Evaluation of an online Diabetes Needs Assessment Tool (DNAT) for health professionals: a randomised controlled trial.

Abstract

BACKGROUND: Continuous medical education is traditionally reliant to a large extent on self-directed learning based on individuals' perceived learning priorities. Evidence suggests that this ability to self-assess is limited, and more so in the least competent. Therefore, it may be of benefit to utilise some form of external assessment for this purpose. Many diabetes educational programmes have been introduced, but few have been assessed for their benefit in a systematic manner. As diabetes is an increasingly prevalent disease, methods for the dissemination and understanding of clinical guidelines need to be explored for their effectiveness. This paper describes the study design of a randomised controlled trial to evaluate the effectiveness of using an interactive online Diabetes Needs Assessment Tool (DNAT), that builds a learning curriculum based on identified knowledge gaps, compared with conventional self-directed learning. The study assesses the effect of these interventions on health professionals' knowledge of diabetes management, evaluates the acceptability of this process of learning and self-reported changes in clinical practice as a result of this novel educational process. METHODS: Following a baseline assessment, participants will be randomised to undergo a 4-month learning period where they will either be given access to the diabetes learning modules alone (control group) or a Diabetes Needs Assessment Tool (DNAT) plus the diabetes learning modules (intervention group). On completion of the DNAT, a personalized learning report will be created for each participant identifying needs alongside individualised recommendations of the most appropriate learning modules to meet those requirements. All participants will complete a Diabetes Knowledge Test before and immediately after the allocated learning and the primary outcome will be the state of knowledge at 4 months. Learners will also be surveyed immediately after the learning period to assess the acceptability of the learning formats and the perceived usefulness and usability of the materials. After a further month, all learners will receive a series of questions to evaluate self-reported changes in clinical practice as a result of this educational experience and asked to include specific examples of any changes in their diabetes care practice. TRIAL REGISTRATION: Current Controlled Trials ISRCTN67215088.

Full text

Page 1

BioMed Central
Page 1 of 8
(page number not for citation purposes)
Trials
Open Access
Study protocol
Evaluation of an online Diabetes Needs Assessment Tool (DNAT)
for health professionals: a randomised controlled trial
Sara Schroter*1, Dean Jenkins2, Rebecca Playle3, Kieran Walsh4,
Courtenay Probert5, Thomas Kellner6, Gerhard Arnhofer7 and David Owens8
Address: 1BMJ Editorial Office, BMJ Group, BMA House, Tavistock Square, London WC1H 9JR, UK, 2BMJ OnExamination, Cardiff Medicentre,
Heath Park, Cardiff, CF14 4UJ, UK, 3South East Wales Trials Unit, Department of Primary Care & Public Health, Cardiff University, 7th Floor,
Neuadd Meirionnydd, Heath Park, Cardiff, CF14 4YS, UK, 4BMJ Learning, BMJ Publishing Group, BMA House, Tavistock Square, London WC1H
9JR, UK, 5BMJ OnExamination, Cardiff Medicentre, Heath Park, Cardiff, CF14 4UJ, UK, 6MSD RBSC GmbH, Richard Reitzner Allee 1, 85540 Haar,
