Basic Concepts and Strategies in Epidemiology PDF

Basic Concepts and Strategies in Epidemiology Edric D. Estrella Principles of Epidemiology Diploma in International Health Program Faculty of Management and Development Studies University of the Philippines Open University  Lea...

Basic Concepts and Strategies in Epidemiology Edric D. Estrella Principles of Epidemiology Diploma in International Health Program Faculty of Management and Development Studies University of the Philippines Open University  Learning Objectives: At the end of the session, the learner should be able to: define what a cause is; discuss the concept of multiple causation of disease; discuss the factors affecting the development of disease; discuss the different ecologic models of disease causation; discuss the natural history of disease; discuss the process of epidemiologic inference; discuss the methods on forming hypotheses; characterize an ideal epidemiological hypothesis; and enumerate the basic epidemiologic study designs. Uses of Epidemiology (Bonita et al., 2002)  Epidemiologic Approach /Epidemiologic Reasoning How does the epidemiologist identify the cause or causes of a disease?  Cause A cause of a disease or injury is an event, condition, characteristic or a combination of these factors which plays an important role in producing the health outcome. It can be any of a large number of characteristic relating to time, place, person, or event.  Cause It must precede an outcome. Sufficient - when it inevitably produces or initiates an outcome Necessary - if an outcome cannot develop in its absence Epidemiologic Study Cycle Analysis of results may Descriptive suggest further studies – data descriptive aggregation and studies or new analysis hypotheses Analytic studies Model building to test and formulation hypotheses of hypotheses  Descriptive Epidemiology studying the characteristics of groups and of individuals determining whether an association exists between a factor (e.g., an environmental exposure) or a characteristic (e.g., an increased serum cholesterol) and the development of the disease of interest.  Analytical Epidemiology determining if the association existing between an exposure and a disease is causal ✓not all associations are causal; some are non- causal (spurious); ✓the stronger the association, the more likely it is to be causal; ✓the strength of association between an exposure and a disease is measured by relative risks, odds ratios, risk differences  Multiple Causation of Disease Ecology - the study of the relationship of organisms to each other and to all other aspects of the environment. Since disease arises within an ecological system, an ecological approach is necessary to explain the occurrence of disease.  Multiple Causation of Disease Disease cannot be attributed to the operation of any one factor. Multiple causation/ multifactorial etiology – the requirement that more than one factor be present for disease to develop Factors affecting disease development Environmental Host factors factors (Intrinsic) (Extrinsic)  Host Factors affects susceptibility to disease The state of the host at any given time is a result of the interactions of genetic endowment with the environment over the entire lifespan  Host Factors 1. Genetic factors – 100,000-fold increase in risk of retinoblastoma among children born with dominant gene compared to non-carriers – 1,000x greater risk of colon CA for persons with polyposis than those without predisposing gene (Knudson, 1977) – Persons with type A blood have an increased risk of gastric CA (Aird et al., 1953)  Host Factors 1. Genetic factors – Persons with type O blood are more likely to develop duodenal CA (Clarke et al., 1955) – Sickle cell trait is associated with decreased risk of falciparum malaria (Allison, 1954) – Patients with xeroderma pigmentosum have a genetically determined inability to repair damage induced by UV light thus at higher risk of developing multiple neoplasms in areas of skin exposed to sunlight. (Robbins et al., 1974)  Host Factors 2. Past environmental exposure Specific immunity – state of altered responsiveness to a specific substance acquired through immunization or natural infection Other type of exposure are chemical in nature like cigarette smoking, occupational exposures (e.g. asbestos), air or water pollution  Host Factors 3. Personality Personality variables partly influence the course of illness because of the tendency to seek medical care and to comply with medical advice. Type A personality (aggressive, competitive, ambitious, restless, sense of time urgency) is predictive of CHD while those with Type B have lower rates of CHD. (Rosenman and Friedman, 1970) Competitive drive and impatience are associated with the development of CHD (Matthews et al., 1977)  Host Factors 4. Social class membership – strongly reflects environmental influence Because developmental experiences and lifestyle are intimately tied to social class, many diseases show a differential frequency among persons of different social classes Social class gradient in cervical CA (rates inversely related to social class) is compatible with class- related behavior as age at 1st coitus and number of sexual partners (Doll and Peto, 1981)  Environmental Factors Influence exposure and sometimes indirectly affects susceptibility 1. Biological environment – includes infectious agent, reservoirs of infection, vector that transmit disease, plants and animals  Environmental Factors 2. Social environment – economic and political organization of the society, social customs, general level of receptivity to new ideas – High degree of integration is protective while social isolation and alienation are productive of disease. – Suicide (MacMahon and Pugh, 1965) and entry into psychiatric care (Stein and Susser, 1969) tend to cluster in period following bereavement.  