The paper by Boyd et al. explores the association between the location of methadone treatment centers (MTCs) and crime in Baltimore, MD, USA 1. Comparing spatial patterns of total crime counts during the period 1999–2001 in 100 m concentric buffers around MTCs, convenience stores, residential locations and general hospitals, the authors argue that MTCs do not act as a focus of crime compared with convenience stores. Online publication of the article attracted the attention of the Baltimore media and the research was immediately deployed to push back against recent community protests over the clustering of MTCs in a disadvantaged area of the city 2. There are, however, a number of methodological flaws in the research that are not acknowledged in the article and render its conclusions questionable. For reasons of space, I highlight just a few of the problems below. Thirteen MTCs in Baltimore were grouped together without taking into account the huge variation in neither the daily number of patients treated in each facility nor the socio-demographic characteristics of the areas surrounding them. The number of patients treated was not compared with the number of customers visiting convenience stores, to the number of people using the hospitals nor to the number of people (permanent residents or ambient population) in residential areas. The 1999–2001 study period was determined by the availability of crime data. However, the telephone survey of the MTCs was conducted much more recently. It is unclear precisely where the recalled information for the MTCs originated, whether it referred to the 1999–2001 period or whether it was verified independently. The location of convenience stores was identified using www.switchboard.com a decade after the 1999–2001 crime data referring to those locations. It is unclear whether the criteria for inclusion on www.switchboard.com introduced potential bias. There is no description of a validation exercise that shows these stores were, in fact, operational in 1999–2001 or that these sites were even used as convenience stores at that date. As Boyd et al. acknowledge, alcohol sales outlets are also known to be foci of crime, yet we are not told whether any of these, or other potential crime-attracting activities such as open-air drug markets, operated in any of the study areas. There may have been biases in the reporting of crimes. Time-sensitive crime control or law enforcement measures may have been operational in and around any of the study sites. For example, eight of the MTCs were located on, or near, hospital campuses. Medical facilities in Baltimore tend to be heavily policed, particularly around the Johns Hopkins and University of Maryland hospitals. Crimes were not disaggregated by type, even though it is well-known that drug addicts most commonly commit acquisitive crimes to fund their addiction 3. It is unclear, then, precisely how the researchers envisaged the causal relationship between MTCs and ‘total crime’. They are extremely critical of the crude popular misconception about the relationship between criminality and addiction, yet they are willing to use a version of it as their research hypothesis. The connections between space, place and crime are highly complex and were inadequately captured in this model. Studies have shown that drug users do commit less crime when they are in treatment. However, the Baltimore AIDS Linked to the Intravenous Experience (ALIVE) prospective cohort study has demonstrated that it takes 9 years of treatment to achieve long-term cessation in about 40% of injecting drug users 4. For 60% of injectors, then, treatment either takes longer than 9 years to work or it fails altogether, perpetuating, for many, a cycle of substance abuse, treatment and crime. Boyd et al. seem to assume that arbitrarily chosen concentric buffers of 100 m—equaling a radius of fewer than three blocks—captures the spatial impact of MTCs and other urban phenomena. They therefore avoid the point about why local communities protest the location of MTCs; it is not just the successful patients that matter to them, it is the failures. Treatment providers (a group to which Boyd belongs) do not routinely share aggregate anonymized information about their programs, such as long-term treatment outcomes or where successfully and unsuccessfully treated patients decide to live. Until they do, and with the collusion of methodologically-flawed studies such as this, communities who protest about the clustering of MTCs are under-informed through no fault of their own and are exposed to convenient accusations of NIMBYism. None.
Read full abstract