how to evaluate an article

Upload: senaa-daughter

Post on 08-Apr-2018

229 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/6/2019 How to Evaluate an Article

    1/38

    Dr. Noha A. DashashMBBS, DPHC, ABFM, SBFMConsultant Family Physician,Director of PHC, Jeddah GovernorateMember of the National EBM CommitteeSupervisor, Jeddah EBM Working Group

    Dr. Noha A. DashashMBBS, DPHC, ABFM, SBFMConsultant Family Physician,Director of PHC, Jeddah GovernorateMember of the National EBM CommitteeSupervisor, Jeddah EBM Working Group

  • 8/6/2019 How to Evaluate an Article

    2/38

    29 May 2011 EBM Jeddah Working Group 2

    In The Name Of AllahThe Most Merciful

    The Most Compassionate

  • 8/6/2019 How to Evaluate an Article

    3/38

    29 May 2011 EBM Jeddah Working Group 3

    Components of an articleComponents of an article

    Title

    Abstract Introduction-background

    Methods (Materials & Methods)

    Results Discussion- Conclusion

  • 8/6/2019 How to Evaluate an Article

    4/38

    29 May 2011 EBM Jeddah Working Group 4

    IMRADIMRAD

    Introduction

    why

    Methodology how

    Results and

    What they found Discussion

    What they THINK the results mean

  • 8/6/2019 How to Evaluate an Article

    5/38

    29 May 2011 EBM Jeddah Working Group 5

    The science of trashing aThe science of trashing a

    paperpaper Does it answer your question?

    Title

    Abstract

    Introduction

    AIM, Objectives

    Study Design: Is it suitable for answering your question?

  • 8/6/2019 How to Evaluate an Article

    6/38

    29 May 2011 EBM Jeddah Working Group 6

    The science of trashing aThe science of trashing a

    paperpaper

    METHODS

    If methodology is not sound

    trash the paper before reading the

    results even if statistically

    significant

  • 8/6/2019 How to Evaluate an Article

    7/38

    29 May 2011 EBM Jeddah Working Group 7

    Critical AppraisalCritical Appraisal

    critique

    1. A critical review or commentary,

    especially one dealing with worksof art or literature.

    2. A critical discussion of a

    specified topic.3. The art of criticism.

    The Free dictionary, by Farlex

  • 8/6/2019 How to Evaluate an Article

    8/38

    29 May 2011 EBM Jeddah Working Group 8

    Critical AppraisalCritical Appraisal

    critique

    review, oranalyze

    a serious examination andjudgment of something;

    "constructive criticism is alwaysappreciated"