Environmental Factors – Being widowed, especially for men, is associated with higher mortality rates (Helsing and Szklo, 1981) – Upheaval associated with geographic mobility with change in cultural milieu (move from rural to urban area) Studies in North Dakota (Syme et al., 1964) and North Carolina (Tyroler and Cassel, 1964) indicate increased rates of CHD in persons exposed to cultural discontinuities.  Environmental Factors 3. Physical environment – includes heat, light, air, water, radiation, gravity, atmospheric pressure, and chemicals  4 Ecologic Models The The Epidemiologic Epidemiologic Triangle Lever The Web of The Wheel Causation Epidemiologic Triangle/ Triad Host Environme Agent nt Epidemiologic Triangle/ Triad  Epidemiologic Lever AGENTS HUMAN HOST AT EQUILIBRIUM LEVER ENVIRONMENT  AGENTS HUMAN HOST Biologic, Nutrient/Chemical, The balance depends on age, Physical, and Mechanical race, sex, habits, customs, genetic factors, personal The balance is determined by defense mechanism. nature and characteristics of these agents in relation to host Change upset balance. and environment. AT EQUILIBRIUM LEVER H A A H ENVIRONMENT A H The aggregate of all H external conditions and A influence affecting the life and development of an organism, human behavior and society. FULCRUM The Wheel of Disease Causation The Web of Causation (MacMahon, 1970) Interrelations of Factors (Ecologic Models) These models depict the ways in which the different environment and host factors influence the occurrence of disease. Regardless of the model, it is important to realize that it is the balance of forces which determines an individual’s state of health at a given time is in a kind of dynamic equilibrium. Interrelations of Factors (Ecologic Models) A potentially harmful change in any of the components of the system may not lead to a detectable disease if the other parts of the system can compensate for the insult. If the existing balance is already precarious, disease may develop after even a small insult. e.g.: – Flying at a high altitude might precipitate thrombotic crisis in persons with sickle cell hemoglobin. – Exposure to organisms that usually cause no damage can be serious for a person with impaired immunologic defenses. Ecologic Concepts and Control of Disease Recognition of the multifactorial nature of disease causation has led to awareness that studies of disease etiology must encompass multiple risk factors and their interactions. – Epidemiologic studies of most chronic diseases are necessarily complex in design, statistical analysis and interpretation. Ecologic Concepts and Control of Disease Although identification of risk factors for a disease is highly desirable, full knowledge of etiologic mechanisms is not necessary for effective control measures. – relationship between diet and pellagra (Goldberger, 1914) – relationship between cigarette smoking and lung cancer (Doll and Peto, 1981) Causes of Tuberculosis Causes of Cholera  Natural History of the Disease process by which diseases occur and progress in the human host over time from onset to resolution, unaffected by treatment Natural History of the Disease (Valanis, 1999) Natural History of the Disease (Mausner and Kramer, 1985) Stages in the natural history of disease 1. Stage of Susceptibility – disease has not yet developed although the factors that favor its occurrence are present Risk factor - factors whose presence is associated with an increased probability that disease will develop later; may be immutable or susceptible to change Stages in the natural history of disease Even when there is strong statistical association between a risk factor and a disease, this does not mean that all individuals with the risk factor will necessarily develop the disease nor that absence of the risk factor will ensure absence of the disease. Adaptation – initial response of the cell or functional system Stages in the natural history of disease Incubation period (infectious) – time after exposure when the organism multiplies to sufficient numbers to produce a host reaction and clinical symptoms (hours to months) Induction/latency period (non-infectious) – time period from exposure to onset of symptoms (years to decades some are instantaneous) e.g.: leukemia in children exposed to radiation – 5 years lung CA from asbestos exposure – 40 years acute episodes of poisoning The end of the incubation/induction period is the point of disease detection, whether by screening or by appearance of clinical signs and symptoms. Stages in the natural history of disease 2. Stage of Presymptomatic disease/ early pathogenesis – individual has no symptoms indicating the presence of illness, adaptation was unsuccessful and pathogenic changes have begun detectable by sophisticated lab test thus, subclinical→ below the level of clinical horizon (imaginary line dividing the point where there are detectable signs and symptoms from that where there are not) Stages in the natural history of disease 3. Stage of Clinical disease – sufficient end-organ changes have occurred so that there are recognizable signs and symptoms Classification of disease is important: – for better management of cases; – for epidemiologic study --- grouping of diseases reduces variation; for evaluation of intervention and for internal comparisons – may be based on morphological (cancer TNM), functional (CVD), therapeutic (heart), etc., considerations Stages in the natural history of disease 3. Stage of Recovery, Disability or Death Disability - any temporary or long-term reduction of a person’s activity as a result of an acute or chronic condition (NHS, 1958)  Recall generation of Epidemiology hypothesis Person description of the Descriptive distribution of disease Place Epidemiology occurrence Analytical identification of disease determinants Time Epidemiology testing of hypothesis Process of Epidemiologic Inference Source: Friis, R.H., Epidemiology 101, JB Learning, 2010.  