“It Eats My Heart”: Identifying knowledge gaps in injection drug-related endocarditis among hospitalized patients

Abstract

Background:

Infective endocarditis (IE) rates have risen sharply, spurred by injection drug use (IDU). Public health interventions to prevent infectious complications of IDU have primarily focused on communicable diseases, such as HIV and viral hepatitis. We assessed patients' knowledge of IE in relation to IDU practice.

Methods:

From 2021 to 2022, we conducted in-person semi-structured interviews (SSIs) with 16 adults hospitalized with IDU-related IE at an academic medical center in North Carolina. SSIs explored participants’ knowledge and experience of IE and IDU practices. The SSIs were audio-recorded, transcribed verbatim, and thematically coded.

Results:

We identified four primary themes: limited knowledge of IE; nonspecific IE symptoms; injection behavior and infection prevention; and knowledge of other IDU-related infections, such as hepatitis C and HIV. Most patients reported little or no knowledge of IE before hospitalization despite previously having soft tissue infections, such as skin abscesses. Presentations were heterogeneous, with many patients not recognizing their symptoms as IE, and often delaying seeking care. In contrast to limited information about IE, all patients reported prior testing for HIV and hepatitis C. While many expressed the importance of not reusing injection equipment, many patients reported reusing injection equipment nonetheless, and several specifically reported that they take no dedicated precautions to prevent infections.

Conclusion:

Our findings suggest prior to hospitalizations, knowledge of IE, its symptoms, and prevention is limited among PWID who acquire IE. Our study indicates an urgent need to disseminate public health messages regarding preventing IDU-related invasive bacterial infections and providing access to sterile equipment to prevent bacterial infections.

Introduction

Cases of infective endocarditis (IE) have risen sharply among young persons who inject drugs (PWID).^1–3 IE is a life-threatening infection that requires four to six weeks of parenteral antimicrobials and, at times, heart valve replacement surgery.⁴ Further, the rate of readmission injection drug use (IDU)-related IE remains high,⁵ with frequent recurrent episodes of IE.^6,7

Despite the increase in the incidence of IE, the extent of patient knowledge and awareness of IE as a risk of injection drug use (IDU) is unclear. One study examining patients' knowledge of severe bacterial infections, such as those requiring surgery or prolonged hospital stays, found that patients have a sufficient understanding of viral infection transmission related to IDU but a limited understanding of bacterial infections.⁸ When examining other IDU-related infections, such as skin and soft tissue infections, patients reported robust knowledge regarding the risks related to IDU practice.^9,10 Interventions aimed at infectious complications of IDU have long focused on communicable infections rather than invasive bacterial ones, such as IE.^11–15 Studies report that such public health efforts have largely been successful, with between 65% and 70% of PWID are aware of their hepatitis C virus (HCV) status.^12,16 The proliferation of syringe service programs (SSPs) has resulted in increased HCV and HIV testing and prevention, as well as a concomitant increase in knowledge of IDU-related HIV and HCV risks among PWID.¹¹ IE and other noncommunicable invasive infections pose a severe and increasingly prevalent risk to PWID,³ but patients’ knowledge of IE and its risk factors remains unclear.

In this study, we interviewed PWID hospitalized with IE to assess their knowledge of and experiences with IE, as well as their drug use practices and infection prevention strategies. Our findings could inform public health efforts to educate about IE across the spectrum of care, from the community to clinical care.

Methods

Study Design and Participant Recruitment

An exploratory qualitative study of patients hospitalized with IDU-related IE occurred at two hospitals, the University of North Carolina Medical Center and a large academic referral center, which were both part of the same health system. Patients were identified as potential participants based on weekly reviews of patients hospitalized on infectious diseases teams. Patients undergoing treatment for a clinical diagnosis of IE, per the treating physician’s assessment, who reported a history of injecting drugs were approached during their hospitalization and given a flyer outlining eligibility criteria and study processes. Eligibility criteria included hospitalization for IE, IDU in the past three months, ≥18 years of age, and English-speaking. The interviewers screened patients who expressed interest in participating in the study to ensure that they met the eligibility criteria, consented, and enrolled in the study. The study was approved by the University of North Carolina institutional review board.

