Transcript Slide 1
Lessons learnt on secondary usage of data from the UK CRC/CfH study …enhancing clinical research Crown Copyright 2007 General Practice Research Database John Parkinson BSc PhD ([email protected]) tel.020 7084 2698 Crown Copyright 2007 Background ▐ Connecting for Health ▐ ▐ ▐ ? The worlds largest IT project Many billions of £ Creating full e-health ▐ ▐ ▐ ▐ ▐ All data held electronically Do it once and record…share Clinically shared as appropriate Make data available for research ▐ Public health ▐ Academic ▐ Commercial e-Health Record Crown Copyright 2007 UKCRC-CfH Simulations ▐ ▐ ▐ ▐ Surveillance Clinical Trials Record Linkage Biobank- Genetic Crown Copyright 2007 European Interoperability Primary use Makes good sense 1. Transfer of FULL record to new doctor in NEW SAME country 2. Emergency useSUMMARY HR across Across Europe the UK Secondary use Some big questions NSAIDs Genetics Small Rx in each country Differences in HC delivery Different doses Different usage patterns Cluster trials Crown Copyright 2007 European Interoperability ▐ In particular, a major benefit of eHealth interoperability lies in improving patient safety. Crown Copyright 2007 Crown Copyright 2007 UKCRC-CfH Must haves- Data Sources- Medicine surveillance 2/3 o care Primary care Hospitalised Clinic Day casescancer haematology NHS Dispensed ETP data Pharmacy Rx Pharmacy care Other care NHS Contracted out care Herbal and other Treatments Non-NHS Self- care via Professional care HealthSpace Dentist Rx Walk in centres Other services OTC medications Aspirin/ paracetamol Crown Copyright 2007 UKCRC-CfH Must haves- What data -GP Coded Patient ANON ID SED001 Symptom SED002 Examination SED003 Diagnosis SED004 Intervention SED005 Management SED006 Administration SED007 Presenting BP code 30,000 lines= 1 file ex 9, 1 practice for 3 months =5 Text BP readings thousand million lines of data/year Crown Copyright 2007 UKCRC-CfH Auto-populate Must haves- e-Yellow Card Patient CVR459 sex age Drug xxxxxxxxx route dose date Event xxxxxxxxx date Outcome xxxxxx Other drugsxxxxx Other information NHSnet MHRA Crown Copyright 2007 UKCRC-CfH Must haves- suitable access Via Specific, high end, healthcare PROVEN, Data manipulation and analysis Tools Business Tools will NOT work GRID enabled for future Crown Copyright 2007 TAKE ONE BETWEEN 4-6 PM TAKE ONE AT NIGHT BETWEEN 7 AND 11PM 1 AT NIGHT REMEMBER TO GET A BLOOD TEST EVERY 4 MONTHS 1 EVERY DAY. REMEMBER TO GET YOUR BLOOD CHECKED EVERY 4 MONTHS TAKE ONE AT NIGHT REMEMBER TO HAVE A BLOOD TEST EVERY 3 MONTHS ONE WITH EVENING MEAL , REGULARLY, AS WELL AS STICKING TO A LOW FAT DIET, TO LOW TAKE ONE DAILY PLEASE SEE DOCTOR FOR REVIEW IN 1 MONTH TAKE ONE AT NIGHT TO LOWER CHOLESTEROL AND REDUCE RISK EVERY FORTNIGHT 1 ON MDU TAKE ONE AT NIGHT TO LOWER CHOLESTEROL AND REDUCE RISK EVERY FORTNIGHT HEART DIS TAKE ONE AT NIGHT FOR TEN DAYS THE 2 EVERY NIGHT. 1 AT NIGHT DISP WEEKLY 1 EVERY NIGHT WEEKLY DISP ONE EVERY NIGHT WEEKLY DISPENSE NOTES FOR PATIENT: PLEASE REPORT ANY UNEXPLAINED ONE EVERY NIGHT WEEKLY DISPENSE VIA DOSETTE BOX NOTES FOR PATIENT: PLEASE REPORT TAKE ONE AT NIGHT DISPENCE WEEKLY TAKE ONE/TWO ALTERNATE DAYS 10N TAKE ONE AT NIGHT 10 PM EVERY DAY...CHECK CHOLESTEROL LEVEL 6 WEEKS 10 EVERY DAY © Crown Copyright 2007 Crown Copyright 2007 Crown Copyright 2007 UKCRC-CfH Comparison of SUPERMARKET and NHS-GP data SUPERM Number of stores Number of users (card holders) Number of ways of using # Stock keeping units (max in store) Factor 2200 3.64 10,000,000 5.00 3 3.33 40000 6.25 No equivs No equivs No equivs No equivs No equivs No equivs No equivs No equivs 18 61 379 500? 5000? *? *? *? *? *? practically zero- bar coded error and change rate of data cumulative factor GPs NHS of GP complexity 8000 Number of practices 50,000,000 NHS Ids 10 #consulation types 250,000 # major READ/ OXMIS/ICD/MEDRA codes used 30,000 1,000 1000 1000 10 10 10 # drug codes # dose instructions # entitry list #expression pick list #intervention types #added data coded data collection areas # added imummisation varaiables up to 10%,little or no bar coding, transcription errors, incorrect birthdays etc error and change rate of data Crown Copyright 2007 UKCRC-CfH Active surveillance Medicines ▐ Move from current PASSIVE system to ACTIVE ▐ ▐ ▐ Availability of e- person level data from: ▐ ▐ ▐ ▐ ▐ ▐ ▐ ▐ ▐ FROM Yellow card + hypothesis testing study TO e Yellow Card +Data Mining on full record + hypothesis testing studies Hospitals HES, Labs, Rx Pharmacy dispensed and prescribed + OTC Pharmacy Other HealthCare Professionals nurses, dentists ++ Other Paradigms HealthSpace aspirin/paracetamol Improved and new Rx decision support tools Automated on screen “instant warning/alert system” Improved, specified standard incident reporting system Maintained position as having the gold standard system Crown Copyright 2007 UKCRC-CfH Active surveillance Devices ▐ ▐ ▐ ▐ ▐ An all incident single NHS reporting system Agreed standards coding Automated delivery to responsible agency Patient reporting via HealthSpace Detailed implanted/integral device coding to be part of e-record Crown Copyright 2007 UKCRC-CfH HEADLINE ISSUES 1. ▐ ▐ ▐ ▐ ▐ ▐ ▐ ▐ RESEARCH LEADS DIRECTLY TO: Improve care, Improved safety Improved efficiency And therefore benefits to patients. It is NOT a SECONDARY use….., RATHER A FUNDAMENTAL PART OF PATIENTS’ OVERALL LIFETIME CARE within a health system such as the NHS. Accepting that it is not usually about 1 on 1 care rather related to groups and cohorts. Additionally, except where consent is obtained, as in interventional studies, the actual research takes place with data subjects (effectively anonymised) rather than patients. Crown Copyright 2007 Helsinki Declaration ▐ ▐ ▐ Currently creates a difficulty as written before the era of her/observational data In the main is about “interventional research” on persons that quite correctly needs consent E-Observational research does not deal with patients ▐ Lines of Data Subject data Crown Copyright 2007 UKCRC-CfH Headline Issues 2. ▐ ▐ ▐ ▐ ▐ ▐ ▐ Mandate use of NHS number on all events Ensure mother-baby and family links are enabled in system Accept a federated approach using record linkage may be the most appropriate and workable solution to maximally enabling the VISIONS Mandate a standardised “Incident reporting “ system be included in all clinical systems Ensure that there is compatibility related to provision of UK wide data (E, W,S, NI), particularly for issues where there is UK wide remit and to gain most of the whole process Build upon existing database/system strengths. (Knowledge, utility, now, historical data) Ensure suitable governance arrangements are in place (Full time staff, related to scientific approval) ▐ ▐ ▐ ▐ ▐ ▐ ▐ All discussions about centralised IT solutions to aid research should be inclusive to gain synergies… some already know more than others. Take on board that there is a large pool of research knowledge about data, data linkage, data validation, data storage, and analysis – the wheel does not need to be reinvented Ensure that clinical trials can benefit from rapid screening to improve recruitment, centralised recruitment and even a portal based clinical trials system directly linked into all clinical systems Talk loud and often about the benefits of the research uses of data- we have nothing to hide Get all hospitals to adopt personal level e Rx and lab systems asap. Prevent/stop silo mentality about use and ownership of data Accept that some types of research (drug related) have a commercial association and are conducted as part of the legally required process. Crown Copyright 2007 UKCRC-CfH Quick wins ▐ Patient safety ▐ Medicine related ▐ ▐ ▐ ▐ Incident reporting…prevention ▐ ▐ ▐ Mandating the inclusion of approved incident reporting software in all clinical systems…..web to central database with adequate coding Disease related ▐ ▐ E-yellow cards into all clinical systems Improved safety related decision support systems in all clinical systems Enablement of close to real time active surveillance (of new medications) Enablement of nationwide reporting systems via all relevant systems Rapidly enable record linkage of existing resources to gain the win-win synergies Foster the creativity that will allow “Privacy enhancing technologies” to maximally enable recruitment to clinical trials and Biobank Crown Copyright 2007 Who owns the data? ▐ ▐ Some commercial IT/database companies seem to have a hold over the data MY data belongs to me ▐ With my consent it may be used for bone-fide research ▐ Implicit consent ▐ based upon anonymous to researcher use Crown Copyright 2007 Data from many systems DB1 DB2 DB3 DB4 SQL Tool Research Dataset Ask the right question……………..get just the data required Tayside, Scotland- GENIE England and Wales - Apollo Crown Copyright 2007 Thank you Crown Copyright 2007