Evaluating the quality and use of economic data in decisions about essential medicines

Abstract Objective To evaluate the quality of economic data provided in applications to the World Health Organization (WHO) Model List of Essential Medicines and to evaluate the role of these data in decision-making by the expert committee that considers the applications. Methods We analysed applications submitted to the WHO Expert Committee on the Selection and Use of Essential Medicines between 2002 and 2013. The completeness of data on the price and cost–effectiveness of medicines was extracted from application documents and coded using a four-point scale. We recorded whether or not the expert committee discussed economic information and the outcomes of each application. Associations between the completeness of economic data and application outcomes were assessed using χ2 tests. Findings The expert committee received 134 applications. Only eight applications (6%) included complete price data and economic evaluation data. Many applicants omitted or misinterpreted the economic evaluation section of the application form. Despite the lack of economic data, all applications were reviewed by the committee. There was no significant association between the completeness of economic information and application outcomes. The expert committee tried to address information gaps in applications by further review and analysis of data related to the application. Conclusion The World Health Organization should revise the instructions to applicants on economic data requirements; develop new mechanisms to assist applicants in completing the application process; and define methods for the use of economic data in decision-making.


Introduction
Essential medicines are those that satisfy the priority health care needs of the population. 1 The World Health Organization (WHO) introduced the first Model List of Essential Medicines in 1977. Countries are encouraged to use the model list as a guide for their decisions on pharmaceutical selection and procurement. Between 1977 and 2007, over 130 countries introduced national lists of essential medicines, modifying the WHO model list for their national context. 2 The model list is updated every two years, following an application and review process by an expert committee. Any individual or institution may submit an application. Before each committee meeting, applications are published on a website for public comment and experts (committee members and external advisors) review each application. 3 These reviews and any public comments are also published on the website before the committee meeting. Since 2002, essential medicines have been selected via an evidence-based process, with due regard to public health relevance, evidence on efficacy and safety and comparative cost-effectiveness. 1 Detailed information on the decision criteria can be found at http://www.who.int/selec -tion_medicines/en/. The model list is divided into a core list and a complementary list. The core list includes medicines that meet the criteria of efficacy, safety and comparative cost-effectiveness, 1 but expensive patented medicines are not necessarily excluded. 3,4 The complementary list includes medicines that may require specialized facilities, that are consistently more costly, or less cost-effective in a variety of settings. 5 Additions to the model list can have a major impact on global and national decisions, with significant budgetary, ethical and health implications. 6,7 Other studies have examined application processes and criteria for national decisions on adopting new vaccines, 8 national health technology assessment programmes 9,10 and financing decisions for health technologies. 11 A recent study found that applications to the model list for mental health medicines generally provided low-quality and incomplete evidence across several required dimensions. 12 There is no global consensus on how to use economic data in decision-making for medicines and health technologies. [13][14][15][16][17] Here, we evaluate the extent to which applicants and the expert committee adhere to the instructions and procedures on economic data and analysis for applications to the model list. We assess the application process rather than the substance of decisions (specifically, compliance with instructions for applicants and the quality of data on price and on cost-effectiveness). We also assess whether economic considerations are included in the final report by the expert committee. The overall goal of this study is to improve the quality, transparency and clarity of the process for reviewing applications to the model list.

Data set
We included applications to the model list for medicines intended for use in adults. We analysed final reports from the twelfth (2002) to nineteenth (2013) meetings of the WHO expert committee; the sixteenth meeting reviewed paediatric medicines only and was not included. We did not review applications for new formulations of existing medicines or applications for reinstatement.