Data Collection and Analysis

Enrolled participants completed a one-time in-person semi-structured interview (SSI) during hospitalization. The SSIs were conducted from June 2021 to May 2022, lasted approximately 60–75 minutes each, and were audio-recorded. Patients completing SSIs were offered a $35 gift card for time and effort. The SSI guide was iteratively developed and refined with input from coauthors with expertise in infectious diseases, substance use epidemiology, medical anthropology, harm reduction, and qualitative research. Questions covered a broad range of topics, including knowledge of IE, drug use practices, and testing for viral diseases. For this analysis, we focused on answers pertaining to knowledge of IE.

We transcribed audio files verbatim, with assistance from NVivo transcription software. Interviewers manually reviewed transcripts for accuracy. Utilizing an inductive thematic analysis,¹⁷ researchers identified themes related to knowledge of endocarditis, drug use practices, and experiences with other IDU-related viral diseases. Two analysts independently read and re-read two SSI transcripts each to familiarize themselves with the data, identify themes, and create codes. Another analyst read and re-read both transcripts, generated a list of codes, consolidated all three code lists, and drafted a preliminary codebook in collaboration with the two analysts. While Braun & Clarke¹⁷ do not discuss using a codebook within their process, we found that the codebook was essential for team communication and collaborative coding, especially as much of this work was done during the COVID-19 pandemic. Two separate pairs of analysts independently coded two transcripts to test the robustness of the identified codes. After the first coding round, inter-rater reliability checks were performed, and the analysts met to discuss and resolve any discrepancies. We used 0.08, an almost perfect agreement, as a cut-off Cohen’s kappa.¹⁸

The analysts conducted two additional rounds of coding, did an IRR check, met to discuss additional discrepancies, and refined the codebook by adding emerging themes and rephrasing code names and definitions for more clarity before resolving the identified coding discrepancies. Subsequently, the three analysts independently coded the rest of the transcripts in real time until thematic saturation was confirmed.¹⁹ Utilizing a semantic approach, code outputs were downloaded into framework matrices in Excel, where we organized themes by grouping codes with similar surface meanings to allow for a rich, inductive description of the data.¹⁷ From there, we sorted themes into primary themes and sub-themes and discussed over a series of meetings to refine and finalize themes presented alongside illustrative quotes. Thematic summaries were reviewed after the enrollment of 16 participants to confirm that, among later participants, themes reached saturation.

Demographic information was self-reported, and clinical data was drawn from medical records. Rurality was determined by mapping the participant’s self-reported county of residence onto a designation of county-level metropolitan status determined based on US Census Metropolitan and Micropolitan Statistical Areas.²⁰

Results

24 patients were approached. One did not have a history of IDU in the past three months; three declined because they were not feeling well enough to speak with the interviewers, and four who were interested were transferred or discharged before enrollment. Sixteen persons hospitalized with IDU-related IE (IDU-IE) were interviewed. Demographic and clinical characteristics of the study population are presented in Table 1. Almost two-thirds (69%) of the participants were young adults between the ages of 20 and 29, with approximately half identifying as female (44%). Most participants had no prior history of IE. 60% of participants were from a metropolitan county, and 40% from a county that is not. Fifty-six percent of participants identified as White, with 19% identifying as Native American, and 13% identifying as Black.

Table 1.

Characteristics of Participants

Participant characteristics		%	(n/N)
Age
	20–29 years	69%	(11/16)
	30–39 years	25%	(4/16)
	40–49 years	6%	(1/16)
Gender
	Man	38%	(6/16)
	Woman	44%	(7/16)
	Other Gender Identity	19%	(3/16)
Self-identified Race
	White	56%	(9/16)
	Black	13%	(2/16)
	Native American	19%	(3/16)
	Other/Unknown	13%	(2/16)
Metropolitan Status of Home County1
	Metropolitan	60%	(9/15)
	Micropolitan	27%	(4/15)
	Neither	13%	(2/15)
Insurance Status
	Medicaid	44%	7/16
	Commercial	6%	1/16
	None	44%	7/16
	Other	6%	1/16
Prior Infective Endocarditis
	No	63%	(10/16)
	Yes	38%	(6/16)

^1.Metropolitan status not included for one participant who resided outside of North Carolina.

We identified four primary themes related to knowledge of IDU-related IE: limited knowledge of IE, nonspecific IE symptoms, injection behavior and infection prevention, and knowledge of other IDU-related infections, such as HCV and HIV.

Theme 1: Limited knowledge of IE

We observed limited general familiarity with IE among PWID. Many reported having little to no prior knowledge of endocarditis, as stated by one of the patients:

No, I’d never [heard of endocarditis before]. I didn’t know anything about it, my foot had swole [/swollen/] up. And ah, so I came to the hospital and thinking I was going to leave with like some antibiotics

(38-year-old, White Male).

