Presentation is loading. Please wait.

Presentation is loading. Please wait.

1 Deterministic Approach to Causality Dr. Shahram Yazdani.

Similar presentations


Presentation on theme: "1 Deterministic Approach to Causality Dr. Shahram Yazdani."— Presentation transcript:

1

2 1 Deterministic Approach to Causality Dr. Shahram Yazdani

3 2 Definition Causality refers to the way of knowing that one thing causes another.

4 3 Sufficient Cause Cluster: a Deterministic Approach A sufficient cause cluster which means a complete causal mechanism, can be defined as a set of minimal conditions and events that inevitably produce effect. Minimal implies that all of the conditions and events are necessary.

5 4 Necessary Cause A necessary cause can be defined as a conditions and events that without which the effect does not occur.

6 5 An effect with one sufficient cause cluster with two component cause A is a necessary cause B is a necessary cause A and B are a sufficient cause cluster A B

7 6 An effect with three sufficient cause cluster U is a necessary cause for the effect Three sufficient cause cluster of a disease U BA U EA U EB

8 7 Sufficient cause cluster When causal components remain unknown, one may be inclined to assign an equal risk to all individuals whose status for some components is known and identical. Thus, men who are heavy smokers are said to have approximately a 10% lifetime risk of developing lung cancer. Some interpret this statement to mean that all men would be subject to a 10% probability of lung cancer if they were to become heavy smokers, as if the outcome, aside from smoking, were a matter of chance

9 8 Sufficient cause cluster We view the assignment of equal risks as reflecting nothing more than assigning to everyone within a specific category In the classic view, these risks are either one or zero, according to whether the individual will or will not get lung cancer.

10 9 Strength of Effect: Three sufficient causes of a disease U BA U EA U EB Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present

11 10 Strength of Effect: case to population ratio in population 1 Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present 2000 case from 4000 population

12 11 Strength of Effect: case to population ratio in population 2 Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present 2000 case from 4000 population

13 12 Strength of Effect: Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present Incidence of A in population 1 =50% Incidence of A in population 2 =50%

14 13 Strength of Effect: Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present Incidence of B in population 2 =50% Incidence of B in population 1 =50%

15 14 Strength of Effect: Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present Incidence of E in population 1 =50% Incidence of E in population 2 =50%

16 15 But we are not aware from all cause clusters and all components of each cause cluster U BA U EA U EB Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present Suppose that we are not aware from the cause component A

17 16 Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 *Suppose that U is always present Suppose that we are not aware from the cause component A 100% of people in both groups with B and E have disease 0% of people in both groups without B and E are healthy 10% of people in group 1 and 90% of people in group 2 with B but without E have disease 90% of people in group 1 and 10% of people in group 2 with E but without B have disease What do you infer about strength of association of B and E with disease if you do the study in population 1 What do you infer about strength of association of B and E with disease if you do the study in population 2

18 17 Strength of Effects Incidence proportions for combinations of Component causes B and E in population 1, assuming that component cause A is unmeasured B=1, E=1B=1, E=0B=0, E=1B=0, E=0 Cases10001009000 Total1000 proportion1.00.10.90 E is a much stronger determinant of incidence than B Because the condition in which E acts as a necessary and sufficient cause –the presence of A or B, but not both- is common (3600 out of 4000 population or 90%)

19 18 Strength of Effects Incidence proportions for combinations of Component causes B and E in population 2, assuming that component cause A is unmeasured B=1, E=1B=1, E=0B=0, E=1B=0, E=0 Cases10009001000 Total1000 proportion1.00.90.10 B is a much stronger determinant of incidence than E Because the condition in which B acts as a necessary and sufficient cause –the presence of A or E, but not both- is common (3600 out of 4000 population or 90%)

20 19 Although the members of these populations have exactly the same causal mechanisms operating within them, the relative strength of factors E and B are completely different in them

21 20 Causal Complement The necessary and sufficient condition for a factor to produce disease is called causal complement of the factor. The condition “A or B but not both” is the causal complement of E in previous example.

