The Cost of Pushing PillsA New Estimate of Pharmaceutical Promotion Expenditures in the United States

Marc‐André Gagnon, Joel Lexchin•View original

OverviewBalancedalloy voice

Let's start with an old picture that still shapes how people see the drug business: on one side, researchers in white coats; on the other, sales representatives in cars. Back in the late 1950s, Senator Estes Kefauver put that split on trial. He argued that patents kept prices high, marketing inflated costs, and many "new" products barely moved the therapeutic needle. Half a century later, we're still asking a very simple question with very large consequences: do pharmaceutical companies spend more driving demand than they do discovering drugs? In two thousand eight, Marc-André Gagnon and Joel Lexchin approached that question with a detective's mindset. They weren't doing a takedown. They were trying to fix the ruler we've been using. The numbers most people cite for "how much pharmaceutical companies spend on promotion" come from a handful of sources, each with blind spots. If you want to set policy, or even have an honest debate, you need a measurement that doesn't leave out big chunks of the story. So they focused on one market and one year where the data were richest: the United States in two thousand four. It was the single biggest market at the time, responsible for about forty-three percent of global drug sales. The premise was straightforward: line up the best data sources, see where they agree, see where they don't, and build a composite that's explicit about what's counted and what isn't. They worked with two lenses. The first is IMS, the industry's long-time scoreboard. IMS compiles an enormous stream of information from companies and distributors; in the early two thousands, it tracked more than a million products and, crucially, said it covered about ninety percent of US prescription drug sales. Those numbers give IMS authority—and also tell you its viewpoint. Much of its spending data comes from surveys of firms, which is good for breadth. But it is vulnerable to underreporting, especially on activities a company might prefer to keep offstage. And IMS's public promotion totals in two thousand four deliberately left out whole categories, like company-sponsored meetings. The second lens is CAM, a physician-focused auditor. Instead of asking companies what they spent, CAM asks doctors what they experienced. In the United States in two thousand four, that meant surveys of about two thousand primary care physicians and four thousand eight hundred specialists, combined with internal newsletters and cross-checks. CAM covers the same familiar categories—detailing by sales representatives, samples, direct-to-consumer ads—and it also tracks company-sponsored meetings, e-promotion, mailings, and some clinical trial activity with promotional intent. It counts the sales call differently, too: not just the representative's time in the office. It covers the full cost of making that call happen—local and regional management, training, and the slick brochures representatives leave behind. Here's the part that makes CAM especially interesting. It runs a validation committee with representatives from major drug makers—think Merck, Pfizer, AstraZeneca, and Sanofi—and, under confidentiality, compares its physician-reported totals to internal corporate spending tallies. When they did that, CAM found something unsettling: about thirty percent of promotional spending didn't show up in the physician surveys. Roughly a third. They attributed about a third of that gap to simple omissions or incomplete disclosure by doctors, and about two-thirds to channels the survey was never going to catch in the first place. That includes promotion aimed at physician groups they didn't sample, unmonitored journals, and possibly practices that shade into the unethical. That thirty percent "unmonitored" slice became a key correction in the final accounting. Now, put the two lenses side by side and you see why Gagnon and Lexchin had to triangulate. Take detailing—the face-to-face visits from representatives that used to be the backbone of drug promotion. IMS pegged detailing for two thousand four at about seven point three billion dollars. CAM put it at twenty point four billion dollars. That gulf isn't a rounding error; it's a definition. CAM priced the whole visit, overhead and all; IMS mostly priced the representative's time in the field. If you think the true cost of a sales call includes the managers who organize it and the materials that fuel it, CAM's number looks more realistic. Samples are the second flashpoint, and here the disagreement isn't just about scope; it's about valuation. IMS reported fifteen point nine billion dollars' worth of samples, using the retail price, and it got its counts by directly monitoring shipments from manufacturers. CAM reported six point three billion dollars, using wholesale price and a survey-based estimate of how many samples physicians actually handed to patients. They then halved that number to account for small starter packs. Which is "right"? For legal and accounting purposes, companies typically value donated products at retail. And monitoring shipments is less error-prone than asking busy doctors to remember how much free stock walked out the door. Gagnon and Lexchin leaned toward the IMS approach on samples for that reason. Other categories line up more neatly. Direct-to-consumer advertising, those glossy television and magazine campaigns, sits at about four billion dollars. Journal ads, the back pages of the medical weeklies, are a relatively tiny five hundred million dollars. These agreements give you a couple of anchor points. But look where IMS has blanks and CAM has numbers: company-sponsored meetings, which CAM puts at about two billion dollars, and a grab-bag of e-promotion, mailings, and some clinical trial activity that CAM totals at around three hundred million dollars. When you add those together with the richer accounting for detailing, you start to see why the CAM view paints a busier promotional machine. Gagnon and Lexchin don't pick a side. They build a composite. For two thousand four, they take CAM's broader detailing figure, IMS's retail-valued samples, and the shared numbers for consumer and journal ads. They bring in CAM's meetings and e-promotion categories that IMS doesn't report. And then they do the uncomfortable but necessary thing: they tack on a line for