Mechanisms and consequences of depression-related multimorbidity over the life course: coordinated analysis of population and primary care data

Multimorbidity (MM), defined as the co-existence of two or more mental and/or physical chronic conditions, represents a complex challenge for clinicians, participants, and their families. Current preventive efforts have partly failed because they have focused on one disease at a time and too late in life. For this reason, there is a critical need to identify groups of individuals at risk of MM earlier in life, and to develop interventions to prevent MM and its adverse health outcomes.
A number of surveys have shown that not only clinical depression the most common illness in any MM cluster, it also is the most important for harming a patient's quality of life. Compared with most diseases that occur (and begin to cluster) in later years, depression tends to begin much earlier in adult life. As most existing research on MM has focused on older adults (over 60 years), this means an important component has been relatively ignored. What exactly is the relationship between the onset of depression in early adult years and later MM clusters? A life course approach that follows groups of people over many years is the best way of addressing this question.
In the proposed study we will analyse three of the world largest birth cohorts (all based in the UK) that contain around 40,000 participants who have been regularly studied in the decades since their birth. We will identify those participants with a record of depression in young and mid-adult years (20 to 64 years of age) and examine how their early onset psychological difficulties interacted with later illnesses, both physical and mental. In particular, we will explore these relationships across time and within different subgroups, defined by gender and socioeconomic background.
In addition, we plan to evaluate the clinical value of the cohort findings with data collected from routine general practice. The CPRD database now contains millions of patients' records and we anticipate examining in excess of 200,000 patients with depression. These large numbers should be sufficient to identify important mechanisms and consequences of depression related MM with reasonable precision. Dr Dregan (PI) has developed a protocol for accessing electronic health records from the primary care database, which will inform this part of the proposed study. Thus, we are confident that we will access to a large number of individuals in order to identify the important risk factors (including physiological, behavioural habits, treatment, and social support) for depression-related multimorbidity clusters across young and mid-adult years. We should be able to determine which depression-related MM clusters have the most harmful impacts during young or mid-adult years and test how these relationships are affected by gender, social class, and specific attributes of depression (such as sleep disturbance).
We will use advanced statistical methods to produce the most reliable results possible. With the large cohort and primary care databases, these important questions that will inform future clinical trials and improve public health, can be addressed at a considerably lower cost than a randomised trial. We will also use the requested resources to enable other researchers to use cohort data to address important research questions.

Background: Depression is the commonest individual disease in multimorbidity clusters and is the most significant in terms of affecting patients' quality of life. Unlike most diseases in multimorbidity clusters, however, depression tends to start earlier in adult life. Its subsequent course and interactions with later diseases, both physical and mental, is relatively under-investigated.

Methods: We will employ cohort data from 1946 National Study of Human Development, 1958 National Child Development Study, and the 1970 British Cohort Study to develop a platform for life course research on depression related multimorbidity. Primary care data from the Clinical Practice Research Datalink (CPRD) will be used to validate findings from the cohort studies. The analytical methods will include latent class analysis, phenotype network analysis, and multiple mediator mediation analyses. Analyses will be stratified be gender and social class to assess the heterogeneity among different population subgroups.

Expected outcomes: The proposed study will generate the necessary evidence to inform the development of future clinical trials aimed at preventing or delaying the onset of depression-related multimorbidity among highly at-risk subgroups. It will also generate novel understanding about modifiable factors with the greatest adverse impact on functioning and healthcare outcomes, supporting the NHS strategy to improving the outcomes of people with multimorbidity. We anticipate that links with King's Health Partners will facilitate the delivery of impact from this research. The research will have an international impact by developing a platform for comparative research methodology, for evaluating general population and clinical multimorbidity data, that can be applied across a wide range of topics of clinical and public health importance.

The most important potential of the proposed project is on informing future clinical trials aimed at improving the health and wellbeing of millions of people living with depression and coexisting disorders worldwide. Specifically, findings from combined analyses will inform future CPRD trials about those patient-reported consequences of depression-related multimorbidity that should be included as primary outcomes. The comprehensive data, covering both general and clinical populations, will provide robust evidence about subgroups of people at greater risk of adverse functional and health outcomes that should be a priority for targeted interventions.

Identification of multiple processes that underline the development of depression-related multimorbidity has the potential to support clinicians to make better informed decisions when discussing and recommending preventative therapies to patients with complex mental and physical health needs. For example, currently the focus on clinical practice is on treating depression with anti-depressant therapy. The rich and individual-level data proposed in this application would allow for an understanding of the main modifiable processes leading from depression to multi-morbid disorders across subgroups of people. If the findings of the study demonstrate that the anti-depressant drugs are associated with a specific morbidity, but not others it would inform the development of future interventions comparing the effectiveness of antidepressant treatment in patients with different clusters of depression-related multimorbidity.

The study findings may also influence on public health initiatives. Currently there is very limited evidence about what clusters of depression-related multimorbidity carry the greatest burden for subgroups and the healthcare system. In the context of depression-related multimorbidity the challenge is identifying the most effective way to prevent the onset of a co-existing disorders across subgroups of people at greater risk of multimorbidity. Our study exploring the combined effects of multiple processes would help identify those modifiable factors that should be the target of future preventative interventions in specific subgroups. This view may need to be revised if our study shows that is the combined effect of multiple processes that are more important for the development of multimorbidity rather than individual factors. Providing answers to such pertinent questions to practice and policy would lead to an exceptionally high and lasting impact on science and public health at a fraction of cost of a single randomised trial.

Despite increased acknowledgment by clinicians and policy makers, there is limited knowledge to guide the development of interventions with the greatest potential to improve the day to day functioning of people with depression-related multimorbidity, particularly among younger adults living with lifelong multiple disorders. The rich nature of our study data would also enable an understanding of those combination of depression-related multimorbidity clusters with the greatest negative effect on patients' wellbeing, such as physical capability and social inclusion. Such evidence would have important implications for patients and their carers understanding of those factors with the greatest negative impact on day to day wellbeing and functioning.

Another potential application of our study findings is with regards to the education and training of future and current healthcare professionals about the main determinants and consequences of depression-related multimorbidity across different subgroups, including women and socially deprived groups. For instance, it could inform the development of technologies (e.g. mobile apps) aimed at enhancing junior practitioners understanding about the complexities of caring for patients with co-existing depression and physical disorders.