Updates from the Pharmacovigilance and Special Access Branch Dr - PowerPoint PPT Presentation

Updates from the Pharmacovigilance and Special Access Branch Dr Grant Pegg Director, Signal Investigation Unit Sarah May Lead Inspector, Risk Management Section ARCS Conference 6 August 2019

Updates from the pharmacovigilance and special access branch • Using new sources of data in Pharmacovigilance • Pharmacovigilance Inspection Program (PVIP) update • International collaboration activities • Adverse Event Management System (AEMS) • Q and A 1

Data in pharmacovigilance • Where are we? • Where do we want to be? 2

Using new sources of data in pharmacovigilance • Expert Review of Medicines and Medical Devices Regulation (MMDR) – Recommendation 27: § The Panel recommends that the Australian government develop a more comprehensive post-market monitoring scheme for medicines and medical devices. Such a scheme to include: • Better integration and timely analysis of available datasets, including analysis of matched de- identified data from the Pharmaceutical Benefits Scheme, Medical Benefits Scheme, eHealth records, hospital records, private health insurance records and device and other relevant registries and datasets 3

Using health data for pharmacoepidemiology • Types of available data – Prescription and dispensing, e.g. Pharmaceutical Benefits Scheme (PBS) data – Medical services, pathology, imaging, e.g. Medicare Benefits Scheme (MBS) data – Hospital discharge data – Birth and Death registries (state-based in Australia) – Australian Cancer Database – General Practice clinical data, e.g., NPS MedicineInsight data – Sales data, eg IQVIA • Linked datasets – Sax Institute 45 and Up study – National Data Linkage Demonstration Project dataset – *coming soon* National Integrated Health Services Information Analysis Asset 4

Using health data for pharmaco-epidemiology Weaknesses Strengths – Lack of certain types of – Potential for large cohort information - residual numbers confounding – “Real-world” exposure – Requires exposure and – Better population coverage outcome to be measurable – Better generalisability in the available dataset 5

Impact of pharmaco-epi on medicines egulation • Examples of studies which have lead to product information document changes 6

TGA use of health data for pharmacovigilance • Two feasibility studies: – Signal detection: using prescription sequence symmetry analysis (PSSA) of PBS data – Signal validation: using the Sax Institute’s 45 and Up study dataset 7

Signal identification: PSSA The prescription sequence can be visualised, as shown below. • PBS data Figure 1: Temozolomide compared to frusemide • Dispensings of a particular medicine (frusemide) ASR: 4.64, lowerlimit of 95% CI 3.26 used as a proxy for an adverse event (heart failure) • Imbalance between dispensings of the proxy medicine before or after initiation of other medicines in the dataset (index medicines) • Expressed as a ratio between the number of patients who received the index medicine before the proxy medicine compared to those who received the index medicine after the proxy medicine. • Positive signals (lower 95% confidence interval >1) investigated further. 8

Signal identificaton: PSSA • Results – 684 medicines included in the final analysis – 26 potential signals for heart failure detected § Majority were indicated for • Cancer • Glaucoma • Migraine – The heart failure signal was verified for one medicine during our internal evaluation; the Sponsor had independently identified the signal and submitted an SRR to update the Product Information during this process 9

Signal validation: 45 and up study cohort analysis • The Sax Institute 45 and Up study – Cohort >250,000 participants, broad consent for linkage of survey data with a large number of state/federal administrative health datasets. • Examined the risk of intracranial haemorrhage with direct-acting oral anticoagulants compared with warfarin. • Results concordant with studies published using international data sources, including FDA mini-Sentinel, and New Zealand population data. • Limited by a small sample size, uncertainty about exposure classification (e.g. duration of warfarin treatment), and lack of information on the indication for treatment. 10

TGA use of health data for pharmacovigilance • Two feasibility studies: – Signal detection using prescription sequence symmetry analysis of PBS data – Signal validation using the Sax Institute’s 45 and Up study dataset • Increased resources for in-house data analysis – Recruitment of a biostatistician and an epidemiologist in PSAB – Access to population level linked administrative datasets 11

