causation

CAUSATION

Thursday, September 17, 15

¡ HUBUNGAN ANTARA DUA KATEGORI KEJADIAN, DIMANA PERUBAHAN SUATU KATEGORI DIIKUTI OLEH PERUBAHAN KATEGORI YANG LAIN

HUBUNGAN KAUSAL


¡ URUTAN WAKTU¡ KUATNYA HUBUNGAN¡ KONSISTENSI¡ BIOLOGIC/EPIDEMIOLOGIC PLAUSIBILITY¡ INSIDENS¡ PAPARAN¡ ELIMINASI¡ PREVENSI PLAUSIBILITY

HUBUNGAN KAUSAL


Cross-sectional Study

n A still photographn Describes particular groups of

people at single hypothetical point in time

n Describes the prevalence of health events at a particular point in time

SNAPSHOT STUDIES

Can not tell why any observed health differentials exist


Advantage of Cross-sectional Studies

¡ Quick, less expensive

¡ Determining associations between variables of interest


Disadvantages of cross-sectional studies

¡ Does not separate cause-effect relationships in the associations established

¡ not useful for rare health conditions¡ not suitable for acute recurrent cases¡ does not identify risk of the occurrence of

diseases¡ not useful for explosive epidemics or acute,

short duration illness


Cross Sectional

Outcome +

Outcome -

Expose + / -

Onset ofthe study TIMENo direction of inquiry

Design


F FEASIBLE

¡ Adequate number of subjects¡ Adequate technical expertise¡ Affordable in time and money¡ Manageable in scope


EXAMPLE : RESEARCH QUESTION

¡ Does exposure to products of combustion or particulates has association with cancers of the respiratory system ?

¡ Is an increased risk of bronchogenic Ca associated with exposure to asbestos


Penyakit

Ya Tidak

a b

c d

a + b

c + d

a + c b + d a + b + c + d

Exposure +

Exposure -

PR = a / a + bc / c + d 95% CI


case control

studyThursday, September 17, 15

CASE CONTROL STUDY

¡ Persons with a given disease (the cases) and persons without the given disease (the controls) were selected

¡ The proportions of cases and controls who have been exposed to possible risk factors are then determined and compared


¡ Best suited to the study of rare cases, the study of diseases for which medical care is usually sought, and diseases with relatively rapid onset .

¡ Weakness: biases in selecting cases and controls;

doubt as which one is the cause and which one is the disease


Retrospective case control study:

¡ cases consist of individuals who have been diagnosed with the disease in the past


CONDUCTING A CASE-CONTROL STUDY



¡ Stating the research question




¡ Clearly defining the disease under study and the exposures of interest





¡ Selecting the case






¡ Defining and selecting a control group







¡ Developing and testing the research instruments







¡ Developing and testing the research instruments

¡ Conducting field operations& Planning the analysis


SELECTION OF CASES

¡ Selected among persons seeking medical care for the disease under study

¡ Newly diagnosed¡ Make sure that the exposure at

least preceded the onset of symptoms

¡ Diagnosed criteria should be established


EXAMPLES: CASE-CONTROL STUDIES

¡ Is an increase risk of myocardial infarction in women caused by recent (past-month) oral contraception use

¡ Does exposure to excessive sunlight cause skin cancer

¡ Does hormone replacement treatment in women increase the risk of breast cancer

¡ Do people with prolonged severe hypertension have an increase risk to suffer early dementia


CASES and CONTROLS¡ Eligibility criteria for the selection * Inclusion criteria * Exclusion criteria

¡ Objective criteria for the diagnosis * standardized definition * combination of signs/symptoms & laboratory test results

¡ Sources of cases and controls : * hospital * community. Methods of selecting cases


SOURCES OF CASES

¡ Hospitals (all, or some in a community)

¡ Disease or tumor registries¡ Vital Statistics Bureau¡ Ambulatory care practices¡ Cases in a defined geographic area (through record linkage system)


DEFINING A CONTROL GROUP

¡ Individuals should be free of disease of interest

¡ Comparability can be partially accomplished in the analysis (statistical adjustment)


SOURCES OF CONTROLS

¡ Hospitals (persons admitted to the same hospitals but with a disease or condition different from that under study)

¡ Community (chosen according to a probability sampling procedure based on census tracts or random digit dialing)

¡ Friends, neighbours, relatives, or associates of cases


SELECTION OF CONTROLS

1) probability samples of the population from which the cases came

2) persons seeking medical care at the same institutions as the cases for conditions believed to be unrelated to the cases diagnosis

3) neighbours of the cases or friends, siblings, schoolmates and fellow workers of the cases


EXCLUSION CRITERIA FOR THE CONTROL SERIES

¡ Individuals who are identified by medical conditions or backgrounds that are known to be associated with the exposure under study

example: a study of the relationship of aspirin

consumption to acute myocardial infarction; patients chronic peptic ulcer should be excluded from the control series (aspirin).


