Supreme Court Strikes Down Statute Restricting Sale and Use of “Prescriber” Data on First Amendment Grounds — Sorrell v. IMS

[Post by Venkat Balasubramani with comments by Eric]

Sorrell v. IMS Health Inc., 10-779 (June 23, 2011) [pdf]

The Supreme Court struck down a Vermont statute restricting the dissemination of “prescriber-identifiable” information for marketing purposes. While this case was viewed as one that could potentially have far-reaching effects on data-mining and privacy, the majority and dissenting opinions ended disagreeing on the level of protection accorded to commercial speech. 

The Vermont statute was aimed at data-mining companies which gather information regarding what drugs doctors prescribed. Drug companies obtained and used this information to better target their marketing efforts to doctors. The statute restricted the sale or transfer this information to make it harder for the drug companies to target. The statute also restricted pharmaceutical manufacturers from using this information for “marketing or promoting a prescription drug” without the prescriber’s consent. The statute included an exception which allowed this information for education and research purposes. Finally, the statute set aside funds for a “prescription drug education program,” which would inform prescribers as to when generic alternatives became available for drugs which they prescribed. (The statute was not aimed at the dissemination of patient information, which the data-mining companies did not disseminate or sell–as Professor Goldman notes, this case was only nominally about privacy.)

Majority: Justice Kennedy wrote that “speech in aid of pharmaceutical marketing … is a form of expression protected by the First Amendment.” In his view, the statute restricted certain speakers from disseminating certain types of content to particular recipients. Because the statute was discriminatory in this respect, it was subject to a heightened level of scrutiny. Applying this scrutiny, he finds a poor fit between the State’s goals and the statute. 

The first justification asserted by the State was prescriber privacy. However, the existence of numerous exceptions to the statute, including an exception for educational and research uses, undermined this objective. Wile the State pointed to the fact that the statute contained an exception for prescriber consent, the Court finds that this merely offers a “contrived choice.” Either the prescriber withholds consent which allows prescriber-information to be used in support of the State’s message, or grants consent and allows for the wide dissemination of the information.

The second justification offered by the State was that the statute would lower the cost of medical care. The Court finds that which this is a laudable and important goal, the State may not accomplish this goal by “restraining certain speech by certain speakers.” If the State wants to tip the balance in favor of generic drugs, this is an acceptable goal, but it cannot accomplish this goal by hamstringing the marketing efforts of the drug companies who manufacture brand-name drugs.

Dissent: Justice Breyer wrote in dissent that since commercial speech was at issue the Court should employ a lower standard, and not require a perfect fit between the State’s asserted goals and the means. In fact, he even seemed skeptical that speech was at issue at all, since the statute regulated the transfer of data, and not necessarily a particular message. In his view, this was just one aspect of the State’s overall regulatory program which the government should have room to pursue.

The big question was whether this decision will have broader effect for data mining or behavioral targeting. I’m guessing it will probably have less effect than what people envisioned. More than anything this case represents a victory for commercial speech, which has steadily inched up the scale in the amount of protection it is accorded. 

The fact that the sale of data is characterized as speech deserving of a high degree of protection may make it tougher for legislators to enact laws which regulate the transfer of consumer information, but what bothered the majority here is that the purported privacy interest was ill-served by the statute and the fact that the state sought to favor one set of products by suppressing the flow of data to its marketers (while allowing the competition to use the information). To use an analogy, the State went beyond restricting the transfer of information to car manufacturers for marketing purposes. It authorized the use of this information only by manufacturers of electric cars.

Will this opinion affect more general laws aimed at the collection, use, or transfer of information for marketing purposes? Some of these already exist in specific contexts (e.g., COPPA for information collection from children under 13; CAN-SPAM includes provisions restricting the transfer of email address in certain contexts; the Video Privacy Protection Act deals with video tape rental records; and there’s of course HIPPA, which deals with patient records). Provided that the regulation is not discriminatory, this case should not present an impediment to enacting this type of legislation.

The Court’s treatment of the consent issue was interesting. Are doctors really powerless from a bargaining standpoint that they can’t take steps in the market to somehow fix the supposed forced consent issue? The majority opinion had a paternalistic tone to it, which may make sense if the statute was dealing with patient records and patient choices, but I found it odd, given that the statutory scheme was about sales pitches to doctors!