Germany, 7Merck Sharp & Dohme Limited, Hertford Road, Hoddesdon, EN11 9BU, UK and 8Diabetes Research Unit, 1st Floor Academic Centre,
University Hospital Llandough, Penlan Road, Penarth, CF64 2XX, UK
Email: Sara Schroter* - sschroter@bmj.com; Dean Jenkins - djenkins@bmjgroup.com; Rebecca Playle - playlera@cardiff.ac.uk;
Kieran Walsh - KMWalsh@bmjgroup.com; Courtenay Probert - cprobert@bmjgroup.com; Thomas Kellner - thomas_kellner@msd.com;
Gerhard Arnhofer - gerhard_arnhofer@merck.com; David Owens - owensdr@Cardiff.ac.uk
* Corresponding author
Abstract
Background: Continuous medical education is traditionally reliant to a large extent on self-directed learning
based on individuals' perceived learning priorities. Evidence suggests that this ability to self-assess is limited, and
more so in the least competent. Therefore, it may be of benefit to utilise some form of external assessment for
this purpose. Many diabetes educational programmes have been introduced, but few have been assessed for their
benefit in a systematic manner. As diabetes is an increasingly prevalent disease, methods for the dissemination and
understanding of clinical guidelines need to be explored for their effectiveness. This paper describes the study
design of a randomised controlled trial to evaluate the effectiveness of using an interactive online Diabetes Needs
Assessment Tool (DNAT), that builds a learning curriculum based on identified knowledge gaps, compared with
conventional self-directed learning. The study assesses the effect of these interventions on health professionals'
knowledge of diabetes management, evaluates the acceptability of this process of learning and self-reported
changes in clinical practice as a result of this novel educational process.
Methods: Following a baseline assessment, participants will be randomised to undergo a 4-month learning period
where they will either be given access to the diabetes learning modules alone (control group) or a Diabetes Needs
Assessment Tool (DNAT) plus the diabetes learning modules (intervention group). On completion of the DNAT,
a personalised learning report will be created for each participant identifying needs alongside individualised
recommendations of the most appropriate learning modules to meet those requirements. All participants will
complete a Diabetes Knowledge Test before and immediately after the allocated learning and the primary
outcome will be the state of knowledge at 4 months. Learners will also be surveyed immediately after the learning
period to assess the acceptability of the learning formats and the perceived usefulness and usability of the
materials. After a further month, all learners will receive a series of questions to evaluate self-reported changes
in clinical practice as a result of this educational experience and asked to include specific examples of any changes
in their diabetes care practice.
Trial registration: Current Controlled Trials ISRCTN67215088
Published: 30 July 2009
Trials 2009, 10:63doi:10.1186/1745-6215-10-63
Received: 5 June 2009
Accepted: 30 July 2009
This article is available from: http://www.trialsjournal.com/content/10/1/63
© 2009 Schroter et al; licensee BioMed Central Ltd.
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0),
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Page 2

