Assessing police officers’ attitudes and legal knowledge on behaviors that impact HIV transmission among people who inject drugs

Abstract

Background

Policing practices such as syringe confiscation and arrest can act as important social-structural drivers of HIV risk among people who inject drugs (PWID). However, police referral to treatment and other services may improve the health of PWID. Little is known about the role of modifiable attitudinal and knowledge factors in shaping officer behavior. Using baseline findings from a police education program (PEP), we assessed relationships between drug policy knowledge and attitudes towards public health interventions with self-reported syringe confiscation, drug arrest, and service referral among street-level police in Tijuana, Mexico.

Methods

Between February, 2015 and May, 2016 we surveyed 1,319 police officers who reported syringe contact. The self-administered survey focused on attitudes, knowledge, and behaviors related to drug policy, public health, and occupational safety. We used ordinal logistic regression to model the odds of syringe confiscation, arrest for heroin possession, and referring PWID to health/social programs.

Results

The sample was mostly male (87%) and had at least a high school education (80%). In the last six months, a minority reported always/sometimes confiscating syringes (49%), arresting someone for heroin possession (43%), and referring PWID to health and social programs (37%). Those reporting needlestick injuries (NSI) had 1.38 (95% CI: 1.02–1.87) higher odds of reporting syringe confiscation. Officers who had favorable views on laws that treat addiction as a public health issue had lower odds (aOR= 0.78; 95% CI: 0.59–1.03) of arresting PWID. Those agreeing that it was their role to refer PWID to health and social programs had higher odds of reporting such referrals (aOR: 3.32, 95% CI: 2.52–4.37). Legal knowledge was not associated with these practices.

Conclusion

Changing drug policy and knowledge may be insufficient in shifting police behavior. Modifying officers’ occupational risks and attitudes towards harm reduction interventions can facilitate efforts to align police practices with PWID health.

Introduction

People who inject drugs (PWID) interface with numerous social and structural factors that increase their vulnerability to bloodborne infections. Criminalization of injection drug use ensures the prominence of drug laws and their enforcement among such factors. Certain policing practices, including syringe confiscation and drug arrests have been shown to be acutely deleterious to the HIV “risk environment” of PWID (Rhodes, 2002). Such policing practices are especially pervasive in settings where drug use is highly criminalized and stigmatized (e.g. United States (Beletsky et al., 2014), Mexico (Pinedo et al., 2015; Pollini et al., 2008), Russia (Sarang et al., 2006), Ukraine (Izenberg et al., 2013), Vietnam (Ahmed et al., 2015), and Malaysia (Michalopoulos et al., 2016)). These are some of the same settings where HIV prevalence among PWID is much higher than in the general population. Conversely, police officers routinely come into contact with PWID, creating opportunities to engage law enforcement in health promotion and HIV risk reduction activities. For instance, police can and do provide this hard-to-reach population with information about and links into harm reduction programs and services (DeBeck et al., 2008; Silverman et al., 2012). Nonetheless, a number of barriers limit the potential of police activities to promote the health of PWID.

The public health importance of police interactions with PWID is especially relevant in Tijuana, Mexico, which has one of the largest PWID populations in the country. Tijuana shares one of the world’s busiest land border crossings with San Diego, California; correspondingly, transmission of blood-borne diseases there has been shown to be bidirectional (Brouwer et al., 2006; Mehta et al., 2015). Reflecting global trends, our research found syringe confiscation to be associated with 2.5 increased odds of HIV seropositivity among PWID (Beletsky et al., 2013a). Further, PWID who reported recent arrest were over twice as likely to report syringe sharing compared to PWID who had not been arrested (Pollini et al., 2008). These practices are likely drivers of localized HIV and hepatitis C virus (HCV) epidemics among PWID in Tijuana, with potential for rapid geographical diffusion through active north- and southward migration through the unique Tijuana – San Diego border gateway (Brouwer et al., 2006; Mehta et al., 2015; Strathdee and Magis-Rodriguez, 2008; Strathdee et al., 2012; White et al., 2007).

