Summary report of a workshop, May 22-23 2008, Ottawa, sponsored by the Canadian Institutes of Health Research and hosted by the University of Ottawa

Definition. "Post-market surveillance is the continued monitoring for, and the study of effects and other safety and effectiveness related aspects of, health products that have been marketed to the public."

(Ref: Alghabban A. (2004) Dictionary of Pharmacovigilance. Pharmaceutical Press, Chicago IL, p. 376.)

Table of Contents

Introduction
Agenda
International experience
Context /issues/needs re: post-market surveillance in Canada
Public Drug Plans
Issues to be resolved
Moving forward
Conclusion

---

Introduction

This was the second of a series of ad-hoc workshops to discuss issues relevant to a proposal to establish a pan-Canadian network of research centres of excellence that would engage in studies related to post market surveillance of prescription drugs. Such a national network would develop trusted, timely and nonbiased evidence to inform pharmaceutical policy decisions and treatment practices. The centres of excellence would help to fill evidence gaps on the safety and effectiveness of medicines which may be currently harming Canadians, develop human resource capacity to conduct on post market surveillance research, and address the need for a rapid response to important policy or safety concerns. The pan-Canadian network would be under the guidance of a national oversight body that would identify gaps in knowledge, evidence and information and help to prioritize areas for research.

The intent of the workshop was to begin to consider how both Health Canada (the regulator) and the country's public drug plans could make optimal use of the real world safety and effectiveness data that would be generated by such a network in order to strengthen decision making practices.

Background: The understanding of a drug's risks and benefits changes over its life cycle. The need for post-market surveillance is acknowledged by stakeholders, as is the need for a coordinated approach and timely, high quality studies. However, no organization now has the mandate and capacity to do post-market surveillance and rectify the evidence gap. Currently, the drivers for post-market surveillance include: high profile withdrawals based on safety concerns (e.g., Vioxx^®); the placement of new, expensive biologics for regular treatment use; the increased use of surrogate markers instead of clinical outcomes for market authorization purposes; and the cost effectiveness of new entities.

Enabling legislation in Canada will assist with the regulatory prerogative necessary to support post-market surveillance . The concept of progressive licensing legislation as outlined in Bill C-51 (2008), with its new emphasis on a life cycle approach to drug regulation , will require post marketing surveillance commitments from market authorization holders (developers and marketers of drugs) and formal evaluations. A modernized Food and Drugs Act will give Health Canada the ability to be reliant on ongoing pharmacosurveillance research to inform decisions about the safety and effectiveness of medicines. Anticipated changes in a new regulatory framework that would enable post-market surveillance and the need for a post-market surveillance research network of centres of excellence include:

• an explicit life-cycle approach to the development and ongoing marketing of a pharmaceutical
• formal incorporation of benefit-risk assessment in addition to assessments on safety, efficacy and quality
• authorization with post-market commitments for the market authorization holder; increased emphasis on product information (labels, product monographs, package leaflet)
• enhanced post-market authority and activities; a more coordinated effort for pharmacovigilance
• risk management activities
• formal re-evaluation of a drug over its life cycle, as necessary
• increased emphasis on evaluation of activities and evaluation of the regulatory framework.

In addition to enabling legislation, Canada is well positioned to realize the potential for a national network of research centres: We currently have rich federal and provincial data sources and the ability to link data sets for a more complete picture of pharmaceutical utilization and its effects. Post-market surveillance research can have some limitations, including the generalizability and validity of findings, selection biases, coding inaccuracies, detection biases and survivor treatment. However, Canadian researchers have developed advanced computing and biostatistical techniques (e.g., data mining), built bridges between disciplines to leverage research methodologies (e.g., econometric approaches), and have a healthy understanding of the limitations and biases involved in the interpretation of pharmacovigilance research findings. Multiple Canadian jurisdictions and agencies have an interest in improving post market surveillance and, given the need to use scarce resources and expertise in an efficient manner, collaborative action on post-market surveillance research is essential.

Agenda

To understand:

• how regulators and drug benefit plans in other countries are using and generating post market information about prescription drugs and the lessons they have for Canada
• Health Canada's current and planned role (as federal regulator) in the post market surveillance of drugs
• How CADTH's Common Drug Review, and the Patented Medicine Prices Review Board may assist with defining the needs for post-market safety and effectiveness information on prescription drugs in Canada
• Federal and provincial drug plans and their use of post market surveillance evidence; case studies of provincial initiatives in post market evaluation
• Issues to be resolved in using evidence from post market surveillance of prescription drugs - views of (1) presenters and (2) workshop participants

