Key Points
Question
What drug characteristics are associated with larger proportions of promotional spending allocated to direct-to-consumer advertising?
Findings
In this exploratory cross-sectional study of 150 prescription drugs with the highest US sales in 2020, a higher proportion of promotional spending allocated to direct-to-consumer advertising was associated with drugs rated as having lower added clinical benefit than for those having higher added clinical benefit (absolute 14.3% increase in proportion) and with total drug sales (absolute 1.5% increase in proportion for every 10% increase in sales).
Meaning
Lower added clinical benefit ratings and higher total drug sales were associated with higher spending on direct-to-consumer advertising as the share of the total; further research is needed to understand the implications of these findings.
Abstract
Importance
Some drugs are heavily marketed through direct-to-consumer advertising.
Objective
To identify drug characteristics associated with a greater share of promotional spending on advertising directly to consumers.
Design, Setting, and Participants
Exploratory cross-sectional analysis of drug characteristics and promotional spending for the 150 top-selling branded prescription drugs in the US in 2020 as identified from IQVIA National Sales Perspectives data. Promotional spending data were provided by IQVIA ChannelDynamics.
Exposures
Drug characteristics (total 2020 sales; total 2020 promotional spending; clinical benefit ratings; number of indications, off-label use; molecule type; nature of condition treated; administration type; generic availability; US Food and Drug Administration [FDA] approval year, World Health Organization anatomical therapeutic chemical classification; Medicare annual mean spending per beneficiary; percent sales attributable to the drug; market size; market competitiveness) assessed from health technology assessment agencies (France’s Haute Autorité de Santé and Canada’s Patented Medicine Prices Review Board) and drug data sources (Drugs@FDA, the FDA Purple Book, Lexicomp, Merative Marketscan Research Databases, and Medicare Spending by Drug data).
Main Outcomes and Measures
Proportion of total promotional spending allocated to direct-to-consumer-advertising for each drug.
Results
The 2020 median proportion of promotional spending allocated to direct-to-consumer advertising was 13.5% (IQR, 1.96%-36.6%); median promotional spending, $20.9 million (IQR, $2.72-$131 million); and median total sales, $1.51 billion (IQR, $0.97-$2.26 billion). Of the 150 best-selling drugs, 16 were missing data and key covariates; therefore, the primary study sample comprised 134 drugs. After adjustment for multiple drug characteristics, the mean proportion of total promotional spending allocated to direct-to-consumer advertising for the remaining 134 drugs was an absolute 14.3% (95% CI, 1.43%-27.2%; P = .03) higher for those with low added clinical benefit than for those with high added clinical benefit and an absolute 1.5% (95% CI, 0.44%-2.56%; P = .005) higher for each 10% increase in total sales.
Conclusions and Relevance
Among top-selling US drugs in 2020, a rating of lower added benefit and higher total drug sales were associated with a higher proportion of manufacturer total promotional spending allocated to direct-to-consumer advertising. Further research is needed to understand the implications of these findings.
This pharmacoeconomic study estimates associations between drug characteristics and the proportion of manufacturer promotional spending allocated to direct-to-consumer advertising.
Introduction
The US and New Zealand are the only countries that allow direct-to-consumer advertising for prescription drugs.1,2 From 1997 to 2016, spending on advertisements for prescription drugs directed to consumers in the US grew from $1.3 billion to $6 billion, with the greatest increases observed among drugs to treat diabetes and endocrine conditions.3
Direct-to-consumer advertising is associated with increased patient requests and increased clinician prescriptions for advertised products.1,4 This advertising strategy may also influence the relationship between patients and clinicians. Increased patient requests for advertised products may present an opportunity for an informed discussion of treatment alternatives. Increased requests can also create additional burden for clinicians or lead to distrust and a worsened relationship between clinician and patient, especially when patient requests are not granted.1,5
This exploratory, cross-sectional study examined drug characteristics associated with shares of total promotional spending allocated to consumer advertising for top-selling branded prescription drugs in order to inform policy conversations about the benefits and harms of this advertising practice.
