Lecture #34 - Pharmacoepidemiology

Question 1

Pharmacoepidemiology

1. What is it?
2. Pharmacoepidemiological studies can be what? Egs?

Answer 1

Use of epidemiological methods to study the use, intended effects (benefits), and unintended effects (benefits or harms) of drugs/medicines in populations.

Pharmacoepidemiological studies can be:

Descriptive, for example:
– Drug utilisation studies (person, place, time - are certain groups using it more than others? Are some ppl missing out? Is the use increasing or not and is it used more often in some places than others?)
– Adherence (patients and prescribers - are clinicians adhering to guidelines for prescribing)
– Case reports, case series

Analytical, for example:
– Randomised controlled trials of treatments (intended &
unintended effects)
– Case-control or cohort studies (to study unintended effects). If wanna see if the drug works then use RCT

Question 2

What happens currently? Like, with the drugs - what do they go through? Yeah, explain all the pre-marketing studies the drug goes through

Answer 2

First animal studies are done - potential drug development.

Then there are pre-marketing trails:
-These trials look at metabolism of drug + how its handled in human system, what’s safe dosage range + exclude any specific toxic reactions that are specific to humans that haven’t been picked up in the animal studies.

1. Phase 1 = first use in humans, healthy volunteers (unless v toxic drugs) and used in only like less than 100 population
2. Phase 2 = Pilot efficacy studies, volunteers with disease (200 - 500)
   - Phase 1 and 2 are usually conducted by pharmacologists
3. Phase 3 = Extensive clinical trials, volunteers with disease (usually a few thousand at the most)
   - Hoping to treat this who volunteer
   - Researches wanna know if drug actually works and have a look at toxicity

Then there is the national drug registration authority (e.g. Medsafe, FDA) who approves drug for distribution and marketing.
-Seek approval to distribute and market

Then, lastly, you have the post-marketing studies of drug safety (Phase 4)
-Phase 4 of monitoring the safety of drugs

At any time, the drug can drop out - may not even make it past the animal testing phase

Question 3

Limitations of pre-marketing studies? (8)

Answer 3

1. Animal models may be inappropriate
   - animals aren’t the same as humans so animals might not be the same
   - “The initial studies conducted on animals may not be applicable to understanding how the drug works in humans.”
2. Healthy human volunteers, followed by highly selected group of patients
   - doesn’t say whats gonna happen in real population
   - The pre-marketing studies start by using healthy human volunteers and then progress to highly selected patient groups. Consequently, the results of these studies may not be widely generalisable.
3. Small numbers in trials
4. Short follow-up period (only months)
5. Primary focus is efficacy, acceptability, common and short-term adverse reactions
   - Small study samples and short follow-up periods are features of pre-marketing studies. This may result in less precise and accurate results, and a lack of information on long-term, or rare but serious, side effects.
6. May not record all adverse events which occur during trial
7. May look at surrogate endpoints only
   - If have drug that might prevent heart attacks and stroke - might just look at whether it lowers cholesterol or BP - not following enough people for long enough time to see whether it lowers risk of heart attacks/stroke
   - Often these studies involve surrogate endpoints (e.g. changes in blood pressure), as the follow-up time is generally not long enough to investigate the longer term, but ultimately more meaningful effects of the drugs (e.g. hard endpoints such as reduction in deaths from coronary artery disease).
8. (Deviation from analysis plan, suppression of results…)
   - Unfavourable results from these studies maybe suppressed, or continually reanalysed to provide a favourable result, if they do not conform to the study sponsor’s business/financial plan.

Question 4

So, pre-marketing studies often….

…..Don’t identify adverse reactions that (4)

….Or (2)

Question 5

What descriptive studies are used for drug safety?