International experience

Concerns about safety and cost effectiveness have led to an international push to step up post market surveillance of pharmaceuticals. While some countries are ahead of Canada in setting up systems to collect information about how drugs are used after they enter the market, it is still early days internationally, as everyone is finding their way (as one conference presenter observed, it is rather "like the wild west"). An abiding issue, internationally, is that the information submitted by manufacturers to support market authorization decisions is not adequate/ appropriate for reimbursement decisions. For example, universally agreed upon current limits of the information submitted to regulators for marketing of a new drug include:

• New drug product submissions that based on surrogate markers or limited clinical evidence, meaning that details about true safety and effectiveness are not known by the time the drug is available for prescribing
• Incomplete safety information
• Clinical trials submitted for market authorization that are based on limited numbers of patients within a fairly homologous group which is not necessarily reflective of the drug's use in the real world
• Exclusion from trials of high risk, complex patients who may be more susceptible to adverse events once a drug is released onto the market

The above limitations are important because, on market approval, the drug may well be prescribed for a wider range of patient types (than those in the trials) exposing them to unknown risks and essentially changing the risk/benefit profile of the drug in the 'real world'. At the same time, information is not accurately/adequately collected post market.

One of the workshop presenters put the above limitations of the clinical trials submitted to regulators in context by explaining the operations of the pharmaceutical manufacturer -- usually a publicly traded private corporation doing business around the world. Much of the evidence developed around a new drug is intended to support a multitude of global commercial decisions outside the regulatory or technology assessment environments, and hence the information about the usefulness of a new drug is very limited. The clinical trials that manufacturers undertake "can answer only one or a few questions per study," while there are a multitude of additional questions surrounding new drugs at the point of market approval. The Phase 3 trials submitted to regulators, which are typically of short duration and based on use by a restricted population, provide only marginal additional safety information to that provided by earlier Phase 2 studies. To restate these limitations, there is little information on: the effects of long term use; any but the most frequent safety issues; drug interactions; use of the drug in its likely full utilization population (e.g., high risk, complex patients); safety and effectiveness compared to existing drugs; and appropriate utilization.

In other words, at the time of completion of Phase 3 studies, we really do not know much about a new drug. We have little predictive knowledge for how a new drug will behave in the "real world" and we have little knowledge for how a new drug will behave in any individual patient. Outstanding questions at the time a new drug is marketed include:

• Where does the product "fit" with respect to other therapies for specific indications?
• What is the cost-effectiveness of the new drug?
• How can the drug best be introduced into sub-populations not studied in the qualifying clinical trials?
• How to collect information on safety for less common and rare, but serious adverse events?

Initiatives in other countries

Most developed countries are aware of the limitations outlined above and are actively pursuing policy initiatives to address them. Workshop participants were reminded that approval should not be the "last call" for realistic and effective regulatory action on drug safety. Several countries have recognized the need for a life-cycle approach to pharmaceuticals, though all face the challenge of putting "evidence through the policy process" to realize beneficial outcomes and many have insufficiently flexible regulatory tools to achieve those outcomes.

Regulatory agencies: Regulatory agencies in France, New Zealand, the United Kingdom and the United States have arrangements with one or more research networks, enabling them to commission research into drug safety and effectiveness. The EMEA is working to establish a network (the European Network of Centres for Pharmacoepidemiology and Pharmacovigilence). Internationally, there are a variety of approaches to pharmacosurveillance.

Active surveillance approaches include tracking patients who are prescribed a targeted drug (prescription event monitoring); setting up cohort studies for those using new drugs; regular surveys of prescribers and pharmacists; disease registries (especially for drugs that present particular risks); database mining (such as that done on the UK General Practice Research Database and the Scottish Medicines Monitoring Unit); and electronic surveillance of emergency ward charts (such as an existing US initiative encompassing 64 hospitals and plans for a new, 25 million patient data base).

Passive approaches include adverse drug reaction (ADR) reporting and monitoring (levels vary internationally, with New Zealand having one of the highest rates); requirement for risk management plans as a condition of market authorization (as required in the European Union, though there are no legal controls for plan completion); requirements for, and monitoring of, phase 4 trials (though most regulators have no regulation or tools to effectively enforce requirements and most phase 4 trials are not completed); networks of pharmacovigilance centres; five year renewals for market authorization and commissioning of post market research projects.

Passive and active approaches are often coordinated. Passive ADR reporting provides signals regarding potential safety issues, but signals generally only generate hypotheses and must be confirmed by methods that allow hypothesis testing through active pharmacovigilance. The US, France and New Zealand take passive reports of ADRs and subject them to a process of formal confirmatory investigation Some of the methodologies used to conduct these confirmatory investigations of ADRs include prescription event monitoring (PEM) and interrogating healthcare databases using data mining techniques.

