Knowledge of Good Samaritan Laws and Beliefs About Arrests Among Persons Who Inject Drugs a Year After Policy Change in Baltimore, Maryland

Abstract

Objectives

Delivering and receiving prompt medical care during an overdose are imperative to ensure survival. Good Samaritan laws encourage people to call 911 during an overdose by providing immunity from selected drug arrests (eg, low-level possession). However, it is unclear whether persons who inject drugs (PWID) are aware of and understand these laws and their implications. We examined awareness among PWID of the 2015 Good Samaritan law in Maryland and their beliefs about whether they could be arrested for calling 911 or having an overdose.

Methods

We surveyed 298 PWID in Baltimore, Maryland. We estimated the proportion who knew what the Good Samaritan law addressed and who believed they could be arrested for calling 911 or overdosing. We used a multivariate model to assess the association between harm-reduction services and knowledge of the Good Samaritan law or beliefs about getting arrested for calling 911 or overdosing.

Results

Of PWID, 56 of 298 (18.8%) knew what the Good Samaritan law addressed, 43 of 267 (16.1%) believed they could be arrested for calling 911, and 32 of 272 (11.8%) believed they could be arrested for having an overdose. After adjusting for demographic characteristics, accessing the syringe services program was associated with accurate knowledge and the belief that PWID could be arrested for calling 911; however, training in overdose reversal was not associated.

Conclusions

Most PWID were unaware of the Good Samaritan law; this lack of awareness is a barrier to preventing overdose deaths. Educating PWID about Good Samaritan laws is essential, and such education should include police to ensure that law enforcement is congruent with Good Samaritan laws and does not perpetuate mistrust between police and PWID.

The opioid crisis has led to a plethora of public health problems in the United States. Rates of nonmedical prescription opioid, heroin, and injection drug use; substance use disorders; drug-involved injuries; and opioid-involved overdoses have all risen substantially.^1-5 In particular, opioid-involved overdose rates have skyrocketed in recent years, increasing from 8048 deaths in 1999 to 47 600 deaths in 2017.^6-9 In 2017, approximately 70 700 deaths were attributed to opioid-involved overdoses in the United States.¹⁰ Nonfatal overdoses treated in hospitals nationally exceeded 140 000 from July 2016 through September 2017.¹ Although this estimate indicates a high burden of nonfatal overdoses, it does not capture a large portion of nonfatal overdoses among persons who do not go to a hospital or call emergency services. In many emergency overdose situations, fears of arrest and other legal consequences (eg, the loss of child custody or housing) often cause bystanders or persons having an overdose to delay or avoid calling 911.^11-13 Barriers to receiving medical care during an overdose are important public health challenges, because swift and scaled-up responses to overdoses are needed to save lives.

When addressing a complex issue such as overdose, multipronged and multilevel responses are essential. Trainings in overdose reversal and naloxone distribution have been broadly implemented and have saved more than 26 000 lives since their inception in 1996.^14,15 Some jurisdictions have introduced drug testing strips so that persons who use drugs can determine whether their drugs have been adulterated with fentanyl and then make behavioral changes to reduce their risk for overdose, such as using drugs more slowly.^16-20 However, these behavior-focused solutions alone are not sufficient, because they do not reach all persons who are at risk or vulnerable to overdoses as a result of structural barriers, such as homelessness and poverty. Overdose prevention programs must be nested within a public health–oriented legal policy framework, because law is often the primary barrier to program access.^21,22

