Literature Evaluation

1
Literature Evaluation

• Types and Quality of Medical Literature
• Study Design

2
Types of Drug Resources- Tertiary

• Tertiary Sources
• Compendia (Facts and Comparisons, textbooks,
  Micromedex, review and meta-analysis articles)
• Advantages convenient, easy to use, quick
  access, compact information, general overviews
• Disadvantages incomplete, outdated, contains
  most human bias, lack of expertise and evidence.

3
Types of Resources Secondary

• Indexing and abstracting services (Medline)
• Use to locate primary drug literature resources
• Advantages easy access to primary resources
• Disadvantage cost, some only have indexing or
  abstracting available

4
Types of Resources Primary

• Articles in journals about original research or
  case reports of actual drug trials/studies.
• Advantages Provides significant detail of drug
  knowledge, research methodology, scientific
  results allowing the reader to evaluate the drug
  in the most non-biased and scientifically based
  way.
• Most current, most unbiased, used for therapeutic
  decisions (actual research data).

5
Types of Literature ArticlesTertiary Resource

• Review articles-- tertiary resource articles (not
  primary)
• Purpose to review a drug in a monograph form or
  in a literature search format where many articles
  are reviewed, clinical studies are looked at and
  conclusions are made that reflect the current
  thinking on a given subject or drug.
• Advantages Excellent beginning resource to
  locate information about a particular subject.
  Rapid access to large amount of information.

6
Types of Literature Articles.Tertiary Resource

• Systematic Review
• Summarizes scientific evidencereviews multiple
  studies
• One or more questions/hypotheses posed at start
  of review
• Studies are collected to review, data is
  extracted and combined/compared Or, if data
  cannot be combined, strength of evidence is
  assessed and used to evaluate results.
• Conclusions are made based on results and
  presence or absence of supporting evidence.

7
Types of Literature ArticlesTertiary Resource

• Meta-Analysis studies tertiary resource
  information.
• A quantitative or statistical method of combining
  the results of independent studies (usually drawn
  from published literature) and synthesizing
  summaries and conclusions, statistically, which
  may be used to evaluate therapeutic
  effectiveness, plan new studies, etc.
• Used when measuring very small effects, to
  resolve conflict or inconclusive results among
  studies, or to evaluate objectives not posed at
  the beginning of the study.
• Always has statistical evidence present.
• Always a retrospective evaluation of collected
  data.

8
Types of Literature Articles in Medical Journals

• Pharmacokinetic studies
• To estimate
• the drugs pharmacokinetic parameters, such as
  volume of distribution, metabolism, elimination.
• bioavailability or bioequivalence of drugs or
  dosage forms.
• Drug interactions (food, drug, genetics)
• Common in phase I or II clinical trials
• animal or human studies (small s, healthy
  volunteers, not placebo controlled, not blinded,
  not randomized)

9
Types of Literature Articles in Medical
Journals...

• Pharmacodynamic Studies
• To evaluate drug concentration-effect
  relationships (if drug conc. Produce a
  therapeutic or toxic effect)
• Can be animal or human studies (small s,
  healthy volunteers, not blinded, not randomized,
  not placebo controlled)
• Common place Pre-clinical and phase I trials.
  Often combined with pharmacokinetic studies.

10
Types of Literature Articles

• Pharmacoeconomic Studies
• Studies identifying cost effectiveness for a
  given dosage regimen, disease state, or drug
  therapy
• Provides additional information about the
  potential impact of a drugs clinical use.
• Example A Study comparing the cost effectiveness
  of tricyclic antidepressants with the newer SSRI
  antidepressants in regards to actual pill costs
  and overall mental health care costs
  (hospitalizations, compliance with medication,
  longer term side effects, etc.)
• Common place phase IV trials

11
Types of Literature Articles...

• Clinical Methodology Studies
• Development of new methods compared to the
  current standards.
• Example A study to estimate the predictive
  performance of a new instrument, sensitivity and
  specificity of a diagnostic test or validation of
  a new drug dosing regimen.
• Purpose to look for more convenient, less
  expensive or less invasive means of making
  measurements.

