Collection to action: How to manage data for better patient outcomes

How can hospital pharmacists use data to improve patient care?

Imagine exiting the elevator and walking onto a hospital unit, and right there on the wall is a sign that shows how many days since the last error. If the sign says 365 days since the last error, then hospital staff has a positive story to share. If the sign says 0 days since the last error, then everyone knows that a patient was potentially harmed. How can hospital pharmacists go from simply recording data to using data to improve patient care?


Optimizing care


“Obtaining data is important, but translating that data into actionable patient outcomes is paramount,” said John B. Hertig, PharmD, MS, CPPS, Associate Director at the Center for Medication Safety Advancement at Purdue University College of Pharmacy. It is possible to integrate data “into processes that optimize patient care at the bedside,” he added.


Hertig and his colleagues gave a presentation at the 2015 American Society of Health-System Pharma-cists Summer Meeting that highlighted ways to collect large amounts of data and then do something useful with that data.


Hertig told attendees that data are an “opportunity and a tool” that can be used to untangle the complexity of today’s health care environment. He presented four steps for managing data for performance improvement.


Collecting data


The first step is focused around data collection. “Systems must be designed so data elements are collected exactly the same way over time,” said Hertig. “There are so many different ways to collect data, but having that standardized approach and speaking that same language [are] so critical.”


A standardized approach to collect​ing data ensures accurate and valid data for quality improvement. Ways to achieve successful and reliable data collection include using proven tools, effective techniques, efficient processes, robust frameworks, and ideally, automation, noted Hertig. “Unless you want to sit and collect data via charts, we need to embrace automation and technology as we collect that data and make it meaningful,” he said.


Make a plan


Hertig pointed out that to standardize the process, a clear data collection plan is essential. The plan should standardize the various processes required to collect and measure data. It should also ensure a consistent process for data review and analysis. We know that getting the same [data] over time will help us [understand] our trending [data],” Hertig explained. “If we’re changing our metrics, changing our data source every time, [then] it’s very difficult to see what we’re doing better or worse.”


The standardization process can be especially tricky if an organization’s electronic systems don’t integrate easily. “They have got to be speaking the same language,” said Hertig. 


One of the best tools for standardizing electronic data are SNOMED CT codes, a common terminology used in software applications to categorize clinically relevant information. According to Hertig, SNOMED CT codes are used in more than 50 countries, giving clinicians the ability to consistently input and output data that are reliable and comprehensive.


‘We can do better’


Data should be tracked and shared with everyone involved. “[It’s a] simple concept, but we don’t all embrace that accountability and that transparency,” said Hertig. He advised that data should be shared at regularly scheduled intervals. Be sure to reward and share successes, and communicate feedback about what can be done better, he noted. 


There are many tools for displaying tracked data, such as a flow chart, run chart, control chart, and root-cause analysis. The tracked data can also be used to create scorecards or dashboards that help measure performance improvement. “[This is] rooted in the idea that we can always do better,” said Hertig. “How do we know if we’re successful if we’re not measuring the data and getting that feedback at baseline when we start, [as well as] during and after the project concludes?”


Analyzing vs. interpreting data


According to Hertig, analyzing and interpreting data are two very different things. “Analyzing is when you review the performance data and determine whether it meets the desired quality level. Did you get the antibiotic up to the patient within 30 minutes? Yes? No?” he said. 


Interpreting data is a process of determining the consequences and figuring out what conclusions you can make from the data. “What does the data actually mean?” said Hertig. “How do we interpret that for the patient? [Interpretation is] sinking deeply into the data, and it allows us to act and preempt any issues before they occur.” 


Taking action


If a hospital or health system is not making progress toward its goals, it’s time to go back to the drawing board, noted Hertig. Take a look to make sure your data systems are reliable, and identify any potential causes for underlying systems problems. Reevaluate the changes you made for improvements or increase the frequency of those changes, Hertig added.


It’s important to keep a positive attitude. “Celebrate your team’s hard work and success,” he said. “After you throw the pizza party, look for other opportunities for improvement.”


Data measurement


Joy Meier, PharmD, BCACP, PA, Clinical Pharmacy Specialist and Data Analyst at Sierra Pacific VISN (21), provided a concrete example of how data can be measured and used to optimize care at the U.S. Department of Veterans Affairs (VA). According to Meier, it is important to have an electronic health record that uses structured and standardized data. This allows the data to be “tagged and accurately identified,” she said. 


For example, there is a “standardized drug list for the entire VA, and each drug has a number,” explained Meier. This allows hospital pharmacists to run reports to see which drugs are most often prescribed.


Other examples of analyzing structured data include creating a report to evaluate the number of hospitalizations for patients on oral chemotherapy or whether patients are meeting glycosylated hemoglobin (A1C) targets.


Think before you act


When it comes time to take and use the data to change outcomes, the biggest thing to think about is “the endgame of what you’re trying to create,” said Karl F. Gumpper, BSPharm, BCPS, CPHIMS, FASHP, Pharmacy Informatics Manager at Boston Children’s Hospital Boston. “Even before you start down the road of implementing something, try to have an idea of what you want to measure.”


It’s a lot easier to implement something if you build it in the beginning instead of trying to figure out where to get the data in the end after the automation was already put in or the electronic health record (EHR) developed, he noted.


Questions to consider


Gumpper provided the following examples of why it is important to think ahead when planning your project. One of the most common safety measurements hospitals are interested in is the number of medication errors that have been prevented. “It sounds like a great measure, but is the data really in the EHR?,” said Gumpper. “No, it [might be] in the reporting system, so how do you tie those systems together?”


If a hospital system wants to look at the number of medication overrides performed, this should be determined ahead of time if hospitals are looking at overrides with automated dispensing cabinets or smart pumps, noted Gumpper. Hospital leadership should also discuss their definition of an override and if there are cases when an override is acceptable.


“Something pharmacy is always trying to get their arm around is medication turnaround time,” said Gumpper. He explained that sometimes hospitals know when a medication was made, when it left the pharmacy, or what time it was given to the patient. “But there are a lot of steps in between that we can improve upon,” he said. 


He encouraged people to think about where data are stored, such as a spreadsheet, computer desktop, or Web portal. “Is it on that mysterious server that is in your hospital that only one person knows how to get to? Does the [information technology] vendor own your data?” 


Looking ahead


With so much data available, hospitals and health systems will need to figure out how to structure and standardize the data so they can be used to improve patient outcomes. 


Gumpper highlighted several technologies on the horizon that could have a large impact on hospital data. Track and trace (ePedigree) is coming very soon, he noted. Medication tracking software or the “FedEx for pharmacy” is also in the not-so-distant future, he noted. 


Hospitals will want to think about how they are going to incorporate radio frequency identification or a real-time location system data; biomedical data such as patient monitors, home medical devices like glucometers, scales, or blood pressure cuffs; and patient-reported data, he suggested.


“Think about the future as we look at how we put our data together,” said Gumpper. “Figure out how you can leverage your data and how to close the loop in your program.”