pharmacoepidemiology.ppt

pharmacoepidemiology and its implications , Special applications of pharmacoepidemiology, Molecular Pharmacoepidemiology

2. “The study of the use and effects of medications
in large numbers of people”
………Strom
“The application of epidemiologic knowledge,
methods, and reasoning to the study of the
effects (beneficial and adverse) and use of drugs
in human populations.”
……Porta and Hartzema

3. Clinical pharmacology
Pharmacoepidemiology
Epidemiology

4. Focus of pharmacoepidemiology and related areas
Discipline Focus Indicators of drug
exposure
Result studied
Clinical
pharmacology
Individual
patients
Clinical effect
Adverse reaction
Drug
effectiveness &
toxicity
Drug utilization Groups Utilization pattern
Appropriateness of
drug use
Excessive /
inadequate use
Quality of care
Drug safety
Pharmaco -
epidemiology
PopulationExposure outcome
relationship
Comparative
effectiveness
Comparative toxicity
Causality
Quantification of
benefit and
risk

5. Pharmacoepidemiology
Pharmacology
Epidemiology Statistics
Therapeutics

6. Health services
research
Health
economicsOutcomes
research
Pharmaco-
Epidemiology
Epidemiology
Economics
Conceptualization by Harry Guess

7. Drug utilization reviews:
“an authorized, structured and continuing program that reviews,
analyses and interprets patterns of drug use against predetermined
standard”.
Medical audit :
“A searching examination of the way in which drugs are used
in clinical practice carried out at interval frequent enough to
maintain a generally accepted standard of prescribing”
It focus on medical practitioners with the aim of improving the
rational use of medicines as therapeutic agent

8. Pharmacovigilance:
“The science and activities relating to the detection,
assessment, understanding and preventing adverse events or
any other drug related problems.”
Passive methods –Case reports and Case series.
Active methods –Analysis of secular trends
Case control & Cohort studies
Randomized control trials
Meta analysis
Prescription event monitoring
Record linkage system.

9. Origin and evolution of
pharmacoepidemiology

10. 1906-Pure, food & drug act-excessive adulteration and
misbranding of the food and drug.
1937-over 100 deaths due to renal failure (sulfanilimide was
dissolved in diethylene glycol).
1938-preclinical toxicity testing, clinical data about drug
safety was required by FDA before drug marketing.
1950s-Chloramphenicol caused aplastic anemia.
1961-“thalidomide disaster”

11. 1962-Kefauver-Harris amendments were passed .
(extensive preclinical pharmacological & toxicological testing)
Data to be submitted to FDA in Investigational new drug
application (IND) before clinical studies could began.
1960’s–Publication of series of drug utilization studies –
marked the beginning of pharmacoepidemiology.
1970-Diethylstilbestrole –cervical, vaginal cancer
Practolol –oculomucocutaneous syndrome.
1977-Computerized Online Medicaid analysis and
surveillance system to perform pharmacoepidemiological
studies.

12. 1980s-Isotretinion (congenital defects), Triazolam ( CNS
disturbances ), fluoxetine (suicidal ideation), ticrynafen (death)
and many more
1990’s–Inclusion of pharmacoepidemiological studies to
study beneficial drug effect, application of health economics to
study drug effects, quality of life studies, meta analysis etc
1996-The International Society for Pharmacoepidemiology
(ISPE).
Drug utilization studies developed in parallel with
pharmacovigilance (adverse events studies)

13. Animal studies
Phase 1
Phase 2
Phase 3
Drug approval
Phase 4
Human subjects
•Human subjects
•Not always required

14. Limitations of premarketing surveillance :
Carefully selected subjects may not reflect real-life patients in
whom drug will be used
Study subjects may receive better care than real-life patients
Short duration of treatment
Study size
Studies with 3000 patients cannot reliably detect adverse
events with an incidence of < 1 per 1000, even if severe
Studies with 500 patients cannot reliably detect adverse
events with an incidence of < 1 per 166, even if severe