On the other hand, frequent police encounters can highlight the potential protective effects of policing on PWID health. In our ongoing research with a cohort of PWID in Tijuana, 76% reported a recent encounter with police (Gaines et al., 2015). Despite this, only 10% reported ever enrolling in a drug/alcohol rehabilitation program and only 3% reported ever receiving information about health or social programs from police officers. Similarly, only 2% reported ever being diverted to drug treatment services by police and a mere 0.3% reported ever being referred to syringe exchange or other health services (Beletsky et al., 2016). It is not clear if the low referral rate pertains to officers’ limited knowledge, negative attitudes, or other factors, such as lack of viable service options.

Taking advantage of police contact to operationalize drug treatment uptake was one of the core elements of drug policy reforms enacted by the Mexican government in 2009. In the context of decriminalizing the possession of small amounts of drugs for personal use, this “narcomenudeo” reform amended the General Health Law to include mandated treatment diversion for habitual users (Russoniello, 2012). Prior research is inconclusive on whether legal reforms or department-level policy changes are capable of such harmonization. Limited supervision, lack of training, and broad enforcement discretion allows officers to continue to engage in practices that are not entirely consistent with formal rules (Beletsky et al., 2011b; Burris et al., 2004; Sabet, 2012). Clarifying the relationship between policy interventions and police practices targeting PWID is especially relevant in Mexico where syringe possession has never been criminalized, but unauthorized syringe confiscation by police remains a major driver of PWID disease risk (Beletsky et al., 2013a)

Amidst a growing focus on criminal law reform and police practices as a structural determinant of PWID health, there is a critical need for research on factors that may shape key behavioral outcomes among police. Despite the recognized epidemiological importance of syringe confiscation, drug arrests, and service referral practices, little is known about the relationship between officer demographics, policy knowledge and attitudinal measures on the one hand and key behavioral outcomes on the other. To close this gap, we analyzed baseline findings from “Proyecto Escudo” (“Project Shield”), a structural HIV prevention intervention that delivered training on occupational safety with legal and harm reduction content. We hypothesized that limited drug policy knowledge would be associated with increased frequency of syringe confiscation and drug arrest. Additionally, we hypothesized that supportive attitudes concerning harm reduction services would be associated with increased frequency of referral to health and social programs.

Methods

Study population

The Secretaría de Seguridad Pública Municipal in Tijuana (Municipal Ministry of Public Safety) is the largest police force in the state of Baja California and employs approximately 2100 active duty officers across 11 districts (Shirk et al., 2015). To become an officer, the individual must be at least 18 years of age, and as of 2009, must have possessed at least a high school education. Officers in Tijuana are among the most highly paid in the country (earning from $800 – $2,000 USD per month) with a moderate turnover (6.2% annually). Based on the PWID-specific focus of this analysis, we included only those officers who reported any contact with syringes while carrying out their law enforcement duties.

Data collection

Study procedures have been described previously (Strathdee et al., 2015). Briefly, from February 2015 – May 2016, over 1,800 officers received a three-and-a-half-hour police educational program (PEP) that included the following training goals: 1) occupational safety related to needlestick injuries (NSI) and HIV 2) narcomenudeo drug reform and legality of syringe possession, and 3) drug pharmacology and harm reduction services. Participants self-administered a pre- and post-training survey that reviewed concepts discussed in the training, attitudes on harm reduction services, and frequency of behaviors in the past six months. The items included in the survey had been drawn from similar training evaluation initiatives in other settings (Beletsky et al., 2011a; Beletsky et al., 2013b). Survey questions were then piloted among senior level Tijuana police officers. We incorporated the officers’ feedback in the final version of the instrument. The study protocol was approved by the UCSD and Universidad Xochicalco institutional review boards and all participants provided written informed consent.

Outcomes

Since we restricted the research to officers who had contact with syringes, we considered three self-reported policing behaviors to be directly relevant to PWID health risk: syringe confiscation, arrest for heroin possession, and referral of PWID to health/social programs. The frequency of these behaviors in the last six months was measured on a 4-point Likert scale ranging from “always”, “sometimes”, “rarely”, “never”.

