Using the Clinical Practice Research Datalink to support ... › wp-content › uploads › 2016 ›...
1
Using the Clinical Practice Research Datalink to support Yellow Cards Katherine Donegan Vigilance and Risk Management of Medicines, Medicines and Healthcare products Regulatory Agency (MHRA), London, UK The Clinical Practice Research Datalink The UK Clinical Practice Research Datalink (CPRD) collates anonymised primary care electronic healthcare record data on patient and practice demographics, as well as individual patient clinical diagnoses, prescrip- tions, referrals to secondary care, test results, and vaccination records for approximately 13.5 million patients in over 680 general practices across the UK. The CPRD is jointly funded by the NHS National Institute for Health Research (NIHR) and the MHRA. How we already use CPRD to support pharmacovigilance? Strengthening signals - Proactive vaccine pharmacovigilance Observed vs expected analyses: 1) Identify key adverse events of special interest based on knowledge of vaccine, prior vaccine campaigns, events common in the target population 2) Use data from the CPRD to calculate age and gender specific background event rates 3) From the start of the campaign gather weekly data on vaccine use (from CPRD) 4) Compare the number of Yellow Cards Yellow Cards (and any other cases identified) to the number expected assuming normal background rates using the MaxSPRT [1]. For example: Chronic Fatigue Syndrome (CFS) in girls aged 12-13 years following vaccination with the HPV vaccine, 2008/09. This signal of a possible risk of CFS with the HPV vaccine was followed up with a more robust pharmacoepide- miological study again using CPRD data which showed that there was no association [2]. This ensured the safe use of this important vaccine. This methodology has been used, by the MHRA, in the phar- macovigilance of all recent newly introduced routine vaccines. The future for CPRD within MHRA’s pharmacovigilance strategy Vaccines are vital in global public health. There have been many scare stories relating to the safety of vaccinations that have affected uptake and hence the success of immunisation campaigns. We need to be able to gather as much data as quickly as possible to expand the knowledge of a vaccine’s safety profile and to act quickly in the event of a Yellow Cards are a valuable tool in pharmacovigilance. The national coverage and range of reporting routes mean that they are particularly useful for detecting rare events or those associated with long term use. However, alone, they provide limited data on the context of the reaction and are subject to under and over-reporting. In the UK, a patient’s GP is the gatekeeper to all other NHS services so the CPRD provides us with the capability to rapidly identify a large cohort of patients and comprehensively monitor their prescriptions and characterise any potential adverse events over time. CPRD is a valuable resource and can be used to optimise the use of Yellow Card spontaneously reported adverse event data and strengthen the MHRA’s routine pharmacovigilance. References & acknowledgements Assumed level of reporting: Putting signals into context - Risk evaluation, communication, and monitoring outcomes CPRD can also provide in-depth drug utilisation data to help with the assessment of signals, assess benefit risk, support clear communications, and monitor the impact of any regulatory actions Further risk evaluation - Pharmacoepidemiological research Finasteride is a type II 5α-reductase inhibitor. A signal was detected from Yellow Cards Yellow Cards for 5mg finasteride associated with male breast cancer. A matched case-control study using the CPRD helped to further evaluate this signal [4]. * Adjusted for smoking status, BMI and alcohol use. Software has been developed to allow the routine extraction and summary of elec- tronic healthcare record data. The MHRA is investigating the potential use of this software and is working with Commonwealth Informatics to incorporate CPRD data into an efficient