A Competitive Advantage: The Framework for Impact

Here I will introduce a framework I have developed in order to help organizations achieve their goals in sustainably managing complex populations(for more context as to how this was developed and why please see the previous article). This framework consists of four steps: Problem precision, evidence-based, adaptation, and actionable measurements.

1.)   PROBLEM PRECISION

? As a rule of thumb the more targeted your intervention is to the particular problem you are trying to solve, the higher the probability of success.

It’s finding the balance between precision and impact.

Defining your target clearly is the sine quo non of effectiveness. Know exactly where you are going and that the activities you employ will get you there. My research shows that depending upon where you want to end up there are different variables such as care team composition, dosing, frequency, intensity and disease modality targets, that will have a higher probability of leading you to a set a specified outcomes( for example preventing 30 day all-cause acute readmission rates with COPD/CHF in seniors is highly correlated with face-to-face visitation v. telephonic only communication). Because each evidenced-based best practice is much more likely to achieve one particular outcome over another, this should heavily inform the development of effective strategies. Intervention activities are not only associated with achieving certain outcomes, but they are correlated at different strengths.

By defining with PRECISION the problem you are solving, you are more likely to make a lasting and measurable impact.

There are other consequences as well--I find many realize that once they define the problem-- they end up wanting to solve a different issue, or that they need to better understand the primary drivers of the issue before investing resources, or that opportunity costs for this particular problem are too great and they need to re-allocate resources into a more worthwhile investment. Without step 1, these considerations are often challenging to tease out. The next question to ask is whether or not the problem is actionable (can we do something about it) and is it impactable (is what we are doing going to make a difference) ?

I very frequently hear about targeting high-cost patients, but that accomplishes little in regard to the specificity needed to be successful. It is convenient to think this way (focusing on cost), but it is not particular useful if you want to understand the underlying characteristics which meaningfully define your population. Instead we want to measure responsiveness, and to a specific solution.

 If you want to look at high-cost, you need to be clear as to what the primary drivers of these costs are and then tailor solutions to mitigate these drivers.

For a very simplified example used to better illustrate this concept: Instead of high-cost ask about the type of costs: inpatient, what type of inpatient costs: readmissions, how do you define re-admissions: 90 day CHF-related, population: seniors with COPD/CHF and behavioral health issues. You can see how this already looks and feels a lot different than “high-cost.” In fact, there is a growing body of research showing that high-cost and high-risk (as defined by high-utilization) are ineffective from a both a targeting standpoint and a cost perspective. (reference the CCNC study linked at bottom) Instead, employing measures of impactability have been shown to double cost-savings and efficacy. 

2.) EVIDENCED-BASED

When I speak about evidence-based interventions, I am referring to three primary sources that must be triangulated in order to achieve optimal results:

1.) Academic literature (broad)

2.)Proprietary research (population-specific refinement)

3.) Empirical observation (local adaptation)

I use meta-analyses to draw my conclusions (more robust and more likely to withstand population-based and geographic variation), but usually those findings , while reliable, are broad so I complement them with social, environmental and behavioral data (usually with data collected in the home environment),this arms me with the information I need to build effective care models that meet the specific needs of the populations we serve.

The frequency of which I see organizations pursuing strategies that have been proven time and time again to be ineffective is unprecedented.

For example, the practice of using telephonic only case management to manage complex populations has been shown (in most cases) to be ineffective in reducing inpatient utilization, yet this is the most widely used strategy for managing costs in elderly co-morbid populations. Using this framework will help your organization navigate around commonly used, but ineffective initiatives.

3.)   ADAPT

Every model is going to need to be tailored to the local environment. But, there are three primary reasons where I see adaptation fail. If you can look out for these three pitfalls you will have you will have already overcome the most common reasons I have identified that result in adaptation failures. 

1.     Taking an intervention/strategy/model where the results depend on a specific population or geography and assuming it can be generalized to your local environment. Demographics, and location (this includes the regulatory environment and inpatient facility infrastructure) play a key role in intervention success.

2.     Differences in payment mechanisms. Incentives must aligned, which includes understanding how the various components of revenue are paid, when and by who. If this is mismatched in a way that misaligns incentives between the setting from which you are adapting a care model it often ends in costly failure.

3.     Differences in network composition. For example, a care model that works well in a multi-specialty clinic staff--like an employed model with co-located behavioral health providers--may not be efficacious in a primary care only model with an associated IPA specialty network (where providers use disparate EMR sources). Yet, I see the extrapolation of interventions into mismatched network settings frequently. This is also quite prevalent with the ACO programs ( NexGen model in MA plans) where there is an attempt to replicate multi-disciplinary team models without the environment to facilitate the needed care, and in settings without sufficient collaboration among the requisite stakeholders (PCP and specialists).

4.)   ACTIONABLE MEASUREMENTS

Programs need to be extremely nimble in response to changing results/metrics, which underscores the importance of having consistent high-quality measurement systems that are re-assessed over time. Needs do not stay consistent and the needs should always drive program or model functions.

It is vital to collect rich meaningful data and select metrics which will drive program sustainability and efficacy over time. This means going FAR beyond the HRA.

An example: We collect information from proprietary assessments which span a multitude of care dimensions including: home environment, provider/patient relationship strength, cognitive health, social support matrices, and many more factors in order to determine the primary needs of the population we serve. Subsequently, we focus our program activities on using evidenced-based methods to serve those primary needs. However, these needs are not stagnant, particularly in an environment with high turnover, and socioeconomic factors which are affecting clinical care.

When I measured the needs of the our population in 2016 most of my findings revolved around environmental safety in the home, which encompasses increased risk as a result of post-discharge needs. But, in 2018, the needs shifted dramatically--from home safety to issues with coordination and access. Because we tracked the needs over time we were able to pivot our activities to continually meet the primary identified needs of the population.

Using this framework has allowed us to create care models for our most complex populations that have reduced bed days 35%, and 30 day all-cause acute readmissions by 67%.

This set of tools can be extrapolated to a variety of challenges across the care spectrum to achieve improved results in the form of better care outcomes, lowered inpatient utilization and costs and increase patient satisfaction.

Please feel free to connect with me to ask questions or provide comments. 

Sources:

https://www.communitycarenc.org/media/files/data-brief-no-4-optimizing-targeting-cm.pdf