1 of 29

Research viva

2 of 29

  • Searching a database
    • PubMed, Cochrane library
    • Search terms, MeSH terms
  • Evaluation of evidence
    • Meta-analysis
    • Trials
    • Applicability
    • Stats – RR, OR, HR, NNT, NNH
  • Audit
  • Recent key research

3 of 29

EBM – research question

  • Population
  • Intervention
  • Comparator
  • Outcome

​

4 of 29

Search engines

  • PubMed
  • Cochrane library
  • clinicalTrials.gov
  • SLJOL

​

  • EMBASE
  • SCOPUS
  • Web of science

5 of 29

PubMed search

6 of 29

7 of 29

MeSH (Medical Subject Headings)

8 of 29

MeSH

  • Vocabulary used in PubMed to describe the content of research paper for indexing purposes

​

  • Searching PubMed using MeSH terms give better focused result than using keywords

9 of 29

Cochrane

10 of 29

11 of 29

12 of 29

Types of articles

  • Reviews
    • Narrative
    • Systematic review +/- Meta-analysis
  • Research
    • RCT
    • Cohort
    • Case control
    • Observational
    • Case series / reports
  • Opinions, commentaries, viewpoints..
  • Guidelines

13 of 29

Accessing full text

  • Open access journals
  • HINARI
  • Institutional subscriptions
  • Journal subscriptions
  • Article subscriptions
  • Write to corresponding author!

14 of 29

Assessing quality of review

  • Methodology
    • Comprehensive search - Several databases, no language restriction, flow chart of search strategy
    • Data extraction – 2 authors
    • Details of primary studies – PICO
    • Quality of primary studies assessed
    • Appropriate summary statistics (heterogeneity)
  • Applicability - PICO
  • Reliable source –
    • Cochrane, high impact journal (NEJM, Lancet, JAMA network, Nature review, BMJ), citations
    • Vs predator journals (Beall’s list)
    • Renowned author

​

15 of 29

Assessing the quality of a trial

  • Question – PICO
  • Relevant population
  • Random allocation and Allocation concealment
  • Comparable groups
  • Equally treated
  • Blinding
  • Outcomes – clinically relevant
  • Results
    • Effect size
    • precision – Cis
    • All participants accounted for at the end – premature stopped, intention to treat (Vs per protocol)
    • – RR, OR, HR, NNT, NNH
  • Benefits Vs risks
  • Other aspects
    • Multi-centre / multi-national
    • Sample size – calculation
    • Appropriate inclusion, exclusion
    • Duration of follow up

16 of 29

Clinical trial phase

  • Pre clinical – animal
  • Clinical
  • I
  • II
  • III
  • IV

17 of 29

Bias in RCTs

  • Selection – allocation concealment, random sequence generation
  • Performance – blinding of patients and personnel
  • Detection – blinding of assessor
  • Attrition – incomplete outcome data

​

​

​

18 of 29

Real world data

  • From registries
    • CVD REAL
  • Non-trial environment
  • Reflects day to day practice
  • Effect on confounding factors

19 of 29

Key trials - DM

  • DCCT, UKPDS
  • ACCORD, ADVANCE, VADT (10 & 15y follow ups)
  • LEADER,
  • EMPA REG OUTCOME, CANVAS, DECLARE TIMI 58

20 of 29

Study (N, yr, FU)

Pt Fx (age, DM, A1c, BMI)

Intensive Vs standard

Outcome

Comments

DCCT, 1500, 1993, 7

< 40y, new T1DM. w/o HTN

7.2, 9.1

Micro 70% reduc

Young people, so macro were rare

EDIC, 1400, 2005, 11y

DCCT participants

~ 8.2% in both cohorts

Macro 42% reduc (57% for severe CVE)

 

UKPDS-33, 3900, 1998, 10y

(glib, glipi, metf, Insulin)

53, new, 7.1, 27.5

7, 7.9

Micro 25% reduc

Mostly retinopathy

BP arms

ateno = capto

144 > 154 for micro, macro, death

ADVANCE,

11 000, 2008, 5

(gliclazide 30-120)

66, 8, 7.5, 28

6.5, 7.3

Death, macro – no diff.

Micro 10% reduc, sp nephron

Small A1c difference

Rosi only 17%

BP arms (peri-inda)

136 > 142 for micro, macro, death

 

VADT, 700, 2008, 5.6

62, 12, 9.4, 32

6.9, 8.4

Death, macro, micro – no diff

Small N, high drop out

ACCORD, 10 000, 2009, 3.5

62, 10, 8.3, 32

6.5, 7.5

Death, macro – increased

Micro not assessed

High rosi (91% Vs 57%) & low ACEi in intensive arm.

More wt gain in intensive arm

BP : 119 V 133. No benefit except transient stroke reduc. More AKI. Underpowered, advanced disease.

LIPID- adding fenofibrate – no mortality reduction

21 of 29

Key trials - IHD

  • ISIS, GUSTO, ASSENT

​

22 of 29

Key trials - stroke

23 of 29

Key trials - HTN

  • SPRINT

​

  • TRIUMPH

​

24 of 29

Key trials – preventive medicine

  • ASPREE ARRIVE ASCEND
  • DPP

 

ASCEND

ARRIVE

ASPREE

Study people

DM

> 55y mod CVDr

>70y healthy

1ry end point

Composite fo CV death, MI, stroke, TIA..

Composite ..

All cause mortality

Outcome

Benefit NNT 110

Increased bleeding – NNH 91 (mostly GI)

No benefit

Major bleeding equal but Increased GI bleeding

Increase all cause mortality, mostly due to cancer related deaths

Limitations

Low PPI use

Low CV event rate

Need long term

25 of 29

Key trials - COPD

26 of 29

Key trials - sepsis

27 of 29

Key trials - NOACs

28 of 29

Lipids

 

Primary prevention

Secondary prevention

Statin

JUPITER, WOSCOPS

4S, LIPID, HPS, CARE, MIRACL, SPARCL

High > mod : PROVE IT TIMI 22, TNT

Ezetimibe

-

IMPROVE -IT

PCSK9i

 

ODYSSEY OUTCOME�FOURIER

other

Helsinki Heart Study - gemfibrozil

FIELD, REDUCE-IT

29 of 29