15. Potential contributions of
pharmacoepidemiology

16. Supplements the information available from premarketing
surveillance:
Higher precision –dose
In patients not studied prior to marketing e.g. the elderly,
children, pregnant women.
Factors such as illness & other drug that can modify the
effects of drug
Its safety and efficacy as compared to the other drug used
for similar indication

17. New type of information not available from
premarketing surveillance :
Discovery of previously undetected adverse and beneficial
effect
Uncommon effects-agranulocytopenia
Delayed effects –vaginal & cervical Ca occurred two
decade later in women exposed to stilbesterol in utero.
Pattern of drug utilization –physician prescribing & pt. usage
Effect of drug overdose
The economic implication of drug use-as its beneficial effects
could reduce the need of medical care, resulting in saving that
can be much larger than the cost of the drug.

18. General contributions of pharmacoepidemiology:
Reassurance of :
Drug safety
Manufacture is fulfilling its organizational duty
ethically and responsible by looking for any
undiscovered problem which may be there.

19. Why to perform pharmacoepidemioloy study
?
Regulatory
Marketing
Legal
Clinical-Hypothesis generating
Hypothesis testing

20. Regulatory
As a requirement for drug approval
To obtain earlier approval to market
As a response to case report of adverse reactions
reported to the regulatory agency
In demonstrating safety of the drug to the regulatory agency in
country which have yet not permitted the marketing of the
drug

21. Marketing
To assist market penetration by documenting the safety of the
drug by clarifying any advantage or disadvantage a drug has as
compared to its competitors.
To increase name recognition as the results of the study would
be publicly presented & published.
To assist in the repositioning of the drug by exploring :
Impact of drug on cost of medical care & pt. quality of life
Effect of drug on different type of patient e.g children, elderly
New indications-unidentified beneficial effect
Lessen the restrictive labeling (toxicity, less efficacy)
To protect the drug from accusation about the adverse effects
Protect investment made in developing new drug.

22. Clinical -Hypothesis testing:
Problem hypothesized on the basis of drug structure/ chemical
class eg. cimitedine related to metinamide (agranulocytosis).
Problem suspected on basis of preclinical or premarketing
animal/human data.
Problem suspected on basis of spontaneous reports eg.
Tolmetine, piroxicam, zomiprac, ketorolac.
Need to better quantitae the frequency of adverse reactions.
eg. Prazosin –frequency of first dose hypotension

23. Hypothesis generating:
A new chemical entity (new important unsuspected effects)
The safety profile of the class is a useful predator of the
experience with the new drug in question.
The relative safety of the drug within the class –less likely to
need supplementary post marketing surveillance
The formulation -if used in institution under close supervision
The disease to be treated, including
Duration –drug to treat chronic illness -long term effects of drug
Common disease –many people exposed to that drug
Severity-if to treat mild / self limited disease –toxicity is not acceptable

24. Legal
In anticipation of eventually having to defend against product
liability suits
Safety v/s. Risk
No drug is completely safe
A drug “is safe if its risk are judged to be acceptable”
Measuring risk is a personal and probabilistic pursuit.
It requires two extremely different kind of activities :
Measuring risk (focus of pharmacoepidemiology)
Judging the acceptability of those risk

25. Case report
Case series
Ecologic
Descriptive
Cohort
Case-control
Non-experimental
RCT
Experimental
Analytic
Epidemiologic study designs

26. Advantages & disadvantages of epidemiology study design
Study design Advantages Disadvantages
Secular trends Provide rapid
answers
No control of
confounding
Case series Easy quantification
of incidence
Cannot be used for
hypothesis testing
Case repots Cheap, easy method
for generating
hypothesis
Cannot be used for
hypothesis testing

27. Study design Advantages Disadvantages
Randomized clinical
trial
(experimental study)
Most convincing
design
controls for unknown,
unmeasurable
confounders
Most expensive,
artificial, difficult
Ethical objections
Focuses on single factor
Cohort study incidence, multiple
causes
Uncommon exposure
Les selection bias
Time consuming
Expensive
Possibly bias outcome
data
Case control studyMultiple exposure
uncommon disease
Easier, faster
Less expensive
Recall bias exposure
Control selection
problem
Biased exposure data

