At Camden Clinical Commissioning Group (CCG), a local health planning organization servicing a London borough with a population of 250,000, we developed a Population Health Management Tool to provide health planners with a greater understanding of the needs of different segments of the population. This move toward population health-based commissioning has led to improved outcomes and reduced costs through the delivery of appropriate care.
While aggregate population data and individual stories play important roles in understanding our population, a tool was required to bridge the gap by segmenting the population according to need and service utilization.
Different segments need different approaches: patients who are mainly healthy will benefit from universal approaches (e.g., improving access), whereas those with the most complex conditions require more individually tailored, integrated approaches in order to maximize value.
Different conditions (e.g., dementia) have different quantifiable impacts on service utilization.
All segments of the population can move in both directions (i.e., toward both higher and lower service utilization); understanding the impact of the disease process or interventions is important in planning future services.
Linking deprivation data to the tool demonstrates the direct adverse impact of deprivation on health, suggesting the need to invest more widely in improving health.
Understanding the population health needs facilitates strategic planning and informs service redesign and impact.
In 2014, one of us (C.S.) and Dr. Thomas Lee detailed the lifetime use of health and social care, and its cost, for three generations of a Camden family. Our findings suggested that a more strategic and thoughtful approach was needed in order to achieve value for patients, that such care went beyond traditional health services, and that it was necessary to recognize the longer-term investments needed to improve lives. To that end, we sought to bridge the gap between aggregate and patient-level information by using service utilization as a proxy for the health needs of different subgroups within the local population. Our primary challenge was to answer three questions:
- Who falls within high-cost and high-utilization groups?
- What drives a patient to fit into one of these groups — disease, demographics, or both — and what factors can influence a move toward better health?
- How do we use the information to drive value and inform service redesign?
Our goal was to develop a more sophisticated tool that would allow us to quantify and better understand the health needs of different segments of the population in order to enable health planners and providers achieve the best “value” (defined as outcomes per pound/dollar spent) for both individual patients and the health system. This “Population Health Management Tool” was to be based on the “Bridges to Health” model.
Following the decision to move to a population-based approach to commissioning services, a team of four specialists (the Camden CCG Sustainable Insights team), supported by clinical input, was assembled to develop a tool that linked hospital data and primary care data. Service utilization was used as a proxy for health needs.
The tool, which is populated with pseudonymized secondary care (hospital) information that has been flagged on a disease register by primary care providers, allows us to look at different segments of the population, from individuals who are generally healthy to those with more complex conditions, as well as at different combinations of health conditions. The ability to evaluate the information in this way helps us to focus health planning (commissioning) activities to the right “segments” of the population at the right place and at the right time.
Six segments of increasing complexity were identified within the Camden CCG population. The most complex segment comprised only 1.21% of the total Camden CCG population but represented 13% of overall spending, suggesting that a different approach to managing this segment could improve value for both the patient and the system.
Splitting the population according to age (children, adults, and the elderly) shifted the proportion of individuals in each segment, as did splitting the population according to the different combinations of diseases; for example, 30% of individuals with dementia were in the most complex segment.
Similarly, the ability to understand and quantify the impact of different disease combinations on service utilization enabled health planners to target resources effectively. For example, for patients with atrial fibrillation, having two or more comorbidities was found to increase the cost of care threefold.
Figure 3 illustrates the cost of emergency admissions for ambulatory care-sensitive (ACS) admissions, which are considered by health planners to be opportunities to achieve better value. The admitted patients fell into three Population Health Management segments: those who were generally healthy, those with long-term conditions, and those with complex conditions.
While 65% of the cost was for the generally healthy segment (representing 20.49% of the total Camden population), the unit cost in that segment was low ($874). In comparison, 31% of the cost was for the most complex segment (representing 1.2% of the total population), in which 74% of the patients had at least one admission for an ACS condition and 44% had two or more admissions. The average cost per patient in this segment was $7,188.