what neither source counted—the unmonitored thirty percent that CAM's validation work revealed, applied to the physician-facing portions of the spending. That adjustment is big. In this construction, it comes to about fourteen point four billion dollars. It's not a fudge factor pulled from thin air; it's the delta between what doctors remember and what companies actually budget, based on comparisons CAM could do under confidentiality. With that in place, three different totals emerge. If you stick with IMS's view alone, two thousand four US promotional spending comes in at twenty-seven point seven billion dollars. If you take CAM's raw, physician-reported total without any correction, it's thirty-three point five billion dollars. If you apply CAM's thirty percent undercount to its own categories and keep wholesale valuations for samples, you're in the high forties—about forty-seven point nine billion dollars. And if you use what Gagnon and Lexchin argue are the most appropriate figures for each category, that means CAM's broader detailing, IMS's retail-valued samples, the shared ad totals, CAM's meetings and e-promotion, and the thirty percent undercount. You land at fifty-seven point five billion dollars. How big is that in real life? Do a quick back-of-the-envelope calculation. In two thousand four, the United States had roughly seven hundred thousand practicing physicians. On the fifty-seven point five billion dollar number, that's about sixty-one thousand dollars in promotion per doctor. Measured against sales, it's twenty-four point four percent of US domestic drug revenue, which totaled about two hundred thirty-five point four billion dollars. Different lenses yield different totals—yet the reconciled view points to a promotional engine that's at least as large as, and very likely larger than, the research effort that's so often front and center in the industry's self-portrait. What's still missing from even that big number matters. IMS's public promotion tallies in two thousand four didn't include company-sponsored meetings. Neither IMS nor CAM fully captures the promotional value embedded in some phase four studies—the post-approval trials that can, in practice, double as extended marketing campaigns. Gagnon and Lexchin flag one analysis for two thousand four that puts thirteen point two percent of US research and development spending—about four point nine billion dollars—into phase four, and notes that roughly three-quarters of that was run by the commercial side of companies, not by clinical development. If a large share of late-stage "research" is led by marketing, it blurs the line we use to separate science from sales. You might ask, why not just open a company's annual report and read off "marketing spend"? Because those documents aren't built to answer this question. They usually merge marketing with administration into a single line, they don't carve out how much was spent in the United States versus abroad, and they bundle in things—like packaging and distribution—that aren't promotion at all. In contrast, both IMS and CAM break out specific promotional channels. Each has blind spots. Together, with a clear correction for the missing third, they give you a much sharper image. Everything comes back to the comparison with research. In the United States in two thousand four, Gagnon and Lexchin put promotion at fifty-seven point five billion dollars. The National Science Foundation's count for domestic industrial pharmaceutical research and development that year is thirty-one point five billion dollars. On sales, promotion sits at twenty-four point four percent; research and development at thirteen point four percent. You don't need to be a cynic to hear the echo of Kefauver's critique. The balance of spending looks tilted toward marketing the products we have, not discovering the ones we don't. This picture isn't an outlier stumbled upon by one team. Marcia Angell, drawing on a different set of sources, pegged US marketing spending at about fifty-four billion dollars in two thousand one. And studies from groups like the Office of Technology Assessment and the Organisation for Economic Co-operation and Development have long suggested that marketing takes roughly a third of industry revenue. The exact totals vary depending on what you count and how you value it—wholesale versus retail, physician-reported versus shipment-tracked—but the scale is consistent. One caveat sits underneath all of this: unmeasured channels still exist. Ghostwritten articles, off-label promotion that skirts the rules, and educational grants that function as soft sales—those are hard to quantify and easy to hide. Gagnon and Lexchin acknowledge this. Even their fifty-seven point five billion dollar estimate likely misses pieces of the puzzle. But compared to the twenty-seven point seven billion that IMS's narrower categories would have you believe, the broader, triangulated view is the better map. If you care about policy, this isn't about scolding; it's about incentives and transparency. When public agencies and payers assess the relative weight of science and sales, they often take the companies' word for it, or they use data sources that are incomplete by design. As Gagnon and Lexchin showed, you can build a more credible baseline by combining physician-level auditing with firm-level validation and by being explicit about what gets left out. Do that, and the conversation about drug pricing, patent protections, and what we reward as a society can take place on solid ground. And there's a practical epilogue here. The US market looms so large—again, about forty-three percent of global drug sales in that period—that how we measure and manage promotion in the United States has outsized global effects. If late-stage "studies" are often steered by commercial divisions, call them what they are and track them accordingly. If a third of promotional activity slips past conventional audits, build the correction into the numbers we cite on Capitol Hill and in medical schools. The science still matters. So does the story we tell ourselves about why pills cost what they do. What Gagnon and Lexchin provided us isn't a final answer; it's a clearer lens. And once you look through it, the outline that comes into focus is hard to unsee: a pharmaceutical market where promotion is not an afterthought or a rounding error, but a center of gravity that's at least as heavy as the research enterprise it sits beside.

