Assessing the economic impact of Rx-to-OTC switches: systematic review and guidelines for future development

Abstract

Introduction:

Switching drugs from prescription to non-prescription status (Rx-to-OTC) presents a unique set of challenges and opportunities to policy-makers and the industry in terms of managing health outcomes, pharmaceutical spending, and steering of consumer choices of therapy. Decision-analytic models are used to address uncertainty and produce reasonable estimates of the economic impact of switches for payers. This article presents a critical literature review of existing models which assess the economic impact of Rx-to-OTC switches, and provides guidelines in which future economic evaluations of Rx-to-OTC switches could be improved.

Methods:

A comprehensive search strategy was implemented in Medline and Embase, to retrieve published economic evaluations on Rx-to-OTC switches from 1995–2010. The research digest of the International Society of Pharmacoeconomics and Outcomes Research (ISPOR) was reviewed for potentially relevant abstracts for the past 3 years. Each model used was critically evaluated in terms of structure, relevance of inputs, methodology used, and robustness of results.

Results:

Worldwide, the economic impact of Rx-to-OTC switches has only been evaluated in a total of 12 peer-reviewed publications. Ten out of 12 studies were US-based, and two European-based. The models covered various disease categories, including allergy, hypercholesterolemia, gastroenterology, contraception, pulmonology, and virology. Seventy-five per cent of the models predicted cost savings for payers and patients. Limitations of the models mainly included use of strong assumptions and non-inclusion of specific populations due to lack of data. Guidelines were developed to help future model development. They cover structural issues on decision context, health states, and clinical outcomes, and other considerations for model specifications.

Conclusions:

Although reviewed studies lacked quality, this review of economic evidence of Rx-to-OTC switches suggests that switches may produce cost savings to public and private payers. This is especially important in light of the trend towards more switches.

Keywords::

• Prescription
• OTC
• Rx-to-OTC switch
• Policy

Introduction

Health insurance typically pays for products when they are prescription products; however, when products move to non-prescription status (also called over-the-counter (OTC) status), they are no longer covered by insurance companies, and patients must pay the full price for OTC products.

Rx-to-OTC switch refers to the switch of prescription drugs (Rx) to OTC status. In 1990, Carl Peck, a former director of the US Food and Drug Administration’s (FDA) Center for Drug Evaluation and Research, proposed a list of 13 criteria to be considered for selection of switch candidates, focusing on efficacy and safety of the drug1. Although these pharmacological interpretations of ‘OTCness’ (amenability to switching) are useful, they are presumably taken strictly from the perspective of the regulatory authority, and not the consumer. In 1998, the director of the division who reviewed OTC drug applications, Robert DeLap, suggested a new list of revised criteria, geared towards the end-user1.

Table 1 presents the list of approved ‘new’ (switched) OTC molecules during the past decade. In the US, 10 molecules switches were approved by FDA. In the EU, the UK has led the way, with 22 switches since 2000, several of which in categories hitherto not considered switch candidates in the US such as statins or triptans. There is an evolving process of harmonization in the EU, which stems from regulations drafted by the European Medicines Agency, specifically on the legal status of products and switching from Rx to OTC2. New regulations will allow a manufacturer to simultaneously gain access to all 27 EU member state markets3. Nevertheless, most switches in Europe have not yet been centralized, and, between countries, the divergence in numbers of switches among switch candidates is significant4. This divergence might reflect different cultural habits, differing roles of pharmacists in each country, regulatory differences with respect to OTC products, and the switch mechanism itself.

Table 1. Rx-to-OTC switches since 2000 in the US, the UK, France, and The Netherlands.