Independent variables

In addition to sociodemographic variables (age, sex, educational level), district where most recently stationed, and history of NSI, the survey also prompted officers for information on their legal knowledge and attitudes related to harm reduction services. The two legal knowledge items of interest were: syringe possession, as defined by the response on the number of syringes that could be possessed legally (none, 2, 5, 15, unlimited), and the specific threshold amount of heroin decriminalized under the narcomenudeo drug law reform (none, 50 mg, 500 mg, 5 g). For both items, we considered legal knowledge on a gradient. For example, the highest level of knowledge was demonstrated by correctly identifying the technically-accurate answer (unlimited amount of syringes, 50 mg of heroin). However, we also considered partial conceptual knowledge reflecting some measure of understanding of the law, as determined by choosing a response other than “none” or “don’t know” in addition to the correct technical answer. Thus, if an officer responded that 500 mg of heroin was allowed (even though the correct amount is 50 mg), this was reflective of the officer understanding drug decriminalization as a concept.

We also assessed whether officers’ attitudes towards harm reduction services were associated with the self-reported behaviors of interest. These attitudinal measures captured to what extent officers agreed with statements that: 1) methadone reduces criminal activity, 2) syringe exchange programs (SEPs) increase police risk of occupational NSI, 3) laws treating addiction as a public health problem make the job of officers easier, 4) people addicted to drugs do not care about their health 5) drug addiction is a disease and 6) it is the role of police to refer PWID to health and social services. Responses to measures were on a 3-point Likert scale, ranging from agree, neither agree nor disagree, and disagree. Statistical analyses

Due to the ordinal nature of the outcome, interpretation of odds ratios should be treated as the higher ordered categories versus the lower ordered categories (e.g. “always” versus the three lower ordered categories [“sometimes”, “rarely”, “never”] or “always/sometimes” versus the two other lower ordered categories, etc.). The Score Test was used to verify the proportional odds assumption and ensure that there were no violations in the final models. The variance inflation factor was used to check for multicollinearity. All multivariable ordinal logistic models adjusted for age, and current district of service. Due to a high proportion of missing responses for the number of years of policing experience (26%), we used age as a proxy in multivariable models. We considered covariates if they were associated with the outcome (p<0.10). The Akaike Information Criterion (AIC) was used to compare and inform the final model selection.

Results

Overall, 1,806 officers received the training, however given its focus, our subsample consisted of 1,319 patrol officers who came into contact with syringes. Sociodemographics, recent policing behaviors, legal knowledge on syringe and heroin possession, and attitudes towards harm reduction services are presented in Table 1.

Table 1.

Baseline characteristics, behaviors, knowledge, and attitudes among those who came into contact with syringes (N=1,319)

Characteristic	N (%) or Median (IQR)
Officer Demographics
Age	38 (32–43)
Male	1150 (87.4)
At least high school education	876 (79.9)
Number of years working as police	11.4 (8.6 – 18.3)
Currently stationed in Zona Centro (downtown district)	204 (15.6)
Policing Behaviors (past 6 months)
Ever been stuck with a needle	171 (13.0)
Frequency of needle or syringe confiscation past 6 months
Always	151 (11.5)
Sometimes	492 (37.4)
Rarely	376 (28.6)
Never	296 (22.5)
Frequency of arrest for heroin possession past 6 months
Always	74 (5.6)
Sometimes	497 (37.8)
Rarely	340 (25.8)
Never	405 (30.8)
Frequency of referring PWID to social or health programs past 6 months
Always	131 (10.0)
Sometimes	360 (27.4)
Rarely	284 (21.6)
Never	540 (41.1)
Legal Knowledge
Syringe possession
Incorrect none/don’t know	485 (36.8)
Partial conceptual knowledge	139 (10.5)
Correct technical knowledge	695 (52.7)
Heroin possession
Incorrect none/don’t know	1174 (89.0)
Partial conceptual knowledge	40 (3.0)
Correct technical knowledge	105 (8.0)
Attitudes on PWID Health and Services
Methadone maintenance programs help reduce criminal activity (agree)	393 (30.2)
Syringe exchange programs increase risk of NSI among police (disagree)	274 (20.9)
Laws treating addiction as public health problem make policing easier (agree)	796 (60.5)
People addicted to drugs do not care about their health (disagree)	107 (8.1)
Drug addiction is a disease (agree)	1110 (85.5)
It is the role of police to refer PWID to health and social services (agree)	739 (56.2)

