Transparency in Risk Regulation: EMA Case Study by Dominic Way, King's College London

This electronic thesis or dissertation has been downloaded from the King’s Research Portal at https://kclpure.uk/portal/ Transparency in Risk Regulation The Case of the European Medicines Agency Way, Dominic Awarding institution: King's College London The copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without proper acknowledgement. END USER LICENCE AGREEMENT Unless another licence is stated on the immediately following page this work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.org/licenses/by-nc-nd/4.0/ You are free to copy, distribute and transmit the work Under the following conditions:  Attribution: You must attribute the work in the manner specified by the author (but not in any way that suggests that they endorse you or your use of the work).  Non Commercial: You may not use this work for commercial purposes.  No Derivative Works - You may not alter, transform, or build upon this work.

Any of these conditions can be waived if you receive permission from the author. Your fair dealings and other rights are in no way affected by the above. Take down policy If you believe that this document breaches copyright please contact librarypure@kcl.uk providing details, and we will remove access to the work immediately and investigate your claim. 2022 Transparency in Risk Regulation: The Case of the European Medicines Agency 2017 Dominic Hugo Patrick Way King’s College London Doctor of Philosophy in Geography (Science) Research FT The copyright of this thesis rests with the author and no quotation from it or information derived from it may be published without proper acknowledgement Dedicated in loving memory of my grandfather, Peter Orchard Williams CBE FRCP (1925-2014) 2 ABSTRACT Transparency has risen to prominence in risk regulation leading government authorities in Europe and North America to introduce an avalanche of new policies.

While transparency promises to achieve many instrumental objectives, there is insufficient empirical evidence on the effectiveness of the actual policies being implemented. This study first explored the ambiguity of transparency in risk regulation. This informed a review of the seriously fragmented literature and an original typology, which clarified what is being made transparent by risk regulators (objects), how (mechanisms), why (goals) and for whom (audiences). An in- depth case study was then conducted on transparency at the European Medicines Agency (EMA), the EU agency responsible for pharmaceuticals.

The four case methods were extensive historical and contemporary documentation; direct observations and interviews conducted between 2012 and 2016; and multi-national surveys (UK, France, Germany, Spain) of medical doctors (specialists and general practitioners) and patients diagnosed with one of five medical conditions (HIV/AIDS, idiopathic pulmonary fibrosis, multiple sclerosis, rheumatoid arthritis, osteoporosis) (N=2,015). After providing an in-depth historical analysis of transparency at EMA (1995-2016), the case study evaluated three salient policies, which all seek to make the data that underpins decision-making in EMA’s scientific committees more ‘visible’ to outsiders. EMA’s policies have important strengths and weaknesses for achieving the regulator’s objectives. Transparency was also found to have serious unwanted effects including significantly increasing legal action against EMA, enabling unrestricted (high and low quality) re-analyses of benefit-risk data, and consuming substantial agency resources.

Simply publishing large quantities of raw data online is unlikely to be the most effective way of achieving the regulator’s transparency objectives. Rather, effective transparency requires understanding (1) the limits of full disclosure, (2) the desired quality of transparency (for multiple audiences), (3) the capacity of expert and non-expert audiences to assess transparent information, and (4) the ability of intermediaries to re-package and communicate benefit-risk information effectively. Choosing the right transparency policies ultimately means resisting becoming captivated by quantity and must take opportunity costs into careful consideration. The thesis concludes by strongly arguing that transparency should be re-conceptualised as a (risk) communication process in order to improve policy effectiveness.

3 ACKNOWLEDGEMENTS First and foremost I would like to thank Ragnar Löfstedt, my primary supervisor, who provided me with tremendous academic support over the past four and a half years. My PhD simply would not have been possible without his invaluable guidance. I extend my thanks to Henry Rothstein, my secondary supervisor, who gave me precious advice and recommendations especially on earlier drafts of this thesis. I would like to express my great appreciation to Frederic Bouder (Maastricht University) and Darrick Evensen (Cardiff University) whose expertise especially on transparency, pharmaceutical regulation and survey methodologies helped guide my research.