Meanwhile, others provided their knowledge in terms of what they learned since being hospitalized, as stated by this patient:

They really um haven't told me much about endocarni- endocarditis. All they told me is that it's a disease that eats my heart and sets on my heart and… um… that… it can kill me if I don't get it taken care of. It can go spread through my whole body and umm… it can kill me.

(22-year-old, White Female)

Despite reporting limited knowledge of IE, many participants said they had known peers with similar symptoms or who had been diagnosed with IE before. For example, one participant commented that they remembered a friend having surgery and seeing them in the hospital, despite not knowing that it was IE:

I heard [of it] from a couple people like [my friend] had it, she had a good friend die from it. Um, and another person we knew had gotten sick from it, but before I came to the hospital, I had no idea that that’s what it was, but I kind of had an idea that maybe it was something from using, but I didn’t know that it was called

(29-year-old, White Male)

Participants reported conflicting views regarding the association between severe infections and IDU. For example, one participant reported:

“[I] would have never thought that this would have happened to me. I thought I was just your typical average 30-year-old drug addict.”

(30-year-old, White self-identified as Bi-Gender).

In contrast, another participant stated that once diagnosed, they felt as though “everyone” associates IE with IDU, saying:

“…as soon as you get it, it’s like, everyone’s like, ‘oh, you know where that came from.’ I hate that…Cuz there are certainly other way you can get it besides shootin’, right? Uh, I don’t know”

(22-year-old, White Female).

Participants’ limited knowledge of IE stands in contrast to their knowledge of symptoms related to other infections that may need acute treatment, such as skin and soft tissue infections (SSTIs). Of those interviewed, majority reported having experienced SSTIs, of which participants reported both self-treatment through drainage and bandaging, as well as seeking medical care.

Theme 2: Diverse and nonspecific symptoms prompting presentation to care.

Participants reported a broad range of symptoms prompting them to seek care, such as fevers, aches, difficulty breathing, and vomiting. Many participants stated that they assumed their symptoms were due to a virus and would pass in time, resulting in delayed care:

You know, I just figured I had the flu until, like, they got me there, and that's when they found out I had, you know what I'm saying, an infection

(49-year-old, Native American Male).

Yeah, last year when I got my heart surgery cuz I [fell] out on the bus, didn't, didn't know why I was, to the point I was almost dead, then, and didn't know what was going on …And when I came to, they was like, “you gotta have surgery, and this that and da da da… you have this, you have that.” And I'm like, “What? ”

(40-year-old, Black Female).

Many participants perceived their symptoms to be less severe than those around them. Some participants reported that a friend or loved one encouraged them to seek medical attention. One participant recalled that their fiancé made them go to the hospital, saying:

“Last time, I waited nine days before I went. Last year if I waited 24 more hours, I would have been dead.”

(49-year-old, Native American male)

Another participant told a story of her roommate calling her grandmother to convince her to go to the hospital:

“And my grandma’s on the phone, he's with my grandma on speakerphone and they're both hollerin’ at me and I'm cussing em and flipping out. And he, she, my grandma was like, [friend’s name omitted], just grab her and throw her in the car… she was like, I don't care what you got to do. She was like, I'm telling you, get that girl in that car and take her to the hospital and dro-drop her off. Thank God that they did because if they wouldn't like I'd be dead right now.”

(22-year-old White female)

In their narrative, participants emphasized someone else's role in getting them to the hospital and acknowledged that they under-assessed the severity of their own symptoms.

A subset of participants sought medical care at least once before being diagnosed with IE. These participants reported being told their symptoms did not merit inpatient care as reported by one of the participants:

“Yeah, they said, “you just have a virus. Go home and just, you know, sleep it off,” …But they did give me a COVID test, and they just were like, ‘well, you don't have the COVID. So, you can go home.’”

(28-year-old, Black/African American Female).

Theme 3: Injection Behaviors and Limited Infection Prevention Access

When asked about IDU practices, participants provided a range of responses. Almost all (14/16) participants reported reusing their needles, with majority (10/16) reporting sharing injecting equipment. When asked how participants accessed clean injecting equipment, participants reported obtaining clean supplies via a syringe service program or purchasing through pharmacies:

“I, I, I got like a whole box of … 100 needles from Wal-Mart for like $13. You can get, I mean, that's pretty cheap.”

