1
Risk Study
Section for Clinical Epidemiology and Biostatistics
Definition
• What is “ Risk” ?
–
The probability of some untoward event
–
The likelihood that people who are exposed to certain factors (risk factors) will subsequently develop a particular disease
–
The proportion of unaffected individuals who, on
average, will contact the disease of interest over a
specified period of time.
Definition
• What is “Risk Factor”?
– Characteristics that are associated with an increased risk of becoming diseased
– Interesting factors which may be associated with unwanted of adverse outcome
• Risk factors
o Inherited: e.g. HLA‐B27 o Infection: e.g. HIV
o Drugs, toxin: e.g. Aspirin overdose o Social, environment: e.g. crowding
o Behavior: e.g. smoking, alcohol abuse, driving without seat belts
THE FIRST APPEARANCE OF THALIDOMIDE
'Thalidomide Babies'
3
Example of Questions
• Are older age pregnant women at increased risk of autistic child ?
• Dose high trans fat consumption lead to increase risk of breast CA?
• Dose Soy Milk increase the risk of developing peanut allergy in children?
• Dose New NSIADs increase risk of GI bleed and CVS than piroxicam?
• Dose BZP increase risk of fall and femoral fracture in old age patients ?
Risk Recognition
• Large risks associated with effects that occur rapidly after exposures are easy to recognize.
• But most morbidity and mortality is caused by chronic diseases, then the relationships
between exposure and disease are far less obvious.
• Impossible for individual clinicians to develop estimates of risk based on their own
experiences with patients (Ethical)
Risk Recognition
•
Long latency
e.g. radiation and CA thyroid
•
Frequent exposure to risk factors
e.g. junk food and coronary heart disease
•
Low incidence of disease: it is difficult to draw
conclusions about the risk in rare disease. e.g. heavy smoker and lung cancer (incidence < 2/1000)
•
Small risk: e.g. birth control pills and breast cancer.
•
Common disease: the risk factors are already known. It becomes difficult to find a new risk factors.
•
Multiple causes and effects
Use of risks
• Prediction the occurrence of disease
• Cause
eg. “marker” = the risk factor that is not cause of disease but increases probability of disease.
• Diagnosis
“Rule in / Rule out by using risk factors”
• Prevention
“Risk removal”
5
Study of Risk
CC‐EBM
Most often used to express the probability that a particular outcome will occur following a particular exposure is “ RISK STUDY”
Others Cross
Sectional Cohort
Case- Control
STUDIES OF RISK
• Cohorts
Cohort study begins with the exposure to cause or
risk and then follow up until the outcome occurs.
STUDIES OF RISK
• Case‐control study
:
Study runs backwards. Researchers enrolled case (having diseases) and controls (having nodisease) and then look back in time to ascertain each person’s exposure status.
Comparing risks
• Basic expression of risk is incidence
(a number of new cases of disease arising in a defined population during a given period of time)
• To compare risks, several measures of the
association between exposure and disease,
called measures of effect, are commonly used.
7
Comparing risks (cont)
Ie = incidence in exposed persons Io = incidence in non‐exposed persons P = prevalence of exposure to a risk factor IT= total incidence of disease in a population
Comparing risks (cont)
Example
• A cohort study of 100 women who had been accidentally exposed to pesticides was followed up overtime and their outcome compared to 100 women who had not been exposed to pesticides.
• 30 women who had been exposed suffered miscarriages.
Compared to10 women who had not been exposed.
– What is the incidence of miscarriage attributable to pesticide exposure?
– How many times more likely are exposed persons to become miscarriage, relative to non‐exposed persons?
– What is the incidence of miscarriage in a population, associated with the occurrence of a pesticide exposure?
– What fraction of miscarriage in a population is attributable to exposure to pesticide?
– To calculate & interpret the OR results ?