I wish to acknowledge David Demeritt and all the PhD students that contributed to King’s College London’s Hazards and Risk reading groups. I would also like to thank my viva examiners, Julie Barnett (Bath University) and George Gaskell (London School of Economics). In collecting data for this thesis, it is a pleasure to thank David Haerry and all the patient group organisations that generously provided their time and assistance including (alphabetically): Asociación De Familiares Y Enfermos De Fibrosis Pulmonar Idiopática; Association Fibrose Pulmonaire Idiopathique; British Lung Foundation; Esclerosis Multiple Espana; European AIDS Treatment Group; Liga Reumatológica Española; Multiple Sclerosis Society; and Pulmonary Fibrosis Trust. The funding for collecting the survey data was provided by a grant from Biogen Idec.

and I would particularly like to thank Carmen Bozic, James Milligan and Jan Kilhamn. I am also grateful to the staff at Ipsos, who collected and collated the survey data, and I wish to acknowledge the help I received from EMA regulators with special thanks to Juan Garcia. Finally, I am deeply grateful and indebted to my family and loved ones, Judy, Patrick, Oliver, Frederick, Jess, Alastair and Simone, for their unwavering support and encouragement throughout. 4 CONTENTS Abstract 3 Acknowledgements 4 List of figures 8 List of tables 10 Acronyms 12 Chapter I: INTRODUCTION 14 1.

The European Medicines Agency 22 1. Contribution of this thesis 25 1. Thesis organisation 27 Chapter II: DEVELOPING A TRANSPARENCY TYPOLOGY 29 2. The need to develop a typology 29 2.

Transparency Audiences 35 Chapter III: TRANSPARENCY IN RISK REGULATION 37 3. Debating transparency policies 38 3. Focusing on input transparency 48 3. Soft risk regulation 48 3.

Scrutinising regulatory data 51 3. Data quantity and quality issues 54 3. Building public trust 60 3. Concluding remarks 64 Chapter IV: METHODOLOGY 66 4.

The case study approach 67 4. Choosing the EMA case 72 4. Explanatory case design 73 4. Observations and interviews 84 4.

Patient and doctor samples 90 4. Recruitment and representativeness 93 4. Analysing the evidence 100 5 4. Limitations 101 Chapter V: TRANSPARENCY AT EMA IN HISTORICAL PERSPECTIVE 104 5.

Background and formation 104 5. Establishing an independent agency (1995-2000) 108 5. Maintaining and strengthening transparency (2001-2009) 112 5. EMA comes under fire (2010) 116 5.

The new transparency era (2010-) 119 Chapter VI: EMA’s INPUT TRANSPARENCY POLICIES 127 6. Clinical trial data 128 6. Clinical trial registration 128 6. Goals and audiences for level 1 130 6.

Tackling non-compliance 133 6. Summary-level trial results 137 6. Goals and audiences for level 2 138 6. Quantity and quality issues 145 6.

Clinical study reports 148 6. Goals and audiences for level 3 149 6. Confidentiality and privacy 156 6. Poor quality re-analyses 160 6.

Capacity of outsiders to re-analyse 161 6. The role of intermediaries after re-analysis 162 Chapter VII: PATIENT AND DOCTOR SURVEYS 166 7. Quantity of publicly available information 167 7. Quality of publicly available information 168 7.

Awareness of regulatory bodies and their activities 169 7. Knowledge of how regulators assess medicines 172 7. Intermediaries as information providers 174 7. When new findings should be made publicly available 176 7.

Patient’s reactions to receiving safety information 177 7. Doctors’ opinions on improving regulatory communication 178 Chapter VIII: EVALUATION OF EMA’s TRANSPARENCY POLICIES 183 8. EMA’s transparency strategy post-2010 184 8. Registration and summary level results policies 186 8.