Drug benefit plans and post market surveillance: In countries such as Australia, Norway, the United Kingdom and France a national authority funds medicines for the whole population (Canada, in contrast, has 19 different provincial and federal public plans that each decide on listing and paying for particular drugs; there is no national Canadian formulary.) The first three countries mentioned have independent or government agencies that scrutinize company clinical trial data to assess the safety, effectiveness and cost effectiveness of new therapies and make recommendations about funding. In France, large scale observational studies are used to inform reimbursement decisions and phase in the introduction of new drugs. The U.S. Veterans Affairs' Center for Medication Safety conducts studies to inform decisions on whether to list a drug on its formulary.

Coordination of regulators and drug plan research: Several countries offer innovative models. In France, the regulator liaises with regional pharmacovigilance centres, and a liaison committee was established to link the regulator and the research centres with the national formulary with respect to post-marketing studies. In New Zealand the regulator (Medsafe) contracts studies to the National Pharmacolvigilance Centre at the University of Otago; the European Medicines Agency (EMEA) is working to establish a network of centres for pharmacoepidemiology and pharmacovigilence; and in the United States, the Centers for Education and Research on Therapeutics (CERTS) and DECiDE (Developing Evidence to Inform Decisions about Effectiveness) centers provide research findings on the post-market use of pharmaceuticals to the Food and Drug Administration. Other innovative models for regulators and health care plans to share research include the arrangement between Veterans Affairs and the Food and Drug Administration in the United States, and the (mentioned earlier) General Practice Research Database (GPRD) in the UK and the Scottish Medicines Monitoring Unit.

Lessons for Canada

A variety of presenters drew the following conclusions, from a survey of international initiatives, about desirable features for a successful post market surveillance system:

• Research networks that:
  • ensure studies are conducted in areas in which manufacturers have a disincentive to investigate (e.g., analyses of whole drug classes for specific indications, head-to-head comparisons)
  • conduct directed research into unanticipated safety problems and report in a timely fashion (such as DECiDE and CERTs in the US, France's Regional Pharmacovilance Centres, and NZ's National Pharmacovigilance Centre)
• Strengthened relationships between regulatory authorities and academic research networks to enhance regulators' capacity to investigate safety and effectiveness issues
• Public oversight of independently conducted post-market research that permits third party review of study protocols, avoids proprietary data conflicts, and allows vetting of industry suggestions to alleviate doubts about the validity of research results
• Phased introduction of new drugs with potential for large scale use. An Only in Research (OIR) assessment can limit the use of publicly funded medicines until 'real-world' safety and effectiveness is determined
• Adoption of risk management plans (RMPs) and post-authorization safety studies as a condition of on-going market authorization could increase industry compliance
• A flexible and enforceable "tool kit" of regulatory options that may be applied at or after approval, e.g., conditions and restrictions on promotion and distribution, postmarketing studies
• Adequate funding to facilitate long-term planning to address emergent issues and threatened access to needed expertise
• Active surveillance to identify and verify the cause of ADRs and unexpected problems. (For example, in the United States, the Food and Drug Administration and Veterans Affairs Memoranda of Understanding to share information attained from VA database mining will enable better targeting of potential safety issues, and responsive feedback to FDA.)
• Regional Pharmacovigilance Centres such as those in France, that offer an important link to clinical care that facilitates prospective observations studies and 'real world' RCTs

Context /issues/needs re: post-market surveillance in Canada

While some post market surveillance activities are being undertaken in Canada, at the federal and provincial/territorial levels, it was evident at the workshop that jurisdictions don't necessarily know what each other is doing. In addition to generating new post-market surveillance research and information, the proposed network of research centres would also serve a coordinating function, facilitating information sharing and minimizing duplication of effort.

The regulator's post-market activities

Health Canada now relies on a variety of sources for information about the safety of the post market experience with prescription drugs. Most of the regulator's efforts are currently concentrated on a passive form of pharmacovigilance due a limitation in regulatory power. They can take action only on specific products and their ability to take action in this regard is currently constrained by legislation (the Food and Drugs Act). Data sources include: Reports of adverse reactions (ADRs) or lack of efficacy; Periodic Safety Update Reports (PSURs); registries (operated by other jurisdictions/groups, not by Health Canada); epidemiologic studies; clinical trials/studies; risk communications from foreign regulatory agencies; scientific literature; industry; international organizations; organizations providing evidence-based reviews of safety & effectiveness (such as the Canadian Agency for Drugs and Technologies in Health (CADTH)); databases & other data retrieval resources; and other Health Canada programs.

Assessment of risk is based on these considerations: International best practices; SNIP Criteria (Strength, Newness (time since the drug went on the market), Importance, Potential for prevention/risk reduction); and what action can be taken (i.e., revise indications, add contra-indications, strengthen warnings, inform public / professionals, suspend or revoke market authorization).

