Since its invention in 2003, the term “population health” has been steadily catching on in the medical community. The corresponding efforts in medical software development started accelerating in the mid-2010s. According to MarketsAndMarkets, the market will grow by $16.5 billion between 2020 and 2025. Statista provides similar evidence—the population health management software market is to be expanding on the global scale by 2025, with downscaling noticeable only in the US:
So what is the promise of population health management (PHM) tech? Is it yet another gimmick or a valid investment that pays off? Let’s explore this in detail.
Public health vs. population health
Before we discuss population health management solutions, let’s first look at what population health is all about as it is often confused with public health. Though the two terms sound very much alike, the concepts behind them are pretty much different.
To cut it short, population health comprises the health outcomes of some group of patients—a certain population. The group can come from just one clinic, a city district, a county, a state, or the entire country. Public health is the approaches we use to make the health of the population better. Besides, there is yet another concept—community health. It’s not only the health or the outcome. It's more about the methods to ensure them through cross-field cooperation pursued with cultural specifics in mind. It's about engaging with the community, using scientific and evidence-based approaches that meet the needs and interests of the community.
We dispelled the terminological fog, but it’s still unclear what population health depends on and how it is related to software at all. We’ll clarify these questions below.
Population health factors
The outcomes of diverse populations’ health do not emerge from nowhere. Health outcomes build upon five factors, namely:
- Genetics. This factor describes a patient’s predisposition to a certain disease or condition based on their genetics.
- Environment. This is about the physical world and its conditions, including housing, availability of healthy foods, clean air and water, exposure to toxins, etc.
- Social conditions. This category comprises income, education, employment, culture, etc.
- Behavior. This is about habits and lifestyles of populations—alcohol consumption, tobacco use in any form, diet, exercising, and more.
- Medical factors. This category describes access to care and its quality.
Given their significance, these factors are often called determinants. But how do they influence populations? This can be illustrated with the county health model prepared by researchers from the University of Wisconsin:
Here comes another question: how did the researchers come to those conclusions? Obviously, checking all available parameters (over 30 items) manually would be exhausting and error-prone. This is where automated tools step in to speed up the process and make it less costly.
3 pillars of population health management solutions
As we’ve figured out, population health management software collects data from diverse source for storage and analysis. These sources comprise the ones used in a clinical setting and raw data sources. So what clinical data sources can be enrolled? Deloitte illustrates them in their concept of a digital health ecosystem:
If you haven’t tapped into population health yet, this complex system of intertwined technologies may frighten you. However, big things start small.
One ‘best’ way to approach [population health] management does not exist; different practices and organizations have succeeded with various approaches to both in reach and out reach. Assess your practice and your resources, and create a model that will work best for you and your patients.
There are three essential elements that make up the core of any population health management solution: it should be designed to analyze big data in order to classify patients into certain groups (populations) by the risks of experiencing some events, manage targeted care delivery, and report on individual and group outcomes.
It’s not enough just to store relevant data. A detailed analysis is needed to stratify patients into cohorts. However, running it without machine assistance is time- and cost-intensive.
Fortunately, there is a number of analytical tools that can speed up these tasks. First of all, there is predictive modeling in healthcare. Such modeling algorithms comb through a multitude of historical data to create models for forecasting potential results.
This analytical approach found its use at Mass Gen Brigham (Boston, MA). The clinic leveraged the internet of medical things by providing their congestive heart failure patients with remote monitoring devices. With the devices, patients could upload real-time updates on their weight, blood pressure, and other metrics, aggregated in the hospital’s intelligent system. Relying on these data sets, the system singled out at-risk patients in need of specific intervention. The tool helped lower the readmission rate as well as the number of nurses necessary to cover patients’ needs, which led to cost reduction.
There is one more application for healthcare big data. As doctors have no extra time to invest in analyzing raw data, they need a solution to provide clear, straightforward insights to work on. This is where healthcare BI proves valid. Advanced BI solutions provide real-time analytics, presenting it via dashboards. This means medical professionals can visualize patient cohort data and the effects of treatments on patients with no delays, which allows making timely corrections and improving care delivery.
In a clinical setting, treating a patient is rarely a single doctor’s responsibility. As a rule, the process involves 2-3 professionals in different medical fields, lab analysts, and other. However, their efforts need to be coordinated, especially when dealing with chronic disease patients. According to the 2019 Commonwealth Fund research, this lack of coordination is the major problem of the US healthcare system. Some users erroneously believe it’s possible to manage coordination through EHR systems. However, it’s not the case. Electronic health records contain detailed information on each patient, but managing this information in the process of multi-touchpoint care is the matter of healthcare interoperability. PHM tools may be of help here.
