Understanding interactions between chronic kidney disease and mental health to improve holistic, intelligence-led care

Description

Background

Individually, CKD and mental health conditions account for a significant burden of ill health. The global CKD prevalence was approximately 9%, with regional variations. In the UK around 3 million people in England were living with CKD. In the UK around 20% of UK adults experienced depression or anxiety symptoms. It is predictable to expect a high prevalence of the combination of both chronic conditions. Individuals with CKD are at an increased risk of experiencing mental health conditions compared to the general population. Studies have shown a higher prevalence of conditions such as depression, anxiety, cognitive impairment, and sleep disorders among CKD patients. There is an increasing recognition of the need to consider the mental health implications of CKD and to incorporate mental health screening, assessment, and interventions into the care of CKD patients as a way for healthcare providers to improve the wellbeing and outcomes of individuals living with CKD, with integrated care and tailored interventions. We currently do not know the prevalence of the co-occurrence of specific mental health conditions and CKD, their spatial variation and stratification by different population groups (e.g. socioeconomic). There is a paucity of longitudinal multidimensional data resources that can describe the temporal aspects of disease progression, onset of comorbidities and frequency of interaction with services.

Objectives

  • To describe the prevalence and incidence of mental health conditions among CKD patients and their distribution by socio-demographic groups in the population of Cheshire and Merseyside.
  • To explore the potentially causal effects of CKD on mental health and modifiers.
  • To explore the potentially causal effects of mental health on CKD development considering interactions with key CKD risk factors (hypertension and diabetes), effects of antidepressants and antipsychotics and socio-economic disadvantage.
  • To describe the progression of CKD in people with mental health conditions, considering interactions with key CKD risk factors (hypertension and diabetes), effects of antidepressants and antipsychotics and socio-economic disadvantage.
  • To explore the effect of adding mental health conditions to the Kidney Failure Risk Equation vs its predictive performance.

Research methods

This research, using population-wide data for the 2.7m residents of Cheshire and Merseyside, will cover a representative 36% of the population of North West of England, with findings generalisable to the national context. It will offer for the first time a quantitative estimation of the burden of co-occurrent mental health and chronic kidney disease, advancing the knowledge of the relationship and mechanisms of interaction between two diseases and informing the development of integrated care models that have a major potential to optimise well-being of renal patients. Additionally, this project will develop a set of open data analytic resources for measuring CKD interactions with comorbidities and monitoring integrated care outcomes.

The project will use group-based multi-trajectory cluster models to identify trajectories of CKD progression within distinct subgroups with recorded mental health conditions, and synthetic controls for the development of risk prediction models. Longitudinal models will need to consider fixed and time-varying covariates, and repeated outcomes. The student will gain experience in a variety of non-trivial statistical methods including join-models. This project will study complex interactions of conditions through variably sparse clinical observations over time. Here, joint models encode multivariate longitudinal data on the course and interrelationship of different indicators of disease progression. Group-based multi-trajectory models are a class of multivariate joint modelling techniques – these allow subdivision of the data into a finite number of clusters that present similar trajectories of disease progression. Understanding of factors that determine differential trajectories and possible inequalities in disease progression will help select factors to include in a risk model. The project will consider a more robust data-driven synthetic control method to build people without CKD and mental health. The student will use longitudinal data on risk factor exposure, renal function, and mental health trajectories to improve the model further.

Learning opportunities

The student will be able to learn the fundamentals of statistical analyses and data processing and progress to advanced biostatistics, causal inference and data engineering knowledge while applying them to real-world data. The novel project addresses a knowledge gap and major clinical needs of global epidemics of CKD and mental health disorders. This collaborative project includes supervision from clinicians and clinical data science researchers. Data science, including advanced biostatistical modelling, can help identify at-risk patients and help target early intervention. However, these methods are underutilised due to a lack of close, iterative communication between clinicians and data scientists – coupling explaining variation with understanding mechanisms. To fulfil the project's objectives, the PhD student will learn data science methods and how to couple their clinical reasoning about mechanisms and contexts with disciplined extraction of statistical structure from the available datasets. This will also teach the student about clinical coding, routinely collected electronic health data engineering and information governance necessary to best use this sensitive data. Training will be adjusted to the student's background data literacy. The supervisory team can facilitate access to data, digital infrastructure, and guidance.