HE2801A Research Methods in Health Sciences Session 6 - October 25, 2024 PDF

Session 6 Design 1 October 25, 2024 HS2801A: Research Methods in Health Sciences Fall 2024 Dr. Afshin Vafaei Department of Health Studies  Today’s Class The idea of design How it fits the research process (Session 1) Case rep...

Session 6 Design 1 October 25, 2024 HS2801A: Research Methods in Health Sciences Fall 2024 Dr. Afshin Vafaei Department of Health Studies  Today’s Class The idea of design How it fits the research process (Session 1) Case report, case series Cross-sectional studies Ecological (correlational) studies In quantitative design sessions specific measures of events and associations, and most pertinent biases will be reviewed  Health Research Process Generat Researc e Collect Analyze Interpret Communic h Hypothe Data Data findings ate Question sis Identify Sample Publish, subjects and the Describe Causatio n present variable population data Exposure/ Bias, outcome Measure Measures of occurrence/ confoundin Report associatio variables associations g n Design Related to the design  Health Research Studies Epidemiologic Biological (population health & (basic medical) clinical) Studies Studies Physiological, Observational Experimental direct cell Test tube Studies Studies observation Descriptive Studies Analytic Studies Case Reports -Ecological Randomized Community Case Series -Cross-sectional Controlled Trial Ecological Case-Control Trial Cross-Sectional Cohort Two Types of Population/clinical Health Research Identify and count cases of 1.To monitor the public’s disease in Case Reports & health populations Series Descriptive Cross-sectional 2.To evaluate the success according to person place, Study of intervention programs Ecologic Study 3.To generate hypotheses and time, and about causes of disease conduct simple Can be also studies analytic Compare group & Experimental 1.To evaluate hypothesis Study Analytic about the causes of systematically  Clinical Trial disease determine: is  Community 2.To evaluate the success there an trials of intervention programs association? Case-Control Study Cohort Study Goals and Design Decisions In which people? How should you How should you define your define your exposure(s)? disease? Conduct health research study to determine the relationship between an exposure of interest and a disease/outcome of interest with validity and precision using a minimum amount of resources. How can you make your study valid? How can you Random error make your study (chance) efficient? Bias Time Confounding Money  Components of a Study 1. Population 2. Exposure 3. Outcome 4. Potential Confounders 5. Analysis 6. Communication of Findings  Population SOURCE POPULATION The population you are interested in knowing more about STUDY POPULATION The population you enroll in your study to represent the source population Generalizability: Assuming the association in a study reflects the truth, to whom does it apply? Internally Valid  Outcome Case Definition = Description of the event you are interested in studying; can be a disease, defect, injury, event, state  A man with cardiovascular disease defined by ICD codes  A woman with first occurrence of breast cancer identified in a cancer registry  An infant with a structural abnormality of the heart identified within the first week of life Goal is to minimize errors (Session 3 & 4)  Exposure Exposure = Determinant of interest upon which an outcome depends; can be constitutional, environmental, or behavioral  Genetic polymorphism, female sex, short stature  Fluoridated water, air pollution, passive cigarette smoke  Cigarette smoker, fitness enthusiast, high fiber diet Goal is to minimize errors  Potential Confounders Confounder = Extraneous risk factor for an outcome; can be constitutional, environmental, or behavioral Pesticide Breast Level Cancer Confounders may lead to distortion of true association between exposure and disease Age Goal is to identify confounders and control them in some way (at design or analysis).  Analysis Analysis = examination of your study data; estimation of measures of disease frequency and association Describe your study population (proportion or rates of the disease, exposures) Crude estimates (if no confounding) Adjusted estimates (if confounding) Standardization Stratified analysis Multivariate analysis  Communication of Findings Communication of Findings = Tell the appropriate persons or community what you found even if you found no association  Peer-reviewed journals  Conferences and meetings  Government reports  What about popular and social media? More in the last session  Hierarchy of Evidence Most health research Good for formulating research questions, theory generating Starting point More in the domain of biology There is no good or bad TYPE of design but poorly vs. properly conducted studies  Case Report and Case Series Case report: a report of a health issues in one patient  A disease, a disorder, or undergoing a procedure Case series: a report that describes a group of individuals who have the same health issue  Case Report and Case Series No theory, no research question, goal is not researching, only reporting  May generate ideas for research questions  The main scientific part: defining the case  Based on a list of the inclusion and exclusion criteria individuals will be classified as a Case (diagnosed)  Sign: an objective indication of disease that can be clinically observed, such as a rash, cough, fever, or elevated blood pressure  Symptom: a subjective indication of illness that is experienced by an individual but cannot be directly observed by others  Several coding systems  International Classification of Diseases (ICD)  Diagnostic and Statistical Manual of Mental Disorders (DSM)  Case Report and Case Series  Mostly in medical settings but data collection usually with standard questionnaires  Ethical issues should be respected  Privacy  Sometimes a group will be labeled unjustly because of the original case series  Few or no data analysis is required  Clinical discussion dominates the report  Health Research Studies Epidemiologic Biological (population health & (basic medical) clinical) Studies Studies Physiological, Observational Experimental direct