Signals are prioritized and classificated as high priority (serious, unknown and/or unlabelled; if confirmed, an intervention likely); medium priority (if confirmed, a potential shift in benefit/risk profile is communicated); or low priority (ADR already known or partially labeled, related to potential confounders, and/or no evidence of an urgent safety problem).

Other considerations for assessment of risk include: the results of consultation with international experts; critical analysis that takes into account all available scientific and regulatory information; the possibility of making recommendations based on sound science; and causality assessments using the World Health Organization algorithm (certain, probable/likely, possible, unlikely, can't be assessed).

HC's post market risk management

When there is a strong association between a drug and adverse reaction (AR) reports, market authorization can be suspended, a process must be followed that includes: informing the manufacturer of the concern and requesting that information be provided within a specified time: if the deadline is not met, or the information is not satisfactory, market authorization can be suspended. There are other options: Letters to request a voluntary withdrawal by the manufacturer must be signed by the Assistant Deputy Minister for the Health Products and Food Branch. As well, the regulator can request cautionary labeling, but cannot now require risk communications from industry.

The regulator has identified some challenges with respect to using post market surveillance information. These include how to:

• determine the level of evidence needed
• integrate data on therapeutic effectiveness into the risk management decision making process
• ensure timeliness for incorporating data in the post market regulatory decision making
• finding the balance between data on health risk and that on therapeutic effectiveness

The Patented Medicine Prices Review Board and the Common Drug Review

These two organizations evaluate a drug shortly after marketing in Canada and may assist with defining the needs for post-market safety and effectiveness information on prescription drugs in Canada

The Patented Medicine Prices Review Board (PMPRB) undertakes systematic reviews of newly patented drugs that have received market authorization to determine the appropriateness of proposed prices. The purpose of their scientific review is to determine: the primary indication/use of a new, patented medicine; the category of the new patented medicine in order to rank therapeutic improvement ("breakthrough", "substantial improvement", "moderate improvement" and "slight or no improvement"); and comparable drug products already on the market as well as comparable dosage regimens. Based on the results of the scientific review, introductory price tests are applied. The PMPRB bases their reviews on the Oxford Centre for Evidence-Based Medicine Levels of Evidence. An evaluation of therapeutic improvement includes consideration of clinical and economic/ pharmacoeconomic factors. The PMPRB has accepted plans to review the maximum non-excessive (MNE) price based on a change in the science surrounding a patented drug, opening the possibility of a "re-setting" the MNE price. The Board's rationale for this decision (after extensive consultation) is that, after a patented drug has been sold in Canada for say 3 to 5 years, additional clinical trials and/or post market surveillance may provide new evidence to better determine the relative category of therapeutic improvement of the medicine. Re-setting the maximum non-excessive (MNE) price would recognize the real value of the medicine. Rather than develop its own review cycle, it has been proposed that the PMPRB adopt a regulatory life-cycle approach in line with the Progressive Licensing initiative of Health Canada (Bill C-51 and changes to Canada's Food and Drugs Act). A post-market surveillance research network could play a role in ascertaining the ongoing value of the drug in the therapeutic armamentarium. A remaining challenge is for the PMPRB to determine what would be the "triggers" for resetting prices - what information or evidence will be required for reconsideration of an MNE.

The Common Drug Review (CDR) is a program funded by the FPT Ministries of Health that operates under the aegis of the Canadian Agency for Drugs and Technologies in Health. Before the creation of the CDR, Canada's federal, provincial, and territorial drug plans had separate processes for conducting reviews and making formulary listing recommendations. The CDR was set up to reduce this duplication and now provides equal access to high level evidence and expert advice, thereby contributing to the quality and sustainability of the 18 participating Canadian public drug plans. (All public plans participate in CDR, except Québec.)

The CDR's process includes a systematic review of the clinical evidence, and a critique of the drug manufacturer's pharmacoeconomic evaluation. These are prepared by a review team tat is typically comprised of clinical and pharmacoeconomic reviewers, a clinical specialist, an information specialist, the CDR Manager of Drug Reviews, a review methodology advisor, a project advisor, and a project manager. The clinical review consists of a systematic review of published and unpublished trials and also includes supplemental issues such as a backgrounder on the condition being treated, the validity of clinical trial outcomes and relevant safety information. The pharmacoeconomic review is a critique of the manufacturer's economic evaluation.

The CDR is aware of the approach of recommending coverage with evidence development (CED) for a drug when there is uncertainty about the therapeutic effect and cost effectiveness. This category of recommendation is often used with technologies and devices. but internationally there has been little experience applying the approach to prescription drugs. The workshop was told that deciding when and how to apply to approach is challenging and the CDR also recognizes the inherent challenges that such an approach would present to public drug plans, e.g. it is very difficult for a plan to withdraw coverage for a drug after it has been grants. However, some of the questions developed to deal with CED are instructive with respect to post market surveillance:

• Will collection of additional information in a real world setting address the uncertainty?
• What is the best trial design to collect the required information e.g., the use of randomized clinical trials (RCTs), observational trials or the use of registries?
• Can CED provide the required information within a reasonable timeframe?
• Are the costs of CED worth the potential for better information (value of information analysis) on the benefit/risk profile and cost-effectiveness?