Theory/ Knowledge Make Inferences Synthesize and re Conceptual Formulate Hypotheses Hunches Conclusions Conceptual and Hypotheses Interpretations Make Inferences An Idealized Conceptualization of the re Operational Scientific Method Design Study Hypotheses (Source: Kleinbaum, Kupper, and Morgenstern, 1982) Empirical Operational Findings Hypotheses Content Categories of Research Analyze Data Collect Data Process or Observations/ Stages of Data Research  Hypothesis Conceptual Hypothesis Operational Hypothesis Coffee consumption is a risk Middle-age women who factor for CHD lived in a certain area in 1970 were more likely to have died during the next 10 years of ischemic heart disease if they reported an average consumption of two or more cups of coffee per day in a 1970 questionnaire than if they reported less consumption  Scientific Hypothesis a supposition, arrived at from observation or reflection, that leads to refutable predictions (Porta, 2014) any conjecture cast in a form that will allow it to be tested and, possibly, refuted (Porta, 2014)  Epidemiologic Hypothesis a testable statement of a putative relationship between an exposure and disease should be: – clear – testable or resolvable – state the relationship between exposure and disease – limited in scope – not inconsistent with known facts – supported by literature, theory, references Source: https://onlinecourses.science.psu.edu/stat507/node/27 (accessed on 10/12/15)  Forming Hypotheses A new and convincing hypothesis can be one of the most powerful forces influencing the direction of future research. The success or failure of this research undertaking depends on the soundness of the hypothesis. Methods on Formulating Hypothesis Difference Agreement Concomitant Variation Analogy  Method of Difference If the frequency of a disease is markedly different under two separate circumstances, and some factors can be identified in one circumstance that is absent in the other, this factor, or its absence, may be the cause of the disease.  Example To test the efficacy of a new drug molecule, scientist will use matched groups. One arm will receive the drug while the other will receive a placebo. The only difference between the two group is the treatment received. Thus, the difference in the results may be attributed to the treatment.  Method of Agreement If a factor is common to a number of different circumstances that are associated with the presence of the disease, this factor may be a cause of the disease  Example Seromarkers of Hepatitis B virus are found with increased prevalence among both homosexuals and prostitutes. – suggestive of sexual transmission of the virus  Method of Concomitant Variation It is not a matter of a factor being present or absent but of its being present in a greater or lesser degree – involves the search for a factor whose frequency or strength varies with the frequency of the disease – quantitative rather than dichotomous approach Example  Method of Analogy The distribution of a disease may be sufficiently similar to that of some other disease that has been more completely and successfully investigated as to suggest that certain causes may be common to both. – uses deductive reasoning (i.e., epidemiological principles already established are applied to other situations)  Example The geographic distribution of Burkitt’s lymphoma in Africa is similar to that of malaria and yellow fever. This led to the hypothesis that like these two disease, an insect vector may be involved in the etiology of this type of lymphoma. Ideal epidemiologic hypothesis should specify the following: 1. The target population to whom the hypothesis applies 2. The cause being considered (exposure) 3. The expected effect (disease/ outcome) 4. The dose-response relationship (amount of cause needed for the outcome to develop) 5. The time-response relationship (time period between exposure and development of the outcome)→ incubation period/latency period  A well-developed hypothesis describes each of the 5 elements with a high degree of specificity. Among adults without Dirty water causes any previous exposure diarrhea. to the disease or its vaccines, the ingestion of a million viable typhoid bacilli will result in an attack rate of typhoid fever of 50% within a 30-day period.  