The median age of the sample was 38 (IQR: 32–43); they were mostly male (87%), and most (80%) had at least a high school education. In the last six months, nearly half (49%) reported (always/sometimes) confiscating syringes and 43% reported (always/sometimes) arresting someone for heroin possession. Correct legal knowledge at baseline was low, with 37% of officers believing it was completely illegal to carry syringes and 71% believing that no amount of heroin had been decriminalized. Knowledge on medication-assisted therapy (i.e. methadone) was similarly low, as only 30% of participants agreed that methadone access helps reduce community-level crime. Only 8% of officers disagreed that PWID do not care about their health. While 56% agreed that it was their role to refer PWID to health and social services, only 37.4% reported doing so (always/sometimes) within the last six months. Correlates of syringe confiscation

Older age (OR: 0.98, 95% CI: 0.97 – 0.99) and female sex (OR: 0.51, 95% CI: 0.38 – 0.68) were significantly and negatively associated with the odds of reporting a higher frequency of self-reported syringe confiscation (Table 2). Conversely, officers who had been stationed in Zona Centro (downtown district) had 3.50 (95% CI: 2.63 – 4.65) higher odds, of reporting syringe confiscation. Report of ever experiencing a NSI was also associated with higher levels (OR=1.36; 95% CI: 1.01 – 1.83) of syringe confiscation. Adjusted odds ratios did not change appreciably after including age, sex, district, history of NSI, and attitudes on addiction in a multivariable model.

Table 2.

Correlates of Syringe Confiscation in past 6 months

	Unadjusted OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Officer Demographics
Age	0.98 (0.97 – 0.99)	0.004	0.98 (0.97 – 0.99)	0.004
Male	Ref		Ref
Female	0.51 (0.38 – 0.68)	<.001	0.50 (0.37 – 0.68)	<.001
At least high school education	Ref
Did not complete high school education	0.82 (0.62 – 1.07)	0.139
Number of years working as police	0.98 (0.97 – 0.99)	0.004
Not stationed in Zona Centro district	Ref		Ref
Stationed in Zona Centro District	3.50 (2.63 – 4.65)	<.001	3.60 (2.69 – 4.81)	<.001
Ever experienced needlestick injury
No	Ref		Ref
Yes	1.36 (1.01 – 1.83)	0.041	1.38 (1.02 – 1.87)	0.036
Legal Knowledge
Syringe possession
Incorrect none/don’t know	Ref
Partial conceptual knowledge	1.47 (1.04 – 2.07)	0.029
Correct technical knowledge	1.04 (0.84 – 1.28)	0.721
Heroin possession
Incorrect none/don’t know	Ref
Partial conceptual knowledge	1.36 (0.76 – 2.41)	0.297
Correct technical knowledge	1.03 (0.71 – 1.47)	0.896
Attitudes on PWID Health and Services
Methadone maintenance programs help reduce criminal activity
Disagree	Ref
Neither agree nor disagree	0.91 (0.71 – 1.15)	0.414
Agree	0.86 (0.67 – 1.11)	0.243
Syringe exchange programs increase the risk of NSI among police
Agree	Ref
Neither agree nor disagree	0.78 (0.60 – 1.02)	0.066
Disagree	0.86 (0.67 – 1.10)	0.233
Laws treating addiction as a public health problem make policing easier
Disagree	Ref
Neither agree nor disagree	1.04 (0.76 – 1.42)	0.807
Agree	0.85 (0.66 – 1.11)	0.232
People addicted to drugs do not care about their health
Agree	Ref
Neither agree nor disagree	1.18 (0.87 – 1.60)	0.297
Disagree	0.72 (0.51 – 1.04)	0.079
Drug addiction is a disease
Disagree	Ref		Ref
Neither agree nor disagree	0.42 (0.24 – 1.27)	0.002	0.36 (0.20 – 0.64)	<.001
Agree	0.91 (0.65 – 1.27)	0.566	0.86 (0.61 – 1.22)	0.401
It is the role of police to refer PWID to health and social services
Disagree	Ref
Neither agree nor disagree	1.31 (0.97 – 1.76)	0.487
Agree	1.09 (0.85 – 1.40)	0.077