cell Test tube Studies Studies observation Descriptive Studies Analytic Studies Case Reports -Ecological Randomized Community Case Series -Cross-sectional Controlled Trial Ecological Case-Control Trial Cross-Sectional Cohort  How to decide on study design? Types of Analytic Studies Observation Experimental Investigator al actively Investigator manipulates observes as which groups nature takes its receive agent course under study -Cross-sectional -Clinical Trial -Cohort Study -Community -Case-Control Trial Study  Example Impact of Statins on lowering cholesterol Observational Experimental Data Collect Create Exposure assessment Inquiry Random Assignment Outcome assessment Inquiry Inquiry Analysis of the Depends,.... Complex but more association straightforward  Cross-sectional (Prevalence) Studies In this type of study a group of people is examined at one point in time Prevalence: the percentage of members of a population who have a given health issue at the time of a study  Point prevalence: the proportion of a population with a particular characteristic at one point in time  Period prevalence: the proportion of a population with a particular characteristic during a defined time period, such as several weeks or several months  Point Prevalence of Arthritis in Postmenopausal Women (session 3) 3,276 postmenopausal women participated in a research study. 597 of these participants had arthritis. The point prevalence of arthritis in this sample is: _18.2___%. The data is a from a cross-sectional study, arthritis was measured at a point of time in all participants ,  Period Prevalence of COVID-19 in Sep. 2022, City A # with disease at specific time period Period PR = Total defined population at same period In city A (population 647,226) there were 255 case of COVID-19 in Sep. 2022 (from Sep 1st to 30th). The Period PR of COVID-19 in this city during Sep 2022 39 cases per 100,000 persons. was ____ 255 ÷ 647,226 X 100,000  Prevalence (rate) ratio relative risk in cross-sectional studies A comparison of the prevalence of a characteristic in two independent populations (or independent subpopulations) that is calculated by taking a ratio of their prevalence rates  Urban versus rural populations  Exposed versus unexposed  A measure of risk of the exposure, association between exposure and outcome Developed Disease Yes No Prevalence Exposed to risk factor a b ratio = Not exposed to risk factor c d [a / (a + b)] [c / (c + d)]  Cross-sectional Studies_Example Target Population: working Canadians Data source: Canadian Community Health Survey data  Contains a variety of health-related factors and outcomes, high quality data, robust sampling with accurate weighting Exposure: support at work Outcome(s): repetitive injuries, most serious injuries  ALL self-reported Is recall bias a huge problem here? Any concern about selection bias?  Results Out of the total population of 4,327,300; 298,800 reported a repetitive injury. Among them, 128,484 individuals enjoyed high levels of social support. 53% of those with no injury worked in high social support environments. 4,028,500*0.53 Outcome Yes No TOTAL Social High 128,484 2,135,10 2,263,58 suppor 5 9 t Low 170,316 1,893,39 2,063,71 5 1 TOTAL 298,800 4,028,50 4,327,30 Prevalence rate ratio=Prevalence in high/ Prevalence in low 0 0 Prevalence rate ratio=(128,484/ 2,263,589)/(170,316/2,063,711)=0.69  Limitations of Cross-sectional Studies Because a cross-sectional study has no time dimension, it cannot be used to assess causality An exposure can be said to be “associated” or “related” to a disease, but a cross-sectional study cannot show that an exposure caused a disease Repeated cross-sectional study: a series of cross-sectional studies that re- sample and re-survey representatives from the same source population at two or more different time points A repeated cross-sectional study design does not track the same individuals forward in time A new set of participants is sampled from the source population for each round of data collection Compare to cohort studies (next session)  Correlational (Ecological) Studies  Correlational Studies In these studies, the unit of analysis is the group, not the individual. The group, or ecological unit, represents an aggregate of individuals Terminology:  Correlational studies  Ecological studies  Aggregate studies **The same*** Example of Correlational Studies: Global, regional, and national burden of Alzheimer’s disease and other dementias, 1990–2016 The variables in correlational studies are usually aggregate (grouped) statistics such as the proportion of a population with a particular characteristic or the average value of the variable in a population Nichols et al. (2019).The Lancet Neurology, Volume 18, Issue 1, 88 - 106 Example #2: Mortality by class of street and house in Manchester, 1844 (from the work of F. Engels) Mortality rate per 100,000 p.a. Class of House 4500 1st 2nd 3rd 4000 3500 3000 2500 2000 1500 1000 500 0 1st 2nd 3rd Social Class Drawn from Krieger. (2001). IJE; 30(4): 899-900  Example 3: Correlational Studies Problem: Trauma surgeons have observed an increased rate of penetrating trauma during periods of increased unemployment. They decided to examine this relationship over a 10-year period in Sedgwich County, Kansas Exposure: Unemployment rate (%) for each year for the county was obtained from the Kansas Department of Human Resources. Outcome: Penetrating trauma (# cases) for each year was obtained from trauma registries at local hospitals. Divided by total population as determined from census. Results: Scatter Plot Represents a year P = 0.014 A positive significant correlation was found  Ecological Fallacy Fallacy: a mistaken belief, especially one based on unsound argument The incorrect assumption that individuals follow the trends observed in population-level data, not a methodological error or bias  The experience of individuals in a population may vary significantly from the population average YEARS INDIVIDUALS I n j u r y Unemployment The results of ecological studies should be interpreted with extra cautious  Next week (Nov. 1, 2024) Two other main types of observational studies Case-control studies Cohort studies Prospective Retrospective Study chapters 10, 11