Trials 2009, 10:63http://www.trialsjournal.com/content/10/1/63
Page 2 of 8
(page number not for citation purposes)
Background
Doctors need ongoing learning in order to stay up to date.
Therefore, they need a method of continuous professional
development that is convenient and practical which ena-
bles them to meet their personalised needs. Online learn-
ing provides learners the opportunity to personalise their
learning, for example, the order in which they read mate-
rial, the speed of their learning, where and when they
choose to learn.[1] In agreement with previous reviews,
[1-3] a more recent systematic review and meta-analysis
[4] of the effects of internet-based learning involving
health professionals in a wide range of learning contexts
and clinical areas reaffirmed that internet-based learning
can be as effective as more traditional methods of learn-
ing. Major positive effects were found compared with no
intervention for both knowledge outcomes, skills, learner
behaviour and patient effects. This suggests that future
research should focus only on direct comparisons
between the different types of internet based interven-
tions.
Continuous medical education is traditionally reliant to a
large extent on self-directed learning based on the individ-
ual's perceived priorities for learning. Evidence suggests
that this ability to self-assess is limited, and more so in the
least competent.[5] There is also evidence that practition-
ers prefer to focus on topics which they are already famil-
iar with and avoid less familiar topics.[6] Current thought
is that interactive CME activities that help individuals rec-
ognise a "teachable moment" and an opportunity for
learning can help encourage engagement in the learning
and motivate learners to pursue the activity.[7] While
there are methods to help you discover your knowledge
and skills gaps, there is no one definitive methodology.
Using a variety of methods often gives a better overall
view.[8] Therefore it may be of benefit to have some form
of formal external assessment.[9] Exclusive reliance on a
formal needs assessment might discourage creativity and
professionalism,[10] but it could be an important part of
the process.
Miller [11] has defined the ideal stages of the develop-
ment of physician skills. The first level is "declarative
knowledge" i.e. the acquisition and interpretation of facts.
This is followed by "procedural knowledge" where indi-
viduals can describe how to do something but may not be
able to actually do it. The third level is about demonstrat-
ing that you know how to do what you have learned, i.e.
developing "competence". The fourth level involves indi-
viduals using the competence they have developed in
their actual clinical practice ("performance"). Moore et
al's[7] conceptual framework of an approach to continu-
ous planning and assessment in continuing medical edu-
cation builds on Miller's[11] model to incorporate an
assessment on improvements in patients' health status
and finally the health status of the population. However,
Curran and Fleet [12] reviewed the types of outcome
measures used in evaluative research in web-based CME
and found that most research has focused on participants'
satisfaction and not on change in clinical practice or
impact on patient and health outcomes. While assessing
the effect on patient's health status can be hampered by
ethical constraints in gaining access to data,[7] assessing
whether learners have tried to use their competence in
practice seems crucial in an assessment of the effectiveness
of a learning intervention.
As diabetes is an increasingly prevalent disease, methods
for the dissemination and understanding of clinical guide-
lines need to be explored for their effectiveness. In educa-
tional research on diabetes much of the focus has been on
the mode of delivery of the education with mixed results
from traditional method of small group teaching to larger
structured learning programmes.[13,14] Internet-based
interventions in diabetes have been shown to increase
guidelines knowledge[15,16] and compliance.[17] Many
education programmes have been examined but few in a
systematic manner and across wide geographic areas
involving large numbers.
There is no previous randomised controlled trial (RCT)
evaluating the effectiveness of different formats of admin-
istering online learning materials in the field of diabetes
on knowledge and subsequent change in clinical practice.
Therefore, we have designed an RCT of two different for-
mats of online learning for health professionals, i.e. an
interactive online learning tool with access to relevant
learning modules (active intervention group) compared
with access to relevant learning modules alone (control
group) to assess effectiveness in terms of diabetes knowl-
edge, satisfaction and self-reported change in clinical prac-
tice. We describe here the design of this randomised
controlled trial.
Methods
Design
This is a randomised controlled trial to evaluate the effec-
tiveness of using an online interactive learning tool, Dia-
betes Needs Assessment Tool (DNAT), in conjunction
with a learning management system (LMS) to improve
health professionals' knowledge of how to manage diabe-
tes; to evaluate the acceptability of this process of learn-
ing; and self-reported changes in clinical practice as a
result of this educational process.
Participants
Volunteers will be recruited through targeted emails and
advertisements inviting English and German speaking
practicing doctors and nurses to take part in an educa-
tional research project. Interested individuals will be