Good Samaritan laws are an important harm-reduction tool in the context of the ongoing opioid crisis. Good Samaritan laws are laws that grant immunity for low-level drug crimes to persons having an overdose and bystanders who call 911 during overdose emergencies.²³ These laws are intended to encourage persons having an overdose and bystanders to seek medical help during an overdose, by allaying fears of legal consequences after calling 911. Forty states and the District of Columbia have Good Samaritan laws that pertain to drug overdoses.²³ The scope of these laws varies by state. Most states offer some form of immunity from arrests, charges, or prosecution for crimes related to possession of drugs or drug paraphernalia, whereas few states offer additional protections against parole or probation violations, restraining order violations, other drug crimes, or civil asset forfeitures.²⁴ In Maryland as of 2015, Good Samaritan laws provided immunity for 6 misdemeanors if evidence was obtained solely from the period during which the person having the overdose or bystanders witnessing the overdose were seeking medical services: possession or administration of controlled dangerous substances, drug paraphernalia, controlled paraphernalia, underage alcohol possession, obtaining alcohol for underage consumption, and furnishing alcohol or underage consumption.²⁵ Although the passage of these laws is a step toward public health–oriented legal policies, the mere existence of these laws is insufficient to affect public health if persons most likely to have or witness an overdose are not aware of these laws.

Education of persons who use drugs and persons who enforce laws on the potential effect of Good Samaritan laws is essential for these laws to save lives. If persons who use drugs are unaware of Good Samaritan laws, they cannot modify their help-seeking behaviors in response to the protections these laws provide. Similarly, if police are not fully informed of the new amnesties provided by these laws, they may still arrest persons at the scene of an overdose, undermining any positive effects of the laws. Educating relevant parties (eg, police, persons who use drugs, likely bystanders) about Good Samaritan laws is essential for the enforcement of these laws to be in good faith to encourage persons to call for medical assistance. In many locales, harm-reduction programs, such as syringe service programs (SSPs) and overdose-reversal programs, also provide health education. Therefore, one might expect that persons who use drugs and access these services would have better knowledge of laws, such as Good Samaritan laws, that affect their rights and health than their counterparts who do not access harm-reduction services.

To date, few studies have assessed what persons who use drugs know about Good Samaritan laws. One qualitative study in Pennsylvania found that none of the 18 substance use treatment attendees who were interviewed and who had experienced overdoses had heard of their state’s Good Samaritan laws.²⁶ Another qualitative study in Baltimore found that 14 of 22 persons who inject drugs (PWID) who were interviewed were unaware of Good Samaritan laws, and many did not believe that these laws would deter police from arresting them.¹² In a study of 198 young adults in Rhode Island who used nonmedical prescription opioids, 90 (45.5%) were aware of the state’s Good Samaritan laws.²⁷ This sample primarily comprised non-Hispanic white men who did not inject drugs; as such, this finding may not be generalizable to communities of color or other marginalized groups of PWID. Although these studies constitute an important first step in understanding how knowledge of Good Samaritan laws has penetrated communities, additional, larger-scale research on marginalized and high-risk populations for overdose is needed.

The objectives of this study were to examine (1) whether PWID in Baltimore understood the state’s 2015 Good Samaritan laws approximately 1 year after they were implemented and (2) whether accessing overdose-reversal trainings or the Baltimore SSP improved knowledge of the local laws.

Methods

Study Design and Recruitment

Our study was a secondary analysis of data collected to evaluate syringe coverage (ie, the percentage of needed syringes that are provided by a program) from the Baltimore SSP among PWID in the city. The cross-sectional survey from the parent study was conducted from April through November 2016 in Baltimore.^28-30 All survey participants were PWID. As of 2016, the Baltimore SSP operated 16 fixed locations throughout the city and provided HIV prevention and other harm-reduction services to its clients. Participants were sampled from all 16 sites, in proportion to the site’s client volume. Study staff members recruited participants after they exited the SSP van and screened them for eligibility. The study team recruited nonclient peers of SSP clients through referrals from previous participants. Client participants were given 3 referral coupons to distribute to their peers and received $5 for each nonclient they referred who enrolled in the study. Eligibility criteria included being aged ≥18, having previously injected drugs, and providing informed consent in English. Participants completed a 30-minute computer-assisted personal interview in English and received a $25 gift card. Our sample included 298 PWID (203 SSP clients and 95 nonclients). The Johns Hopkins Bloomberg School of Public Health Institutional Review Board approved this study.

