If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Violence Prevention Research Program, Department of Emergency Medicine, University of California Davis School of Medicine, 2315 Stockton Blvd, Sacramento, CA, 95817, USA
Violence Prevention Research Program, Department of Emergency Medicine, University of California Davis School of Medicine, 2315 Stockton Blvd, Sacramento, CA, 95817, USA
Division of Health Policy and Management, Department of Public Health Sciences and Center for Healthcare Policy and Research, University of California Davis, 4900 Broadway, Suite 1430, Sacramento, CA, 95820, USA
Division of Trauma, Acute Care Surgery, and Surgical Critical Care, Department of Surgery University of California Davis Medical Center, 2335 Stockton Blvd, Sacramento, CA, 95817, USA
The analysis of total hospital cost per visit is an effective hypothesis-generating strategy.
•
Trends in visit cost and frequency suggest clinically distinct groups requiring separate intervention strategies.
•
High-frequency, low-cost return visits to the emergency department are prevalent.
•
Older white males demonstrate different cost trends than younger men of color.
Abstract
Background
Analysis of the costs associated with emergency department (ED) visits after discharge for violent injury could highlight subgroups for the development of cost-effective interventions to support healing and prevent treatment failures in violently injured patients.
Methods
A retrospective cohort review was conducted of all patients with return ED visits within 90 days of discharge after treatment for a violent injury occurring between July 1, 2016, and June 30, 2018. Hospital costs were calculated for each incidence and analyzed against demographic and injury type variables to identify trends.
Results
218 return ED visits were identified. Hospital costs showed a high frequency of low-cost visits. For more complex visits, distinct cost patterns were observed for Black and LatinX males compared to White males as a function of age.
Conclusions
Analysis of hospital cost per visit identified trends among different subgroups. Underlying etiologies presumably vary between groups, but hypothesis-driven further investigation and needs assessment is required. Understanding the driving forces behind these cost trends may aid in developing effective interventions.
1. Introduction
Return visits to the emergency department (ED) after initial hospitalization for violence-related trauma represent a significant burden on patients and their support systems, as well as healthcare systems more broadly.
As EDs continue to process high volumes of patients, identifying opportunities for preventing return visits remains critically important. Violently injured patients are of particular concern, as they are more likely to revisit the ED within 90 days of discharge than trauma victims overall.
Current literature on ED revisits focuses heavily on recidivism for new violent injuries. Less data is available to describe patterns of early return to the emergency department after violent injury for issues that may be related to the index trauma.
Known risk factors for all-cause ED revisits include public insurance and low-income zip code; both of which are prominent within the violently injured population and are consistent across race/ethnicity, age, and sex/gender demographics.
et al. used hospital costs to compare the utility of hospital-based interventions, making it an attractive baseline for a needs assessment.
Our objective was to conduct an analysis of costs per visit for early ED revisits following violent injury, and to identify patterns within demographic subgroups for further analysis. We chose to focus on the 90 days following discharge for the initial violence-related injury to identify patterns more strongly associated with treatment failure rather than violence recidivism. Treatment failures leading to revisits such as injury site infections or inadequate pain control may occur due to miscommunication, misunderstanding, or inability to follow recommendations secondary to social, structural, or economic factors. These factors may include mental health issues including acute stress disorder and PTSD, lack of access to necessary components of care, and competing priorities.
Using hospital costs to identify subpopulations with frequent or costly treatment failures could help streamline needs assessment and intervention efforts.
Revisit costs represent an opportunity to estimate the frequency and complexity of these occurrences and identify areas for improvement with respect to cost savings, patient experience, and discharge planning.
2. Methods
This study was designed as a retrospective cohort review of ED return visits among violently injured patients admitted to a high-volume academic level 1 trauma center in inland California. Hospital costs for all ED revisits occurring within 90 days of discharge were considered. Data were obtained from the hospital trauma registry and billing record databases, respectively. Prior to the identification of subjects, the protocol was reviewed by the local Institutional Review Board and was determined to be exempt.
2.1 Sample
The study group was identified through the hospital trauma registry database as patients with an index admission between July 1, 2016, and June 30, 2018, and an ICD-10-CM external cause of morbidity code of assault (X92-Y04 and Y08–Y09). Patients in legal custody at index admission, victims of domestic violence or child abuse (ICD-10-CM code Y07), and those with self-inflicted injuries (ICD-10-CM codes X71-X83) were excluded. Patients in legal custody were excluded as these patients are often managed by correctional facility hospital staff after discharge and are not able to return to the ED on their own. Victims of domestic abuse and self-harm were excluded due to preexisting intervention efforts already focused on these subpopulations.
