Webbased data collection and processing in international medical research

1
Web-based data collection and processing in
international medical research
Centre for International Health Friday 9th
November 2001
2
EBM
Arild Bjørndal Signe Flottorp Atle
Klovning Medisinsk kunnskapshåndtering Flommen
av ny medisinsk kunnskap er vanskelig å håndtere.
Denne boka gjør oppgaven lettere, og oppøver
evnen til å stille spørsmål som er relevante for
klinisk og samfunnsmedisinsk praksis. Den viser
hvordan man kan finne pålitelig kunnskap, og
kritisk vurdere forskningsmetoder og -resultater.
Forfatterne viser hvordan man kan implementere ny
kunnskap i klinisk og samfunnsmedisinsk praksis.
Målgrupper Leger under spesialistutdanning,
leger i praksis og medisinstudenter. ISBN
82-417-1062-3
3
References and resources

• This presentation
• WEB www.medicalwebresearch.org/projectorg/cmwr.ht
  m
• Download www.medicalwebresearch.org/projectorg/cm
  wr.ppt
• The forms presented
• WEBSURVEY Sleep www.e-survey.dk/ndno/sleep_no/slee
  p_no.html
• WEBSURVEY Winter Depression www.e-survey.dk/ndno/S
  PAQ/spaq.html
• WEBEPI-UI www.e-survey.dk/ndno/ui/ui2.html
• Centre for Medical Web Research
• www.cmwr.org or www.medicalwebresearch.org
• Finding answers to questions in EBM
• www.uib.no/isf/people/atle/ebm.htm
• OECD CD-ROM of comparative health data

4
EBM
Best evidence
Clinical expertise
Patient values
5
EBM 1-2-3 overview

• 4-part question
• Patient, intervention, comparison, outcome
• 7-step search typical mistake to start off in
  Medline
• (www.uib.no/isf/people/atle/ebm.htm)
• Cochrane
• Secondary publications www.ClinicalEvidence.org
• ________________________________
• Medline
• Primary journals
• Guidelines
• Internet google, yahoo,
• E-post lists www.jiscmail.ac.uk
• 3-step critical appraisal
• intern validity, result, extern validity

6
EBM The 4-part question (PICO)
7
Why Centre for Medical Web Research
(www.CMWR.org)?

• a need for an experimenting web site with no risk
  of jeopardising the UiB site, and having control
  over the technology
• an idea of a virtual, web-based collaboration of
  researchers
• protocols should be open for scrutiny
• development of a technology bit for data
  collection and handling, since most available
  software is inadequate
• applying Evidence Based Medicine
• Validation of methods
• RCT data collection an follow-up
• Surveys, epidemiological studies
• Diagnostic procedures in developing countries
• knowledge-sharing of web research experience

8
Methods

• PROTo provide a web-based networking of
  researchers conducting
• Surveys
• RCT data collection
• Validation of diagnostic tests
• Model clinical decision-making
• PUBLICTo validate direct-to-web-users research
• Effect of static pointer placing on various
  hosting web-sites
• Effect of dynamic banner distribution of forms
  (run-of-sites, banner networks)

9
Protocol phase

• Setting up the protocol
• Formulating research questions
• Choosing design
• Blinding
• Concealment of allocation
• Triple blinding?
• Developing forms etc.
• Web-based sample size calculations
• Ethics and person data management

10
Conducting the study

• Avoid bias in presenting the form
• Presenting the form
• Static on a hosting web site
• Dynamically by means of banners
• Input control and feedback
• Sequential analysis

11

• Build your form
• Deploy your form to an ftp-location, containing
  the cgi-script
• Download from ftp-site and clean up the HTML-code
• Upload again
• Distribute your form
• Collect responses (pump/autopump)
• Use the built-in analyser
• Quantitative Univariate and bivariate data
• Qualitative Text field analysis
• Export data to Excel, Word, HTML, SPSS for any
  further analyses

