`Prescriptions
`Author(s): Natalie Mizik and Robert Jacobson
`Source: Management Science, Vol. 50, No. 12 (Dec., 2004), pp. 1704-1715
`Published by: INFORMS
`Stable URL: https://www.jstor.org/stable/30048061
`Accessed: 30-04-2019 22:35 UTC
`
`JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
`range of content in a trusted digital archive. We use information technology and tools to increase productivity and
`facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.
`
`Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
`https://about.jstor.org/terms
`
`INFORMS is collaborating with JSTOR to digitize, preserve and extend access to Management
`Science
`
`This content downloaded from 208.85.77.1 on Tue, 30 Apr 2019 22:35:27 UTC
`All use subject to https://about.jstor.org/terms
`
`Opiant Exhibit 2171
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685
`Page 1
`
`
`
` MANAGEMENT SCIENCE
` Vol. 50, No. 12, December 2004, pp. 1704-1715
` ISSN 0025-1909 1 EISSN 1526-5501 104150121 1704
`
` infimeR
` DOI 10.1287/mnsc.1040.0281
` f 2004 INFORMS
`
` Are Physicians "Easy Marks"? Quantifying the
` Effects of Detailing and Sampling on
` New Prescriptions
`
` Natalie Mizik
` Graduate School of Business, Columbia University, 3022 Broadway, New York, New York 10027-6902, nm2079@columbia.edu
` Robert Jacobson
` Department of Marketing, School of Business, University of Washington, Seattle, Washington 98195-3200,
` yusho@u.washington.edu
`
` Much public attention and considerable controversy surround pharmaceutical marketing practices and their
` impact on physicians. However, views on the matter have largely been shaped by anecdotal evidence or
` results from analyses with insufficient controls. Making use of a dynamic fixed-effects distributed lag regres-
` sion model, we empirically assess the role that two central components of pharmaceutical marketing practices
` (namely, detailing and sampling) have on physician prescribing behavior. Key differentiating features of our
` model include its ability to (i) capture persistence in the prescribing process and decompose it into own-growth
` and competitive-stealing effects, (ii) estimate an unrestricted decay structure of the promotional effects over
` time, and (iii) control for physician-specific effects that, if not taken into account, induce biased coefficient esti-
` mates of detailing and sampling effects. Based on pooled time series cross-sectional data involving three drugs,
` 24 monthly observations, and 74,075 individual physicians (more than 2 million observations in total), we find
` that detailing and free drug samples have positive and statistically significant effects on the number of new
` prescriptions issued by a physician. However, we find that the magnitudes of the effects are modest.
`
` Key words: pharmaceutical marketing; salesforce effectiveness
` History: Accepted by Linda Green, public sector applications; received March 8, 2004. This paper was with the
` authors 1 month for 1 revision.
`
` Introduction
` industry's promotional tactics lead to an increase in
` appropriate versus inappropriate use of drugs in a
` As the cost of prescription drugs continues to escalate,
` cost-effective manner.
` increased public attention is being focused on the
` Concern that pharmaceutical marketing practices
` marketing practices of the pharmaceutical firms as
` have exacerbated increases in public health costs has
` one source of the problem. Direct-to-physician activ-
` prompted government actions at the federal and state
` ities account for the bulk of U.S. pharmaceutical
` levels. For example, in 2002 the federal government
` firm promotional spending. IMS Health (2003) esti-
` issued a warning to the drug industry to curtail some
` mates that over $5.8 billion was spent in 2002 on
` of their marketing practices (Washington Post 2002).
` detailing, i.e., pharmaceutical sales representatives
` H.R. 2356, which calls for ongoing annual funding of
` (PSRs) visiting physicians to promote their firm's
` $75 million to conduct comparative cost-effectiveness
` drugs. In addition, the retail value of the free drug
` drug studies, was introduced in Congress in June
` samples distributed during these visits is estimated
` at $11.5 billion.
