Discussion on Data/Statistical Analysis for the Bronx diabetes data
Virtual class: March 2022 Discussion on Data/Statistical Analysis for the Bronx diabetes data
The third assignment (submission- NO later than Sunday April 10th before 5pm Period.)
at least 2 charts and 2 tables of the five-year trend analyses.
In all charts and tables, you should include the field of “Year” plus at least 2 other fields.
detail explanations of your research findings on these charts and analyses.
Compare 2015 data (treat it as baseline) with the rest of 4 years. Describe the pattern of changes in these five-year trends.
Example: what happened of your findings? Why or how does it happened? When, and where etc. Be creative to describe your research findings from your data analyses.
The Final Examination Power Point and Presentation during class (no later than April 11st Monday Morning before noon, period):
A complete project paper including “at least” a. Introduction, b. Research Findings with at least 2 charts and tables from the third assignment, and c. Conclusion.
(assuming the first 2 assignments was the proposal of applying research grant)
Of curse, the more creative the better , you can add more paragraphs , such as your recommendations, etc., etc.
Put your project paper on the Power Point. You have 10 mins to present your project to the class.
Discussion on Statistical Analysis for the Bronx diabetic data *REQUIREMENTS:* A. Required: Five year-trends for the comparison purpose is required for all charts. You can treat 2015 data as the baseline, compare with the rest of the 4-year trends. B. You have to submit “at least” 2 analyses with 2 charts and 2 tables for your third assignment and your final examination: Power Point Presentation Suggestion: C. Run frequency tables of the fields (variables) you pick to see how clean are the data
Data analysis guidelines: You can pick whatever variables you think are significant or logical for this data analysis.
Example : Not a requirements
1. Discuss what dataset you want to focus on: Inpatient , Outpatient, Emergency Department or all three.
2. Report card for each facilities ; Comparing Facilities on ED use, or mortality rate, readmission rate?
3.compare the changing pattern; (any before and after intervention? Such as wellness program)
Discussion on Statistical Analysis for the Bronx diabetic data 4. Focus on all kind of demographic characteristics, gender, race, age group zip codes etc 5. If you have clinical background, you can compare different type of diabetics- Diagnosis descriptions, e.g. Type I, Type II or severity of the illness 6.Using the census data, NYCDOH or NYSDOH data to figure out the target population of diabetic in Bronx as denominator, using the Bronx RHIO data as the numerator. (Hint: to calculate the number of in-care or unmet need etc)
Discussion on Statistical Analysis for the Bronx diabetic data 7. Based on the meaningful use of “pay for performance” think about how to create a report card (Hint: by hospitals) 8.Think about the relationship between the number of encounter and the number of unique (unduplicated count of patient) . (Hint: level of utilization, frequency of utilization) 9.Duration of service (Hint: LOS, ALOS)
Discussion on Statistical Analysis for the Bronx diabetic data 10. Mortality rate (identify the denominator; target population, which hospital) 11.Comparing type of insurance, Medicaid, Medicare, Private 12.How often patient use ED, who and which hospital provided more ED services
Discussion on Statistical Analysis for the Bronx diabetic data 13. Comparing age GROUP, (how to break down age groups) by race, gender etc 14.Categorize type of services by demographic characteristics 15.comparing zip codes by type of services, race, languages 16.Does language and zip codes has any relationship
Discussion on Statistical Analysis for the Bronx diabetic data 17. diagnosis are (1) more significant, (2) in which zip codes, (3) affecting which demographic characteristics (4) utilization of ED (5) ALOS (Hint: Descending order of diagnosis 18. How to compare the baseline data with the trends of other years. 19. Set the baseline of ED utilization, compare the ED use with the rest of the years 20. City by ED utilization, mortality rate, type of insurance, diagnosis description, ALOS, languages etc.,
Graphical Representation of Data
11
Health IT Workforce Curriculum Version 3.0/Spring 2012
p <0.001
Look at this graphic. There clearly has been a statistically significant improvement before and after the change.
But what other information can you gather from this graph?
Was the improvement due to the change?
Is the improvement holding over time?
It is very difficult with this type of graphic representation of data to determine what happened overtime and to get a good understanding of what is happening to the system.
11
Run Charts
12
Health IT Workforce Curriculum Version 3.0/Spring 2012
These three scenarios reflect the data in the histogram. They all have a pre-change average of 70 and a post change average of 30. However, as you examine the display of the outcome over time you will realize they tell very different stories.
In the blue chart you can see that the outcome hovered around 70 before the change and although there is forty point range the outcomes after the change hovered around 30. The change seems to have produced an improvement in the outcome.
In the green chart there is a progressive decrease of the value of the outcome that started before the implementation of the change. Although there seems to have been an improvement, it’s not due to the change implemented.
Finally, in the maroon chart there is an improvement after the change, but it seems to be short lived since, after the March measure, the outcome seems to worsen again.
12