Public Drug Plans

Canadians covered by all the public drug plans (five federal plans and 14 provincial/territorial, including Quebec's) will potentially benefit from post market surveillance of drugs. However, the federal plan that covers drugs for First Nations and Inuit faces particular challenges. For one thing, the plan, which is payor of last resort for First Nations and Inuit peoples, interacts with all the federal/provincial/ territorial plans and so faces the challenge of sorting out responsibilities with multiple players, each with different coverage schemes. But more importantly, the population covered by the federal plan has a higher than average burden of ill health, including a very high rate of diabetes. Post market surveillance of the use of prescription drugs for that condition alone would be particularly useful in helping define interventions to relieve the burden of illness. However, there is a history of sensitivities about collecting data and other specific administrative challenges in this population.

Provincial drug plans: producing and using post-market surveillance information

Provincial initiatives, such as those described in the following section, provide examples of what the proposed network would do on an ongoing basis. As well as producing evidence, the network will help to coordinate efforts and share information, making better use of scarce resources and minimizing duplication.

British Columbia: Experience with dementia drugs. Background: All provinces except B.C. and Newfoundland provided some coverage for cholinesterase inhibitors (for Alzheimer's dementia). At least 3 recommendations from B.C.'s Drug Benefit Committee of Pharmacare between 1999 to 2005 were against providing benefit status and, subsequently, the drugs were not covered in that province during this time period. The reasoning was the limited randomized controlled trial evidence of clinical effectiveness, and this lack of evidence was also acknowledged in the US, Germany and the UK, where the National Institute for Clinical Excellence (NICE) had recommended against National Health Service coverage of the drugs. Even among those arguing for positive effects, there is widespread recognition that the effects are small and that the clinical significance may be marginal.

However, the cholinesterase inhibitors are the only drugs for Alzheimer's dementia and the Alzheimer's Society and British Columbia geriatricians encouraged a three-year trial to evaluate these drugs in the "real world". Five projects have been initiated with the goal of providing sufficient information to Pharmacare to rewrite (if appropriate and according to the evidence produced) the terms for provincial eligibility for coverage. The five projects are: a cost and utilization project, clinical epidemiology project (with a link to the drug data base to provide information on other drugs that patient are taking), a Randomized Controlled Trial (without a placebo since BCMA would not accept a placebo arm), a clinical assessment of dementia, and a caregiver appraisal (to compare their assessment of drug impact with the clinical evidence). The deliverable for the project is to rewrite the formula for Pharmacare coverage for the drugs. The projects are a joint initiative of the Universities of Victoria and British Columbia, B.C.'s Pharmacare program and, at arm's length, industry, which is providing the first three months of the drug at no cost. The projects, however, ran into some political obstacles (see below, obstacles) once they became operational.

Nova Scotia: Making decisions about the use of anti-cancer drugs. Background: A comprehensive review of cancer drugs was launched in order to make fair and reasonable coverage decisions. At issue was the fact that in recent years, the cost of such drugs had been increasing at an annual rate of 25-30%. A Cancer Systemic Therapy Policy Committee (CSTPC) was formed to assist in allocation of limited provincial funding for expensive cancer drugs. Despite the fact that the drugs typically afford small treatment benefits, and that there is a lack of robust clinical and economic data, there is intense public pressure to fund the drugs with little or no concern for sustainability. Therefore, it was felt that a decision-making framework with broad input was necessary, and that decisions about allocation of the expensive drugs should be based on a combination of evidence, economics, and ethics. CSTPC members included public/community members, persons living with cancer, health care providers/researchers, a health economist and health care ethicist as well as participants from Cancer Care Nova Scotia, the Nova Scotia Department of Health, and Public Health. The values and principles that were established to guide decision making were beneficence (minimizing harm), health equity, efficiency, sustainability, and justice.

