Treatment Considerations And Bias

1
Treatment Considerations And Bias

• Dosages, potency, inclusion/exclusion, study
  setting

2
What is Bias?

• Systematic (or introduced) error that leads to
  data distortion (skewing) and hence, to an
  incoherent conclusion
• Subject selection bias
• Study design
• Random sampling
• Randomization are groups equal
• Inclusion/exclusion criteria

3
Data Collection Bias

• Method selected for data collection
• Measuring method
• Validity and reproducibility of instrument used
  to measure
• Blinding

4
Investigator Bias

• Journal study is published in
• Drug company funding? (dont automatically
  reject the study!)
• Value statements in results/discussion that are
  unsupported by literature (citations, statistics
  from current study)
• Statistical vs. clinical significance
• Use of the word significance

5
Bias with Significance

• Obstructive symptom score in Men with BPH treated
  with Finasteride or Placebo
• Group Baseline score 12 mo. Score
• Placebo 6.7 3.5 5.9 3.8
• 1 mg Finas. 7.4 3.8 6.0 3.8
• 5 mg Finas. 7.0 3.6 5.1 3.6
• Statistically significantly different from
  placebo.

6
Confounding variables

• Something the investigators did NOT do or failed
  to consider, which may have influenced the
  outcome/results.
• Pharmacokinetics
• Side effects
• Demographics of study population
• Pharmacogenetics race, gender, age
• Compliance
• Dosages
• Study setting

7
Dosages

• Must be administered in a dose likely to produce
  benefit
• There is no single or standard relationship
  between the intensity of response and dose (dose
  response curve) which can be uniformly applied to
  all drugs.
• Variation
• small increase in dose can cause large response
• large increase in dose can cause small response

8
Potency

• Because 1 drug is more potent than another on a
  mg to mg basis, this does not imply clinical
  superiority.
• Potency is not important when comparing efficacy
  if they are equivalent doses
• Doses for active controls should also be
  comparable to those of the study drug
• Use of fixed doses could be inappropriate

9
Fixed Doses

• Problems with a fixed dose study
• might not be applicable to real world
• a therapeutic dose in one patient might not be
  adequate in another.
• If patients have widely varying body weights, the
  amount of drug each receives (on a mg to mg
  basis) could vary widely.

10
Therapeutic Concentration

• For many drugs there is no proven exact
  correlation between the concentration of the drug
  and clinical response.
• Beta blockers- propranolol, metoprolol
• anti-inflammatories- ibuprofen, naprosyn
• Some antidepressants- Paxil, Celexa, Zoloft
• Cholesterol/Triglyceride lowering agents-
• Clonidine (BP), Ketoconazole (antifungal)

11
Therapeutic Concentration

• For many, it is easier and more clinically
  relevant to measure the response, or another
  marker
• Warfarin- PT/ INR, Heparin- PTT
• Cholesterol lowering drugs- blood cholesterol
• Diuretics (Dyazide, Diuril)- urinary output
• Anti-diabetic drugs (actos, miglitol, Glipizide,
  glucotrol, metformin)- blood glucose
• Antibiotics (PCN, TCN, Ery., keflex)- blood
  bacteria levels

12
Therapeutic Concentration

• For some drugs, assaying levels is expensive or
  inconvenient
• Free phenytoin levels vs. bound phenytoin levels
• Drug concentrations can be used to help indicate
  whether or not the patient has been compliant
  with medication regimen BUT, no guarantee of
  compliance with regimen.

13
Therapeutic Range

• If drug has a therapeutic range, then doses
  should be employed which will assist patients
  achieving concentrations in the therapeutic range
• Drugs with narrow therapeutic ranges Digoxin,
  Lithium, Dilantin, Theophylline, Tegretol,
  Vancomycin, Gentamicin, Tobramycin, Procainamide,
  Lidocaine, Quinidine

14
Pharmacokinetic Properties

• Absorption Efficacy measurements taken at
  appropriate times, ie. after absorption
• Bioavailability can be altered by
• Concomitant ingestion of food- antibiotics
• Diurnal variation- pravastatin, prednisone
• Alterations in GI tract motility- Reglan (ac)
• pH differences- enteric coated ASA, bisacodyl

15
Absorption in relation to meals

• Pen VK with food slows rate but not overall
  absorption
• Pen G has decreased absorption w/ food
• Ketoconazole has increased absorption w/ food
• Griseofulvin absorption is increased w/ high fat
  meal.

