Descriptive Epidemiology PDF

Learning Objectives Define the term descriptive epidemiology Person, Place, Time Compare types of descriptive epidemiologic studies. Discuss limitations of descriptive epidemiology study designs Describe epidemiologic study types Person, Place, and Time Descriptive Epidemiology Definiti...

Learning Objectives Define the term descriptive epidemiology Person, Place, Time Compare types of descriptive epidemiologic studies. Discuss limitations of descriptive epidemiology study designs Describe epidemiologic study types Person, Place, and Time Descriptive Epidemiology Definition and Aims Person, Place and Time Unequal distributions of health and disease in populations To determine why health conditions vary throughout populations, one must answer the following questions: What is the condition of interest? Who was affected? (Person) Where did the (health) event occur? (Place) When did the (health) event occur? (Time) Definition: Descriptive Epidemiology A descriptive epidemiologic study is one that is “...concerned with characterizing the amount and distribution of health and disease within a population.” The field of descriptive epidemiology classifies the occurrence of disease according to the following variables: Person (who is affected) Place (where the condition occurs) Time (when and over what time period the condition has occurred) Aims of Descriptive Epidemiologic Studies Aims of descriptive epidemiology Permit evaluation of trends in health and disease Provide a basis for planning, provision, and evaluation of health services Identify problems to be studied by analytic methods and suggest areas that may be fruitful for investigation (hypothesis generating) HIV prevalence | 2021 | Ages 13 years and older | All races/ethnicities | Both sexes | All transmission categories | United States Example of a descriptive epidemiology graphic from CDC Footnotes: Data for 2022 and 2023 are preliminary and based on data received by CDC as of September 2023. Inclusion of preliminary data in trend assessments is discouraged. See Notes on second slide for details. Prevalence data for 2021 are preliminary and based on death data received by CDC as of December 2022. Prevalence data prior to 2010 are based on residence at diagnosis; prevalence data from 2010 to present based on most recent known address. ^ Jurisdiction with incomplete reporting of deaths for most recent year. NA - Not Applicable. Person Variables: Examples Age Sex Gender Identity Race/ethnicity Socioeconomic status Marital status Nativity (place of origin) Migration Religion Age Perhaps the most important factor to consider when describing occurrence of disease or illness Age-specific disease rates usually show greater variation than rates defined by almost any other personal attribute. Examples of Age Associations The incidence of and mortality from chronic diseases increase with age. Some infections (e.g., mumps and chickenpox) occur more commonly during childhood. The leading cause of death among young adults is unintentional injuries. Maternal age is associated with rates of diabetes and related complications. Sex Epidemiologic studies have shown sex differences in a wide scope of health phenomena including morbidity and mortality. Examples: All-cause age-specific mortality rates are higher among males. Differences in cancer rates (e.g., cancers of the genital system). NIH-AARP Diet and Health Study 1995- 2011 Cancer 2022; https://doi.org/10.1002/c ncr.34390 Nativity Place of origin of the individual or his or her relatives Subdivisions include: Foreign-born Native-born Race/Ethnicity Categories Five major categories in Census 2010: White Black or African American American Indian and Alaska Native Asian Native Hawaiian and other Pacific Islander Census 2000 was the first to allow respondents to check a multiracial category. Census 2020 expanded these categories Proposal to expand even more for 2030: White, Hispanic or Latino, Black or African American, Asian, American Indian or Alaska Native, Middle Eastern or North African, Native Hawaiian or Pacific Islander Race/Ethnicity: Other Considerations Somewhat ambiguous classification. Tends to overlap with nativity. Many propose that Race is a social construct rather than a biological construct. Used to track various health outcomes in epidemiological study designs Race/Ethnicity (cont.) Examples of racial/ethnic differences in health characteristics: Lower frequency of asthma reported among Hispanic Americans. Non-Hispanic white and non-Hispanic black individuals less frequently report that they have no usual source of medical care than do Hispanic Americans. Difficulties in physical functioning are highest among adults classified as Native Hawaiian and Pacific Islander. Worse maternal outcomes among pregnant black mothers Lower cancer mortality among white individuals Socioeconomic Status (SES) Defined as a “descriptive term for a person’s position in society,…” Often formulated as a composite measure of the following dimensions: Income level Education level Type of occupation A single dimension of SES (e.g., poverty level) may be used. Socioeconomic Status (SES) (cont.) The social class gradient Strong, inverse association of SES with levels of morbidity and mortality. Those in the lowest SES positions are confronted with excesses of morbidity and mortality from numerous causes. Place Variables International National (within-country) Urban–rural differences Localized patterns of disease (state county, city) International World Health Organization (WHO) studies: Both infectious and chronic diseases show great variation from one country to another. Climate, cultural factors, national dietary habits, and access to health care affect disease occurrence. Variations in life expectancy across countries National (Within Country) Regional differences may affect the prevalence and incidence of disease. Factors include: Climate Latitude Environmental pollution Example: Regional variations in stroke or cancer or HIV prevalence HIV diagnoses | 2021 | Ages 13 years and older | All races/ethnicities | Both sexes | All transmission categories | US Map-State Level Alaska Hawaii District of Columbia Puerto American Gua N Mariana US Virgin Rico Samoa m Islands Islands Rate per 100,000 among selected population 0 7 - 65 68 - 233 253 - 501 528 - 899 909 - 4,399 Footnotes: Data for 2022 and 2023 are preliminary and based on data received by CDC as of September 2023. Inclusion of preliminary data in trend assessments is discouraged. See Notes on second slide for details. NA - Not Applicable. Urban–Rural Differences Urban and rural sections of the United States show variations in morbidity and mortality related to environmental and lifestyle issues. Urban example: Elevated occurrence of lead poisoning among children who live in older buildings Rural example: Pesticide exposure and farming injuries among agricultural workers Localized Patterns of Disease Associated with specific environmental conditions that may exist in a particular geographic area Examples: Lung cancer and radon gas Naturally occurring arsenic in water supply Presence of disease vectors: Dengue fever along the Texas– Mexico border Characteristics of Time Cyclic fluctuations Point epidemics Secular time trends Clustering Temporal Spatial Cyclic Fluctuations Periodic changes in the frequency of diseases and health conditions over time Examples: Birth rates Higher heart disease mortality in winter Influenza Unintentional injuries Meningococcal disease Related to changes in lifestyle of the host, seasonal climatic changes, and virulence of the infectious agent Cyclic Fluctuations Figure 4-27 Percentage of Rotavirus Tests with Positive Results, by Surveillance Week: Participating Laboratories, National Respiratory and Enteric Virus Surveillance System, United States, July 2000–June 2009* *A median of 67 laboratories (range: 62–72) contributed rotavirus testing data to NREVSS during July 2000–June 2009. Reproduced from Centers for Disease Control and Prevention. Reduction in rotavirus after vaccine introduction—United States, 2000– 2009. MMWR Morb Mortal Wkly Rep. 2009;58(41):1148. Point Epidemics “The response of a group of people circumscribed in place and time to a common source of infection, contamination, or other etiologic factor to which they were exposed almost simultaneously.” Examples: Foodborne illness Responses to toxic substances Infectious diseases Secular Time Trends Refer to gradual changes in the frequency of a disease over long time periods Example is the decline of heart disease mortality in the U.S. May reflect impact of public health programs, dietary improvements, better treatment, or unknown factors Clustering Case clustering: Refers to an unusual aggregation of health events grouped together in space or time Temporal Clustering Definition: The clustering (grouping) of cases in time. Examples: Postvaccination reactions Adverse reactions to vaccines Postpartum depression May occur after a woman gives birth Spatial Clustering Definition: Concentration of disease in a specific geographic area Spatial clustering may reflect shared environmental exposures correlated with adverse health outcomes Example of spatial clustering: Geographic clustering of cases of Hodgkin’s disease Descriptive Epidemiology: Study Types Types of Descriptive Epidemiologic Studies Case reports Case series Cross-sectional studies Ecological studies The Simplest Descriptive Studies CASE CASE REPORT SERIES Detailed report on Detailed report on a one patient; usually group of patients with new or unusual same symptom or symptom or problem; usually new problem or unusual Useful for: 1. Recognition and description of new diseases, new manifestations of old diseases 2. Detection of drug side effects 3. Provide insight into disease mechanisms 4. Provide information that may help develop hypotheses The Simplest Descriptive Studies CASE CASE REPORTS SERIES Detailed report on Detailed report on a one patient; usually group of patients with new or unusual same symptom or symptom or problem; usually new problem or unusual Main Limitation: No explicit comparison group CDC MMWR with 2 Cases of TB reported  1981 Case Series Pneumocystis Pneumonia –Los Angeles MMWR, CDC In the period October 1980 – May 1981, five young men, all active gay men, were treated for biopsy confirmed Pneumocystis carinii pneumonia at three different hospitals in LA. Two patients died. All five patients had laboratory confirmed previous or current cytomegalovirus infection….The patients did not know each other and had no known common contacts or knowledge of sexual partners who had had similar illnesses. Two reported having frequent sexual contacts with various male partners. All five reported using inhalant drugs, and one reported parenteral drug abuse. Editorial Note: Pneumocystis pneumonia in the US is almost exclusively limited to severely immunosuppressed patients. The occurrence of pneumocystosis in these five