(32-year-old Female, Unknown Race).

However, participants listed difficulty in affording new injecting equipment as a barrier to using sterile syringes and reported continuing to use needles that were damaged during use:

“…Sometimes I couldn't afford to go up there and buy [clean needles]. So, I have to reuse the ones I had already. Not other people's, but reuse the ones I already had, you know?”

(21-year-old White Male)

Having a low supply of injecting equipment on hand led to riskier injection practices:

“The needles are so easy to [mess] up, you're only supposed to use them once and usually you just have a couple or somethin'. So, you keep using those until you, you know, get more.

(34-year-old White Female).

Of the 14 participants who reported reusing their needles, most participants reported methods by which they cleaned their needles. Cleaning practices varied across participants. Some reported using only warm or hot water, while others used soap, alcohol, or bleach to sanitize their needles.

“…I put some warm water in it, rinse the blood out of it about two or three times. Then I put my tops back on it back, put it back in a bag. I never use alcohol or none of that. I just use water, rinse 'em out and put it back.”

(28-year-old, Black/African American Female)

“You gotta rinse [the needle] out with alcohol, but really, I'm just lazy, and I just do with water.”

(34-year-old, White Female)

However, despite strategies to clean needles to avoid infection, participants also discussed injecting practices that predisposed them to risks of acquiring certain infections. Many participants reported using nonsterile water sources to clean, mix, and cook their substances:

“Umm, most of the time, I have to find something to dump my shit in a-and, and a clean cotton and a needle, and I pull it up and get clean water from wherever sink or clean water bottle or whatever… I have gotten off the ground like a puddle, I've gotten toilet water before. I don't know how dirty that is.”

(38-year-old, White Male)

“I'd rather use [bottled water] than sink water, you know? but- I mean, I've used soda to shoot up before. I've used spit to shoot up before. Like, I mean it just, it just depends on the situation that I'm in.”

(30-year-old, White Bi-Gender)

Theme 4: Better Knowledge of HCV and HIV Risk

Regarding communicable diseases, all participants reported having been tested for HIV and Hepatitis C, either previously or during the current hospitalization and knew their status.

Interviewer: So it sounds like you're saying that, you know that sharing needles that can get you sick?

Participant: Yeah. Like Hepatitis B or C or whatever. Mm hmm. Yeah.

Interviewer: You mentioned hepatitis, you ever been tested for hepatitis?

Participant: Uhh, not until I got up here this time. (21-year-old White Male)

Most patients discussed that they knew about their HIV status before being hospitalized, with some patients asking hospital staff to provide testing.

Interviewer: OK. And have you ever been tested for hepatitis?

Participant: Yeah, I'm negative.

Interviewer: Negative. OK. Have you ever been tested for HIV?

Participant: Yeah, negative. They just did that test, and uh, I asked them to do it… cuz I wanted to see if I had hepatitis C, you know, just to see, cuz my ex-boyfriend does have it. (33-year-old, White Female)

They also conveyed a clear understanding that sharing needles was increasing the infection risk:

“Now I gave people my needles that I've used but I've never used a needle after somebody. Like, I gave them my dirty needles because they're, they're not [going to care], but like, as far as me [sharing needles], that’s a no, negative. I'd snort it before I use a needle after somebody.”

(30-year-old, White Bi-Gender).

Discussion

Our qualitative study sought to explore patients' familiarity with IE, experience with its symptoms, and knowledge of risk behaviors for acquiring IE. Our results capture the experience of PWID hospitalized with IE and can inform public health strategies to address prevention and awareness of IE among individuals at risk.

First, participants presented with limited prior knowledge of IE as a potential consequence of IDU. However, while there appears to be limited discussion of IE among PWID before diagnosis, participants noted that once diagnosed, they recognized the symptoms in friends or remembered they had heard of others diagnosed with IE. Even among those diagnosed previously with IE, there appeared to be little comprehension of the root causes of IE as related to IDU. This finding, in combination with our finding that participants were aware of their HIV and HCV status and that sharing needles increases that risk, suggests that public health messaging is needed to address IDU-related risk of IE. Studies have identified that PWID positively favors receiving HCV and HIV education, screening, and referrals to treatment through harm reduction programs, such as SSPs.^21,22 Integrating education regarding severe bacterial infections, such as IE, into harm-reduction programs could help increase awareness among PWID and help empower individuals with the knowledge to seek care promptly. Further, such initiatives can address infection prevention strategies such as promoting the use of sterile water, skin sterilization, and avoiding reuse of injecting equipment.