Other coverage:

CDT Statement on Supreme Court Decision in Sorrell v. IMS Health (CDT)

Information is not Beef Jerky (info/law)

Supreme Court Rx Records Case: Not So Bad (info/law)

Court’s data-mining ruling: big change on commercial speech? (First Amendment Center)


Eric’s comments:

I agree with Venkat’s comments but I wanted to add a few more:

Let’s start with two basic premises: (1) healthcare costs have spiraled out of control, and (2) doctors’ medical decisions are a big part of that. For example, it turned out that Wisconsin healthcare costs were unusually high because Wisconsin doctors are more likely to require various tests/diagnostics than doctors in other areas. It was unclear if this was because local doctors had a heightened fear of malpractice liability, different regional norms, different assessments of medical best practices or something else. However we get there, the cumulative effect of Wisconsin doctors’ choices was dramatic: enormous healthcare insurance premiums (and heaven help you if you didn’t have medical insurance).

Therefore, it’s quite logical for a state to look more closely at doctors’ drug prescribing choices both as a matter of public health and fiscal responsibility. If a state could identify systematic drug prescribing judgment-calls that unnecessarily jack up medical costs, it would be in the public interest to curb those.

The theory behind Vermont’s statute (and other states in the Northeast that adopted similar laws) is that doctors are overprescribing branded pharmaceuticals when they could be prescribing generic drugs instead; and that doctors are overweighting branded drugs because drug reps are bending their ears to persuade them to prescribe the branded drug in preference to the generics; and that the drug reps are successfully persuading doctors to make this choice because the drug reps are armed with the doctor’s past prescribing practices and therefore can make a more effective but socially unwanted sales pitch that is overriding the doctor’s own medical judgment that would otherwise lead the doctor to prescribe the generic drug.

Stated this way, we see that the statute is targeting a problem (high medical costs) through a very indirect means (suppress a doctor’s past prescribing practices from drug sales reps). Should any inference in this logic chain be wrong, then the statute is, at best, ineffectual. However, there are a broad range of other ways the state could try to remediate the problems with branded drugs jacking up medical costs, including monkeying with the states’ reimbursement policies for branded vs. generic drugs; counter-educating doctors about the merits of generic drugs; educating patients about the bioequivalence of branded and generic drugs so they could make their own substitutions or push their doctors to prescribe generics when available; etc. The state was trying some of these as well.

There are two other aspects of the unique situation of doctors that I feel get lost in the top-line headlines. First, the whole concern here is face-to-face meetings between doctors and drug sales reps. Given how hard it is for us as patients to see our doctors face-to-face, it is a little shocking that doctors are voluntarily choosing to spend discretionary time with the drug reps for meetings that the doctors know are sales pitches. Why are the doctors allocating their time this way?

Putting aside the odds that the drug sales rep is very attractive and charming (have you ever noticed that on the Survivor TV show, the former beauty queens all list their job title as “pharmaceutical sales”?), it’s presumably because doctors find the meetings valuable to them. Indeed, even Justice Breyer in dissent acknowledges that the drug sales reps impart valuable information in those meetings. The state statute very explicitly tried to make those meetings less useful to doctors by making the drug sales rep less well-prepared. If the drug sales reps wanted to provide tailored information to the doctor’s practice, the drug reps would have to take time out of these meetings to interrogate doctors about their prescribing practices; and if the doctors concluded that the meetings weren’t productive because they took too much time on irrelevant or uninteresting chatter, the doctors would simply skip the meetings entirely and perhaps lose the other valuable information being exchanged in the meetings. So before we get too worked up about the evilness of the drug reps working against the consumer interest, we should not forget that very busy doctors are voluntarily choosing to take these meetings, and doctors can and will choose otherwise when it doesn’t make sense for them.

Second, the opinions talk a lot about “privacy,” and this baffled me. Everyone agrees we’re not talking about patient privacy. Instead, there is some back-and-forth on DOCTOR privacy in their prescribing patterns. What??? In this situation, doctors are business operators making business decisions. Tracking their prescribing decisions is similar to tracking how other businesses interact with third party vendors. We might have concerns about how tracking these decisions exposes trade secrets or competitive intelligence, but we wouldn’t talk about business decision-making as being covered by “privacy” concerns. So the notion that this case teaches us anything about “privacy” law confuses me greatly.

In the end, what we really want to know is whether this case will enable more First Amendment challenges to behavioral advertising or other privacy statutes. I personally don’t feel any more knowledgeable about that question after reading the majority and dissenting opinions. Part of this reflects my cynicism about the Supreme Court’s First Amendment’s jurisprudence, which still seems to me that it’s developed case-by-case instead of forming a coherent body of jurisprudence. Part of this reflects the quirks of Vermont’s statute, which suffered from two easily targeted defects. First, it sought to regulate the doctor-drug rep conversation, setting up the possibility of content-based and perhaps even viewpoint-based review. Second, Vermont changed its position about who could get access to the database of prescribing information, and this flip-flopping gave the majority extra reasons to suspect the state’s policy rationales. So I suspect that Vermont or other states could find a way to draft around this opinion if they chose to; and I’m skeptical that other behavioral advertising or privacy laws would set off the justices’ First Amendment hackles like this statute did.