` 2003. A primary intent of this legislation is to pro-
` vide objective scientific evidence to "reduce doctors'
` A detailing visit typically lasts two to five minutes
` reliance on marketing information from the pharma-
` during which time a PSR discusses one to three of
` ceutical industry" (Pear 2003). Given the fact that one
` the company's drugs. Information (and, at times, mis-
` of every five dollars spent on pharmaceutical drugs in
` information) about a drug's composition, therapeutic
` value, proper dosage, and potential side effects is
` the United States is paid for by a state program, state
` governments have also taken steps to counter PSR
` communicated (Zigler et al. 1995). Often, PSRs will
` influence. Most notably, several states have under-
` also dispense samples and possibly offer small gifts
` taken counterdetailing initiatives (Gold 2001). State
` to the physician. At issue is whether these interac-
` employees visit physicians in hopes of persuading
` tions with PSRs compromise physician integrity and
` them to switch from prescribing branded drugs to
` affect their prescribing behavior. More precisely, the
` key public policy issue is the extent to which the
` prescribing lower-cost generic drugs.
`
` 1704
`
`This content downloaded from 208.85.77.1 on Tue, 30 Apr 2019 22:35:27 UTC
`All use subject to https://about.jstor.org/terms
`
`Opiant Exhibit 2171
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685
`Page 2
`
`
`
` Mizik and Jacobson: Quantifying the Effects of Detailing and Sampling on New Prescriptions
` Management Science 50(12), pp. 1704-1715, v 2004 INFORMS 1705
`
` bias if left uncontrolled, we assess the effect of detail-
` Prescription drug expenditures are projected to
` remain the fastest-growing sector of health care
` ing and sampling on physician prescribing behavior.
` The large number of observations in the database (it
` expenditures. They are expected to account for 14.5%
` involves a total of more than 2 million observations)
` of $3.1 trillion health care expenditures by 2012
` (compared with approximately 10% in 2001). With
` allows us to accurately pinpoint the impact that inter-
` actions with PSRs have on the number of new pre-
` recent legislation providing a Medicare drug benefit
` scriptions issued by physicians.
` expected to cost the federal government $534 billion
` We find that, although detailing and free drug sam-
` over the next decade, it is no wonder that the impact
` ples have a positive and statistically significant asso-
` of pharmaceutical industry marketing practices is of
` ciation with the number of new prescriptions issued
` keen interest to policymakers, the business commu-
` by a physician, the magnitudes of the effects are mod-
` nity, and the general public.
` est. As such, our results challenge the two domi-
` Two competing views have dominated discussion
` nant views and support the contention that, rather
` on the matter. The prevailing view contends that
` than being easy marks, physicians are tough sells.
` PSRs significantly influence physicians' prescribing
` This realization is important because the public policy
` behavior and that this influence has negative effect
` debate continues over how best to address the high
` on patients' welfare, in that PSRs encourage physi-
` cost of prescription drugs.
` cians to prescribe more expensive branded drugs.
` Many public policy organizations and consumer
` advocacy groups adhere to this view (see, for exam-
` PSR Influence on Physicians
` ple, www.nofreelunch.org). The prominent alternative
` Most discussions of PSRs have focused on the factors
` view argues that PSRs do influence physicians' pre-
` facilitating their influence. Unquestionably, PSRs pro-
` scribing behavior, but that this influence is positive
` vide physicians with information about new drugs,
` in that PSRs provide physicians with valuable infor-
` new indications, dosages, and interactions for existing
` mation. As a result, physicians are better informed
` medicines. Azoulay (2002) finds evidence that detail-
` and make better choices for their patients. Pharma-
` ing diffuses product information. Avorn et al. (1982)
` ceutical companies and industry groups advocate this
` report that 20% of surveyed physicians view informa-
` second view.
` tion provided by PSRs as "very important" in influ-
` Despite the substantial resources that pharmaceu-
` encing their prescribing behavior. Furthermore, PSRs
` tical companies invest in promoting their products
` are trained in persuading physicians. Detailing takes
` and the controversy associated with pharmaceutical
` the form of presenting facts and, as has been doc-
` marketing practices, surprisingly little is known about
` umented (Zigler et al. 1995), misrepresenting facts
` the magnitude of the impact that PSR visits and free
` about the drug in an effective manner. Finally, mere
` drug samples have on physician prescribing behavior.
` exposure or salience effects might lead to a temporary
` Narayanan et al. (2003) report a pharmaceutical exec-
` increase in prescribing following a PSR visit. Numer-
` utive as stating, "No one is really sure if sending the
` ous studies have reported high physician responsive-
` legions of reps to doctors' offices really works. Every-
` ness to PSR activity attributed it to PSR persuasive-
` one is afraid to stop it, because they don't know what
` ness (Avorn et al. 1982, Powers 1998).
