Recidivism at the puerto rico trauma hospital

Trauma kills over five million people globally in a year, generating 9% of the world’s mortality [1]. In the United States (US), trauma contributes to two of the ten leading causes of death—unintentional injuries and suicides—and the leading cause of death in individuals aged 1–44 years old [2]. Non-fatal injuries represent 68% ($456.9 billion dollars) of all medical and life productivity loss costs; while fatal injuries account for nearly 33% of all medical and life productivity loss [3, 4].

There is limited data regarding a rare population of trauma patients that present an injury on two or more separate occasions (recidivists). The rates of recidivism vary largely, ranging from 0.38% to 44% depending on location [5, 6]. Several characteristics including sex, race, socioeconomic position (neighborhood level), substance use, health insurance status, employment status, and civil status have been associated with recidivism [5‐9]. In addition, a couple of studies reported recidivists were likely to present the same mechanism of traumatic injury in a subsequent admission as in the first injury [5, 10]. However, due to the heterogeneous classification of mechanisms of injury, the prevalence and trends of mechanisms associated with recidivism remain unclear.

Moreover, when comparing recidivists to non-recidivists, there have been either contradictory findings or little findings regarding the outcome variables—the Injury Severity Score (ISS), the Glasgow Coma Scale (GCS), the hospital length of stay (LOS), the amount of days in the trauma intensive care unit (TICU), the requirement for mechanical ventilation (MV), the MV days, and mortality [5, 8, 12, 13]. The window for reinjury has been estimated to vary largely between 7.9 months and 45 months [5, 14]. Defining time to reinjury among recidivists is instrumental due to its implications in prevention efforts.

Puerto Rico (PR) is an unincorporated territory of the US, with a largely predominant Hispanic population. In 2016, about 43.5% of the population was below the poverty level, more than triple the national average of the US and double the poorest state in the US‒Mississippi [15]. The Puerto Rico Trauma Hospital (PRTH) attends all levels of injury-related medical emergencies, being the only tertiary level II trauma center in PR. The hospital offers a unique setting to evaluate factors that may predispose to trauma recidivism. Thus, the present study aimed to determine the rate and independent risk factors of trauma recidivism, as well as comparing clinical and injury-specific features of the first and second traumatic event in recidivist patients. Findings could encourage policy advancement, and the development of tailored cost-effective intervention programs at hospital and community level.

A total of 24,650 patients were admitted to the PRTH during the study period. The rate of trauma recidivism was 14 per 1000 patients discharged alive. Table 1 depicts the comparison of sociodemographic characteristics, injury profile, hospital course, and outcomes between the study groups. Recidivists were predominantly male (95.4% vs. 82.6%; p < 0.001), and frequently within the 15- to 24-year-old age range (40.1% vs. 25.4%; p < 0.001) when compared to non-recidivists. Furthermore, those patients with more than one trauma-related admission were mostly from Metro-North (25.2% vs. 20.3%), San Juan (18.6% vs. 15.0%), East (17.3% vs. 16.3%), and Northeast (14.1% vs. 13.7%) Health Regions when compared to their counterparts with only one trauma-related admission (p = 0.013). Additionally, the proportions of patients with public health insurance (39.4% vs. 26.5%) or uninsured (12.3% vs. 10.6%) were higher in the recidivism group than in the non-recidivism group (p < 0.001).

Regarding the type of injury, penetrating trauma was more common among recidivists (39.5% vs. 26.9%; p < 0.001). The leading mechanism of injury overall was motor vehicle accident (MVA), occurring mostly in the non-recidivism group (36.8% vs. 35.2%). Nevertheless, there was a greater proportion of gunshot wounds (GSWs) (30.3% vs. 19.0%) and stab wounds (SWs) (9.5% vs. 7.7%) among recidivist trauma patients (p < 0.001). Violence-related injuries were the most prevalent mechanisms of trauma among recidivists (44.6% vs. 29.4%; p < 0.001).

As to vital signs (i.e., breathing, SBP, temperature, and heart rate) and ABD, there were no statistically significant differences (p > 0.05) between the study groups. Similarly, the proportion of patients admitted to TICU, TICU LOS, MV days, and hospital LOS were comparable between recidivist and non-recidivist patients (p > 0.05). However, the relative frequency of patients requiring MV was lower for those with more than one trauma-related admission than for their counterparts with only one trauma-related admission (15.0% vs. 19.9%; p = 0.031).

