12 key steps to a successful population health strategy

12 key steps to a successful population health strategy

No single vendor in the current healthcare IT market meets all the requirements of population health management that providers are looking for, says Dale Sanders, who spent 22 years as a healthcare chief information officer and now serves as senior vice president at Health Catalyst, an analytics vendor. To help organizations make sense of the available vendor options and understand what is required to effectively manage populations, Sanders offers 12 criteria to plan PHM strategy and evaluate vendors.

Sanders says these criteria follow a logical progression. Criteria listed earlier are foundational; subsequent criteria depend on the earlier steps. The latter criteria are the most complicated, technically and culturally.

Determining which patients to include in pop health registries

Building an accurate population registry is foundational to effective population health management. Registries are the gatekeepers to accuracy—without precise definition of patient populations, everything else in the strategy suffers. Traditionally, population cohorts have been defined using billing data, specifically ICD-9 codes. But relying on codes to segment populations could cause the organization to miss as much as 40 percent of the patients that should be included in a group.

Picking strategies and algorithms to assign patients to clinicians

One of the most complicated tasks of population health is knowing who is really responsible for the patient. Who constitutes the care team and what are their roles? Even though one doctor is the primary care, the patient may visit other doctors, primarily specialists. One way to attribute the clinician-patient relationship is to analyze a patient’s visit patterns. This task is a key in assigning financial risk and performance incentives to the doctors who are responsible for the patient’s care.

Identifying ways to flag patients who are hard to manage or should be excluded

It is difficult to define patients who should be included in a registry and assigning them to accountable physicians. Also challenging is identifying patients in the registry who will be difficult to manage. Every electronic medical record should be able to capture data that reflects the socio-economic indicators that impact health and well-being. These include language barriers, cognitive or physical inability to participate in a care protocol, income and economic barriers, home and local violence, geographic inability to participate, refusal to participate, and medication contraindications to participating in a care protocol.

Monitoring clinical effectiveness and total cost of care

An organization needs to measure the practice of medicine against clinical protocols and continue to measure the variability in care. It should build dashboards around specific patients and populations of patients. Measurements should not focus exclusively on clinical quality but also track total cost of care for specific patients and on a per-capita basis across the population.

Establishing triage and clinical protocols for disease states

A good population health system defines how it will manage each population cohort. The problem with evidence-based medicine is a lack of applicability of the evidence that may be collected but lies outside the controlled clinical trial that generated the evidence in the first place. Clinical trials are rarely generalizable in complex patient profiles. In the future, trial evidence will be replaced with evidence from analysis of local data sourced from the enterprise data warehouse.

Stratifying work queues for care management teams

Risk stratification enables an organization to analyze, and aim to minimize, the progression of a disease and development of comorbidities. After patients in the registry are stratified and monitored, an organization can develop strategies to identify and intervene with patients on a high-risk trajectory. Over time as data get richer, profiling and proactively treating patients before they become members of the disease registry is the top goal, attempting to prevent disease rather than reactively treating it.

Accessing patient data from outside the delivery system

Most organizations rely on claims data to get a view of care outside their delivery system. But most organizations struggle with poor data quality, missed critical data elements and frequency of data delivery, which exacerbates the already inherent lag associated with claims data. So, healthcare organizations should negotiate with insurers to get a full claims file for all enrolled members and all services regardless of where the care was rendered.

Enabling patient-reported outcomes measurement systems

Patient-reported outcomes data is one of the most important pieces of data typically missing from current records systems. Little has been done in this area yet because of limitations in current records systems. However, organization can make significant progress on population health while the technology develops and evolves to enable the incorporation of patient-generated data.

Establishing a clinician communication and patient management system

Members of the care team need an automated way to communicate in order to manage populations. Effective strategies treat every high-risk patient as the center of a project plan. All members of the patient’s care team should have access to the overall project plan, next milestones and the responsibilities of each member. In the future, EHRs will incorporate project management concepts into their functionality.

Treating comorbid patients via triage and clinical protocols

The industry has not yet developed effective protocols for treating comorbid patients, relying on applying multiple single-disease protocols to these cases. But population health initiatives need to target comorbid patients, in particular Medicare beneficiaries who on average have three chronic conditions. At present, there are not many sources for clinical protocols that can account for comorbid patients.

Creating a patient education material and distribution system

This is significantly more complicated than enabling the patient to exchange messages with the care management team. Today, patient engagement revolves around the patient portal or personal health records. However, future patient interactions need to be personalized to meet individual patients’ specific needs.

Engaging patients, creating a system to communicate with them

Platforms need to evolve into a personal health project management system. Current options for this criterion are fragmented and immature but will improve during the next three years, Sanders believes. However, the patient engagement platform of the future will be owned completely by the patient with no dependency on the electronic health record or a single healthcare organization.