As distribution-channels tighten, pharma manufacturers are looking for better sales and marketing data
Who is prescribing pharmaceuticals, where does the patient get the drugs that are prescribed, and how much gets paid for filling the prescriptions? Answering these seemingly simple questions has given growth to the multibillion-dollar industry of providing sales data to manufacturers, retailers and payers (including governments).
The biopharma industry has a bottomless appetite for data that tell it what medical professionals are thinking, how sales channels are functioning, what the competitive landscape is, and where the new opportunities are.
Part of this appetite is simply the complexity of the pharmaceutical value chain: while the ultimate customer is the patient, the ultimate payer is usually an insurer or health plan, and the ultimate “buyer” (in the sense of making the product choice) is the physician. Another part is health statistics: the medical conditions of the population; the range of care options of healthcare providers; the tradeoffs between cost and performance that go into formulary designs and co-pay schedules.
Standing in the middle of this data swamp is the Health Insurance Portability and Accountability Act (HIPAA), the federal law that requires patient data to be kept private. The pharma data business has developed numerous techniques to comply with the law—although some states want to go farther in restricting access to the information (see box).
IMS Health defines the universe
Prescription data surveying is a global enterprise, but at least in North America, it is dominated by three players: IMS Health (Norwalk, CT), Wolters Kluwer NV (U.S. HQ: Yardley, PA) and Verispan (Yardley, PA). IMS Health is the far-and-away leader, at least on a global basis, grossing $1.96 billion revenue in 2006. When Wolters Kluwer acquired NDC Health in late 2005 (NDC having been a close competitor of IMS Health), it paid $382 million to buy a business that it said was generating revenues of $163 million, and added that to existing services for a Wolters Kluwer Health unit that grossed just over $1 billion in 2006. (The $1 billion includes businesses in medical publishing and claims processing.)
.jpg)
Verispan is joint venture between McKesson and Quintiles International and doesn’t release revenue information. But a sense of its scale might be approached by looking at employment figures: IMS Health has 7600; Wolters Kluwer Health about 2700, and Verispan, 500.
In a May 2007 investors-day conference, IMS Health presented data indicating it views the global pharmaceutical data market (see figure) at $7 billion. But the data markets and applications it counts don’t align precisely with others; some include data for retail pharmacies, OTC products, or claims processing and back-office functions for healthcare providers.
The dataset of most immediate interest to pharma marketers calculates prescription sales by territory, drug, and class of trade; movement of these figures are often tied into the compensation plans of sales reps. Another group of services tries to calculate the effects of promotional spending on pharma sales, the better to guide such decisions as DTC campaigns, drug sampling programs and the like. A third set revolves around consulting and monitoring services of drug pricing, discounts and rebates.
Although the data-delivery business seems to grow at a healthy clip, staying on top of the business has been a challenge. NDC Health was sold when it encountered legal and technical problems in its services. IMS Health entertained a proposed merger with Dutch publisher VNU in 2005 which was ultimately cancelled. In early January of this year, the company announced a 10% staff cutback, saying that it will “streamline its cost structure in response to healthcare marketplace dynamics.” In its 2006 annual report, Wolters Kluwer noted that its Healthcare Analytics unit had “renewed its service contracts for 2007 with key pharmaceutical customers, but experienced significant price compression in its traditional targeting and compensation product lines.” On the other hand, both IMS and Wolters appear to have a healthy appetite for acquisitions: IMS had completed 26 between 2004 and 2006, and Wolters acquired a medical claims business, ProVation Medical, in addition to NDC Health.
Where the money is
Pharmaceutical manufacturers lose sight of product once it leaves their shipping dock. After wholesalers take possession, medicines land at any of 300,000 distribution points – often with no geographic connection to the wholesaler’s location.
Understanding distribution down to the last patient is critical for uncovering illegalities like diversion and re-importation. A good deal could be written on that topic, but for the time being, at least, pharmaceutical companies are more interested in data to support their $30-40 billion marketing efforts. Consequently most data collection focuses on activities near the point of use.
The most significant event in sales data utilization occurred in the early 1990s with the advent of prescriber-level information. Before that, sales reps collected whatever information they could from pharmacies and doctors, “but pharmaceutical companies had no idea who was prescribing their drugs,” observes Chris Wright, Managing Principal at marketing consulting firm ZS Associates (Evanston, IL).
