1. © 2010 Jones and Bartlett Publishers, LLC

2. Design Strategies and Statistical Methods Used in Descriptive Epidemiology Chapter 4

3. © 2010 Jones and Bartlett Publishers, LLC  The program that directs the researcher along the path of systematically collecting, analyzing, and interpreting data.  It allows for descriptive assessment of events and for statistical inference concerning relationships between exposure and disease and defines the domain for generalizing the results. What is a study design?

4. © 2010 Jones and Bartlett Publishers, LLC  Means of organizing, summarizing, and describing epidemiologic data by person, place, and time  Descriptive statistics can take on various forms, including tables, graphs, and numerical summary measures  Application of statistical methods makes it possible to effectively describe the public health problem Objective: Define descriptive epidemiology

5. © 2010 Jones and Bartlett Publishers, LLC  Provides information about a disease or condition  Provides clues to identify a new disease or adverse health effect  Identifies the extent of the public health problem  Obtains a description of the public health problem that can be easily communicated  Identifies the population at greatest risk  Assists in planning and resource allocation  Identifies avenues for future research Why is descriptive epidemiology helpful?

6. © 2010 Jones and Bartlett Publishers, LLC  Four types of descriptive studies: 1. Ecologic studies 2. Case reports 3. Case series 4. Cross-sectional surveys Objective: Describe uses, strengths, and limitations of descriptive study designs

7. © 2010 Jones and Bartlett Publishers, LLC  Involves aggregated data on the population level  Ecologic fallacy Ecologic study

8. © 2010 Jones and Bartlett Publishers, LLC Correlation between eating 5 or more servings of fruits and vegetables per day and being overweight (in the United States and US territories, 2007)

9. © 2010 Jones and Bartlett Publishers, LLC  A case report involves a profile of a single individual  A case series involves a small group of patients with a similar diagnosis  Provide evidence for larger scale studies (hypothesis generating) Case reports and case series

10. © 2010 Jones and Bartlett Publishers, LLC  Conducted over a short period of time (usually a few days or weeks) and the unit of analysis is the individual  There is no follow-up period Cross-sectional survey (sometimes called prevalence survey)

11. © 2010 Jones and Bartlett Publishers, LLC  Can be used to study several associations at once  Can be conducted over a short period of time  Produce prevalence data  Biases due to observation (recall and interviewer bias) and loss-to-follow-up do not exist  Can provide evidence of the need for analytic epidemiologic study Cross-sectional survey Strengths

12. © 2010 Jones and Bartlett Publishers, LLC  Unable to establish sequence of events  Infeasible for studying rare conditions  Potentially influenced by response bias Cross-sectional study Weaknesses

13. © 2010 Jones and Bartlett Publishers, LLC  Cross-sectional surveys that are routinely conducted U.S. Census Behavior Risk Factor Surveillance System National Health Interview Survey National Hospital Discharge Survey Serial surveys

14. © 2010 Jones and Bartlett Publishers, LLC Percentages of overweight and obese adults in the United States between 1990 and 2007. From CDCs BRFSS annual survey.

15. © 2010 Jones and Bartlett Publishers, LLC ChinaChina: 1.32 billion (19.84%) IndiaIndia: 1.13 billion (16.96%) United StatesUnited States: 304.0 million (4.56%) IndonesiaIndonesia: 231.6 million (3.47%) BrazilBrazil: 186.5 million (2.8%) PakistanPakistan: 163 million (2.44%) BangladeshBangladesh: 158.6 million (2.38%) NigeriaNigeria: 148 million (2.22%) RussiaRussia: 142 million (2.13%) JapanJapan: 127.8 million (1.92%) MexicoMexico: 106.5 million (1.6%) PhilippinesPhilippines: 88.7 million (1.33%) VietnamVietnam: 87.4 million (1.31%) GermanyGermany: 82.2 million (1.23%) EthiopiaEthiopia: 77.1 million (1.16%) Approximately 4.3 billion people live in these 15 countries, representing roughly two-thirds of the world's population.

16. © 2010 Jones and Bartlett Publishers, LLC 7/1/2007 6,600,411,051 8/1/2007 6,606,949,106 6,538,055 9/1/2007 6,613,487,162 6,538,056 10/1/2007 6,619,814,313 6,327,151 11/1/2007 6,626,352,369 6,538,056 12/1/2007 6,632,679,520 6,327,151 1/1/2008 6,639,217,576 6,538,056 2/1/2008 6,645,755,632 6,538,056 3/1/2008 6,651,871,878 6,116,246 4/1/2008 6,658,409,934 6,538,056 5/1/2008 6,664,737,085 6,327,151 6/1/2008 6,671,275,141 6,538,056 7/1/2008 6,677,602,292 6,327,151 Monthly World Population

17. © 2010 Jones and Bartlett Publishers, LLC  Nominal data (dichotomous or binary)  Ordinal data  Discrete  Continuous Objective Define the four general types of data