…towards national level linked health data • Investigating signals of interest – Rapid investigation § Number of individuals in the dataset exposed to the medicine § Number of relevant outcome events – If sufficient numbers of exposed individuals and outcome events, proceed to a full study with appropriate confounder management (different methodologies possible depending on the question, e.g. cohort, case-control, case-crossover). 12

TGA use of health data for pharmacovigilance • Two feasibility studies: – Signal detection using prescription sequence symmetry analysis of PBS data – Signal validation using the Sax Institute’s 45 and Up study dataset • Increased resources for in-house data analysis – Recruitment of a biostatistician and an epidemiologist in SIU – Access to population level linked administrative datasets • Academic partnerships and collaboration – Experts in pharmacoepidemiology from a number of Australian Universities 13

Other types of data • Sales data – Useful for products not on the PBS, or where medicines are being prescribed privately, and for over-the-counter products. • Using IQVIA sales data to analyze the impact of the upscheduling of codeine – Converted number of packs to mg of codeine supplied to the market Projected quantity of codeine that would have been – sold if no upscheduling, based on supply trends over the previous 4 years. – Amount supplied following upscheduling was 46% less than projected, equivalent to a decrease in over 6900kg codeine supplied. • www.tga.gov.au/media-release/significant-decrease- amount-codeine-supplied-australians 14

Lessons learnt • Importance of appropriate expertise • Value of collaboration with academic partners • Using health data for this purpose is time intensive • Not all medicine safety signals can be analysed using this method • Australia is a small country � insufficient sample size for rare adverse events for highly specialised medicines. 15

Future directions • Build rapid response capability • Concurrent analysis of Australian and Canadian linked administrative health data • Collaborate with international agencies on ‘big data’ in PV 16

Pharmacovigilance Inspection Program (PVIP) update • Overview of PVIP to date • findings of interest • common findings around significant safety issues - a review of – Legislation – Management – Reporting – TGA actions 17

Pharmacovigilance requirements • Therapeutic Goods Act 1989 (section 28(5e), 29A and 29AA) • Therapeutic Goods Regulations 1990 (Regulation 15A) • Pharmacovigilance responsibilities of medicine sponsors – Australian recommendations and requirements (Pharmacovigilance guidelines) • Conditions – standard and specific (Applying to registered or listed therapeutic goods under Section 28 of the Therapeutic Goods Act 1989) 18

Overview of the PVIP PVIP PILOT September 2017 to July 2019 October 2015 to May 2016 Sponsors selected based on 2018 risk Sponsor selection: volunteers assessment survey and internal intelligence 16 routine PV Inspections 10 PV Inspections Average inspection time: 3 days Average inspection time: 2.5 days Conducted by 2-4 inspectors Conducted by 1 - 2 inspectors 60 significant findings (critical, major) 25 significant findings (critical, major) 37 other findings (minor, observations) 18 other findings (minor, observations) 19

PVIP metrics 2017-2019 20

Findings of interest (2019) Social listening- collection of ADRs Rogue social media sites- business rules for set up and management including monitoring of any social media by ANY staff member Sales Managers/Representatives using up to date marketing materials and PI/CMI- ensure timely communication of new material and that out of date material is returned/destroyed Due diligence in medical enquires- where individuals are enquiring about an ADR or use in as a special situation always ask if the product has been used? Is there an ADR? Request and collect Aboriginal and Torres Strait Islander ethnicity data 21

Findings of interest (2019) Dangers of automation (automated reporting, seriousness, follow up rules etc.)- continual review of business rules to ensure accuracy and compliance with Australian requirements Company clinical services not being considered as part of the sponsors remit Due diligence in Market research activities- ensure contractors have valid contracts (with PV language), regular training of all staff, reconciliation of any ADRs and review of data. Use of CRM software and free text fields- who is monitoring this for ADRs 22