    The Free dictionary, by Farlex

  • 8/6/2019 How to Evaluate an Article

    9/38

    29 May 2011 EBM Jeddah Working Group 9

    Critical AppraisalCritical Appraisal

    Appraisal: the classification of

    someone or something with

    respect to its worth

  • 8/6/2019 How to Evaluate an Article

    10/38

    29 May 2011 EBM Jeddah Working Group 10

    Critical AppraisalCritical Appraisal

    Critical appraisal: an appraisal

    based on careful analytical

    evaluation

  • 8/6/2019 How to Evaluate an Article

    11/38

    29 May 2011 EBM Jeddah Working Group 11

    (critically appraise) evaluation for

    validity and usefulness

  • 8/6/2019 How to Evaluate an Article

    12/38

    29 May 2011 EBM Jeddah Working Group 12

  • 8/6/2019 How to Evaluate an Article

    13/38

    29 May 2011 EBM Jeddah Working Group 13

    Measures of StudyMeasures of StudyReliabilityReliability

    B ias, I nternal validity,

    C onfounding,

    E xternal validity,

    P ower

  • 8/6/2019 How to Evaluate an Article

    14/38

    29 May 2011 EBM Jeddah Working Group 14

    BiasBias

    A systematic tendency to produce anoutcome that differs from the

    underlying truth

    factorarisingfrom Design

    Conductofa study

    skewsthe data in one particulardirection,eitherawayfromortowardsthetrue

    population valuethatis beingestimatedwiththestudy

  • 8/6/2019 How to Evaluate an Article

    15/38

    29 May 2011 EBM Jeddah Working Group 15

    BiasBias

    Selection bias

    example, we select sicker patients to

    receive active treatment and fitter

    patients to receive inactive treatment

    How can it be overcome? Randomization CONCEALMENT

  • 8/6/2019 How to Evaluate an Article

    16/38

    29 May 2011 EBM Jeddah Working Group 16

    BiasBias

    Observer bias

    example, we know that the patient

    had active treatment so wesubconsciously encourage her to rateher quality of life as higher than itreally is.

    How can it be overcome? Blinding

  • 8/6/2019 How to Evaluate an Article

    17/38

    29 May 2011 EBM Jeddah Working Group 17

    BiasBias

    Participant bias

    example, in a study of aspirin versus

    no treatment, people allocated to notreatment take aspirin anyway

    Contamination

  • 8/6/2019 How to Evaluate an Article

    18/38

    29 May 2011 EBM Jeddah Working Group 18

    BiasBias

    Withdrawal bias ordrop out bias

    when we lose people to follow up,

    those that remain for analysis at theend of the study may not be

    representative of the group originally

    included at the start of the study.

  • 8/6/2019 How to Evaluate an Article

    19/38

    29 May 2011 EBM Jeddah Working Group 19

    BiasBias

    Recall bias

    example, mothers of children withleukemia may remember living nearhigh voltage power cables becausethey fear a link between power linesand cancer, while mothers of childrenwithout leukemia are likely to forget

    whether they lived near a power line,because they regard it as a trivial fact.

  • 8/6/2019 How to Evaluate an Article

    20/38

    29 May 2011 EBM Jeddah Working Group 20

    BiasBias

    Instrument or measurement bias

    E.g. blood pressure measurement

  • 8/6/2019 How to Evaluate an Article

    21/38

    29 May 2011 EBM Jeddah Working Group 21

    BiasBias Publication bias (in Systematic reviews/MA)

    results from researchers and journalsbeing biased towards publishing onlypositive results

    If negative or equivocal results remainhidden, then clinicians can get amisleadingly excessive view of howeffective a treatment really is (or of how

    harmful an exposure might be).

    authors of good systematic reviewsmake a point of tracking downunpublished research.

  • 8/6/2019 How to Evaluate an Article

    22/38

    29 May 2011 EBM Jeddah Working Group 22

    WorkWork--up Biasup Bias (Verification Bias)(Verification Bias)

    In Studies of Diagnostic Tests:

    Subjects with a negative test are sometimes not

    referred to the 'gold standard' assessment.

    In some cases, the investigators may refer a small

    random sample of subjects with a negative test to

    the 'gold standard' assessment in order to

    determine if they really do not have the disease.

    'gold standard' results are not available on anentire population,

    work-up bias,

    equations for sensitivity and specificity give biased

    results.

  • 8/6/2019 How to Evaluate an Article

    23/38

    29 May 2011 EBM Jeddah Working Group 23

    Internal validity ofInternal validity ofmeasurementmeasurement

    how well a study was run (researchdesign, operational definitions used, howvariables were measured, what

    was/wasn't measured

    Wasitreallythetreatmentthatcausedthe difference between thesubjectsin

    thecontrol andexperimental groups?"In descriptive studies (correlational, etc.)

    internalvalidityrefersonlytotheaccuracy/qualityofthestudy

  • 8/6/2019 How to Evaluate an Article

    24/38

    29 May 2011 EBM Jeddah Working Group 24

    results of study are due to the unusual typeof people who were in the study.

    Or, it might only work because of the unusualplace you did the study in

    (e.g. you did your educational study in a collegetown with lots of high-achieving educationally-oriented kids).

    Or, study done in a peculiar time. For instance, smoking cessation study done the

    week after the General Surgeon issues the well-publicized results of the latest smoking and cancerstudies,

    you might get different results than if you had doneit the week before.