Arrest for heroin possession

Similar to syringe confiscation, we found strong and significant negative associations between age and female sex and arrest for heroin possession in multivariable analysis (Table 3). Additionally, officers stationed in the downtown precinct Zona Centro district had three times higher odds (aOR = 3.06, 95% CI: 2.28 – 4.11) of arresting someone for heroin possession compared to officers not stationed in Zona Centro. The best fitting multivariable model also included several attitudinal measures. Officers who agreed with the statement “Laws that treat drug addiction as primarily a public health problem make the job of police easier” had 0.78 times (95% CI: 0.59 – 1.03) the odds of making arrests for heroin possession, as compared to their counterparts. Conversely, officers who neither agreed nor disagreed that drug addiction is a disease had 0.52 times (95% CI: 0.29– 0.94) the odds of making arrests for heroin possession.

Table 3.

Correlates of arrest for heroin possession in past 6 months

	Unadjusted OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Officer Demographics
Age	0.97 (0.96 – 0.98)	<.001	0.96 (0.95 – 0.98)	<.001
Male	Ref		Ref
Female	0.53 (0.39 – 0.71)	<.001	0.46 (0.34 – 0.64)	<.001
At least high school education	Ref
Did not complete high school education	0.77 (0.59 – 1.01)	0.056
Number of years working as police	0.97 (0.95 – 0.98)	<.001
Not stationed in Zona Centro district	Ref		Ref
Stationed in Zona Centro District	3.07 (2.30 – 4.10)	<.001	3.06 (2.28 – 4.11)	<.001
Ever experienced needlestick injury
No	Ref
Yes	1.23 (0.92 – 1.65)	0.169
Legal Knowledge
Syringe possession
Incorrect none/don’t know	Ref
Partial conceptual knowledge	1.39 (0.98 – 1.96)	0.065
Correct technical knowledge	1.07 (0.87 – 1.33)	0.512
Heroin possession
Incorrect none/don’t know	Ref
Partial conceptual knowledge	1.39 (0.78 – 2.48)	0.267
Correct technical knowledge	1.21 (0.84 – 1.75)	0.299
Attitudes on PWID Health and Services
Methadone maintenance programs help reduce criminal activity
Disagree	Ref
Neither agree nor disagree	0.83 (0.63 – 1.05)	0.112
Agree	0.81 (0.65 – 1.05)	0.122
Syringe exchange programs increase the risk of NSI among police
Agree	Ref		Ref
Neither agree nor disagree	0.75 (0.57 – 0.97)	0.031	0.69 (0.52 – 0.91)	0.008
Disagree	0.93 (0.72 – 1.19)	0.548	0.83 (0.64 – 1.08)	0.163
Laws treating addiction as a public health problem make policing easier
Disagree	Ref		Ref
Neither agree nor disagree	0.92 (0.67 – 1.27)	0.620	0.95 (0.68 – 1.32)	0.744
Agree	0.73 (0.57 – 0.95)	0.021	0.78 (0.59 – 1.03)	0.078
People addicted to drugs do not care about their health
Agree	Ref
Neither agree nor disagree	1.03 (0.76 – 1.41)	0.829
Disagree	0.74 (0.52 – 1.07)	0.111
Drug addiction is a disease
Disagree	Ref		Ref
Neither agree nor disagree	0.58 (0.33 – 1.02)	0.058	0.52 (0.29 – 0.94)	0.029
Agree	1.06 (0.76 – 1.49)	0.726	1.03 (0.73 – 1.47)	0.852
It is the role of police to refer PWID to health and social services
Disagree	Ref
Neither agree nor disagree	1.25 (0.92 – 1.68)	0.150
Agree	1.00 (0.78 – 1.29)	0.998

