Business Intelligence Blog

Latest Stories

Featured Stories

Filter By Categories
Chris Dickson
By
August 16, 2016

Population Health Management Part 1: Self service analytics

populationhealthmanagement

What is Population Health Management?

Before I get into the details I just want to clarify my discomfort with this term and its definition.

The term is generally defined as the use of data and algorithms to identify individuals and population segments for whom a specific intervention would be beneficial. If you pop the term into a search engine you are presented with the following explanation:

 

"Population Health Management is the aggregation of patient data across multiple health information technology resources, the analysis of that data into a single, actionable patient record, and the actions through which care providers can improve both clinical and financial outcomes."

 

It should be noted that management of population health cannot be simply an information based activity. It may seem obvious to state that the involvement of clinicians and service managers is essential to successful change in health systems. I would be more comfortable if the word management were replaced with "Analysis" or "Profiling" for the data element.

 

We currently see a push in health services towards Population Health Management, a required element of the Local Digital Roadmaps being submitted to NHS England currently. This is based on the belief that with correct application of informatics (the collection, management and application of various analytical tools) the NHS will improve patient care and outcomes, reduce inefficiency and deliver financial savings through effective targeted intervention.  

 

Over a short series of posts I will explore the elements which organisations need to consider when implementing their roadmaps. 

A short history

We could go back over 30 years to the work of Dr Barbara Starfield, but that was all very academic and future thinking. I'll start where my story with Population Health Management began, although it wasn't defined as such then. In 2005, while working in the East of England we were asked by Essex PCT to help them assess a tool they had been working on with the Kings Fund and Health Dialog to predict patients at risk of re-hospitalisation. The tool was built in MS Access, the interface was a bit clunky but on testing it had a reasonable hit rate, Over the coming years PARR+ and PARR++ were developed, adding more data feeds and more complex algorithms. At around the same time the Hampshire Health Record came into existence. This was the first care record in the UK bringing together data from Primary, Acute, Mental and Community care settings into a single searchable record for the purpose of providing a more joined up service to patients, alongside this an anonymised version of the Hampshire Health Record was made available to public health analysts and educational bodies for research purposes. These two events - the development of algorithms to identify patients who may benefit from specific interventions and the creation of a care record bringing a holistic approach to health informatics and analysis - were the earliest shoots of this type in the NHS.

 

Fast forward ten years and;

  • the information age is starting to be leveraged by health services
  • super-fast storage and processing of vast amounts of care record information is now possible
  • the discipline of data science has gained recognition as something the NHS should be adopting and exploiting
  • risk stratification algorithms are moving from highlighting individual patients to receive individual care packages to identifying sub-populations for whom a type of management should be adopted (some risk stratification tools always did this but the output wasn't leveraged in that way)

New models - old thinking

Currently NHS organisations that are implementing these systems and principals are applying very similar staffing structures to this work as they with previous health informatics work. A team of knowledgeable analysts gather requirements from clinicians, extract relevant data, apply some analysis and present their finding back to the clinicians - this process has the potential to iterate ad infinitum as answers to questions inevitably lead to further questions.

Self-service analytics

A fundamental movement in analytics over the last few years has been a shift towards knowledgeable staff directly interrogating data about their area of expertise, supported by intuitive technologies that no longer need technical expertise to use. This has allowed the information professionals to focus on providing the best quality data possible to these professionals on a stable managed platform. 

 

Clearly there is a strong belief in the BI market place that this is the way forward, as there isn't a vendor on the market that hasn't now brought out their version of software for this. Where Tableau led being built from the ground up for visual self-service analytics and is still the clear leader in the technology type, Qlik, Business Objects, Microstrategy and Microsoft have all followed. In addition, there are any number of smaller niche offerings now available as cloud technologies (Good Data, Sisense etc.)

Self-service analytics in population health

As mentioned earlier, NHS organisations looking to implement these processes need to reinforce the role of the clinician and service managers in the process and, in fact, go further in their engagement. By implementing a BI solution that has self service analytics at its heart it is possible to allow experts with the correct knowledge and experience to create their own analytics reports. With the correct management of the environment and data integration layers combined with a tool such as Tableau these experts can explore the healthcare data of individuals and populations, create their own analysis, define their own population segmentation (there will be a later post specifically about this) and create reports that can be used by themselves and other clinicians and managers to identify opportunities and improve outcomes. They can also create the reports they will need to monitor the outcomes of their work.

 

New Call-to-action

 

Subscribe Email