Healthcare Data Analytics Consulting

The difference between the potential of AI in health care versus what is actually deployed is more about the failure of execution than the absence of data. Despite the fact that funding to support AI pilots and the development of dashboards has occurred, clinical and operational workflows have continued to remain stagnant as gaping holes continue in the system.

Hutchins Data Strategy Consultants has an objective of closing the gaps for health systems, payers, and life sciences organizations. This objective is constructed on the premise that data and AI investments are made to optimize the quality of care while enhancing the operational and financial performance of the organization. It is a strategic failure before it is a technological failure if this premise is not achieved.

What Goes Wrong

People in analytics roles in health care systems have come to rely on the present being predictable. Resources are allocated to a system that is new to the health care unit, a new team is formed, and new report frameworks are established. Within the year, the leadership is asking about the same thing, what is the return on this investment?

The systems needed to establish a reliable data trust have not been constructed. Capacity for analytics is built with little to no operational guidance, and the needed transformation to facilitate the adoption of insights from dashboards to workflows is grossly underestimated.

As AI algorithms become increasingly widespread, we are starting to see what might be called a fourth failure mode. Traditionally, analytics teams decoupled their models and deployed them within the contexts of supportive oversight practices. This would allow teams to assess the propriety, safety, and validation of models within clinical and operational domains. Without oversight, we are left with pilots that eat analysts’ time while they stagnate and rot the analytics function.

This is primarily a lack of strategy. In order to address these problems, we need to hire someone who has practiced within the environment where the problems are being adjudicated. This is not a case where we can hire someone from the outside who is selling software or frameworks.

This has presented a gap in the marketplace that will be filled by consulting. Many analytics-related consulting engagements lead to products that are designed to be presentation artifacts, rather than artifacts that will facilitate the execution of strategic initiatives. Proprietary maturity models, capability assessments, and technology roadmaps are useful diagnostic frameworks, but are ultimately devoid of strategy. The analytics team will adopt a strong strategy in order to define the actions they will take, the outputs they will realize, and the impact these will have on the organization.

Why Healthcare Analytics Fails at the Data Layer

Most analytics issues manifest in the creation of dashboards. However, they can actually be classified as data issues. These data issues can be a report arriving late, or two different teams creating two different reports containing two different measures, and each team will refuse to take responsibility for saying which report is accurate. In each of these cases the same underlying data issue is present. Data for each of these reports has been captured in two or more different systems, using definitions of the data that have not been agreed upon by the teams that captured the data.

The cost of disparate data streams becomes clear when you look across project portfolios. Each project gathers its own data stream and works around the gaps. Most projects don't utilize all the data they collect, so nothing seems to stop them from progressing. The weight of these project-based data streams becomes apparent when outputs must be reconciled across the project portfolio. This can take a long time. Employees devote weeks of effort to validating the data and determining what it means. Projects contain this data work and label it "planned complexity." This ultimately amounts to the cost of the unresolved data problems carried forward to the next work initiative.

It is not uncommon for separate project teams to be completely unaware that they are building duplicate data processes. Often, this duplication is only discovered once the data processes are in production. At that point, the cost of reconciling the two data processes exceeds the cost of maintaining them both. Underlying data dashboards or modeling that is validated in one environment can be produced in another environment and begin to show different results for the same data, leading to a deterioration of data quality. Doing analytics consulting without validating the data thus far just leads to fancier dashboards on the same unreliable data.

All work is focused on determining the where and why of the previous attempts that have stalled. Healthcare organizations have not completely stopped all work. There are existing investments and staff, as well as the political climate. Without an understanding of all these aspects, your strategy will be immediately obsolete.

Our approach is governance-first. Until we verify the establishment of governance preconditions, like data quality, access control, and decision rights, and the organization’s readiness to utilize its data, we will not propose any new technologies, hiring, or initiatives.

We then focus on the following:

Analytics to drive operations

We have to identify how we integrate the analytics function to operations and how we can maximize the potential impact of our analytics efforts. Every analytic function needs a consumer and should drive a decision. There needs to be a feedback loop to ensure a value creating analytics function.

We understand the operational AI use case challenges that our healthcare clients experience. As a result of our understanding, we create frameworks for healthcare clients that provide greater clinical oversight to operational AI use cases and greater internal review to ensure that AI is implemented in a safe and fair manner.

Data strategy creation

This work is tied to data-connectivity strategies that link organizational goals to data investments over time, including the organization’s data architecture and design, and strategies for workforce, vendors, and continuum oversight. We develop operational plans that include owners, goals, and deadlines. These plans are actionable and not made to be forgotten.

Operational Integration & Change Management

Most analytics investments are wasted due to insights inaction. These insights may exist in the form of abandoned dashboards, predictive model outputs that never reach decision-makers, or reports that become irrelevant because they are delivered after the fact. Analytics outputs must be designed to be situated within workflows. To design analytics outputs for workflows, a comprehensive understanding of team structures and operations is required. This cannot be achieved by forcing employees to work in a way dictated by the technology.

Who We Work With

Our clients include health systems currently undergoing digital transformation, payers constructing analytics and value-based care models, and life science clients utilizing real-world evidence in the clinical value and market access framings. All of our clients are looking to understand the operational impacts of their data and AI investments.

Engagements cover CDOs, CIOs, CMO, and business leaders accountable for the outcome the analytics will deliver. To conceptualize ideas that will fundamentally transform operations, the business and technology sectors of the organization must be concurrent.

We work with organizations that are developing analytics capabilities. Many community hospitals and regional health systems do not have analytics teams, but understand the value in developing one. These organizations have specific challenges, such as deciding the order of their initial investments, what to develop internally vs. externally, and how to avoid the challenges that larger systems encounter. This advisory work pulls upon the author's experience with the specific challenges analytics functions face as they evolve from basic reporting to a much-needed and appreciated strategic function.

How This Work is Different

The analytics consulting industry is full of companies that create evaluations, maturity models, and technology roadmaps. Here at Hutchins Data Strategy Consultants, we focus our work on the client, not a specific framework, and this is the main differentiator from our competitors. This is especially true in healthcare analytics, where the application of frameworks often results in project failure.

Every health system has their own implementations of EHR, data storage, staffing, oversight, and leadership maturity. Analytics consulting that uses the same approach with every client ignores the factors that ensure how analytics is actually used. Our consulting work is based on the client's individual context, and is not a copy and paste of other work.

Selecting a Healthcare Analytics Consulting Firm

There are many healthcare analytics consulting firms and they all say they do similar things. The real difference comes later and can be seen with the outcome of some recommendations vs. others. A few questions can separate consultants that provide insights vs. those that provide another version of a framework.

Where have the consultants actually worked? The majority of healthcare consultants have never set foot in a clinical area. Consulting firms that you should consider can describe the pathways of a clinical indicator/report and the steps they took to resolve issues at every stage.

What things will you not do? Good consultants will say you are not ready for a new hire or analytics platform and will offer to fix the data and the analytics framework. Firms that will do anything are selling time not results.

What does your analytics consulting look like? The best engagements take your team to a point of being able to do consulting on their own. If the end goal is a presentation instead of recommendations, the consulting has ended prematurely.

At Hutchins Data Strategy Consultants we take a governance and listening to decisions approach to healthcare data analytics consulting. We find the data and definition issues and partner with you and your team to build analytics solutions that address your current everyday decisions.

Authoritative sources

• Agency for Healthcare Research and Quality
• HIMSS
• ONC HealthIT.gov
• HL7 FHIR interoperability standard