Page 3

Trials 2009, 10:63http://www.trialsjournal.com/content/10/1/63
Page 3 of 8
(page number not for citation purposes)
asked to complete an eligibility questionnaire online (see
inclusion/exclusion criteria) and those deemed eligible
will be given the opportunity to register for the trial. We
have produced two parallel websites, one for English
speaking participants and the other for German speaking
participants. Participants are able to choose whether they
want to receive the materials in English or German.
Participants will be informed that the time spent on learn-
ing activities for the project will contribute to their Con-
tinuing Professional Development (CPD) requirements
and that they will receive a personalised certificate of
learning on completion of the course. Participants will be
told that they will receive their test results and correct
answers, and access to the most effective learning package
at the end of the study free of charge. In addition, a choice
of access to one of three BMJ knowledge related products
(English speaking participants) and some QUAIME AG
online learning modules (German speaking participants)
will be provided.
Inclusion and exclusion criteria
On registering for the study, potential participants will be
asked a series of screening questions ensuring that only
eligible participants are recruited i.e. participants must be
either English or German speaking practising doctors or
nurses managing at least one patient with diabetes per
week. Note – German nurses, except diabetes nurses and
assistants for family physicians, will be excluded from tak-
ing part as they are not licensed to receive training where
information about pharmacological
included.
treatment is
Ethical approval and trial registration
The Research Ethics Committee for Wales confirmed that
the study does not require full ethical review (personal
communication 07 January 2009). The research will be
carried out in compliance with the Helsinki Declaration.
All participants will be asked to give their consent to take
part in registration. The trial was registered with Current
Controlled Trials before any participants were recruited
(Current Controlled Trials ISRCTN67215088). The study
adheres to the Good Publication Practice guidelines for
pharmaceutical companies.
Assessments and procedures
All assessments, learning materials, and surveys will be
administered to participants online at the study website.
Participants will be asked to give their consent to take part
in a randomised controlled trial designed to test the effec-
tiveness of different learning materials and methods and
will be informed that they will be required to use the
online learning materials for at least five hours during the
4-month learning period. Figure 1 is a schematic represen-
tation of the key components of the trial.
Once registered, participants will be asked to complete the
first Diabetes Knowledge Test before being randomised to
a study group. At the start of the 4-month learning period
they will be directed to their randomly allocated learning
materials. Learners in both groups will receive the same
six automated reminders to look at the learning materials
during the learning period (4, 8, 10, 12, 14, and 16 weeks
after the start of the learning period). The intervention
group will not be reminded to look specifically at the
DNAT during the learning period.
Immediately after the learning period, access to the learn-
ing materials will be closed and all learners will be asked
to complete the second Diabetes Knowledge Test (Test B)
and the Acceptability Survey. After a further month, all
learners will be asked to complete the Practice Change
Survey. The administration of these tests and surveys will
be automated so all learners receive them at the same
interval after receiving the learning materials. Non-
responders will be emailed a series of reminders to com-
plete the test and survey. Learners who do not complete
the Diabetes Knowledge Test B and/or the Acceptability
Survey will still be followed up and asked to complete the
Practice Change Survey.
Data collection and handling
Baseline data will be collected automatically when partic-
ipants register online for the study. The following infor-
mation will be recorded in the database: health
professional role (doctor or nurse), preferred language to
receive study materials (German or English), years since
qualification, baseline Diabetes Knowledge Test score and
whether they are registered users of the webservice uni-
vadis® and/or BMJ Learning. Baseline information will be
used to check comparability between study arms. Diabe-
tes Knowledge Test B scores and follow-up questionnaires
will be automatically collected online. Anonymised and
blinded data will be exported in a format appropriate for
statistical analysis.
Interventions
Participants will not be told which group they have been
allocated to. All participants will be given access to the
same online Diabetes Learning Modules (see below) on a
Learning Management System (LMS). An LMS is a techni-
cal platform which tracks all user behaviour including
how long and how often materials are used, and also
ensures that learners follow the materials in a fixed
sequence. In addition, the intervention group will be
administered a novel interactive learning tool, the DNAT.
A link tracking system will capture which resources indi-
vidual learners visit during the learning period. As the
learners progress with the modules they will also be
prompted to indicate which they completed and estimate
the time they spent using them.

Page 4

Trials 2009, 10:63http://www.trialsjournal.com/content/10/1/63
Page 4 of 8
(page number not for citation purposes)
Schematic representation of components of the intervention and assessment points
Figure 1
Schematic representation of components of the intervention and assessment points.
Time line Diabetes Learning Modules
(Control group)
Diabetes Learning Modules plus
DNAT (Intervention group)
Recruitment and enrolment
Baseline (time 0)
Randomisation
Months 1 to 4 {
Measurement of outcomes
Month 5
Month 6
a
a
b
b
c
dede
ff
Diabetes Knowledge Test. A
Diabetes Learning Modules
Diabetes Needs Assessment Tool (DNAT). A computerised adaptive test
plus feedback and tailored suggestions to do modules that are specific to
the user who completed the DNAT.
Diabetes Knowledge Test B
Acceptability Survey
Practice Change Survey
a
b
c
d
e
f