22 21 Causal Complement and Strength The strength of a factor’s effect on a population depends on the relative prevalence of its causal complement; and this strength is independent of the biologic mechanism of the component’s action

23 22 Causal complement for A U BA U EA U EB Exposure* Outcome Exposure Frequency ABE Population 1Population 2 1111100900 1101100900 1011 100 1000900100 0111900100 0100900100 0010 900 0000100900 Select Those Combinations that when turn the value of A from 0 to 1 the outcome value also turns from 0 to 1 Exclude Combinations Containing A Exclude Combinations Which Leads to Effect

24 23 In epidemiology, the strength of a factor’s effect is usually measured by the change in disease frequency produced by introducing the factor into population.

25 24 for any cause component, observed strength of effect is an epidemiologic concept and not a biologic one.

26 25 Despite U is a necessary cause component you don’t see any association between it and the effect, and the strength of effect for U in both populations is Zero

27 26 Interaction among causes Biologic interaction can be defined as the participation of two component causes in the same sufficient cause cluster. Such interaction is also known as causal co-action or joint action. The joint action of the two component causes does not have to be simultaneous action.

28 27 Interaction among causes: example Suppose a traumatic injury to the head leads to a permanent disturbance in equilibrium. Many years later, the faulty equilibrium may lead to a fall while walking on an icy path, causing a broken hip. These two component causes have interacted with one another, although their time of action is many years apart.

29 28 Observable interaction among causes The degree of observable interaction between two specific component causes depends on how many different sufficient cause clusters produce disease and the proportion of of cases that occurs through sufficient causes in which the two component causes both play some role.

30 29 Interaction among causes Suppose that G does not exist in a population. Consequently, no disease would occur from sufficient cause cluster 2, and factors B and F would act only through the distinct mechanisms represented by sufficient cause clusters 1 and 3. Thus, B and F would be biologically independent. A B CD E A B FG H A C FI J 1 23

31 30 Interaction among causes Now suppose G is present; then B and F would interact biologically through cause cluster 2. Furthermore, if C is completely absent, then cases will occur only when factors B and F act together in the mechanism represented by sufficient cause cluster 2. Thus, the extent or apparent strength of biologic interaction between two factors is dependent on the prevalence of other factors. A B CD E A B FG H A C FI J 1 23

32 31 Proportion of Disease due to specific causes Assuming that the three sufficient causes in the diagram are the only ones operating, what fraction of disease is caused by U ? All of it This is not to say that all disease is due to U alone or that a fraction of disease is due to U alone. No cause component acts alone; rather, these factors interact with their complementary factors to produce disease. U BA U EA U EB

33 32 Proportion of Disease due to specific causes A report written by the scientists at the NIH proposed that as much as 40% of cancer is attributable to occupational exposure; many scientists thought that this fraction was unacceptably high and argued against this claim.

34 33 Proportion of Disease due to specific causes There is a tendency to think that the sum of the fractions of disease attributable to each of the causes of the disease should be 100%. For example, Doll and Peto in their work Causes of Cancer created a table giving their estimates of the fraction of all cancers caused by various agents, the total for the fractions was nearly 100%

35 34 Proportion of Disease due to specific causes For any disease, the upper limit for the total of the fraction of disease attributable to all the component causes of all the causal mechanisms that produce it is not 100% but infinity. Only the fraction of disease attributable to a single component cause cannot exceed 100%.

36 35 Induction Period If in sufficient cause 1, the sequence of action of the causes is A, B, C, D, and E and we are studying the effect of B, we do not observe the occurrence of disease immediately after B acts. The interval between the action of B and the disease occurrence is the induction time for the effect of B. A B CD E A B FG H A C FI J 1 23

37 36 Induction Period Lengthy induction period in the cause – effect relation between exposure of a female fetus to Diethylstilbestrol (DES) and the subsequent development of adenocarcinoma of the vagina between ages 15 and 30. Some evidence suggests that hormonal action during adolescence may be part of the mechanism (Rothman, 1981)

38 37 Induction Period It is incorrect to characterize a disease itself as having an induction period. The induction time can be conceptualized only in relation to a specific component cause. Thus we say that the induction time relating DES to clear cell carcinoma of vagina is 15-30 years, but we cannot say that 15-30 years is the induction time for clear cell carcinoma in general.

39 38 Each component cause in any causal mechanism have its own induction time.

40 39 By definition, the induction time will always be zero for at least one component cause, the last to act.