28. Special applications of
pharmacoepidemiology

29. Studies of drug utilization:
Drug utilization studies can be performed to quantify and
identify problem in drug utilization, monitor changes in
utilization patterns, or evaluate the impact of interventions.
It may be conducted on an ongoing basis in programs for
improving the quality of drug use.
Assessing the appropriateness of drug utilization requires data
on indication for the treatment, patient characteristics, drug
regimen, concomitant disease and concurrent use of other
medications.
When assessing quality of care drug utilization studies must
often rely on multiple sources of data

30. Evaluating and improving physician prescribing:
Quality problems in prescribing exist at the level of medication -
over use –e.g. antibiotics for viral respiratory infection in
adult and children
misuse –e.g. coxibs for patients at low risk upper
gastrointestinal bleeding or high risk of cardiac events
under use-e.g. bisphosphonates in the secondary prevention
of osteoporosis related fractures or inhaled corticosteroids for
reactive airways disease.
(The vast majority of prescribing problem are related to under
use of proven effective treatments.)

31. Passive interventions such as dissemination of printed or emailed
guidelines, drug utilization reviews and medication profiles or
traditional didactic continuing medical education lectures are
unlikely to improve practice.
More active intervention strategies (e.g. point-of-care reminders,
educational outreach, achievable benchmarks with audit and
feedback) especially when combined together to overcome
barriers at the level of the system, the physician will be able to
modestly improves the quality of prescribing.

32. Drug utilization reviews :
There is evidence that some forms of drug utilization review
can modestly impact prescribing but evidences that it improves
clinical outcomes is lacking.
Drug utilization reviews programs can have unintended
consequences and thus should be subjected to rigorous
evaluations.

33. Methodological issues in pharmacoepidemiologic
studies of vaccine study:
There are still substantial gaps and limitations in our
knowledge of many vaccine safety issues.
A high standard of safety is required for vaccines as large
number of people are exposed, some of them are compelled by
law or public health regulation.
New research capacity such as vaccine safety Datalink
provides powerful tool to address many safety concerns.

34. Pharmacoepidemiologic study of device:
Medical device are of public health importance and its use
have diverse characteristic and are complex.
Existing data have limited utility for medical devices
epidemiology because complete documentation of device use
is not routine.
Another barrier to good documentation of medical device use
is -lack of a detailed identification system for medical devices.

35. Studies of drug induced birth defects:
Unlike most drug effect, human teratogenesis can only be
identified on postmarketing surveillance.
There is a remarkable ignorance about the teratogenic risk of
the vast majority of prescription and non-prescription drugs
taken by pregnant women.
Most human teratogens increase risks of specific defects (not
birth defect overall)

36. Pregnancy registries are well suited to identify “high risk
teratogens” (e.g. thalidomide, isotretinoin) but do not have
power to identify “moderate risk teratogens”.
Case control approach has the power to identify it and also the
relative safety.
Combining the complementary strengths of cohort & case
control approaches can provide a comprehensive design to
identify human teratogen and to establish ranges of risk/safety
for the wide range of medication taken by pregnant women.

37. Pharmacoepidemiology and risk management:
Risk management encompasses those strategies intended to
shift the balance of benefit and risk for a drug to an acceptable
level.
To date, most strategies have been found to be of no or
marginal effectiveness.
Additional methods are needed that effectively will optimize
the balance of benefit and risk and will succeed in modifying
physician prescribing behavior.

38. Pharmacoepidemiology to study medication error:
Medication errors are very common, compared to adverse drug
events, and relatively few results in injury.
The epidemiology of medication errors and adverse drug events
has been fairly well described for hospital adult, but less
information is available for specific population , and for
ambulatory settings

39. It is now possible to detect many medication errors using large
claims database and it has becomes possible to link these data
with more types of clinical data (including especially
laboratory and diagnosis data), thus it will be feasible to more
accurately asses the frequency of medication errors across
population.
The increasing use of electronic health record will have a
dramatic effect on ability to do research in this area using
pharmacoepidemiological technique.