Definition of variables
We assessed the extent to which applicants complied with the instructions provided. The instructions ask applicants to provide "a range of costs of the proposed medicine" that Objective To evaluate the quality of economic data provided in applications to the World Health Organization (WHO) Model List of Essential Medicines and to evaluate the role of these data in decision-making by the expert committee that considers the applications. Methods We analysed applications submitted to the WHO Expert Committee on the Selection and Use of Essential Medicines between 2002 and 2013. The completeness of data on the price and cost-effectiveness of medicines was extracted from application documents and coded using a four-point scale. We recorded whether or not the expert committee discussed economic information and the outcomes of each application. Associations between the completeness of economic data and application outcomes were assessed using χ 2 tests. Findings The expert committee received 134 applications. Only eight applications (6%) included complete price data and economic evaluation data. Many applicants omitted or misinterpreted the economic evaluation section of the application form. Despite the lack of economic data, all applications were reviewed by the committee. There was no significant association between the completeness of economic information and application outcomes. The expert committee tried to address information gaps in applications by further review and analysis of data related to the application. Conclusion The World Health Organization should revise the instructions to applicants on economic data requirements; develop new mechanisms to assist applicants in completing the application process; and define methods for the use of economic data in decision-making.
reflect "average generic world market prices", with a clearly specified source of data. The WHO-recommended sources for these data include the International Drug Price Indicator Guide, the United Nations Children's Fund, Médecins Sans Frontières or WHO itself. We allocated one point for each of the following: price given; offered range of price or average or median price; clear source; WHO-recommended source. The application was classified as complete if the application attained a score of four; applications with scores between one and three points (inclusive) were classified as incomplete. If no pricing data were provided, this was considered missing (zero points). If the applicant explicitly mentioned that no pricing data exist, this was classified as not available.
Completeness of economic evaluation data was scored based on WHO instructions for comparative cost-effectiveness presented as range of cost per routine outcome, and a clear citation, preferably of WHO-recommended sources. We also accepted citations from the published scientific literature as one of the recommended data sources. Adherence was classified as complete if applications met all of the following criteria: economic evaluation value provided; comparator given; clear source; WHO-recommended source. As with price, applications scoring between one and three inclusive were classified as incomplete. Omitted economic evaluation sections were classified as missing; cases where the applicant explicitly noted the lack of available data on economic evaluation were classified as not available. There were instances in which the applicant provided administration costs for medicines, without an outcomesbased denominator, under the heading of cost-effectiveness. These applications were classified as having provided economic evaluation data that were not applicable.
Based on the final reports from each meeting, the outcome was coded as a dichotomous variable: either rejected or deferred (coded 0) versus accepted to the core or complementary lists (coded 1). Applicant type and therapeutic class were categorized based on information provided in the application. We recorded the proportion of applications reviewed by the committee that were added to the model list, by therapeutic class. We also recorded whether the discussion and decisions mentioned price or economic evaluation.

Data entry and analysis
A data entry form was generated based on the variables listed above. There were two independent data reviewers, each of whom extracted data from the applications; discrepancies were resolved by consensus. Application data were entered using CSPro version 5 (United States Census Bureau, Washington, United States of America).
Qualitative information from discussions and application decision categories were coded by hand. We extracted the narrative sections about the reviewed medicines from the final report for each committee meeting; text concerning economic considerations was collated in an Excel spreadsheet. We identified whether the report mentioned economic data or the absence of such data, as well as which data sources were cited (i.e. the application itself or new additional data).
Associations between the completeness of economic data and the application outcomes were assessed using χ 2 tests in Stata version 12.1 (StataCorp. LP, College Station, USA). We conducted sensitivity analyses that dropped each scoring criterion in turn.

