Epidemiology 101: Ratios, Proportions, and Rates in Montana

1. Data Rocks! Epidemiology 101 Epidemiologists

2. Summary Measures • Ratios • Proportions • Rates

3. Ratio • Comparison of any two numbers • Calculated by dividing one quantity by the other • The numerator and denominator are separate and distinct quantities (neither is included in the other) Student Teacher 8 1 = 8

4. Ratio Example • What is the ratio of abortions to live births in Montana? • 1,800 induced abortions • 12,000 live births • Ratio = 1,800/12,000 = .15 • Montana’s abortion ratio is 150 abortions per 1,000 live births (or 15 abortions per 100 live births)

6. Proportion • Terminology • Proportion = Percent = Prevelance • Comparing two numbers • Numerator is part of the denominator • Styles • Percent: 10% • Decimal: .10 4 6 = .667 = 66.7%

7. Causes of Unintentional Injuries in Montana, 2010-2013 Other/ Undetermined 28%

8. Rates • A rate measures the number of events that occur in a defined population, with respect to some unit of time.

9. Crude Rates • The numerator is part of the denominator • Numerator: count of events in a time period • Denominator: population at risk during that same time period

10. Crude Rates • Asthma Emergency Department (ED) Visits in Montana in 2016 • Count: 2,100 cases • Population at Risk: 85,000 people with asthma • Constant: 10,000 • Rate = times 10,000 = 247 ED visits per 10,000 persons with asthma per year

11. Is the numerator included in the denominator? Proportions, Rates, and Ratios Yes No Is time included in the denominator? Yes No RATE RATIO PROPORTION

12. Applying Summary Measures to Disease • Incidence • Prevalence • Mortality

14. Incidence • Measures NEW cases of disease among a population at risk of disease over a period of time • Examples of diseases/events measured with incidence • Cancer • Influenza, pertussis, or other notifiable disease • Hospitalizations • Often reported as a NUMBER or a RATE

15. Prevalence • Measures existing cases of a disease at a particular point in time or over a period of time • Often reported as a PERCENTAGE • We use point prevalence • Ex: The prevalence of current adult smokers in Montana is 19.0%.

16. Mortality • Measures DEATH due to a particular cause among a population over a period of time. • Often reported as a NUMBER or a RATE.

17. Age-adjusted rates • Disease or death is often associated with age. • Age-adjusted rates are a way to make more fair comparisons between groups with different age distributions. • Use when: • Comparing geographic areas Ex: MT to U.S.; or MT to counties; or County A to County B • Comparing time periods Ex: 1990-1999 to 2000-2009

18. Why age-adjust?

19. Why age-adjust?

20. Why age-adjust?

21. Measures of Uncertainty • Statistics (from samples) estimate the true value of a parameter (from population). • There is unavoidable variability because of this called sampling variability. • Confidence intervals are most common in health data.

22. Confidence Intervals We can’t ask all one million Montanans about their health habits. So we sample. • BRFSS sample n=9,300 • 2015 – 24.6% obese (23.4 – 25.8) • If we look at all of the different samples of 9,300 people, and we produced an interval estimate for each sample, then 95% of those intervals would contain the true estimate.

23. Absolute obesity prevalence = 24.3% (If we were able to ask ALL Montanans) 24.3 23.8 (22.6 – 24.9) 24.6 (23.4-25.8)

24. Confidence Intervals • 2015 – 24.6% obese • Margin of error: 1.2% • 24.6 -1.2 = 23.4 24.6%+1.2= 25.8 • 95% confidence interval (23.4% – 25.8%)

25. Why do they matter? • Confidence intervals are one method to tell if two estimates are significantly different or not • Overlapping confidence intervals indicate that there is no significant difference in the two estimates • Non-overlapping confidence intervals indicate that there is a significant difference between the two estimates

26. Confidence Intervals • Are these significantly different? • 11.3% (10.0-13.7) of youth in MT use smokeless tobacco versus 7.3% (6.1-8.6) of youth nationwide • 88.2% (81.4-95.0) of mothers in Flathead county initiate breastfeeding versus 76.5% (70.3-82.7) of mothers in Dawson county. 11.3% 7.3% YES 8 80 6 75 85 10 12 90 95 14 NO 88.2% 76.5%

27. Confidence Intervals BIGGER ≠ BETTER To your confidence interval: Increase sample size decrease standard error confidence level (from 95% to 90%)

28. Activity Female smoker Female Female Male Male Male smoker

29. Data Rocks! Public Health Data Sources Part 1

30. Outline • Population Estimates • Vital Statistics • Hospital and Emergency Department Discharge Data • IBIS

31. National Center for Health Statistics (NCHS) Population Estimates

32. NCHS Population Estimates • Derived from Census or American Community Survey estimates • Provided by the National Center for Health Statistics • Bridged Race Population Estimates • Single year data for age and county level population estimates • No economic information • Available via Montana Indicator-Based Information System (IBIS)

33. NCHS Population Estimates • Age • Single Year though 85, 85+ • Race • 4 Categories: • White • Black • American Indian / Alaska Native • Asian / Pacific Islander • Hispanic Ancestry (Hispanic / Non-Hispanic) • County of residence • Sex • Single year (1990-2016)

34. Accessing Population Estimates • http://ibis.mt.gov/

36. NCHS Population Estimates • Cody Custis, ccustis@mt.gov • 406-444-6947 • http://dphhs.mt.gov/publichealth/epidemiology

37. Vital Statistics

38. Death Data • Data on all resident deaths and deaths that occurred in state • Cause of death • underlying vs. multiple causes • Location of residence • down to county • Location of death • hospital, home, nursing home, or other • Demographic information • Did tobacco use contribute to death • yes, no, probably, unknown • Manner of death

39. Birth Data • Data on all resident births and births that occurred in state • Parental characteristics • Location of birth • hospital, birthing center, home etc. • Prenatal care • Smoking/alcohol use during pregnancy • Method of delivery • Birth outcomes • weight, apgar score, abnormal conditions • Breastfeeding status at discharge

40. Accessing Vital Statistics • CDC Wonder • http://wonder.cdc.gov/ • National Vital Statistics • http://www.cdc.gov/nchs/nvss.htm • Montana IBIS • http://ibis.mt.gov/

41. Accessing Vital Statistics • http://ibis.mt.gov/

44. Vital Statistics • Montana Vital Statistics Office • Todd Koch, tkoch@mt.gov, 406-444-1756 • http://www.dphhs.mt.gov/publichealth/epidemiology/

45. Hospital and Emergency department discharge data

46. ED Encounters and Inpatient Admissions • Data on ED and hospital discharges from majority of hospitals in state • Voluntary system through MHA • Does not represent individual people and cannot be de-duplicated • A person may have been transferred to another hospital, which would be counted twice as two discharges

47. ED Encounters and Inpatient Admissions • Age • Sex • Length of stay • Date of admission • Primary diagnosis • Secondary diagnoses • Up to 8 total • Procedure codes • E-codes • Up to 3 • Total charges • County of residence

48. Accessing Hospital Discharge Data • http://ibis.mt.gov/