When evaluating injury severity markers, recidivist subjects were significantly less likely to have an ISS above or equal to 25 (7.9% vs. 18.1%; p < 0.001) and a GCS below or equal to 8 (6.3% vs. 13.9%; p < 0.001) relative to their non-recidivist counterparts. In-hospital mortality was also lower among patients with more than one trauma-related admission (7.2% vs. 10.7%; p = 0.045).

The results of the multivariate regression analysis revealed factors independently associated with trauma recidivism, as displayed in Table 2. However, none of the interaction terms tested in the model had a significant effect. Males were 3.88 (95% CI: 2.21–6.80) times more likely to suffer from a second injury episode; and individuals aged between 15 and 24 years old were 3.80 (95% CI: 2.24–6.46) times more likely. Moreover, patients with public health insurance had a 51% [adjusted odds ratio (AOR) = 1.51; 95% CI: 1.12–2.03] increased risk of experiencing another injury. An ISS above or equal to 25 (AOR = 0.44; 95% CI: 0.28–0.68) and a GCS below or equal to 8 (AOR = 0.56; 95% CI 0.34–0.92), conversely, were demonstrated to be independent protective factors from a second traumatic event.

The present study primarily sought to compare recidivist and non-recidivist trauma patients in terms of their sociodemographic characteristics, injury profile, and hospital outcomes. In the PRTH, 1.4% of the population experienced recidivism. Previously, rates of recidivism had been found to differ from 0.38% to 44% depending on location [5, 6]. This is a considerable variation that could be explained by several studies including less severe injuries; by including evaluations made in the periphery; and by excluding mechanisms of injury unrelated to violence. A male majority among these patients in our institution is also consistent with the existing literature [5, 12, 17, 18]. Interestingly, the descriptive analysis suggests that a larger proportion of recidivists present at an earlier age, with 40% of cases being reported among the 15–24 age group. This was comparable to findings published by Strong et al., in which 37.3% of recidivists were aged 18–29 years old [13], distinct from other pieces that identified recidivists to be over 30 years old [5, 10, 19]. Our population of recidivists frequently used public health insurance, contrast to other populations that are predominantly uninsured [9, 12, 18]. A possible explanation for this discrepancy is the accessibility to public health insurance among the population served by our institution. When admitted to our hospital, uninsured patients are often offered the opportunity to acquire public health insurance with assistance from social workers.

In PRTH, the leading mechanism of injury was MVA, notably among non-recidivists, likewise to other findings throughout the years [5, 8, 9, 17]. Violence-related injuries overall occurred largely among recidivists, as previously described [8, 19]. However, the percentage of violence-related injuries among recidivists in PRTH (44.6%) was considerably higher than the 25% seen by Dixon et al. and the 15% documented by Erdogan et al. [8, 19]. Furthermore, a previous study revealed recidivist patients were likely to present with the same mechanism of injury in the first and second traumatic events, with rates of up to 75% [5]. In our population, 65.7% of recidivists presenting with violence-related injuries during the first event returned for the same type of injury. Although McCoy et al. used different classifications for violent injuries, it was reported that 34% of GSW/SW victims and 37% of assault victims tend to recur with the same mechanism of injury [18]. Further analysis suggests PRTH recidivists suffer from more violence-related injuries on their second event when compared to their first event; whilst the contrary occurred with motor vehicle-related trauma.

The ISS, GCS, MV requirement, and mortality have been considered questionable due to limited evaluation. Consistent with our ISS results, Caufield et al. demonstrated recidivists suffered less severe injuries than non-recidivists, as did Dixon et al. and Erdogan et al. [5, 8, 19]. Regarding the GCS, our recidivists were less likely to have a severe state. In a 2007 study, Toschlog et al. documented a higher GCS in recidivist patients as compared to their non-recidivist counterparts, whereas Caufield et al. found no such difference [5, 20]. In addition to these severity markers, the PRTH patients with one trauma related admission had a greater need for MV than their recidivist counterparts. Previously, no differences in in-hospital mortality based on the recidivism status of patients were detected in many scientific works [8, 19, 20]. Yet, our analysis suggests PRTH recidivists were less likely to die during their hospital stay. Kwan et al. described a similar finding, with non-recidivists suffering from higher in-hospital mortality [12].