` difference it's making" (p. 4).
` Less attention has been paid to the factors limiting
` In point of fact, much of the evidence on PSR
` PSR effectiveness. The key consideration here is that
` effectiveness is anecdotal. The empirical studies
` PSRs are not the only or even the primary source
` investigating the issue have been subject to data or
` of information about drugs for physicians. Scientific
` methodological limitations that restricted their ability
` papers, advice from colleagues, and a physician's own
` to control for potential biases and have come to
` training and experience also influence prescribing
` contradictory conclusions regarding even the central
` practices. Indeed, most physicians view these other
` issues: the effects of detailing on prescriptions (e.g.,
` influences as far more important than that of PSRs
` Parsons and Abeele 1981 versus Gonul et al. 2001),
` (Peay and Peay 1990).
` of detailing on price elasticity (e.g., Rizzo 1999 versus
` PSR influence is limited by the fact that many
` Gonul et al. 2001), and even of price on sales (e.g.,
` physicians have skeptical or negative attitudes toward
` Rizzo 1999 versus Gonul et al. 2001).
` PSRs (Lichstein et al. 1992, McKinley et al. 1990).
` We have obtained access to a unique database
` Attribution theory suggests that with low source
` that allows us to undertake econometric analysis
` credibility, which is determined by factors such as
` that overcomes a number of fundamental limitations
` a source's trustworthiness and expertise (Dholakia
` existing in past research. In particular, making use
` and Sternthal 1977), arguments in a message will
` of a dynamic fixed-effects distributed lag model that
` be discounted (Eagley and Chaiken 1975). Physicians
` accounts for physician-specific effects likely to induce
` recognize that PSRs are neither experts nor com-
`
`This content downloaded from 208.85.77.1 on Tue, 30 Apr 2019 22:35:27 UTC
`All use subject to https://about.jstor.org/terms
`
`Opiant Exhibit 2171
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685
`Page 3
`
`
`
` Mizik and Jacobson: Quantifying the Effects of Detailing and Sampling on New Prescriptions
` 1706 Management Science 50(12), pp. 1704-1715, r2004 INFORMS
`
` estimated main effect of detailing was negative. The
` pletely trustworthy. They realize that information pre-
` most dominant explanatory factor in the model is
` sented is biased toward the promoted drug and is
` sales lagged one period, which would reflect per-
` unlikely to be objective or even accurate (Connelly
` et al. 1990). Thus, physicians will discount informa-
` sistence in behaviors and carryover effects magnify-
` tion received from a PSR.
` ing the influence of detailing. Alternatively, lagged
` Some additional characteristics of physicians would
` sales could be reflective of territory-specific effects
` that are not modeled and, as such, could lead to
` seem to make them particularly tough sells. Friestad
` and Wright's (1994) persuasion knowledge model
` biased estimates. Wotruba (1982), for example, raises
` suggests that targets of persuasion use their knowl-
` this possibility of territory-specific effects to question
` edge about the persuasion agent and can effectively
` the reported effects of detailing.
` Rizzo (1999) uses annual data for the period 1988-
` cope with and even achieve their own goals during a
` persuasion attempt, e.g., obtaining free drug samples
` 1993 and for 46 drugs to estimate a brand-level model
` linking prescriptions for a drug for a given year
` that can be later distributed to patients. Campbell and
` Kirmani's (2000) tests of the persuasion knowledge
` to pharmaceutical company marketing activities. He
` model reveal that busy targets with accessible agent
` finds that price is negatively related to sales and that
` motivation (a profile that would fit most physicians)
` detailing is anticompetitive in that it decreases price
` sensitivity. Detailing is also found to have a direct
` are particularly effective in resisting persuasion.
` When cast within the workings of other sources
` positive effect on sales. Surprisingly, no consideration
` of influence, we would expect the ability of PSRs to
` is given to the dynamic properties of sales. The classic
` influence physician behavior to be relatively small. As
` spurious regression characteristics, i.e., very high R2
` such, we hypothesize a relatively small effect of PSR
` in the presence of substantial unmodeled autocorre-
` activity on physician prescribing behavior.
