Infectious disease triggers of chronic disease exacerbations

This Data Sharing Agreement permits the retention and processing of the data provided under previous iterations of this Agreement.

Respiratory viruses (like colds and flu) trigger asthma attacks, chronic obstructive pulmonary disease (COPD) exacerbations, and heart attacks. Other factors like air pollution can trigger these attacks too. Virus surveillance data on its own is not good enough to work out what proportion of these attacks are from viral infections or are caused by other factors. The research team at the London School of Hygiene and Tropical Medicine (LSHTM) need to know which factor is most important so that they can try to do something to prevent these attacks.

The pattern of viral circulation in populations is affected by how people mix together and pass viruses to each other. Children have a lot of contacts each day and pass a lot of viruses to each other, to their parents, and to other people they meet. School holidays are especially important, because children pass viruses less when they are not in school, because they meet fewer people. By combining what is known about how viruses transmit and how people mix together, it is possible to better understand the circulation of viruses.

This can then be used with the viral surveillance data (which is not good enough on its own) to understand the effect of each factor on the number of attacks of asthma, COPD, and heart disease each day. This kind of analysis is scientifically really hard to do, because it depends on the small differences in school holidays from place-to-place, and different amounts of pollution in one place compared to another. These small differences in lots of places over many years add up, and so scientists can calculate the effect of each factor. This is the benefit of “big data”, which lets scientists do studies that would not work otherwise, and is how NHS data can be used to help other patients.

It should be noted that the entire population is of interest for this study, not just children.

Further details:

This study examines the timing of hospitalisations for three major chronic diseases to detect associations between known viral and environmental triggers for these conditions.

The health impacts of asthma, chronic obstructive pulmonary disease (COPD), and coronary heart disease (CHD) are greatly increased by acute episodes of worsening symptoms (exacerbations). Exacerbations are triggered by environmental factors e.g. poor air quality and temperature, and by acute respiratory infections. The large (and growing costs) of exacerbations provide considerable motivation to improve understanding the triggers of exacerbations.

Current methods do not include the dynamic risk of exacerbation caused by respiratory virus transmission, which means that estimates of risk from other variables may be unreliable. This project will begin to tackle the need to understand how viral triggers affect population-level timing of exacerbations. The project will use interdisciplinary methods to develop a novel quantitative framework to assess the population-level drivers of chronic disease exacerbations.

The outcome under study is the daily timing of inpatient chronic disease exacerbations for three chronic diseases under study. Therefore, this study is requesting these data from NHS digital for this analysis. The study is also requesting information from NHS digital on infectious respiratory inpatient hospitalisations to aid in the analysis of the chronic disease exacerbations.

Specific information:

The London School of Hygiene & Tropical Medicine (LSHTM) requires Hospital Episodes Statistics data for use in project: “Novel methods in data science to quantify viral and environmental triggers of chronic disease exacerbations”.

There are two organisations involved in this work: the London School of Hygiene & Tropical Medicine, and Public Health England (PHE). LSHTM is the lead for the study, and instigated it as part of the research project being carried out by the Principal Investigator of this project, who is substantially employed by the LSHTM. Members of staff at PHE are also involved in an advisory capacity to the Principal Investigator, due to their expertise in respiratory infections and respiratory virus surveillance. They have no further role in the study, and will not have access to the data provided by NHS Digital. The LSHTM is, therefore, the sole data controller who will also process data.

The raw data will only be viewed, accessed and analysed by direct substantive employees of the London School of Hygiene & Tropical Medicine (LSHTM). PhD students may use aggregated time series containing small numbers but will not access the record-level data. This must be aggregated with small numbers suppressed in line with the HES Analysis Guide. Only substantive employees at LSHTM will have access to the record-level data requested from NHS Digital.

LSHTM established the Electronic Health Records Research Group to undertake health research using electronic health records. This project was proposed as part of a request for projects using UK health data, funded by Health Data Research UK (HDR-UK). LSHTM responded to a call for research in this area, and the project proposed by the lead researcher for the project was successful.

LSHTM/lead researcher applied for and secured funding from Health Data Research UK (HDR-UK) to undertake this work. HDR UK is a joint investment led by the Medical Research Council, together with the National Institute for Health Research (England), the Chief Scientist Office (Scotland), Health and Care Research Wales, Health and Social Care Research and Development Division (Public Health Agency, Northern Ireland), the Engineering and Physical Sciences Research Council, the Economic and Social Research Council, the British Heart Foundation and Wellcome. It supports world-leading research to develop cutting-edge analytical tools and methodologies to address the most pressing health research challenges.

