Impact of HbA1c test on diabetes treatment from 1999 to 2008

Jingyu Zhang
Posted on Jul 23, 2016

Diabetes mellitus is a chronic disorder associated with disturbances in carbohydrate, fat, and protein metabolism and characterized by hyperglycemia. It is one of the most prevalent diseases, affecting approximately 24 million individuals in the United States.The motivation of this study is to use data science to help focus treatment development on more effective forms of treatment. In this work, I use the Health Facts database (Cerner Corporation, Kansas City, MO), a national data warehouse that collects comprehensive clinical records across hospitals throughout the United States. [1]  The data is an extract representing 10 years (1999–2008) of clinical care at 130 hospitals and integrated delivery networks throughout the United States. The final data includes 69,051 patients’ distinct visits with 50 variables.  

Five factors are investigated for revealing the information of diabetic patients and the relation of HbA1c testing result and patients’ readmission ratio back to hospital.

1). The diabetic population distribution in age and race

The plot below shows that patient numbers dramatically increase with age, and reaches a peak in the range of 80-90 years old.
Within the diabetic population, the highest count is Caucasian, then African American, Hispanic and Asian respectively. Because the data is collected in the USA, the result strongly depends on the specific demographics of this country.  


Diabates distributions in age


2). Primary diagnosis of diabetic patients

In our dataset, each observation corresponds to a unique patient diagnosed with diabetes. However, not every patient had diabetes as a primary diagnosis.  In many cases other diseases were more dominant in terms of symptoms, and so even people with diabetes might have a different primary diagnosis.  The plot shows the breakdown by primary diagnosis of patients who also have diabetes.  Circulatory disease ranks first  followed by respiratory disease, neoplasms, and digestive conditions. Diabetes as a primary diagnosis ranks 5th.  Since different diseases receive different medications, I categorized the data by primary diagnosis.


A subset of the original 69,051 diabetics was chosen for further study.  The selection was based on including only those patients whose initial diagnosis was performed using the HBA1c test, as this is test is unique in being able to identify how well the disease is being controlled.  12,000 subjects, all diagnosed with HBA1c remained in this study for further investigation.


These 12,000 subjects were further broken out into three groups: those whose diabetes is  under control (“normal”), and two groups whose diabetes is not under control.  In one instance (“high without change”) the doctor did not change medication despite the high occurrence of diabetic symptoms, and in the third group (“high with change”) the doctor changed medication in response to the diabetic symptoms.


The objective of this study is to understand the relation of HbA1c test result and the medication numbers.


3). HbA1c impact on medication number

The distinct medication number with HbA1C test  is analyzed in box plot. With primary diagnosis of diabetes, the median medication numbers are lower  than that of circulatory by HbA1c test result, respectively. Among the top ranked primary diagnosis, such as circulatory, respiratory, neoplasms and digestive, patients with high HbA1C test result and medical changed received more distinct medications, compared with those with normal test result. Oppositely, those with high test result, but no medical change received less distinct medications.  This information combined with readmitted ratio back to hospital proves the treatment results according to HbA1C test.



4). HbA1c impact on readmitted ratio

One of the features to demonstrate an effective treatment is the readmission rate of patients back to the hospital within 30 days. This bar plots show the relation of HbA1c test and readmitted ratio by primary diagnoses.

readmission ratio for HbA1C

Circulatory patients overall have higher readmission rates among all the patients. The diabetic patients with high test results show reduced readmitted rates only after being treated.  The respiratory, injury and musculoskeletal patients with normal HbA1c test result show a relatively lower possibility of going back to the hospital, compared those with high test result. We assume the patients with high test results are more serious than those with normal results. For the patients with primary diagnosis of diabetes, it appears that treatment reduces the rate at which they return to the hospital.  It is the opposite for  circulatory, respiratory, digestive, diabetes, musculoskeletal and genitourinary diagnoses.  In these cases patients appear to do better if there is no change in their medication.  

For neoplasms and injury patients with high test result, the changed medication reflects a lower readmitted ratio than the previous medication.

