Precision public health: imprecise thoughts

What is precision public health?

• A recently coined term first appearing in print in 2014 (Dolley 2018)
• Sparse literature
• Various definitions but all encompass a set of consistent ideas:
  • Use of big data and new technology to improve health
  • Data used to give more precise descriptions of populations and individuals
  • Application of new techniques and methods
  • Speed, accuracy and scale
  • “the application and combination of new and existing technologies, which more precisely describe and analyse individuals and their environment over the life course, to tailor preventive interventions for at-risk groups and improve the overall health of the population.” (Weeramanthri et al. 2018)
  • “improving the ability to prevent disease, promote health, and reduce health disparities in populations by applying emerging methods and technologies for measuring disease, pathogens, exposures, behaviors, and susceptibility in populations; and developing policies and targeted implementation programs to improve health” (Khoury and Galea 2016)
  • “requires robust primary surveillance data, rapid application of sophisticated analytics to track the geographical distribution of disease, and the capacity to act on such information” (Dowell, Blazes, and Desmond-Hellmann 2016)
  • “Precision public health is characterized by discovering, validating, and optimizing care strategies for well-characterized population strata” (Arnett and Claas 2016)

Terminology

Biro et al have reviewed the literature on related terms and suggest 3 potential terms:(Bíró et al. 2018)

• Individualised prevention
• Personalised prevention
• Precision prevention

Their definitions are:

• Individualized prevention is a form of prevention in public health, in which health professionals consider the characteristics, lifestyle, family history, anamnesis, risk status and medication of the client when making proposals to maintain or improve the individual’s quality of life.
• Personalized prevention is a form of prevention in public health, which includes the activities of individualized prevention and in which health professionals also consider biological information and biomarkers at the level of molecular disease pathways, genetics, transcriptomics, proteomics and metabolomics of the client when making proposals to maintain or improve the individual’s quality of life.
• Precision prevention is a form of prevention in public health, which includes the activities of personalized prevention and in which health professionals also consider the socioeconomic status or the opportunities offered by psychological and behavioral data of the client when making proposals to maintain or improve the individual’s quality of life.

Current and possible uses

Current literature has identified 4 main use cases for precision public health:

1. Disease surveillance and signal detection
2. Risk prediction
3. Targeting treatment interventions
4. Improved disease understanding

Source:(Dolley 2018)

These are largely in environmental health and communicable disease control. The only studies identified by Dolley in non-communicable disease or broader public health relate to diabetes.

Source:(Dolley 2018)

Examples developing of:

• Synthesizing data, academic literature and expert opinion for priority setting (https://www.frontiersin.org/articles/10.3389/fpubh.2017.00125/full)
• Combining big and “smaller” data (https://www.frontiersin.org/articles/10.3389/fpubh.2016.00248/full)

Precision public health vs precision medicine

There has been some debate about the relationship or otherwise between precision medicine and precision public health. There are number of strands:

1. The 4 Ps of precision medicine (Predictive, Preventive, Personalised, Participatory) should be extended to 5 with the addition of P for “population”…
2. Some argue that the two are unconnected
3. Some argue that the precision medicine is the antithesis of public health and that precision medicine may increase inequality, and divert resources and focus away from population health and prevention (Taylor-Robinson and Kee 2018)
4. Many agree that the distinction is unhelpful and that they are two sides of the same coin.

Critiques and controversies

http://www.phgfoundation.org/blog/personalised-prevention-and-public-health-an-urgent-agenda

Implications for health intelligence

In some ways these ideas are nothing new - our users and stakeholders continually want more granularity and timeliness, more comprehensive data and insight, and actionable data linked to intervention. We have analytical techniques like population segmentation, risk profiling and geodemographics. This requires modernisation of surveillance, epidemiology, information systems and targeted interventions with population health perspective.

Precision and precision public health implies:

• More granular data for individuals and populations - build up from smallest analytical unit by default (e.g. SOA for populations, GP for clinical populations).
• More timely data - latest available data by default
• More granular time data - smallest feasible time unit (e.g. day, week or month)
• Incorporating big and small data
• Balancing against issues of information and data governance (re-identification risk)
• Large scale ecological linkage (e.g. large datasets of small area data - phenomics)
• Cohort development - requires linkage e.g. people with specific conditions
• Better measures of disease frequency - incidence for prevention, prevalence for burden and planning => data linkage
• Improved risk factor estimation and surveillance
• Combining multiple datasets - “n-tangulation”
• Application of new techniques - time-series/ forecasting, predictive models and machine learning (unsupervised and supervised)
• Complexity/ network analysis

## Warning in stri_replace_all_regex(string, pattern,
## fix_replacement(replacement), : argument is not an atomic vector; coercing

Precision PH in the news

https://www.dailymail.co.uk/health/article-6246191/GPs-text-patients-urging-change-unhealthy-habits.html

Patients could be sent text messages urging them to change unhealthy habits under radical proposals from the Health Secretary., Matt Hancock wants to use data from patients’ own medical records – as well as more general information on population trends – to work out their risk of developing certain illnesses., Patients would then be sent targeted public health advice via an app, email or text message which is based on their specific risk., Health secretary Matt Hancock wants to use data from patients’ own medical records, alongside population trends, to work out their risk of developing certain illnesses, Mr Hancock believes it would be far more effective than blanket campaigns on obesity, smoking and drinking., A middle-aged adult with high blood pressure – who was overweight and drank a glass of wine a night – would be urged to cut back on alcohol, diet and take more exercise., The policy, known as ‘predictive prevention’, is only in the very early stages and officials at Public Health England are yet to determine exactly how it will work in practice., But the idea of using patients’ confidential medical records to find out about their behaviour and lifestyle is likely to fuel both privacy and nannying concerns., Mr Hancock outlined plans for predictive prevention at a series of talks at the Conservative Party Conference in Birmingham earlier this week, Two years ago the Government was forced to abandon the hugely controversial ‘Care.data’ scheme, which intended to harvest medical records for research purposes. , Campaigners feared data would be hacked or passed to insurers., Mr Hancock outlined plans for predictive prevention at a series of talks at the Conservative Party Conference in Birmingham earlier this week., He pointed out that in the past, family doctors knew their patients so well that they were able to give them targeted advice., But times have changed and many patients see different GPs, often a locum they have never met., Under the new proposals patients could be sent text messages urging them to change unhealthy habits via an app, email or text message, Mr Hancock said: ‘Public health has essentially dealt with populations as a whole – the anti-smoking campaign on TV is targeted at everybody., ‘But using data, both medical data – appropriately safeguarded of course for privacy reasons – and other demographic data, you can work out somebody might have a higher propensity to smoke., ‘Then you can target interventions much more closely.’, Mr Hancock wants Government officials to work out which illnesses and unhealthy habits groups of patients are most likely to develop., Middle-aged middle-class adults tend to drink more alcohol than the general population. , Obesity rates are typically higher in more deprived parts of the country., Last night, Professor Helen Stokes-Lampard, chairman of the Royal College of GPs, said the scheme was ‘interesting’.

Concerns about big data

https://jamanetwork.com/journals/jama/article-abstract/2683125

• Electronic health records (EHRs) are a major source of big data for health and healthcare.
• Both structured data, and use of NLP to mine unstructured data (notes, text fields)
• Data linkages
• Quantitative risk prediction not new based on classical statistical learning (e.g. 1000 cvd prediction models - but few in routine care)
• Is it likely that ML will do better?
• Four barriers
  • Better decision making - are decisions sensitive to predictions of models
  • Calibration (Achilles heel of prediction)
  • User trust, transparency and commerical interests
  • Data quality and heterogeneity