12
Build
13
Defining structure
14
Deploy-collect responses-analyze
15
Deploy survey
16
Distribute as e-mail form
17
Distribute the link by e-mail
18
Place a survey banner on a hosting web sites
front page
19
Collect responses
20
Collect responses
21
Analyze

• Univariate
• Bivariate
• Text analysis
• One-click button to export to
• Excel
• Word
• HTML
• SPSS

22
Univariate analysis
23
2x2 table
24
Qualitative information
25
Web epidemiology  WEBEPI-UI

• A study comparing data from classic
  epidemiological research is being prepared to
  validate web-based epidemiological research
• Survey questions related to urinary incontinence
  are being used, and made comparable with data
  collected in a large epidemiological study in
  Norway (HUNT)
• Bristol Urological Institute have deloped a
  similar form we are going to validate

26
WEBEPI-UI
http//www.e-survey.dk/ndno/ui/ui2.html
27
Web RCT

• Randomised controlled trial of web-based vs.
  teacher guided EBM-course

28
Web security

• CIP Cookie and IP-controlled medical web
  research
• Cookies have many functions, and this research
  project aims at looking at different ways of
  using cookies for research purposes
• Practical, ethical and technical aspects will be
  addressed

29
Web security

• STORED Secure Transport Of REsearch Data
• One major concern has been secure transport of
  research data, and the assumption that any data
  stored online may be hacked or sabotaged.
• This project aims at developing a secure method
  of web-based handling of e-survey data.
• Emphasis is laid on web to e-mail, post-mail
  handling and storing offline.
• Security issues like log reporting, fault
  detection and hacker security will be addressed.

30
Web surveys

• Sleep
• This project aims at comparing self-reported
  sleep dysfunction with existing guideline
  criteria.
• This project has run, data have been collected,
  and are being analysed. (N414 in 4 days)
• Winter Depression
• The Winter depression projects was one of our
  first e-survey projects, testing a rather complex
  web survey.
• Data have been collected and will be analysed and
  published.
• One of the research questions related to this
  study was too see how complex web surveys could
  be performed.

31
Sleep
http//www.e-survey.dk/ndno/sleep_no/sleep_no.html
32
Winter Depression
http//www.e-survey.dk/ndno/SPAQ/spaq.html
33
EBM

• Finding ansers to questions in evidence-based
  medicine
• http//www.medicalwebresearch.org/ebm/index.html

34
HONcode approval
35
Ethics
36
Ethical guidelines 1 

• Centre for Medical Web Research (CMWR) is founded
  on the vision of creating a virtual researcher
  community of highest quality.
• CMWR aims to improve public health through
  medical research and teaching using methods in
  accordance with evidence-based medicine.
• All web resources at CMWR are made available for
  the use of collaborators at all times.
• CMWR will ensure that collaborators also aim to
  exchange knowledge and co-operation across by
  means of this Web.
• CMWR is legally and morally committed by the
  medical oath not to pass on information that is
  entrusted by patients/site visitors.
• Information or data relating to individual
  patients and visitors of the CMWR Web site,
  including their identity, is respected by this
  site.

37
Ethical guidelines 2

• The independence and integrity of collaborating
  researchers is of vital importance to CMWR.
• Research at CMWR must fulfil the highest medical
  and ethical standards.
• Any project funders of CMWR will always be
  clearly identified, and commercial or
  non-commercial organizations that have
  contributed, will be acknowledged.
• Any competing interests will be declared.

38
Ethical guidelines 3

• CMWR wishes to continually improve its quality
  through means generated by MWR.COM projects, and
  to provide funding for MWR.ORG projects not
  fundable in other ways.

39
Ethical guidelines 4

• Any information given by CMWR is not meant to
  replace the relationship between a patient/site
  visitor and his/her existing physician.

40
Ethical guidelines 5

• CMWR will claim ownership of data and the right
  to publish no matter what results may show. 

