Improving population health requires multiple interventions over time, tailored to meet the needs of individuals at each stage of their life, whilst simultaneously reducing the inequalities that exist in health outcomes across a population.1
Health and care systems have a key role in improving population health. One way in which their role can be understood is through their enablement of infrastructure, intelligence and interventions to support integrated care and population health management, as described by NHS England:2
- Infrastructure: For example, developing integrated health and social care models such as primary care networks (PCNs) that enable greater provision of proactive, personalised and coordinated care.
- Intelligence: The use of shared data via a normalised population health and care record to understand variation and support quality improvement across health and care services, making the best use of collective resources available.
- Interventions: Supporting more personalised care to help people manage their own health and wellbeing.2
In this short paper, one perspective is provided on how PCNs – a model of care emerging globally within integrated care systems (ICSs) – can make use of population health data and technology to support evidence-based improvement across large health and care systems.
Building from the ground up
Whilst recognising population health improvement requires multiple interventions across systems, there is a strong evidence base that universal, high-quality primary care provision, with its roots in local communities, improves overall population health outcomes, lowers per capita healthcare costs and reduces health inequalities.3,4,5
One new model of care emerging globally is the development of PCNs. These and similar models, such as neighbourhood care networks, can loosely be defined as groups of general practices (GPs) working together with a range of local providers across primary care (including pharmacy, optometry and dentistry), community services, social care, the voluntary sector and other public services to offer more personalised, coordinated health and social care services to local populations of around 15,000-150,000 people, addressing local population health needs and inequalities. Such models include improved access to medical, nursing, allied health, mental health and social care teams, plus other staff groups within an integrated care approach.6,7
The development of PCNs in various forms are reported as a key health policy aim across more than 15 OECD countries, including the UK, Australia, France, Switzerland, and the United States.5 Such care models may be supported by other primary care and out of hospital services being delivered at scale. For example, in the case of GP and primary care federations in the UK, GP federations are often members of ICSs, representing local general practice and primary care together with local medical committees (as the statutory organisation occupying this role).8
GPs and PCNs are a good place to start stakeholder engagement and involvement when embarking on a population health journey in order to create the foundation of any ICS or health network. Taking the UK primary context there are a number of reasons for this:
- As the first point of contact for anyone entering the health system, primary care accounts for around 80-90% of healthcare contacts. Contacts are increasingly being delivered via digital means, such as through online apps and telephone triage, with COVID-19 resulting in face-to-face GP appointments dropping from 82% in April 2019 to 47% in April 2020, and telephone visits increasing from 14% to 48%, when comparing the same two months.
- In the UK, GPs’ records are the primary record of care for an individual across their lifetime, thus being a central source of information to support direct care and population health management.
- Services are focused on the whole person, providing continuity of care across a lifetime and universal access to critical public and preventative health interventions such as vaccination, immunisation and screening.
- Focused primary and secondary prevention in primary care is effective in reducing health inequalities. A recent OECD report demonstrates that those in low income groups are 5% less likely than those in high income groups to see a GP; this extends to a 12% difference to see a specialist. The UK Department of Health has also demonstrated that inequalities in life expectancy can be narrowed by at least 10% for focused primary care interventions. Such interventions, like optimising effective therapies for cardiovascular risk in low income groups, can be implemented quickly and effectively and are enhanced by appropriate electronic case finding.5,10,11
- Services also play a critical role in connecting people for health and wellbeing, for example via digital social prescribing. This includes coordinating referral to services addressing the wider determinants of health, such as income, housing and community support.
- Primary care teams have an increasing role in precision medicine and pharmacogenomics, for example identifying patients with, or at risk of, a genetic condition; supporting clinical management of genetic conditions; and communicating genetic information that can inform precision medicine.
If new models of care like PCNs are going to be successful, access to joined-up, population health data to support integrated care and personalised interventions will be critical.
Developing and applying intelligence to improve care
Population health management is concerned with the use of aggregated data to predict, identify and manage individual health risk through a single, longitudinal record, whilst targeting new models of care to improve population health outcomes. Without access to a population health data platform with joined-up health and care data as a critical enabler to population health management, health and care stakeholders are flying blind in their attempts to understand and improve the health of their population.
A key feature of population health management programmes is identifying groups of the population with similar characteristics or needs that would benefit from action to support improvement in health outcomes, commonly through risk stratification applied at a population level or for sub-segments of the population. For example, using combined health and care data allows health and care providers to identify those at risk of frailty in near-real time and at the point of care to see whether the person’s care has been assessed, their goals discussed and interventions such as home, lifestyle and personalised prescribing and nutritional support put in place. Using data in this way will also support emerging and future approaches to care, such as precision medicine that takes into account individual variability in genes, environment, and lifestyle for each person. By combining all available data in a standardised, normalised and scalable way, health and care stakeholders, together with the individuals and communities they serve, can make the best use of their collective resources and assets to improve health outcomes.
