured through a constructed wellbeing score, a standardized continuous composite score capturing quality of employment, housing, social relationships and connectedness, and mental health. The social wellbeing outcome was measured at a study interview approximately one year after treatment discharge.
1
Pre-specified sub-group analyses included analysis of main trial outcomes according to stratification group (region and residence), sex, and injecting history.
In addition to monitoring of trial outcomes, baseline data were collected about the following:
• Basic demographic characteristics: age, sex, region of residence, urban/rural residence type.
• Treatment entry characteristics: treatment referral source, housing situation, duration of opiate use,
injecting history.
• Initial treatment success: whether initial treatment programme was completed free of substance
dependence.
Aims
NHS England have commissioned you to prepare a report on the trial outcomes.
Firstly, they are interested in analysis of the baseline enrolment data to better understand factors associated with treatment programme success.
1. NHS England are particularly interested in whether duration of opiate use before entering therapy and injecting history are associated with treatment programme success.
Secondly, report the main trial outcomes about the effectiveness of the peer mentor programme:
2. Whether the peer mentor programme affected relapse to opiate use following treatment discharge.
3. Whether the peer mentor programme affected social wellbeing one year following treatment discharge.
Instructions
Design and conduct a statistical analysis to address the three questions outlined in the Aims above.
Your report should not exceed 2500 words (excluding figures, tables, and references) and should take the format of a journal paper with four sections Introduction, Methods, Results, and Discussion.
Your report should include a structured abstract of not more than 250 words with sections Background, Methods, Results, and Conclusions.
Use figures and tables to aid the communication of your results, but not excessively. There is not a fixed requirement or limit for the number of figures and tables, but a typical paper would not have more than 5-6.
Focus your efforts on the design, analysis, and interpretation of your statistical analysis. You are not expected or encouraged to substantially research this topic outside of the information provided here and comprehensiveness of referencing will not form part of the project marks. You do not need to reference statistical methods or software covered in the ISTDA course.
You may optionally submit an annotated R script alongside your report, which may be used to clarify what statistical analysis was conducted if there is ambiguity in the report. The R script is not required, and a top marked paper would be self contained and fully describe the methods such that consulting the R script is not required.
The exam is to be completed independently. The project should not be discussed with colleagues on the course or others outside of the course.
2
Dataset description
The study dataset is contained in the file peermentor.csv. The table below describes each of the variables in the dataset.
Variable
id
gender
age
region
residence
duration_use
injecting_status
housing_status
referral_source
rehab_success
intervention
wellbeing1yr
relapse_days
last_study_obs
last_obs_outcome
Description
Participant identification number
Gender (Male, Female)
Age in years
NHS region of residence (North East, North West, South East, West Midlands) Urban/rural residence
Duration of regular opiate use before entering therapy (months)
Injecting status prior to entering therapy (Never injected, Previously injected, Currently injecting)
Housing status at time of entering therapy (No problem, Housing problem, Urgent housing problem, Other)
Source of referral to therapy (Self/family/friend, Health service/social care, Criminal justice)
Abstained from opiate use through duration of rehabilitation therapy programme
(Yes, No)
Post-therapy intervention arm allocation (Peer mentoring, Standard of care) Wellbeing composite score measured one year post therapy
Days post therapy of first drug relapse
Days post therapy of last study visit
Final study outcome (Study completion, Loss to follow-up)
3
Reviews
There are no reviews yet.