Healthcare is awash in data but starved for information. Progress has been slow in adopting analytics within healthcare. A key ingredient that could speed up the development and adoption of analytic approaches appropriate for healthcare is clarity in both terminology and technical products.
The road ahead requires a classification of analytical needs. Healthcare analytics is all over the map at the present time. It is unclear if this problem is due to the fact that both vendors and healthcare organizations alike are unable to articulate the scope of analytical needs. Problem may also simply be a by-product of what is a rapidly changing market and the struggle that all stakeholders are facing in trying to position themselves for a future state that still remains undefined. Given the lack of clarity in defining analytical needs, vendors offer varied and sometimes random solutions hoping to address problems.
Fixing this problem will require clarity of thought – including a taxonomy and definitions. For simplicity, the bastion of big data—life sciences should be kept separate. My discussion of healthcare analytics relates to patient and population centric data, including data from numerous sources e.g., clinical, claims, etc.
Commonly touted claims from software vendors state that their analytical tools will help physicians make better decisions and improve quality at point of care. However, the claims are often not specific enough. My interviews with multiple healthcare providers and even some technology executives suggests that many are confused by the claims made by tech companies today. “I find this confusing,” is a common refrain. Conversations with numerous vendors in this emerging industry elicited disparate responses to any single problem.
It is hard for target customers to understand the differences among the various offerings in the market. Healthcare providers are under immense pressure to improve clinical outcomes while reducing costs. In this highly competitive arena, effective analysis and use of data (clinical, financial and operational) to improve quality and outcomes while reducing costs will give healthcare organizations a foothold for profitability and success.
Software vendors need to acknowledge openly that no single analytics solution can tackle all the needs of a given healthcare organization. Then, they should identify specific problems that their technology excels at addressing and equally importantly, problems that their technology does not solve well. By doing so, vendors will bring greater clarity to the market as to what their solutions are capable of performing.
In addition, standardizing clinical data entry and providing clarity around this data so that it is truly credible for analysis is going to be critical. Success of analytical tools is highly dependent on data integrity. A single erroneous data has the potential to have life threating consequences. It will be the responsibility of healthcare organizations and their chosen partners to provide this clarity of data for the analytics solutions chosen.
In this nascent industry, vendors should consider incorporating a ‘checks and balance’ tool to catch errors. Clinical data entry can vary due to human oversight. A simple example, an adult patient may see multiple doctors over a period of 6 to 8 weeks. Across all these appointments, it may not be surprising if this patient’s height has been entered incorrectly. As industry veteran, Larry Yuhasz, Director for Strategy and Business Development, Truven Health Analytics (formerly Thomson Reuters Healthcare) recently pointed out, “EMR data is all over the place. If analytics for clinical support has to be optimized, clinical data has to be standardized.”
It will take an unique group of thought leaders in healthcare organizations and vendors alike to intuitively and pragmatically understand the analytics’ challenges within healthcare. Classification of needs will help vendors build appropriate tools that can be nimbly customized as organizational needs change. The products developed by this unique group will own significant market share once the dust settles in this sphere.
An excellent point. Massive data entered by thousands of people often results in unusable data.
If anyone can translate the chaos of unmanaged data, into usable information it may be Thomson, who has been doing it in financial services for more than 20 years.