Drug	US	UK	France	Netherlands
Adenosine		2004
Alclometasone		2009
Ambroxol				2004
Amorolfine		2006	2010
Azelastine			2005
Azithromycin		2008
Beclometasone			2002
Cetirizine	2007
Chloramphenicol		2005
Clobetasone		2001
Docosanol	2000			2008
Fenticonazole		2001
Fexofenadine				2008
Flurbiprofen		2001	2005
Fluticasone		2002
Ketotifen	2006
Ipratropium bromide				2009
Levocabastine			2009
Levonorgestrel	2006	2001		2005
Loratadine	2002		2006
Naproxen		2007
Omeprazole	2003	2004	2010	2008
Orlistat	2007	2009*	2009*	2009*
Nicotine	2002	2001		2000
Nizatidine	2009
Pantoprazole		2009*	2009*	2009*
Penciclovir		2006	2000
Polyethylene glycol	2006
Prochlorperazine		2001
Simvastatin		2004
Sumatriptan		2006
Tamsulosin		2009
Terbinafine		2001	2002
Tranexamic acid		2010

*EU centralized switch, applicable to all 27 EU members. Only two successful pan-European regulatory approvals using the centralized process.

The number of applicants for Rx-to-OTC switches in Europe and the US is likely to increase in the coming years5,6. First, switches will likely continue to be derived from pharmaceutical firms’ desire to extend the lifecycle of drugs in their portfolio, particularly those that are so-called ‘switch candidates’7. A large number of prescription medicines potentially eligible for OTC status will go off patent in the next decade8, and as Rx-to-OTC switches create an opportunity for manufacturers to extend the lifecycle of certain drugs, switches might constitute a strategy to pre-empt generic competition. Second, this may also be partly payer-driven, a departure from standard practice, in which a party other than a drug manufacturer requests a switch. The case of non-sedating antihistamines is a case in point: loratadine was approved by FDA in 1993 and quickly became a blockbuster. The regulatory petition for OTC status for non-sedating antihistamines, including loratadine, came not from the sponsoring companies, but from the insurer WellPoint. WellPoint successfully argued that second-generation antihistamines were safer than existing OTC compounds and that the switches would save the US healthcare system billions of dollars. The FDA granted WellPoint’s request but could not force the manufacturers to switch until they were ready (when patents expired).

A balance must be struck between the increased demand for switches and their attendant safety risks: from a regulatory perspective, in order to be eligible for a switch, a switch candidate must be shown to be safe and effective when used without the supervision of a healthcare practitioner9. However, for policy-makers and payers the assessment of the potential clinical and economic impact of switches is critical. In most cases, there is not enough information to accurately anticipate implications of Rx-to-OTC switches. Therefore, an examination of existing decision models constitutes an appropriate way to address uncertainty and provide reasonable estimates of the economic impact of switches.

To date, no systematic review focused on economic evaluations of Rx-to-OTC switches has been published. The first objective of this article is to review existing decision-analytic models developed to assess the potential impact of Rx-to-OTC switches. Empirical and normative aspects of Rx-to-OTC switches will be addressed using existing decision-analytic models as points of departure for our critical analysis. The disparity of the results will be examined to improve the understanding of various cost components incorporated in the published models.

The second objective is to provide guidelines to improve future economic evaluations, as, with the rising interest in Rx-to-OTC switches, more economic evaluations will be conducted in the near future.

Methods

Search strategy

A systematic literature review was performed in order to identify all relevant published economic evaluations on Rx-to-OTC switches, in North America and Europe. Medline and Embase databases were used to retrieve economic publications. The search was initially limited to the years 2000–2010. We completed the systematic search by reviewing relevant bibliographical references of selected articles between 1995–2000. The research digest of the International Society of Pharmacoeconomics and Outcomes Research (ISPOR) was reviewed for potentially relevant abstracts for the past 3 years. In addition, Google Scholar was used to identify any relevant literature that may not be found via conventional sources. Table 2 presents the search strategy used.

Table 2. Search strategy.