Answers
9
Answers
Answers
Answers
Answers
OR = Odds that diseased person has been exposed Odds that non‐diseased person has been exposed
= {(a/a+b) / (b/a+b)} / {(c/c+d)/(d/c+d)}
= (a/b) / (c/d)
= ad/bc
= (30 x 90) / (70 x 10)
= 2700/700
= 3 (95 % CI = 0.2 – 9.8) is not significant.
11
RISK AND PREVENTION
• If the exposure is prevention,
so that Ie (incidence among exposed) is less than Io (incidence among unexposed),
the attributable risk is meaningless.
The prevention fraction (PF) can be defined.
PF = Io – Ie Io
CAUSE?
• Does it really cause a disease?
• Is this exposure really causing disease or ill‐health in the population at risk?
• Marker is a risk factor that does not cause a disease.
How can we tell a cause of a disease?
Criteria for judgment of causal
• Temporal sequence Did exposure precede outcome?
• Strength of association How strong is the effect, measured as relative risk or odds ratio?
• Consistency of association Has effect been seen by others?
• Biological gradient (dose‐response relationship)
Does increased exposure result in more of the outcome?
• Specificity of association Does exposure lead only to outcome?
• Biological plausibility Does the association make sense?
• Coherence with existing knowledge
Is the association consistent with available evidence?
• Experimental evidence
Has a randomized controlled trial been done?
• Analogy Is the association similar to others?
From Hill AB. Principles of medical statistics
EBM : Critical Appraisal
• Are the results valid?
• In a cohort study, aside from the exposure of interest did the exposed and control groups start and finish with the same risk for the outcome?
– Were patients similar for prognostic factors that are known to be associated with the outcome (or did statistical adjustment level the playing field)?
– Were the circumstances and methods for detecting the outcome similar?
– Was the follow‐up sufficiently complete?
• In a case‐control study, did the case and control group have the same risk (chance) for being exposed in the past?
– Were case and controls similar with respect to the indication or circumstances that would lead to exposure?
– Were the circumstances and methods for determining exposure similar for case and controls?
Users’ Guides for an Article About Harm || Gordon Guyatt
13
EBM : Critical Appraisal
• What are the results?
– How strong in the association between exposure and out‐come?
– How precise is the estimated of the risk?
• How can I apply the results to patient care?
– Were the study patients similar to the patient in my practice?
– Was follow‐up sufficiently complete?
– Is the exposure similar to what might occur in my patient?
– What is the magnitude of the risk?
– Are there any benefits that are known to be associated with exposure?
Users’ Guides for an Article About Harm || Gordon Guyatt
Clinical Question in OPD
• DM patient(65 years old) in OPD, was recently lab of HbA1c =9 many times after had symptom of polyuria and high blood sugar.
• The doctor prescribe Pioglitazone oral to patient after failed treatment with Glipizide & metformin and clinical well and controlled DM condition with FBS 120 mg/dl and HbA1c =5
• Patient was read Newspapers and found association of CA bladder and New
drug(Pioglitazone that he treated and concerned
the ask you about this ?
Transform Clinical Question to PICO
• P : DM Patient
• I : Pioglitazone (Thiazolidinedione group)
• C : Other anti‐diabetic drugs (Others than thiazolidinedione group)
• O : CA bladder event
Searching
• PubMed
• SCOPUS
• EMBASE
15
Search strategy
•
("Diabetes Mellitus"[Mesh] AND ("Thiazolidinediones"[Mesh] OR
"pioglitazone"[Supplementary Concept])) AND
"Urinary Bladder Neoplasms"[Mesh]
Why select ?
• Recent update in 5 years
• Large cohort of UK
• Good Methodologic :
Nest case control study based on well designed cohort study
• BMJ High impact factor
(impact factor of 17.4 (June 2015)
• Relevant to our PICO
17
Are the results valid?
• In a case‐control study, did the case and control group have the same risk (chance) for being exposed in the past?
– Were case and controls similar with respect to the indication or circumstances that would lead to exposure?
