HSC4501-Template for Correlation Analysis (1)

County Risk factor X Risk factor 1 = Alachua 10.8 49.5 % Slope Baker 7.9 37.7 % Intercept Bay 12.8 50.5 % Correlation coefficient Bradford 0.0 41.4 % Brevard 8.4 48.6 % R-squared Broward 8.0 54.7 % Sample size Calhoun 0.0 48.3 % T-statistics Charlotte 6.5 46.4 % Citrus 11.9 39.1 % Clay 12.5 42 % Collier 9.2 47.9 % Columbia 6.2 35.6 % Miami-Dade 10.0 52.7 % De Soto 10.1 36 % Dixie 22.1 38.1 % Duval 13.4 54.7 % Escambia 10.0 51.3 % Flagler 8.5 35.4 % Franklin 9.9 42.2 % Gadsden 16.3 53.5 % Gilchrist 12.6 40 % Glades 0.0 36.9 % Gulf 0.0 37.8 % Hamilton 21.5 43.4 % Hardee 0.0 46.9 % Hendry 7.3 47.9 % Hernando 8.7 37.3 % Highlands 14.1 40.5 % Hillsborough 10.3 50.3 % Holmes 0.0 49.6 % Indian River 11.5 42.1 % Jackson 0.0 46.8 % Jefferson 0.0 34.9 % Cervical Cancer age-adjusted rate, per 100,000 Example: Disease Y and Risk factor X 1. Replace cells in orange with your data from FLHealthCHARTS. Geographic Correlation Study - An association between average exposure level and d If p ≥ 0. correlati If p < 0. correlati magnitu If p= 00 to 0. 2. Change the names for disease frequency, risk factors IN RED according to your research. Then, they will be updated in the chart automatically. 3. These numbers should be form 'number' to allow computation o

Lafayette 0.0 40.4 % Lake 6.9 40.9 % Lee 11.4 46 % Leon 5.9 58 % Levy 14.1 39.3 % Liberty 29.2 45.5 % Madison 11.0 45.5 % Manatee 4.4 39.4 % Marion 13.0 47.1 % Martin 13.8 50.6 % Monroe 2.8 39.8 % Nassau 26.5 39.2 % Okaloosa 8.0 46.1 % Okeechobee 16.5 40.3 % Orange 10.7 51.3 % Osceola 11.4 51.5 % Palm Beach 7.2 59.6 % Pasco 11.8 47.7 % Pinellas 10.1 40.8 % Polk 13.2 55.1 % Putnam 17.8 34.1 % St. Johns 4.9 54.7 % St. Lucie 7.2 49.9 % Santa Rosa 6.9 46.4 % Sarasota 11.3 49 % Seminole 5.8 45.5 % Sumter 18.4 37.3 % Suwannee 11.4 44.7 % Taylor 22.9 44.2 % Union 13.3 41.7 % Volusia 8.1 43.3 % Wakulla 13.4 54 % Walton 6.7 43.2 % Washington 0.0 41.3 % 5 41.3

60 65

years Risk Factor X 1) Risk factor 1 = 2019 7.9 101.2 slope m= 2018 8.3 99.6 Intercept b= 2017 8.1 101.6 correlation coefficient r= 2016 7.7 100.9 2015 7.4 101.9 r-squared r2= 2014 7 101.4 sample size N= 2013 6.9 99.1 t-statistics t- 2012 6.7 99.1 p-value= 2011 6.6 95.1 2010 6.6 93.9 2009 5.9 94.8 2008 5.8 93 2007 5.9 91.1 2006 5.5 90 2005 5.8 87.1 2004 5.4 85.7 2003 5.8 80.5 2002 5.4 76.7 2001 5.9 76.6 2000 5.4 74.1 years Risk Factor x 2) Risk factor 2 = Risk Factor Disease Y mortality rate, per 100,000 Risk Factor X Disease Y mortality rate, per 100,000 Example 2) When the data points are available for only specific years. You have to include only those years in the analysis. Please make sure that the sample size matches with the number of years included Here are two examples for the time-series study 1) When the data points are available for all years. 2) When the data points are available for specific years. 1. Replace cells in orange with your data. 2. Don't forget to change the names for disease frequency, risk factors according to your research. It will Time Trends (or Time-Series) Study - Changes in disease rate in a single population Example 1) When the data points are available for all years If p ≥ 0.05, it indicate correlation is NOT sig If p < 0.05, it indicate correlation is signific for the magnitude of correlation.

0.084 -1.222 0.801 0.642 20 5.680 0.000 r x 40 50 60 70 80 90 100 0 1 2 3 4 5 6 7 8 9 f(x) = 0.083784249056576 x − 1.2223942355445 Chart Title Risk Factor X Disease Y mortality rate, per 100,000 r= p= Risk Factor x r= p= d. If not, please get assistance from Dr. Lee l be updated in the chart automatically. over time es that the gnificant. es that cant. Look f the Make sure that the names of axis reflect your study. 3. Update the correlation coefficient and p-value acco and Copy and paste this chart to Results Section 3.1

0.558 0.888 0.977 0.954 5.000 7.913 0.001 0 1 2 3 4 5 6 7 8 9 f(x) = 0.557834290401969 x + 0.888187038556195 Risk Factor x Disease Y mortality rate, per 100,000 p=

How to find the correlation coefficient and p-value Step 1: To find the mortality data 1. Go to www.FLHealthCHARTS.com >> HEALTH INDICATORS >> Chronic Diseases, and click the link for deaths on chosen disease 2. The bottom right chart displays the counties’ death rates. Choose age-adjusted rates. Export to Excel. 3. Use data from all counties if you want to compare by county (do not choose some of them, because you lose statistical power). Step 2: To find the data for a risk factor 1. Go to www.FLHealthCHARTS.com >> COUNTY PROFILES>> Healthiest weight >> Healthiest Weight Profile or FLHealthCHARTS.com >> HEALTH INDICATORS >> Behavioral Risk Fator Data tab to search available data for potentially modifiable risk factors. You can also search these factors from the search box in the right upper corner on the webpage. 2. Choose age-adjusted rates for counties. Export to Excel. 3. Please note that unlike disease statistics, these behavioral risk factors are not collected every year. For some risk factors, the most recent data is from 2016. You do not need to use the same year when the disease frequency was measured. Step 3: To find a correlation coefficient and p-value 1. Input exported data (mortality data and risk factor) into the Excel template/COUNTY tab. Make sure all data is in the right spot and lines up correctly by county name. 2. Once you input all data into the spreadsheet, the corresponding correlation coefficient and p- value should compute automatically. 3. However, the correlation coefficient and p-values will not appear automatically in the chart. You have to update them with the calculated values from your data. 4. Once you make sure that the slope and the intercept of the line in the chart match with the values you get from the automated calculation, it’ good to use in the paper. Copy and paste the chart to your paper and descibe them. *You can also examine time-trends over time for two factors . 1. Export data from the left side, data for years and paste into the Excel template/TIME tab. 2. As mentioned earlier, risk factors are not collected for all years. Just use as data available. But make sure you match up by year as you did for county's name. NOTES: Make sure you are using age-adjusted rates for comparison between counties or over time. Make sure that the order of data for the disease and the risk factor are the same. When you copy and paste data from extracted to the Excel template, paste them as 'number' not 'text' for numeric computation. These will fix some of the errors. RP3 quiz will provide a practice .