Methods
Sample and Data
The Bloomberg School of Public Health Institutional Review Board (IRB) determined that this study represented nonhuman subjects research obviating IRB oversight.
IQVIA National Sales Perspectives (a data set that projects US national estimates based on 90% of the pharmaceutical market)6 was used to identify the 150 branded drug products with the highest 2020 US sales, which were selected because they represented 60% of all US sales. Excluded were 2 devices, 4 vaccines, 2 products with unreported total promotional spending, and 1 product with indeterminate direct-to-consumer advertising spending. IQVIA ChannelDynamics data6 provided promotional spending for individual products across 6 subcomponents: product sampling, clinician contacts, meetings and events, journals, mail, and direct-to-consumer advertising. Advertising directly to consumers includes the following media: digital, magazine, newspaper, outdoor (eg, billboards), radio, and television. Using these data, we computed the percent of total promotional spending allocated to advertising directly to consumers for each product.
We collected data for the following explanatory variables: total 2020 product sales, added clinical benefit (high vs low), number of indications (single vs many), off-label use (yes vs no), molecule type (biologic or biosimilar vs small molecule), nature of the condition treated (chronic, acute, or both), administration type (clinician vs self), generic availability (yes vs no), US Food and Drug Administration (FDA) approval year (since 2012 vs earlier), World Health Organization (WHO) anatomical therapeutic chemical classification, 2020 Medicare annual mean spending per beneficiary, the percent of manufacturer sales attributable to the drug product, market size, and market competitiveness (competitive, oligopolistic, or monopoly or near monopoly).
Added clinical benefit was identified using ratings from health technology assessment agencies in France and Canada.7,8 These countries consider several criteria when determining a drug’s benefit for a particular indication relative to existing treatments (Box).9,10 Both countries prioritize evidence from randomized trials that compare the new drug to an existing treatment.9,10 The countries issue a rating of added benefit that qualitatively summarizes existing comparative effectiveness evidence. These added benefit ratings have been used in previous studies to assess drug quality, including a study of pharmaceutical promotion aimed at clinicians.11,12,13,14,15 Previous research found that France and Canada agreed in 85% of assessments.11
Box. Criteria Considered in Added Benefit Assessments.
France (Haute Autorité de Santé)
Quality of research evidence
Effect size (efficacy, quality of life, safety)
Clinical relevance
Medical need
Innovation (defined as novelty in the mechanism of action that responds to an inadequately met medical need)
Canada (Patented Medicine Prices Review Board)
Increased efficacy
Reduction in incidence or grade of important adverse reactions
Route of administration
Patient or caregiver convenience
Compliance improvements leading to improved efficacy
Duration of usual treatment course
Time to optimal therapeutic effect
Success rate
Percentage of affected population treated effectively
Disability avoidance or savings
France assigns 5 possible ratings of added benefit: major, important, moderate, minor, and none. These ratings inform pricing decisions in France for all drugs reimbursed by the National Health Insurance scheme: drugs with at least moderate added benefit may be priced according to prices in various international markets, whereas drugs with minor or no added benefit are priced according to the prices of therapeutic alternatives in the French domestic market.16 Drugs are re-assessed every 5 years or earlier if new evidence becomes available.17
Canada assigns 4 ratings: breakthrough, substantial, moderate, and slight or none. These ratings are used to inform list prices of patented medicines at market entry and for determining when prices of existing products may be excessive.10 Drugs receiving the 2 highest ratings may have prices set according to the median price in 7 international markets, including the US.10
The top 3 ratings from each country were aggregated into “high added benefit” and the remaining ratings into “low added benefit.” For each product, we identified the most favorable rating given to the drug (eg, “major” is more favorable than “minor”) for any indication and regardless of the assessment date. We used Canada’s ratings for 27 drugs for which there was no French rating available.
We identified indications, whether off-label use is indicated and whether a generic is available, using the Lexicomp database. We determined molecule type by identifying biologics or biosimilars in the FDA Purple Book. We identified approval year using the Drugs@FDA database and dichotomized the variable as before or after 2012, the median year of approval for the 150 drugs, to achieve an equal balance of products on either side of the threshold. We identified whether a drug primarily treats chronic or acute conditions using the maintenance indicator variable in Merative MarketScan’s outpatient drug claims.