16
Onset of Action

• Immediate release tablet/capsule has fast onset
• Delayed release has slower onset
• Liquids have fast onset
• Injectables have fastest onset
• Transdermal patches have slow onset

17
Time Until Onset of Effect Differences

• Nicotine patch- 2-4 hrs
• Clonidine patch - 6-12 hrs
• Duragesic patch 8-12 hrs

18
Timing of Dose in Relation to Blood Levels

• Antibiotics should be dosed around the clock for
  maximum bioavailability
• Gentamicin is renal toxic if levels are not
  allowed to drop before re-dosing
• Nitrates need to have blood levels drop to a drug
  free period to be most effective

19
Time To Therapeutic Effect

• Need blood levels drawn at appropriate time
• Dependent on half life of drug
• Steady state achieved w/in 4-5 half lives
• Takes 4-5 half lives to remove drug from body
• Peak and/or trough blood levels are appropriate
  for some toxic drugs (Gentamicin, Vancomycin,
  Tobramycin)

20
Concurrent Medications

• Subjects may take non study drugs
• because they have another medical condition
• if they were taking them before this study began.
• Consider drug interactions with study med
• increase bioavailability may make study drug look
  more effective than it is
• affect underlying disease state being studied
• taken equiv. amounts in tx and control group?
• Do they contribute to adverse effects observed?

21
Study Setting- Inpatient Setting

• The more rigidly controlled the environment is,
  the more difficult to extrapolate the findings to
  the outside population
• Inpatient
• close supervision - improved compliance
• less exercise/ activity - more tests
  needed
• more evaluations needed - less protocol
  variation
• less protocol variation
• less environmental variability

22
Outpatient Setting

• More environment variability
• Less compliance control
• More diet variability
• Less testing
• less supervision
• more protocol variation

23
Artificial Setting

• Arranged environment (NH, VA hospital, clinical
  research labs, dorms, hotel rooms, prisons)
• Used more often in studies which recruit
  volunteers.
• This environment is primarily used to control a
  greater of variables

24
Compliance of Drugs

• 80 of patients take less than 80 of their
  prescribed doses
• Those with mild disease have worst compliance
• Those with severe disease have moderate
  compliance
• Those with moderate disease have best compliance

25
Compliance Bias

• Occurs when differences in tx lead to different
  degrees of compliance by patient
• Bias exists when it can be concluded that one of
  the drugs lacked efficacy when in actuality, it
  might not have been taken.
• Example Using Nitroglycerin oint vs. patch
• Ointment messy, more frequent applications
• Patch once per day at bedtime (easier)
• Compliance is usually a greater concern in
  studies involving out-patients than in-patients

26
Methods to Measure and Assess Compliance

• Observing patients take their doses
• pill counts
• use of a non-toxic inert marker in study med
• asking the patient directly and frequently
• measuring the concentration of drug in blood
  stream
• use of an electronic device in the vial cap
• review of Rx records- refills, patient diaries
• Physiological evidence- HR, BP, urine color

27
How Can Non-Compliance Effect the Study Results

• A drugs therapeutic response will be less than
  expected or absent
• A drug will be considered less potent than it
  actually is
• One drug could be assumed to be similar or less
  effective than another drug
• Non-compliance increases the sample size needed
  to detect a difference between groups

28
How to Analyze Non-Compliance

• Exclusion Drop non-compliant patients from
  evaluation of the results.
• Intent-to-treat Include everyone even if they
  had their therapy altered (non-compliant)
• taking patients last score at time they dropped
  out
• taking average score for entire group
• taking worst score for entire group

29
Number Needed to Treat (NNT)

• NNT number of individuals that need to be
  treated in order to prevent one adverse event or
  one outcome. NNT 1
• ARR
• Ex study determine efficacy of drug preventing
  cancer. Incidence of cancer in placebo 15, in
  treatment group 5
• 15-5 10 1/10 10NNT (10 pts needed to
  treat to prevent 1 case of cancer
• NNT 1/ placebo - treatment group

30
Number Needed to Harm (NNH)

• NNH 1/ treatment- placebo group
• Ex Headache occurred in 25 of placebo patients
  and 75 of patients taking drug X.
• The NNH 75-25 50 1/0.5 2
• Only 2 patients would need to be treated with
  drug X in order to cause a headache occurrence.