12
Types of Literature Articles...

• Pharmacoepidemiologic Studies
• These studies evaluate the use and effects of
  drugs in large numbers of people. Used to assess
  drug use patterns and to document clinical
  efficacy and toxicity of drugs.
• Have led to the discovery of uncommon, but
  serious side effects of drugs, leading to their
  withdrawal from the marketplace.
• Most common type of design cohort, case-control.
• Common place phase IV trials

13
Types of Literature Articles.

• Quality of Life Studies
• performed to evaluate the functional effects,
  beneficial or adverse effects that a drug or
  treatment regimen may have on a patient and their
  quality of life.
• Used to assess treatments that have minimal
  effects, to evaluate a drug in a disease state
  where few effective therapies exist, or to
  differentiate between drugs of the same class.
• Most often in phase IV trials.

14
Editorial Board

• Editorial boards are made up of people with
  expertise to critique the type of manuscript
  submitted for publication.
• Editorial board members
• Advise the editor about the suitability of
  submitted manuscripts
• Assist the editor in setting journal policy

15
Peer Review

• Experts in their field, asked to review, critique
  manuscripts before approval for publication.
• Most journals do not identify peer reviewers for
  specific articles.
• Peer reviewers may or may not be part of the
  editorial board.

16
Peer Review

• How to tell if journal article has been peer
  reviewed
• 1. Look in Guidelines for submitting
  manuscripts. This will tell you how many copies
  of their manuscript they need to send in, and
  will often indicate if peer review is used. (If
  more than 2 copies, assume it is peer reviewed)

17
Peer Review.

• 2. Check either the bottom of the first page or
  the last page of the published study. Some
  journals will print the date they received a
  manuscript, the date it was revised and the date
  it was accepted for publication. The use of peer
  review will often mean that 3 months or more have
  lapsed from the time a manuscript is received by
  the journal to the time it is accepted for
  publication.

18
Peer Review.

• 3. If the manuscript has been revised, it
  usually means it has been peer reviewed and the
  author has been asked to make changes.
• 4. Some journals will put a special symbol in
  the side margins or at the end of those articles
  that have been through the peer review process.

19
Advertising in Journals

• The amount of advertising (ads for specific
  drugs, books, equipment, etc) varies from journal
  to journal, and ranges from none to the majority
  of the contents.
• Advertising does not in itself affect the quality
  of a journal, but it does lower the cost of the
  journal for its readership.

20
Controlled Circulation

• Several journals are distributed free of charge,
  with their cost entirely absorbed by advertisers.
  
• These journals focus on a specific target
  audience (community practitioners, retail RPh,
  etc) and distribution is targeted to these
  individuals. This method of targeting is called
  controlled circulation.
• Controlled circulation journals usually do not
  contain comprehensive research or clinical
  studies.
• They can provide topic overviews of interest as
  review articles.

21
Citations

• Citations (references) indicate the author has
  researched the subject and has used previous
  published work as a guide to this research.
• Caution- Sometimes an author will cite his own
  work from a previously written article on the
  same/similar work-- can create bias and lower the
  quality of the reference

22
Review of Clinical Trial Phases

• Phase I to identify preferred route, safe dosage
  range, toxicity, pharmacokinetics. Uses healthy
  volunteers w/ no pre-existing conditions. Uses
  low dosages.
• Phase II to evaluate efficacy in subjects with
  proposed disease state identify group most
  likely to benefit from drug.

23
Review of Drug Trial Phases

• Phase III Compare to current therapy,
  multicenter studies, serve as basis for NDA,
  large numbers of patients tested.
• Phase IV Post marketing surveillance studies, to
  look for side effects, new indications, new
  dosages, long term effects, drug interactions in
  defined populations.

24
Drug Study Design
25
Purpose of a Study Design

• To answer a research question.
• The research design determines what can be
  concluded from a particular study and the amount
  of inference required to interpret the data.
• Research design influences the selection of the
  statistical procedures used to analyze and
  interpret results.

