basic lecture on literature types, importance of primary literature (papers,article) , study designs, and organization of scientific paper. p value and assessment of a new test is additional topic.

2. Critical Appraisal
How to read a scientific paper ?
Mohammed AbdElWadood
MD Urology, MRCS Eng.

9.  Appraisal For:
1. Relevance ‫النفعية‬ :
2. Validity (Accuracy) ‫المصداقية‬: free from bias,
before study: selection bias, during study:
measurement bias (absent randomization,
absent control, non-blinded), after study:
analysis bias (improper statistics).
3. Reliability (Consistency) ‫الموثوقية‬: results similar
in different studies (replicable).
4. Applicability (Importance) ‫القابلية‬‫للتطبيق‬ :
statistical significance and also clinical
importance.

11. 
Original Articles Review Articles Textbooks

12. Review Articles
• Analysis of 1ry
(many papers)
• Comprehensive
/ focused.
• Older.
• More
established
Original Articles
• 1ry Source.
• focused.
• Updated.
• Needs
appraisal.
Textbooks
• Analysis of well
accepted 1ry.
• More
comprehensive
(broader).
• Even older.
• Well established.
high academic standard
Not peer-reviewed
(less scientific standard)Best to cite for research

18. Case-Control
Cohort
Cross-sectional
(survey)
Case report
Case series

20. Prospective or respective study are called
longitudinal studies.
Exposure Outcome
Exposure Outcome

23.  Describe an observation only (No control, No
intervention, no randomization).
 Value: generate hypothesis, that is tested by other
study designs.The results are not applicable yet.

24.  stories that are interesting or unusual in some way.
 Case report: describes (history) of a single patient as a
story.
 LE 5: No interv., Not Randomized, Retrospective, No
Control Group.
 Case series: describes a series of patients with an
outcome of interest.
 LE 4: No interv., Not Randomized, Retrospective, No
Control Group.
 Example: Students mentoring students in a service-
learning clinical supervision experience: an
educational case report. Lattanzi JB, et al. Phys Ther.
2011 Oct;91(10):1513-24.

25.  Observe difference between 2 or more groups
(samples), with control (No intervention, no
randomization).
 Value: determine association but can’t determine
causality (cause-effect) because not randomized.

26.  Researchers choose group with a particular outcome
(case group) and group without that outcome (control
group) and interview them or check their records to
detect what different exposure between groups.
 LE 3: No interv., Not Randomized, Retrospective,
Control Group.
 Example: Non-use of bicycle helmets and risk of fatal
head injury: a proportional mortality, case-control
study. Persaud N, et al. CMAJ. 2012 Nov
20;184(17):E921-3.
Exposure Outcome

27.  Group exposed to condition or received treatment
(exposure group) and group without condition
(control group) compared and followed up over time
for different outcome between groups.
 Level 2 or 3: No inrev., Not Randomized, Prospective
(usually) or Retrospective (rare), Control Group.
 Sleep Quality, Sleep Duration, and the Risk of
Coronary Heart Disease: A Prospective Cohort Study
With 60,586 Adults. Lao, X., et al. (2018). Journal Of
Clinical Sleep Medicine, 14(1), 109-117.
Exposure Outcome

28.  Observation of a defined population at a single
point in time.
 Exposure and outcome are determined
simultaneously.
 Level 4: Epidemiological ‘Snapshot’, No
Control Group.
 Value: prevalence numbers and correlations.

29.  Researcher interfere with one group (intrervention group)
and not other (control group) and follow up difference in
outcome.
 Value: most valid, determine causality.

30.  A controlled clinical trial that randomly (by
chance) assigns participants to 2 or more groups
and intervention (exposure) to one group and
follow up outcome. Usually double-blinded.
 High validity.
 Difficult, expensive and time-consuming.
 Example: Meditation or exercise for preventing
acute respiratory infection: a randomized
controlled trial. Barrett B, et al. Ann Fam Med.
2012 Jul-Aug;10(4):337-46.
intervention Outcome

31. Blinding
Randomization
Control

32.  combining data from many different research
studies (statistical process).
 Most valid (highest LE) is meta-analysis of
RCTs.
 Example: Anxiety outcomes after physical
activity interventions: meta-analysis
findings. Conn V. Nurs Res. 2010 May-
Jun;59(3):224-31.

33.  A summary of the clinical literature.
 critical assessment and evaluation of all research
studies that address a particular clinical issue.
 systematic review may also include a
quantitative pooling of data (meta-analysis).
 Example: Complementary and alternative
medicine use among women with breast cancer:
a systematic review. Wanchai A, Armer JM,
Stewart BR. Clin J Oncol Nurs. 2010
Aug;14(4):E45-55.

34.  Animal Research Studies
 Studies conducted using animal subjects.
 Example: Intranasal leptin reduces appetite and
induces weight loss in rats with diet-induced
obesity (DIO). Schulz C, Paulus K, Jöhren O,
Lehnert H. Endocrinology. 2012 Jan;153(1):143-
53.
 Test-tube Lab Research
"Test tube" experiments conducted in a
controlled laboratory setting.