We determined administration type by computing the product’s proportion of total Medicare claims that are in Medicare Parts B and D using publicly available Medicare data on spending by drug. Products with at least 90% of claims in Part B were classified as clinician-administered, and products with at least 90% of claims in Part D were classified as self-administered. For the drugs without clear Part B or Part D use, 2 researchers (M.J.D. and C.C.D.) reviewed FDA medication guides to determine administration type when the drug entered the market. The mean annual spending per beneficiary for each individual drug was determined using the Medicare data on spending by drug; for drugs classified as clinician administered, we used Part B spending data, and for drugs classified as self-administered, we used Part D spending data.
Manufacturer sales were drawn from the IQVIA National Sales Perspectives. Market size for each drug was calculated as the sum of all sales in the drug’s class defined using level 3 of IQVIA’s Uniform System of Classification (eg, nonnarcotic analgesics), which is a therapeutic classification of pharmaceutical products.18,19 Market competitiveness was assigned for each drug based on calculating the Herfindahl-Hirschman Index at the drug’s Uniform System of Classification level 4 (eg, synthetics, nonnarcotic).18,19 We defined Herfindahl-Hirschman Index of less than 2500 as competitive, greater than or equal to 8000 as a monopoly or near monopoly, and all others as oligopolistic.19 All spending and sales data were from 2020.
Analysis
All analyses were conducted in Stata version 14.2 (StataCorp LLC).
For the primary analysis, a generalized linear model with a log-link function and gamma distribution was used because the dependent variable (percent of total promotional spending allocated to direct-to-consumer advertising) was nonnegative and right-skewed. The following independent variables exhibiting right-skewness were log transformed: total promotional spending, total sales, Medicare spending per beneficiary, and the percent of manufacturer sales attributable to the drug product.
Independent variables included in the primary analysis were identified through Akaike information criterion–based best subsets selection using the gvselect module.20 The model with the lowest Akaike information criterion was selected and included the following independent variables: log-transformed spending per beneficiary, log-transformed total sales, added clinical benefit, the nature of the condition treated, the anatomical therapeutic chemical classification, years from approval, number of indications, and administration type. To these variables, molecule type was added as a potential confounder between administration type and the outcome. The relationship between molecule type and administration type was assessed using the Fisher exact test.
Variance inflation factors were computed for all independent variables included in the primary analysis to check for multicollinearity. Sixteen drugs were excluded from the primary analysis due to missing data. Data on added clinical benefit was missing from 10% of the observations because neither France nor Canada had assessed the drug. Missingness for other variables was trivial, occurring in each case in less than 2.7% of observations. Both unadjusted and adjusted mean marginal effects were computed for each covariate. The mean marginal effect can be interpreted as the mean absolute percent change in the proportion of promotional spending allocated to direct-to-consumer advertising per unit change in the independent variable. The threshold for statistical significance was set at α = .05 using a 2-sided test.
Several sensitivity analyses were conducted. First, we used backward selection instead of the Akaike information criterion–based best subsets approach to select variables for inclusion in the generalized linear model described above. Second, a model that included all covariables was run. Third, to further investigate the association between added clinical benefit and the outcome, the primary model was run including all covariables statistically significantly associated with added clinical benefit (eTable 1 in Supplement 1). Fourth, multiple imputation was used to account for missing added clinical benefit data. A logistic model was used to impute the missing added clinical benefit data based on the proportion of total promotion allocated to direct-to-consumer advertising and all reported covariables. One hundred imputations were performed, and a pooled analysis was completed using the primary model described above.
Results
Summary Drug Characteristics
Drug characteristics are summarized in Table 1. The 2020 median proportion of promotional spending allocated to direct-to-consumer advertising was 13.5% (IQR, 1.96%-36.6%); median promotional spending, $20.9 million (IQR, $2.72-$131 million); median total sales, $1.51 billion (IQR, $0.97-$2.26 billion); median share of manufacturer sales attributable to the drug, 7.9% (IQR, 4.2%-17.7%); median market size for each drug’s class, $17 billion (IQR, $5.68-$31.1 billion); and median annual spending per beneficiary by Medicare, $17 900 (IQR, $3410-$46 200).