Answer 5

1. Case reports - doctors write in journal about possible adverse symptom, case series - selection of case reports (voluntary reporting systems, journals)
   - common pattern to how adverse effects energy - usually through disruptive studies - case reports
2. Monitoring routine statistics

Question 6

(Descriptive studies of drug safety) Voluntary reporting systems
-tell me about it

Voluntary reporting systems

• what are the 3 advances
• what are the 3 limitations

Answer 6

So like NZ Pharmacovigilance Centre is a centre for adverse reactions are reported. Also provide reactions to over-the-counter medication etc
-It’s important to ask patients not just about prescribed medication but also what else they might be taking too (e.g. herbal products cause serious adverse effects or they reduce effectiveness of prescribed medicine)

Advantages:

1. Prove early warnings
2. Inexpensive
3. Educational role

Limitations:

1. Relies on suspicion of health professionals (for them to see the link between patient and drug and also relies on them having time to report - effort to report)
2. Under-reporting (see above thing with time)
3. Can’t measure incidence (don’t have all people who suffered reaction i.e. no numerator. And don’t have denominator bc don’t know how many are taking the drug so don’t know what to make of the findings)

Question 7

(Descriptive studies of drug safety) Monitoring of statistics

1. E.g……
2. One advantage
3. Three limitations

Answer 7

1. E.g morbidity and mortality trends, prescribing trends
   - look at trends, hospital admissions, death and look at them alongside prescribing trends for drug of interest
2. Inexpensive
3. Limitations:
   - Very insensitive, only major hazards detected
   - Delay in availability of data esp mortality (e.g. cause of death is 2 years behind)
   - Uncertainty whether drug to blame or other changes e.g. in risk factors, prescribing practices (or in ppl). Like, if the trend for death increases for certain condition and this is in parallel with drug - can’t be sure it’s because of drug. It might be because doctors prescribing it to more “at risk” people over time

Question 8

Descriptive studies

1. What is the hypothesis generated by these studies?
2. What is the problem?

Answer 8

1. “Drug x increases occurrence of outcome y”
2. Problem is that outcome ‘y’ also occurs in people not taking drug x.
   - so analyse whether y occurring more in people taking the drug than not taking the drug

Question 9

Analytic studies of drug safety - what three are used?

Test hypotheses and determine what 4 things?

Answer 9

Case-control, cohort and intervention

Test hypothesis and determine (see if link bw drug and outcome):

1. Magnitude of harm/benefit (look at RR and absolute risk)
   - maybe drug increases risk x5 but if absolute risk is 1 in 10 million then increasing it but 5 won’t be too bad but if background rate (absolute risk) was 1 in 100 then x5 is big.
2. Any effect modification
   - e.g. RR of heart attack in ppl taking pill is higher in smokers than non-smokers so smoking is effect modifier. So would want to know if patient smokes or not
3. Whether harm/benefit only during use or persists
4. Whether magnitude of harm/benefit related to dose or duration of use (pre important practical thing analytical studies help with)

Question 10

Discuss the various epidemiological study designs that are used in the post- marketing surveillance of drug safety.

Answer 10

Descriptive information is gathered from case reports, case series, and routinely collected data in order to generate hypotheses that may require further investigation. These are not considered study designs in PUBH192, but are the initial steps toward understanding if a problem that deserves investigation might exist.

The three analytic study designs that are used in post-marketing drug studies are case- control, cohort, and randomised controlled trials.

• Case-control studies allow for the efficient investigation of rare adverse effects.
• Cohort studies allow for long follow-up periods, which may include the investigation of hard endpoints.
• RCTs may not be possible if there are strong suspicions that the drug is causing a harmful side effect (there is a lack of equipoise).

Question 11

How can bias arise in pharmacoepidemiological studies? What steps can be taken to reduce the possibility of bias distorting the results of these studies?

Answer 11

As with other types of studies, selection, information and publication bias are all possible.
• Selection bias – randomly select participants from the source population (to increase external validity); regular contact with participants involved in follow-up studies; focus on achieving good response rates; use anonymised routinely collected data.
• Information bias – use objective measures of exposures and outcomes (e.g. routinely collected data); when collecting information by self-report use structured questionnaires, trained interviewers, memory triggers.
• Publication bias – ensure that results of studies are available and published; compare the methods used in the study with what was specified in the protocol; search for grey literature.