Measures

Knowledge of Good Samaritan laws and arrest beliefs

We prefaced questions about the Good Samaritan law with the statement, “In 2015 there was a change in the Good Samaritan law, which has to do with overdose and the police, and we want to know what people know about the law.” Participants were first asked, “If you know about the Good Samaritan law, can you tell me what it deals with?” Nonexclusive response categories were: (1) protecting from arrest for low-level drugs, (2) changes in arrest for drug paraphernalia, (3) protecting from arrest in the event of an overdose, and (4) don’t know. We generated a variable for accurate knowledge, in which identifying the law’s protection from arrest in the event of an overdose was coded as correct and incorrect response options were coded as incorrect. We then asked participants the following yes/no questions: “If somebody overdoses, can you be arrested just because you called 911?” and “If you overdose and get medical help, can you be arrested just because you overdosed?” Refusal to answer or “don’t know” responses were recoded as missing for the arrest belief variables. According to the laws in Maryland, persons should not be arrested for either event, so beliefs that a person could be arrested for calling 911 or having an overdose would be in contrast to the official policy.

Use of harm-reduction services

In Baltimore, overdose-reversal trainings can be accessed through the local health department and other harm-reduction organizations, and naloxone is broadly available under the jurisdiction-wide blanket prescription issued by the city health commissioner. We assessed whether participants accessed the SSP and whether they had received training in overdose reversal. For SSP usage, we identified participants as current clients or nonclients of the Baltimore City SSP at recruitment. We recruited clients directly after their visit to the SSP, and we asked nonclient referrals whether they were SSP clients. For overdose-reversal training, we asked participants whether they had participated in a training where they learned about overdose reversal and received naloxone as part of that training in the previous 6 months (yes/no).

Demographic characteristics

Participants reported their age (18-34, 35-44, 45-54, ≥55), sex (male, female), race/ethnicity (non-Hispanic white, non-Hispanic black, other), and education (<high school degree, high school degree or equivalent, ≥some college).

Analysis

First, we calculated the prevalence of accurate knowledge about the Good Samaritan laws and arrest beliefs. Next, we assessed associations between knowledge and beliefs by using tetrachoric correlations (ie, a correlation between 2 binary variables). Then, we tested for differences in knowledge and beliefs by demographic characteristics and use of harm-reduction services by using Pearson χ² tests. We also estimated bivariate logistic regression models for accurate knowledge on each demographic characteristic and service use variable. Finally, we estimated a multivariable logistic regression model, regressing accurate knowledge of the law on demographic characteristics and harm-reduction service usage, including both accessing the SSP and attending overdose-reversal training. We conducted all analyses using Stata release 14.³¹ We considered P < .05 to be significant.

Results

Of 298 participants, 205 (68.8%) were male and 172 (57.7%) were non-Hispanic black (Table 1). Most participants had a high school degree or equivalent (n = 130, 43.6%) or <high school degree (n = 113, 37.9%). Seventy-one (23.8%) participants were aged 18-34, 64 (21.5%) were aged 35-44, 108 (36.2%) were aged 45-54, and 55 (18.5%) were aged ≥55. Most participants were SSP clients (n = 203, 68.1%) and had received overdose-reversal training in the past 6 months (143 of 297, 48.1%). Only 56 (18.8%) participants had accurate knowledge of Maryland’s Good Samaritan laws. Of the 267 participants who answered the question about being arrested for calling 911, 43 (16.1%) believed that a person could be arrested; of the 272 participants who answered the question about being arrested for having an overdose, 32 (11.8%) believed that a person could be arrested for having an overdose.

Table 1.