26
Research Design

• Organized around three general concepts
• Description
• Analysis
• Evaluation
• 2 basic types of medical research studies
• Observational
• Experimental

27
Hierarchy of Research Design

• Lowest levels of evidence
• Descriptive
• Cross Sectional
• Case-Control
• Higher levels of evidence
• Cohort
• Highest levels of evidence
• Experimental (randomized controlled study)

28
Types of Medical Studies...

• Observational
• prospective (happening in present and future)
• retrospective (looking backwards for data)
• observing only, no direct intervention
• Experimental
• always prospective
• investigators of study have direct intervention

29
Observational Designs

• Descriptive (Case-Series, Case Report) Study
• describes situation in group of patients, or
  interesting or unique phenomena.
• No comparison or control patient groups are used.
• Example 15 patients are described in whom
  resolution of herpes zoster occurred following
  the initiation of Tagamet therapy for peptic
  ulcer. All pts. began therapy with Tagamet 300 mg
  QID for peptic ulcer disease during the time they
  were having acute episodes of herpes zoster. The
  zoster resolved in 3 days

30
Descriptive (Case Series, Case Report)

• Advantages
• Relatively easy and inexpensive to perform
• Provides preliminary information that could prove
  valuable upon further study- raises questions
  that may prompt more studies
• Disadvantages
• Patients might not be representative of others
  with the specific condition
• Information provided is subject to many potential
  biases.

31
Observational Design- Cross Sectional Study

• Sometimes called Prevalence or Survey Study.
• Used to determine the existence of specific
  conditions or characteristics at one given point
  in time. It identifies a sample population and
  makes an assessment of outcome all at the same
  time.
• No controls used

32
Cross-Sectional Study

• Example A study selects patients which have had
  an adverse drug reaction to a glaucoma medicine.
  The study looks for risk factors at the time the
  study patients were selected.
• Example Study to determine whether serum
  creatinine increases with age. SCr levels were
  measured in 400 healthy subjects, between 18-70
  y.o. The relationship between age and serum
  creatinine were then analyzed. The analysis
  showed the strength of the relationship or
  association and the direction of the relationship
  or association.

33
Cross-Sectional Study Design

• Leads to the Chicken and the Egg dilemma of
  which came first. Does the patient have the ADR
  because of the risk factor or is the risk factor
  a manifestation of the disease, ADR, or a
  confounding factor?

34
Cross-Sectional Study Design

• Advantages
• Less time consuming to perform
• Least expensive to perform
• Easy to complete (sometimes survey form)
• Appropriate for evaluating the current status of
  a condition or characteristic.
• Useful for measuring prevalence, association, and
  estimating risk with odds ratios.

35
Cross Sectional Study Design

• Disadvantages
• Bias may be introduced if the measured outcome
  depends on chart review or recall information
  (selection bias, subject bias, investigator bias)
• Transient effects If the condition being
  monitored can change over time, the results can
  be inaccurate or misleading.
• Does not establish causation
• No control group

36
Observational Design- Case Control Study

• Case-Control Study
• Used to determine the possible factors
  influencing (risks), or the causes of, an
  observed event or outcome. Since one is looking
  back in time to identify the factors or causes of
  the observed event or outcome, the design is
  always retrospective.
• The cases are the patients who have the
  observed event or outcome.
• The controls are similar to the cases, except
  that they do NOT have the observed event or
  outcome. The factors that are different between
  the cases and controls are then examined in terms
  of the ability to influence the event or outcome.
• Always 2 groups of patients one group with
  outcome, one group without outcome.

37
Case-Control Study

• Study Example
• A study was conducted to evaluate the risk of
  leukemia resulting from the use of cancer
  chemotherapy to treat breast cancer. The patient
  files of 90 women with breast cancer diagnosed
  from 1973 to 1985 who developed leukemia were
  examined. Two to 3 additional patients with
  breast cancer but without leukemia were matched
  to each of the other 90 patients and examined. It
  was found that leukemia was uncommon in patients
  with breast cancer, although the risk of
  developing it was significantly enhanced by the
  use of regional radiation.