Population health management software allows creating clinical pathways for mapping diverse patient journeys. PHM tools don’t automatically transfer paper-based documents to a digital environment and let them be. Powered by machine learning technologies, these tools rely on the uploaded clinical documents to set and coordinate tasks all across teams in the care continuum. This way, they help clinicians accelerate the care cycle and refine its quality, which also ensures a better patient experience.
Seeking to improve care coordination, Banner Health (Phoenix, AZ) introduced a platform called Navigation Accelerator that lets doctors access patients’ medical history and social data right at a care facility. With this tool, clinicians gain access to comprehensive data about patients’ physical and behavioral health to cover an array of their needs in full, from drug prescription to referrals. However, it’s not only patients who may benefit from the tool:
…It really is a first-of-its-kind shared platform that gives providers in different practices a full and current picture of a member's comprehensive health status.
This effort to bridge healthcare data gaps for the sake of patients and medical professionals proved efficient, and not exclusively in Arizona. In 2019, the Accelerator was deployed in over 70K clinics across the country.
Engagement and collaboration
Fruitful cooperation requires a common goal, and in healthcare it’s in reducing the costs. According to Peter G. Peterson Foundation, the US healthcare will eat up about 20% of the country’s GDP by 2025. So what can key stakeholders (patients, clinics, insurance companies) do to contain the growth?
For patients, the task is clear—they need to take a proactive approach to their health management and prevent gaps in their clinical data. At the same time, contacting a doctor or a clinic whenever they detect some missing lab analysis is exhausting. The majority of patients will let go of their health management until the next hospitalization. Nevertheless, statistics say patients do want to control their health. According to Statista, in 2019 65% of Parkinson’s disease patients took some steps to study the disease and/or actively engage in their health.
This is where providers may join the game. For example, they may introduce a healthcare mobile app to let patients connect to their EHRs and make informed decisions about their health from anywhere. Such an app can also become a part of the clinic’s digital environment, allowing teams to access EHRs, supervise patients’ efforts in managing their health, and intervene when necessary. At the same time, providers can pilot population health management and break patients into several cohorts regarding their conditions and associated risks. This can help streamline inventory management and workload to deliver better care.
When it comes to payers, let’s turn to statistics again. The World Health Organization reports that cardiovascular diseases (CVDs) are the most frequent cause of death worldwide. The American Heart Association adds that roughly every 40 seconds someone falls victim to a CVD. Besides, Ipsos reports that the expected CVD death rate in the US in 2019 was 19% lower than the actual one. Naturally, for insurers, this potentially leads to increased reimbursements to the families of the insured individuals. Finally, CVD deaths are preventable, and prevention is to the payers’ interest as well.
When it comes to PHM, all the three parties are on the same page. Patients strive to better their health and avoid incidents, providers are eager to offer early detection and prevention for at-risk patients to reduce costs, and payers wish to reduce the number of payments through enhanced risk detection in patient cohorts.
The main problem is to make sure all parties clearly understand their benefits. Patient education, seminars for employees, and meetings with key payers may be of help.
PHM: a broader picture
PHM is applicable not only to clinics. With the present industry focus on prevention over hospitalization, it makes sense to go the extra mile and ensure that certain data is available at a larger scale. This can be solved with a health information exchange (HIE). It collects data on patients over a region and lets medical professionals from different locations access it via a single API. Thanks to such solutions, medics can analyze populations, stratify them by risks, and plan interventions.
However, in this case, some factors that impact health, such as social ones, go ignored. While medics unanimously recognize their importance for patient outcomes, they can hardly gain access to them. EHRs collect data pertinent to patients’ health only, leaving aside the data on food habits, smoking, and more. Clinicians have only one way to get this data—personal interviews with patients, which means extra time and crammed schedules.
Fortunately, PHM can help here too. This is what the Indiana Network for Population Health, deployed in 2020, does. The tool collects social determinants of health (SDOH) data across multiple sources and links it to patients’ clinical data stored in the HIE. As a result, doctors get full access to information on individuals and patient populations and are able to provide better prevention or timely intervention at a lower cost.
Now as we’ve looked at PHM and the software required to facilitate it, it’s time to answer the key question: is it worth going for? It absolutely is. Population health management assists with delivering personalized care to high-risk individuals, preventing relapses timely and improving care outcomes. It also allows caregivers to fine-tune their resource management and lower costs. As for the software, it significantly speeds up analytics and care administration.
Deploying PHM solutions across a large territory enables clinicians to analyze the information on patients coming from many sources, not only EHRs. What’s more, with the growth of adoption rates, the instances of medical data loss will gradually disappear for good.