Achieving the most effective and comprehensive health care outcomes for each segment and reducing costs incurred in expensive hospital settings requires different strategies: those who are mainly healthy may benefit from population-level interventions (i.e., access to rapid assessment and treatment) to return them to health and function as quickly as possible; those with long-term conditions or chronic illness require supportive self-management and health services delivered in the community to manage their condition, maintain health, and prevent complications; and those with the most complex segments require tailored interventions and multidisciplinary input across health and social care to meet their specific needs.
Application of the Tool in a Clinical Pathway
CCG Commissioners, along with provider and patient partners, have used the tool to support the redesign of services for the purpose of creating better value. For example, the Diabetes Integrated Practice Unit used the tool to create a community-based service for identifying and managing patients with diabetes and achieving value through improved outcomes and reduced hospital use. Figure 4 demonstrates the increase in the number of adults diagnosed with diabetes (information that allowed supportive self-management interventions, care planning, and community specialist support to be put in place), and Figure 5 demonstrates the decrease in the number of unplanned hospital admissions for hypoglycemia/hyperglycemia.
The tool has been used to redesign services for more complex, frail, elderly patients. Specifically, the development of multidisciplinary teams based in the community, combined with support by care coordinators, has enabled 74% of such patients to spend more time at home (defined as the most important outcome by patients and caregivers) and an 18% decrease in hospital bed usage.
Application of the Tool in the Analysis of Patient Segmentation
Most interesting of all, further refinement of the Population Health Management Tool has identified movement between segments, both to greater complexity but also to better health, even among the elderly. This finding raises further questions about what factors influence such movement. We are exploring whether this movement is related to the nature of the disease process itself (the way in which the disease progresses or remits) or to the impact of interventions that have been put in place to support and manage patients.
Application of the Tool in the Analysis of Utilization and Deprivation
Linking the Camden population risk profiles to deprivation data has enabled us to start to quantify the impact of deprivation on the utilization of health resources. In all segments, we have observed that deprivation is associated with a disproportionate increase in health utilization.
Within Camden, significant health inequalities exist between the poorer and more affluent parts of the borough, contributing, for example, to an 11.6-year mortality gap among men. Understanding and measuring the impact of deprivation on health outcomes suggests that, as health planners, we need to address some of the wider determinants of health. Examples of specific steps that we have taken include providing support to vulnerable families (resulting in improved school attainment), providing domestic violence support workers to Accident and Emergency and Gynecology departments, and training communities to identify mental health illness.
The Camden CCG Sustainable Insights Team consisted of four individuals: a health economist, an operational researcher, a statistician, and a health information specialist.
The relevant metrics are illustrated in Figures 1-A through 7. All data have been sourced from Camden CCGs Population Health Management Tool, which is based on pseudonymized National Health Service (NHS) Secondary Uses Services (SUS) data and clinical information flagged in a disease register by general practitioners. Camden CCG is an Accredited Safe Haven allowing data with appropriate permissions to be used to populate the tool. The cost of developing the tool was split between software and specialist resources, which amounted to £5k and £55k (approximately $6,100 and $68,000), respectively. There are no license costs associated with sharing the tool.
Where to Start
To maximize the impact of good health care planning, the first challenge is to develop a deep understanding of your population and how services are utilized. The NHS is very good at collecting transactional data; historically, however, these data have not been used to drive service transformation. Through the use of the Population Health Management Tool, Camden CCG has developed an understanding of its population and service utilization, which has allowed us to bridge the gap between epidemiological research and individual patient feedback and thereby develop focused interventions to transform services.
We need tools to enable better understanding of the needs of our population if we are to achieve greater value for patients and the NHS. These tools can then be used to support both the commissioning and delivery of more strategic and integrated models of care across health and social care. Planning at an individual level is not practical, but the ability to group patients with similar health needs and to find ways to respond more effectively and efficiently to these needs can help us to deliver measurable improvements in patient outcomes and reduce costs to the system.
Data are needed to understand population needs and to plan services effectively. To achieve safe and secure access to these data, technical issues such as the information governance concerns of individuals and organizations need to be addressed. Once the technical challenges are overcome, cultural challenges exist. These challenges involve sharing and understanding data and then using these data in a different way to transform care beyond organizational boundaries.