Processing...
Δ
Healthcare
Data Engineering
Python
The company provides multiple software solutions for the healthcare sector. One of their projects is a platform that offers enhanced benchmarking through executive-level summaries, detailed quality and operational views, and system-level and facility comparisons, accessible by stakeholders across the organization.
The NIX team uses data from different applications, defines the key parameters, and classifies hospitals by size, internship availability, and other data. There are three extensive studies per year that study cardio hospital, inpatient medical facilities, and systems databases. Each study consists of two databases – one processes data for one year and the other for five.
The one-year database determines the winners of the current year. For each hospital, we calculate the final score, which shows the hospital’s rating. The hospitals with the best performance results are sorted into top 15, top 50, and top 100 ratings. After study databases are created, results are ready and tested. PDF reports are then created with hospital performance results based on the gathered data. Any hospital participating in a study can request a custom report with additional settings for its metrics. We then create separate current and trend databases based on the originals to generate results for the PDF report.
Setting up the Python framework for each study considering updates in logic or metrics and making sure all data in reports are completely checked
Providing quality assurance services for huge database tables
The NIX team collects data from public datasets by different metrics: extensive studies for one and five years, financial data, patient’s experience, mortality, and complications, time spent in the hospital, profit and expense indicators, and much more. We create some source databases each quarter, others once a year, and then use them for three studies per year.
Hospitals’ data for chosen metrics are collected from prepared source databases. Data is filtered and insufficient or invalid information is discarded. The team distributes hospitals by class, and their data are processed and compared, corresponding to the hospital’s class.
When we get KPIs, the NIX team applies a custom methodology to determine hospital ratings within each class. Each class of hospitals will receive a definite number of winners according to winner counts by a course set in requirements for the current year. All yearly changes for measures or logic of data processing are implemented in Shell scripts, which create databases and tables.
Every database has a development and quality assurance areas to check if changes to the data or logic were properly and fully evaluated.
Databases are created on the Informix platform and main study databases are roughly:
When everything is checked and development databases contain valid processed data, the team generates PDF reports with performance results on each hospital from a study based on Informix data mart. Reports include graphs, metrics, and performance comparisons with winner and non-winner hospitals’ median values by each measure within the corresponding hospital class. While processing and analyzing hospitals filtered metrics for five years in SAS, we use linear regression to find the best-fitting straight line that helps determine if hospital performance is improving, worsening, or remains the same over time by each metric.
Output files for business analysts are created with extracted database information about metrics and hospital data along with tested output files.
Each report can contain up to 40 pages and one large study can have:
Python framework tests are performed if generated PDF reports for hospitals contain appropriate data by comparing reports and data from databases (logs with extracted hospital values).
Reports are uploaded to the platform and we use the Robot framework to check if all reports are fully uploaded and correctly displayed. Finally, the reports can be used by the client company’s management for future distribution.
The solution provides collected, filtered and processed data for receiving useful results which can be used by hospitals to analyze their performance and compare their metrics with winner and non-winner hospitals.
By analyzing the KPIs and rating reports they receive, hospitals understand where they can improve their performance and specific metrics, get information about their overall level of performance compared to other hospitals in their class, and, after comparative analysis, see highest and lowest values.
For example, if their hospital has several times more deaths from the disease than others in their class, they need to pay attention and review the protocols used to treat that disease.
3 Data Engineers, 2 QA Engineers, Shell Developer
Python, Unix, Shell, Robot Framework, SQL, IBM Informix
Data Analytics Platform for 3D Educational System
Publishing
Data Processing BI System for Healthcare
AWS-based BI Platform for Marketing Insights
Marketing & Advertising
Cloud Analytics for Retail Chain
Retail & eCommerce
Web Platform with Predictive Algorithms
Schedule Meeting
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.