Numerous factors confound intelligent use of prescriber data. Physicians write millions of prescriptions per year that go unfilled, or which are automatically converted to a generic equivalent. “Traveling prescriptions” (written in New York, filled in Tennessee) and “snowbirds” (filled for months on end at a winter residence) are common, and lead to the wrong sales rep receiving credit.
Until relatively recently, even with prescriber data available, actually collecting it was problematic. In the early days of prescriber data sales reps photographed prescriptions at pharmacies, and the contents would be copied into databases back at the office. As one can imagine, errors abounded.
“Automation has been the key enabler here,” notes Paul Grey, President of Archi-Tech (West Trenton, NJ). Archi-Tech is one of a growing number of third-party analytics firms that turns raw data from top aggregators into “something actionable,” according to Grey. Near-universal computerization systems finally enabled top aggregators to capture prescriber data accurately, in near real-time.
At which point the back door of pharmaceutical distribution was revealed in all its mystery and complexity: What was previously identified as a bucket of prescription cough syrup sold to a chain pharmacy became dozens of transactions, tracked back to prescribers.
PAUL GREY, ARCHI-TECH
Buying data
It might be a truism that, for enough money, someone will collect some level of data to provide an answer or at least reasonable estimate for any question a pharma marketer might ask. More conventionally, though, the data houses identify who and where prescribers are (separating them into deciles based on prescribing activity), and what the trends are in prescribing activity.
A specialized provider like SK&A Information (Irvine, CA), for example, provides rapidly updated prescriber identity and location data. In the fall, the company announced that it had already refined the new NPI (National Prescriber Identification) list published by CMS and based on Medicare reimbursement data. According to the company, almost 20% of the NPI numbers were inaccurate on the day the dataset was released, primarily because physicians hadn’t updated their personal files in a timely fashion. “This study on NPI data quality highlights the challenges that healthcare payers and marketers will face when trying to accurately match . . . the federally mandated NPI number,” noted Dave Escalante, SK&A president. SK&A telephone-verifies prescriber data on a continuous, rolling basis, making some 6,000 calls daily, it claims.
Professional associations like AMA and others also provide doctor ID data; according to one source, providing those lists is a $40-million/year+ revenue stream for AMA.
Sales trend data are collected from payers, retail pharmacies, hospitals and GPOs, and from manufacturers themselves. The latest data source is the wholesaler inventory and sales data being generated as a result of the fee-for-service agreements between leading wholesalers and manufacturerS. These “forward movement” data of sales to retailers and others can, in theory, be matched with the “backward movement” data from scrips and retail sales to provide a truer picture of what is going on in the marketplace, although cleaning and aligning the data is a daunting challenge.
None of the data houses are very clear on where they get their data, and how much it costs. Some of the data-sharing agreements include some bartering some of the value for market-trend reports from the data houses. IMS Health referred indirectly to a “data cost” of about $390 million in 2006 at its investor conference, although it’s not clear whether that figure included software investments as well as data purchases, and a company spokesman says “IMS does not generally report on specific levels of spend related to our quality investments or changes in supply channels, and we do not believe that a significant data spend equates to high-level data quality.” He adds that IMS Health invests considerable resources in “supplier programs in training and education” to improve the initial quality of datasets.
Verispan (Yardley, PA) acquires raw data from pharmacies, payors, and other sources, either purchasing it outright or providing information services in exchange. Verispan increasingly works with smaller, third-party data firms, to provide specialized data products underwritten by a pharmaceutical industry client.
Verispan provides data in conventional spreadsheet or database formats. “It’s very rare to have to do much work on the back end when we aggregate the data into a more usable format,” says VP for product management Jody Fisher. Occasionally, clients ask for pure, unformatted data, which usually leads to additional work since customers typically lack the tools for assimilating this type of data.
Even when clients opt for easily-digestible information products, the quantity of data and the speed at which it arrives can overwhelm. Fisher thinks there is a great need for training and education on interpreting top-level data and applying it to business decisions. “Home product stores have catalogs, but no such thing exists in our industry. Customers spend a lot of time trying to determine what data are useful and how they apply to a business problem.” The effect, Fisher says, is a longer, steeper, more expensive learning curve.
