Manual for Handsearchers

Cochrane Wounds Group

Contents
Page
The Cochrane Collaboration 3
Eligibility criteria & classification of Trials 4
What is a Randoised Controlled Trial! "
Methods of allocation #
$ther potentially useful trials %
What not to include &
Handsearching Guidelines ''
Coding of Trials '4
What to photocopy '"
Chec(list for )ournal handsearchers '#
*ournal search report '%
Glossary of ters '+
2
The Cochrane Collaboration
The Cochrane Collaboration is an international organization that aims
to help people make well-informed decisions about healthcare by
preparing, maintaining and promoting the accessibility of systematic
reviews of the effects of health care interventions. It also aims to keep
these reviews up to date as new evidence becomes available.
To do this, reports of andomised Controlled Trials !CTs" are
collected from books, #ournals and conference abstracts.
These are found by thoroughly searching the medical literature both
electronically and by hand.
eports of CTs are$ -
%. sent out to reviewers so they can update reviews
2. added to the &ounds 'roup (pecialized egister
Aim of Handsearching
The aim is to identify all CTs and (ystematic eviews !()s" in the
medical literature. It is important not to e*clude a study that may
possibly be an CT.
+
Eligibility Criteria and Classification of Trials
What is an RCT?
,n CT is a trial of any aspect of healthcare including screening,
treatment, prevention etc, where the individuals !or other units"
followed in the trial were definitely or possibly assigned prospectively
to one of two !or more" alternative forms of health care using$ -
andom allocation - where random means that the participants in the
trial were assigned to different interventions by a chance process - they
have e-ual chances of being in one group as in another.
The units of randomisation may be individuals, groups !communities
or hospitals", organs !eyes", or other parts of the body !legs, arms".
Controlled Clinical Trial - CCT
These are trials using some form of -uasi-random allocation - such as
date of birth or case record number. These studies do not use true
randomisation but these they may be important and need to be recorded.
.
What is a Randomised
Controlled Trial?
One group of people


Control Intervention
Experimental
Intervention

Follow up Assessment
!/0- blinding"
Follow up Assessment
!/0- blinding"
1

2
Examples of methods of allocation - as they may be
described in the literature
RCT
3 Patients were randomly allocated to the two treatment groups using a
random numbers table
3 Patients were allocated to the test or control group using numbers in
sealed envelopes
3 Patients were allocated to the test or control group by use of a
random code
CCT
3 Patients were allocated to the test or control groups by their NHS
number - odd numbers were the control
3 Patients were allocated alternately to the the test or control group
Allocation methods not stated
3 Patients admitted to the study were split into two groups
3 Patients were allocated to the test or control group
3 All patients were given either treatment tablets or placebo tablets of
similar appearance
4
Other potentially useful trials
Trials which mention double,blinding or as(ing where neither the
participant or the assessor is aware of which groups they are in are
almost always CTs. ,n e*ample of this is where a trial involves the
use of a placebo.
Cross o-er studies - where the first intervention may be randomized or
-uasi-randomized should be included
(ome studies may not even mention the word random so all
prospective trials that mention a control group should be included.