As part of the process a local expert gives a clinical presentation to the committee - providing information about the cancer, the therapy, the results of studies and existing clinical guidelines. He/she cannot discuss money or funding or give an opinion about funding. The Committee's health economist provides an 'understandable' plain language summary of his/her conclusions/opinions arising from a critical appraisal of the best available pharmacoeconomic analysis(es) of the cancer therapy . There is also discussion of other relevant information, such as: particular social groups with high risk of the cancer and/or increased vulnerability if the therapy were not to be funded; current status of funding in other jurisdictions, e.g., other provinces, UK, Australia; and the present provincial and Canadian 'social consensus' with respect to public funding of this and similar cancer therapies, if known or determinable. The committee then considers recommendation options, which include: approval of funding for use of therapy as per clinical guidelines established by the relevant cancer site team; approval of funding for use of therapy with further restrictions; approval of 'in-between' options, such as partial coverage with amount determined by sliding scale of income and/or other indices of disadvantage/vulnerability; and denial of coverage (which may lead to further action or attempts to negotiate the drug cost downwards with pharmaceutical company provider). The options are analysed in terms of benefits and burdens (e.g., drug-only cost per patient per median therapy duration; anticipated human and infrastructure resource costs; cost per gained QALY; and budget impact analysis). At this point, a member of comparator analysis working group provides a brief summary of actual (or projected) costs of selected, comparable (funded and non-funded) cancer and non-cancer therapies, and, as appropriate, early intervention initiatives, e.g., non-funded screening programs for the cancer.

Recommendation options are identified and evaluated (a facilitated ethics dialogue based on the Committee's underlying values is included as part of process) and voting is secret (by an electronic ballot). The chair then prepares for the Deputy Minister a report with a summary of the key deliberations, the voting results, the committee's recommendations and, if appropriate, dissenting opinions. A suggested communication strategy and relevant briefing notes are also included. In at least one recent situation, the elected Minister of Health chose to override the decision-making process and provided funding for a drug not approved by CSTPC.

Ontario: Drug Innovation Fund for projects to fill in the gaps. Background: The Ontario Public Drug Program has tried to implement "listing agreements" and "trial programs" in order to manage the data gaps that are inevitable after market approval. Listing agreements require additional data collection by the manufacturer in exchange for listing the drug for payment under the public plan. However, there are several concerns regarding this policy arrangement. For one thing, observational study data were often not at the level needed for making decisions. Ideally, a randomized controlled trial would be necessary to support the observational data - in order to provide the level of detail needed by government -- but it is difficult to segregate benefits in a public drug plan. Also, if the results from observational data were negative and if the drug were for a chronic condition, there could be hundreds of thousands of Ontarians being covered who would be affected by a decision to de-list the drug. In the Ontario government's opinion, it is easier to restrict access (i.e., to make listing changes or require restrictions) from a safety perspective than for reasons based on cost-effectiveness or the place of the new drug in ongoing therapy for a specific condition. These issues may become more challenging with the advent of progressive licensing and increased post marketing surveillance of drugs and, by extension, conditional listing policies.

In trial programs, under which reimbursement is based on an individual or group of individuals attaining a clinical effect, it proved to be difficult to reconcile funding based on patients attaining the benefit of the drug in question. Ontario used one such trial, with respect to cholinesterase inhibitors for Alzheimer's, but has no plans to use more trial programs since these involved the difficult challenge of how to give access to a benefit and seek to remove it. A Citizens' Council has been recently announced in Ontario to assist with some of these difficult policy decisions and it will be interesting to see how the process works and what advice or recommendations it will suggest.

Finally, based on the above two experiences, Ontario has put its resources towards a Drug Innovation Fund. Drug manufacturers often claim that use of their drugs will lead to downstream savings in the health care system, and the province hopes to investigate this claim by generating strong, high-quality, independent scientific evidence on the impact and value of new and existing drugs across the health care system. As well, it is hoped that funded projects will support linkages between researchers, clinicians and drug policy decision makers with an aim to make better decisions and, further, to support and develop research capacity in the area of drugs and health outcomes. Last year $3.8-million was allocated for studies with a focus on drug access and utilization, optimal use of drugs and drug adherence.

Issues to be resolved

In general: Workshop presenters indicated that some of the obstacles to undertaking effective post market surveillance of drugs include pressures from the pharmaceutical industry, the media and the public in general or patient advocacy groups, specifically, and political mandates. For example, the B.C. initiative to study dementia drugs faced opposition when the provincial premier decided he would announce the project and ruled that it could not be discussed before his announcement. As a result, doctors in the province were taken by surprise (they could not be informed in advance) and reacted by demanding payment for their participation (filling out forms) since this was a research project. Other "big picture" issues include privacy concerns, and need for cooperation between regulators and the governments that pay for prescription drugs.

Canada specific: The various players in Canada (federal and provincial/territorial governments, the PMPRB etc.) have different needs with respect to the type of post-market surveillance information and research required, as well as its timing and availability. This is a hurdle for leveraging Canada's post-market surveillance strengths in personnel and data. Determining the funding for post-market surveillance activities is also an issue. As well, there are the abiding issues of the need for coordination and cooperation among federal and provincial jurisdictions, and with other public agencies like Canadian Agency for Drugs and Technologies in Health and the Canadian Patient Safety Institute to negotiate different interests and needs towards the same end: better and safer use of prescription drugs. This is especially true in light of the effects that progressive licensing (C-51) will have on the life cycle of a drug product in Canada.

