June 29th, 2006
For Sale: Physicians' Prescribing Data
By Robert Steinbrook, M.D.
New England Journal of Medicine
Since the early 1990s, health care information companies have bought electronic records of prescriptions from pharmacies and other sources and linked them with information about doctors that is licensed from the Physician Masterfile of the American Medical Association (AMA). These information companies, the largest of which is IMS Health of Fairfield, Connecticut, have then compiled and sold individual physicians’ prescribing data to pharmaceutical manufacturers. The business is lucrative. But a growing number of physicians have rebelled after becoming aware that drug companies have access to their data — in some cases because zealous sales agents have confronted them with their prescribing histories.1 The abuses have occurred despite “best practice guidelines” from the AMA that include admonitions that industry and its representatives should keep prescribing data confidential, that companies should prohibit disclosure “by sales representatives to any other party,” and that “the use of prescribing data to overtly pressure or coerce physicians to prescribe a particular drug is absolutely an inappropriate use.”
The ferment has had several consequences. First, the AMA has created a Prescribing Data Restriction Program. Physicians are now able to deny all sales representatives access to their individual prescribing data. The restriction is limited to sales representatives and their direct supervisors; physicians will not be able to deny access to other officials at pharmaceutical companies.2
Second, the California Medical Association (CMA), in conjunction with IMS Health, is planning a program that will allow physicians who do not restrict access to their information to see their own data, comparative data, and educational material that will focus on prescribing for common diseases as well as categories of medications that are heavily prescribed. After pilot testing is completed this year, the program is expected to be available in 2007. The CMA is likely to make the program available to all California physicians, not just association members. Eventually, similar programs may be available to physicians in other states.
Third, some states may enact legislation that would supersede the AMA program. They may ban outright the use and sale, for most commercial purposes, of prescription data in which prescribers are identified. Typically, the use of such data for marketing, promotion, or any activity related to sales agents would be prohibited. However, the noncommercial use of an individual’s prescribing data for health care research or public health purposes would be permitted, as would the commercial use of data in which providers were not identified and were grouped by ZIP Code, geographic area, or specialty. In 2003 and 2004, a legislative ban was seriously considered in California at the request of the CMA. After the CMA and IMS Health agreed to develop the forthcoming program, the ban was not enacted. Earlier this year, the New Hampshire legislature approved a ban. As of mid-June, the legislation was on its way to Governor John Lynch, who has said he will sign it. The restrictions would go into effect immediately.
The Masterfile includes current and historical data on physicians from the time of their undergraduate medical education onward; the database encompasses information not only about AMA members, but also about nonmembers and many graduates of foreign medical schools who reside in the United States. As of December 2004, the Masterfile contained data on 820,000 living physicians. It does not contain prescribing data.
Companies that license data from the Masterfile are subject to AMA rules and regulations. According to a recent article by Robert Musacchio, of the AMA, and Robert Hunkler, of IMS Health, the rules of the new program “allow the industry to retain access to prescribing data for most purposes, but they require companies to police their own sales forces."2 They warned that unless companies “comply in letter and spirit with the requirements,” they “will sabotage the success of the program and pave the way for legislation that imposes stricter measures on the industry.”
The current controversy is about collecting and selling physicians’ prescribing information, not data that identify patients, whose confidentiality is protected under the Health Insurance Portability and Accountability Act. Prescribing data can be linked to individual physicians in various ways — by means of Drug Enforcement Administration registration numbers, state medical license numbers, or internal identifiers used by pharmacies, for instance.
According to a national survey of physicians conducted by the Gallup Organization for the AMA in 2004, 77 percent were aware that pharmaceutical companies had access to physicians’ prescribing data, and 66 percent opposed the release of these data to sales representatives. However, 68 percent believed that the release of the data had “neither a particularly positive nor a particularly negative impact on their medical practices”; 77 percent indicated that their concern would be alleviated if they were given the choice of preventing the release of their prescribing data to sales representatives.3
Of course, prescribing data have other uses, some of which have little to do with marketing. For example, they can be used by public health agencies and in studies about drug interactions, prescribing trends, long-term effects of medication, and the prevention of medical errors. In a 2004 report, the Board of Trustees of the AMA, while acknowledging the concern about the release of prescribing data and their commercial uses, noted that pharmaceutical companies believe they “would struggle without the data, resulting in physicians seeing an increase in sales calls, less targeted educational information, and fewer or less relevant drug samples. The public good uses of the data would be severely cut back or eliminated as [health care information organizations] would no longer have a financial incentive to maintain the data."3
In 2005, IMS Health had operating revenue of $1.75 billion, substantially all from sales to the pharmaceutical industry, according to its annual report.4 Forty-eight percent was from “sales force effectiveness offerings,” which include “sales territory reports and prescription tracking reports.” Pharmaceutical manufacturers use such reports to decide how much to pay their sales representatives as well as for many other purposes. IMS Health collects information on about 70 percent of all prescriptions filled in community pharmacies and projects nationally representative data. Some organizations, such as Kaiser Permanente, do not make their prescription data available because they object to their use in influencing prescribing patterns; the limit, however, applies only to prescriptions filled in their pharmacies.