Page 5

Trials 2009, 10:63http://www.trialsjournal.com/content/10/1/63
Page 5 of 8
(page number not for citation purposes)
Diabetes Learning Modules
Participants in both groups have online access to the same
Diabetes Learning Modules on the LMS. All the content of
these learning modules is applicable to European practice
and material comes from BMJ Learning, Excerpta Medica,
a division of Elsevier Health Sciences, and the Interna-
tional Diabetes Federation and Elsevier Health Sciences.
The modules include current evidence-based guidelines
(particularly those of the European Association for the
Study of Diabetes and the European Society of Cardiol-
ogy) on diabetes, pre-diabetes and cardiovascular disease;
important clinical areas and common difficulties in prac-
tice (Type 1, Type 2, diabetes in pregnancy and secondary
causes of diabetes).
Diabetes Needs Assessment Tool (DNAT)
The DNAT is a new web-based diabetes interactive learn-
ing tool. It is a computerised adaptive test comprised of
clinically rich case problems developed by a group of dia-
betologists/educationalists. The 253 DNAT test items
have already been validated through use with a large
number of learners with a range of ability registered on
BMJ OnExamination. The items cover six categories: Prin-
ciples of Diabetes (53 items), Lifestyle (49 items), Drug
Treatment (64 items), CVS/Macrovascular (34 items),
Acute Complications (21 items), Microvascular Compli-
cations (32 items). The DNAT can be completed over sev-
eral sessions if required and takes the learner
approximately 90 to 120 minutes to complete. Not all
learners will see the same questions as the test adapts to
knowledge level as individuals complete it. On comple-
tion of the DNAT, a personalised learning report is created
for each learner identifying learning needs alongside indi-
vidualised recommendations of the most appropriate
Diabetes Learning Modules to meet those needs. At any
stage this personalised report can be viewed listing the
performance of the learner at that point.
Outcome measures
Primary Outcome measure
Diabetes Knowledge Test
The Diabetes Knowledge Test contains 20 multiple choice
questions (in the format of single best answer out of five
possible options). There are two versions of the Diabetes
Knowledge Test – Test A and Test B. All learners will
receive the same baseline Diabetes Knowledge Test (Test
A) prior to randomisation. Test B is a slightly different test
of equal difficulty to be administered at the end of the
learning period. The primary outcome will be Diabetes
Knowledge Test scores at four months. Note-while the
knowledge tests contain 20 items, only 19 items will be
used in the knowledge test scores as one of the items in
each test is a "rogue" item not on diabetes.
The Diabetes Knowledge Tests were developed from the
same large pool of calibrated items as the DNAT but there
is no overlap between the DNAT and the knowledge tests.
Secondary outcomes
Acceptability Survey
Immediately after the learning period, all learners will be
administered a survey containing a series of questions
about their level of satisfaction with the materials received
and the acceptability of the learning formats. They will be
asked about the perceived usefulness and usability of the
materials. Participants in the intervention group will be
asked additional specific questions about the acceptability
of the DNAT.
Practice Change Survey
Five months after the start of the learning period, all learn-
ers will be administered a series of questions on self-
reported changes in clinical practice as a result of the
learning experience. They will be asked to give specific
examples of any changes in their diabetes management.
This will be used as an indicator of the transference of
competence into clinical behaviour.
Randomisation and allocation concealment
Eligible registered participants who complete Diabetes
Knowledge Test A will be randomised to the control and
intervention groups of the study. Optimal allocation with
a ratio of 1:1 will be used. Randomisation will be bal-
anced for language (English or German), ability (based on
Diabetes Knowledge Test A scores), doctor vs nurse, and
number of years since qualified at medical or nursing
school, and whether they are registered users of the web-
service univadis® and/or BMJ Learning, using a minimisa-
tion technique.[18] The total sample of health
professionals recruited will be divided into blocks of 24
and within each block a process of optimal allocation will
be undertaken. This involves obtaining all possible alloca-
tions and calculating a balance statistic.[19] 1000 alloca-
tions with the greatest degree of balance will be identified
and passed to a statistician within the South East Wales
Trials Unit (SEWTU) at Cardiff University, (who is inde-
pendent of the trial) who will randomly select a single
allocation for each block. This will then be returned to the
trial statistician (RP) and the study database manager
informed of the allocations.
Blinding
All outcome measures are tests and self-report surveys and
will be administered automatically online. All data will be
held in the online database. The statistical analysis of the
trial data will be conducted by an independent statistician
at SEWTU blinded to the group allocation of participants
until the analyses are completed.