Table 1. Descriptive Characteristics of 150 Top-Selling Drugs as Measured by 2020 US Drug Sales.
| Spending and market variables | Median (IQR) |
|---|---|
| Proportion of total promotional spending allocated to direct-to-consumer advertising, % | 13.5 (1.96-36.6) |
| Total promotional spending, millions, $ | 20.9 (2.72-131) |
| Total product sales, billions, $ | 1.51 (0.97-2.26) |
| Medicare spending per beneficiary, thousands, $ | 17.9 (3.41-46.2) [n = 147] |
| Proportion of manufacturer sales attributable to the drug product, % | 7.9 (4.2-17.7) |
| Market size, billions, $a | 17.0 (5.68-31.1) |
| Drug variables | No. (%) (n = 150) |
| Added clinical benefit (low vs high)b | 92 (68) [n = 135 assessed] |
| Approval year (since 2012 vs earlier) | 76 (51) |
| Administration (clinician administered vs self-administered) | 43 (29) |
| Market competitivenessc | [n = 146] |
| Competitive | 79 (54) |
| Oligopolistic | 58 (40) |
| Monopoly or near monopoly | 9 (6) |
| Molecule type (biologic or biosimilar products vs small molecule) | 59 (39) |
| Single indication (yes vs no) | 63 (42) |
| Off-label use (yes vs no) | 65 (43) |
| Generic availability (yes vs no) | 28 (19) |
| Conditions targetedc | [n = 149] |
| Chronic | 76 (51) |
| Chronic or acute | 44 (30) |
| Acute | 29 (19) |
| Anatomical group | |
| Alimentary tract and metabolism | 25 (17) |
| Anti-infectives for systemic use | 14 (9) |
| Antineoplastic and immunomodulating agents | 53 (35) |
| Nervous system | 15 (10) |
| Respiratory system | 16 (11) |
| Otherd | 27 (18) |
Market size for each drug was calculated as the sum of all sales in the drug’s class defined using level 3 of IQVIA’s Uniform System of Classification (eg, nonnarcotic analgesics).
Added clinical benefit was identified using assessments from health technology assessment agencies in France and Canada. See the Box, Methods section, and Table 2 footnotes for details. Canada’s ratings for 27 drugs were used when there was no French rating available.
Data were not available in the sources used.
World Health Organization anatomical chemical classification system categories: blood and blood forming agents (n = 8), systemic hormonal preparations (n = 6), cardiovascular system (n = 4), musculoskeletal system (n = 3), sensory organs (n = 3), genitourinary system and sex hormones (n = 2), and dermatological conditions (n = 1).
Ninety-two of 135 drugs (68%) with a clinical benefit assessment were rated as having a low added benefit. Seventy-six of the 150 top-selling drugs (51%) have been approved since 2012. Forty-three drugs (29%) were clinician-administered; 79 (54%) were in a competitive market; 59 (39%) were biologic or biosimilar products; 63 (42%) had 1 indication only; 65 (43%) were used off-label; 28 (19%) had generic availability; 76 (51%) targeted a chronic condition; and 53 (35%) were classified as antineoplastic and immunomodulating agents. The clinician-administered drugs in the study sample were statistically significantly more likely to be biologic or biosimilar products than were self-administered drugs (P < .001).
Characteristics of the 12 products with more than 80% promotional spending allocated to direct-to-consumer advertising are summarized in Table 2. (eTable 2 in Supplement 1 summarizes characteristics for all 150 study drugs.)