Referral to health/social programs

Table 4 presents the unadjusted and adjusted odds ratios for variables found to be significantly associated with frequency of referral to health and social programs. Officers who had accurate technical knowledge on syringe possession laws had lower odds of referring PWID to health and social services (aOR: 0.79, 95% CI: 0.63 – 0.98). Additionally, we found a strong association between officers who agreed that it was their role to refer PWID to health and social services and actually referring PWID to health/social programs (aOR: 3.32, 95% CI: 2.52 – 4.37). Conversely, officers who disagreed that SEPs increase NSI risk among police officers had significantly lower odds of referring PWID to health and social programs (aOR: 0.76, 95% CI: 0.59 – 0.99).

Table 4.

Correlates of referring PWID to health/social programs in past 6 months

	Unadjusted OR (95% CI)	p-value	Adjusted OR (95% CI)	p-value
Officer Demographics
Age	1.02 (1.01 – 1.03)	<.001	1.02 (1.00 – 1.03)	0.013
Male	Ref
Female	0.73 (0.54 – 0.98)	0.039
At least high school education	Ref
Did not complete high school education	1.08 (0.82 – 1.41)	0.584
Number of years working as police	1.02 (1.00 – 1.03)	0.035
Not stationed in Zona Centro district	Ref		Ref
Stationed in Zona Centro District	1.29 (0.98 – 1.69)	0.067	1.37 (1.04 – 1.81)	0.028
Ever experienced needlestick injury
No	Ref		Ref
Yes	1.33 (1.00 – 1.79)	0.054	1.30 (0.96 – 1.76)	0.085
Legal Knowledge
Syringe possession
Incorrect none/don’t know	Ref		Ref
Partial conceptual knowledge	1.21 (0.86 – 1.70)	0.285	1.27 (0.89 – 1.80)	0.184
Correct technical knowledge	0.72 (0.58 – 0.89)	0.002	0.79 (0.63 – 0.98)	0.033
Heroin possession
Incorrect none/don’t know	Ref
Partial conceptual knowledge	1.38 (0.78 – 2.44)	0.268
Correct technical knowledge	1.12 (0.78 – 1.60)	0.558
Attitudes on PWID Health and Services
Methadone maintenance programs help reduce criminal activity
Disagree	Ref
Neither agree nor disagree	0.85 (0.67 – 1.08)	0.164
Agree	1.20 (0.93 – 1.54)	0.188
Syringe exchange programs increase the risk of NSI among police
Agree	Ref		Ref
Neither agree nor disagree	0.83 (0.64 – 1.09)	0.175	0.92 (0.70 – 1.21)	0.551
Disagree	0.71 (0.55 – 0.92)	0.008	0.76 (0.59 – 0.99)	0.043
Laws treating addiction as a public health problem make policing easier
Disagree	Ref
Neither agree nor disagree	1.15 (0.84 – 1.83)	0.377
Agree	1.41 (1.08 – 1.83)	0.011
People addicted to drugs do not care about their health
Agree	Ref
Neither agree nor disagree	1.15 (0.85 – 1.56)	0.371
Disagree	0.92 (0.62 – 1.32)	0.641
Drug addiction is a disease
Disagree	Ref
Neither agree nor disagree	0.65 (0.37 – 1.15)	0.139
Agree	1.22 (0.87 – 1.72)	0.249
It is the role of police to refer PWID to health and social services
Disagree	Ref		Ref
Neither agree nor disagree	1.66 (1.22 – 2.28)	0.002	1.68 (1.22 – 2.32)	0.002
Agree	3.39 (2.59 – 4.44)	<.001	3.32 (2.52 – 4.37)	<0.001