JODY FISHER, VERISPAN
Third-party data adds value
A second significant data spigot opened in the late 1990s, when managed care plans began disclosing formulary status and co-pay tiers for specific drugs. This data becomes increasingly relevant as pharmacy benefit providers proliferate and cater to individual plans’ cost structures.
“Whether a patient pays $20 or $90 is a huge factor in them filling the prescription,” says Chris Wright of ZS Consultants. ZS combines sales data with “activity data,” which consists of sales calls, sampling, and marketing data.
Unfortunately, the sheer number of formularies makes this data “a bit suspect,” resulting in imperfect calculation of which rebate status maximizes a product’s profitability. Drug companies would ideally like to view past transactions to see if patients still fill the prescription when the co-pay is $40 vs. $20, “but that’s difficult to determine because the data isn’t clean,” Wright laments.
The combined efforts of data aggregators and specialty data firms are helping to zero in precisely on prescriber and patient activity. An upshot of ever-increasing granularity is that data becomes a commodity, says Paul Grey, president of Archi-Tech (West Trenton, NJ).
Archi-Tech integrates data from top aggregators and other sources into a single view. Its flagship product, DART (Data Analysis & Reporting Tool), can track product as it moves through the distribution chain. Individual DART modules cover sales trends through prescribers, distributors, and managed care.
“It’s possible, within a very short time period, to target the top prescribing physicians,” or those whose previous behavior suggests they might be influenced if armed with the right information. Grey mentions “innovator” and “traditionalist” prescribers as, respectively, likely or unlikely to act on sales messages.
Verispan has put an emphasis on providing longitudinal patient data, which tracks prescription-consumption data over months or years as patients deal with long-term health issues. The longitudinal data business both is valuable both in its own right, and as a means of tracking drug switching, patient adherence and—ultimately—health outcomes. Surveillance Data Inc. (Plymouth Meeting, PA) is now competing against Verispan and other longitudinal data providers by mining claim forms and other sources to help clients identify trends at the state and national level.
John Ross, VP of client services, says that by matching “de-identified” claims forms and prescription data, the company is able to link diagnostic procedures with prescriber choices and patient status (regimen compliance, subsequent health diagnoses, etc.) These datasets can be examined to determine therapy choices, referral patterns for practice specialties, usage splits for drugs with multiple indications, off-label use, duration of therapy and the like, by disease state, geographic region or healthcare setting. “We provide insight into how a physician uses a product for a particular indication,” Ross adds.
There is clearly the potential to use such longitudinal data for monitoring for unforeseen safety issues, a la Vioxx, but as Ross explains, it’s not a simple matter of toting up health outcomes by medication. “Such studies have to be set up very carefully to have a full continuum of care, and for a qualified set of patients,” he says. “Hospitals don’t routinely collect all the data that might be relevant. The process of qualifying patients, care and health conditions is part art, part science. “Professionals looking at the same datasets can draw different conclusions,” he says.
Pushing the limits of granularity
An example of developing marketing data to guide promotional spending comes from the relatively new firm MindSet Marketing Solutions (Scottsdale, AZ), founded by industry veterans Michael Weintraub and Shel Silverberg.
MindSet’s main product, MindScores, rates geographic locales according to diseases and treatments. MindSet pushes the limits of data granularity with a mathematical algorithm that pinpoints 9-digit zip codes (a small cluster of households) containing a preponderance of patients with conditions or exhibiting an identifiable pattern of pharmaceutical usage. MindSet specifically looks for “brand demanders” (motivated patients) while shunning those of the “doctor knows best” persuasion.
MindSet’s calculations sequentially process and overlay 9-digit zip code data from SDI (profiled in this article), the census bureau, and market research firm MediaMark. SDI contributes medical and prescribing data from claim forms, MRI provides consumer tastes and preferences, and the census bureau furnishes demographics. Each data set undergoes an independent mathematical massage that maps observations, responses, or transactions to 9-digit zips, which are over-laid to identify marketing “hot spots.”
Weintraub claims to beat traditional data sources at identifying advertising targets by 10-15%. The only way to get better numbers is through patient opt-in.
The consumer product companies have been doing this kind of market research for at least 15 years, Weintraub explains, “but nothing like this exists in pharma.” Big pharma is partial to national television and magazine advertising, which according to Weintraub is inefficient. “Every household in the country has toothpaste in the medicine cabinet. Very few, relatively speaking, have diabetes drugs.” MindSet’s service provides a rationale for targeted marketing through the mail, in print media, and through the airwaves.
