SELECT subject_id, charttime, value, valuenum, row_id
FROM labevents
LIMIT 5Analysis Report One - What’s Your Data Strategy
Executive Summary
Summarize your core themes, findings and recommendations into a concise, high-level overview tailored for busy healthcare executives.
Introduction
In “What’s Your Data Strategy?” by Leandro DalleMule and Thomas H. Davenport, the arguement that stood out to me is that organizations should view data as a strategic business asset rather than simply a technology resource DalleMule and Davenport (2017). The authors introduce a framework centered on balancing data defense and data offense DalleMule and Davenport (2017). Data defense focuses on protecting and governing data through security, privacy, compliance, and even quality management DalleMule and Davenport (2017).
All of this is happening while data offense is on the other side. Data offense emphasizes using data to generate business value through analytics, innovation, and improved decision-making DalleMule and Davenport (2017). DalleMule and Davenport state that organizations must determine the appropriate balance between these two priorities based on their industry, competitive environment, and other strategic objectives DalleMule and Davenport (2017).
One of the concepts that stood out to me within the article is the relationship between the SSOT (Single Source of Truth) and Multiple versions of the Truth (MVOTs). The SSOT serves more of a centralized, authoritative source of organizational data, which ensures consistency and reliability across the enterprise DalleMule and Davenport (2017). MVOTs allow different departments to transform and interpret the data according to their own specific needs, while still attempting to maintain alignment with the original source DalleMule and Davenport (2017). All together, these concepts can provide a framework that balances control and flexibility, which then enables organizations to maintain strong governance while supporting innovation, agility, and data-driven decision-making.
The Healthcare Context
Although “What’s Your Data Strategy?” by Leandro DalleMule and Thomas H. Davenport doesn’t cover a regular healthcare-focused front, the challenges they identify are extremely relevant to modern healthcare organizations. DalleMule and Davenport discuss the need to balance data defense and data offense DalleMule and Davenport (2017). Healthcare organizations face this exact challenge everyday. Hospitals must protect sensitive pation information and comply with other security and privacy regulations under HIPPA while also leveraging data to improve patient outcomes, reduce costs, and health initiatives that would improve majority of the general population in the US DalleMule and Davenport (2017).
DalleMule and Davenports overall concept of a SSOT and MVOT’s are particularly applicable because healthcare providers often struggle with fragmeneted patient records that spread across electronic health record systems, laboratories, billing platforms, and specialty clinics DalleMule and Davenport (2017).
I wanted to expand on this research, and according to the Office of the National Coordinator for Health Information Technology (ONC), many health organizations still struggle to exchange and integrate patient data effectively, despite hospital interoperability improving (faridoon2024healthcare?). This has created massive information silos that can negatively affect care coordination and decision-making (faridoon2024healthcare?). Additionally, three-quarters of hospitals report many barriers to electronic data exchange, including technical complexity and integration costs (faridoon2024healthcare?). To the current day, the Healthcare industry is the one of the most targeted industries for cyberattacks, accounting for 23% of major data breaches in 2024 and a little over 22% in 2025 (xu2025trends?). All of this research showed to me that the data strategy framework that was proposed by Dallemule and Davenport remains very relevant in healthcare, where organizations must continously balance protecting patient data with using it to drive clinical and operational improvements.
Data Visualizations
Offensive Visualization
Explanation:
The first thing I am explaining is that I do not understand how to fully use the GGPlot, as I tried watching the tutorial video for it, and it kept on freezing. However the visual I was trying to paint for the offensive visualization using the Labevents and the subject_id, charttime, and valuenum was to build a visualization of how the labvalue is changing over time across patient stays. With the one patient that had stayed and had worked with the lab of the facility they stayed at, they stayed for 20 hours and 21 minutes. Small sample size, but if more subjects are used in different tables with these same sub-groupings, it will be able to show the value of time that patients stay within a lab, and how to efficiently treat and move them to bring in new patients. A more subjected chart would help spot more abnormal trends.
ggplot(data = reportquery2, aes(x = charttime )) +
geom_bar()
dbListFields(mydb, "DIAGNOSES_ICD")[1] "row_id" "subject_id" "hadm_id" "seq_num" "icd9_code" SELECT subject_id, seq_num, icd9_code
FROM diagnoses_icd
LIMIT 10ggplot(data = reportquery3, aes(x =icd9_code, y = subject_id ))
geom_bar()geom_bar: just = 0.5, width = NULL, na.rm = FALSE, orientation = NA
stat_count: width = NULL, na.rm = FALSE, orientation = NA
position_stack Defensive Visualization Explanation:
The reasoning I wanted to use the diagnoses_icd table for the defensive visualization is because it is a core compliance and data-quality table that would in theory check completeness and consistency of the coding of private information within a facility. I know I am losing points for not accurately using the ggplot graphs (hopefully the coded tables above them will suffice), however, I wanted to show the importance of Diagnoses_icd and how it is a foundational reference data that other facilities when transferring information, and how they need the information from any ICD code to be as accurate as possible.
Recommendations for Industry
Based off of the research I have seen, and the information that I have gathered, I believe that the best reccomendations for the industry regarding an overall data strategy plan that will imporve efficiency, on the defensive end would be to use the DIAGNOSIS_ICD visualization (better than I did I hope) and show that the diagnostics coding has a lot of gaps. Finding these gaps, you can then invest in more standardized coding training, regular coding audits, and staff training on staff accuracy.
On the offensive side, using the LABEVENTS table like I did (hopefully once again better), using the table can show what becomes possible once data is reliable, and finding early identification of abnormal lab trends, and even predicting the flagging of certain risk patients, and then using the right resources to treat them and bring in new patients. This will help generate more efficient patient trips, and generate more revenue for the facility that uses it.
References
(article?){dallemule2017s, title={What’s your data strategy}, author={DalleMule, Leandro and Davenport, Thomas H}, journal={Harvard business review}, volume={95}, number={3}, pages={112–121}, year={2017} }
(inproceedings?){faridoon2024healthcare, title={Healthcare data governance, privacy, and security-a conceptual framework}, author={Faridoon, Amen and Kechadi, M Tahar}, booktitle={EAI International Conference on Body Area Networks}, pages={261–271}, year={2024}, organization={Springer} }
(inproceedings?){xu2025trends, title={Trends in US Healthcare Data Breaches}, author={Xu, Li}, booktitle={2025 IEEE International Conference on AI and Data Analytics (ICAD)}, pages={1–8}, year={2025}, organization={IEEE} ::: {#refs} :::
References
DalleMule, Leandro, and Thomas H Davenport. 2017. “What’s Your Data Strategy.” Harvard Business Review 95 (3): 112–21.