Knowledge about the 2015 Good Samaritan law in Maryland and beliefs about arrest among persons who inject drugs, by characteristics, use of syringe services programs, and receipt of overdose-reversal training, Baltimore, Maryland, 2016

Characteristics	Total	Accurate Knowledge of the Good Samaritan Law		Believed That a Person Could Be Arrested for Calling 911a		Believed That a Person Could Be Arrested for Having an Overdoseb
	No. (%)	No. (%)c	χ2 (P Value)d	No. (%)e	χ2 (P Value)d	No. (%)e	χ2 (P Value)d
Full sample	298 (100.0)	56 (18.8)	—	43 (16.1)	—	32 (11.8)	—
Accurate knowledge
Yes	56 (18.8)	—	—	6 (11.1)	1.3 (.26)	2 (3.9)	3.7 (.05)
No	242 (81.2)	—		37 (17.4)		30 (13.6)
Sex
Male	205 (68.8)	33 (16.1)	3.1 (.08)	30 (16.4)	0 (.85)	24 (12.7)	0.5 (.47)
Female	93 (31.2)	23 (24.7)		13 (15.5)		8 (9.6)
Race/ethnicity
Non-Hispanic white	111 (37.2)	29 (26.1)	6.9 (.03)	10 (10.1)	3.4 (.11)	10 (9.9)	0.6 (.75)
Non-Hispanic black	172 (57.7)	26 (15.1)		31 (20.0)		20 (12.7)
Otherf	15 (5.0)	1 (6.7)		2 (15.4)		2 (14.3)
Education
<High school degree	113 (37.9)	19 (16.8)	4.7 (.10)	14 (13.7)	12.0 (.37)	8 (7.9)	2.6 (.27)
High school degree or equivalent	130 (43.6)	21 (16.2)		23 (19.7)		16 (13.1)
≥Some college	55 (18.5)	16 (29.1)		6 (12.5)		8 (16.3)
Age, y
18-34	71 (23.8)	17 (23.9)	3.3 (.35)	6 (9.4)	3.2 (.37)	6 (9.5)	0.4 (.93)
35-44	64 (21.5)	14 (21.9)		9 (16.1)		7 (12.1)
45-54	108 (36.2)	18 (16.7)		19 (19.6)		12 (12.2)
≥55	55 (18.5)	7 (12.7)		9 (18.0)		7 (13.2)
Is an SSP client
Yes	203 (68.1)	47 (23.2)	7.9 (.01)	36 (19.6)	5.3 (.02)	20 (10.8)	0.6 (.45)
No	95 (31.9)	9 (9.5)		7 (8.4)		12 (14.0)
Had training in overdose reversalg
Yes	143 (48.1)	35 (24.5)	5.7 (.02)	21 (16.2)	0 (.87)	12 (9.2)	1.7 (.20)
No	154 (51.9)	21 (13.6)		21 (15.4)		20 (14.2)

Abbreviations: —, does not apply; SSP, syringe services program.

^aBelieved that someone could be arrested for calling 911 in the case of an overdose; 31 participants were missing information on this question.

^bBelieved that someone could be arrested for having an overdose; 26 participants were missing information on this question.

^cRow percentages.

^dUsing the Pearson χ² test of significance, with P < .05 considered significant.

^ePercentages are row percentages, after excluding persons with missing data.

^fOther includes all other racial categories, including but not limited to Hispanic, Asian/Pacific Islander, and American Indian/Alaska Native.

^gOne participant was missing information on this question.

We found no significant correlations between knowledge about the Good Samaritan laws and arrest beliefs (Table 1). Race/ethnicity and education were significantly associated with knowledge of Good Samaritan laws but not with arrest beliefs. More non-Hispanic white participants (n = 29) than non-Hispanic black participants (n = 26) or participants of other races/ethnicities (n = 1) had accurate knowledge of the laws (χ² = 6.9; P = .03; Cramer V = 0.15).

Being a client of the SSP was associated with accurate knowledge of Good Samaritan laws (χ¹ = 7.9; P = .01; Cramer V = 0.16) and a belief that someone could be arrested for calling 911 (χ¹ = 5.3; P = .02; Cramer V = 0.14) (Table 1). Being an SSP client was not associated with a belief that someone could be arrested for having an overdose. Having attended overdose-reversal training in the past 6 months was significantly associated with knowledge of Good Samaritan laws (χ² = 5.7, P = .02; Cramer V = 0.14) but not with either measure of arrest beliefs.