` lation, appear to be present. As such, questions exist
` about the validity of both the point estimates and
` standard errors reported in the analysis.
` Previous Empirical Research
` Gonul et al. (2001) use data involving 1,785 patient
` The various studies assessing the effect of PSR activ-
` visits to estimate a multinomial logit model assess-
` ity on physician prescribing behavior have generated
` ing factors influencing physician prescribing behav-
` conflicting results. Indeed, on some of the most cen-
` ior. Exactly opposite to the findings of Rizzo (1999),
` tral issues-ranging from the effects of detailing on
` they report that price has a positive effect on prescrip-
` prescriptions, of detailing on price elasticity, and even
` tion probabilities and that detailing increases price
` of price on sales-studies have come to diametrically
` sensitivity. They find positive effects of detailing and
` opposite conclusions. Data and methodological limi-
` sampling, but do not discuss the implications of their
` tations, however, raise concerns about the inferences
` magnitudes. These magnitudes, calculated based on
` drawn from these analyses.
` descriptive statistics, imply elasticities that are sur-
` A few quasi-experimental studies of the issue orig-
` prisingly large.' The elasticity estimates for the seven
` inate in the medical community. These studies com-
` drugs studied, evaluated at the mean level of detail-
` pare physicians who did not see PSRs or were visited
` ing and sampling, average 41% for detailing and 48%
` less frequently by PSRs to physicians who saw PSRs
` for sampling. Particularly for samples, which have a
` or were visited more frequently by PSRs (Chren and
` negligible marginal cost, their estimated coefficients
` Landefeld 1994, Powers 1998). The limitation of these
` imply enormous returns to enhanced PSR activity. In
` studies is that they are not randomized: PSRs do
` point of fact, these substantial effects could arise, not
` not determine which physicians to visit on a random
` from the influence of PSR activity, but rather as an
` basis. Rather, PSRs tend to see physicians who are
` outgrowth of a joint correlation with an omitted factor
` more likely to utilize the drug or who prescribe in
` from the model, e.g., larger practices prescribe more
` higher volume. This consideration invalidates these
` and receive more free samples.
` attempts to assess the effect of PSRs independent of
` A concern, which Gonul et al. (2001) explicitly
` controls accounting for motivation influencing PSR
` acknowledge, is over the role of physician-specific
` behavior.
` effects that can induce a bias in the estimated coeffi-
` The ability to potentially control for other influ-
` cients. They state (p. 84),
` ences is an advantage of regression-based analysis.
` Past research has made use of different regression
` techniques to assess PSR influence. Unfortunately,
` it has been inadequate in controlling for physician-
` specific effects. Parsons and Abeele (1981) use data
` for 24 months and 14 territories to model the number
` 1 The elasticity of prescription probability Pj to covariate Xjk in a
` of prescriptions sold in a given territory for a given
` conditional logit model is calculated as (dPj/Pj)/(aXjk/Xjk) = f3k * Xjk *
` (1- P).
` month as a function of sales calls. Interestingly, the
`
` prescription behavior patterns might be strongly influ-
` enced by factors other than the explanatory vari-
` ables we include in our model. Examples are physi-
`
`This content downloaded from 208.85.77.1 on Tue, 30 Apr 2019 22:35:27 UTC
`All use subject to https://about.jstor.org/terms
`
`Opiant Exhibit 2171
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685
`Page 4
`
`
`
` Mizik and Jacobson: Quantifying the Effects of Detailing and Sampling on New Prescriptions
` Management Science 50(12), pp. 1704-1715, l2004 INFORMS 1707
`
` to 11 years; annual sales range from under $0.5 billion
` cians' unobservable personal characteristics.... Ignor-
` ing these factors might bias the coefficients of the
` to more than $1 billion; they come from different ther-
` included explanatory variables.
` apeutic areas. Although the effect of detailing can
` vary across drugs, analysis of these three drugs offers
` The extent to which their estimates are biased by
` some generalizable insights, not only because they
` the failure to control for unobservable factors remains
` provide a cross-section of drugs in the marketplace,
` unanswered, but this is one consideration that might
` but because they represent more than 4 million PSR
` account for the large estimated effects.
` interactions with physicians.