previously healthy individuals without apparent underlying immunodeficiency is unusual…. All of the above suggest the possibility of…immune dysfunction related to a common exposure that predisposes individuals to opportunistic infections…. Cross-Sectional Studies and Surveys Definition: A study or survey that examines the relationship between an exposure and disease at a single point in time. Takes a snapshot A type of prevalence study Measures exposure prevalence in relation to disease prevalence Many government surveys are cross- sectional. Example of a Cross Sectional Study The Behavioral Risk Factor Surveillance System (BRFSS), which conducts an ongoing survey of the health-related behaviors of U.S. residents. Cross-sectional Survey of Coronary Heart Disease (CHD) Among Farm Owners by Occupational No. with No. Physical Prevalence of CHD CHD Activity Examined at Time of Survey Not Physically 14 89 14/89 or Active 157/1,000 Physically 3 90 3/90 or Active 33/1,000 Prevalence Ratio: (157/1,000) / (33/1,000) = 4.75 Farm owners who are not physically active have 4.75 times the prevalence of CHD as farmer owners who are physically Source: McDonough active. et al., 1965 Limitations of Cross-Sectional Studies Problem inferring temporal sequence of exposure & outcome “Which came first?” ?? Physical Inactivity Coronary Heart Disease Physical Inactivity Cross-sectional studies are problematic when Coronary Heart Disease exposure is a changeable characteristic (e.g., smoking, drinking, physical activity). They are OK for: Immutable characteristics (e.g., genetic traits, blood type) Measures of long-term exposure (e.g., lead in bones) Limitations of Cross-Sectional Studies Problem with Using Prevalent Cases Remember: P ≈ IR * D Prevalence is not ideal for etiologic research because it combines incidence and duration. Limitations of Cross-Sectional Studies Prevalent cases detected on cross- sectional surveys tend to be cases of long duration. X-------------------------------------------------------------------------------------------D X----D X----------D X---------------------------------------------------D X-----D X=Case diagnosed X-------D X---------------------------------------D D= Case died X-------D X----------------------------------------------------------------------------------D X-----------------D X-----------------------------------------------------------------D X---D Time -------------------------------------------------------------------------------------------------------- Cross-sectional survey Who will be picked up? done here: Limitations of Cross-Sectional Studies Prevalent cases detected on cross- sectional surveys tend to be cases of long duration. X-------------------------------------------------------------------------------------------D X---------------------------------------------------D X---------------------------------------D X=Case X----------------------------------------------------------------------------------D diagnosed X-----------------------------------------------------------------D D= Case died Only 5 of 12 cases will be captured. These long duration cases may be a biased subset of all cases. Thus, measure of association from a cross-sectional study may be biased. Strengths of Cross-Sectional Studies Relatively quick and inexpensive Highly generalizable if based on general population (e.g., gov’t surveys) Problem with temporal inference can be avoided for inalterable, long-term, and historical exposures Ecological Studies Definition: A study that examines rates of disease in relation to a population-level factor Unit of observation is the group (e.g., country, state, neighborhood) rather than the individual Exposure and outcome data at group level  Underlying Question: Do smokers have higher death rates from coronary heart disease than non-smokers? Exposure = individual cigarette smoking habits Outcome = individual deaths due to coronary heart disease Ecological Study: Do US states with higher smoking rates have higher death rates from coronary heart disease than states with lower smoking rates? Exposure = per capita average cigarette sales in each state Outcome = coronary heart disease mortality “rates” in each state Ecological Studies Underlying Question: Do people who consume meat have higher rates of colon cancer than people who don’t consume meat? Exposure = individual meat-eating habits Outcome = individual occurrences of colon cancer Ecological Study: Do countries with higher meat consumption rates have higher colon cancer rates than countries with lower meat consumption rates? Exposure = per capita (average) meat consumption in each country Outcome = colon cancer incidence rates in each country Ecological Fallacy or Bias Exposure is measured as average for a population, not a person, so there is no real link between exposure and disease. But our interest is in individual-level effects. Strong associations were seen in the prior examples, but are we certain that people who smoked were the ones who died from coronary heart disease? Or that the people who ate more meat were the ones who developed colon cancer? The ecologic fallacy means that the group-level association may not transfer to the individual level. Example of Ecologic Fallacy Worldwide, richer cities have higher rates of coronary heart disease (CHD) than poorer cities. It would be incorrect to infer that richer individuals have higher rates of CHD than poorer individuals. In fact, in industrialized cities poorer people have