Outcomes research
Ehlert does not criticize drug firms for purchasing harvested data. He places blame on the information’s gatekeepers, who know the data is used to influence prescribing decisions. “Doctors deny that marketing influences their prescribing practices, but if it didn’t why do drug companies spend $29 billion per year on physician marketing?” Ehlert describes the practice as “a wholesale sellout of science-based medicine” that brings physicians to the edge of a “slippery slope” of split loyalties.
Chris Wright, managing principal at ZS Associates (Evanston, IL) doesn’t think much of the over-promotion concern. “The premise that prescriber-level data fosters over-promotion is questionable.” Lacking such data, Wright holds, companies will simply spend more to reach the same physicians, as they did before prescriber data became available. (Think consumer marketers with poor mailing lists and their inevitable junk mail).
Richard McIntyre of PA Consulting is another privacy skeptic. “All that data, properly managed, can provide huge benefits for sales while helping to promote better health outcomes.” McIntyre thinks integrating patient data from prescriptions and medical interventions might cut the cost of drug development as well by reducing the need for huge clinical trials.
McIntyre’s “ideal world” position is not as far-fetched as it must sound to privacy advocates. The United Kingdom’s National Health Service (NHS) is assembling a database of electronic health records comprising every prescription, condition, and provider contact for 51 million Britons. While concerned about privacy, NHS believes improved care will result.
RICHARD MCINTYRE, PA CONSULTING GROUP
The building blocks for a parallel system in the United States already reside at insurance companies, pharmacies, distributors, and in patient records, but the competing interests of various “stakeholders” (and the generally poor state of our medical information technology infrastructure) prevent aggregation into a single database.
A universal U.S. health database, McIntyre believes, can exist only under the oversight of an agency that is totally independent of today’s stakeholders. That is easier to achieve in the UK, with its sixty-year history of socialized medicine. Staffing such an agency in the United States with high-minded professionals, and keeping it free from financial interests, will be difficult. PC
STATES AND INDUSTRY TUSSLE OVER ACCESS TO PHYSICIAN PRESCRIBING DATA
New Hampshire’s and Maine’s legislative attempts are derailed by federal courts
While patient-level data are sanitized to remove all taint of personality, data firms take no such precautions with prescriber data, a policy that riles ethicists and privacy advocates.
Maine and Vermont have passed laws that seriously restrict use of prescriber information. In August, 2007, IMS Health, Wolters Kluwer, and Verispan sued in U.S. District Court to overturn the legislation, which becomes effective on January 1, 2008. IMS had earlier prevailed in overturning a similar New Hampshire law. The states hold that mining physician data leads to over-promotion of newer, more-expensive drugs that work no better than older medicines. The “big three” aggregators claim the legislation “block[s]…vital healthcare information from the public view,” according to a jointly-issued press release.
Prescriber data should serve to answer scientific questions, says Michael Ehlert, MD, president of the American Medical Student Association (AMSA; Reston, VA), which claims to be the first national medical group to accept no sponsorships, advertising, or freebies from drug-makers. “Science is just a sidebar here,” Ehlert adds. “The driver is detailing.”
Whether or not that’s the case, it’s hard to follow the rationale that physicians need to be protected from pharma reps, or that their prescribing habits are not deserving of review or analysis by outsiders. While HIPAA rules do protect patient privacy (and the data aggregators say that they have extensive internal controls to “de-identify” records), those privacy rules do not extend to the physicians treating the patients. IMS Health has also pointed out that physicians can restrict access to their prescribing history through the AMA Prescription Data Restriction Program, which went into effect last spring, but which has had only lukewarm participation from doctors.
On Dec. 21, U.S. District Judge John Woodcock ruled that the Maine law “amounts to an unconstitutional abridgement” of First Amendment rights, and granted a preliminary injunction preventing the law from going into effect until a full hearing can be held. That was the same rationale used by Judge Paul Barbadoro in the New Hampshire case, who overturned that law. New Hampshire has filed an appeal which has yet to be heard.
Next up is Vermont’s law, for which the three data aggregators had filed for injunctive relief in August.