2.2 Measures
Sex/gender was defined as male or female and a binary categorization was employed. Race/ethnicity was defined using National Institutes of Health racial and ethnic categories: non-Hispanic American Indian (AI) or Alaskan Native (AN), Asian, Black or African American, Native Hawaiian or Other Pacific Islander (PI), white and other, and Hispanic/Latinx. Due to their small sample size, patients who were categorized as Asian, AI/AN, Native Hawaiian or Other PI, and other were grouped together into an “other” category. Age was dichotomized (35 years and older or less than 35 years) for the purposes of subgroup analysis. Mechanism of injury was also dichotomized: penetrating injuries primarily included stab wounds and wounds associated with firearms, whereas non-penetrating injuries primarily included injuries with blunt weapons such as bats or injuries from assault without weapons (e.g., punches and kicks). Hospital costs, defined as the expense incurred to deliver health care services to the patient, was chosen as the economic variable due to its consistency across payor status and greater generalizability when compared to charge or reimbursement. Costs were analyzed per individual visit.
2.3 Analysis
Standard descriptive data analysis techniques were used to review the frequency of return ED visits by categories of age, sex/gender, race/ethnicity, and mechanism of index violent injury. Total hospital costs were obtained for each visit. The distribution of these costs was determined. T-tests were used to compare the total hospital costs between subgroups.
Racial/ethnic subgroups were further divided into youth/young adult and adult categories by age (<35 vs ≥ 35), consistent with standard cut points used by the Centers for Disease Control and Prevention. Due to the small number of female patients, they were excluded from additional subgroup comparison. Descriptive techniques were used to review cost patterns within sex, race, and age based subgroups.
3. Results
450 patients met the inclusion criteria of violent injury. Within those patients, 218 return ED visits across 127 unique patients were identified within 90 days of discharge. Most patients had similar payor status, with 77% covered by Medicaid. Overall, the median hospital cost for a return visit was $641.
The number of return visits by patients identified as female was markedly lower than the number of return visits by patients identified as male. Only 17 return visits involved female patients compared to 201 visits for male patients.
3.3 Age
The age of patients included in the study ranged from 11 years to 68 years, with a mean age of 36 years. The average age among the racial/ethnic groups was notably different. Black and Hispanic/Latinx patients tended to be younger with mean ages of 33.1 years and 32.6 years, respectively. White males tended to be over 35 years of age (45 years on average).
3.4 Race/ethnicity
The racial/ethnic composition of patients with return ED visits included 85 Black, 67 White, 29 Hispanic/Latinx, and 37 patients who identified outside these three categories.
3.5 Index injury
More than half (124 or 57%) of return ED visits were secondary to an initial (or index) penetrating trauma such as gunshot wounds or stab wounds and 86 (43%) were secondary to non-penetrating trauma such as blunt trauma from baseball bats or other non-edged weapons or assault with punches or kicks. Index violent injuries are stratified by race/ethnicity in Table 1. The percentage of index injuries due to penetrating trauma varied by race, with penetrating trauma responsible for 57% and 46% of index injuries in the returning Black and “all other” racial/ethnic subgroups, respectively, compared to 29% and 10% in the returning Hispanic/Latinx and white subgroups, respectively. Returning white patients were more often victims of blunt assault at index admission, with 66% falling within this category (see Table 2).
Of the 218 return ED visits, 15.6% (34 visits) resulted in a hospital readmission. The race/ethnicity-based subgroups with larger number of ED revisits also had the greatest number of readmissions. Black patients had 12 readmissions representing 14.6% of all ED revisits in this subgroup, and white patients had 11 readmissions representing 17.2% of all ED revisits in this subgroup. Hispanic/Latinx patients had 5 readmissions, while the category “Other” had 6 readmissions.
3.7 Average hospital costs
The average hospital cost overall was $4230. The average cost per visit was not statistically different based on sex, age, race/ethnicity, or injury type.
3.8 Distribution of hospital costs
The distribution of cost frequency is shown in Fig. 1. Fig. 1 demonstrates the variation in hospital costs for ED revisits ranging from $189 to $91,000. Low cost of ED revisits occurred at high frequency. Overall, 58% of costs were below $1000 with 31% percent of total visits below $500. “High cost” visits, defined as >$10,000, represented 10% of all ED visits. These high-cost visits did not appear confined to a single demographic subgroup but rather their intersection. High-cost visits were distributed such that greater numbers of these visits were seen in different age groups when stratified by race/ethnicity (Fig. 2).
Fig. 1Distribution of hospital cost per visit for each return visit to the Emergency Department. (A) displays all visits, demonstrating the majority are <$10,000. (B) distribution of visits with costs under $10,000, the majority are less than $1000. (C) shows the distribution of all visits under $1000.
Specifically, among male, Black, non-Hispanic patients, higher cost visits were seen in patients less than 35 years of age. When costs for this group are viewed as a function of age, decreasing age correlates with increasing cost as seen in Fig. 3. Conversely, among male, white, non-Hispanic patients, higher re-visit costs were seen among those aged greater than 35 years. When cost is viewed as a function of age, increasing age correlates with increasing cost for this subgroup (Fig. 3).
Fig. 3Graph demonstrating cost trends across racial subgroups. Hispanic patients show a similar trajectory as African American patients while the other subgroup (red line) demonstrates both age effects. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)
ED revisits among older white males were most often preceded by anon-penetrating violent injury at index visit as opposed to young Black male patients who had a higher proportion of penetrating trauma at index visit. Return ED visit costs among male, Hispanic/Latino patients followed a similar age-specific pattern as their non-Hispanic Black counterparts. The non-Hispanic, other racial/ethnic subgroup demonstrated a mixed cost pattern with a trend toward increased costs at both margins of age (Fig. 3).