interface (LODEX) for MHRA scientists and assessors to use so that the data can be available to support signal detection. CPRD data extracted in this routine way could also be used in other ways to sup- port wider pharmacovigilance. To help prioritise signals by providing drug exposure data to input into the as- sessment of the potential public health impact. There is an increased focus on monitoring the impact of regulatory actions. Basic drug exposure data can often be used as a crude measure of this. Help assess the feasibility of conducting a more robust epidemiological study helping both our own research but also in assessing industry sponsored Post Authorisation Safety Studies (PASS). Cases (n = 180) Controls (n = 878) OR adjusted* (95% CI) Finasteride Never exposed 173 (96.1%) 849 (96.7%) - Ever exposed 7 (3.9%) 29 (3.3%) 1.29 (0.44-3.75) Finasteride duration Never exposed 173 (96.1%) 849 (96.7%) - Up to 2 years 3 (1.7%) 19 (2.2% 0.52 (0.06-4.24) More than 2 years 4 (2.2%) 10 (1.1%) 2.04 (0.59-7.10) Antipsychotics: initiative to reduce prescribing to older people with dementia Concerns over recent years about the use of antipsychotics to treat the behavioural and psychological symptoms of dementia. 0 20 40 60 80 100 120 140 01/00-03/00 04/00-06/00 07/00-09/00 10/00-12/00 01/01-03/01 04/01-06/01 07/01-09/01 10/01-12/01 01/02-03/02 04/02-06/02 07/02-09/02 10/02-12/02 01/03-03/03 04/03-06/03 07/03-09/03 10/03-12/03 01/04-03/04 04/04-06/04 07/04-09/04 10/04-12/04 01/05-03/05 04/05-06/05 07/05-09/05 10/05-12/05 01/06-03/06 04/06-06/06 07/06-09/06 10/06-12/06 01/07-03/07 04/07-06/07 07/07-09/07 10/07-12/07 01/08-03/08 04/08-06/08 07/08-09/08 10/08-12/08 01/09-03/09 04/09-06/09 07/09-09/09 10/09-12/09 01/10-03/10 04/10-06/10 07/10-09/10 10/10-12/10 01/11-03/11 04/11-06/11 Quarter Number of children aged <18 years with a prescription for desmopressin per 100,000 patients in GPRD Nasal spray Oral tablet Sublingual The prevalence of prescribing of desmopressin in children aged <18 years in the UK 01/2000-06/2011 split by formulation (CPRD) [1] Kulldorff M, et al. A maximised sequential probability ratio test for drug and vaccine safety surveillance. Sequential Analysis. 2011; 30: 58–78. [2] Donegan K, et al. Bivalent human papillomavirus vaccine and the risk of fatigue syndromes in girls in the UK. Vac- cine. 2013; 31: 4961-7. [3] Seabroke S, et al. Development of a novel regulatory pharmacovigilance prioritisation system: an evaluation of its performance at the UK Medicines and Healthcare products Regulatory Agency. Drug Safety. 2013; 36: 1025-32. [4] Seabroke S, et al. A case-control study to investigate the association between finasteride and male breast cancer using the GPRD. Drug Safety. 2009; 32: 963-4. [5] Cederholm, S, et al. Structured Assessment for Prospective Identification of Safety Signals in Electronic Medical Records: Evaluation in The Health Improvement Network. Drug safety. 2015; 38: 87-100. Thanks go to Suzie Seabroke for the data presented on finasteride. The screenshots from the LODEX software have been provided by Commonwealth Informatics CPRD has proven itself to be a valuable resource for supporting the Yellow Card scheme. However, the approach taken has been, to date, somewhat ad- hoc. In order to better utilise CPRD to support the Yellow Card and routine pharmacovigilance activities a more consistent approach should be taken. LODEX functionality: Overall summary of CPRD database Drug specific summary Medical event summary Drug-event summary Chronographs are also availa- ble for individual drug-event combinations. These can be used to look for temporal pat- terns. Here a signal of Eusta- chian tube disorder with gen- tamicin could be dismissed as there is a clear increase in risk prior to treatment [5]. As seen for vaccines, this data can help put Yellow Cards into context and examine trends over time as well as examine under-reporting. © Crown Copyright 2015
Transcript of Using the Clinical Practice Research Datalink to support ... › wp-content › uploads › 2016 ›...