36.  Put forth by experts in the field.
 Example: Health and health care for the 21st
century: for all the people. Koop CE. Am J
Public Health. 2006 Dec;96(12):2090-2.

38.  Journal, Title, Author(s) and Abstract
 Body of paper (I M R A D - format)
1. Introduction,
2. Methods (Materials and Methods),
3. Results and
4. Discussion
5. Conclusion: considered part of discussion.
 End of paper
5. Conflict of interest.
6. References.

40.  Journal
 Title: concise, informative and attractive.
 Author(s): names with academic
qualifications.

41.  Abstract: brief summary of the research paper
(usually 150 – 250 words) including:
1. Introduction (background)
2. aim (objective) (question?),
3. methods summary,
4. results summary
5. Conclusion.

43. Abstract reading determine if paper is of interset
to you to continue reading, but abstract alone
is never enough as it won’t reveal the strengths
and weaknesses of the research.

44.  Relevance (importance).
 Review of recent literature in this subject:
importance and limitations with references.
 Aim (objective) of study: (the question the
author is trying to answer).

45.  Study Design: RCT, cohort, case series, case
report, prospective vs. retrospective.
 Setting: type of hospital / location.
 Patients: number (sample size), inclusion
criteria and exclusion criteria.
 Factors: Description of prognostic factors
considered.
 Statistical Methods (data analysis):
 Outcomes: expected.
 Follow-up: duration, completeness.

47.  Was the patient sample clearly defined,
representative of clinical practice, and captured
at a similar point in disease progression?
 Was duration of follow-up sufficient? Were all
patients accounted for?
 Were outcome criteria objective and unbiased
relative to the prognostic factors?
 Was their adjustment for linked prognostic
factors?
 Were patients in the study treated similarly?
 Do the study population characteristics describe
your patient?

48.  Finding presentation objectively (without
interpretation).
 Statistical results: cases number, mean,
variation, significant or insignificant results.
 Figures.
 Tables.

51. Null hypothesis is correct.
Null hypthesis is accepted.
 Null hypothesis has to be assumed correct
before accepting or rejecting it.

52.  Probability is likelihood of event occurring in
proportion to total number of possibilities.
 Probability ranges from 0.0 (never happens) to
1.0 (certain to happen).
 p value (probability value): probability of
obtaining result by chance assuming null theory
is correct. (Calculated using statistical tests).
 0.05 (result by chance = 1/20) is accepted
threshold for statistical significance.
 P value < 0.05 (<1/20) is significant (reject null
theory).
 P value > 0.05 (>1/20) is insignificant (accept null
theory).

54.  Statistical significance : intervention effect
can't be explained by chance only.
 Clinical significance : how effective is the
intervention or treatment in real life.
 A study that is found to be statistically
significant may not necessarily be clinically
significant (relevant).
 This can be due to bias in sample size or
statistical tests.

56. (TP fraction of All diseased) (TN fraction of All non-diseased)
By gold standard

57.  Ability to identify those with the disease (TP of
all diseased).
 Highly sensitive tests used as screening test
(rule out) disease. Good –ve test.
 If +ve, a more specific test is needed.
 Independent on prevalence of disease.
 a test with 90% sensitivity will correctly return a
positive result for 90% of people who have the
disease (TP), but will return a negative result
(FN) for 10% of the people who have the disease
and should have tested positive.

58.  Ability to identify those without the disease (TN
of all non-diseased).
 Highly specific test used as diagnostic test (rule
in) disease. Good +ve test.
 Independent on prevalence of disease.
 a test with 90% specificity will correctly return a
negative result for 90% of people who don’t
have the disease (TN), but will return a positive
result (FP) for 10% of the people who don’t have
the disease and should have tested negative.

59.  Positive PredictiveValue (PPV): ability of
test to detect presence of disease (TP of all
+ve tests).
 Directly related to prevalence of disease.
 Negative PredictiveValue (NPV): ability of
test to detect absence of disease (TP of all
+ve tests).
 Inversely related to prevalence of disease.

62. By gold standard
Prevalence = 30 / 2030
= 0.0147

63.  Explanation of findings.
 Interpretation of findings in light of other
available literature and/or current standard-of
care.
 Evaluation of study (Validity): Strengths and
Weaknesses of Study.
 Recommendations: if results can change
management or counseling of the patient?
 Suggestions for future: Next steps for further
study of this problem.

64. 

66.  Financial:
 Sponsors:

67.  the last section in the paper.
 References are formatted in the journal’s
specific format.
 Citation is abbreviated alpha-numerical
expression (in the text) that refers to a
detailed reference on the reference section.

68. Vancouver Style: standard in papers
 Citation (in text): number in brackets or superscripted
 Reference List/Bibliography: Sorted Numerically, in
the order in which they appear in the text

69.  Harvard Style: standard in essays
 Citation (in text):
 Reference List/Bibliography: Sorted Alphabetically

70.  In short, a fast way to review an article is to read
1.Title.
2. Abstract.
3. Scan the Methods section for data source and
design
4. Review tables and charts in Results section. If
these are unclear, go to 5 then back to results.
5. Read first few paragraphs of the Discussion
section, often summarizing the results.
 6. Read conclusion.
 7. if wish or if unclear, re-read the entire article.