higher CHD rates than richer ones. Other Limitations of Ecological Studies Usually can’t adjust for confounding factors --i.e., other factors that may account for the observed association Could the association between cigarette sales and heart disease be due to another factor? What factors might be confounders? Complex relationships can be masked. Strengths of Ecological Studies Inexpensive and fast, conducted on available data Good for early stage of knowledge Wider range of exposures (especially for international studies) than other types of studies May wish to study ecologic relationships  For example, the effect of motor cycle helmet laws on motorcycle fatalities in a state Summary of Descriptive Studies Case Report or Ecological Study Cross-sectional Case Series Study Definition Report on one or Study that “Snapshot” study more patients examines rates of examines same symptom or disease in relation exposure disease problem, usually to population-level relationship at new or unusual factors same/single time point (compares exposure prevalence vs. disease prevalence) Strengths Simple, fast, Simple, fast, Fast, inexpensive inexpensive, inexpensive, often provides valuable wide exposure insights range, may be interested in ecological exposures Limitations No explicit No info on Unclear temporal comparison group individuals, group relationship if level associations exposure is may not translate changeable, to individual level tendency to Epidemiologic Inferences from Descriptive Data Descriptive epidemiology and descriptive studies provide a basis for generating hypotheses. Epidemiologic inference tested in analytic research is initiated with descriptive observations.  Figure 5-4 Process of epidemiologic inference. Reprinted with permission from Aragón T. Descriptive epidemiology: Describing findings and generating hypotheses. Center for Infectious Disease Preparedness, UC Berkeley School of Public Health. Available at: http://www.idready.org/slides/feb_descriptive.pdf. Accessed August 16, 2008. Conclusion Descriptive epidemiology classifies the occurrence of disease according to the variables of person, place, and time. Descriptive epidemiologic studies aid in generating hypotheses that can be explored by analytic epidemiologic studies. Descriptive studies include case reports, case studies, and cross-sectional studies. Epidemiological Study Designs Rationale for the Choice of Study Designs How much is already known about a health issue? Use existing data if possible. As knowledge increases, use more complex designs. Rationale: Differences Among Study Designs Number of observations made Directionality of exposure Data collection methods Timing of data collection Unit of observation Availability of subjects Observational Versus Experimental Approaches in Epidemiology Manipulation of study factor Was exposure of interest controlled by investigator? Randomization of study subjects Was there use of a random process to determine exposure of study subjects? Typology of Epidemiologic Research Manipulation Randomization Study Type Yes Yes Experimental Yes No Quasi- Experimental No No Observational Typology of Epidemiologic Research Overview of Study Designs Used in Epidemiology Experimental studies Quasi-experimental studies Observational studies Experimental Studies In comparison with quasi-experimental and observational studies, they maintain the greatest control over the research setting. The investigator: Manipulates the study factor Randomly assigns subjects to the exposed and nonexposed groups An example of an experimental study is a clinical trial. Experimental Studies: Uses of Clinical Trials Test efficacy of new therapies Study surgical procedures Evaluate new drugs Study the effects of interventions to modify health status Quasi-Experimental Studies Include community trials (community interventions) Examples: interventions for smoking cessation, control of alcohol use, weight loss Involve manipulation of the study factor Do not include randomization of study subjects (to the study groups) Oriented toward education and behavior change Can be used to evaluate the extent to which programs meet public health goals Observational Studies Used when experimental designs are impractical or unethical Do not entail manipulation of a study factor Do not involve randomization of study subjects Include both descriptive studies and analytic studies Observational Studies (2 of 2) Descriptive studies: Case reports Case series Cross-sectional surveys Analytic studies Case-control studies Cohort studies The 2 by 2 Table (1 of 2) Tool for evaluating the association between exposure and disease status Columns: Represent disease status or outcome (yes or no) Rows: Represent exposure status (yes or no) The table cross-classifies exposure status and disease status. The 2 by 2 Table (2 of 2) Disease Status Yes No (People (People Total with without Disease) Disease) A B Yes (exposure (exposure A+B (total (exposure & disease present, number present) present) but no exposed) Exposure disease) Status C (no D (no C+D (total No (no exposure, disease, no number exposure) disease exposure) with no present) exposure) A+C (total B+D (total number number N (sample with without total) disease) disease)  Epidemiological Study Designs Studies of human subjects Observational studies Experiments Descriptive Randomized Analytic Clinical Trials Cross- Ecologic Case series Case-control Cohort sectional studies studies studies studies Prospective Retrospective 72