Half a century later, we're still asking a very simple question with very large consequences: do pharmaceutical companies spend more driving demand than they do discovering drugs?

In two thousand eight, Marc-André Gagnon and Joel Lexchin approached that question with a detective's mindset. They weren't doing a takedown. They were trying to fix the ruler we've been using.

The numbers most people cite for "how much pharmaceutical companies spend on promotion" come from a handful of sources, each with blind spots. If you want to set policy, or even have an honest debate, you need a measurement that doesn't leave out big chunks of the story. So they focused on one market and one year where the data were richest: the United States in two thousand four.

It was the single biggest market at the time, responsible for about forty-three percent of global drug sales. The premise was straightforward: line up the best data sources, see where they agree, see where they don't, and build a composite that's explicit about what's counted and what isn't.

They worked with two lenses. The first is IMS, the industry's long-time scoreboard. IMS compiles an enormous stream of information from companies and distributors; in the early two thousands, it tracked more than a million products and, crucially, said it covered about ninety percent of US prescription drug sales.

Those numbers give IMS authority—and also tell you its viewpoint. Much of its spending data comes from surveys of firms, which is good for breadth. But it is vulnerable to underreporting, especially on activities a company might prefer to keep offstage.

And IMS's public promotion totals in two thousand four deliberately left out whole categories, like company-sponsored meetings.

The second lens is CAM, a physician-focused auditor. Instead of asking companies what they spent, CAM asks doctors what they experienced. In the United States in two thousand four, that meant surveys of about two thousand primary care physicians and four thousand eight hundred specialists, combined with internal newsletters and cross-checks.

CAM covers the same familiar categories—detailing by sales representatives, samples, direct-to-consumer ads—and it also tracks company-sponsored meetings, e-promotion, mailings, and some clinical trial activity with promotional intent. It counts the sales call differently, too: not just the representative's time in the office. It covers the full cost of making that call happen—local and regional management, training, and the slick brochures representatives leave behind.

