The use of DRG for identifying clinical trials centres with high - - PowerPoint PPT Presentation

The use of DRG for identifying clinical trials centres with high recruitment potential: a feasability study

Philippe Aegerter1,2, Noelle Bendersky2, Thi-Chien Tran1,2, Jacques Ropers2, Namik Taright3 and Gilles Chatellier4

• 1. Univ. Versailles St-Quentin EA 2506, FRANCE
• 2. AP-HP, Hôpital A. Paré, Unité de Recherche Clinique, Boulogne, FRANCE
• 3. AP-HP, Département d’Information Médicale, Paris, FRANCE
• 4. AP-HP, Hôpital G. Pompidou, Unité de Recherche Clinique, Paris, FRANCE

MIE 2014 Istanbul - 02-09-2014

Rational Objectives Material and methods Results Discussion

Clinical Trials

The mean for assessing (internal validity) and enhancing standards of care (recommandations) But external validity requires

large sample of heterogeneous patients numerous and different centres

Reality is cruel

• ne half of trials achieved original recruitment target [Sully 2013]
• ne third to one half of centres did not recruit (or so few)

Having a good estimate of the sample size and of recruitment has important consequences

ethics validity cost, notably opening centres

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 2 / 15

Rational Objectives Material and methods Results Discussion

Clinical Trials Recruitment Support System

Why ?

correlation between number of treated patients with the real inclusion capacity [London 2013] centres with pre-trial recruitmennt assessment more often reach the target

How ?

cohort identification using electronic medical records hot topic : 28 systems in a 2011 review [Cuggia], 78 in 2014 [Köpke], EHR4CR initiative

But

very few CTRSS were assessed against a large number of trials most of them relied on proprietary HIS

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 3 / 15

Rational Objectives Material and methods Results Discussion

Aim

Developpement of a pre-recruiment system for estimation of the set of eligible patients

(not a real-time inclusion alert)

using commonly available data Feasability study involving a large number of trials Evaluation criteria

content coverage precision influence of trial characteristics

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 4 / 15

Rational Objectives Material and methods Results Discussion

Material : EHR

French national DRG database (PMSI)

each stay in each acute hospital standardised case record compiled at discharge extensive quality checks

Contains

demographics : gender, age, dates of stays, hospital id diagnoses : coded with ICD-10 procedures : CCAM, multiaxial classification treatments : some, if expensive no free-text privacy : stays of one patient are linked but anonymous at the national level

the playground : AP-HP

Paris university care institution = 25 acute care hospitals around 1.2 million discharge records a year

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 5 / 15

Rational Objectives Material and methods Results Discussion

Material : CR

Protocols eligibility criteria

inclusion : all are mandatory

• demographics, nosology (diagnosis, findings, scores), consent

non inclusion : one is enough

• demographics, organ failures, drugs contra-indications

Protocols under evaluation

multicentre (within AP-HP) various clinical areas : oncology, cardio-vascular, neurology, rheumatology, ... closed in 2010-2011, in order to compare estimated and observed recruitment

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 6 / 15

Rational Objectives Material and methods Results Discussion

Material : Protocols eligibility criteria

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 7 / 15

Rational Objectives Material and methods Results Discussion

Methods : the go-between

Formal representation of eligibility criteria

9 categories : demographic, medical history, clinical scores and symptoms, diagnoses, laboratory tests, procedures, treatments, adverse events and others corresponding sources and items in the database

• perators : boolean and choronological (delays)

Transcoding

performed by 2 professionals

• n a criteria standardised form

independance between data and program

Executable

SQL queries automaticaly generated from criteria form

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 8 / 15

Rational Objectives Material and methods Results Discussion

Methods : criteria standardised form

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 9 / 15

Rational Objectives Material and methods Results Discussion

Results : ressources

85 protocols library of precoded common situations time needed for transcoding 1 protocol : 1-2 days time needed for query on AP-HP DRG database : 1-2 hours

numbers are medians

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 10 / 15

Rational Objectives Material and methods Results Discussion

Results : content coverage

inclusion / non inclusion criteria : 60 % / 80 % translated most criteria were diagnostic criteria, 94 % translated

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 11 / 15

Rational Objectives Material and methods Results Discussion

Results : precision

strong reduction of eligible patients (mostly after 5-6 criteria) final number of records would allow manual verification but depending on clinical areas

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 12 / 15

Rational Objectives Material and methods Results Discussion

Discussion

What has been achieved

large feasability study good content coverage

What is still to be done

comparison with whole patient record → precision, sensitivity at the patient level

Limits

”low tech", rely on coder ability poor performance when scores or lab tests rely on structured data

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 13 / 15

Rational Objectives Material and methods Results Discussion

References

Cuggia M, Besana P, Glasspool D. Comparing semi-automatic systems for recruitment of patients to clinical trials. Int J Med Inform. juin 2011 ;80(6) :371388. Köpcke F, Prokosch HU. Employing computers for the recruitment into clinical trials : a comprehensive systematic review. J Med Internet Res. 2014 Jul 1 ;16(7) :e161. doi : 10.2196/jmir.3446. London JW1, Balestrucci L, Chatterjee D, Zhan T. Design-phase prediction

• f potential cancer clinical trial accrual success using a research data mart. J

Am Med Inform Assoc. 2013 Dec ;20(e2) :e260-6. doi : 10.1136/amiajnl-201 Sully BG, Julious SA, Nicholl J. A reinvestigation of recruitment to randomised, controlled, multicenter trials : a review of trials funded by two UK funding agencies. Trials. 2013 Jun 9 ;14 :166. doi : 10.1186/1745-6215-14-166.

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 14 / 15

Rational Objectives Material and methods Results Discussion

Aknowledgements

Thank you for your attention

CeNGEPS for funding F Willig, Clinical Research Dpt (DRCD) AP-HP, for help

MIE 2014 philippe.aegerter@apr.aphp.fr Pred-Inclus 02-09-2014 15 / 15