Discussion

Our findings show low levels of knowledge regarding legal and public health issues among one of Mexico’s largest police forces. Indeed, 37% of officers incorrectly believed that carrying syringes was illegal—despite Mexican criminal law never censuring syringe possession. Similarly, less than 30% were aware that heroin had been decriminalized six years prior to the study. This echoes other research on the gaps in police knowledge of laws related to drugs and syringes (Banta-Green et al., 2013; Beletsky et al., 2005; Beletsky et al., 2012b). In view of clear shortcomings in dissemination of policy information to those charged with its enforcement, trainings and other communication modalities should marshal special attention and resources.

Nevertheless, contrary to our expectations, measures of legal knowledge assessed in this study were not associated with practices that correspond to these provisions. We found that correct knowledge of syringe possession policy was negatively associated with referral of PWID to services. However, knowledge on the syringe law did not appear protective against syringe confiscation and knowledge of heroin decriminalization did not appear protective against arrest for heroin possession. It is possible that police culture that encourages extrajudicial practices supersedes formal policy, irrespective of whether the officer was aware of the law—a finding that is bolstered by qualitative research from other settings (Beletsky et al., 2005). Additionally, based on unpublished in-depth interviews, officers in Tijuana reported confusion about the jurisdictional relevance of the state laws governing drug and syringe possession. Indeed, municipal ordinances give officers broad authority to detain anyone they see as “encouraging vice” or consuming drugs or alcohol in public (Secretario de Gobierno Municipal del XVII Ayuntamiento de Tijuana, 2002). In this context, shifting incentives and performance metrics away from arrest and towards referral to care may be a strategy to consider, as legal knowledge may not be sufficient to align these policing practices with public health imperatives. We will determine if similar trends in arrest and syringe confiscation persist in our prospective cohort sample post-PEP.

Our findings have additional important implications. The significant positive relationship found between ever experiencing a NSI and syringe confiscation reinforces the need to mount training and other interventions that safeguard officer occupational safety. While we cannot establish temporality due to the cross-sectional nature of these data, if experiencing a NSI reinforces negative attitudes towards PWID and promotes syringe confiscation, then training officers on preventing NSI could serve a dual health benefit for both officers and PWID. We have reported similar findings in Baltimore, MD, where officers who had experienced a NSI and had received a PEP were less likely to change their attitudes towards better public health practice compared to officers who had never experienced a NSI (Beletsky et al., 2012a). This analysis affirms the rationale for conducting PEP that bundle occupational safety with public health content, since improved public safety among police can improve the HIV risk environment for PWID.

Several demographic correlates of behavioral outcomes are noteworthy. In our analyses, female respondents were significantly less likely to engage in both syringe confiscation and arrest for heroin possession. Such a pattern is consistent with previous research suggesting that women in policing are less likely to engage in use of force, weapons, and controlling behaviors, including searches and arrest (Bazley et al., 2007; Rabe-Hemp, 2008; Schuck and Rabe-Hemp, 2007). While department-wide police trainings should seek to increase public health education as a way to foster greater empathy with PWID, female officers may play a critical role in reducing detrimental extrajudicial practices that harm the health of PWID. Indeed, other global entities, including the U.S. Department of Justice, have recommended increasing the representation of female officers in an effort to improve community – law enforcement relations (Copple, 2017). We also found that older officers were significantly less likely to engage in syringe confiscation and arrest, but more likely to refer to health/social programs. In past research, older officers expressed disillusionment with criminal justice tools to combat drug-related harms (Beletsky et al., 2005). Our findings reaffirm the possibility that, having understood the instrumental futility of arrest and incarceration, older police officers may be more open to alternative approaches.