Consideration in the Formation of Hypotheses 1. New hypotheses are commonly formed by relating observations from several different fields. – Epidemiological findings are most profitably viewed in the light of clinical, pathological, and laboratory observation. – Example: Snow on cholera Semmelweiz on puerperal fever among women who delivered by medical students  Consideration in the Formation of Hypotheses 2. The stronger a statistical association, the more likely it is to suggest a causal hypothesis. – Strength refers to the degree to which a disease that is entirely absent in one circumstance and invariably present in another is approached. – Measured by relative risk (RR) – Example: smoking and lung CA (causal association)  Consideration in the Formation of Hypotheses 3. Observations of change in frequency of a disease over time have been very productive. – particularly those that occurred over short period of time – Example: thalidomide (1959-1961)  Consideration in the Formation of Hypotheses 4. An isolated or unusual case should receive particular attention in the formation of hypotheses. – Example: Hampstead widow  Consideration in the Formation of Hypotheses 5. Observations that appear to be in conflict or to create a paradox are particularly worthy of consideration. – Example: Goldberger on pellagra among inmates Selection of Hypothesis for Evaluation 1. The value of a hypothesis is inversely related to the number of acceptable alternatives. – The greater the number of separate associations that can be explained by the hypothesis, the fewer the number of acceptable alternatives. – The more closely two variables found to be associated with a disease are associated, the less their independent value is in the formation of a hypothesis. – Association with certain variables may be more stimulative of hypotheses than those with other variables because of the unique environmental circumstances associated with them. Selection of Hypothesis for Evaluation 2. It is useful to make a deliberate search for specific demographic information that may be relevant, particularly that which may appear to refute or at least contradict the hypothesis – Example: genetic predisposition to stomach CA among Japanese versus role of the environment Circumcision hypothesis on cervical CA Selection of Hypothesis for Evaluation 3. A hypothesis need not explain all existing observations. – Inconsistences may be due to: Multiple independent causes of a disease – Lung CA among those who never smoked Crudity of disease classification – Etiological differences among subentities of a disease may produce inconsistencies with a hypothesis that is valid for only one of them. Epidemiologic Approach (From the lecture slides of Dr. O.P. Saniel for Epidemiology 201) Nomenclature of Epidemiologic study designs Case report/series Ecologic/Correlational Descriptive studies Prevalence studies Epidemiologic Study Cross-sectional studies Designs Observational Case-control studies prospective Cohort studies Analytical retrospective Randomized Control Trial Interventional Community Trial Hierarchy of Evidence Systematic Review and Meta-analyses Experiments Cohort studies Case-Control studies Cross-sectional analytic studies Prevalence surveys Ecologic studies Case Series Case Reports  Dichotomizing Terminologies Descriptive versus Analytical Observational versus Interventional Cross-sectional versus Longitudinal Prospective versus Retrospective  Case report/ series describe the unusual or idiosyncratic experience in terms of person, place and time  Ecologic/ Correlational Studies unit of observation: population 2 ecologic variables are contrasted to examine their possible association – ecologic measure of exposure versus aggregate measure of disease 2 general types: – comparing disease frequencies among different groups during the same period of time – one group at different points in time Ecologic fallacy  Prevalence Surveys unit of observation: individual members of a defined population determine the prevalence of disease in a given population in a specified place and time Temporal ambiguity  Cross-sectional Studies unit of observation: individual members of a defined population comparison group both exposure status and disease status are measured in one point in time in the life of the study participants Source: http://blog.manuscriptedit.com/author/twarita-chakraborty/  Cohort Studies unit of observation: individual members of a cohort who are disease- free at the start of the study Longitudinal: follow up Inquiry is on the disease Prospective vs. retrospective Source: http://blog.manuscriptedit.com/author/twarita-chakraborty/  Case-control Studies unit of observation: individual members of a study base Longitudinal Inquiry is on the exposure Source: http://blog.manuscriptedit.com/author/twarita-chakraborty/  Experimental Studies Randomization/random allocation – on the average, control all other factors (both known and unknown) that may affect disease risk Quasi-experiment Ethical concerns http://www.pitt.edu/~super1/lecture/lec19101/index.htm Study Designs in relation to time  The decision to use a particular design strategy is based on: – features of the exposure and disease – current state of knowledge – logistic consideration (time and resources)  References: Aschengrau, A. and G. Seage, III, Essentials of Epidemiology in Public Health, 2nd ed., 2008. Bonita, R. et al., Basic Epidemiology, 2nd ed., 1993. Friis, R. and T. Seller, Epidemiology for Public Health Practice, 5th ed., 2014. Mausner, J. and S. Kramer, Epidemiology: An Introductory Text, 2nd ed., 1985. Valanis, B., Epidemiology in Health Care, 3rd ed., 1999. Lecture on Basic Concepts in Epidemiology of O.P. Saniel in Principles of Epidemiology. 2014.  References: MacMahon, B. and D. Trichopoulos, Epidemiology: Principles and Methods, 2nd ed., Little, Brown and Company, Boston, 1986. Hennekens, C.H. and J.E. Buring, Epidemiology in Medicines, Little, Brown and Company, Boston, 1987. Friis, R.H., Epidemiology 101 (Essential Public Health), JB Learning, Maryland, 2010. Kleinbaum, D.G., L.L. Kupper, and H. Morgenstern, Epidemiologic Research: Principles and Quantitative Methods, Lifetime Learning Publications, London, 1982. Szklo, M. and F. J. Nieto, Epidemiology: Beyond the Basics, 3rd ed., JB Learning, Maryland, 2014.