Table 2. Select Characteristics of Top-Selling Drugs as Measured by 2020 US Drug Sales and With More Than 80% of Total Promotional Spending Allocated to Direct-to-Consumer Advertising.
| Product (nonproprietary name) | Share of promotional spending on direct-to-consumer advertising, % | Total, $ | Indication (WHO anatomical therapeutic chemical classification) | Added clinical benefita | Original assessment, country | |
|---|---|---|---|---|---|---|
| Promotional spending, millions | Product sales, billions | |||||
| Gammagard Liquid (immune globulin) | 99.9 | 0.275 | 1.02 | Hypogammaglobulinemia; multifocal motor neuropathy (anti-infectives for systemic use) | Low | None, France |
| Yervoy (ipilimumab) | 97.6 | 82.5 | 1.14 | Various cancers (antineoplastic and immunomodulating agents) | Low | Minor, France |
| Ocrevus (ocrelizumab) | 97.4 | 97.1 | 3.54 | Multiple sclerosis (antineoplastic and immunomodulating agents) | High | Moderate, France |
| Ibrance (palbociclib) | 95.8 | 131 | 3.79 | Breast cancer (antineoplastic and immunomodulating agents) | Low | Minor, France |
| Opdivo (nivolumab) | 92.2 | 95.3 | 3.95 | Various cancers (antineoplastic and immunomodulating agents) | High | Moderate, France |
| Odefsey (rilpivirine-emtricitabine-tenofovir alafenamide combination) | 91.7 | 0.52 | 1.44 | HIV-1 (anti-infectives for systemic use) | Low | Slight or none, Canada |
| Entyvio (vedolizumab) | 87.4 | 85.6 | 2.83 | Crohn disease; ulcerative colitis (antineoplastic and immunomodulating agents) | Low | None, France |
| Keytruda (pembrolizumab) | 86.5 | 144 | 8.31 | Various cancers (antineoplastic and immunomodulating agents) | High | Moderate, France |
| Neulasta (pegfilgrastim) | 85.5 | 10.4 | 2.83 | Hematopoietic radiation injury syndrome; prevention of chemotherapy-induced neutropenia (antineoplastic and immunomodulating agents) | High | Major, France |
| Genvoya (elvitegravir-cobicistat-emtricitabine-tenofovir alafenamide combination) | 84.3 | 0.647 | 3.38 | HIV-1 (anti-infectives for systemic use) | Low | None, France |
| Hemlibra (emicizumab) | 84.2 | 0.564 | 1.33 | Hemophilia A (blood and blood-forming organs) | High | Important, France |
| Mavyret (glecaprevir-pibrentasvir combination) | 80.1 | 50.2 | 1.15 | Hepatitis C (anti-infectives for systemic use) | Low | Minor, France |
Abbreviation: WHO, World Health Organization.
Added clinical benefit was identified using assessments from health technology assessment agencies in France and Canada. The 5 possible ratings from France are major, important, moderate, minor, and none; the 4 ratings from Canada are breakthrough, substantial, moderate, and slight or none. The top 3 ratings from each country were aggregated into high added benefit and the remaining ratings into low added benefit. For each product, the most favorable ratings given to the drug (eg, major is more favorable than minor) was identified for any indication and regardless of the assessment date. Canada’s ratings for 27 drugs were used when there was no French rating available. See the Box and Methods section for details.
Share of Promotional Spending for Direct-to-Consumer Advertising
Adjusting for all variables in the primary model, the mean change in the proportion of total promotional spending allocated to direct-to-consumer advertising was an absolute 14.3% higher for drugs with low added clinical benefit than for drugs with high added clinical benefit (95% CI, 1.43%-27.2%; P = .03) and an absolute 1.5% (95% CI, 0.44-2.56; P = .005) higher for each 10% increase in total drug sales (Table 3). A statistically significantly lower proportion of total promotional spending on advertising directly to consumers was allocated for drugs in the WHO alimentary tract and metabolism classification category (including insulins and antihyperglycemics; eTable 2 in Supplement 1) than on all other classification categories (range, 19.4% to 38.6%; P < .05; Table 3). No other variables were statistically significantly associated with the share of promotional spending on consumer advertising.
