CTU surveyDr Oluseun AdeogunNIHR Research Methods Fellow in Health InformaticsUniversity Of Liverpool / North West Hub for Trials Methodology Research / MRC HeRC North / Farr Institute

23rd June 2014 1

Purpose of survey

• Aims: • To discover types of resources used by CTUs

across the country• Identify experiences of using electronic

databases to support clinical trials:• Feasibility assessment• Identifying and contacting eligible patients• Outcome data collection

• Questionnaire completed by 35 of 45 CTUs• Overall response rate of 78% 2

Feasibility assessments using routine data sources

Have you used electronic databases to estimate the

numbers of eligible patients?

Yes: 22 (63%)

No: 13 (37%)3

Distribution of electronic databases used to assess feasibility

4Lo

cal Gen

eral P

ractice

Electr

onic Hea

lth Reco

rd sy

stems

Local

Hospita

l Elec

tronic H

ealth

Record

syste

ms

CPRD GOLD/ G

PRD

QResearc

h

THIN data

baseHSC

IC

Other ele

ctronic d

atabase

(s)0%

5%

10%

15%

20%

25%

30%

Types of electronic databases

% o

f CTU

s who

hav

e us

ed e

lect

roni

c dat

abas

es

Snapshot: ways in which results were used

“To estimate the number of GP practices required to approach enough

potential study participants to meet target recruitment”.

“Estimate likely number of eligible patients. Estimate potential number of

sites.”

“Decided not to - data accuracy concerns”.

“…to estimate the likely number of people consulting with the condition of

interest over a specified time period. We are always very cautious to not take the figures from this feasibility exercise to represent the likely recruitment rate

as we cannot always apply our inclusion/exclusion criteria to this data

nor account for the consent rate”.5

Identifying and contacting eligible patients

Have you used electronic databases to identify eligible

patients for studies?

Yes: 19 (54%)

No: 16 (46%)6

7Contact data

providerContact GP Other Contact data

provider and GPContact data

provider, GP and other

Contact GP & other

Contact data provider and

other

0%

5%

10%

15%

20%

25%

30%

35%

40%

45%

50%

Ways in which CTUs contacted eligible patients%

of C

TUs

Options: Contact the data provider to act as a third party to contact patients | Contact the general practice surgeries who contacted the patients | Other

Categories: ways of approaching eligible patients

Sending letters of invitation/study

information by mail

10 (56%)

Health professionals approached patients at

clinics/GP surgeries

6 (33%)

Study publicised using several methods: social

networking, local media, public

engagement events

1 (5.5%)

Popups on GP electronic systems to

alert GP during consultation when

appropriate codes are entered

1 (5.5%)

8One response said that patients were not approached since it was a cluster trial

Outcome data collection using the EHR

Have you used electronic databases to obtain outcome

data in trials?

Yes: 25 (71%)

No: 10 (29%)9

Categories: Outcome data collection using the EHRLab Data• 4

Health Economics Data• 2

Death/Mortality Data• 9

Unscheduled Medical Contact

• 1

Cancer Data• 3

Pregnancy Data

• 1

Hospital Data• 6

Therapeutic Data

• 1

Not Specified• 5

10

Outcome data collection using the EHR

Was there another source of the same outcome data e.g.

paper CRF?

Yes: 18 (72%)

No: 7 (28%)

If Y (to previous question), was there a comparison made to

the electronic database?

Yes: 16 (89%)

No: 2 (11%)11

Comparisons of electronic outcome data versus data on paper

11 responses give details of data

comparisons made

12

These will be followed up to see how they compare with: Tudur Smith, C., Stocken, D.D., Dunn, J., Cox, T., Ghaneh, P., Cunningham, D. & Neoptolemos, J.P. (2012) 'The Value of Source Data Verification in a Cancer Clinical Trial'. PLoS ONE, 7(12).

Snapshot: Advantages of using electronic databases for trial feasibility and participant identification

Accuracy, ease of use

Allows large numbers of participants to be

identified and invited at low cost

Cheaper to use and less invasive of time of

participating centres

More reliable than surgeon guess on likely

numbers - surgeons over estimate!

Speed, once access obtained

Useful to judge feasibility of doing the

research

It has reduced workload of trial unit

staff13

Snapshot: Disadvantages of using electronic databases for trial feasibility and participant identification

Accessing some of the databases and potential

inaccuracies. Delays between data entry and

availability.

Difficult as real data may not give you exact eligible

population but an approximation based on what data are available.

Cost, especially for relatively small studies as the set-up

cost is high.

It is difficult to establish eligibility if data is not

complete.

Lack of granularity in data. Timeliness of data

Need to be careful with queries as easy to over- or

under- include

Restricted by the variables in the database…may often

not able to apply all eligibility criteria when

estimating patient numbers.

Routine data is not so clean as trial data

14

Summary

• Described the results of the survey• Highlighted the following:• Feasibility assessment• Identifying and contacting eligible patients• Outcome data collection• Summarised the advantages and

disadvantages of using electronic databases for trial feasibility

15

Acknowledgements

• Saeeda Bashir for disseminating the questionnaire• All the 35 CTUs that responded • Co-applicants of the grant:• Prof P. Williamson, Prof C. Gamble, Dr C. Tudur-

Smith, Prof M. Clarke, Prof J. Brown, Dr D.O’Reilly, Dr M. Lay, Dr A. Lane, Prof J. Carpenter, Prof H. Snooks, Prof I. Ford, Prof L. Smeeth

• MRC Hub for Trials Methodology Research Network for funding 16

Questions?

17