Fundamentals of Data Analysis in Healthcare

The recent proliferation of connected devices, sensors, and other equipment has made it almost too easy for healthcare organizations to acquire data. The potential benefits are evident: mining this data to make more informed decisions about their internal operations and patient care.

The problem however is that just having access to data does not in itself produce results, either because it is not reliable or not easily understood. Healthcare institutions need to focus on the fundamentals of data analysis to uncover the relevant nuggets of insights which help drive decision-making.

Bad data muddles up analytics, and bad presentation of data can put the focus on the wrong things or miss the mark altogether – it’s essential that the data be trusted and actionable. So how can healthcare facilities identify more meaningful insights that ultimately improve patient care? Here are some things to keep in mind:

Make data available in real time
The emergence of real-time data sources is having a dramatic impact across all industries. Data analysis no longer has to be a retrospective waiting game. It’s now enabling organizations to ask, “What’s going on right now, and what can we do about it?”

Take for example the challenge of improving patient satisfaction scores. Hospitals have a pretty good idea of what contributes to a positive patient experience: short wait times, meaningful interaction with caregivers, and effective communication with patients and family.

But a report on essential metrics like on-time start percentage and patient/provider contact time that arrives even a day later isn’t very helpful. It’s hard after the fact for caregivers and managers to link these stats to specific events and thus gain insight on how to do better.

If the information is delivered in real-time by leveraging technologies such as real-time location systems (RTLS), caregivers can respond immediately to a patient who has been waiting too long and managers can better anticipate and eliminate bottlenecks in the overall patient flow. Data created by an automated RTLS can be much more accurate and timely than that entered in manually, often well after the fact and also much less accurately.

Present data in a simple, meaningful way
Simple, visual dashboards are essential to make the data actionable. They enable staff to more easily monitor and understand a patient’s care process in a manner that is intuitive and doesn’t require data analysis expertise, while quickly identifying pain points of procedural inefficiencies to get ahead of a problem before it occurs.

As an example, take a look at hand hygiene procedures in a typical hospital. Healthcare-Associated Infections (HAIs) cost organizations over $35 billion annually and are a pervasive threat to patient safety. Using real-time monitoring and dashboards, however, hospitals can show staff members how they are doing individually, and show managers how the unit or hospital is doing overall – all through visual analytics. This allows for immediate action or longer term interventions such as further education or mentoring.

Dashboards take real-time information about patients, staff and assets, and provide faster insight into how to improve the patient experience – whether by resolving issues that impact wait time or seeing what’s causing reduced staff contact time. Doctors and nurses can even access dashboards on-the-go on a tablet or wall-mounted display to gain real-time visibility into what’s happening in the OR, waiting rooms or post op rooms to make sure everything is running smoothly.

Today, dashboards loaded with predictive analytics are becoming a reality as historical data uncovers trends. Looking at the facility’s records, organizations can better predict trends in the coming hours, weeks and months to make more informed decisions, such as using previous data on infusion pump deployments to identify how the devices should be distributed and when more will need to be ordered or rented.

Bring on a data expert

There’s a movement to bring self-service analytics to the masses. Business intelligence (BI) and data visualization tools like Tableau, Qlik, Microsoft and SAP are paving the way for non-technical individuals to analyze and make sense of data. But the simplicity of these solutions for users masks great sophistication on the back end both in terms of managing the data and building dashboards that non-data experts can rely on to make strategic decisions.

Organizations need to be able to differentiate between “good” and “bad” data if they hope to avoid confusing or non-correlated results. An experienced analytics team knows that data integrity is the key to success.

Healthcare institutions should either look into third-party vendors to handle and manage data analysis, or find a data expert to bring in-house. There are increasingly more BI teams emerging within hospitals as of late, due to the value deep data analytics provides. As the availability and applications for analytics solutions continues to grow, it’s safe to say this trend will only intensify. Often these in-house teams will partner with vendor teams who are experts in their solutions as a starting point, then take over day-to-day operation of the BI systems once launched.

Spot and breakdown data silos
When implementing new technologies, it’s also important to consider data silos; particularly how to avoid creating new silos and how to eliminate old ones.

Data silos are repositories of data that are isolated from other parts of the organization. Healthcare groups should use all of the data that is available to them to drive more informed decisions and ultimately help improve patient care.

For instance, being able to combine RTLS data with clinical performance data provides caregivers with a more complete picture of the patient journey. Ensuring that the RTLS system is integrated to the clinical system and that a common “key” exists is essential to being able to blend and analyze this type of data.

The good news is that data silos are being broken down more and more, and larger organizations are leading the charge with tools like Tableau, Qlik and SAP. By taking advantage of the connectors in BI solutions, organizations can easily combine SQL, Oracle, Excel data and more to gain holistic discoveries.

The amount of data flooding through hospitals today is unprecedented. But that doesn’t mean it’s being used effectively. Organizations need to develop a strategy that delivers on the basics of data analysis — understanding the data at hand, ensuring its quality, finding the relevant bits to combine and analyze, and presenting it in a way that’s easy to consume. It’s an approach that’s already delivering results in better care and higher efficiency in forward thinking healthcare organizations.

Image:  Stuart Miles,

Source –

Focusing on the fundamentals of data analysis in healthcare