Page 6

Trials 2009, 10:63http://www.trialsjournal.com/content/10/1/63
Page 6 of 8
(page number not for citation purposes)
Translation of study materials
All study materials were originally developed in English
and subsequently translated into German. All recruitment
and marketing material, communication to research par-
ticipants, instructions on how to participate and website
information were translated by an independent company
and this was then reviewed by one of the bilingual mem-
bers of the study team. The independent companies were
selected by BMJ Group and contracted to do the work.
While these companies do regular work for the pharma-
ceutical industry, they do not regularly work for MSD. The
company did the work under their own editorial princi-
ples and MSD had no influence on the work that they pro-
duced. The DNAT and Diabetes Knowledge Tests
underwent a more rigorous multi-step process. Content
was translated by an independent company (from English
to German) and the translated text from this company was
then sent to a second independent company for proofing.
The text was then medically reviewed by bilingual mem-
bers of the research team and the texts finalised.
Sample size
We estimated that for a moderate effect size of 0.4 we
needed a minimum of 100 participants in each arm (at a
significance level of 0.5 and power of 80%). For an effect
size of less than that, e.g. 0.3 the sample size increases to
176 per group. These needed to be inflated to account for
drop out.
Since recruitment was to take place via online and print
advertisement and registration, recruitment rates, learning
material uptake and drop out were unknown, we decided
to allow up to 1000 participants per country to complete
the registration and baseline test until the cut off date for
study recruitment. Obtaining a larger sample size than
stated above will provide greater power for smaller effect
sizes and allow the country subgroup analyses to have
more power.
Statistical Analysis
All data will be checked during data cleaning for outlier
data. Range checks from the online data collection will
eliminate any out of range data. Any observed outliers will
be checked with the database management team. Inten-
tion to treat analysis will be carried out for the primary
analysis. Participants will be encouraged to complete Dia-
betes Knowledge Test B even if they have not spent as
much time learning as they had planned. Complete case
analysis will be used as a secondary analysis.
Summary statistics on participant assessment, eligibility,
recruitment, withdrawal and dropout will be collated for
both arms of the trial and form the CONSORT flow dia-
gram for clinical trial reporting. The primary analysis will
use Analysis of Covariance (ANCOVA) to compare the fol-
low-up diabetes knowledge test scores between trial arms
adjusting for baseline score as a covariate. This analysis
makes distributional assumptions about the data and
these will be checked prior to analysis. Variables used in
the balancing algorithm will also be considered for inclu-
sion as covariates. Secondary outcome analysis will com-
pare the acceptability and practice change outcomes
between trial arms. T-tests and Chi square tests will be
used where appropriate for questionnaire items.
Missing data
In the first instance missing follow-up Diabetes Knowl-
edge Test Score data (Test B) will be assumed to be miss-
ing at random. Baseline checks will be carried out to test
this assumption. Missing follow-up Diabetes Knowledge
Test Scores (Test B) will be assumed to have remained
unchanged for the intention to treat analysis. A complete
case analysis will also be carried out excluding those miss-
ing follow-up test scores. A sensitivity analysis will be car-
ried out to assess the difference between these analyses.
Subgroup analysis
Planned subgroup analyses involve the investigation of
the learning outcomes within language group. This will be
achieved by the addition of an interaction term to the pri-
mary analysis.
Trial organisation
The principal investigator will co-ordinate the trial with
the assistance of the Chief Information Officer (CIO) at
BMJ OnExamination and chair regular steering group
meetings with the research co-investigators. The co-inves-
tigators will regularly monitor the study's progress.
SEWTU will provide input at management group level for
the design and conduct of the study together with statisti-
cal input in the design of the websites to ensure quality
data collection. SEWTU will validate the data collection
and randomisation together with the CIO from OnExam-
ination. The CIO is responsible for the management,
recording, and storage of data. The study protocol is man-
aged centrally by the CIO and any protocol deviations will
be documented and reviewed by the steering group at reg-
ular intervals.
Data Quality Assurance
All study tests and surveys will be administered to partici-
pants online and the data collected centrally on an inde-
pendent platform. Participants will be asked for their
email address at enrolment and will be prevented from
enrolling more than once using the same email address.
This should prevent the majority of duplication. Shortly
after randomisation a check will be made for those that
enrolled with identical firstnames, surnames, and lan-
guage choice but different email addresses. The first regis-
tered email address that had a completed baseline