Database	Search terms
Embase	([english]/lim OR [french]/lim) AND [2000–2011]/py) AND (‘switch’ OR ‘move’ OR ‘moving’ OR ‘changing’ OR ‘status’ OR ‘changes’) AND (‘otc’ OR ‘over-the-counter’ OR ‘counter’ OR ‘nonprescription’ OR ‘non-prescription’ OR ‘self medication’/exp OR ‘self medication’ OR ‘self care’/exp OR ‘self care’) AND (rx OR ‘prescriptions’/exp OR ‘prescriptions’ OR prescribed) AND (‘economics’/exp OR ‘economics’ OR ‘health economic evaluation’ OR ‘costs’ OR ‘cost’ OR ‘cost’/exp OR cost OR ‘analysis’ OR ‘analysis’/exp OR analysis OR costing OR costly OR ‘budget impact’ OR ‘decision making’/exp OR ‘decision making’ OR ‘cost-benefit’/exp OR ‘cost-benefit’ OR ‘managed care’/exp OR ‘managed care’ OR ‘managed care programs’/exp OR ‘managed care programs’ OR ‘cost-of-illness’/exp OR ‘cost-of-illness’ OR ‘payers’ OR ‘impact’ OR ‘out-of-pocket’ OR ‘co-payment’)
Medline	(Limits: English, French, published in the last 10 years) AND (Economics OR costs OR cost OR costing OR costly OR ‘budget impact’ OR ‘Decision Making’ OR ‘comparative study’ OR ‘cost-benefit analysis’ OR ‘cost-minimization analysis’ OR ‘Cost’ OR ‘managed care’ OR ‘Managed Care Programs’ OR ‘Decision Support Techniques’ OR ‘Cost-of-Illness’) AND (Rx OR prescriptions OR prescribed OR prescription) AND (‘OTC’ OR ‘over-the-counter’ OR ‘counter’ OR ‘nonprescription’ OR ‘non-prescription’ OR ‘self medication’ OR ‘self care’)
ISPOR	‘OTC’, ‘Over the counter’, ‘Non prescription’ [abstract]
Google Scholar	(OTC or ‘over the counter’ or ‘non prescription’ or ‘self medication’) and (prescription or prescribed or Rx) and (switch or move or change) (date 2000–2010)

Article selection and data extraction

References were screened by two independent reviewers. Once references were screened, data from the models were extracted: authors, publication type, geographic scope, switched drugs, disease area, population considered, type of analysis, costs included, time horizon, and stakeholder perspective. The data were transcribed onto a Microsoft® Excel data extraction form.

Each model was critically evaluated by the reviewers with respect to structure, relevance of inputs, methodology used, and robustness of results yielding an overview of the strengths and limitations of the reviewed models.

Guidelines development

Once the models were critically evaluated, two experts (authors of this paper) reviewed strengths and limitations of each model, and developed draft guidelines. The approach was then refined until the broad structure was agreed. Finally, all authors wrote the details of the different points considered.

Results

Overview of the reviewed publications

We found 12 economic evaluations of Rx-to-OTC switches10–21, as presented in Table 3. Ten studies concerned the US market10–12,14–20, and two were European-based (one the UK13, and one representing all of Europe21). The studies covered various diseases, dominated by allergy (loratadine)10–12, dyslipidemia (statins)13–15, and gastroenterology (famotidine, ranitidine and cimetidine)16,17. The other papers concerned contraception (oral contraceptive pills)18, smoking cessation (nicotine replacement therapy (NRT))19, and virology (valacyclovir)20. The EU study did not focus on any specific disease13.

Table 3. General characteristics of studies included in the review.