With different and, at times, competing interests at stake, the lack of a sense of ownership with respect to post-market surveillance research and its findings will be a major obstacle to moving the idea of a research network of centres of excellence forward. Regulators and the market authorization holders (pharmaceutical companies), the provinces and payors all have different perspectives on how post-market surveillance research should be done, who should have responsibility for carrying it out, and how the resulting evidence should be used.

For the regulator, Health Canada, a key challenge is integrating data on therapeutic effectiveness into the risk management decision-making process. Decisions on risk-management under new legislation (Bill C-51 and progressive licensing) will have to follow on the risk-management plans as developed and operationalized by the drug company. Health Canada will have to be prepared and have the ability and expertise to critically evaluate a company's risk management plan and to ascertain the levels of evidence needed to inform on any decision that needs to be made about a drug's safety and effectiveness.

Other issues that need to be resolved include the need to find out how to:

• undertake contemporaneous assessment of safety and therapeutic effectiveness so that it is relevant and helps to define the benefit / harm profile of a drug
• decide who does the work, who asks and prioritizes the questions and on what basis these decisions are made
• ensure timeliness with respect to answering the questions based on the data available (e.g., ADR signal assessment, setting up real world clinical trials, etc.)
• establish and maintain transparency of the post-market surveillance process
• eommunicate the process and outcomes of post-market surveillance research with the public and other key stakeholders

Technical issues with post-market surveillance evidence. A number of methodological and technical issues were addressed with respect to producing and using post-market surveillance evidence.

• Multiple sources of drug-related research data are possible: expert opinions, clinical trials, case reports and case series, clinical registries, systematic reviews and meta-analysis, secondary databases. It becomes a challenge then to ascertain which is the most appropriate source(s) of data for the post-market surveillance question being asked.
• It may be necessary to establish a rule of hierarchies of evidence although, despite limitations, all data sources are of some value and their worth will depend on the post-market surveillance research question being asked and the strength of synthesis used.
• We will need to determine how differences between randomized controlled trial (RCT) evidence and real world safety and effectiveness (RWSE) evidence should be interpreted or leveraged upon to pose new post-market surveillance research questions.
• The resolution of conflicting evidence must have a process for consideration. For example, when clinical/cohort studies showed both a decreased and increased cardiovascular mortality following Hormone Receptor Therapy use, the need for interpretation was paramount for informed care for women post-menopausally.
• A determination of how best to study what's happening in the real world needs to be made and may depend on the question being asked: should intervention trials, cross-sectional time-series, and/or case-control study designs be used?
• A number of key challenges present themselves to researchers charged with undertaking post-market surveillance research: Understanding data sources (local expertise); understanding and applying appropriate methodology; creation of infrastructure and data access; governance and operational models and the need for high quality information in a timely manner.

Moving forward

Workshop participants were divided into two groups and asked to consider two questions.

Public drug plan group

This group was to consider: 1) What are the obstacles (barriers) from a practical perspective to using evidence on drug safety and effectiveness in drug plan management after drugs are approved for market? 2) How can these obstacles be overcome?

1) Obstacles:

The lack of flexibility and agility in the system to respond to information (for example, how to stop paying for a drug based on post market evidence).

Having/developing expertise in the area: drug plans are not set up to deal with signals/safety issues after a drug is listed on the formulary.

Inconsistency in formulary policy and management across the country: not all plans have the capacity to approve and fund limitlessly and many plans are looking for ways to make the decisions clearer for a subset of patients needing certain interventions. How resources can be coordinated among larger and smaller drug plans.

Privacy and confidentiality rules can be an obstacle to transparency.

How to balance conflicts of interest.

Balancing overall societal interests with those of individuals, and managing expectations of different groups (for example, the issue of patient demand and the government's response for access to a drug even when biomedical evidence is lacking or negative).

There is a lack of alignment between clinical practice guidelines and reimbursement guidelines.

Policy constraints exist to responding to post market information.

Regulators look at one drug at a time, so how can the safety and effectiveness of a class of drugs be determined (e.g., even CDR does not have comparator information when they make their recommendations).

2) Overcoming obstacles

Transparency in making decisions and education of the public, the media, patients, and politicians (often patient pressure goes to politician) about the issues. This education should facilitate the public's acceptance of the decision making process.

Involvement of the public in drug plan decisions: for example, Ontario's new advisory body, a Citizens' Council or the use of citizen juries.

Speedy response to media reports. In the United Kingdom, independent science media centres draw on experts who agree to respond within two hours to provide comments and context on news events and reports.

A process for bringing (drug impact) signals together, for evaluation and application to the drug plan.