Page 7

Trials 2009, 10:63http://www.trialsjournal.com/content/10/1/63
Page 7 of 8
(page number not for citation purposes)
Diabetes Knowledge Test will be assigned as their account
and any other email addresses disabled. We will then
write to these participants to tell them that we noticed
they had registered more than once and that they should
use the selected email address from now on as we will dis-
able the other(s).
Publication Policy
We will submit the results of the trial for publication in a
biomedical journal irrespective of outcome.
Timetable
February to March 2009: Recruitment of learners. Regis-
tered learners gave consent and were administered Diabe-
tes Knowledge Test A
25th to 31st March 2009: Stratified randomisation
1st April – 31st July 2009: 4-month learning period
1st August 2009: Administration of Diabetes Knowledge
Test B and the Acceptability Survey
1st September 2009: Administration of the Practice
Change Survey
25th September 2009: Close of data collection
26th September 2009: Data to be given to independent
statisticians for analysis
Abbreviations
DNAT: Diabetes Needs Assessment Tool; BMJ Group: BMJ
Publishing Group; SEWTU: South East Wales Trials Unit;
Cardiff University
Competing interests
SS is a full time employee of the BMJ Group. She regularly
conducts research into all aspects of publishing for BMJ
Group and will not benefit financially from the outcome
of this study.
DJ is a Director of BMJ OnExamination. He is the Princi-
pal Investigator for the study and developed the DNAT
and Diabetes Knowledge Tests from test items generated
by BMJ OnExamination. He will not benefit financially
from the success of the DNAT if it is shown to be effective.
RP: No competing interests.
KW is editor of BMJ Learning and works for the BMJ
Group. He is paid a fixed salary.
CP is Chief Information Officer for BMJ OnExamination.
He developed the DNAT and Diabetes Knowledge Tests
from test items generated by BMJ OnExamination. He
also developed the websites for the study. He will not ben-
efit financially from the success of the DNAT if it is shown
to be effective.
TK is a full time employee of MSD RBSC GmbH. He
helped design the study and helped with the marketing
campaign to promote it to potential participants. He will
not have access to the data until after it has been analysed
by researchers at Cardiff University and then will only be
given access to anonymised data.
GA is a full time employee of Merck Sharp & Dohme Lim-
ited. He helped design the study and helped with the set
up of the Learning Management System and the provision
of the online learning modules. He will not have access to
the data until after it has been analysed by researchers at
Cardiff University and then will only be given access to
anonymised data
DO has conducted lectures, nationally and internation-
ally and served as a consultant for a number of pharma-
ceutical companies for which he has received honoraria.
He will not benefit financially from the success of the
DNAT if it is shown to be effective.
Authors' contributions
SS formulated the research question, designed the study,
helped with trial management and wrote the first draft of
this manuscript. DJ is Principal Investigator and guarantor
of this paper. DJ helped formulate the research question,
developed the DNAT and helped revise the manuscript.
CP developed the DNAT and the technology to run the
study and deliver the learning interventions, and contrib-
uted to the study design. RP contributed to the study
design, formulated the statistical analysis plan and helped
draft this manuscript. KW contributed to the study design
and helped revise the manuscript. TK contributed to the
study design, managed the recruitment plan for the Ger-
man participants, and helped revise the manuscript. GA
contributed to the study design and helped in the techni-
cal provision of the learning interventions and the Learn-
ing Management System. DO contributed to the study
design and helped revise the manuscript. All authors read
and approved the final manuscript.
Funding
Merck Sharp and Dohme Regional Business Support
Center (MSD RBSC) GmbH (Germany) gave BMJ Group
£110,000 to help fund this study. All other expenses for
the study, except translation costs, were paid by BMJ
Group. While staff at MSD RBSC GmbH were involved
with designing the study and the recruitment campaigns
for learners, they will not have access to the trial data until
the end of the study and will not be involved in the statis-

