Exploring clinical pathways in large-scale electronic healthcare record data: applying sequence analysis to pre-arthroplasty pathways in osteoarthritis

Study type
Protocol
Date of Approval
Study reference ID
24_003806
Lay Summary

Many patients get knee replacement for osteoarthritis, but we don't know exactly what kind of medical care they receive leading up to it. Even though there are guidelines for different treatments, we don't know how people actually experience care – when they get it, and what kind of treatments they get.

To understand this, we will look at the medical records of people who have had a knee replacement. We will study what healthcare they received over the ten years prior to their joint replacement, including doctor visits, medicines they were prescribed, and other healthcare events. The goal is to simplify this information into easy-to-understand types of care patterns. Then, we will see which types of care patterns are more common in different groups of people, like older or younger patients, men or women, and those who are poorer or richer. This way, we hope to learn more about the early signs of knee problems and find out if there are differences in how people are taken care of.

Technical Summary

Background:

Primary total knee replacement is the definitive and effective intervention for end-stage knee osteoarthritis. However, there is a need to enhance conservative care earlier in the condition. The NICE outlines a care pathway from diagnosis, encompassing self-management, non-surgical treatments (information, exercise, weight loss), and various nonpharmacological and pharmacological interventions before considering joint surgery. The actual paths patients take are unclear but likely to be complex, variable, and influenced by multiple factors. Understanding real-world healthcare utilisation patterns is essential for patients, healthcare professionals, and health service planners.

Aim:
Identify different care patterns before primary total knee replacement for osteoarthritis and explore variations and inequalities in these patterns.

Objectives:
1. Derive various care pattern typologies through sequence analysis of retrospective care records of the above case series. Patients who had knee replacement surgeries will be identified using linked data from CPRD- Aurum and HES-APC data. Primary and secondary care strategies will be identified using linked data from CPRD-Aurum and HES-APC data.

2. Explore the association between care pattern typologies and patient characteristics (e.g., age, sex, ethnicity, geographical region, body mass index, comorbidities, area-level deprivation, and lifestyle risk factors like smoking and drinking). Patients’ socioeconomic characteristics will be identified using linked data from CPRD-Aurum and patient-level Index of Multiple Deprivation.

Methods

Study 1: A descriptive analysis will be applied to estimate the prevalence of each type of exposure in the ten years leading up to the index condition, using different time intervals (1, 3, 6 and 12-month time windows).
Study 2: Sequence analysis will be used to construct distinct typologies of care patterns.
Study 3: Multinomial regression analyses to test the association of each derived typology and patient characteristics.
Study 4: Examine trajectories of care strategies over time using latent class growth models and compare the results with sequence analysis results.

Health Outcomes to be Measured

Study 1
Prevalence of exposures at different time interval.

Study 2
Distinct typologies of care patterns prior to primary total knee replacement.

Study 3
Association of patients’ characteristics with different typologies of care patterns and outcome will be the odds ratios and confidence intervals.

Study 4
Comparison between trajectories of care strategies derived using latent class growth models and care patterns derived using sequence analysis.

Collaborators

Dahai Yu - Chief Investigator - Keele University
Smitha Mathew - Corresponding Applicant - Keele University
Emma Parry - Collaborator - Keele University
George Peat - Collaborator - Keele University
James Bailey - Collaborator - Keele University

Linkages

HES Admitted Patient Care;Patient Level Index of Multiple Deprivation