,rticles or reports mentioning trials which are planned, ongoing
incomplete or have been abandoned should be included
.ysteatic Re-ie/s should also be recorded as they are potentially
useful sources of information about CTs relevant to the sub#ect area
being reviewed
5
67&,7$ -
The term random selection does not necessarily mean random
allocation - further evidence is needed
Where random selection OE! "OT mean random
allocation
One group of people
!ample of patients selected at random
Treatment given
Follow up Assessment
What not to include
Articles hich state that they are! -
+. historical control
.. population control
1. animal studies
OR
"escripti#e studies
2. Correlation .tudies - two variables are plotted to see if there is a
pattern, for e*ample, smoking patterns in population and rates of lung
cancer. This is not e*perimental - nothing is being allocated.
4. Case Report0.eries - people may be surveyed or given an
intervention 68T there is no control group
5. Cross,section0 Pre-alence .ur-ey , this is a survey of a population.
It is not e*perimental 9 no intervention is being allocated to the
populations
continued on following page:
OR
Obser#ational analytic studies
;. Cohort .tudies - which are long follow up studies of treated and
untreated groups. These are only of use if there is random allocation
at the start.
%<. Case Control .tudies - two or more different populations are
compared, for e*ample, cases of hypertension and controls of healthy
individuals. To confuse matters, sometimes the intervention for each
group is randomised - this then makes it two different randomised
controlled trials
$nly trials /hich in-ol-e li-ing huans are included
Handsearching - $uidelines
<= It is useful to photocopy the table of contents as a checklist
%= The whole #ournal should be searched page by page in order to
identify the ma*imum number of relevant reports
Articles
The layout of these is similar in many #ournals, in most cases there will be$
%%. a title
%2. an abstract or summary
%+. an introduction0background
%.. methods section
%1. results
%2. discussion
%4. conclusion
%5. acknowledgements
%;. references
1ou /ill rarely ha-e to read the /hole article
2<. first look at the title - if it mentions randomised-controlled trial it should
be photocopied. !(ee >What to Photocopy?"
2%. if not - look at the abstract0summary and then the methods - if it is still
not clear then continue to read until you are sure whether it meets the
eligibility criteria or not
22. check the references at the end of the article - look for any mention of
other CTs
Re#ie articles
2+. those describing CTs or CCTs should be photocopied
2.. also photocopy articles which are (ystematic eviews of the research
evidence
%etters
21. these may mention CTs, CCTs or (s - planned, complete or ongoing -
if they do, photocopy the whole letter
Abstracts
22. photocopy these if they meet the criteria
Conference &roceedings
24. may also mention CTs, CCTs or (s and if so, should be photocopied
'oo( Re#ies
25. check for any mention of CTs, CCTs or (s
!searching each #ournal issue should not take more than about 2< minutes -
some will take less time"
2f in doubt , photocopy3
The -uality of the study does not need to be #udged at this stage.
It is probably easier to mark each relevant page in each #ournal being
searched and then do all the photocopying at once
$n the photocopy4 ,
2;. @ighlight the evidence, i.e. where it states that this trial is, or may be an
CT or CCT.
+<. @ighlight the full citation, i.e. the #ournal name, year, month, volume,
issue and page numbers - in most #ournals these are already printed on
each page, if not, please add them
+%. In the top right hand corner of the first page, indicate whether the trial is
CT, CCT or (ystematic eview !(". If you are not sure or it is not
stated, classify this as (A( !second opinion sought"
+2. make sure photocopies are clear and readable
Coding of Trials
3 andomised Controlled Trial CT
3 Buasi-randomised Controlled Trial CCT
3 Cossibly andomised or Buasi-randomised
Controlled Trial CCT
3 (ystematic eview or Deta-analysis (
3 (econd Apinion (ought (A(
What to photocopy
The &ounds 'roup are developing a database of trials relevant to the
group. The scope is as follows$ -
2= Cressure (ores - treatment E prevention
+= Feg 8lcers - to include venous ulcers, rheumatoid ulcers, diabetic
foot ulcers
.= 6urns
++. Traumatic &ounds
+.. Dalignant &ounds
+1. (urgical &ounds
+2. (kin ,bscess
+4. 6ites !mammalian"
+5. (carring
(tudies pertaining to gastric or duodenal ulcers, corneal ulcers, se*ually
transmitted diseases, dental wounds, eye in#uries are not within the scope of
the 'roup.
+;. If items are found which mention trials which are within the &ounds
'roup scope, the full article, letter, review etc, should be photocopied
.<. If articles do not come within the &ounds 'roup scope, then the first
page and the part of the article showing that it may be an CT or CCT
should be photocopied. ,gain, highlight the evidence and make sure the
citation is complete.
.%. If it is a letter, photocopy the whole letter.
.2. If any of the references at the end of the article mention randomised-
controlled trials but the article does not, then #ust photocopy the
references, highlighting the relevant references.
Then!
.+. Gill in the Hournal (earch eport (heet, remembering to include the
month of the #ournal and also if there are missing #ournals or missing
pages in the #ournal
... Gill in the checklist so you know where you are up to with each #ournal
.1. (end in the photocopies along with the Hournal (earch eport sheet
Chec(list for )ournal handsearchers
.2. photocopy the table of contents
.4. e*amine each issue page by page to identify all reports of
CTs and CCTs and (ystematic eviews, remembering
they may be in articles, letters, news columns, editorials,
references, etc.
.5. if the item is within the scope of the &ounds 'roup,
photocopy the whole item, if not then #ust photocopy the
first page
.;. highlight the evidence - the key terms that show the item
is an CT, CCT or a possible CT
1<. ensure the full citation is available on all photocopies -
either highlight what is already printed or add to the copy
1%. code each item as CT, CCT, ( or (A( - in the right
hand corner of the first page
12. complete the Hournal (earch eport (heet - remembering
to include the month and any missing #ournals or pages
*ournal .earch Report 5or
.earchers 6ae4
Phone nuber4
*ournal Title4
1ear0Month4 7olue4
2ssue4 Pages4
Classifications4 enter the total number of trials or reviews found in the
ournal and supported by attaching the necessary photocopies
Rele-ant to Wounds
Group
$thers
RCT
CCT
.R
.$.
,ny comments from (earcher to &ounds 'roup$
,ny pages0parts missing from the Hournal$
Iate search completed and returned to &ounds 'roup$
G8$..9R1 : for reference