41 40 Induction Period: initiator vs. promotor In carcinogenesis, the terms initiator and promotor have been used to refer to component causes of cancer that act early and late, respectively, in the causal mechanism. Cancer has been considered as a disease process with a long induction time. This characterization is a misconception, because any late-acting component in the causal process, such as a promotor, will have a short induction time.

42 41 Similarly there is no difference between: –Risk factors vs. etiologic factors –Condition and context vs. etiology

43 42 Latent Period Disease, once initiated, will not necessarily be apparent. The time interval between disease occurrence (completion of one of sufficient cause clusters) and detection has been termed latent period.

44 43 Reducing latent vs. induction period The latent period can be reduced by improved methods of disease detection. The induction period on the other hand cannot be reduced by early detection of disease, since disease occurrence marks the end of induction period.

45 44 Catalyst vs. Cause Some agents may have a causal action by shortening the induction time of other agents. Suppose that exposure to factor A leads to epilepsy after an interval of 10 years, on the average. Exposure to a drug B would shorten this interval to 2 years.

46 45 Catalyst vs. Cause Is B acting as a catalyst or as a cause of epilepsy? The answer is both a catalyst is a cause B causes the onset of the early epilepsy

47 46 Catalyst vs. Cause Without B epilepsy would not occur after 2 years. The time of occurrence is part of our definition of an event. Without B epilepsy occurs later only if the individual survives an additional 8 years, which is not certain.

48 47 Catalyst vs. Cause Not only does agent B determine when the epilepsy occurs, but it can also determine whether it occur at all. Any agent that acts as a catalyst of a causal mechanism, speeding up an induction period for other agents, is considered a cause.

49 48 Postponement vs. Prevention Any agent that postpones the onset of an event, drawing out the induction period for another agent, is preventive. Postponement = Prevention

50 49 Causal Criteria Mill (1862) (1) Strength (2) Consistency (3) Specificity (4) Temporality (5) Biologic gradient (6) Plausibility (7) Coherence (8) Experimental evidence (9) Analogy Mill suggested that the following aspects of an association be considered in attempting to distinguish causal from noncausal associations

51 50 Nine guidelines for judging whether an association is causal Hill (1965) Temporal relationship Strength of association Dose response relationship Replication of the findings Biologic plausibility Consideration of alternate explanations Cessation of exposure Specificity of the association Consistency with other knowledge

52 51 Deterministic Approach to Causality A Researcher’s Point of View

53 52 Observation about disease P Prevalence of disease P is 18% 61% of patients with disease P have factor A 21% of people with factor A have disease P Suppose that the association of A and P is causal Draw the causal map of disease P? A What kind of research do you propose for clarification of etiology of disease P ? A?A?

54 53 People With A Without A With disease P Without disease P With disease P Without disease P Difference

55 54 Observation about disease P (2) 100% of A+/P+ people have B 19% of A+/P- people have B 57% of A+/B+ people have P 100% of A-/P+ people have B 63% of A-/P- people have B Suppose that the association of B and P is causal Draw the causal map of disease P? What kind of research do you propose for clarification of etiology of disease P A BB

56 55 People A+/B+ A-/B+ With disease P Without disease P With disease P Without disease P Difference B-

57 56 Observation about disease P 63% of A+/B+/P+ people have C 0% of A+/B+/P- people have C 100% of A+/B+/C+ people have P 100% of A-/B+/P+ people have D 0% of A-/B+/P- people have D Suppose that the association of C and D with P is causal Draw the causal map of disease P? What is your inference if you know that all of the 37% of A+/B+/P+ people without C are D+? A BC D B A B 

58 57 A BC D B 39% 89% ABCDP prevalence 111115 111012 110114 110008 1011016 101007 100106 100004 011112 0110025 010115 010001 001103 001006 000101 000005 Prevalence of A: 52% Prevalence of B: 52% Prevalence of C: 66% Prevalence of D: 42% What is the perceived association between A and B If factor C is present in all people 73% 61% 100% 0% What is the perceived association between A and B If factor D is absent in all people 57% of A+/B+ people have P 100% of A+/B+ people have P 57% of A+/B+ people have P

59 58 Thank You ! Any Question ?


Download ppt "1 Deterministic Approach to Causality Dr. Shahram Yazdani."

Similar presentations


Ads by Google