Using the Clinical Practice Research Datalink to support Yellow Cards
Katherine Donegan Vigilance and Risk Management of Medicines, Medicines and Healthcare products Regulatory Agency (MHRA), London, UK
The Clinical Practice Research Datalink
The UK Clinical Practice Research Datalink (CPRD) collates anonymised primary care electronic healthcare record data on patient and practice demographics, as well as individual patient clinical diagnoses, prescrip-tions, referrals to secondary care, test results, and vaccination records for approximately 13.5 million patients in over 680 general practices across the UK. The CPRD is jointly funded by the NHS National Institute for Health Research (NIHR) and the MHRA.
How we already use CPRD to support pharmacovigilance?
Strengthening signals - Proactive vaccine pharmacovigilance
Observed vs expected analyses: 1) Identify key adverse events of special interest based
on knowledge of vaccine, prior vaccine campaigns, events common in the target population
2) Use data from the CPRD to calculate age and
gender specific background event rates 3) From the start of the campaign gather weekly data on vaccine use (from CPRD) 4) Compare the number of Yellow Cards Yellow Cards (and any other cases identified) to the
number expected assuming normal background rates using the MaxSPRT [1].
For example: Chronic Fatigue Syndrome (CFS) in girls aged 12-13 years following vaccination with the HPV vaccine, 2008/09.
This signal of a possible risk of CFS with the HPV vaccine was followed up with a more robust pharmacoepide-miological study again using CPRD data which showed that there was no association [2]. This ensured the safe use of
this important vaccine. This methodology has been used, by the MHRA, in the phar-macovigilance of all recent newly introduced routine vaccines.
The future for CPRD within MHRA’s pharmacovigilance strategy
Vaccines are vital in global public health.
There have been many scare stories relating to the safety of vaccinations that have affected uptake and hence the success of immunisation campaigns.
We need to be able to gather as much data as quickly as possible to expand the knowledge of a vaccine’s safety profile and to act quickly in the event of a
Yellow Cards are a valuable tool in pharmacovigilance. The national coverage and range of reporting routes mean that they are particularly useful for detecting rare events or those associated with long term use.
However, alone, they provide limited data on the context of the reaction and are subject to under and over-reporting.
In the UK, a patient’s GP is the gatekeeper to all other NHS services so the CPRD provides us with the capability to rapidly identify a large cohort of patients and comprehensively monitor their prescriptions and characterise any potential adverse events over time.
CPRD is a valuable resource and can be used to optimise the use of Yellow Card spontaneously reported adverse event data and strengthen the MHRA’s routine pharmacovigilance.
References & acknowledgements
Assumed level of reporting:
Putting signals into context - Risk evaluation, communication, and monitoring outcomes CPRD can also provide in-depth drug utilisation data to help with the assessment of signals, assess benefit risk, support clear communications, and monitor the impact of any regulatory actions
Further risk evaluation - Pharmacoepidemiological research
Finasteride is a type II 5α-reductase inhibitor. A signal was detected from Yellow Cards Yellow Cards for 5mg finasteride associated with male breast cancer. A matched case-control study using the CPRD helped to further evaluate this signal [4].
* Adjusted for smoking status, BMI and alcohol use.
Software has been developed to allow the routine extraction and summary of elec-tronic healthcare record data. The MHRA is investigating the potential use of this software and is working with Commonwealth Informatics to incorporate CPRD data into an efficient interface (LODEX) for MHRA scientists and assessors to use so that the data can be available to support signal detection.
CPRD data extracted in this routine way could also be used in other ways to sup-port wider pharmacovigilance. To help prioritise signals by providing drug exposure data to input into the as-
sessment of the potential public health impact. There is an increased focus on monitoring the impact of regulatory actions.
Basic drug exposure data can often be used as a crude measure of this. Help assess the feasibility of conducting a more robust epidemiological study
helping both our own research but also in assessing industry sponsored Post Authorisation Safety Studies (PASS).