Table 3. Absolute Change in the Proportion of Total Promotional Spending Allocated to Direct-to-Consumer Advertising.
| Variable | Unadjusted mean marginal effect (95% CI)a | P value | Adjusted mean marginal effect (95% CI)a | P value |
|---|---|---|---|---|
| Total product sales (per 10% increase) | 0.92 (0.16 to 1.68) | .02 | 1.50 (0.44 to 2.56) | .005 |
| Added clinical benefit (low vs high)b | 1.58 (–8.75 to 11.9) | .77 | 14.3 (1.43 to 27.2) | .03 |
| Approval year (since 2012 vs earlier) | 13.0 (3.40 to 22.7) | .008 | 10.7 (–0.61 to 21.9) | .06 |
| Molecular type (biologic or biosimilar vs small molecule) | 13.9 (3.32 to 24.5) | .01 | 15.5 (–1.27 to 32.3) | .07 |
| Single indication (yes vs no) | –2.53 (–11.7 to 6.60) | .59 | 12.2 (–2.07 to 26.4) | .09 |
| Anatomical group | ||||
| Alimentary tract and metabolism | [Reference] | [Reference] | ||
| Anti-infectives for systemic use | 36.8 (8.84 to 64.7) | .01 | 38.6 (8.42 to 68.9) | .01 |
| Antineoplastic and immunomodulating agents | 24.2 (12.9 to 35.4) | <.001 | 22.6 (7.84 to 37.3) | .003 |
| Nervous system | 7.84 (–3.14 to 18.8) | .16 | 21.6 (0.96 to 42.2) | .04 |
| Respiratory system | 9.59 (–1.99 to 21.2) | .11 | 21.9 (1.33 to 42.4) | .04 |
| Otherc | 7.90 (–0.80 to 16.6) | .08 | 19.4 (4.95 to 33.9) | .009 |
| Condition targeted | ||||
| Chronic | [Reference] | [Reference] | ||
| Chronic or acute | 0.91 (–8.73 to 10.6) | .85 | 10.4 (–4.46 to 25.2) | .17 |
| Acute | 10.6 (–4.15 to 25.3) | .16 | 16.4 (–4.20 to 37.0) | .12 |
| Medicare spending per beneficiary (per 10% increase) | 0.72 (0.32 to 1.12) | <.001 | 0.31 (–0.30 to 0.92) | .32 |
| Administration (clinician administered vs self-administered) | 16.4 (3.38 to 29.5) | .01 | 6.25 (–13.3 to 25.8) | .53 |
The mean marginal effect can be interpreted as the mean change in the share of total promotional spending allocated to direct-to-consumer advertising for a change in the independent variable.
Added clinical benefit was identified using assessments from health technology assessment agencies in France and Canada. See the Box, Methods section, and Table 2 footnotes for details. Canada’s ratings for 27 drugs were used when there was no French rating available.
See Table 1 footnotes for details on the other included categories.
In 3 sensitivity analyses that varied by independent variables (using backward selection to select drug characteristic variables for model inclusion, all variables, and variables associated with clinical benefit), added clinical benefit and total drug sales remained statistically significantly associated with the proportion of promotional spending allocated to direct-to-consumer advertising (eTable 3 in Supplement 1). Across these analyses of direct-to-consumer advertising spending, the mean absolute percent increase allocated for drugs with low added clinical benefit ranged from 14.8% to 19% compared with spending for drugs with high added clinical benefit, and from 1.5% to 1.8% per 10% increase in drug sales. Multiply imputing missing values for added benefit did not substantively alter the results of the primary analysis (eTable 4 in Supplement 1).
Discussion
In this exploratory cross-sectional study, a higher proportion of manufacturer promotional spending allocated to direct-to-consumer advertising was associated with drugs rated as having lower rather than higher added clinical benefit and with higher total drug sales. Drugs in the WHO alimentary tract and metabolism classification category, including insulins and antihyperglycemic agents, had a significantly lower share of total promotional spending on advertising directly to consumers vs drugs in other classification categories.
Direct-to-consumer advertising may increase patient requests for advertised products and the likelihood of their prescription by clinicians.1,4 Allocating a greater share of promotional spending to consumer advertising (vs promotions that target clinicians) may therefore reflect a strategy to drive patient demand for drugs that clinicians might be less likely to prescribe because either there are several similarly effective alternative treatments available or there is a more effective alternative available. Drugs with lower added clinical benefit may also be more likely to have lower cost alternatives, be part of a step-therapy regimen, or have unfavorable formulary placement. Whether including information on added clinical benefit in direct-to-consumer advertising would affect the frequency of patient requests for the advertised product may be a future research question.