Ref	Type	Authors	Year	Country	Sponsor	Therapeutic area	Objective	Type of analysis	Type of model
[10]	Full paper	Cohen and DiMasi	2001	US	Independent	Allergy	To estimate the impact of a loratadine switch by using an economic model to compare two treatment pathways (pre- and post-switch) in term of cost, effectiveness, and safety	BIA	Decision tree
[11]	Full paper	Sullivan et al.	2003	US	Independent	Allergy	To assess the cost-effectiveness of switching loratadine from Rx-To-OTC status	BIA + CEA	Decision tree
[12]	Full paper	Sullivan and Nichol	2004	US	Independent	Allergy	To assess the economic impact of payer policies following an Rx-to-OTC switch of loratadine	BIA + CUA	Decision tree
[13]	ISPOR poster	Amirsadri-Naeini and Jackson	2008	UK	Independent	Dyslipidemia	To consider the cost-effectiveness of POM statins, OTC statin, and plant sterol/stanol products from the perspective of middle-aged males at moderate CVD risk when used according to current UK guidance for the primary prevention of CVD	CUA	Markov model
[14]	ISPOR abstract	Gandhi et al.	2006	US	Independent	Dyslipidemia	To compare the cost-utility of statins available via prescription vs hypothetical OTC status in the US	CUA	NA
[15]	ISPOR abstract	Gandhi et al.	2006	US	Independent	Dyslipidemia	To evaluate the cost-effectiveness of statins available by prescription vs OTC from a societal perspective	CEA + BIA	NA
[16]	Full paper	Kalish	1997	US	Johnson & Johnson Merck Consumer Pharmaceuticals	Gastro-enterology	To assess the societal costs of the treatment of dyspepsia prior to and after to OTC availability of the H2RAs	BIA	Decision tree
[17]	Full paper	Kunz et al.	1996	US	Johnson & Johnson Merck Consumer Pharmaceuticals	Gastro-enterology	To calculate the potential cost savings for a managed care organization (MCO) after switching appropriate patient from a prescription to OTC treatment of heartburn/non-ulcer dyspepsia	BIA	Decision tree
[18]	Full paper	Zhu et al.	2001	US	Independent	Contraception	To assess the economic value of moving oral contraceptives to OTC status	BIA	Decision tree
[19]	Full paper	Keeler et al.	2002	US	Pfizer	Smoking cessation	To assess the social benefit from the RX-to-OTC switch of nicotine replacement drugs (nicotine patches and gum) based on available data from the literature and estimations methods	BIA	Micro-simulation
[20]	ISPOR poster	Chang and Nichol	2009	US	Independent	Virology	To assess the cost implications of transitioning valacyclovir to OTC status and examine episodic and suppressive treatment models for genital herpes	BIA	Decision tree
[21]	Report	AESGP	2004	Austria, France, Germany, Italy, Portugal, Spain, UK	Independent	General	To develop a model assessing the economic impact of an Rx-to-OTC switch for European health care systems as well as the effects on public health	BIA	NA

AESGP, Association of the European Self-Medication Industry.

Type of analysis

The 12 economic studies included 10 budget impact analyses (BIA)10–12,15–21, three cost-utility analyses (CUA)12–14, and two cost-effectiveness analyses (CEA)11,15.

Type of model

Of the 12 economic studies, seven models were simple decision trees10–12,16–18,20, one was a Markov model13, one a micro-simulation model19; and we could not determine which type of model was used in three of the publications14,15,21.

Populations

Eight studies solely examined patients who were on the same prescription-only drug before the switch, and are considered as those who would likely turn to the OTC product12–15,17,19–21. Four models took a more comprehensive approach by including all different populations (currently on treatment, switching from other Rx drugs in the same therapeutic class, currently on an OTC in the same class, untreated, diagnosed, or undiagnosed)10–12,18. The majority of models took a narrow view with respect to the population dynamic, assuming that only patients who were already on the Rx medication would turn to the OTC product without considering the possibility of some turning to the OTC product from other Rx drugs in the same therapeutic category. Most of the models did not take into account the differential impact switches may have on the market uptake of drugs in the same therapeutic class by different sub-populations.

Costs

All studies included drug acquisition costs, nine considered physician visits10,14–16,18–20, one included hospitalizations16, six considered costs related to adverse events11^--13,18,21, and six studies tallied indirect costs (productivity loss) as well. Other costs were drug- or disease-specific and included, for example, the cost of motor vehicle accidents or other occupational injuries due to sedation associated with first generation antihistamines11,13.