This work is a study on inpatient hospitalisations resulting from exacerbations of 3 chronic diseases. Inpatient hospitalisation from respiratory infections will be an input to the study. Inpatient data are requested from NHS digital.

The aim is to better understand the triggers of chronic disease exacerbations, and to do this LSHTM need to link the circulation of respiratory viruses to the patterns of exacerbations. Inpatient hospitalisation data are needed to achieve this aim and deliver the study results and benefits.

This study will develop new methods that allow estimation of both the parameters of dynamic transmission models for viruses, and the contribution of environmental factors, at the same time.

This is a new study, and no data have been supplied for this study before.

No additional data will be provided under this Agreement.

Pseudonymised HES data was transferred from NHS Digital to LSHTM using the secure data transfer portal.

This data is stored on the secure server at LSHTM which can be accessed only by the LSHTM study team using a unique network password. No-one else outside of the LSHTM study team will have access to any of the NHS Digital data under this Agreement.

There are no subsequent flows of data.

All organisations party to this agreement must comply with the Data Sharing Framework Contract requirements, including those regarding the use (and purposes of that use) by “Personnel” (as defined within the Data Sharing Framework Contract i.e.: employees, agents and contractors of the Data Recipient who may have access to that data).

The data will not be linked with any record level data.

There will be no requirement nor attempt to reidentify individuals from the data.

The data will not be made available to any third parties other than those specified except in the form of aggregated outputs with small numbers suppressed in line with the HES Analysis Guide.

This project requires Hospital Episode Statistics Admitted Patient Care data. The project minimised the data requested by limiting to specific health conditions and diagnosis codes. This is a large study, and needs to be, because the differences from city-to-city in viral circulation and pollution are expected to be quite small. Therefore “big data” are needed to give the statistical power to estimate these effects.

The project requires 13 years of data. This number of years is needed for 4 major reasons:

1) School holidays affect transmission of viruses and hence risk of exacerbations. There are small changes to school calendars each year, and so as many years as are possible are needed to be able to detect the effect of these small changes in holiday timing.

2) Over long time periods we expect demographic change in the populations (e.g. increase or decrease in number of children, or fraction of older adults). These changes can alter viral circulation, i.e. a higher proportion of children in the population increases the circulation of respiratory viruses, because children have higher contact rates. This is a slow process, and therefore a fairly long time period is needed over which to estimate these effects. If the time period is too short, the estimation procedures will not have the power to detect the effect of demographic changes on viral circulation. Therefore the public health benefit may not be met.

3) The fraction of older adults with chronic conditions has been increasing through time in the UK. These changes are quite slow and therefore a fairly long time period is needed over which to estimate these effects. If the time period is too short, the estimation procedures will not have the power to detect the effect of demographic changes on number of individuals at risk. Therefore the public health benefit might not be met.

4) Affecting all 4 previous reasons, the study period contains two major events in influenza circulation: the 2009 pandemic, and the phased introduction of the paediatric influenza vaccination program (2013 onwards). Influenza is known to be a trigger for chronic disease exacerbations, so data from before and after these events is needed in order to properly assess the baseline and the effect of these events on viral circulation. Without this, the effect of viral circulation may be confounded by these events/changes and the study will not achieve the aims and therefore the public health benefit.

Data is required nationally because there is variation in school calendar timing and air quality in different regions, and this is what is being studied. Using national-level data will maximise the number of hospitalisation events in each school calendar year which will increase power to detect the effect of each factor on exacerbation risk.

Data are minimised by filtering to limited specific conditions of relevance. There are 4 categories: 1) Asthma and similar, 2) COPD and similar, 3) CHD and similar, 4) respiratory infections and similar. Categories 1, 2 & 3 are the target conditions, and Category 4 is an input to the model to understand the exacerbations of 1, 2 & 3.

Daily resolution is required to detect effects of air quality on exacerbation risk, because air quality changes every day. Age of patients (month-year) is required for assigning to correct year of school. Location is required for assigning to geographic areas covered by particular school calendars, but only first part of postcode is needed.