The findings regarding added clinical benefit rely on assessments conducted by health technology assessment agencies in France and Canada. There is currently no government entity in the US that conducts similar assessments. However, the Inflation Reduction Act will require the Centers for Medicare & Medicaid Services to consider the comparative effectiveness of drugs that are selected to undergo Medicare price negotiations starting in 2026.21 This legislation provides an opportunity for the federal government to begin assessing the added clinical benefit of drugs as is being done in other countries.
Limitations
This study has several limitations. First, we performed a cross-sectional analysis of a limited number of product characteristics using 1 year of promotional and sales data. Other models and other years of analysis might yield different findings. Second, our analyses relied on market data from IQVIA. Although such data have been previously used to analyze pharmaceutical marketing and promotion,3,22 technical information regarding how the data are gathered, curated, and harmonized or how quality control is performed is limited. Third, these data do not capture certain promotional elements, such as spending on patient coupons and patient assistance programs that, if included, could impact estimates of the share of promotional spending on advertising to consumers directly. Fourth, the added benefit assessments from France and Canada may reflect different value judgments about the strength or clinical significance of evidence. Fifth, because drugs often enter the market or may be used for new indications earlier in the US than in other international markets, our measure of added benefit may omit some drugs or drug uses. However, we mitigate this limitation by incorporating assessments conducted after 2020 to capture potential cases for which market entry or changes to clinical practice occur later than in the US. Sixth, the CIs around the finding of association of lower added benefit drugs with higher proportions of promotional spending on direct-to-consumer advertising spending is wide, and statistically compatible with an increase in the proportion of promotional spending on direct-to-consumer advertising of as little as 1.43 percentage points and as much as 27.2 percentage points.
Conclusions
Among top-selling US drugs in 2020, a rating of lower added benefit and higher total drug sales was associated with a higher proportion of manufacturer total promotional spending allocated to direct-to-consumer advertising. Further research is needed to understand the implications of these findings.
eTable 1. Characteristics of top-selling drugs by level of added benefit (high vs low)
eTable 2. Select characteristics of 150 top-selling drugs, organized by anatomical therapeutic chemical
eTable 3. Mean percentage point change in the share of total promotional spending allocated to direct-to-consumer advertising, adjusted associations from three alternative models
eTable 4. Adjusted associations between explanatory variables and share of total promotional spending allocated to DTCA, primary model with and without multiple imputation
Data Sharing Statement
References
- 1.Khokar B, Weathers S, Mattingly TJ II. Direct-to-consumer advertising and the advancement of the quadruple aim: a narrative review. Clin Pharm Forum. 2022;5(4):459-466. doi: 10.1002/jac5.1599 [DOI] [Google Scholar]
- 2.Guideline on the regulation of therapeutic products in New Zealand. Medsafe. Updated June 20, 2022. Accessed January 14, 2022. https://www.medsafe.govt.nz/regulatory/current-guidelines.asp
- 3.Schwartz LM, Woloshin S. Medical marketing in the United States, 1997-2016. JAMA. 2019;321(1):80-96. doi: 10.1001/jama.2018.19320 [DOI] [PubMed] [Google Scholar]
- 4.Franquiz MJ, McGuire AL. Direct-to-consumer drug advertisement and prescribing practices: evidence review and practical guidance for clinicians. J Gen Intern Med. 2021;36(5):1390-1394. doi: 10.1007/s11606-020-06218-x [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.DeFrank JT, Berkman ND, Kahwati L, Cullen K, Aikin KJ, Sullivan HW. Direct-to-consumer advertising of prescription drugs and the patient-prescriber encounter: a systematic review. Health Commun. 2020;35(6):739-746. doi: 10.1080/10410236.2019.1584781 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Available IQVIA data. IQVIA . Accessed September 1, 2022. https://www.iqvia.com/insights/the-iqvia-institute/available-iqvia-data