`
` 6 6
`
` Empirical Analysis
` Model
` A key benefit of utilizing pooled time series cross-
` We employ the following dynamic fixed-effects dis-
` sectional (panel) data is the ability to test for and
` tributed lag regression model to assess the effect of
` control the effect of unobserved fixed factors. These
` detailing and sampling on new prescriptions:
` unobserved factors, if left uncontrolled, can induce
` Prescribeit
` bias in the coefficient estimates of the explanatory fac-
` tors included in the model. Past research has either
` =- a + E j * Detailsit-j + E yj * Samplesit-j
` not used panel data or not made full use of the ben-
` j=0 j=0
` efits of panel data analysis. We make use of pooled
` 6 6
` time series cross-section observations (24 months of
` + E Aj * Competitort- + E * Prescribeit-
` observations across 74,075 physicians) and panel data
` =O i=1
` statistical methods (i.e., a dynamic fixed-effects dis-
` T 11
` tributed lag regression model) to assess the effect
` + , * Time(r) + i, Ks Specialty(s) * Trendt
` of detailing and sampling on physician prescribing
` +=1 s=1
` behavior.
` + Eit' (1)
` Data
` where Prescribeit, Detailsit, Samplesit, and Competitorit
` are, respectively, the number of new prescriptions
` Access to the data was gained from a U.S. pharmaceu-
` issued, the number of PSR visits, the number of free
` tical manufacturer with the only condition of ensur-
` ing the anonymity of the firm and the drugs in the
` drug samples received, and the number of new pre-
` study. Two different sets of data were merged to form
` scriptions issued for competitive drugs by physician i
` the database. One data set pertains to the number
` at time period t. Time(7) is an indicator function that
` of new prescriptions for the studied drugs and their
` takes on the value 0 prior to the time period r and 1
` competitors issued by physicians during a month.
` from the time period r on, Specialty(s) is an indi-
` cator function that takes on the value 1 when the
` The new prescription measure reflects both new and
` repeat usage, but does not reflect refills accompany-
` specialty area of the physician is s, 0 otherwise (i.e.,
` ing the prescriptions. These data cover a 24-month
` separate dummy variables for each of the 11 spe-
` cialty areas), and Trend is the observation number
` period for three widely prescribed drugs. The second
` data set pertains to detailing and sampling activity
` for a given month and year. Because it includes both
` by PSRs for the same three drugs. The two data sets
` current-term and lagged variables in the model, Equa-
` were merged into one database containing prescrib-
` tion (1) allows for a wide range of possible effects and
` ing and promotional activity information by month
` influences, e.g., serial correlation (current-effects) and
` and physician.
` state-dependent (persisting) dynamic relationships.
` To reduce the possible influence of extreme val-
` A key characteristic of Equation (1) is that it allows
` ues (outliers) that would arise from, for example,
` for a physician-specific effect, i.e., the intercept ai
` data entry errors and the common practice of one
` is allowed to vary by physician. This consideration
` physician signing for all samples that later get dis-
` acknowledges that physician behavior patterns are
` influenced by unobserved or unobservable factors,
` tributed to a group of physicians attending a confer-
` ence, we excluded the top 0.5% of observations for the
` e.g., physician characteristics. To the extent that these
` number of details, samples, and new prescriptions.
` unobserved factors are correlated with detailing and
` We later undertook sensitivity analysis on alternative
` sampling, analysis not controlling for their effects will
` definitions of outliers (e.g., 0%, 1%, 5%) and found
` result in biased estimated effects for the marketing
` results in close correspondence across these alterna-
` phenomena. Although a Hausman (1978) specification
` tive samples.