18. © 2010 Jones and Bartlett Publishers, LLC DescriptionExamples NominalCategorical – unordered categories Two levels – dichotomous More than two levels – multichotomous Sex, disease (yes, no) Race, marital status, education status OrdinalCategorical – ordering informativePreference rating (e.g., agree, neutral, disagree) DiscreteQuantitative – IntegersNumber of cases ContinuousQuantitative – Values on a continuum Dose of ionizing radiation Types of data

19. © 2010 Jones and Bartlett Publishers, LLC  Ratios, proportions, and rates are commonly used measures for describing dichotomous data  The general formula for a ratio, proportion, or rate is x/y*10 n 10 n is called the rate base, with typical values of n = 0, 1,..., 5 Objective: Define ratio, proportion, and rate

20. © 2010 Jones and Bartlett Publishers, LLC  In a ratio the values of x and y are independent such that the values of x are not contained in y  The rate base for a ratio is typically 1 Ratio

21. © 2010 Jones and Bartlett Publishers, LLC  In a proportion, x is contained in y  A proportion is typically expressed as a percentage, such that the rate base is 100 Proportion

22. © 2010 Jones and Bartlett Publishers, LLC  A rate may be thought of as a proportion with the addition that it represents the number of health-related states or events in a population over a specified time period Rate

23. © 2010 Jones and Bartlett Publishers, LLC Rate equations

24. © 2010 Jones and Bartlett Publishers, LLC Rate equations

25. © 2010 Jones and Bartlett Publishers, LLC  Diseases or events that affect a larger proportion of the population than the conventional incidence rate. Cumulative incidence rate (attack rate)

26. © 2010 Jones and Bartlett Publishers, LLC  The crude rate of an outcome is calculated without any restrictions, such as by age or sex, on who is counted in the numerator or denominator  These rates are limited if we try to compare them between subgroups of the population or over time because of potential confounding influences, such as differences in the age-distribution between groups Objective: Distinguish between crude and age-adjusted rates

27. © 2010 Jones and Bartlett Publishers, LLC  In 2002, the crude mortality rate in Florida was 1,096 per 100,000 compared with 579 per 100,000 in Utah  The crude mortality rate ratio is 1.9, meaning the rates in Florida are 1.9 times (or 90%) higher than in Utah  However, the age distribution differs considerably between Florida and Utah. In Florida 6.3% of the population is under five years of age and 16.7% of the population is 65 years and older. Corresponding percentages in Utah are 9.8% and 8.5%. Example of the importance of age-adjustment

28. © 2010 Jones and Bartlett Publishers, LLC  Using the direct method of age- adjustment based on the 2000 US standard population yielded rates of 762 in Florida and 782 in Utah per 100,000  Thus, after adjusting for differences in the age distribution, the rate in Florida is 0.97 times that in Utah Example of the importance of age-adjustment (continued)

29. © 2010 Jones and Bartlett Publishers, LLC US crude and age-adjusted (to the 2000 US standard population) rates for all-cause mortality and all malignant cancers according to year. Data from the National Cancer Institute.

30. © 2010 Jones and Bartlett Publishers, LLC  Direct  Indirect Two methods for calculating age-adjusted rates

31. © 2010 Jones and Bartlett Publishers, LLC MaleFemale AgeCountsPopulationRate per 100,000 CountsPopulationRate per 100,000 <408759305609622992842921118632 40-4910,6289,537,477111184559487526195 50-5927,0997,624,126355344107919559434 60-6949,4224,535,8841090389974873955800 70-7964,0532,984,78721464453936427181223 80+56,6092,322,37724385765238612741493 Total2165705756561337620333758996218345 Age-specific and overall all cause malignant cancer incidence rates among males and females in 1999-2001

32. © 2010 Jones and Bartlett Publishers, LLC  Suppose that we want to know the rate for females assuming they had the same age- distribution as males. To do this we multiply the age-specific female cancer rates by the age- specific population values for males to get expected number of cases for females for each age group, assuming they had the same age distribution as males. These expected counts are then summed and divided by the total male population. Direct method

33. © 2010 Jones and Bartlett Publishers, LLC The age-adjusted rate ratio for males to females is now 1.28. This means that if females had the same age-distribution as males, malignant cancer incidence would be 28% higher for males than females, as opposed to 9% higher found using crude rates. The crude rate is 1.09 times (or 9%) higher for males than females The resulting malignant cancer rate for females age-adjusted to the male population is: Direct method

34. © 2010 Jones and Bartlett Publishers, LLC Population A Age (years)Population# deathsAttack Rate 15-1910002424/1000=.024 20-2440001616/4000=.004 25-296000121121/6000=.020 Total11000161161/11000=.0146 Population B Age (years)Population# deathsAttack Rate 15-195000120120/5000=.024 20-2420001010/2000=.005 25-295001010/500=.020 Total7500140140/7500=.0187 Example 2