Public health messaging about recognizing IE symptoms may be challenging. Patients reported that their IE symptoms prompting care were nonspecific. Nonspecific symptoms are common with IE, for which symptoms may be indistinguishable from those of other systemic inflammatory responses, as seen in certain viral infections, such as influenza or COVID-19, complicated urinary tract infections, or other syndromes.²³ Additionally, cotton fever, a noninfectious syndrome due to the injection of bacterial endotoxins, can present with a similar inflammatory and sepsis-like syndrome but will spontaneously resolve.²⁴ Although IE can present with focal symptoms due to either heart failure or embolic phenomena (e.g. stroke, septic arthritis, lung abscess, etc.), such localizing symptoms were not typically reported in our interviews. Nonetheless, given the increasingly strong association between IDU and serious invasive infections,³ patients and providers should be educated to evaluate seriously any symptoms potentially indicative of sepsis (e.g. fever, chills, myalgias), even though these might more commonly be attributed to less aggressive syndromes.

The nonspecific symptoms of IE can delay care. This can occur because patients and providers may not recognize the potential severity of the symptoms and because patients lack access to services.^23,25 In North Carolina, where Medicaid was not expanded at the time of this study, many patients did not have insurance. PWID experience barriers to accessing medical care due to stigma or discrimination, untreated withdrawal, or feeling their medical questions and concerns are dismissed, resulting in further avoidance of seeking timely medical care.^22,26,27 This stigma becomes then magnified and compounded through intersectional lenses of race and gender.²⁸ It is worth noting that approximately half of the study participants identified their sex as female. Research has enumerated how women, regardless of drug use history, often experience their symptoms being dismissed when they present for medical care, with much research focusing on cardiac symptoms.²⁹ As other studies indicate that women who inject drugs make up a greater percentage of PWID who are diagnosed with IE,^2,30 future work is needed to understand how intersecting identities, such as social status, gender, and race, may compound, creating barriers to earlier screening and diagnosis. To address PWID’s access to health care, some harm reduction models have utilized mobile health units or embedded basic healthcare screenings into community-based harm reduction programs to address barriers to PWID accessing care.^31,32 Such programs may help PWID stay connected to the healthcare system and help facilitate earlier care for patients with symptoms of severe infections.

Lastly, PWID may have limited access to injection supplies. Even participants who obtain injecting equipment from syringe service programs, especially those in rural North Carolina, reported difficulty accessing sterile drug use supplies.³³ Research has reported that the COVID-19 pandemic disrupted the methods by which persons used drugs, resulting in increased injection drug use.³⁴ State policy regarding harm reduction further complicates this matter by limiting the types of drug use supplies a harm reduction program can provide.^34–36 States, including North Carolina, have offered limited immunity regarding drug paraphernalia to allow for syringe service public health initiatives. Yet despite this limited immunity, a study by Morrisey et al. (2022) found that law enforcement often doubts or questions PWID's rightful acquisition of these supplies, which can further prevent PWID from accessing sanitary supplies. At a minimum, healthcare providers and harm reduction programs may want to provide education on the risk of IE for PWID and information on how to reduce that risk.

This study contains several limitations. These data may not reflect the experiences of those who were uncomfortable participating in the interview, as they may not have consented to the study. Research indicates that many PWIDs experience stigma within the hospital context and, therefore, may not be willing to disclose or discuss drug use as it relates to IE.³⁷ Further, participants may have been only partially forthcoming of their knowledge of IE out of fear of stigma for continuing IDU despite that knowledge. It also stands pertinent to note that these perspectives are of those who sought care; thus, they may have had a higher awareness of the need for care as opposed to those with IE who did not seek care. We also recognize that PWID exists within political, structural, social, and economic environments that help shape their knowledge.^14,38 As we focused specifically on participants’ articulation of their understanding of IE, further research should excavate the broader environments in which that knowledge is produced. Lastly, all participants were recruited from a single health system. Research is needed to understand the experiences of PWID who are hospitalized with IE and how stigma might impact IE risk.

Overall, our findings indicate that efforts aimed at IE education and prevention are needed. While IE symptoms are heterogeneous and nonspecific, greater awareness of the risk of IE may help PWID in seeking timely medical care. Ensuring access to safe and sterile drug use supplies, as well as education regarding safer methods of drug use, may support PWID in preventing IE. Implementation studies are needed to develop and pilot educational initiatives focused on IE.