5). Medical specialty on readmitted ratio

The readmitted ratio by medical specialty is plotted in the bar chart.  Patients with hematology/oncology, oncology,  and vascular surgery ranked among the highest readmitted rates.  This suggests that these diabetic patients are very likely to return to the hospital in short time. On the bottom of the chart, the patients with obstetrics/gynecology and otolaryngology issues show very low readmission rates.

medication VS readmitted ratio

6). Conclusion

This study looked at diabetic patients from 1999 to 2008. First, we found that the diabetic population is increasing with age.  The rates vary for different ethnic populations, from Caucasian (the highest) to  African American, Hispanic and then Asian, respectively. Second, we found that many patients with diabetes have other primary diagnoses, such as circulatory, respiratory, neoplasms, digestive,  injury, musculoskeletal, and genitourinary in sequence by population. Third, compared with normal test results, when HbA1c high test with medication change, the number of medications increased. However, for those high test result without medication change, the number of medications is decreased. Fourth, we looked at the HbA1c test impact readmitted ratio. We found strong evidence of when there is a need to change medication, based on the diagnosis.  Fifth, the diabetic patients  with different primary diagnoses show different readmission rates.


I believe that for the data from 1999 to 2008, people realize how important of HbA1c test is for the diabetic patient’s treatment. Further data analysis for diabetic patients from 2008 to present with large amount of HbA1c test is ongoing.




About Author

Jingyu Zhang

Jingyu Zhang

Jingyu Zhang has Ph. D degree in Electrical Engineering with emphasis on optics and electronics. She had been working as display scientist at Sharp Labs of America for four years. Now she is passionate about data science and...
View all posts by Jingyu Zhang >

Leave a Comment

township hack online November 23, 2016
Really excellent post, I surely adore this website, keep on it.
Maryjo November 9, 2016
This is a very great hints especially to those new to blogosphere, short and exact info... Thanks for sharing this one. A must read article.
Hba1c Diabetes Ranges Chart | Play GTA for free September 18, 2016
[…] Impact of HbA1c test on diabetes treatment from 1999 to 2008 – Diabetes mellitus is a chronic … readmitted ratio even though we are lack of HbA1c test results corresponding to each medical specialty. The readmitted ratio by medical specialty is plotted in the bar chart. I find the patients with hematology … […]

View Posts by Categories

Our Recent Popular Posts

View Posts by Tags

2019 airbnb Alex Baransky alumni Alumni Interview Alumni Reviews Alumni Spotlight alumni story Alumnus API Application artist aws beautiful soup Best Bootcamp Best Data Science 2019 Best Data Science Bootcamp Best Data Science Bootcamp 2020 Best Ranked Big Data Book Launch Book-Signing bootcamp Bootcamp Alumni Bootcamp Prep Bundles California Cancer Research capstone Career Career Day citibike clustering Coding Course Demo Course Report D3.js data Data Analyst data science Data Science Academy Data Science Bootcamp Data science jobs Data Science Reviews Data Scientist Data Scientist Jobs data visualization Deep Learning Demo Day Demo Lesson Discount dplyr employer networking feature engineering Finance Financial Data Science Flask gbm Get Hired ggplot2 googleVis Hadoop higgs boson Hiring hiring partner events Hiring Partners Industry Experts Instructor Blog Instructor Interview Job Job Placement Jobs Jon Krohn JP Morgan Chase Kaggle Kickstarter lasso regression Lead Data Scienctist Lead Data Scientist leaflet linear regression Logistic Regression machine learning Maps matplotlib Medical Research Meet the team meetup Networking neural network Neural networks New Courses nlp NYC NYC Data Science nyc data science academy NYC Open Data NYCDSA NYCDSA Alumni Online Online Bootcamp Open Data painter pandas Part-time Portfolio Development prediction Prework Programming PwC python python machine learning python scrapy python web scraping python webscraping Python Workshop R R language R Programming R Shiny r studio R Visualization R Workshop R-bloggers random forest Ranking recommendation recommendation system regression Remote remote data science bootcamp Scrapy scrapy visualization seaborn Selenium sentiment analysis Shiny Shiny Dashboard Spark Special Special Summer Sports statistics streaming Student Interview Student Showcase SVM Switchup Tableau team TensorFlow Testimonial tf-idf Top Data Science Bootcamp twitter visualization web scraping Weekend Course What to expect word cloud word2vec XGBoost yelp