Here's the part that makes CAM especially interesting. It runs a validation committee with representatives from major drug makers—think Merck, Pfizer, AstraZeneca, and Sanofi—and, under confidentiality, compares its physician-reported totals to internal corporate spending tallies. When they did that, CAM found something unsettling: about thirty percent of promotional spending didn't show up in the physician surveys.

Roughly a third. They attributed about a third of that gap to simple omissions or incomplete disclosure by doctors, and about two-thirds to channels the survey was never going to catch in the first place. That includes promotion aimed at physician groups they didn't sample, unmonitored journals, and possibly practices that shade into the unethical.

That thirty percent "unmonitored" slice became a key correction in the final accounting.

Now, put the two lenses side by side and you see why Gagnon and Lexchin had to triangulate. Take detailing—the face-to-face visits from representatives that used to be the backbone of drug promotion. IMS pegged detailing for two thousand four at about seven point three billion dollars.

CAM put it at twenty point four billion dollars. That gulf isn't a rounding error; it's a definition. CAM priced the whole visit, overhead and all;

IMS mostly priced the representative's time in the field. If you think the true cost of a sales call includes the managers who organize it and the materials that fuel it, CAM's number looks more realistic.

Samples are the second flashpoint, and here the disagreement isn't just about scope; it's about valuation. IMS reported fifteen point nine billion dollars' worth of samples, using the retail price, and it got its counts by directly monitoring shipments from manufacturers. CAM reported six point three billion dollars, using wholesale price and a survey-based estimate of how many samples physicians actually handed to patients.

They then halved that number to account for small starter packs. Which is "right"? For legal and accounting purposes, companies typically value donated products at retail.

And monitoring shipments is less error-prone than asking busy doctors to remember how much free stock walked out the door. Gagnon and Lexchin leaned toward the IMS approach on samples for that reason.

Other categories line up more neatly. Direct-to-consumer advertising, those glossy television and magazine campaigns, sits at about four billion dollars. Journal ads, the back pages of the medical weeklies, are a relatively tiny five hundred million dollars.

These agreements give you a couple of anchor points. But look where IMS has blanks and CAM has numbers: company-sponsored meetings, which CAM puts at about two billion dollars, and a grab-bag of e-promotion, mailings, and some clinical trial activity that CAM totals at around three hundred million dollars. When you add those together with the richer accounting for detailing, you start to see why the CAM view paints a busier promotional machine.

Gagnon and Lexchin don't pick a side. They build a composite. For two thousand four, they take CAM's broader detailing figure, IMS's retail-valued samples, and the shared numbers for consumer and journal ads.

They bring in CAM's meetings and e-promotion categories that IMS doesn't report. And then they do the uncomfortable but necessary thing: they tack on a line for what neither source counted—the unmonitored thirty percent that CAM's validation work revealed, applied to the physician-facing portions of the spending.

That adjustment is big. In this construction, it comes to about fourteen point four billion dollars. It's not a fudge factor pulled from thin air; it's the delta between what doctors remember and what companies actually budget, based on comparisons CAM could do under confidentiality.

With that in place, three different totals emerge. If you stick with IMS's view alone, two thousand four US promotional spending comes in at twenty-seven point seven billion dollars. If you take CAM's raw, physician-reported total without any correction, it's thirty-three point five billion dollars.

If you apply CAM's thirty percent undercount to its own categories and keep wholesale valuations for samples, you're in the high forties—about forty-seven point nine billion dollars. And if you use what Gagnon and Lexchin argue are the most appropriate figures for each category, that means CAM's broader detailing, IMS's retail-valued samples, the shared ad totals, CAM's meetings and e-promotion, and the thirty percent undercount. You land at fifty-seven point five billion dollars.

How big is that in real life? Do a quick back-of-the-envelope calculation. In two thousand four, the United States had roughly seven hundred thousand practicing physicians.