Independent guidance to doctors, pharmacists and the public about how a drug should be used when it is released.

Using the network to collectively discuss available evidence.

Establishing a network of drug plan managers (the information users) to facilitate use of evidence from the proposed drug effectiveness and safety network.

Include a requirement for pharmacovigilance data in provincial electronic health record initiatives and in Canada Health Infoway, thereby helping to facilitate "surveillance by design."

Use of an Australian example that has patients tried on conventional therapy before becoming eligible for new (high cost) biologics. (Hence the ability to find those most likely to benefit from a new drug, and track the adverse drug reactions amongst them.)

Regulatory context group

This group was asked to consider: 1) what is the best mix of safety and effectiveness data (appropriate balance) to strengthen the drug regulatory decision-making practice? 2) what level of value added is there to look at safety and effectiveness data for subgroups of the population who are at risk of adverse effects or any other health product related problem?

1) Safety/effectiveness mix to strengthen decision making

In order to have the needed information on benefits and risks associated with a specific drug, consistent appropriate methodologies should be used. For example, if one tries to look at benefit/risk data across different drug classes, or across different subpopulations, the data may not be derived using similar enough methods to draw any conclusions.

Alignment to international standards is important if they exist, but frequently there are none. Canada may have an opportunity to lead, given the data characteristics produced by our system. We may be able to help facilitate the development of international standards of post market surveillance.

Need transparent and evidence-based decisions. The transparency with respect to the balance of benefit risk information used in the decision-making process is important so those who see the decision would understand how much information was available, how it was used and how the decision was made.

Where possible, standardized consistent use of certain data sets to address certain types of questions would facilitate identification of gaps and would feed back into integrated health decision making. We do not have a general physician research data base in Canada like in the UK, where they resource a set of physicians to make sure decision makers have a consistent source of clinical data. The rationale for having something like this in Canadian practice settings might be stronger and its creation more likely if there were systematically identified gaps in care.

There needs to be communication and established information-sharing pathways between researchers and decision makers. These pathways could be facilitated by a network of centres sharing the finding of this type of work and using a mix of safety and effectiveness data in an appropriate manner to make transparent, evidence-based decisions. One of the problems, from a federal regulatory perspective, is that when federal regulators try to answer a safety issue, they are often criticized for not consulting with agencies such as the Institute for Clinical Evaluative Sciences, or experts in the other provinces. But the fact is the regulator does not know what research is underway, so having better a communication network would facilitate methods development, best practice sharing, and leveraging of policy decisions re: privacy regulations, data base linkages, and gap awareness.

2) Looking at safety/effectiveness data for subgroups of the population

In Canada there are challenges facing our ability to collectively answer scientific questions that would inform on the benefit/risk profile of pharmaceuticals, but our ability to do so is even worse with respect to answering questions of import about certain subpopulations (children, small population cancer, various metabolic diseases, personalized medicine issues, First Nations/Inuit clients and others). Having a more clearly identified process and an agreement on best practices, as well as "go- to" people around certain questions with respect to subpopulations, would help. With small subgroups of patients with cancer, for example, it is often imperative to collect national data to have a reasonable sample. We need to identify the level of evidence required, and whether this can be achieved. The importance of having a national perspective on the types of information needed to make decisions is key. Sharing these discussions nationally would facilitate any network to produce usable research to answer scientific questions. Another example of the potential benefits of a network of centres is the possibility of conducting head-to-head comparisons of therapies.

The design of the fields used in electronic health records has to be considered so the information will be useable to answer the kinds of scientific questions we are asking. There is concern that there are disconnects among the provinces and territories and decision makers in this regard. There may be an opportunity for the proposed network to identify gaps and collectively bring data needs forward. With respect to off-label use, there can be problems tracking information because of differences in provincial and territorial data sets. In this case, it may be possible to answer questions by specifically designing studies of a subpopulation, then applying the findings more widely.

In general, subgroups of interest should not be limited to those based on disease and/or demographics, but should also include those to whom prescription drugs have been appropriately, and not appropriately, prescribed.

Conclusion

The regulatory and the drug plan participants recognized that this workshop was a first step towards figuring out how to make optimal use of the evidence gathered by a post marketing surveillance network. Some of the key ideas that emerged included recommendations to:

• include the requirement for pharmacovigilance data in provincial electronic health record initiatives and in Canada Health Infoway, in order to encourage "surveillance by design" and
• establish information sharing pathways between researchers and decision makers. A better communication network would facilitate method development, best practice sharing, leveraging policy decisions with respect to privacy legislation, data base linkages, and gap awareness.

Finally, there may be greater willingness on the part of decision makers and the public to listen to the collective voice of an arm's length, independent drug effectiveness and safety network, as compared to the voice of individual stakeholders.

Appendix: list of participants