In all our models, we found that officers currently stationed in Zona Centro were more likely to have confiscated syringes, arrested someone for heroin possession, and referred PWID to health/social programs. This locale, which includes Tijuana’s “red light district”, is a concentrated hotspot of drug activity (Gaines et al., 2015; Gaines et al., 2017). Because of this concentration, it may also present higher occupational risk for officers, including HIV risk (Kori et al., 2014). Indeed, this is a substantial occupational risk for police in Tijuana as 15% reported a lifetime history of NSI, several-fold higher than among officers in the United States (Mittal et al., 2016). However many harm reduction programs, including SEPs are also sited in this district. Thus, future interventions on improving officer safety and facilitating officer referral to evidence-based addiction treatment and harm reduction services should be prioritized in this locale.

Reframing police as agents of public health outreach can function as a key component in drug policy reform efforts. The 2009 drug policy reforms that decriminalized small amounts of drugs for personal consumption in Mexico, recast drug use as a health problem and stipulated that police must refer PWID to drug treatment upon their third infraction. The reforms in Mexico were loosely modeled on the Portuguese framework, whereby minor drug possession was decriminalized in 2001 and the response strongly focused on referral to drug treatment rather than criminal punishment (Hughes and Stevens, 2010). However, poor drafting, insufficient operational planning, and flawed implementation of the reforms in Mexico underscore the need for additional interventions, such as PEPs to buttress these reforms. Furthermore, police continue to play an integral role in settings without decriminalization laws but with robust harm reduction infrastructure, such as safe injecting facilities. For example, police in Vancouver have supported the city’s safe injecting facility, as evidenced by greater police referral of public injecting PWID to this site, in part, because of its benefit to the public order (DeBeck et al., 2008). Previous studies have documented shifts in police culture and attitudes towards health of PWID, particularly as a result of the evolving opioid overdose crisis (Green et al., 2013; Wagner et al., 2016). In our study, we found that officers who believed that it was their role to refer PWID to health and social services were significantly more likely to report doing so, suggesting a high degree of self-efficacy to engage in this behavior. While the construct of “health and social programs” is quite broad, we chose this expansive framing because Tijuana suffers from a paucity of services tailored to address the needs of PWID (Bazzi et al., 2016; Werb et al., 2015). Improving referral mechanisms will depend on the ability to increase the scale of high quality, low threshold evidence-based services in the city.

Some associations between attitudinal variables and outcomes in our multivariable models were contrary to our expectations. For instance, officer attitudinal ambivalence—rather than positive attitudes—was a significant correlate of self-reported behavior. Specifically, those who felt ambivalent about whether SEPs increase their risk of NSI were significantly less likely to report recent heroin arrest. Similarly, respondents who felt ambivalent about the characterization of drug addiction as a disease were less likely to confiscate syringes. It is possible that officers who do not report strong views on issues related to substance use and harm reduction are also more ambivalent towards elements of police culture that may drive harsh enforcement practices. This unanticipated relationship warrants further exploration in follow-up and ethnographic research.

Limitations

These data were based on self-report; thus, social desirability and recall bias may have led to an underreporting of extrajudicial behaviors or over-reporting of referral to health/social programs. However, surveys were self-administered and may have reduced the influence of social desirability since no interviewer was present. Additionally, behavioral data were collected prior to the PEP, so officers who were not aware of the legality of their behaviors would still most likely have answered truthfully. While our original intention was to capture the behavioral outcomes more quantitatively, we received feedback from senior officers during the piloting of the survey to utilize an ordinal scale. Thus, there may be some imprecision and subjectivity with officers’ response to these behaviors. Since these baseline data were cross-sectional, we cannot determine the temporality of officers’ knowledge and attitudes on these practices. Lastly, we limited this to only patrol officers who reported some contact with syringes, however there may have been officers who underreported incidents with syringes. These findings may also not be generalizable to officers with other primary duties or officers who reported no contact with syringes due to personal preferences.

Conclusions

Modifying officer occupational risks and attitudes towards harm reduction interventions is an important element in efforts to align police practices with public health objectives. Our study highlights the need to improve officers’ public health knowledge and attitudes since this may reduce the frequency of arrest and syringe confiscation and promote referral to health and social programs.