In 2005, the AMA received $44.5 million in revenue — about 16 percent of its total revenue — from the sale of database products.5 The association does not release information indicating what percentage of its database revenue comes from licenses to IMS Health and its competitors, such as Dendrite International, Verispan, and Wolters Kluwer.
Prohibiting the release of prescribing data to sales representatives will not put an end to another practice to which some physicians object: the use of such data by managed care or pharmacy benefit managers. These entities have sources of information that are independent of the AMA Masterfile. It also will not stop visits from sales agents, which doctors have always had the right to refuse, nor will it curtail the marketing of drugs. According to the AMA, the potential effects of restricted release may include a reduction in the number of “offers physicians currently receive from the pharmaceutical industry, such as drug samples, CME programs and speaking engagements.”
Although broad notification efforts are planned, many physicians remain unaware that they can restrict access to their data, so it is impossible to predict how many will do so, or whether the monitoring system to prevent abuses will be effective. Programs that would let physicians view their data along with comparative data and educational materials are not yet available, so their features and value are not known. It is also uncertain whether all physicians in the United States will eventually have access to them. If the AMA were to stop licensing the Masterfile to aid the compiling of prescribing data, it would separate itself from the practice, but its revenue would decline, and health care information companies would no doubt seek to obtain similar data from other medical societies and commercial entities. Finally, fueled by general objections to the practice, other states may follow New Hampshire and enact broad bans on using and selling prescribing data. Thus, the fallout from this controversy is likely to remain uncertain for at least several years.
The Prescribing Data Restriction Program
The AMA’s Prescribing Data Restriction Program (http://www.ama-assn.org/ama/pub/category/12054.html) allows physicians in the United States, regardless of whether they are members of the AMA, to deny all pharmaceutical sales representatives access to their individual prescribing data. The Web-based system also allows physicians to report concerns about the inappropriate use of prescribing data by a specific pharmaceutical-company representative or a specific company and to make “general observations” about the uses of the data. The program became available in May 2006.
“Restricted data” are defined as “any product-level data regarding the prescribing practices of an identifiable physician who has [restricted access to his or her data]. This includes measures of prescription volume in absolute and percentage terms, the associated dollar value of a physician’s prescribing, any indicators of change in these measures, as well as any means of ranking, benchmarking, or grouping physicians that may reveal prescribing habits for a particular product."2 Efforts to identify the individual physicians associated with restricted prescribing information are also prohibited.
A restriction is valid for three years and can then be reinstated; it can be removed at any time. A restriction applies equally to all sales representatives, regardless of their company. Pharmaceutical manufacturers are required to check the program database at least quarterly; it may take as long as 90 days from each such check for them to cut off their sales representatives’ access to data newly marked as restricted. The new program is separate from other restrictions that physicians can already place on the uses of their Masterfile data, such as the “no contact” list, which prevents doctors from receiving solicitations through the mail.
The AMA says that it will gauge compliance by monitoring physicians’ complaints and that companies with isolated infractions will receive warnings. Those with a pattern of abuse may lose access to the data altogether.2
Dr. Steinbrook () is a national correspondent for the Journal.
An interview with Dr. Jack Lewin, chief executive officer of the California Medical Association, can be heard at http://www.nejm.org.
1. Saul S. Doctors object to gathering of drug data. New York Times. May 4, 2006:A1.
2. Musacchio RA, Hunkler RJ. More than a game of keep away. Pharmaceutical Executive. May 1, 2006. (Accessed June 8, 2006, at http://www.pharmexec.com/pharmexec/article/articleDetail.jsp?id=323311.)
3. Reports of board of trustees of the American Medical Association. Use of physician and patient prescribing data in the pharmaceutical industry (Resolution 606, I-03). Interim meeting of the AMA House of Delegates, December 2004. (Accessed June 8, 2006, at http://www.ama-assn.org/meetings/public/interim04/bot_reports.pdf.)
4. Health IMS. 2005 Annual report. (Accessed June 8, 2006, at http://www.imshealth.com/ims/portal/front/indexC/0,2478,6599_1825,00.htm.)
5. American Medical Association. 2005 Annual report. (Accessed June 8, 2006, at http://www.ama-assn.org/ama/pub/category/12528.html.)