Page 8

Publish with BioMed Central and every
scientist can read your work free of charge
"BioMed Central will be the most significant development for
disseminating the results of biomedical research in our lifetime."
Sir Paul Nurse, Cancer Research UK
Your research papers will be:
available free of charge to the entire biomedical community
peer reviewed and published immediately upon acceptance
cited in PubMed and archived on PubMed Central
yours — you keep the copyright
Submit your manuscript here:
http://www.biomedcentral.com/info/publishing_adv.asp
BioMedcentral
Trials 2009, 10:63http://www.trialsjournal.com/content/10/1/63
Page 8 of 8
(page number not for citation purposes)
tical analysis or interpretation of data. All statistical anal-
ysis will be performed by independent researchers at
Cardiff University and they will be blinded to group allo-
cation.
Acknowledgements
We thank Ms Nicola Rylett, Mrs Vanessa Edmonds, and Mr Edward Briffa
at BMJ Group for help with planning and coordinating the study. We also
acknowledge Dr Peter Posel at QUAIME AG for his early contributions to
the study design. We thank Ms Donna Fish at Medelect Limited for her
research assistance, and Dr Ben Carter at the South East Wales Trials Unit
(SEWTU) at Cardiff University for the randomisation and help with the
proposed data analysis. Thanks also to the technology and marketing teams
at BMJ OnExamination. Finally, we thank Dr Sturm of HBF – Unabhängiges
Institut für hausärztliche Bildung und Forschung for his help with the
recruitment of the German participants and reviewing translated content.
References
1.Ruiz JG, Mintzer MJ, Leipzig RM: The Impact of e-learning in
medical education. Acad Med 2006, 81(3):207-212.
2. Wutoh R, Boren SA, Balas EA: eLearning: a review of Internet-
based continuing medical education. J Contin Educ Health Prof
2004, 4(1):20-30.
3.Curran VR, Fleet L: A review of evaluation outcomes of Web-
based continuing medical education. Med Educ 2005,
39(6):561-7.
4.Cook DA, Levinson AJ, Garside S, Dupras DM, Erwin PJ, Montori VM:
Internet-based learning in the health professions: a meta-
analysis. JAMA 2008, 300(10):1181-96.
5.Colthart I, Bagnall G, Evans A, Allbutt H, Haig A, Illing J, McKinstry B:
The effectiveness of self-assessment on the identification of
learner needs, learner activity, and impact on clinical prac-
tice: BEME Guide no 10. Medical Teacher 2008, 30(2):124-45.
6.Sibley JC, Sackett DL, Neufeld V, Gerrard B, Rudnick KV, Fraser W:
A randomized trial of continuing medical education. N Engl J
Med 1982, 306(9):511-5.
7.Moore DE Jr, Green JS, Gallis HA: Achieving desired results and
improved outcomes: integrating planning and assessment
throughout learning activities. J Contin Educ Health Prof 2009,
29(1):1-15.
8.Walsh K: How to assess your learning needs. J R Soc Med 2006,
99:29-31.
9.Davis DA, Mazmanian PE, Fordis M, Van Harrison R, Thorpe KE, Per-
rier L: Accuracy of Physician Self-assessment Compared
With Observed Measures of Competence: A Systematic
Review. JAMA 2006, 296(9):1094-1102.
10.Grant J: Learning needs assessment: assessing the need. BMJ
2002, 324:156-9.
11.Miller GE: The assessment of clinical skills/competence/per-
formance. Acad Med 1990, 65(9 Suppl):S63-7.
12. Curran VR, Fleet L: A review of evaluation outcomes of Web-
based continuing medical education. Med Educ 2005,
39(6):561-7.
13.Gerstein HC, Redd SSK, Dawson KG, Yale JF, Shannon S, Norman G:
A controlled evaluation of a national continuing medical
education programme designed to improve family physi-
cians' implementation of diabetes-specific clinical practice
guidelines. Diabetic Medicine 2001, 16(11):964-9.
14.Perria C, Mandolini D, Guerrera C, Jefferson T, Billi P, Calzini V, Fio-
rillo A, Grasso G, Leotta A, Marrocco W, Suraci C, Pasquarella A:
Implementing a guideline for the treatment of type 2 diabet-
ics: results of a Cluster- Randomized Controlled Trial (C-
RCT). BMC Health Serv Res 2007, 7:79.
15. Stewart M, Marshall JN, Østbye T, Feightner JW, Brown JB, Harris S,
Galajda J: Effectiveness of case-based on-line learning of evi-
dence-based practice guidelines. Fam Med 2005, 37(2):131-8.
16.Wiecha JM, Chetty VK, Pollard T, Shaw PF: Web-based versus
face-to-face learning of diabetes management: the results of
a comparative trial of educational methods. Fam Med 2006,
38(9):647-52.
17.Weston CM, Sciamanna CN, Nash DB: Evaluating online contin-
uing medical education seminars: evidence for improving
clinical practices. Am J Med Qua 2008, 23(6):475-83.
Altman DG, Bland JM: Treatment allocation by minimisation.
BMJ 2005, 330:843.
Carter BA, Hood K: Balance algorithm for cluster randomized
trials. BMC Med Res Methodol 2008, 8:65.
18.
19.