&ords or phrases marked by an = indicate that they are potentailly relevant trials.
Types of trial and ters relating to trial design /hich you ay find
useful /hen reading the )ournal3
;linding
In clinical trials this means that the patients and0or the investigators are not aware of
which treatment the patient is getting. (ee double blind and single blind.
Case <History= .tudy
, report on one or more people, often with an unusual disease.
Case,control .tudy
!,lso called case comparison study, case compare study, case history, case referent
study, and retrospective study=
Compares people with a disease !cases" with a group of people without the disease
!controls". 7*amines possible causes or risk factors for the disease.
7*ample$ If you were interested in finding out whether thalidomide caused birth defects
you could select a group of children with birth defects !cases" and a group of children
without birth defects !controls". Jou could then find the number of children in each
group that had been e*posed to thalidomide through their mothers taking the tablets.
This is sometimes called a retrospective study because it is looking back in time.
Clinical Trial >
!,lso called therapeutic trial, intervention study".
, trial which tests out a drug or treatment to see whether it is effective or safe. This
general term encompasses randomised controlled trials, controlled clinical trials,
randomised clinical trials, and other trials where there may or may not be a control
group.
Cohort .tudy
This type of study takes a group of people and follows them up over a period of time to
see who gets a disease or how a disease progresses.
7*ample$ If you were interested in risk factors for heart disease, you could follow a
group of people without heart disease for several years. Jou could see which people
developed disease and which people did not.
Control
@as several meanings which can cause confusion$
%" In clinical trials comparing two or more treatments, control refers to the coparison
group that receives one of the following$
a" a placebo
b" no treatment0care
c" usual form of care0treatment !e.g. the normal surgical operation for a hip
replacement, the normal e*ercise programme following a stroke"
d" another drug 0 treatment.
2" In case-control studies, the control refers to the comparison group /ithout the
disease
+" In studies of infectious diseases, control means programmes aimed at eliminating the
disease.
Controlled Clinical Trial>
efers to a study that compares one or more treatment groups to one or more
comparison !control" groups. &hilst not all controlled studies are randomised, all
randomised trials are controlled.
Gor the purposes of handsearching and coding, if there is doubt as to whether the
patients were randomly allocated treatments, code as CCT
Cross sectional .tudy
This is like taking a )snapshot) of a group of people at one point in time and seeing the
prevalence of certain diseases, for e*ample heart disease, in that population. Jou could
also try to find out how many !for e*ample" smoked, drank etc.
Cross,o-er Trial ?esign>
, method of comparing two or more interventions where sub#ects are randomly
allocated to receive one intervention and subse-uently receive the other!s". The order of
the interventions is random.
7*ample$ If comparing aspirin with a placebo, half of the patients would be randomly
allocated to receive aspirin first, followed by the placebo. The other half of the patients
would receive a placebo first, followed by aspirin. (ometimes all patients receive a
placebo, before they are randomly allocated to the treatments. (ee also washout period.
?ouble blind
Keither the patients nor the investigator are aware of the treatment the patient is being
given. The purpose of this is to reduce bias. , double blind trial is a usually a
randomised controlled trial or a controlled clinical trial.
Epideiology
The study of the distribution and determinants of disease fre-uency in the population.
2ncidence
This is the number of new cases of a disease in a population
2nter-ention .tudy>
(ee Clinical Trial
Parallel Group Trial ?esign>
These trials compare two or more interventions with sub#ects allocated to one of the