In the bivariate models, non-Hispanic black race/ethnicity (vs non-Hispanic white race/ethnicity) was associated with inaccurate knowledge of Good Samaritan laws (odds ratio [OR] = 0.50; 95% confidence interval [CI], 0.28-0.91) (Table 2). Being an SSP client (OR = 2.88; 95% CI, 1.35-6.16) and overdose-reversal training (OR = 2.05; 95% CI, 1.13-3.73) were associated with accurate knowledge of Good Samaritan laws. In the multivariable model, being an SSP client (adjusted odds ratio [aOR] = 2.24; 95% CI, 0.99-5.03) and previous overdose-reversal training (aOR = 1.82; 95% CI, 0.95-3.49) were associated with accurate knowledge of Good Samaritan laws, although these relationships were not significant.

Table 2.

Bivariate and multivariate logistic regression results for correlates on accurate knowledge of Good Samaritan laws among persons who inject drugs (n = 298), by characteristics, use of syringe services programs, and receipt of overdose-reversal training, Baltimore, Maryland, 2016

Characteristics	Bivariate Models		Multivariate Model
	OR (95% CI)	P Valuea	aOR (95% CI)	P Valuea
Age, y
18-34	1.00 [Reference]		1.00 [Reference]
35-44	0.89 (0.40-1.99)	.78	1.03 (0.39-2.69)	.96
45-54	0.64 (0.30-1.34)	.23	1.56 (0.56-4.33)	.39
≥55	0.46 (0.18-1.21)	.12	0.85 (0.23-3.10)	.81
Sex
Male	1.00 [Reference]		1.00 [Reference]
Female	1.71 (0.94-3.12)	.08	1.27 (0.62-2.59)	.51
Race/ethnicity
Non-Hispanic white	1.00 [Reference]		1.00 [Reference]
Non-Hispanic black	0.50 (0.28-0.91)	.02	0.48 (0.21-1.10)	.08
Otherb	0.20 (0.03-1.60)	.13	0.24 (0.03-2.10)	.20
Education
<High school degree	1.00 [Reference]		1.00 [Reference]
High school degree or equivalent	0.95 (0.48-1.88)	.89	0.83 (0.38-1.79)	.63
≥Some college	2.02 (0.95-4.35)	.07	2.53 (0.99-6.51)	.05
Is an SSP client	2.88 (1.35-6.16)	.01	3.70 (1.39-9.80)	.01
Had training in overdose reversal	1.78 (0.90-3.54)	.10	1.32 (0.61-2.85)	.48

Abbreviations: aOR, adjusted odds ratio; OR, odds ratio; SSP, syringe services program.

^aUsing the Pearson χ ² test of significance, with P < .05 considered significant.

^bOther includes all other racial categories, including but not limited to Hispanic, Asian/Pacific Islander, and American Indian/Alaska Native.

Discussion

We found low levels of knowledge of Good Samaritan laws among PWID in Baltimore 1 year after the laws’ implementation. At the same time, most participants did not believe that a person could be arrested for calling 911 or for having an overdose. Generally fearing any police involvement has been identified as a major barrier to calling for help immediately during an emergency.^11-13,32 It is imperative that all persons at risk for an overdose know that they cannot be arrested for calling 911 or having an overdose, consistent with local Good Samaritan laws. It is the responsibility of public health practitioners to ensure that PWID are fully educated about their rights during a health emergency.

Knowledge of Good Samaritan laws among PWID should be contextualized within the general population’s understanding. A 2018 study suggests that as many as 78% of lay responders who obtained naloxone from their local health department were aware of their state’s Good Samaritan law.³³ A low level of awareness of changes in Good Samaritan laws exists among police and paramedics, leading to confiscation of drugs or paraphernalia at one-quarter of overdose scenes where police responded in Seattle, Washington.³⁴ The difference between awareness and understanding of Good Samaritan laws is meaningful. Although we did not measure awareness of Good Samaritan laws in our study, we can conclude that most PWID did not understand what the Maryland Good Samaritan law entails even if they had previously heard of it.