On the fifty-seven point five billion dollar number, that's about sixty-one thousand dollars in promotion per doctor. Measured against sales, it's twenty-four point four percent of US domestic drug revenue, which totaled about two hundred thirty-five point four billion dollars. Different lenses yield different totals—yet the reconciled view points to a promotional engine that's at least as large as, and very likely larger than, the research effort that's so often front and center in the industry's self-portrait.

What's still missing from even that big number matters. IMS's public promotion tallies in two thousand four didn't include company-sponsored meetings. Neither IMS nor CAM fully captures the promotional value embedded in some phase four studies—the post-approval trials that can, in practice, double as extended marketing campaigns.

Gagnon and Lexchin flag one analysis for two thousand four that puts thirteen point two percent of US research and development spending—about four point nine billion dollars—into phase four, and notes that roughly three-quarters of that was run by the commercial side of companies, not by clinical development. If a large share of late-stage "research" is led by marketing, it blurs the line we use to separate science from sales.

You might ask, why not just open a company's annual report and read off "marketing spend"? Because those documents aren't built to answer this question. They usually merge marketing with administration into a single line, they don't carve out how much was spent in the United States versus abroad, and they bundle in things—like packaging and distribution—that aren't promotion at all.

In contrast, both IMS and CAM break out specific promotional channels. Each has blind spots. Together, with a clear correction for the missing third, they give you a much sharper image.

Everything comes back to the comparison with research. In the United States in two thousand four, Gagnon and Lexchin put promotion at fifty-seven point five billion dollars. The National Science Foundation's count for domestic industrial pharmaceutical research and development that year is thirty-one point five billion dollars.

On sales, promotion sits at twenty-four point four percent; research and development at thirteen point four percent. You don't need to be a cynic to hear the echo of Kefauver's critique. The balance of spending looks tilted toward marketing the products we have, not discovering the ones we don't.

This picture isn't an outlier stumbled upon by one team. Marcia Angell, drawing on a different set of sources, pegged US marketing spending at about fifty-four billion dollars in two thousand one. And studies from groups like the Office of Technology Assessment and the Organisation for Economic Co-operation and Development have long suggested that marketing takes roughly a third of industry revenue.

The exact totals vary depending on what you count and how you value it—wholesale versus retail, physician-reported versus shipment-tracked—but the scale is consistent.

One caveat sits underneath all of this: unmeasured channels still exist. Ghostwritten articles, off-label promotion that skirts the rules, and educational grants that function as soft sales—those are hard to quantify and easy to hide. Gagnon and Lexchin acknowledge this.

Even their fifty-seven point five billion dollar estimate likely misses pieces of the puzzle. But compared to the twenty-seven point seven billion that IMS's narrower categories would have you believe, the broader, triangulated view is the better map.

If you care about policy, this isn't about scolding; it's about incentives and transparency. When public agencies and payers assess the relative weight of science and sales, they often take the companies' word for it, or they use data sources that are incomplete by design. As Gagnon and Lexchin showed, you can build a more credible baseline by combining physician-level auditing with firm-level validation and by being explicit about what gets left out.

Do that, and the conversation about drug pricing, patent protections, and what we reward as a society can take place on solid ground.

And there's a practical epilogue here. The US market looms so large—again, about forty-three percent of global drug sales in that period—that how we measure and manage promotion in the United States has outsized global effects. If late-stage "studies" are often steered by commercial divisions, call them what they are and track them accordingly.

If a third of promotional activity slips past conventional audits, build the correction into the numbers we cite on Capitol Hill and in medical schools.

The science still matters. So does the story we tell ourselves about why pills cost what they do. What Gagnon and Lexchin provided us isn't a final answer; it's a clearer lens.

And once you look through it, the outline that comes into focus is hard to unsee: a pharmaceutical market where promotion is not an afterthought or a rounding error, but a center of gravity that's at least as heavy as the research enterprise it sits beside.

More in Pharmacology, Toxicology and Pharmaceutics →

Related lectures