At a health system level, one way in which population health platforms can support health and care system improvement is through supporting data-driven audit and feedback of key evidence-based health measures that can be actioned at the individual level to a wider population level for an ICS. A Cochrane review of 140 randomised trials found that audit and feedback produced a median 4.3% absolute improvement (interquartile range 0.5% to 16%) in health and care professionals’ compliance with desired practice, such as recommended investigations or prescribing.12 These improvement gains from audit and feedback may appear modest, but “cumulative incremental gains through repeated audit cycles can deliver transformative change” across large health and care systems.12-17 Furthermore, there is a range of evidence to suggest that electronic audit and feedback of care quality measures via electronic dashboards and registries results in improved care quality, is cost-effective relative to usual care and is suitable to support system-level transformation.12-17
The use of joined-up, population health data to improve audit and feedback in this way can be understood in the context of prescribing. Prescribing, the most common patient-level intervention across health and care, is second only to workforce in terms of total expenditure and a common focus for PCNs for quality improvement.5 Prescribing data from a single source only (e.g. GP data) is likely to vastly under-report prescribing activity compared with joined up data from multiple sources in venues where prescribing occurs, such as community pharmacies, community services and wider secondary and tertiary services. Variation in prescribing codes, for example between proprietary EHR codes and recognised standardised terminologies, necessitate the provision of normalised, standardised population health records to present data back to individuals in a meaningful way. Equally, the ability to case find for structured medication reviews and understand the impact of prescribing interventions on care utilisation, employment, relationships and wider health and life outcomes, is enhanced through combing data on personal risk factors and therapy indications. For example, identifying those most at risk of stroke using multiple risk factors that will benefit from more personalised anticoagulation therapy or targeting antibiotic use by understanding individual characteristics and symptom to reduce inappropriate antibiotic prescribing.
Personalising care by measuring what matters
Population health management will only deliver a step-change in health and care delivery if care is personalised to the individual and their unique context. Central to the delivery of personalised care within new models of care such as PCNs is the use of evidence-based, person-reported outcome measures (PROMs) that can be used within wider data sets to support risk identification and assess health outcomes from the perspective of the individual. Such measures can help to understand the unique psychological constructs that influence, and are predictive of, the use of health and care services, in addition to supporting outcome-based care models and evaluation of health system interventions using real-world data.
One such resource is the Patient Activation Measure (PAM), which gauges the knowledge, skills and confidence an individual has in managing their own health. PAM has strong psychometric properties and association with a number of other health outcomes, demonstrating reliability and validity for use in primary care and population-based care delivery settings.18-21 Critically, the measure is also actionable, where improvements in PAM have been shown to be influenced by evidence-based, community health interventions that support behaviour change and improve an individual’s ability to manage their own health. Such benefits are well aligned to the Quintuple Aim of improving health outcomes: improving the experience of care received by citizens, reducing unnecessary health and care costs, affecting inequalities in care (through improving risk identification and resource use) and improving staff experience of care delivery through more person-centred care.18-21 A study completed by the Health Foundation demonstrated that patients who were most able to manage their own health, as assessed through PAM:
- Had 38% fewer emergency admissions than the patients who were least able to manage their own health
- Had 32% fewer attendances at A&E
- Were 32% less likely to attend A&E with a minor condition that could be better treated elsewhere
- Had 18% fewer GP appointments
- Were likely to have improved medication adherence18
Linking back to the previous example around prescribing, measures such as PAM could be combined with clinical and care data to support personalised treatment approaches that also recognise individual barriers to medication adherence. For example, targeting structured medication reviews and patient education in primary care to those with lower PAM scores to increase medication adherence and improve outcomes as part of medication optimisation programmes.
Person-reported outcome measures like PAM may also support cost effectiveness and economic evaluations of health and care interventions. Technology has the ability to capture and report PROMs at a scale previously impossible to achieve, whilst integrating them within health and care workflows, such as digital patient portals and population health tools like analytics and registries, to create outcome-based care pathways.18-21
Considerations for getting started with supporting PCNs with population health management to be understood within the context of health system governance, transformation programme, structures and processes:
- Identify and map PCN organisations and stakeholders across a defined population, e.g. those registered and resident within a defined geography. Utilise publicly available data sets and existing local needs analysis to identify early variation in health inequalities and outcomes, priorities for population health management and specific population segments with higher risk of poor health outcomes for quality improvement.