Time horizon

Half of the models employed a 1-year time horizon11,12,18–21. When considering a chronic disease such as hypercholesterolemia, a longer time horizon was used13,14. Studies looked at the long-term impact of the switch once equilibrium was reached, after peak sales had occurred. None included impact information for the time period between the switch and reaching equilibrium.

Switch rate

Seventy-five per cent of models provided switch rates. Switch rates were not available for three of the studies10,14,15. Across the models, data robustness varies. The data sources were dominated by the secondary literature in most of all the models. Sales and market share data were also used in three models12,17,19, while limited sets of national utilization data were included12,18,20. Some authors generated switch rates based on assumptions owing to the lack of available evidence or existing benchmarks11,21. These assumptions were generally based on expert recommendations.

Handling of misuse

Costs of potential misuse due to lack of doctor’s supervision or inappropriate indication were rarely taken into account: four studies discussed potential inclusion of misuse10,16,19,21, but none included associated costs. Several economic models ignored the cost of adverse events as well as the risk of sub-optimal therapy while others stated they did not include them due to the lack of evidence.

Sensitivity analysis

Most of the models incorporated sensitivity analyses to address uncertainties regarding the variables used. Deterministic one-way sensitivity analyses were presented in eight studies11–13,16–20, and probabilistic multivariate sensitivity analyses were presented in two studies11,12. One study did not present any result10, and information was not available in two studies14,21.

Economic impact associated with switches

With several exceptions16,20,21, economic studies reported savings to payers. Table 4 summarizes cost-saving opportunities associated with Rx-to-OTC switches.

Table 4. Economic impact associated with Rx-to-OTC switches.

Ref	Authors	Year	Country	Therapeutic area	Drug	Perspective	Estimated savings	Observation	Strengths	Limitations
[10]	Cohen and DiMasi	2001	US	Allergy	Loratadine	Insured patient, uninsured patient, payer	Uninsured patients would save 80% following the switch; unclear impact for insured patient	Fewer physician visits, lower drug prices	Includes algorithm taking into account treatment failures due to lack of effectiveness or safety	Unclear how safety is incorporatedAssumes that each physician visit to obtain a prescription will be avoidedAssumes that untreated patients are not eligibleTheoretical model
[11]	Sullivan et al.	2003	US	Allergy	Loratadine	Societal	Estimated annual savings of $4 billion	Savings derive largely from avoiding adverse events from first-generation anti-histamine use	Includes utilization rates of first- and second-generation antihistamines before and after the switch	Only reports total savings, without a breakdown into cost components
[12]	Sullivan and Nichol	2004	US	Allergy	Loratadine	Managed care organization, employer, and Medicaid	OTC with policy change would save Medicaid $0.02 per member per month (PMPM)	By providing limited OTC coverage while retaining generous Rx coverage for other second-generation antihistamines	Shows potential impact of various payer policies on the overall budget impact of an OTC switch	Projection of potential impact in the absence of real data
[13]	Amirsadri-Naeini and Jackson	2008	UK	Hypercholesterolemia	Statins	Patient	From the patient’s perspective, the most cost-effective intervention to prevent CVD is POM statins (£1827.82/QALY)	Assumption is that lower dose OTC statin has same effectiveness as higher dose	No detailed description of the methodology to identify strengths	No detailed description of the methodology to identify limitations
[14]	Gandhi et al.	2006	USA	Hypercholesterolemia	Statins	Societal	OTC statins cost-utility of $1434/QALY compared to $1867/QALY for Rx statins	OTC statins were observed to be the optimal strategy choice.	No detailed description of the methodology to identify strengths	More costs could be included in the overall assessment such as lab monitoring costs, adverse events monitoring costs, compliance, and adherence measures
[15]	Gandhi et al.	2006	US	Hypercholesterolemia	Statins	Societal	From the patient’s perspective, the most cost-effective intervention to prevent CVD is a POM statins (£1827.82/QALY)	Assumption is that lower dose OTC statin has same effectiveness as higher dose.	Different statin agents were compared in the analysis	No detailed description of the methodology to identify limitations
[16]	Kalish SC, Bohn RL, Avorn J	1997	US	Gastro-enterology	Cimetidine, Ranitidine, Famotidine	Payer and societal	No substantial cost saving (payer and patient perspective)	The possible increase in costs following problems with OTC H2RAs use	Evaluation of possible increase in costs following the failure of OTC H2RAs use	Only considers the symptomatic population
[17]	Kunz et al.	1996	US	Gastro-enterology	Famotidine	Managed care organization	25% reduction in the overall cost	From avoiding unnecessary physician visits, laboratory tests, and prescriptions	As the model considers patients receiving chronic treatment, results were presented for 5 years that corresponded with new and continuing patients	Does not consider episodic patients
[18]	Zhu et al.	2001	US	Contraception	Oral contraceptive (OC) pills, Ethinyl Estradiol	Societal	$12.6 billion in savings annually	Savings from unintended pregnancies avoided	Analysis of costs comparing the cost impact of OTC pill availability	No detailed sensitivity analysisAll costs obtained from literatureSwitch rates may not be accurate to all geographic locations and some local differences may exist
[19]	Keeler et al.	2002	US	Pulmonology	Nicotine replacement therapy (NRT)	Societal	Estimated annual net ‘social value’ was ∼$1.8–2 billion in 1995	Calculations were made from the statistical value of a life saved	A demand function was developed for patches and gumHas to estimate the demand after the switchDemonstrated no significant price-sensitivity	No consideration of costs of misuse
[20]	Chang and Nichol	2009	US	Virology	Valacyclovir	Societal	Potential savings is $736 billion in direct medical expenditures	The savings are due to lower drug prices and a decrease in physician visits	Comparison of two treatment regimens (episodic and suppressive) both as a prescription and OTC status	No consideration of costs that could impact savings (such as herpes education, viral resistance, self-diagnosis, indirect cost, and the financial and quality-of-life impact of delayed treatment)Not all the population were included (only Rx population)
[21]	AESGP	2004	Austria, France, Germany, Italy, Portugal, Spain, UK	General	General	Societal	Saving to public funds just over €16.4 billion	The switch of 5% of Rx drugs to OTC status within Europe would result in fewer physician visits and an increase in worker productivity	Switch impact both from a local level (some EU countries) and from an EU perspective by projecting the results of the studied countries	No consideration of indirect costs due to inappropriate useNo consideration of potential adverse events