Re-analysing CSRs 190 8. Desirability and feasibility of full disclosure 190 8. Capacity to re-analyse CSRs 192 8. Conveying the results of re-analyses 194 8.

Implications of survey findings 196 8. Patients and doctors want more and better quality information 197 8. Patient and doctors are confused about the contribution of the regulators 198 6 8. Patients and doctors disagree over when re-analysed data should be conveyed 200 8.

Patients will have varying reactions to re-analysed data 201 Chapter IX: CONCLUSION AND RECOMMENDATIONS 203 9. Captivated by full disclosure 204 9. Re-conceptualising transparency as a risk communication process 207 9. Implications and recommendations 209 9.

Establish an independent benefit-risk communication advisory 211 board at EMA 9. Build relationships with trusted intermediaries 212 9. A conference on evidence-based transparency 212 9. Conduct more empirical research on effectiveness 213 9.

A call for proposals from risk researchers 213 REFERENCES 215 APPENDICIES 274 A. Geographic regions of patient and doctor samples 274 B. Four example screenshots of adrreports. Example summary level results document 309 F.

Statement on Dominic Way’s contribution to research at KCRM 329 7 LIST OF FIGURES 2.1 Diagram showing three objects of transparency that risk regulators might seek to make more visible. Adapted from Heald (2006a: 30) 31 6.1 Screenshot providing an example of a clinical trial record in EMA’s register (clinicaltrialsregister. Note: the trial protocol can be accessed (on the website) by clicking the blue ‘GB’ symbol.2 Bar chart showing the number of requests and distribution of requestors for access to EMA documents that were either refused (blue shading) or granted (orange shading) between 2010-2013. Parentheses show the number of documents released for each requestor.

EC denotes European Commission. EU denotes European Union. Source data: Bonini et al.1 Bar chart comparing the extent to which patient (N=1,010) and doctor (N=1,005) respondents agree or disagree with three statements regarding information about medicines and health information 170 7.2 Bar chart comparing the extent to which patients (N=1,010) and doctors (N=1,005) agree or disagree with four statements regarding information about medicines and how that information is communicated.3 Bar chart comparing the extent to which patient and doctor respondents agree or disagree with the statement ‘I have good knowledge of how [NCA/EMA] assesses the safety of [medical condition] medicines’. Separate questions were asked for their relevant NCA and EMA.

Each relevant medical condition was piped in where applicable.4 Bar charts for (A) patients and (B) doctors that compare: (1) respondents (%) that indicated ‘very easy’ or ‘somewhat easy’ to the question: “How easy is it for you to obtain information about medicines from each of the following sources of information?” (light shading) and; (2) respondents (%) that indicated ‘very trustworthy’ or ‘somewhat trustworthy’ to the question: “How trustworthy 8 do you believe the following sources are in providing you with advice on the side effects associated with specific medicines?” (dark shading). Results are sorted by ease of obtaining information (light shading). SMPC denotes summary of product characteristics.5 Bar chart comparing answers to the question “At what stage do you think information should be conveyed to the public about a possible safety issue of a medicine that they use or may use?”.6 (A) Respondents’ answers (%) to the question: “If the information you personally received (via letter, telephone, e-mail etc.) points to safety problems with a [insert sample group medical condition] medicine you are currently taking, do you think you are more likely to…” (N=1010). (B) Four choropleth maps comparing regional variations for patients that indicated ‘stop taking your medicine’ or ‘reduce the dose of your medicine’.

The darker the shading the higher the number of respondents in that geographic region that indicated stop or reduce (see key).7 Bar chart comparing respondents divided into country and medical condition groups that answered either reduce your dose of the medicine (light shading) or stop taking your medicine (dark shading) (%) for the question: “If the information you personally received (via letter, telephone, e-mail etc.) points to safety problems with a [relevant medical condition] medicine you are currently taking, do you think you are more likely to… (a) stop taking your medicine, (b) reduce your dose of the medicine, (c) continue taking your medicine as usual, (d) seek additional advice about the medicine or (d) don’t know”.