The median time to reinjury for recidivist trauma patients was 42 months. A study of a Canadian recidivist population exhibited an average 41 months to reinjury, which is closest to our marks [19]. Closer to one of the first reinjury analyses, with a median time of 7.9 months, Kaufmann et al. demonstrated a median time of approximately 10 months to reinjury [14, 21]. There is a wide range of times to reinjury reported in the scientific literature, possibly due to either limitations or wide access to healthcare center locations of various levels. Moreover, the time to reinjury could vary according to the type of trauma, as alluded in our institution. The time to reinjury for non-penetrating trauma and penetrating trauma differed by about a year, with penetrating injuries occurring earlier. A possible explanation is that non-penetrating trauma, although commonly involves a third party, has more elements under the control of the possible recidivist. These factors may include seatbelt usage, traffic vigilance, prevention of falls through avoidance of clutter at ground level, stairs, and protected ladder usage, among others. Penetrating trauma, however, involves third party behavior commonly beyond the victim’s control.

As for predictors of reinjury, male sex was a significant risk factor for all forms of recidivism, consistent with literature on violent recidivism [11, 21, 22]. Furthermore, the 15–24-year-old patients in PRTH were found to be almost four times as likely to be recidivists. This was a stronger association than previously reported by Algham et al., in which 18–25-year-old patients were almost twice as likely to be recidivists [23]. Public health insurance was also independently associated with an increase in all forms of reinjury, as depicted in other literature [23]. Interestingly, however, a severe GCS or ISS could be a protective factor for our population. Patients with a GCS less than or equal to eight or an ISS greater than or equal to 25 were half as likely to become recidivists, thereby serving as potential protective factors. In other general trauma recidivism studies, no predictive effect has been found for the GCS or ISS [10]. A violent recidivism study by Kaufmann et al. also conveyed this lack of value [21]. Yet, Nygaard et al. stated a lower ISS was a predictor of violent recidivism [22]. The protective effect found in the severity markers could be due to consequential lifestyle or functional impairments which led to avoidance of the situation or environment, precautionary measures instilled, possible death due to trauma before a second admission, or loss of follow-up due to emigration. Nevertheless, further investigation of long-term mortality and functional impairments of these patients, including all mechanisms of trauma, would be of benefit to clarify these risk or protective factors.

The early identification of potential recidivist trauma patients by hospital personnel, and an in-depth history assessment, including substance abuse and psychiatric disorders, could expand our knowledge base and promote investigation. This could in turn be used to tailor prevention platforms targeting road safety, violence, and mental health, among others. Previously, health care-based violence interventions directed towards recidivism prevention have shown some improvements [13]. Moreover, the establishment of community based, mental health, and substance abuse treatment programs could be useful before the recurrence. Finally, time to reinjury supplies valuable information on the window of prevention to establish these methods efficiently and effectively.

There were limitations to rendering a definitive picture of recidivism. Several records were lacking information, whereas others could have had variables erroneously evaluated. Most files were missing variables related to substance use, including alcohol use, which led these to be excluded from further study. Self-inflicted injuries and repeated attempted suicides were also not explicitly included in the study. Although the aforementioned categories represent a very small fraction of the population, exclusion of them could contribute to bias in our findings. Recurring patients may have also had their information incorrectly entered under a new record number and, consequently, been identified mistakenly as a non-recidivist. Additionally, the island has had a high emigration rate throughout the years, suggesting patients leave before a possible recurrence. Recidivists could have also arrived to institutions in the periphery and have been effectively managed. However, this is less likely due to high referral rate to our trauma center, the only tertiary level II center on the island. Death of patients on scene either early after the initial injury, or due to comorbidity, could have affected the results and information available. Moreover, some recurring patients might have not been identified because the follow-up period was not drawn out enough for such an event to occur. Because of this, there could have been a misrepresentation of information and, subsequently, in the classification of recidivists and non-recidivists.

Recidivists are more likely to be male, present at an earlier age (15–24-year age group) and use public health insurance. Additionally, recidivists were more likely to suffer violence-related injuries, suffer less severe injuries, be evaluated with a higher GCS, and have less need for MV. This PRTH subgroup also exhibited less in-hospital mortality. Comparing the first and second traumatic incidents of PRTH recidivists, findings disclosed more violence-related injuries on their second occurrence; whilst the contrary occurred with motor vehicle-related trauma. Second traumatic injuries were inclined to be graver than the first episodes, with a decline in neurological status, increasing severity scores, a marginally longer hospital stay, and greater requirement for MV. Risk factors of reinjury include the male sex, being 15–24 years old, and possessing public health insurance. However, a severe GCS or ISS could be protective factors. The median time to reinjury was 42 months, with penetrating trauma reoccurring about a year earlier than non-penetrating trauma, giving a window of opportunity for intervention. Future studies could be directed toward investigating long-term mortality, the evaluation of deaths on scenes, and verifying outcomes or presence of possible recidivists in surrounding hospitals. The usage of this information could promote effective interventions and awareness around the existing problem.