SELECT itemid, label, category
FROM D_items
WHERE category in ("Hematology", "Chemistry")Analysis Report One: What’s Your Data Strategy
Executive Summary
Healthcare managers are looking down the barrel of an inundating amount of information, but most of them do not know where to begin to decipher the data they have been given. According to a 2013 study from SRM university there are 3 main categories including storage, management of the data, and financial aspects related to the data. In order to properly understand all this data the healthcare industry, along with other industries, have worked to create C-suite positions revolving completely around data management. Having an individual such as a Chief Information Officer (CIO), may be vital as the amount of information contained in a healthcare system is massive. These individuals will be able to better manage the data to ensure the healthcare system has an appropriate data defense where all required regulations are met, as well as, data offense where the CIO, along with other C-Suite individuals extract and use this data to form a business plan for the future.
Introduction
Healthcare systems create and manage a lot of data revolving around patients, their care, admissions, re-admissions, etc. Unfortunately, most healthcare leaders are unable to extract this data in a manner that is useful to back decisions relating to patient care or business planning. Many companies have moved towards creating an entirely new role that handles the system’s data. This is to ensure that the proper protocols and regulations are met in order to decrease any risk that may effect the data such as a data breach. Secondly, on the other hand this individual will also need to ensure they have an offensive plan devised from the data. This ensures that the data is being put to use to help make strategic decisions about what the healthcare system should do. As with any data management there are always trade offs. If the health system is too defensive about its data it may not be able to produce a profitable strategy. If the company is too offensive there may be lack of defense of the data possibly putting the data at risk of a breach.
Along with the management of data these information mangers will have to be able to use data architecture that includes both Single source of the truth (SSOT) and Multiple versions of the truth (MVOTS). SSOT typically is involved with data defense while MVOT is involved with data offense. To key to success is to integrate the two to create understandable, meaningful reports for other health managers to be able to use help them make decisions. Having a CIO will be a game changer for hospitals where the general theme around data is to protect it at all costs (a defensive strategy). However, having an individual with the ability to carefully extract pieces of data from all that a hospital collects will be priceless in understanding a hospital’s metrics, costs, etc. (Huer (2018))
The Healthcare Context
As with other organizations, healthcare systems are struggling to handle the mass amount of data that they collect. This data includes information such as patients’ ages, marital status, diagnoses, medications, insurance, and blood lab values. Thankfully most Electronic Health Records (EHRs) are good at compiling all this data into a user-friendly platform to ensure the patient receives the best care. However, putting data into these EHRs is the easy part. It is very difficult to find a leader in healthcare that has the know how to extract data from the EHRs and create actionable reports to help drive initiatives forward. (Balaraman and Kosalram (2013))
This problem not only plagues healthcare systems but retailers and even banks. Our world today constantly produces data whether that be in the form of transactions, entries my a nurse, articles posted by the news, etc. Data is everywhere and there is a lot of it. This may lead to the installation of Chief Information Officers (CIOs) in companies where they can work the back end of the database and formulate aesthetically pleasing reports that provide those who make the decisions for the company a quick “snapshot” of what the company has been doing. This is beneficial across all industries but, especially in healthcare. The Centers for Medicaid & Medicare Services (CMS) implement standards upon healthcare systems that have to be met in order to receive payment from these insurance plans. For example, patients who come in for a heart failure exacerbation are closely monitored by CMS. (Hentel, Menard, and Khorasani (2017)) If this patient’s health problems resolve but they return within 30 days of their discharge the hospital receives no payment from CMS for the second visit. This is why data analysis and being able to extract it in a meaningful way is the future of healthcare system management. It not only helps the systems to see how they have been doing but can help predict how the future may look for the company.
Data Visualizations
Visualization One - Offensive
ggplot(data = items_2,
aes(x = category)) +
geom_bar() +
theme_minimal() +
labs(
title = "Visualization showing relative frequencies of item categories",
subtitle = "Data taken from D_items table within MIMIC-III",
x = "Item Category",
y = "Count of Items",
caption = "Source: MIMIC-III Clinical Database v1.4"
)
This visualization helps to graphically explain how many items have been used that are categorized as “chemistry” or “hematology.” Most would think these are pretty much the same thing as most tests completed in a hospital involved blood (hematology) that is then put through a machine to analyze the chemistry of that blood (i.e. how much sodium is in a patient’s blood at a specific time) Being able to look at this data could help the healthcare system better manage funds that revolve around lab work, blood collection, etc. For example, if a hospital has a contract with a certain producer of blood sample tubes that is about to expire the amount of all the tubes may be important in their negotiations of the contract with the current and/or other companies.
Visualization Two - Defensive
SELECT subject_id, language, marital_status
From ADMISSIONS
WHERE marital_status LIKE "widowed%"
AND language LIKE "RUS%"ggplot(data = myquery2,
aes(x = language)) +
geom_bar()
Often times patients enter the United States healthcare system by happenstance. People from all over the world come to be a tourist in the US but unfortunately some of them become patients. For example, the data in this data portrays how many widowed, Russian speaking men were admitted to the health system. This information falls under the patient information protection act, HIPAA. Thus, this data has to have an offensive strategy in place to ensure it is not shared or breached.
Recommendations for Industry
Overall, healthcare leaders need to take a step back and realize just how important data they possess is to the success of the organization and safety of patients. Most healthcare leaders make decisions based off of data that is produced by someone else. There is no “fact-checking” they run on the information before they make a decision pertaining to the healthcare system. Healthcare systems do a great job on defending their data but it is time they start to mine that data and use it offensively. (DalleMule and Davenport (2017)) Healthcare has become a business and every business has to have a strategy.
References
Balaraman, Premkumar, and Kalpana Kosalram. 2013. “E-Hospital Management & Hospital Information Systems-Changing Trends.” International Journal of Information Engineering and Electronic Business 5 (1): 50.
DalleMule, Leandro, and Thomas H Davenport. 2017. “What’s Your Data Strategy.” Harvard Business Review 95 (3): 112–21.
Hentel, Keith, Andrew Menard, and Ramin Khorasani. 2017. “New CMS Clinical Decision Support Regulations: A Potential Opportunity with Major Challenges.” Radiology 283 (1): 10–13.
Huer, Jonathan. 2018. “A Brief History of the CIO.” EDUCAUSE Review (Online).