Yes,
UK general practice research database It contains the complete primary care medical record of more than 10 million people enrolled in more than 600 general practices. The geographical distribution of the practices has been shown to be representative of the UK population
Cohort DB: Patient enrolled before had CA bladder, Time of enroll was tie of any types anti‐diabetic was started. Patient are equally chance to receive any type anti‐diabetic drug by primary doctor.
19
Are the results valid?
• Were the circumstances and methods for determining exposure similar for case and controls?
Yes,
Carefully diagnosis was done by the Read classification of UK. Anti‐diabetic Prescriptions was prescribed and recorded base on prescription Pricing Authority Dictionary.
Patients were classified into 5 groups to compare between Case : Control
1.Exclusive categories for use of thiazolidinediones:
2.Exclusive ever use of pioglitazone 3.Exclusive ever use of rosiglitazone
4.Ever use of both pioglitazone and rosiglitazone (mainly switchers from one drug to the other),
5.Control group: never use of any thiazolidinedione, served as the reference category in the analyses
Anti‐Diabetic Drugs vs. CA Bladder
RANGE OF VALIDITY
Not valid 25% 50% 75% Valid
• Case and controls similar with respect to the indication or circumstances that would lead to exposure
• Circumstances and methods for determining exposure similar for case and controls
What are the results?
– How strong in the association between exposure and out‐come?
The primary analyses were restricted to 376 cases, matched to 6699 controls (1: 20), with at least one year of follow‐up between cohort entry and index date to account for latency.
– Exclusive ever use of pioglitazone group was associated with an 83%
increased rate of bladder cancer (adjusted rate ratio 1.83, 95% CI 1.10 to 3.05). *
– This corresponded to an absolute adjusted rate difference of 74 per 100, 000 person years (95% CI 9 to 140) *
– This effect was not observed for exclusive ever use of rosiglitazone (adjusted rate ratio 1.14, 95% CI 0.78 to 1.68), the other thiazolidinedione available in the United Kingdom during the study period.
21 Thiazolidinediones and risk of bladder cancer among
cases of bladder cancer and matched controls*
What are the results?
• Researchers compared exclusive ever users of pioglitazone with exclusive ever users of rosiglitazone. The adjusted rate ratio was numerically increased but did not reach statistical significance (adjusted rate ratio 1.60, 95% CI 0.88 to 2.90)
• Found the evidence of adose‐response relation between pioglitazone use and the rate of bladder cancer with the highest rate observed in users of more than 24 months (adjusted rate ratio 1.99, 95% CI 1.14 to 3.45)
• Statistically significant association was observed in patients who received more than 28,000 mg (2.54, 95 %CI 1.05 to 6.14)
Pioglitazone cumulative duration of use and cumulative dosage and risk of bladder cancer among cases of bladder cancer and
matched controls
What are the results?
– How precise is the estimated of the risk?
Yes, Narrow ranges of 95 %CI
NNH is –vE show the control condition is more likely to harm than the treatment condition expose
non expose
23
RANGE OF RESULTS
Poor 25% 50% 75% Good
• Strong in the association between exposure and out‐come
• Good Precision
How can I apply the results to patient care?
– Were the study patients similar to the patient in my practice?
Yes, but differences of ethnics
(should be compare incidence of CA bladder between UK and Thai and ware for difference life styles, Socio‐
economic‐environment through genetic differences) Local made also are available in Thailand .
– Was follow‐up sufficiently complete?
Yes, but not clearly mention‐only declare 1 cases of control to exposed to Pioglitazone)
– Is the exposure similar to what might occur in my patient?
Yes
– What is the magnitude of the risk?
Pioglitazone group was associated with an 83% increased rate of bladder cancer (adjusted rate ratio 1.83, 95% CI 1.10 to 3.05) and Significant in dose‐response relationship.
– Are there any benefits that are known to be associated with exposure?
Yes, for awareness and regulation of use.
Should I Stop exposure or alternating exposure to keep high benefit and low risks ?
• Yes, We can stop this exposure by use other anti‐diabetics drugs. And wait for further studies or meta‐analysis to answers this question.