AESGP, Association of the European Self-Medication Industry.

First, it was shown that switches generated direct savings to payers, mainly due to shifting drug acquisition costs of the switched drug as well as of other Rx products in the same therapeutic class onto consumers since OTC products are not covered by health insurance companies. Other direct savings included the cost of doctor visits to obtain a prescription and savings related to early relief of acute symptoms, e.g., avoidance of emergency room visits or hospitalizations. Also, long-term improvement in health outcomes has been cited, as illustrated by smoking cessation nicotine-replacement OTC therapy19. Finally, introduction of an OTC with an improved safety profile compared to the already available OTC was sometimes yielding to cost savings. This was demonstrated with the switch of loratadine, a safer therapy than the OTC 1st-generation antihistamines10,12.

Strengths and limitations

We also identified strengths and limitations of the models in Table 4. Mainly, limitations included use of strong assumptions in models to solve the problem of missing information, and non-inclusion of specific patient populations.

Discussion

Critical analysis

We have observed in this review that very few switch impact studies have been published during the past 15 years. It appears to be a novel area where more research is required. This may naturally arise in the next decade as the current trend toward more Rx-to-OTC switches continues.

This review also shows that the main source of economic benefits is shifting the costs of drugs from payers to patients. At the same time, there were potential risks associated with Rx-to-OTC switches, related to a patient’s misdiagnosis, inappropriate self-medication, or lack of physician supervision. These risks could be translated into costs (i.e., clinical worse outcomes), resulting in more utilization of healthcare resources as a consequence of an increase in physician, emergency room, or hospital visits.

