Privacy protections and law enforcement use of prescription drug monitoring databases

• Published date: 01 July 2021
• Author: Anne E. Boustead
• Date: 01 July 2021
• DOI: http://doi.org/10.1111/lapo.12174

ORIGINAL ARTICLE

Privacy protections and law enforcement use of

prescription drug monitoring databases

Anne E. Boustead

School of Government and Public Policy,

University of Arizona, Tucson, Arizona, USA

Correspondence

Anne E. Boustead, School of Government and

Public Policy, University of Arizona,

315 Social Sciences Bldg. P.O. Box 210027,

Tucson, AZ 85721-0001, USA.

Email: boustead@arizona.edu

Abstract

Prescription drug monitoring programs (PDMPs) are

databases that can be used by healthcare professionals

to identify problematic drug-seeking behavior. Law

enforcement officers can also obtain PDMP informa-

tion, raising significant privacy concerns. In this paper,

I use regression analysis to explore the association

between state PDMP protections and law enforcement

information requests. I find that while requiring law

enforcement to meet a specified standard of proof prior

to accessing PDMP information is associated with fewer

requests, other methods of regulating law enforcement

access are not. These findings provide important and

novel evidence about law enforcement behavior in

response to privacy protections.

1|INTRODUCTION

As harms due to the misuse of opioids—a category of drugs including both prescription analgesics

such as OxyContin and illicit narcotics such as heroin—have skyrocketed over the past two decades

(Rudd et al., 2016), policymakers have struggled to respond. Forty-nine states have implemented

prescription drug monitoring programs (PDMPs), which are “centralized database[s] . . . to which

pharmacies report dispensed medications . . . [along with] date, prescriber, and patient”(Deyo

et al., 2013, p. 603). These databases can be used to identify individuals with multiple overlapping

opioid prescriptions from different physicians and pharmacies, a potential sign of nonmedical use.

While they were originally intended as a mechanism for policing controlled substances, PDMPs

have recently been hailed as part of the new “public health”approach to addressing the worsening

opioid crisis (Haffajee, 2015, p. 1657).

While PDMPs have been embraced as a tool for preventing opioid misuse, they also raise

significant privacy concerns. PDMPs contain hundreds of millions of prescription records, and

they allow the user to track medical information for an individual over time and across pro-

viders. In many states, prescriptions that must be reported to PDMPs include drugs whose use

allows inferences about a wide range of highly sensitive medical information, including

DOI: 10.1111/lapo.12174

©2021 University of Denver and Wiley Periodicals LLC.

Law & Policy. 2021;43:229–261. wileyonlinelibrary.com/journal/lapo 229

treatment for mental illness and HIV status (Mejia, 2014). Furthermore, the majority of states

allow law enforcement to access PDMP data without obtaining a warrant, instead relying on

less stringent mechanisms to protect the significant privacy interests at stake. While law enforce-

ment officers have reported that PDMP information is useful for identifying potentially illegal

activities and communicating with prescribers and pharmacists (Perez et al., 2017), allowing

law enforcement access to sensitive medical information may threaten important privacy inter-

ests. In particular, without strong legal protections, there may be concerns about law enforce-

ment accessing sensitive medical data arbitrarily, or using knowledge gleaned about medical

conditions to inappropriately target people for investigation.

Variation in PDMP protections provides a rare opportunity to explore whether requiring a

warrant or subpoena prior to pursuing evidence impacts law enforcement information seeking.

Despite the doctrinal and normative importance of using justification standards such as the

warrant requirement to regulate law enforcement surveillance, there is currently little empirical

evidence on whether these standards change law enforcement behavior. Consequently,

addressing these issues in the context of PDMP information will provide insights into law

enforcement behavior that are useful for understanding the effectiveness of privacy protections

more broadly.

In this paper, I explore variation in PDMP protections across states and analyze whether

this variation is associated with the frequency of law enforcement requests to access PDMP

information. I start with a general overview of our current understanding of law enforcement

behavior in response to legal protections on information, focusing on the rational choice model

that explicitly or implicitly underlies judicial assumptions about this behavior. I also justify the

importance of considering whether and how law enforcement behavior changes in response to

legal protections. I then describe state PDMP programs and how access to the data generated

by these programs has been regulated, emphasizing the importance of state statutory law. Next,

I quantitatively analyze state administrative data describing law enforcement requests for access

to PDMP information across the country, as well as substate data from Utah. This analysis

focuses on addressing two main questions. First, do law enforcement requests for access to

PDMP data vary based on the legal protections afforded to that data under state statutes? Sec-

ond, based on evidence from Utah, do the observed differences before and after the imposition

of a warrant requirement vary based on agency or locality characteristics? I find that state laws

requiring high standards of proof to obtain PDMP information are associated with lower rates

of law enforcement requests for that information. Evidence from Utah suggests that the imposi-

tion of a warrant requirement may have less of an impact on requests for information by

narcotics-related agencies than on requests made by other types of agencies. I conclude by dis-

cussing the implications of my findings for the regulation of law enforcement access to informa-

tion, arguing that these findings both support and complicate the application of rational choice

theory to law enforcement surveillance decision making.