  • 8/6/2019 How to Evaluate an Article

    25/38

    29 May 2011 EBM Jeddah Working Group 25

    Internal validity ofInternal validity ofmeasurementmeasurement

    1. How good (VALIDITY of) ourassessment tools are for measuring

    the things we want to measure in thestudy (outcomes)

    Quality of life: asking people how are

    you on a scale of 1 to 10?

    or

    Using a validated Questionnaire(e.g.QALY)

  • 8/6/2019 How to Evaluate an Article

    26/38

    29 May 2011 EBM Jeddah Working Group 26

    Internal validity ofInternal validity ofmeasurementmeasurement Cont.

    Glucose control assessment

    monthly urine dipstick test?

    Or

    glycosylated hemoglobin

  • 8/6/2019 How to Evaluate an Article

    27/38

    29 May 2011 EBM Jeddah Working Group 27

    Internal validity ofInternal validity ofmeasurementmeasurement Cont.

    Smoking

    Asking people how much do you smoke?

    Or

    asking people to keep their empty

    cigarette packets for measuring tobacco

    exposure.

    Whichhasmoreinternalvalidity

  • 8/6/2019 How to Evaluate an Article

    28/38

    29 May 2011 EBM Jeddah Working Group 28

    Internal validity ofInternal validity ofmeasurementmeasurement Cont.

    2. Proxy or intermediate endpoints For example a study may suggest that a treatment

    reduces heart disease, when the only outcome

    assessed was blood cholesterol.

    The internal validity of cholesterol measurements for

    measuring heart disease is poor

  • 8/6/2019 How to Evaluate an Article

    29/38

    29 May 2011 EBM Jeddah Working Group 29

    ConfoundingConfounding

    is an exposure that the

    researcherdid notmeasure that

    independently affectstheoutcomes of the study andwhich

    is associated with the exposure

  • 8/6/2019 How to Evaluate an Article

    30/38

    29 May 2011 EBM Jeddah Working Group 30

    ConfoundingConfounding

    Example: study of the relation between lung

    cancer and coffee drinking

    association is found between the two.

    cause coffee drinkers are more likely to

    smoke

    real risk factor for lung cancer

    smoking. In this study, the left-out variable ofsmoking

    confounds the relationship between coffee

    and lung cancer

  • 8/6/2019 How to Evaluate an Article

    31/38

    29 May 2011 EBM Jeddah Working Group 31

    External validityExternal validity

    External validity is related to

    generalizing

  • 8/6/2019 How to Evaluate an Article

    32/38

    29 May 2011 EBM Jeddah Working Group 32

    Threats to External ValidityThreats to External Validity

  • 8/6/2019 How to Evaluate an Article

    33/38

    29 May 2011 EBM Jeddah Working Group 33

    Threats to External ValidityThreats to External Validity

    if you did your smoking cessation study the

    week after the Surgeon General issues the

    well-publicized results of the latest smoking

    and cancer studies

    type of people who were in the study

  • 8/6/2019 How to Evaluate an Article

    34/38

    29 May 2011 EBM Jeddah Working Group 34

    PowerPower

    the ability of the study to detect

    an effect if in truth there is an

    effect (ie were we to perform an ideal studyof the entire population). We can

    increase the power of the study

    by:

  • 8/6/2019 How to Evaluate an Article

    35/38

    29 May 2011 EBM Jeddah Working Group 35

    We can increase the power ofWe can increase the power ofthe study by:the study by:

    (a) increasing the number of

    outcome events we observe

    by increasing the sample size

    or duration of the study

  • 8/6/2019 How to Evaluate an Article

    36/38

    29 May 2011 EBM Jeddah Working Group 36

    We can increase the power ofWe can increase the power ofthe study by:the study by:

    (b) improving the precision of our

    measurements

    by increasing the sample size

    or duration of the study

  • 8/6/2019 How to Evaluate an Article

    37/38

    29 May 2011 EBM Jeddah Working Group 37

    In a NutshellIn a Nutshell

  • 8/6/2019 How to Evaluate an Article

    38/38

    THANK YOUTHANK YOU