Care Transformation Center Blog

3 pitfalls to avoid when assessing patient risk

Most population health managers already know that prioritizing patients by risk is a critical component of effective care management. So why is it so difficult to segment patients into "high," "moderate," and "low" risk categories?

There are several risk stratification tools that hospitals use to assess patient risk, but in many cases, care managers fall prey to three common missteps when developing and selecting the questions they ask.

1. The assessment tool isn't accurate

The term "patient risk" begs the question, "At risk for what?" Risk can mean different things to different people. The first step to improving patient risk segmentation is tying the assessment tool a specific purpose, be that avoiding preventable readmissions, mitigating unnecessary ED utilization, or reducing mortality rates.

Next, a patient risk assessment should predict patient outcomes. In many cases, the relationship between risk factors and clinical outcomes is not always clear. There are still significant gaps in the literature, particularly when it comes to information on the interactions between different risk factors.

Identifying true patient risk, especially for the moderate- to low-risk patients, requires a multidimensional assessment. Beyond severity of clinical diagnoses and symptoms, organizations should incorporate non-clinical factors into risk assessments to determine the root cause of risk. These non-clinical factors may include psychosocial or demographic indicators, like transportation access or housing stability.

2. The assessment tool isn't user-friendly

A comprehensive risk assessment tool does not necessarily ask patients about every facet of their lives. Exhaustive risk assessments are resource-intensive, may frustrate staff and patients, and often result in complicated or unwieldy outputs.

Best practice assessment tools consist of 30 questions or fewer and can easily be administered across care settings. Likewise, these tools should only include indicators that are easily observable or verifiable through discussion with the patient or caregiver. The results of the assessment should be easy for providers to interpret, and should clearly delineate participants into a particular risk category.

3. The assessment tool is too broad and isn't goal-oriented

Identifying a patient's risk is only the first step. To better manage patient populations, a risk assessment tool should be directly linked to targeted, evidence-based intervention. For example, a particular risk designation may be used to connect patients to existing resources, identify gaps in current practices, or inform the development of new programming.

Lastly, patient risk segmentation is not a static process. Providers should be able to track the tool's predictability to see if it is effectively improving patient outcomes over time. Population health managers and care management teams therefore benefit from embedding performance management and quality improvement processes when applying patient assessment tools.

How to build a better risk assessment tool

To avoid these common pitfalls, population health managers can take three immediate action steps:

  1. Clarify the scope of the tool in terms of the type of risk being assessed, targeted patient populations, and desired health outcomes.

  2. Ask the right types of questions by including clinical, demographic, and psychosocial criteria.

  3. Get frontline staff involved early and often by soliciting staff input and feedback to evaluate the tool's relevance, reliability, and ease of use.

Need to learn more about why focusing just on high-risk care management isn't enough? Watch our two-minute video.

Get the reports you need to change opioid prescribing behaviors

Successful population health managers need on-demand access to clinical analytics in order to impact change across the care continuum. Learn how to quickly identify instances of unnecessary opioid prescriptions with the custom analytics in our Enhanced Analytics Reports Library on October 24.

Register for the Webconference