We identified several weaknesses in the models. First, most of the publications did not provide evidence to support assumptions on issues such as OTC-switch rates. Second, the treatment algorithm (set of choices facing the consumer following a switch) was not clearly specified. Each switch may have a differential impact on consumer behaviour; some will opt for the new OTC product, while others will choose a different Rx-only product in the same therapeutic class. Finally, a majority of models did not realistically reflect the impact of the switches, omitting several aspects of the disease.

Although a lower price does not necessarily guarantee an improvement in social welfare, especially in the case of drug classes in which there are patent-protected products22, switches generate direct cost savings to payers and society, they do so by lowering the price of the switched drug (OTC products tend to be have lower prices, ceteris paribus) and shifting the costs onto consumers. The demand for OTC products is likely to be more elastic due to lack of insurance coverage for OTC products. Hence, drug makers would be inclined to charge a lower price, especially for chronic diseases. Other direct cost savings included fewer doctor’s visits to obtain a prescription and fewer emergency room visits and/or hospitalizations. Switches also generate indirect cost savings through improved productivity (less absenteeism).

Generally, models tended to over-estimate cost savings as they ignored certain potential costs that could partly offset the total estimated savings. For example, when estimating the savings that accrue from fewer physician visits, most studies did not consider the possibility that the switch may also result in additional physician visits. Such visits may occur as a result of misdiagnosis, because of increased awareness of the disease that induces the patient to consult his doctor, or because of a move to a more expensive medicine that is still reimbursed, allowing the patient not to pay for the OTC product, but not saving any physician visit.

Guidance for future economic models

Currently, no guidelines exist regarding an economic assessment of switches. In this section we attempt to provide guidance on how to improve economic models. The guidelines we develop aim to help critical appraisers of economic models optimize credibility and accuracy of future model development. Guidelines cover three distinct components of the modelling process: structural issues on decision context, health states and clinical outcomes, and other considerations for model specifications.

Structural issues on decision context

Pathology

Developers should assess whether the model structure is appropriate and relevant to the disease in question. A Markov model is likely to be more relevant for chronic progressive disease (e.g., osteoarthrosis), while a decision tree will be more relevant for an acute disease (e.g., influenza). The natural course of disease should be well described, as well as the rationale for model choice, to ensure the model structure is transparent.

Drug

Several drug characteristics should be accounted for with each Rx-to-OTC switch, such as dose range, effectiveness (and its relationship to dosing), side-effects (and its relationship to dosing), and risk of misuse should be systematically anticipated.

Populations

A Rx-to-OTC switch model should include all important populations concerned. This includes patients already treated with the Rx drug or other substitutable Rx or available OTC. In addition, it is also critical to identify the user profile. Patients using the OTC switched drug may be classified under various populations depending on their prescription type (switching from the prescription to the OTC form of the same drug; switching from other prescription drugs—e.g., same class, or substitutes; switching from other OTC drugs). Further classification is possible, including diagnosis status before the switch and previous treatment. A combination of epidemiologic, bibliographic, and sales data is to be used to define those populations.

Health states and clinical outcomes

Epidemiology

Critical information such as incidence and prevalence of disease(s) being targeted by the switched product should be documented, in particular, to address budget impact implications of Rx-to-OTC switches.

Health outcomes

Important aspects of the disease should be taken into account, such as clinical states and adverse events. Outcomes should be considered from a public health or societal perspective, implying that both direct and indirect costs and benefits accruing to the population as a whole should be included in the analysis.

Resource utilization.

We recommend the inclusion of several distinct cost components in order to broaden the perspective of the switch on budget savings. These can be divided into direct and indirect costs. For direct costs, Rx drug acquisition cost should be considered, as well as all costs related to the patient’s treatment algorithm, such as general practitioner (GP) costs, hospitalizations, diagnostic tests, etc. In addition to payer cost savings, other potential switch benefits include improved health outcomes. Nevertheless, future researchers should also account for the possibility that switches may generate certain costs related to the presence of adverse events due to the lack of physician monitoring. If a societal perspective is adopted, indirect costs should also be included: work productivity-related costs such as the time taken off work, or the productivity gains due to more expedient access. The importance of those components depends on the nature of the disease, and in some disease areas (e.g., migraine) this is likely to be the major economic benefit. A significant degree of uncertainty has emerged with respect to the impact of switches on patient costs: at the time of the model development, the price, size of package, dose, and other characteristics of the OTC product may be unknown.