- Understand GP and primary care data sets in use (including those for digital and online providers), which will be foundational to population health management regardless of chosen population health strategies. Identify any known early information and data quality gaps, for example, from local digital maturity assessments, in addition to data processing and sharing considerations for early use cases.
- Identify the analytical and quality improvement capacity to support early engagement and light-touch quality improvement support.
- Define, baseline, audit and feedback a core set of indicators (which may already be collected) to support understanding around variation and quality improvement within the PCN setting – aligned to locally identified population segments and priorities – for example, care quality measures for those identified as at risk of frailty in older adult age groups. Avoid measurement fatigue and confusion by focusing on a smaller set of 5-10 measures that are applicable across the scale of the health system to focus early engagement.
- Identify any common person-reported outcome measures that may be useful to collect for population health purposes and to support personalised care.
Next steps
Primary care networks and similar globally emerging models of care are well placed to support population health management and a good place to start early engagement for population health programmes. Critical to their development is the use of joined-up data and technology which can create the environment for evidence-based quality improvement and help to re-imagine primary care in the recovery from the global coronavirus outbreak. This paper demonstrates one perspective on how to get started.
References:
1. Kindig, D., & Stoddart, G. (2003). What is population health? American Journal of Public Health, 93(3), 380–383.
2. https://www.england.nhs.uk/integratedcare/ (Accessed online 30th June 2020)
3. Macinko, J., B. Starfield and T. Erinosho (2009), “The Impact of Primary Healthcare on Population Health in Low- and Middle-Income Countries”, Journal of Ambulatory Care Management, Vol. 32/2, pp. 150-171.
4. Macinko, J., B. Starfield and L. Shi (2003), “The contribution of primary care systems to health outcomes within Organization for Economic Cooperation and Development (OECD) countries, 1970-1998”, Health Services Research.
5. OECD (2020), Realising the Potential of Primary Health Care, OECD Health Policy Studies, OECD Publishing, Paris.
6. Fisher, R., Thorlby, R., Alderwick, H. (2019). Understanding Primary Care Networks. Context, benefits and risks. London. The Health Foundation.
7. https://www.england.nhs.uk/primary-care/primary-care-networks/ (Accessed online 30th June 2020).
8. Charles, A., Wensel, L., Kershaw, M., Ham, C., Walsh, K. (2018). A year of integrated care systems. Reviewing the journey so far. London. The King’s Fund.
9. https://bjgplife.com/2020/06/29/how-are-we-going-to-track-national-progress-in-new-ways-of-working/ (Accessed online 10th July 2020).
11. http://www.nao.org.uk/wp-content/uploads/2010/07/1011186.pdf
13. Peiris D, et al. Effect of a computer-guided, quality improvement program for cardiovascular disease risk management in primary health care: the treatment of cardiovascular risk using electronic decision support cluster-randomized trial. Circ Cardiovasc Qual Outcomes. 2015;8(1):87–95.
14. Thomas KG, et al. Use of a registry-generated audit, feedback, and patient reminder intervention in an internal medicine resident clinic--a randomized trial. J Gen Intern Med. 2007;22(12):1740–4.
15. Carlhed R, et al. Improved adherence to Swedish national guidelines for acute myocardial infarction: the quality improvement in coronary care (QUICC) study. Am Heart J. 2006;152(6):1175–81.
16. Brinkmann, L., Gavan, S., Peek, N., Ashcroft, D., Elliot, R. (2020). Cost-effectiveness of an electronic audit and feedback intervention to improve medication safety using electronic health records and quasi-experimental methods. Value in Health. Volume 23, SUPPLEMENT 1, S250-S251, May 01, 2020.
17. Foy, R., Skrypak, M., Alderson, S., Ivers, N., McInerney, B., Stoddart, J., Ingham, J., Keenan, D. Revitalising audit and feedback to improve patient care. British Medical Journal. 2020;368:m213 doi: https://doi.org/10.1136/bmj.m213 (Published 27 February 2020).
18. https://www.health.org.uk/sites/default/files/Reducing-Emergency-Admissions-long-term-conditions-briefing.pdf
19. https://www.ncbi.nlm.nih.gov/pubmed/21078231
20. https://hijournal.bcs.org/index.php/jhi/article/view/9/26
21. Murphy et al (2018). Identification, description and appraisal of generic PROMs for primary care: a systematic review. BMC Family Practice (2018) 19:41
