Originally published March 29, 2007
There is a general consensus among industry experts that healthcare organizations (HCOs) – hospitals, specialty clinics, and other medical facilities – have lagged behind other vertical industries in embracing business intelligence (BI) and adopting it as a means to drive care improvement and cost efficiencies. While some HCOs have transcended the early stages of BI adoption and begun deploying important information to business stakeholders, the vast majority continues to struggle to educate the business on the use and promised efficiencies of business intelligence and how it can drive business value.
Of the myriad articles published on how HCOs can leverage business intelligence, only a few explore the reasons behind their lagging or late entry. Understanding the reasons why HCOs have struggled elucidates the specific cultural, technical, and functional challenges they have faced, and prepares companies at similar stages of the BI life cycle to tackle them in a deliberate way.
The following issues are the main reasons for healthcare’s sluggish BI adoption. As widespread as they are in healthcare, companies in retail, telco, manufacturing, insurance, and pharmaceutical and life sciences confront the similar challenges as they embark on or revisit their BI efforts. Indeed, some might sound very familiar:
Proprietary Systems as the Norm: Healthcare systems are typically packaged applications that automate patient care, physician/provider administration, contract management for health plans, and supplies management. It is not uncommon for an HCO to have a couple of dozen heterogeneous operational applications coexisting, and often trying to share data. Some even have multiple systems handling single business processes, for instance, claims processing.
Moreover, many health plan providers have grown through acquisition, further exacerbating the problem of rampant and proprietary data silos. These siloed systems in turn invite redundant data entry, which, of course, drives up the costs. Adding to the complexity, many of these systems have back-end proprietary databases and data structures. Many of the software vendors focused on the healthcare market offer HCOs scant documentation, no data extract interfaces, and no data models or data dictionaries that can help annotate the systems and their functions. This fact alone causes major data integration challenges and hampers BI efforts to extract common data from potentially scores of disparate systems with little in common except for the generation of unsynchronized and mismatched business data.
Immaturity in Data Usage: HCOs spend an immense amount of time accessing and understanding operational reports. Most of the time, integrated reports and dashboards are built by senior data experts who manually “plug in” values into spreadsheets – a classic way to introduce data quality issues. Manual report building practices are a key indicator of a low BI maturity level. (Baseline notes such practices as “code reds” in our BI Scorecard assessments.) This is in no way intended to play down the importance of operational reports to make day-to-day tactical decisions: hospitals live and die on their operational metrics. But HCOs have begun adopting more disciplined approaches to generating operational reports that can ultimately support strategic planning and decision making driven by (not merely supported by) data. For example:
Lower Maturity: Report on how nursing staff are allocated in each unit today. Also, show patient bed assignments today.
Higher Maturity: How can I predict the number of nurses and the skill levels of those needed in each unit based on my staff utilization, bed assignments and capacity, and projected patient flow (admissions, discharges and transfers) for any week I chose?
Lower Maturity: What is the number of patients with DVT (blood clots), infections or overstayed patients in a given unit last month?
Higher Maturity: How can I predict which patients are at risk of facing harm events based on patterns or combinations of patient medical history, symptoms, diagnostics, medications, procedures and practices followed based on the historical data and outcomes data? The need for this type of capability is driving new business cases for data mining using decision trees, statistical analysis and neural network modeling. It also mandates granular and historical data.
As with all industries, HCOs are working toward data-driven decision making and using data to enable strategy fulfillment. Both goals are business-driven, and should thus be inspired by executives on the business side.
IT Treated as a Cost Center: Despite their maturing visions for the use of information, most HCOs still view IT as facilitating application and systems support. In reviewing the results of many of our assessments, we’ve found that IT organizations in other verticals demonstrated more mature BI practices and successful implementations and are frequently part of overall company strategic initiatives. HCO IT groups ultimately have more to offer their business counterparts. As each HCO line of business specializes in a narrow area such as neonatal intensive care, cardiac intensive care, surgery OR , and radiology, they become completely focused on their specific areas and the applications (and data) they use. This results in a lack of an “enterprise” view of data as cross-functional and reusable. Immature HCOs don’t view data as a supply chain in the enterprise, but as a by-product of their respective silos. IT can become the agent of change for the HCO to empower the organization with integrated information to support the business through business intelligence. HCOs should elevate IT beyond systems support and entrust IT organizations to begin gathering requirements for information, thus deploying business intelligence as a business solution.
Managed Care, Unmanaged Data: Data management and governance are essential practices to manage data as an enterprise asset. Because the majority of HCOs are still focused on applications, few are data-centric. But as they continue to accumulate data about their patients, the need to manage that data in a deliberate and structured way becomes more critical. This includes understanding issues of availability, integrity, and access and security, all of which fall within the boundaries of data management. Companies in other vertical industries have largely made the leap to viewing corporate data in terms of having a life cycle that needs to be managed. Several recent compliance regulations, HIPAA among them, are compelling HCOs to adopt a more holistic view of data and institute data governance practices in order to develop and enforce policies around the data. Rigorous data management practices – like business-driven data architectures and data quality processes – represent the execution of those governance policies which, when realized, will enhance BI capabilities too.
Different Locations, Different Incarnations: Data is dispersed far and wide, and in different formats within HCOs. It can still be found in:
The Ongoing MDM Challenge: As businesses, healthcare companies engage not only patients and physicians, but other clinicians, vendors, drug company reps, private insurers, managed care divisions at both state and federal levels, regulatory bodies and others. HCO information also encompasses various drugs, drug brands, drug interactions, drug compounds, and a variety of disease classifications. Business-driven BI programs rely on high-quality, cross-referenced master data for a range of subject area data. The very nature of healthcare mandates that cleansed, matched, and reconciled master data be propagated not only to BI systems but to operational systems that continue to proliferate and disseminate their own versions of information. Initiatives like master patient index in healthcare are evolving to become enterprise master person indexes, registries capable of reconciling party and product data in real time. Thus the entire system and business infrastructure of a healthcare organization can reap huge rewards from MDM.
Although HCOs face often-daunting challenges on their way to reaching BI goals, understanding the various information usage needs across the healthcare system, instituting formalized data management practices and organizational roles, and ultimately hiring experienced BI skills can result in huge payback in terms of cost savings, revenue generation, and healthier, happier patients.