interventions only.
Phase 2 0 Phase 22 .tudies
These studies are concerned with drug safety, drug dosage and drug effectiveness. They
are not normally randomised or controlled, and therefore not of interest to us. !@owever
within a crossover trial the first period is sometimes referred to as Chase %."
Placebo
, harmless substance given as an alternative to the intervention e.g. a sugar pill. (hould
be identical to the intervention being tested i.e. the appearance, dosage, administration
instructions should be identical to the active drug.
Placebo effect
Catients report a favourable response to any treatment regardless of whether it is the real
thing, or a placebo.
Pre-alence study
Jou are trying to find out how many people have a particular disease in the population.
(ee prevalence and cross sectional study
Pre-alence
This is the number of e!isting cases of a particular disease at any point in time.
Prospecti-e
In clinical trials@ this means that patients are divided into groups before treatment is
given3 That is, the investigators plan in advance what treatments are to be given.
(ub#ects are then selected and treatments allocated, either randomly or non-randomly.
Auasi,rando
Dethods of allocating patients to a treatment which are not statistically rando
Buasi-random methods of allocation include$
Iate of birth
Iays of the week
Dedical record number
Donths of the year
Arder in which the patients are included in the study
Auasi,randoised Trial=
, trial using a quasi-random method of allocating patients to different forms of care.
Gor the purposes of handsearching and coding, code as CCT
Rando
'overned by chance
Rando selection
andom selection refers to how the group of patients was selected from a population. It
is not related to how the patient was allocated a treatment.
Rando allocation 0 assignent 0 Randoisation0
andom assignment implies that each individual has the same chance of receiving each
of the possible treatments. ,lso the probability that an individual will receive a
particular treatment is independent of the probability that any other sub#ect will receive
the same treatment.
Dethods of randomisation include$
6locks of ten
Coin tossing
Ieck of cards
Codes
Computer generated numbers 0 computerised 0Dinimisation
Cre-ordered list
andom number tables
emote randomisation
(ealed envelopes
(tratified randomisation
(ee also quasi random
Telephone randomisation
Randoised Controlled Trial>
, trial in which sub#ects are randomly allocated into groups, to receive, or not to receive
a treatment. This treatment could be a drug, a surgical procedure, a dressing, a diet, an
e*ercise programme etc. Can also compare two or more types of drug. (ee also
Control
Gor the purposes of handsearching and coding, code as RCT
Randoised Clinical Trial>
(ee randomised controlled trial
(hould be coded in the same way as a Randomised Controlled Trial
Retrospecti-e
Fooking back in time. In terms of collection of data, the treatment, or event has already
happened.
7*ample$ , study may compare two different drugs for cancer, by e*amining the
medical records of people who have had the two different treatments in the past. This
means that the patients cannot have been randomised to the two treatments because they
were administered at the time when the investigation begins.
Retrospecti-e study
(ee case-control study"
Re-ie/
This is an article that looks at a number of different studies and may draw conclusions
about a particular treatment. The review may look at several randomised controlled
trials and compare the results. It is not usually a randomised-controlled trial in itself.
Accasionally the author may report methods and results of a trial that he0she is involved
with.
Ris( 5actor
,n aspect of a person)s lifestyle or environment which increases the probability of
occurrence of a disease. Gor e*ample, cigarette smoking is a risk factor for lung cancer.
.ingle ;lind
The investigator is aware of the treatment the patient is getting, but the patient is
unaware.
Therapeutic Trial
(ee #linical $rial%
Washout period0phase
The stage in a crossover trial when treatment is withdrawn so that patients body returns
to normal.