38
Case-Control Study

• Always retrospective.
• Control group used (without outcome)
• Reports odds ratios and risk ratios
• Outcome is known- you are looking for factors
  influencing the development of the outcome.
• Measures risks.
• Sometimes called Trohoc studies

39
Case-Control Study

• Advantages
• Relatively quick to perform
• Generally inexpensive
• Appropriate for studying rare diseases or
  conditions which take several years to develop.

40
Case-Control Study

• Disadvantages
• Dependent upon accuracy and completeness of
  existing records of past events.
• Cannot evaluate cause and effect of measure
  incidence-- only prevalence and association.
• Easily biased by the retrospective nature.
• Berksons Bias Difficult to obtain controls
  which are truly comparable to the cases
• Protopathic Bias A disease leads to exposure of
  risk factor rather than visa versa.

41
Berksons Bias

• The more similar the case and control groups are
  to each other, the more likely a difference found
  between the groups will be due to the variable
  being assessed instead of random variation.
• Example A study who uses subjects drawn from
  hospitalized populations whose exposure to the
  risk factor may differ from a population in the
  community.

42
Protopathic Bias

• A disease leads to exposure of the risk factor
  rather than the exposure of the risk factor
  leading to the disease.
• Example Vaginal bleeding may be an early sign of
  uterine cancer. But vaginal bleeding leads to the
  use of progesterone. A study may say that
  progesterone caused cancer but, maybe the cancer
  preceded the progesterone use.

43
Terminology

• Prevalence rate The proportion of existing cases
  of a medical condition in the population at risk
  at a specified point in time. For example, the
  prevalence of hospitalized-acquired pneumonia is
  the number of cases divided by the total number
  of patients hospitalized at a particular point in
  time.

44
Terminology

• Incidence Rate The ratio of the number of new
  cases of a condition to the total number of
  persons in the population at risk for the
  condition during a specified period. For example,
  the incidence of a condition developing may be
  reported as the percentage per person per year.

45
Terminology

• Odds Ratio A relative measure of disease
  occurrence that provides an estimate of the
  relative risk. It is defined as the frequency
  with which the event occurs divided by the
  frequency with which it does not occur and is
  usually calculated for case-control studies where
  the incidence is not known.

46
Observational Design- Cohort Studies

• Longitudinal study and observational in nature.
• Example study looking for drug-drug
  interactions.
• Cohort group of patients who have not yet
  experienced the outcome under study.
• Cohort Study characteristics of the cohort are
  recorded over time as the investigator observes
  them waiting for the outcome to occur.
• Does use a control group

47
Cohort Study

• Prospective Cohort follows or observes patients
  both with and without certain characteristics/feat
  ures over time period to determine what outcomes
  will be.
• Retrospective (historical) Cohort Patients are
  followed both with and without certain features
  but all information is collected from the past
  (records, charts, billing claims etc)

48
Cohort Study

• Advantages of Cohort Study
• Usually a clear relationship between risk factors
  and the outcomes can be determined.
• Allows for uniform data collection and avoids
  recall bias (prospective cohort).
• Useful for measuring incidence, relative risk,
  association and causation
• Uses a control group -ranks higher for levels of
  evidence in evidence base medicine)

49
Cohort Study

• Disadvantages of Cohort Study
• Exact causes of an outcome cannot be definitively
  proven since only observations are being made (no
  intervention)
• Expensive if patients are followed over long
  period of time
• Drop out rate can be higher if study is over long
  period of time.

50
Hawthorne Effect

• Subjects change their behavior while being
  observed (if they know they are being watched,
  they will perform differently)
• This can increase compliance and increase side
  effects.

51
Cohort Study Example

• In order to calculate the relationship between
  high serum cholesterol concentrations and the
  risk of developing heart disease, 1000 healthy
  men between 35-50 years were enrolled into a
  study. They were divided into 2 groups- those
  with cholesterol lt180 and those gt180. Men were
  given medical checkups yearly. At the end of 10
  years, the results indicated men with high
  cholesterol were 2x more likely to develop heart
  disease.