2|LAW ENFORCEMENT BEHAVIORAL RESPONSES TO

SURVEILLANCE REGULATION

The Fourth Amendment’s protection of the right to be secure “against unreasonable searches

and seizures”(U.S. Const. Amend. IV) is implemented through privacy interests. Under the test

established by the Supreme Court in Katz v United States (1967), whenever the government

engages in activities that violate an individual’s reasonable expectation of privacy, they are con-

ducting a Fourth Amendment search and generally must first obtain a warrant based on proba-

ble cause. However, despite the doctrinal significance of the warrant requirement and the

centrality of privacy in Fourth Amendment protections, very little is currently known about

whether and how the warrant requirement changes law enforcement practices. In this section,

230

I review the current research on law enforcement search behavior, demonstrating that common

jurisprudential assumptions about this behavior are based on an application of rational choice

theory that has not been empirically verified in the context of technological surveillance. I then

argue that a clearer and more empirically supported understanding of how justification stan-

dards such as the warrant requirement change law enforcement behavior is necessary to ensure

that the Fourth Amendment serves its intended role in ensuring individual privacy, regulating

police conduct, and promoting judicial oversight.

2.1 |Rational choice theory and technological surveillance

In crafting Fourth Amendment limitations on law enforcement use of technological surveil-

lance, courts often assume that the frequency with which law enforcement uses surveillance

depends heavily on its cost (legal or otherwise). In United States v Jones (2012), Justice Alito

noted that location tracking was traditionally “difficult and costly and therefore rarely under-

taken”(Jones, 2012, p. 965). The New Jersey Supreme Court has stated that cell phone location

data can “function as a substitute for 24/7 surveillance without police having to confront the

limits of their resources”(State v Earls, 2013, p. 642). These assumptions are predicated on a

model of law enforcement officers as rational consumers of technological surveillance, shifting

their consumption in the face of shifting surveillance costs.

In making these assumptions, judges are explicitly or implicitly extending the rational choice

theory of decision-making commonly used in the law and economics literature to law enforce-

ment use of surveillance. Under rational choice theory, social-level phenomena are thought to

be the result of many individual actors making choices that would maximize their utility within

their circumstances (MacDonald, 2003). While discussions of rational choice theory frequently

focus on easily quantified financial and time costs, there are also “thick”models of rational

choice theory that allow for the impact of intangible values on decision-making (Hechter &

Kanazawa, 1997). When judges note that expensive forms of surveillance will only be used

“during cases of unusual importance”(Jones, 2012, p. 963), they are implicitly assuming that

the decision to use technological surveillance is in part a function of a cost–benefit analysis

through which officers determine that the information they expect to obtain will be worth the

time, expense, and hassle of conducting surveillance.

Under this model, legal requirements for obtaining surveillance can be seen as imposing a

special type of cost relating to the resources necessary to develop evidence and complete admin-

istrative requirements. When a state decides to impose a warrant requirement on law enforce-

ment access to information sources such as PDMPs, they increase the expense necessary to

obtain that data. Consequently, courts can use legal standards as a corrective mechanism when

new technologies reduce the cost of conducting surveillance, ensuring that the ability of the gov-

ernment to obtain information remains stable relative to the individual’s ability to conceal it

(Kerr, 2011) and that law enforcement does not engage in patterns of surveillance that inconve-

nience large numbers of innocent people (Bambauer, 2014). For example, Bankston and

Soltani (2013) argue that application of the Fourth Amendment to emerging forms of techno-

logical surveillance should depend on the change in the cost of obtaining information.

Although rational choice theory appears to underlie many arguments surrounding the appli-

cation of Fourth Amendment protections to technological surveillance, there is currently little

empirical evidence concerning whether—and under what circumstances—law enforcement use

of surveillance changes in response to changes in legal requirements or cost. Minzner and

Anderson (2013) used formal economic modeling to argue that the time and expense require-

ments impose greater restraint on law enforcement use of wiretaps than the warrant

requirement. Slobogin (2011, p. 331) reviewed the empirical literature on whether requiring law

enforcement to obtain a warrant improves the probability that a search will return relevant

BOUSTEAD 231