Results
We analysed 134 applications for new medicines from 2002 to 2013 (Table 1). Application authors varied by year, but were most commonly submitted by WHO departments (25.4%; 34/134), academic institutions (22.4%; 30/134) and nongovernmental organizations (21.6%; 29/134). Therapeutic classes of medicines also varied between meetings; applications for cancer and diabetes medicines were more common in recent years. Of 134 applications analysed, 52 were rejected and decisions were deferred for an additional 10. Among the 72 applications added to the model list, 67 were added to the core list and five to the complementary list.
Across all meeting years, only 6.0% of applications (8/134) provided complete price and economic evaluation data ( Table 2). Since no time trends were apparent, we report totals for all years; analyses by year are available from the corresponding author. If we include applications that reported economic data were not available, assuming that these applicants made an effort to include data but that none were available, 13.4% (18/134) of applications fully followed the instructions on price and economic evaluation. Only 20.9% (28/134) and 18.7% (25/134) of applications provided complete price and economic evaluation information, respectively. Over a third (36.6%; 49/134) included no economic evaluation information, one-fifth (20.9%; 28/139) stated that economic evaluation information was unavailable and 17.9% (24/134) presented financial information that was unrelated to economic evaluation (e.g. total treatment cost).
All applications were reviewed by the committee. For the applications with complete economic data, 52.0% (13/25) was added to the model list. When we included applications that stated that economic data were not available, 50.9% (27/53) was added to the model list. Among applications with no price data or no economic evaluation, over one-third and over half, respectively, of these applications were added to the model list.
There was variation in the extent of economic information provided by applicant type. Academic authors were most likely to provide complete price and economic evaluation information. Applicants from WHO did not include economic evaluation information in half of their applications and industry applicants very rarely provided complete information on price or economic evaluation. Detailed data on the completeness of economic information by applicant type and therapeutic class are available from the corresponding author.
For applications with complete price data, 64.3% (18/28) were added to the model list, a percentage that was not significantly higher than for applications with incomplete or absent price information (50.5%; 50/99; P = 0.2). For applications with complete economic evaluation information 52.0%, (13/25) were added to the model list, compared to 54.1% for those with incomplete, missing, inapplicable or unavailable information (59/109; P = 0.9).
We explored whether the committee's discussions of economic factors reflected the content of the applications. As shown in Table 3, price data or economic evaluations were discussed in some cases, even where this information had not been included in the application. Our qualitative analysis suggests four Research Essential medicines Corrina Moucheraud et al.
ways in which this happened. First, the committee sometimes noted the lack of information in the application (two applications with missing price data, seven applications with missing economic evaluation); in three of these cases, the medicines were nonetheless added to the core list. Second, the committee sometimes referenced economic information, suggesting an alternative, although unspecified, source of supplementary data (six applications). Third, the committee sometimes did further research: for example, the report mentioned a review of cost-effectiveness data prepared by the secretariat, or indicated that the secre-tariat (usually comprised of WHO staff who support the committee process) performed its own economic evaluation (four applications). Fourth, in two instances where an economic evaluation was missing, the committee discussed price data included in the application but referred to it as "cost-effectiveness," suggesting a misinterpretation.
We explored whether the distinction between core and complementary lists was followed in the decision process. A feature that distinguishes the complementary list is its inclusion of medicines requiring specialized diagnostic or monitoring facilities and/or specialist medical care and/or specialist training. 5 According to our analysis, 50% (46/92) of applications that we classified as belonging to the complementary list were nonetheless added to the core list by the committee. This included all psychiatric medicines and antiretroviral medicines for treating human immunodeficiency virus infections.
We conducted sensitivity analyses, to examine the relative importance of each scoring criterion. The most influential variables were: for price data, the average generic world price and for economic evaluations, the choice of comparator (i.e. presenting relative or   WHO: World Health Organization. a If applicant explicitly noted that price and/or economic evaluation data did not exist for the medicine, these were classified as not available. To be classified as missing, the applicant did not provide any price or economic evaluation information. Applications classified as not applicable were those in which the economic evaluation sections did not include outcomes-based measures (i.e. neither cost-effectiveness, cost-benefit nor cost-utility analysis).