` test can empirically assess the role played by fixed
` Table 1 presents basic background information and
` effects, we have a priori reason to believe that these
` unobserved factors will in fact be correlated with mar-
` descriptive statistics for the drugs included in our
` study. The drugs differ on a variety of dimensions:
` keting activity. For instance, larger practices will gen-
` erate more prescriptions and will also attract more
` They have been on the market from less than 1 year
`
`This content downloaded from 208.85.77.1 on Tue, 30 Apr 2019 22:35:27 UTC
`All use subject to https://about.jstor.org/terms
`
`Opiant Exhibit 2171
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685
`Page 5
`
`
`
` Mizik and Jacobson: Quantifying the Effects of Detailing and Sampling on New Prescriptions
` 1708 Management Science 50(12), pp. 1704-1715, w2004 INFORMS
`
` Table 1 Drug Profiles
`
` Drug A Drug B Drug C
`
` Sales range (US$) 0.5 to 1 billion over 1 billion under 0.5 billion
` Time on the market at the 3 years 11 years 6 months
` beginning of the study period
` Estimated number of competitors 12 18 11
` in the respective therapeutic area
` Mean number of details per 1.73 1.98 1.73
` physician per month (1.75) (1.70) (1.44)
` Mean number of free drug samples 4.34 7.79 4.02
` per physician per month (9.76) (13.71) (7.73)
` Mean number of monthly new 13.18 8.82 2.27
` prescriptions per physician (14.81) (10.43) (3.58)
` Average number of refills following 2-3 2-3 2-3
` one new prescription
`
` Recommended duration of therapy 3 months, and as maintenance with 3 months and up to 9 months Not yet established
` periodic patient reexamination
` Cost of therapy relative to other Average Above average Average
` drugs in the therapeutic area
` Mean number of monthly new 42.91 48.80 22.46
` prescriptions per physician in (43.63) (53.73) (19.03)
` the respective therapeutic area
` Therapeutic area is Relatively new Well established Well established
` Specialty area of top prescribers Psychiatry Primary carett Primary careft
` Number of physicians 10,516 55,896 30,005
` Number of data points 252,384 1,341,504 720,120
`
` Note. Standard errors are in parentheses. The sum of the number of physicians in each of the three data sets is greater than the total
` number of physicians in the study (74,075) because some physicians are in the upper 60 prescribing percentile for more than one of
` the drugs in our study.
` ttThe primary care specialty area includes family practice, general practice, internal medicine, and osteopathy.
`
` involve both a direct effect on prescriptions and an
` detailing. As such, a spurious positive correlation, i.e.,
` unrelated to any potential effects of detailing, will
` indirect effect that arises through persistence in physi-
` exist between detailing and prescriptions due to a
` cian behavior. The total effect of detailing and sam-
` pling can be calculated as
` joint correlation with practice size.
` Because the effects of PSR activity are unlikely to
` 6 6 6 6
` be limited to the month when the visit occurred, we
` /Pj/ 1-Bb-[ j and y/ 1-yyj ,
` allow for current and lagged effects for both detailing
` j=O j=l j=0 j=1
` and sampling. The number of observations available
` respectively.
` in our data sample allows us to directly estimate sep-
` We include both lagged own prescriptions and
` arate effects for lagged terms. As such, we do not
` lagged competitors' prescriptions, thus the model
` need to impose a specific decay pattern as a neces-
` separates lagged total demand dynamics into two
` sity for preserving degrees of freedom. We do, how-
` key components: competitive substitution and own
` ever, need to specify the length of time that a PSR
` demand growth. Lagged own prescriptions will have
` visit might influence physician behavior. We select a
` a positive effect on prescriptions. Lagged competitors'
` six-months lag length under the view that the effect
` prescriptions will have a negative impact on prescrip-
` of a visit will dissipate substantially over this period,
` tions because they capture the substitution effects
` but we also test for longer-term lagged effects. Still,
` that physicians make between competing drugs. The
` we expect the effects of detailing to be largest in the
` current competitive prescriptions, however, will cap-
` months directly following the visit. The cumulative
` ture two different phenomena. They will reflect not
` direct effect of detailing and sampling can be obtained
` only substitution effects, but also changes in total
` demand due to market expansion or contraction. In
` simply by summing the coefficients, i.e., E61=0j and
` this regard, current-term competitive effects will act
` as a proxy for two different phenomena with oppo-
` Lagged values for new prescriptions (Prescribeit_)
` capture autocorrelation in the series that arises
` site effects, i.e., negative substitution effects and posi-
` tive market demand effects. As such, the current-term
` through inertia and persistence in physician behavior.
` These autoregressive effects play a key role in that
` coefficient (AO) will depend on the relative magnitude
` they magnify the effects of detailing and sampling.
` of the two conflicting effects and, therefore, the sign
` That is, the total effects of detailing and sampling
` of the effect cannot be postulated a priori.