In our study, accessing the SSP was associated with accurate knowledge of the Good Samaritan laws, but overdose-reversal training was not. Although accessing the SSP was associated with accurate knowledge of Good Samaritan laws, it was paradoxically associated with the belief that one could be arrested for calling 911. One possible explanation for this finding is that SSP clients may be more likely than nonclients to have negative experiences with the police in general, thus having more negative expectations about legal outcomes. Police harassment of SSP clients is common, despite SSPs being legal.^35-38 Experiences of police harassment and systematic abuse have left lingering fears about being arrested, regardless of legal protections.¹² These fears act as a barrier against the effectiveness of Good Samaritan laws. Our findings and the findings of others suggest that working with police is an important part of maximizing the benefits of Good Samaritan laws.

For Good Samaritan laws to be effective, members of law enforcement and the criminal justice system must act in good faith. Imperfect implementation of Good Samaritan laws, whether or not intentional, may exacerbate the structural risk (eg, homelessness, poverty) the punitive policy environment creates for PWID and undermine effectiveness of the laws if PWID believe they are likely to be arrested if they call 911.²¹ Such flawed implementation would worsen PWID’s legitimate fears of police interactions because of their regular exposure to violence by police, stigmatizing interactions with police, and degradation of their rights.^35,39,40 Concerns about police implementation may also make harm reductionists hesitant to engage in educational campaigns to preserve their clients’ trust, even though the laws are officially enacted. Harm reductionists should partner with police when trying to increase accurate knowledge of Good Samaritan laws among PWID, to start to repair PWID’s mistrust of police.²¹ Community–police partnerships that seek to educate both police and citizens of policies, so that all parties are aware of the laws and their rights, are essential for effective implementation of Good Samaritan laws.

Whether persons understand a Good Samaritan law has important implications for their behavior during an overdose emergency, especially their willingness to call for help. An early evaluation of Washington State’s Good Samaritan law found that 88% of persons who used opioids in the state would be willing to call 911 in an overdose emergency after being educated about the Good Samaritan law, indicating that improving knowledge of the law can change a person’s willingness to call 911.⁴¹ More harm-reduction services and other services for PWID need to incorporate education on Good Samaritan laws. Overdose-reversal trainings can potentially incorporate this education. In our study, recent participation in overdose-reversal trainings was not significantly associated with knowledge of the Good Samaritan laws, suggesting that this topic may not have been addressed, at least not in-depth, during the training. In Baltimore, as in many cities, overdose-reversal training is readily available to the public and, therefore, is an ideal conduit for education on Good Samaritan laws.⁴² Trainings should actively educate participants about the laws and their rights to encourage the best emergency responses possible. Such education requires relatively few resources yet can have a large effect on a person’s behavior during overdose emergencies, thus saving lives.

Limitations

Our study had several limitations. First, this analysis used a convenience sample, so we cannot inherently assume that our findings are representative of all PWID. Second, our measure of knowledge was relatively simplistic; as such, future studies would benefit from including more complex measures that capture nuances in understanding. Still, our study provides new insight into PWID’s understanding of Good Samaritan laws. Additional studies are needed to fully explore how PWID understand Good Samaritan laws and how harm-reduction services affect this understanding. Future studies should also explore other factors in addition to knowledge, such as attitudes toward police, and shifts in police practices that would enable PWID to more readily feel comfortable calling for help in an overdose emergency. Understanding other ways to facilitate calling 911 in addition to better education on Good Samaritan laws will help achieve the goals of such laws and save lives.

Conclusions

Our study found that most PWID in Baltimore do not know about Good Samaritan laws, although most have accurate beliefs about the potential of being arrested. Harm-reduction services should incorporate education about Good Samaritan laws into their programs to encourage PWID to call for help during an overdose emergency. Improving education on this issue is essential for Good Samaritan laws to save lives. Pairing educational campaigns for PWID with proper police implementation of laws and genuine efforts to repair the trust between PWID and police can help maximize the benefits of Good Samaritan laws. Good Samaritan laws are a single, but important, aspect of a comprehensive public health response to the overdose crisis.

ORCID iD

Kristin E. Schneider https://orcid.org/0000-0001-5813-1327