Cases (n = 180)
Controls (n = 878)
OR adjusted* (95% CI)
Finasteride
Never exposed 173 (96.1%) 849 (96.7%) -
Ever exposed 7 (3.9%) 29 (3.3%) 1.29 (0.44-3.75)
Finasteride duration
Never exposed 173 (96.1%) 849 (96.7%) -
Up to 2 years 3 (1.7%) 19 (2.2% 0.52 (0.06-4.24)
More than 2 years 4 (2.2%) 10 (1.1%) 2.04 (0.59-7.10)
Antipsychotics: initiative to reduce prescribing to older people with dementia
Concerns over recent years about the use of antipsychotics to treat the behavioural and psychological symptoms of dementia.
0
20
40
60
80
100
120
140
01
/00
-03
/00
04
/00
-06
/00
07
/00
-09
/00
10
/00
-12
/00
01
/01
-03
/01
04
/01
-06
/01
07
/01
-09
/01
10
/01
-12
/01
01
/02
-03
/02
04
/02
-06
/02
07
/02
-09
/02
10
/02
-12
/02
01
/03
-03
/03
04
/03
-06
/03
07
/03
-09
/03
10
/03
-12
/03
01
/04
-03
/04
04
/04
-06
/04
07
/04
-09
/04
10
/04
-12
/04
01
/05
-03
/05
04
/05
-06
/05
07
/05
-09
/05
10
/05
-12
/05
01
/06
-03
/06
04
/06
-06
/06
07
/06
-09
/06
10
/06
-12
/06
01
/07
-03
/07
04
/07
-06
/07
07
/07
-09
/07
10
/07
-12
/07
01
/08
-03
/08
04
/08
-06
/08
07
/08
-09
/08
10
/08
-12
/08
01
/09
-03
/09
04
/09
-06
/09
07
/09
-09
/09
10
/09
-12
/09
01
/10
-03
/10
04
/10
-06
/10
07
/10
-09
/10
10
/10
-12
/10
01
/11
-03
/11
04
/11
-06
/11
Quarter
Nu
mb
er o
f ch
ild
ren
ag
ed <
18 y
ears
wit
h a
p
resc
rip
tio
n f
or
des
mo
pre
ssin
per
100
,000
pat
ien
ts i
n
GP
RD
Nasal spray Oral tablet Sublingual
The prevalence of prescribing of desmopressin in children aged <18 years in the UK 01/2000-06/2011 split by formulation (CPRD)
[1] Kulldorff M, et al. A maximised sequential probability ratio test for drug and vaccine safety surveillance. Sequential
Analysis. 2011; 30: 58–78. [2] Donegan K, et al. Bivalent human papillomavirus vaccine and the risk of fatigue syndromes in girls in the UK. Vac-
cine. 2013; 31: 4961-7. [3] Seabroke S, et al. Development of a novel regulatory pharmacovigilance prioritisation system: an evaluation of its
performance at the UK Medicines and Healthcare products Regulatory Agency. Drug Safety. 2013; 36: 1025-32. [4] Seabroke S, et al. A case-control study to investigate the association between finasteride and male breast cancer
using the GPRD. Drug Safety. 2009; 32: 963-4. [5] Cederholm, S, et al. Structured Assessment for Prospective Identification of Safety Signals in Electronic Medical
Records: Evaluation in The Health Improvement Network. Drug safety. 2015; 38: 87-100.
Thanks go to Suzie Seabroke for the data presented on finasteride. The screenshots from the LODEX software have been provided by Commonwealth Informatics
CPRD has proven itself to be a valuable resource for supporting the Yellow Card scheme. However, the approach taken has been, to date, somewhat ad-hoc. In order to better utilise CPRD to support the Yellow Card and routine pharmacovigilance activities a more consistent approach should be taken.
LODEX functionality: Overall summary of
CPRD database Drug specific summary Medical event summary Drug-event summary
Chronographs are also availa-ble for individual drug-event combinations. These can be used to look for temporal pat-terns. Here a signal of Eusta-chian tube disorder with gen-tamicin could be dismissed as there is a clear increase in risk prior to treatment [5].
As seen for vaccines, this data can help put Yellow Cards into context and examine trends over time as well as examine under-reporting.
© Crown Copyright 2015