40. Hospital pharmacoepidemiology:
A substantial proportion of the medical care provided is in a
hospital, treating very ill patients with drugs (some of which
are more toxic than drugs used in outpatient.)
While pharmacoepidemiology began as a field with hospital
data , recently developed automated databases rarely include
the drugs administered in hospitals.
Hospital are complex organizations, and drug used throughout
a hospital is equally complex, administered by multiple types
of personnel and recorded in multiple record forms. The task
of measuring hospital drug exposure fully and accurately is
daunting.

41. Intensive hospital based surveillance consisted of routine
prospective recording of demographic and clinical information
on hospitalized patients, including all drugs administered
throughout their hospital stay.
Then, by comparing the rates of events occurring in these
patient and performing cohort studies, one can detect adverse
reactions.
Hospitals now have ad hocadverse drug reaction monitoring
and drug use evaluation programs.

42. Molecular Pharmacoepidemiology

43. Genes can effect a drug response via: altering of drug
pharmacokinetics & pharmacodynamics of a drug.
It is a study of the manner in which molecule biomarker alter
clinical effects of medication in populations.
It answers questions related to:
Population prevalence of SNP’s & other genetic variants
Evaluating how these SNP’s alter disease outcome
Assessing the impact of gene-drug and gene-gene
interactions on disease risk
Evaluating the usefulness and impact of genetic test in
population exposed or to be exposed, to drugs.

44. The methodological challenges of molecular
pharmacoepidemiology are closely related to the issues of
statistical interactions, type 1 and type 2 errors, and
confounding.
Case only study can be used to measure the interaction
between genetic variants and medications, thus eliminating the
difficulty and inefficiency of including a control group.
The ability of genes and other biomarkers to improve patient
care and outcome needs to be tested in properly controlled
studies, including randomized controlled trials.

45. Advantageous & disadvantages of
pharmacoepidemiology models

46. Advantages of pharmacoepidemiology models:
They can be used in any cases where any other model cannot.
Eg. Pregnant or pediatric patients, those having concomitant
disease or using other drugs.
Generate effective data e.g.. Examine drug use under everyday
condition rather than randomized controlled trial.
These type of studies can be used to alert clinicians,
manufactures, regulators of potential problems.
For monitoring drug use by health care providers, health
authorities, or in the regulation of clinical practice.
They can use data from a variety of sources (records and
charts, patient interviews and computerized database.

47. Disadvantages of pharmacoepidemiology models:
Mostly nonexperimental and tend to be correlational hence,
causation is difficult to establish.
A number of biases and confounding variables can threaten the
validity of pharmacoepidemiological data and has been
categorized as selection bias, information bias and
confounding.
Incomplete records and missing data are also common.
In rare disease the sample size are often very small with
anecdotal information , making confirmation of result difficult.

48. The Future of pharmacoepidemiology

49. The discipline of pharmacoepidmiology has been growing and
is likely to continue further within academia, industry and
government.
Methodologic advances are expected to continue in order to
support this study, as well as new approaches such as risk
management programs and molecular pharmacoepidemiology.
Content areas such as drug utilization reviews, hospital
pharmacoepidemiology, pharmacoeconomics, medication
adherence, patient safety, intermediate surrogate markers will
grow as interest and need for these foci.
Both computerized database and denovostudies will continue
to be important to the field and will serve as important
complementary to each other

50. Challenges -
limiting funding opportunities,
regulatory restriction
privacy concerns surrounding human research
limiting training opportunities
inadequate personnel resources

52. Given concerns of type 1 errors (along with the other
methodological concerns such as uncontrolled confounding,
publication bias and linkage disequilibrium) the key issue here
is to replicate association study findings.
Because of genetic variability leading to phenotypic
expression of complex diseases results from relatively small
effects of many relative low prevalence genetic variants. the
ability to detect gene-response relationship is likely to require
relatively large sample sizes to avoid type 2 errors.

53. Population stratification can distort the gene-
medication response association. Although unlikely
to be a significant source of bias in well controlled
epidemiological association studies, a number of
analytical approaches exist to either circumvent
problems imposed by population genetic structure, or
that use this structure in gene identification.
The ethical, legal, and social implications of genetic
testing must be considered and addressed, just as they
must be considered for all research.