Discussion
Very few applications complied fully with the instructions on providing price data and economic evaluations. The quality of information provided in applications to the Model List of Essential Medicines can be improved substantially. Despite the majority of applications being submitted without the required economic information, the committee reviewed all applications. There was no significant association between the completeness of economic information and the outcome of applications: provision of economic data in an application is not necessary for a positive decision by the committee.
Our qualitative findings suggest that in some cases, the committee found alternative ways to address information gaps -such as internal literature reviews and analyses. However, such data collection and analysis requires additional time on the part of the committee and reviewers. Since the committee still considered and approved incomplete applications, this may have reinforced applicants' decisions to not comply with instructions. There is little evidence that applicants have improved their efforts to submit requested information over time.
There are several possible reasons why applicants often did not provide complete or robust economic data. First, this information may be hard to collect: high-quality price and economic evaluation data that are relevant for resource-poor settings may be difficult to find in the literature; applicants may lack the technical skills to conduct their own economic evaluation. Such issues should, however, be less of a barrier for applicants from private companies, who have access to proprietary data, but applications from companies had some of the lowest rates of provision of economic information. Second, the application instructions may be unclear or insufficiently detailed, as suggested at a recent committee meeting. Application instructions for the recent twentieth committee meeting were revised, 18 but still do not fully incorporate recent recommendations, such as asking applicants to provide comprehensive search strategies or stating that only complete applications would be reviewed. Third, the committee itself may not be clear about whether economic studies should be a necessary component of the decision process. Our qualitative analysis suggests that other considerations -such as safety and tolerability -are sometimes given higher priority even in cases of unfavourable price or economic evaluation data. However, it is often unclear which criteria were emphasized during the committee's review and decision process. There are also chal-lenges in using economic criteria, such as comparative cost-effectiveness, in decision-making for medicines. These challenges include information deficiencies and a lack of universal standards on appropriate thresholds.
Our study has some limitations. We did not assess the accuracy of information provided and when an applicant noted that no data were available, we did not attempt to verify this. Our analysis of committee discussions and decisions was limited to the publicly-available meeting reports, which may not capture all aspects of the meeting. To overcome the various data limitations, we relaxed certain criteria in the application instructions to allow more flexibility in assessing adherence to the instructions: for example, we expanded the average generic world market price criterion to permit inclusion of median prices and prices per country-group (e.g. low-income countries); we allowed other published data sources (e.g. peer-reviewed manuscripts) or internal sources (e.g. clinical trial dossiers) for economic evaluation.
In conclusion, we have three recommendations to improve the review and decision-making process for the Model List of Essential Medicines. WHO should: (i) provide clear and detailed instructions about how much and what kind of economic data are required; (ii) develop mechanisms to assist applicants in completing an application; and (iii) provide clear rules about how economic data will be used in making decisions about applications and about the consequences of not providing economic data (Box 1). To implement these recommendations, the current application form needs to be revised and examples of high-quality applications and an interactive, guided, application process should be provided. Many countries are debating how to use economic data and analysis in decision-making processes about essential medicines. WHO   a) For price data, the current instructions ask applicants to provide a "range of costs of the proposed medicine and to show medicine prices from a range of settings where the product is registered. " This should be clarified: a range of unit prices (not costs) should be provided for specific countries and for specific sellers and data should include low-and/or middle-income countries whenever possible. The application should clearly identify the source of price data, the years for the data and conditions that apply to the prices (bulk purchasing, payment method, etc.). In addition to unit prices, the applicant should provide meaningful per-patient prices, for example, per treatment duration or per full vaccination course. If no price data exist, this should be explicitly noted as such and the efforts to find price data should be documented. c) For both price data and economic evaluation information, the distinction between required and optional information should be explicitly stated in the application instructions. For price data, required items might include unit and clinical prices (e.g. per course of treatment); and for economic evaluation, clear outcomes and comparative, incremental analyses (e.g. incremental cost-effectiveness ratios).
d) For price data reported directly in the application and for any price data used in de-novo economic evaluation calculations, WHO should seek to extend its partnerships with agencies that collect economic information -currently WHO recommends data sources such as the International Drug Price Indicator Guide and the United Nations Children's Fund. This is a longer-term goal, but would provide a standardized, high-quality source for applications that present price data.
2. WHO should develop mechanisms to assist prospective applicants in completing an application. We recommend four new mechanisms, which may also require additional financing.
a) The application instructions should be reviewed using focus groups and feedback from past applicants and reviewers, to improve the clarity of instructions around data types and formats (as suggested in recommendation 1a-d).
b) Examples of high-quality, complete applications from the past should be provided to prospective applicants, to illustrate appropriate responses to specific sections of the application instructions.
c) An interactive, guided, form-based application process, with detailed application sections and mandatory fields, could be developed to improve the process of applying and assure that all sections of an application are completed before submission. d) WHO should develop an online training course and require that this course be completed by applicants before submission of a new application. This would include information about the model list, its objectives and role, as well as technical guidance in completing a high-quality application, including identifying, interpreting and using economic data. The World Health Assembly passed a resolution to improve policies and practices for adopting national-level model lists, including promoting collaboration and information-sharing about best practices for selection procedures20; WHO could extend such technical assistance and education programmes to aid national decision processes as well.
3. WHO should provide clear rules about how economic data will be used in making decisions and about the consequences of not providing economic data.
a) Incomplete applications should be identified as such and should not be reviewed. Applicants should be informed that their application is not complete and will not be reviewed unless all instructions are met, including providing specific price data and economic evaluation studies as identified via recommendation 1c.
b) All criteria used to guide application assessments should be accounted for in documentation of the committee' s decisions: just as a structured application form/ process could help ensure uniform provision of information, a template of the process for committee decisions could increase transparency. For instance, the committee report should provide a brief snapshot summary of information provided in terms of the medicine's safety, efficacy, quality and economic profile. A model for this already exists, in the expert review process before the committee meeting -and this type of form could be expanded and incorporated into the committee decision process. There is a short timeline after each meeting in which the report is generated (with decision justifications, etc.); standardized reporting might facilitate a quicker and more efficient process for writing the report, but rapid turnaround should be considered as an important factor in adopting such a change.
c) WHO should provide clear decision rules about what information and criteria are used to assign an application for consideration under the core list or the complementary list, including the role of economic data in doing so. These determinations should also be clearly accounted for in documentation of the committee's decisions.
d) All documents used in the application process before the committee meeting should be publicly available on a WHO website, including initial application, all expert reviews, any application revisions, all public comments and all additional information sought from applicants or others. All information used in the process should be fully documented and publicly available, including details of the committee' s decision process, including assessment of information provided as per items 3a-b and any additional data or analyses prepared by the Secretariat, to assure full transparency about how decisions are made and to strengthen the evidence base. The committee should continue to require only publicly-available information in applications.

Research
Essential medicines Corrina Moucheraud et al.