As we get older, we become more likely to develop multiple long-term health conditions, which is referred to as ‘multimorbidity’. As a result of people living for longer and general improvements in medical care, more people that have a heart attack (also known as a ‘myocardial infarction’) will have multimorbidity. This is a problem because patients with heart attacks receive specialist care from expert heart clinicians that may not be best equipped to care for patients with multiple, non-heart-related conditions. In addition, heart attack treatments were developed in clinical trials for heart attack patients without other long-term health conditions, so we do not know how well they work for patients with different patterns of multimorbidity.
This study aims to use anonymous GP and hospital records to find out what effect multimorbidity has on heart attack treatments and outcomes. We hope to use this information to find ways to improve how patients with heart attacks and other long-term conditions are cared for. We will look at the number of patients that have multiple long-term conditions at the time of a heart attack and how this has changed over time, study combinations of conditions that may lead to worse outcomes, and evaluate the safety and effectiveness of heart attack treatment for people with multiple long term conditions. The results of this study will enable doctors to provide more tailored and effective medical care after a heart attack, taking into account the other long-term conditions a person has.
As a result of population ageing and general improvements in medical care, an increasing proportion of patients that have an acute myocardial infarction (MI) have multiple pre-existing long-term health conditions. This presents significant challenges. Firstly, post-MI care is delivered by highly specialised cardiology teams, which may not be equipped to manage patients with a greater burden of complex, multi-system disease. Secondly, there may exist a risk-treatment paradox: patients with multiple long-term conditions at greatest risk of post-MI adverse outcomes may be less likely to receive appropriate treatments, due to greater perceived risks of treatment-associated harm. Lastly, patients with multiple long-term conditions were excluded from many landmark trials of MI treatment, making the optimal treatment strategy uncertain.
We will perform a retrospective cohort study to evaluate the impact of multimorbidity (≥ 2 long-term conditions) on post-MI clinical outcomes. Our primary outcome of 1-year all-cause mortality will be derived from linked ONS Death Registry data and assessed by multimorbidity status, using multivariable flexible parametric survival models. The study will include individuals in CPRD Aurum with an index MI from 1st January 2000 onwards. We will use linked Hospital Episode Statistics data to ensure robust ascertainment of MI hospitalisation. In addition, small area-based Index of Multiple Deprivation (IMD) data will be used to adjust for the confounding effect of socioeconomic deprivation. Secondary objectives include: (i) quantifying the prevalence and determinants of multimorbidity among patients presenting with MI (and in a similar group of patients, without MI), (ii) characterising specific patterns of multimorbidity for individuals with MI (and non-MI comparators), and (iii) evaluating the impact of multimorbidity on guideline-recommended MI treatment equity, safety and efficacy. We envisage the findings of this study will enable the delivery of more appropriate, person-centred care to people with MI and multiple long-term conditions; improving their long-term clinical outcomes.
The following outcomes will be reported by multimorbidity status (i.e. those with ≥ 2 pre-existing long-term conditions at the time of MI, vs. those with < 2 pre-existing long-term conditions at the time of MI), for individuals that had an index MI during the study period (1st January 2000 onwards).
Primary outcome:
The cumulative incidence of all-cause mortality at one year of follow-up, ascertained using linked ONS Death Registry data.
Secondary outcomes:
1) The cumulative incidence of long term, all-cause mortality during the entire period of available follow-up, to be ascertained using linked ONS Death Registration data.
2) The length of stay during the index hospital admission (measured in days), to be ascertained using HES Admitted Patient Care data).
3) The cumulative incidence of hospital admissions for any cause at one year post-MI and during the total follow-up period, to be ascertained using HES Admitted Patient Care data.
4) The number of primary care consultations for any cause at one year post-MI and during the total follow-up period (to be ascertained using CPRD Aurum data).
5) The cumulative incidence of diagnosis of specific disease outcomes relevant to the post-MI context (including recurrent MI, stroke, heart failure and major bleeding), to be ascertained using both CPRD Aurum and linked HES Admitted Patient Care data.
6) The proportion of patients with MI that receive each component of ‘guideline-recommended’ MI treatment. This includes: in-hospital coronary angiography +/- percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) and the commencement of regular secondary preventative medication (i.e. a dual antiplatelet therapy, a β-blocker, an angiotensin converting enzyme inhibitor [ACEi] or angiotensin II receptor blocker [ARB] and a statin).
In addition, a comparator group that did not have an MI during the study period will be identified using risk-set matching. In this comparator population, the primary outcome and selected secondary outcomes (1, 3, 4 and 5) will be reported by multimorbidity status, in order to evaluate the interaction of MI and multimorbidity compared with a matched non-MI, non-multimorbid group. This will enable us to quantify the excess risk associated with MI in both patients with and without multiple long-term conditions that did not have an MI.
The following exploratory (hypothesis-generating) outcomes will be reported for all individuals with MI, vs. all individuals in the risk-set matched, non-MI sample:
1) The proportion of individuals with multimorbidity (the prevalence of ≥ 2 long-term health conditions) and complex multimorbidity (≥ 3 chronic conditions affecting ≥ 3 body systems) at study entry and during follow-up.
2) The median number of long-term conditions diagnosed at study entry and during follow-up. The prevalence of individual long-term conditions at the time of study entry and the incidence of new conditions during follow-up will also be reported.
3) The proportion of individuals belonging to a given multimorbidity phenotype at the time of study entry and during follow-up, as ascertained using an unsupervised machine learning approach.
Marlous Hall - Chief Investigator - University of Leeds
Jonathan Batty - Corresponding Applicant - University of Leeds
Charlotte Sturley - Collaborator - University of Leeds
Christopher Hayward - Collaborator - University of Leeds
Mohsin Masood - Collaborator - University of Leeds
HES Admitted Patient Care;ONS Death Registration Data;Patient Level Index of Multiple Deprivation;Practice Level Index of Multiple Deprivation