“You can have data without information, but you cannot have information without data.”
— Daniel Keys Moran
Now more than ever, understanding how to turn data into actionable information is key to driving major decisions and thus revenue. For example, COVID-19 has generated an even heightened awareness of the medication use process from leadership outside of pharmacy, particularly on medication supply reliability and drug shortages. In order to move work forward collectively as a system, it is imperative that data is understood by the different audiences that will consume it for insight. Let us review an example of hospital administrators asking for how long shortage medications may last.
Being able to understand the data is key, beginning with data literacy. Although no single definition exists, it is generally thought of as the ability to read, work with, analyze, and communicate with data.
Reading Data
Being able to read data begins with understanding basic data types and structure, as well as understanding how data is generated, and how it is stored. Data may be structured or it may be unstructured. Structured or discrete data is stored in a standardized way and format. An example of structured data would be the weight of a patient, such as kilograms. An example of unstructured data is free-form progress notes. Data stored using structured methods are more easily queried.
Understanding how data is generated provides context and makes it easier to speak to different audiences. For example, a patient will only have a single birth date, but has many weights over time. Other patient parameters may depend on practice settings. Blood pressures are commonly taken many times during inpatient admissions, but may only be available sparsely and at irregular intervals in the ambulatory setting. Consider how some of this data is generated next time you or your team is working with data analysts or other technical staff.
Working with Data
Data is generated in various related systems, which may or may not live in the same location. Managing your contact list of everyone you know demonstrates this; your contact data probably lives in a combination of these places: your personal email, your work email, business cards, social media platforms, scribbled on a napkin somewhere, and ones you keep in your head. Similarly, medication shortage related data may live on different databases and/or servers (e.g. in the Electronic Health Record (EHR), Automated Dispensing System (ADS), and/or wholesaler purchasing software).
In recent years, a more concerted effort to consolidate the databases of these systems into Enterprise Data Warehouses (EDW) has helped bridge the gap of data connectivity. Having an awareness of the source of data and how they relate to each other is paramount in guiding data requests and data queries. Data dictionaries (a document that provides all the attributes of a data pull and defines them) may help provide this context for interrogators of data; however it is not often provided with each report.
Returning to how long the shortage medication will last, the starting point is often pulling the number of patients with orders for the medications. Typically this data needs to be combined meaningfully with the number of dispenses, average number of dispenses, rate of replenishment, and perhaps many other details that are tracked in different locations.
Analyzing Data
The act of taking data and turning it into meaningful and actionable knowledge requires many considerations. A good baseline knowledge of pharmacy practice and operations is incredibly helpful in preparing reports or graphics that will answer questions or provide insight. For example, thorough medication dispensing reports require knowing the workflow within an institution and the ability to extract data from any technology that facilitates drug dispensing.
Most pharmacy technologies create their own silo of data in a “raw” or unrefined state. Sometimes, data investigation only requires looking at one source of pharmacy data. Typically informaticists need to process the data into a final product that is relevant to the audience in question. Spreadsheet tools, like Microsoft Excel, help with this task. Through manual manipulation or automated features, spreadsheet tools can do many things. For multi-system analysis, exporting data from different systems into a single location is often required. The one location/database can be a single spreadsheet, Access database, a single server, or an EDW.
Data analysis can show trends in the data through Pivot tables. Spreadsheets from different systems can be merged together through VLOOKUP or Power Query functions. Data can also be harmonized and cleaned to a format that standardized to meet reporting and presentation requirements.
If you have large amounts of data that Excel and the amount of RAM you have is unable to handle it, you can use other analytics tools like Python. Using pandas dataframes and other analytics packages like scikit learn and matplotlib can help you process, analyze, and visualize data easily.
Communicating the Data as Information
After the data has been interrogated, the data is organized and presented in a format that meets the requirements of the requester. Many different pharmacy systems have their own reporting tools. The type of question, the results of the data analysis, etc. influence the final presentation. A simple data table with the right rows and columns may be acceptable. Line graphs can graphically show trend data. Sharing mock-ups of the analysis is a good idea to make sure all questions and needs are addressed. In the end, the knowledge obtained from the data will help drive pharmacy decisions.
Having an understanding of data begins and ends with data literacy. Similar to how clinical pharmacy requires lifelong learning, data literacy and reporting skills also require constant upkeep and experience with various tools and knowledge areas. Now that almost everything we do involves technology and the generation of data (e.g. wearables, Internet of Medical Things, devices, smart pumps, mHealth), the understanding of how to convert data into information is a foundational skill for pharmacists.
Authors
From a Workgroup of the Clinical Applications Section Advisory Group of the Section of Informatics & Technology:
Chezka “Mimi” Baker, Pharm.D. Pharmacy Informatics Specialist. Banner Health. Phoenix, AZ.
Raymond Chan, Pharm.D. Pharmacy IS Specialist; PGY2 Residency Program Director. Sentara Healthcare. Virginia Beach, VA.
Andrew Liu, Pharm.D., CPHIMS Informatics Pharmacist. Rush University Medical Center. Chicago, IL.
Clement Ng, Pharm.D., CAHIMS Clinical Pharmacist, Informatics. University of Maryland Medical Center. Baltimore, MD.
References
- Bhargava R, D’Ignazio C. Designing Tools and Activities for Data Literacy Learners. Journal presented at In Wed Science: Data Literacy Workshop. Oxford, England, UK; 2015 Apr 1.
- Dataversity. The Importance of Data Literacy. https://www.dataversity.net/the-importance-of-data-literacy/ (accessed 2020 Apr 28).
- Federer L, Foster ED, Glusker A, Henderson M, Read K, Zhao S. The Medical Library Association Data Services Competency: a framework for data science and open science skills development. J Med Libr Assoc. 2020;108(2):304-309.
- Gartner. Champion data literacy and teach data as a second language to enable data-driven business. https://www.gartner.com/smarterwithgartner/a-data-and-analytics-leaders-guide-to-data-literacy/ (accessed 2020 Apr 28).(accessed 2020 Apr 28).
- Harvard Business Review. Boost Your Team’s Data Literacy. https://hbr.org/2020/02/boost-your-teams-data-literacy (accessed 2020 Apr 28).
- National Library of Medicine, Health Services Research Information Central. Data Literacy & Management. https://hsric.nlm.nih.gov/hsric_public/topic/data_literacy/ (accessed 2020 Apr 28).
- National Library of Medicine, National Network of Libraries of Medicine. Data Literacy. https://nnlm.gov/data/thesaurus/data-literacy (accessed 2020 Apr 28).
- Qin J, D'Ignazio J. Lessons learned from a two-year experience in science data literacy education. Journal presented at the Annual International Association of Scientific and Technological University Libraries Conference. West Lafayette, IN; 2010 Jun.
- Towards Data Science. 25 Sapient Principles for Better Data Literacy. https://towardsdatascience.com/25-sapient-principles-for-better-data-literacy-5ff5d42480a5 (accessed 2020 Apr 28).
Other Resources
- Code Academy. https://www.codecademy.com/
- Coursera. https://www.coursera.org/
- edX. https://www.edx.org/
- Khan Academy. https://www.khanacademy.org/
- LinkedIn. https://www.linkedin.com/learning/me
- openHPI. https://open.hpi.de/
- SoloLearn. https://www.sololearn.com/
- Udacity. https://www.udacity.com/
- Udemy. https://www.udemy.com/courses/