Other considerations for model specifications

Patient behaviour

Past switches or available retrospective cost studies could be used to strengthen assumptions about patient’s behaviour; i.e., their anticipated responses to switches. Studies should use nation-specific data whenever possible, as the local context for the OTC market. Experience shows that some switches were quick, but most occurred over time due to monitoring of safety issues and the increasing marketing efforts targeting disease awareness. As a consequence, patient’s behaviour changes gradually over time and a long time horizon (3 or 5 years) is recommended to capture the long-term effects of Rx-to-OTC switches.

OTC switch rate

The number of OTC users will drive the payer savings and company revenue. As a result, it is crucial to consider the behaviour of various sub-groups of consumers following a switch. Once patient sub-groups are clearly identified, rates of switching should be applied, considering substitution and expansion effects that take place across different patient sub-groups. In other words, it is necessary to understand and quantify how patients will behave following the switch in order to obtain an estimate of the switched population. In 2005 a study examined consumer perceptions regarding the risk of OTC drugs, and a minority of the 553 surveyed appeared willing to accept considerable risk to gain greater access to pharmaceuticals23. Another survey of over 1500 consumers revealed a more empowered American public, with about three-quarters (73%) preferring to try to treat conditions themselves rather than go to a doctor24. Future models should also consider patients who will refuse the OTC product, and turn to another prescription drug. Reasonable and well-documented assumptions should be developed, and uncertainty addressed through sensitivity analysis.

Misuse

Risk of misuse and associated costs should be carefully considered. Indeed, while patients will not require a doctor visit, they should have a minimal level of support to avoid drug misuse25. First, it is possible that the OTC drug is being used by patients who are not suffering from the condition. This will not generate any benefit, and will expose patients to potential side-effects and might delay proper diagnosis. Such side-effects have to be listed, and taken into account when relevant. Second, a sub-optimal treatment of the condition will delay the institution of appropriate treatment. It is necessary to discuss this potential delay. Third, chronic over-use of OTC drugs may be associated with the risk of side-effects that needs to be listed. Finally, some drugs can interact with foods and beverages, or with health conditions such as diabetes, kidney disease, and high blood pressure. Modellers should make sure this has been addressed with clinicians. Discussions about mode of action, pharmacodynamics, observations from case studies and real world evidence, clinical trial evidence, pharmacovigilance, and likelihood of consumers being prescribed a more expensive drug (or larger quantity or higher dose) to avoid paying for the OTC version should be developed.

Conclusions

In addition to health benefits, the potential economic gains associated with Rx-to-OTC switches are large. Policy-makers rely on economic models to predict potential cost savings associated with Rx-to-OTC switches, but such models should be carefully developed, and nation-specific. To date, there have been few studies conducted, and those that have been conducted lacked quality.

The development of high quality comprehensive models to assess public health and budgetary impact of Rx-to-OTC switch from the payer perspective may contribute to highlight the opportunity for a more systematic assessment of Rx-to-OTC switches.

Transparency

Declaration of funding

This article was funded by Sanofi.

Declaration of financial/other relationships

Aurélie Millier and Slim Karray are employees from Creativ-Ceutical which received funding from Sanofi to develop this review. Joshua Cohen and Mondher Toumi have no relevant financial or other relationships to disclose. JME Peer Reviewers on this manuscript have no relevant relationships to disclose.

Acknowledgements

No writing assistance is to be declared in the preparation of this manuscript.

Previous presentation

A summary of this research was presented at the ISPOR 14th Annual European Congress, held in Madrid, Spain, 5–8 November 2011, Poster Code: PHP31.