41
Models

• Virtual world-wide network of researchers
• Each researcher has access to the web-based form
  to enter data
• Direct-to-web-users
• Forms are placed in various media
• Banners are used to distribute the links to the
  forms

42
I Virtual world-wide network of researchers
43
CMWR Sleep project
II Direct-to-web-users

• e-survey to analyze sleep disorders among
  internet users

44
Aims

• This project will
• create a set-up for e-care (diagnosis and
  treatment) of patients with sleep disorders
• be a scientific research project, designed as a
  randomised controlled trial to validate this
  intervention
• There is great technology enthusiasm for
  web-based care, but lack of evidence of the
  efficacy/effectiveness of these web-based
  interventions according to a recent Cochrane
  overview of telemedicine
• Sleep disorders are very common, and often
  ill-diagnosed, and could easily be better
  diagnosed and treated

45
procedure

• questions based on SLKs national therapy
  guideline

46
Survey banner on a hosting web front page
47
SOL published it on their front page
48
and on the top of SOL helse
49
e-survey banner
50

51
An example

• Sleep disorders research

52
Sleep disorders medicine

• Sleep disorders medicine is a clinical specialty
  which deals with the diagnosis and treatment of
  patients who complain about disturbed nocturnal
  sleep, excessive daytime sleepiness, or some
  other sleep-related problem

53
Sleep medicine

• Sleep medicine is currently in a state of
  transition.
• Growing field sleep medicine has long relied
  upon the help and expertise of physicians who
  work in the field, but are not sleep specialists
• GPs should be able to diagnose a majority of
  sleep disorders on their own and then only send
  the more difficult cases to a sleep specialist
• Realities All people with sleep disorders should
  IDEALLY be diagnosed through a sleep clinic,
  staffed with sleep specialists and sleep
  technologists under their supervision, because
  the intricacies and subtleties of some sleep
  disorders can be completely missed without proper
  testing- the problem is capacity
• Ref. http//www.lis.uiuc.edu/course/fall1999/lis4
  33b/needs_assess_exem1.html

54
My story

• http//www.sleepnet.com/insomnia/messages/856.html
  
• Posted by Sakina Gerhard on December 06, 1998 at
  215011 I wanted to update you folks about my
  situation. I had a Stanford appt Friday
  afternoon, and went with my boyfriend to Palo
  Alto.
• Got the tangible results back from my (11, not 14
  day) wrist actigraph.
• Here's what it conclusively said (I got a big
  printout of the whole thing)
• I am sleeping. Only I am sleeping very lightly
  all the time.
• I am definately always only in stage one or two
  sleep.
• Never any deep sleep at all.
• Also, I have about 50 wake ups a night.
• To achieve stage one sleep takes me usually 45
  minutes.
• Usually I sleep a few minutes at a time in stage
  one.
• My wake ups are normally one and a half minutes
  long

55
Actiwatch
56
Actiwatch with reader
57
Testimonial cutting edge

• Cambridge Neurotechnology are pleased to announce
  that they have been awarded the prestigious
• Millennium Products status for their Actiwatch
  range of monitors by the British Government.
• This is in recognition of the use of cutting edge
  technology and design concepts in the development
  and manufacture of Actiwatch monitors
• http//www.camntech.co.uk

58
Technology example Information everywhere
using WAP
News
Diseases
Pollen
Tests
59
Internet Disease Monitoring
Asthma Pt.
Peak Flow-meter
WAP
E-record
GP monitoringand giving feedback
Online access
GP
60
Projects

• Actiwatch-based sleep disorder management
• sleep disorders
• PLMS
• Interactive sleep diary
• calculates sleep efficiency
• SPAQ
• Kappa agreement between
• self-reported winterdepression
• SPAQ
• HAD

61
Problems

• You need access to a computer and the Internet
• Internet security hazards
• Selection bias in direct-to-web-users studies
• Existing software not yet sufficient
• How could this be implemented in developing
  copuntries?
• ?

62
Advantages

• Once the Internet access is in place, time is
  saved for many aspects of research
• Easy export of data to multiple platforms
• Easy to conduct multi-national studies
• ?

63
Finale

• We should not be techno-optimists
• All technology should validated
• Web-based research solves some problems, but
  creates others
• The web is only a medium- good research still
  needs a solid scientific basis based on properly
  formulated research questions