35. © 2010 Jones and Bartlett Publishers, LLC Population A Age (years)PopulationAttack Rate Pop. BExpected 15-191000x.024=24 20-244000x.005=20 25-296000x.020=120 11000164 Age-adjusted rate: 164/11000=.0149 Crude rate ratio:.0146/.0187=.7822% lower in population A Adjusted rate ratio:.0146/.0149=.982% lower in population A Example 2 Continued

36. © 2010 Jones and Bartlett Publishers, LLC  In situations where age-specific rates are unstable because of small numbers or some are simply missing, age- adjustment is still possible using the indirect method Objective: Define the standardized morbidity (or mortality) ratio

37. © 2010 Jones and Bartlett Publishers, LLC Standardized morbidity (or mortality) ratio (SMR)

38. © 2010 Jones and Bartlett Publishers, LLC  SMR = 1 The health-related states or events observed were the same as expected from the age-specific rates in the standard population.  SMR > 1 More health-related states or events were observed than expected from the age- specific rates in the standard population.  SMR < 1 Less health-related states or events were observed than expected from the age- specific rates in the standard population. Interpretation of the SMR

39. © 2010 Jones and Bartlett Publishers, LLC  Suppose that some or all of the female age- specific counts are unavailable, but that the total count is available  Further suppose that the age-specific rates for males can be calculated  Now multiply the age-specific rates in the male (standard) population by the age-specific female population values to obtain the expected number of all malignant cancer cases per age- specific group (see following table) Example of SMR

40. © 2010 Jones and Bartlett Publishers, LLC AgeMale Rate per 100,000 Female Population Expected Counts <4029292111868471 40-49111948752610531 50-59355791955928114 60-691090487395553126 70-792146364271878173 80+2438386127494138 Total272554 Data for calculating the age- adjusted malignant cancer rate for females using the indirect method

41. © 2010 Jones and Bartlett Publishers, LLC This ratio indicates that fewer malignant cancer cases (about 25%) were observed in females than expected from the age-specific rates of males Sum the expected counts to obtain the total number of expected malignant cancers in the comparison population Example of SMR

42. © 2010 Jones and Bartlett Publishers, LLC Population A Age (years)Population# deathsAttack Rate 15-1910001212/1000=.012 20-2420002020/2000=.010 25-2930009191/3000=.030 Total6000123 Population B Age (years)Population# deaths 15-19400085 20-24250Not Available 25-29750Not Available Total500095 Example 2 – Indirect Method

43. © 2010 Jones and Bartlett Publishers, LLC Age (years)Population BAttack Rate AExpected Deaths 15-194000.01248 20-24250.0102.5 25-29750.03022.5 500073.0 SMR = Observed/Expected = 95/73 = 1.3 The ratio indicates 30% more deaths than expected, based on the age-specific rates of population A (standard population) Example 2 Continued

44. © 2010 Jones and Bartlett Publishers, LLC  Tables Line listing Frequency distribution  Graphs Bar chart, pie chart Histogram Epidemic curve Box plot Two-way (or bivariate) scatter plot Spot map Area map Line graph Objective: Be familiar with tables, graphs, and numerical methods for describing epidemiologic data

45. © 2010 Jones and Bartlett Publishers, LLC Breast cancer incidence rates white women in Utah (by LDS status) and SEER (without Utah) by year of diagnosis Poor (5-year RSR<30%) Medium (5-year RSR 30-80%) Good (5-year RSR>80%) Low Incidence <20StomachCervix NHL Melanomas-Skin Medium Incidence 20-100 Lung & Bronchus Colorectal High Incidence 100+Breast (female) Prostate Prevalence of Selected Invasive Cancer Sites According to Their Incidence and Relative Survival Rate Combinations

46. © 2010 Jones and Bartlett Publishers, LLC  For U.S. Whites, 2003-2005 Female breast cancer rate = 129.4 per 100,000 Male prostate cancer rate = 155.3 per 100,000 What cancer is more common? Breast cancer in women Prostate cancer in men

47. © 2010 Jones and Bartlett Publishers, LLC  Measures of central tendency Mean Median Mode  Measures of dispersion Range Inter-quartile range Variance Standard deviation Coefficient of variation Empirical rule Chebychev’s inequality Numerical methods

48. © 2010 Jones and Bartlett Publishers, LLC  For discrete and continuous variables Correlation coefficient (denoted by r) Coefficient of determination (denoted by r 2 ) Spearman’s rank correlation coefficient Slope coefficient based on regression analysis Slope coefficient based on multiple regression analysis  For nominal and ordinal variables Spearman’s rank correlation coefficient Slope coefficient based on logistic regression analysis Slope coefficient based on multiple logistic regression analysis Objective: Be familiar with measures for evaluating the strength of the association between variables

49. © 2010 Jones and Bartlett Publishers, LLC  Under analytic epidemiologic studies, the risk ratio (also called relative risk) and odds ratio are commonly used to measure association, as will be discussed in a later chapter Other measures of association