`
`This content downloaded from 208.85.77.1 on Tue, 30 Apr 2019 22:35:27 UTC
`All use subject to https://about.jstor.org/terms
`
`Opiant Exhibit 2171
`Nalox-1 Pharmaceuticals, LLC v. Opiant Pharmaceuticals, Inc.
`IPR2019-00685
`Page 6
`
`
`
` Mizik and Jacobson: Quantifying the Effects of Detailing and Sampling on New Prescriptions
` Management Science 50(12), pp. 1704-1715, @2004 INFORMS 1709
`
` The other variables in Equation (1) are time-period-
` data for antiulcer drugs.2 We found that, after making
` specific indicators and specialty-specific trends. The
` use of competitive sales as a proxy variable, monthly
` time-period-specific indicators (the coefficients 8,)
` changes in brand detailing exhibited little correlation
` allow for the fact that the number of prescriptions
` (0.01) with monthly changes in competitor detailing.
` Given the absence of correlation at the brand level, the
` can shift across time periods. These intercepts capture
` not only seasonal effects but all brandwide influences
` correlation at the physician level can also be expected
` to be similarly small and, indeed, even smaller. As
` that shift prescribing behavior across all physicians
` (e.g., price changes, changes in the set of alternative
` such, the magnitude of this bias in the estimated effect
` medications available, changes in advertising cam-
` of detailing caused by the exclusion of competitive
` paigns, etc.). That is, the inclusion of the time-specific
` detailing from the analysis can be expected to be min-
` indicator variables will capture all effects common
` imal, or even completely absent.
` across physicians, which would include the diffu-
` We estimate Equation (2) using instrumental vari-
` sion pattern for the drug, research reports in scien-
` able estimation as ordinary least squares will generate
` tific journals, any negative or positive publicity for
` biased estimates of the coefficients for APrescribeit-1
` the brand or its competitors, etc. The 11 specialty-
` and ACompetitorit. By construction, APrescribeit_1 will
` specific trends K, capture influences that shift pre-
` be correlated with the differenced error term qit
` scribing behavior across all physicians in a particular
` and, just as substitution effects cause competitor pre-
` specialty. After we take first differences, the time-
` scriptions to influence own prescriptions, own pre-
` period-specific indicators and the specialty-specific
` scriptions will influence the amount of competitor
` trends are transformed into time-period-specific inter-
` prescriptions. Following Anderson and Hsiao (1982),
` we use lagged values of the levels of the series
` cepts and specialty-specific intercepts.
` To remove the influence of physician specific effects
` (values at time period t - 2 and earlier) to gener-
` (i.e., ai), we take first differences of Equation (1) to
` ate instrumental variable estimates for APrescribeit-1
` obtain
` and ACompetitorit. This procedure generates consis-
` tent (i.e., asymptotically unbiased) estimates of the
` APrescribeit
` parameters and their standard errors.3
` 6 6
` = p fj3 * ADetailsit-j + yei * ASamplesitj
` Results
` j=0 j=O
` For each of the three drugs in our study, we estimated
` 6 6
` the Equation (2) regression model. Table 2 reports
` + E Aj * ACompetitorit-j + ~ * APrescribeit-j
` the estimated coefficients. Figures 1 and 2 graphically
` j=0 j=1
` depict the estimated direct effects of detailing and
` T 11
` + , * ATime(r) + K * Specialty(s) + it. (2)
` sampling, respectively.
` Persistence in Prescribing Behavior. For each of
` Of notable absence from our model, due to lack of
` the three drugs in the study we observe signifi-
` available data, is competitive marketing effort. How-
` cant persistence in physicians' prescribing behavior.
` ever, several considerations suggest that the poten-
` Although the first-order autocorrelation is the most
` tial bias in the estimated effects of firm detailing and
` substantial for all three drugs, significant higher-order
` sampling stemming from this omission will be minor.
` effects are present as well. For Drug A the estimated
` First, the inclusion of competitor sales will capture
` coefficients for months 1 through 6 of 0.357, 0.205,
` some of the effects of competitive marketing activi-
` ties and thus reduce potential omitted variable bias.
` 2 We thank Ernst R. Berndt for granting us access to these data. See
` In essence, the competitor sales variable in the model
` Berndt et al. (2003) for a complete description of the data. Consis-
` will act as a proxy variable (Wickens 1972) for com-
` tent with lack