MATCHING

¡ One way of controlling for confounding factors

¡ A term used to refer to a number of procedures for selecting a comparison group in such a way that it is comparable to the study group with respect to factors that are not of major interest to the investigator but that would confound the disease-exposure association if left uncontrolled.


DEVELOPING THE RESEARCH INSTRUMENT

¡ Construct a questionnaire¡ Start with the construction of a list

of pertinent variables, including the level and extent of information needed

¡ Then write the questions that will obtain the required information as it is likely to be used in the analysis


¡ A reliable questionnaire is one that collects information that is replicable

¡ Assessing reliability : a. re interviewing a sample of the respondents

and comparing the first and second results (by the same interviewer or different interviewers)

b. repeating questions in a slightly different form at different points in the interview in order to check for consistency of response

RELIABILITY


Tips

¡ Phrase the question in a manner that can be understood by the respondents

¡ Don’t make an ambiguous questions: “Do you have trouble with your muscles?”

¡ A way of handling sensitive or embarrassing topics: presenting cards that carry the possible responses; questions are introduced in a non-judgmental way

¡ Consider the advantage and disadvantage between self-administered questionnaire as compared with using a trained interviewer for your study


¡ A valid questionnaire measures “that which it is purported to measure”

¡ Three components: a. Relevance: does the questionnaire

obtain the information it was designed to seek

b. Completeness: was all desired relevant information obtained

c. Accuracy: can reliance be placed upon the responses to the questions


Case Control Studies

Onset ofthe study

timeDirection of inquiry

Design

Control

Case

Exposed

Unexposed

Exposed

Unexposed


Disease

Case Control

a b

c d

a + b

c + d

a + c b + d a + b + c + d

Exposure +

Exposure -

OR = a / cb / d

95% CI


COHORT STUDIES


COHORT STUDIES

¡ Start with a group of individuals apparently free of the disease(s) of interest; divide this cohort into those exposed to a possible risk factor and those not exposed, and then follow through time in order to determine the incidence rate (or mortality rate) among the exposed and unexposed


COHORT STUDIES


COHORT STUDIES¡ PROSPECTIVE : the investigator collects

information on the exposure status of the study subjects at the time the study begins, and identifies new cases of disease (deaths) from that time on.


COHORT STUDIES¡ PROSPECTIVE : the investigator collects

information on the exposure status of the study subjects at the time the study begins, and identifies new cases of disease (deaths) from that time on.

¡ RETROSPECTIVE: exposure status is esta-blished from information recorded at some time in the past, and disease incidence or mortality determined from then until the present.


Cohort

Outcome +

Outcome -

Outcome +

Outcome -

Exposed

Unexposed

Cohortselectedfor study

Onset ofthe study


Design


Historical Cohort

Outcome +

Outcome -

Outcome +

Outcome -

Exposed

Unexposed

Recordsselectedfor study

Onset ofthe study


Design


RETROSPECTIVE COHORT AND CASE-CONTROL STUDY



¡ Retrospective cohort study: study subjects are selected based on their exposure status (Good for exposures of extremely low prevalence)



¡ Retrospective cohort study: study subjects are selected based on their exposure status (Good for exposures of extremely low prevalence)

¡ Case-control study: study subjects are selected according to whether or not they have the disease ( Good for rare diseases)


COMPARISON GROUP


COMPARISON GROUP

¡ Internal comparison group :


COMPARISON GROUP

¡ Internal comparison group : if a population is sampled and study

subject are classified on exposure status as measured by the investigator, then the natural comparison group is an internal comparison group of those from the same sample who do not have the exposure


COMPARISON GROUP



¡ External comparison group:


COMPARISON GROUP



¡ External comparison group: from general population or a group with

other particular exposure


ANALYSIS


ANALYSIS

¡ Two by two table :


ANALYSIS

¡ Two by two table : relative risk; survival analysis


ANALYSIS


¡ Confounding factor:


ANALYSIS


¡ Confounding factor: ** Stratified table


ANALYSIS


¡ Confounding factor: ** Stratified table ** Logistic regression


Penyakit

Ya Tidak

a b

c d

a + b

c + d

a + c b + d a + b + c + d

Exposure +

Exposure -

RR = a / a + bc / c + d


Attributable Risk :

Memberi informasi tentang efek paparan

Contoh studi kohort: AR = a/a+b - c/c+d = 27/482 - 77/1908 = 0.01566

Insiden penyakit yg dapat dicegah bila paparan dihentikan: 1566/ 100,000


causation

Documents