A final report of results will be submitted to HDR-UK in February 2021. This will cover key findings of the study including: methodological developments, scientific findings, policy implications.

Academic paper(s) will be published in open-access, peer-reviewed journals, and on the organisation’s website on the following topics:

• impact of air quality, viral circulation and other covariates on daily exacerbation rate for each condition;

• methodology in using “big data”;

• cost and effectiveness of potential vaccination strategies.

Target dates for submission will be minimum 2 per year, starting mid 2019.

Where possible, the project will target general public health and/or epidemiology journals with a broad audience (e.g. Lancet Respiratory Medicine, Lancet Global Health, Lancet Public Health. PLOS Medicine, PLOS Computational Biology. BMJ, BMJ Open. International Journal of Epidemiology, American Journal of Epidemiology, Epidemiology.). Because analyses are likely to be of interest not just in public health but in methodological advance, the project will also consider specialist journals in statistical methods for large datasets. Dissemination at national and international conferences will adopt a similar strategy of aiming for as broad as possible a reach. They will include disease-focused meetings such as Chest, COPD, and Asthma, and modelling meetings such as Epidemics. The project will also seek presentations at specialty conferences where results have relevance to those audiences, as well as meetings where public health decisionmakers are likely to be represented.

For each paper published, a presentation will be developed to summarise the findings for a range of stakeholders, e.g. scientists, patient groups, public engagement events, outreach, policymakers. Findings will be presented at appropriate events.

A simplified version of the findings will be generated for sharing with charities/patient groups of interest, and publishing on the organisation’s website.

Findings from the study like this will also be shared in posters, presentations, and online. They will be promoted through Twitter and the University website. Where appropriate they will be included in the public engagement work the PI already does (e.g. New Scientist Live, TEDx talks, other events).

The LSHTM website provides links to LSHTM open access papers. All publications and conference presentations are promoted on twitter, via the @ehr_lshtm account (Electronic Health Records Research Group, >200 followers), @cmmid_lshtm account (Centre for Mathematical Modelling of Infectious disease, >600 followers) and sometimes the @LSHTM account (main University account, >19,000 followers). They may be promoted through the @HDR_UK account (the funder, >1300 followers), or my personal account (>200 followers).

All outputs will contain only data that is aggregated with small numbers suppressed in line with the HES Analysis Guide.

Decreasing the cost of chronic disease exacerbations is a public health priority, because those costs are rising in the UK and worldwide. To do this, LSHTM need better scientific understanding of the factors that can trigger exacerbations.

There is a lot of research from studies of individuals with chronic conditions that viral infections can trigger exacerbations of their conditions. However, public health policy is made at the level of populations, not individuals. There is a real need to determine if the effects we see in individuals are true at the population level.

Previous research has shown that environmental variables affect the population-level patterns of exacerbations, but no study has included both environmental triggers and infectious triggers in the UK. The benefits of this study will therefore be in bringing scientific understanding to the interactions between environmental and infectious triggers.

The study is generating new knowledge that will not only benefit researchers but will benefit the wider community and society as a whole. Analyses of the patterns of these serious conditions will help to better understand the risks and causes of ill health, especially in these populations that already have serious chronic conditions. Rigorous epidemiology like this study will allow design of better preventive strategies, to help patients and populations decrease the burden of chronic disease exacerbations.

Studies of long time series of cases from around England and Wales will allow understanding of whether the exacerbations from one factor, for example, air pollution, have gotten more or less likely over time. It also allows estimation of whether the baseline rate of exacerbations has improved or gotten worse. Short studies cannot detect these kind of changes. Studies of long time periods allow evaluation of interventions that have been made, such as vaccination programs, but to properly estimate their impact, the analysis must have enough data before the intervention before estimating the impact of the new intervention.

Results will be shared in open-access scientific articles, in reports, and in talks, and promoted as widely as possible. They will be shared with scientists, and through links with PHE, with public health officials. The results will also be shared publicly and written so that non-specialists can understand and interpret the results. Findings on environmental triggers, especially air quality, will feed into evidence of the role of air pollution on health.

These benefits are achieved through the full use of hospital data on exacerbations, public air quality sources, and surveillance data on the viruses involved. These can then be used in population modelling and quantifying public health outcomes

The yielded benefits section will be updated upon submission of a further renewal of this Agreement.