Healthcare Data Strategy

A data strategy should answer one question: how will this organization utilize data to improve what it does? In healthcare, there's more pressure to answer this question than in other industries. The data connects to the patients, the policies have little room for error, and the operational challenge of just a medium-sized healthcare system makes most data frameworks useful to a limited degree.

Still, the majority of data strategies in healthcare operate, at least formally, as though the organization exists in a vacuum. They describe target architectures, oversight models, and maturity assessments without defining the specific operational challenges the data is meant to address. The end result is a document that, in board meetings, is articulate and informative, and offers very little in the months that follow.

Hutchins Data Strategy Consultants take a different approach to healthcare data strategies. We define our strategy as tackling the question of what decisions are being made incorrectly because the required data is unavailable, of poor quality, or not used, and what are the quickest means to resolve these issues. The operational gaps the strategy aims to address take precedence over a perfect data architecture.

Why Most Healthcare Data Strategies Fail

The pattern of failure is consistent and predictable. A healthcare system realizes that it is failing in the analytics domain. Then the leadership makes a budgetary decision. Subsequently, a consulting firm is engaged or an internal activity is undertaken to produce a strategy. The strategy includes a target state, technology recommendations, structures of oversight, and phased plans. About twelve months after this has been done, little progress has been made.

Most failures are structural rather than technical. Operational leaders, who would actualize the strategy, were excluded from the strategy's formulation. Assumptions about the oversight model were that there was order and discipline in the organization, which in fact, was absent. Recommendations were given for technologies that would serve the vendor's agenda rather than the organization's problems. The roadmap was developed without considering competing demands of the budget and the political order which determines the prioritization of initiatives.

This seems to be the general case with the healthcare data strategies. There is a thriving consulting business that develops healthcare data strategies that are internally coherent and externally so, but are disconnected from the realities in which healthcare systems function.

What a Working Strategy Requires

There are conditions in healthcare data strategies that yield desirable outcomes, and the great majority of strategic planning exercises entirely miss them.

Grounding in operational reality is the first thing that needs to be done. The strategy needs to begin with the leaders in charge of the clinical quality and the revenue cycle; the supply chain; and the population health management. Their pain is what determines the priorities. Without their input, the strategy is largely focused on the systems architecture with little or no impact for the organization.

Before data technology can be developed and leveraged, data governance must be established. True data governance is the layer of oversight that allows organizations to build reliable data, control access to data, make decisions regarding the data, and define the appropriate actions to take when data governance malfunctions. Technology investment is futile in the absence of a working layer of governance. Conversely, when oversight is treated as an afterthought, even the best technologies will underperform.

The capacity of an organization delineates the type of governance that can be established. For example, the capabilities of regional 200-bed hospitals are vastly different from a network of 15 regional hospitals. Likewise, governance strategies that work for one will almost certainly not work for the other. One of the biggest pitfalls in healthcare data governance is the application of a governance strategy that was designed for a different scale.

Every initiative should be directed toward an observable outcome. Recommendations that cannot be correlated with an improvement in healthcare delivery, financial outcomes, or operational efficiency are not fitting for the overarching strategy. This is more about discipline than an attempt to scale down. Within organizations that maximize their data investments, the most distinguished leaders are those who do not invest in initiatives that cannot be substantiated in terms of their anticipated outcomes.

The Development of Data Strategies in Healthcare

Our process begins with a more extensive review than most companies would expect. Our assessment includes a detailed analysis of existing data systems, alongside a review of the systems within the organization that utilize the data. For example, in what areas of the organization are leadership and the clinical and operational teams making data driven decisions? What data initiatives have been implemented in the past, and what obstacles have prevented their success?

Our process consists of three layers to develop the healthcare data strategy.

The first layer is the oversight foundation, which includes decision making frameworks, data quality frameworks, data access frameworks, data stewardship frameworks, and data conflict resolution frameworks. In the case of healthcare organizations that utilize protected health information, this layer includes oversight for compliance with HIPAA and respective frameworks. The risk frameworks are the minimum standards for the oversight layer and serve as the foundation for downstream analysis and artificial intelligence.

The second layer is an analytic capability model. This layer focuses on the organization's priorities and areas of focus that were identified within the assessment. This layer is an organization's plan for the data and includes the outputs, the required skills and roles, and the supporting tools and technology. The plan includes the outputs and the intended impacts. The plan details the sequence of activities, and ownership with a defined timeframe.

AI Readiness

The third layer focuses on AI readiness. From an AI readiness perspective, it assesses the AI use cases given the current state of data and maturity, oversight posture, and organizational readiness. This includes model validation, integration into clinical workflows, and vendor assessments, as well as the ethics of AI in healthcare. This layer of the framework addresses the two most significant pitfalls in healthcare AI: moving too quickly without oversight, and moving too slowly due to lack of focus.

Who This Work Is For

We designed this work for healthcare systems, payers, and life sciences organizations, particularly for those investing in data-focused strategies and systems but expect no or little return. This also serves those organizations that are about to make substantial investments in data and technology. For those peer organizations, the implementation of data strategies and systems is both costly and strategically harmful.

Most of the time, the leadership we work with are the Chief Data and Chief Technology Officers and other leaders accountable for data outcomes. Engagements are most effective and easy to implement when both the technology and operations sides are aligned. The co-creation of strategies from both sides ensures a high likelihood of effective implementation.

We partner with organizations undergoing leadership transitions, merger integrations, and EHR migrations. We treat these as opportunities to develop data strategies. The appointment of a new Chief Data Officer triggers a series of strategic data decisions that will shape the organization for years — including how it integrates new facilities. Data strategies during a merger are of greater value than the strategies that attempt to redirect new resources and costly data infrastructure that result from the merger.

The sentiments of every customer are the same. They believe data strategies have not delivered the expected value. Recognition of the failure is the first step. Success is measured by how the losses from the investments are minimized.

Recognizing Mistakes

A data strategy is confused when a data warehouse is seen as strategy. While a data warehouse is essential to a strategy, the infrastructure creates systems of information, the result of which is reports, the use of which no members of the organization will act on.

Filling the Chief Data Officer position without restructuring institutional power is another significant issue. The CDO role is now somewhat commonplace in healthcare. However, the role is only effective if the organization is ready to alter how it makes decisions concerning data. A CDO who lacks the power to allocate budget resources, a place on the Operating Committee, or an active sponsor from the leadership team will generate plans that are perfectly designed from a technical standpoint. However, those plans will be irrelevant to the institution.

Another example is the underinvestment in data quality. It is a known fact that all analytical initiatives are fundamentally reliant on data quality. Attributing a certain degree of confidence in the quality of data to a specific analytical initiative is common in the industry. However, that is not reflected in practice. Healthcare data strategies are, therefore, constructed on an ever-shifting foundation.

Moving Beyond The Strategy Document

A document that resides on a shared drive is nothing more than a report. The true measure of strategic work is the operational changes that occur as a result of the document. For this reason, each consulting engagement comes with an implementation roadmap with defined milestones, owners, and a cadence of monitoring that maintains visibility of the work to leadership.

Authoritative sources

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