- 7.Base de Données Publique des Médicaments [Public drug database]. Haute Autorité de Santé . Accessed August 9, 2022. https://base-donnees-publique.medicaments.gouv.fr/telechargement.php
- 8.New patented medicines reported to PMPRB. Patented Medicine Prices Review Board. Accessed August 10, 2022. http://www.pmprb-cepmb.gc.ca/pmpMedicines.asp?x=611
- 9.Haute Autorité de Santé . Transparency Committee Doctrine: Principles of Medicinal Products Assessment and Appraisal for Reimbursement Purposes. Updated 2020. Accessed August 14, 2022. https://www.has-sante.fr/upload/docs/application/pdf/2019-07/doctrine_de_la_commission_de_la_transparence_-_version_anglaise.pdf
- 10.Patented Medicine Prices Review Board . Compendium of Policies, Guidelines and Procedures. Published 2017. Accessed August 14, 2022. http://www.pmprb-cepmb.gc.ca/CMFiles/Compendium_Feb_2017_EN.pdf
- 11.DiStefano MJ, Kang S-Y, Yehia F, Morales C, Anderson GF. Assessing the added therapeutic benefit of ultra-expensive drugs. Value Health. 2021;24(3):397-403. doi: 10.1016/j.jval.2020.10.021 [DOI] [PubMed] [Google Scholar]
- 12.Kyle M, Williams H. Is American health care uniquely inefficient? evidence from prescription drugs. Am Econ Rev. 2017;107(5):486-490. doi: 10.1257/aer.p20171086 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Kyle MK. Are important innovations rewarded? evidence from pharmaceutical markets. Rev Ind Organ. 2018;53:211-234. doi: 10.1007/s11151-018-9639-7 [DOI] [Google Scholar]
- 14.Lexchin J. The relation between promotional spending on drugs and their therapeutic gain: a cohort analysis. CMAJ Open. 2017;5(3):E724-E728. doi: 10.9778/cmajo.20170089 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Frank RG, Avorn J, Kesselheim AS. What do high drug prices buy us? Health Affairs blog. April 29, 2020. Accessed January 14, 2022. doi: 10.1377/hblog20200424.131397 [DOI]
- 16.Rodwin MA. Pharmaceutical price and spending controls in France: lessons for the United States. Int J Health Serv. 2020;50(2):156-165. doi: 10.1177/0020731419897580 [DOI] [PubMed] [Google Scholar]
- 17.Sauvage P. Pharmaceutical pricing in France: a critique. Eurohealth (Lond). 2008;14(2):6-8. [Google Scholar]
- 18.IQVIA . The Uniform System of Classification (USC). Published 2018. Accessed August 4, 2022. https://www.iqvia.com/-/media/iqvia/pdfs/institute-reports/the-uniform-system-of-classification.pdf
- 19.Kang S-Y, Sen AP, Levy JF, Long J, Alexander GC, Anderson GF. Factors associated with manufacturer drug coupon use at US Pharmacies. JAMA Health Forum. 2021;2(8):e212123. doi: 10.1001/jamahealthforum.2021.2123 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lindsey C, Sheather S. GVSELECT: Stata module to perform best subsets variable selection. Ideas. Accessed August 1, 2022. https://ideas.repec.org/c/boc/bocode/s457816.html
- 21.Inflation Reduction Act of 2022. Pub L No. 117-169, 136 Stat 1818. Accessed October 26, 2022. https://www.congress.gov/bill/117th-congress/house-bill/5376/text
- 22.Kornfield R, Donohue J, Berndt ER, Alexander GC. Promotion of prescription drugs to consumers and providers, 2001-2010. PLoS One. 2013;8(3):e55504. doi: 10.1371/journal.pone.0055504 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eTable 1. Characteristics of top-selling drugs by level of added benefit (high vs low)
eTable 2. Select characteristics of 150 top-selling drugs, organized by anatomical therapeutic chemical
eTable 3. Mean percentage point change in the share of total promotional spending allocated to direct-to-consumer advertising, adjusted associations from three alternative models
eTable 4. Adjusted associations between explanatory variables and share of total promotional spending allocated to DTCA, primary model with and without multiple imputation
Data Sharing Statement