4. Discussion
Hospital cost analysis of ED revisits can be an efficient tool for directing organizational needs assessment and intervention. Our retrospective analysis of ED revisits following violently injury demonstrated two distinct cost patterns representing clinically different groups for further investigation.
4.1 Low cost revisits
The present analysis found a high frequency of low-cost ED revisits. Nearly 50% of all revisits were assigned costs of < $500, with most around $200. These figures likely represent the institution's choice of emergency department facility fee and include few if any tests or interventions. Nonetheless, ED care is significantly more costly compared to other sources of care both in time and resource utilization.
While preventing individual low-cost visits may be seen as having little utility in cost savings for a hospital system, unnecessary ED visits utilize valuable time and resources that could be focused on sicker patients. The high frequency of these return visits ultimately adds up to a significant share of the total cost burden. Revisits for wound checks, dressing supplies, pain medicine refills, or suture and staple removal could all be handled in an outpatient setting, for example. Increasing patient access to non-ED options for care could provide a solution. Ultimately, further investigation is warranted into the specific reasons for these revisits following violent injury discharge so appropriate interventions can be determined.
4.2 High cost revisits
Ten percent of revisits were of remarkably high cost (>$10,000). These return visits were not associated with any demographic subpopulation alone, but rather demonstrated an interesting relationship at the nexus of sex/gender, race/ethnicity, age, and index injury type, as demonstrated in Fig. 3. Two distinct high-cost patterns deserve further comment. For patients identified as male, clear trends in revisit costs emerged based on age, race/ethnicity, and nature of index injury (penetrating versus non-penetrating): Black and Latinx males had a revisit cost pattern which decreased with increasing age, whereas white males had a revisit cost pattern which increased with age. Older ( ≥ 35 years) white male patients with ED revisits primarily presented with an initial blunt assault injury, while younger (<35 years) patients of color were more likely to have penetrating index injuries, with gunshot injuries being the most common. Differences between these groups suggest that for the studied region, at least two distinct intervention strategies would need to be developed to address ED revisits and corresponding costs following index violent injury, with the caveat that the findings likely have some reporting bias due to access to care challenges.
For the young Black and Latinx male subpopulation, understanding contributing factors to the relatively high incidence of index penetrating trauma could be profoundly impactful when designing intervention and prevention initiatives. Past research demonstrates that community violence intervention strategies, including hospital-based violence intervention programs (HVIPs), have cost benefits in reducing violent injury recidivism as well as ED revisits.
Critically, HVIPs often focus on clearly defined, relatively narrow patient populations as a means of maximizing program impact, usually in the context of resource limitations. The results of this study suggest that distinct patient subgroups exist beyond those typically served by HVIPs. These subgroups may benefit greatly from an additional intervention program. While extending support to a broader spectrum of age groups within existing interventions is a possibility, this is likely to be inadequate. Ideally, intervention strategies specific to each intersectional patient subgroup would have greater impact.
For white males ages 35 years and older who are victims of blunt assault, there are fewer (if any) hospital-based violence intervention programs. This population has been identified in another study to have higher rates of follow-up noncompliance, which could represent an opportunity for improvement and thereby cost reduction.
Although further needs assessments is necessary to determine the causal factors for increased costs among this group; mental health issues, substance abuse, or/and homelessness are possible explanations.
Critical limitations of the current study include sparse numbers within certain subgroups, lack of data from surrounding treatment centers, lack of information regarding the cause of revisits, and deficiencies associated with hospital cost as a measure. There were a sparse number of visits by patients identified as female and small numbers of visits by patients identified as having race/ethnicity other than Black, white, and Hispanic/Latinx. Additionally absent are patients who accessed another care facility after discharge from our medical center. Although shared medical records capture ED visits to most local hospitals, hospital cost data for these institutions is not made available and as such, these visits were not included in our analysis but have the potential to influence or explain certain cost trends. These “shared medical records” also do not include certain urgent care facilities or smaller private emergency rooms. For these reasons, our estimates of the broader societal cost burden of ED revisits may be conservative.
This analysis has an inherent selection bias for patients with access to the health care system. Critically, one cannot rule out the possibility that unique access to care issues may exist for patient groups with low observed numbers that impact return visits to the ED. Such issues might include real or perceived language barriers, primary child-care responsibilities, or perceived issues of citizenship status. While these results suggest that interventions with the most significant impact might be focused on younger patients identified as Black and male or older patients identified as white and male, care must be taken to avoid over-interpretation. Further investigation of subgroups with low visit numbers in this study, therefore, remains a warranted priority. Further, binary categorization of sex/gender is potentially a limitation that would be more apparent in subsequent subgroup evaluation. Issues of implicit bias and access to care impact patients differently who do not identify with the categories as defined.
The duration of the study is also an important limitation that will become increasingly important in follow-up studies going forward. As time increases, the impetus for return ED visits possibly shifts such that the component arguably driven by initial treatment failure will decrease while the component driven by violent injury recidivism may increase. It is worth noting that while our study chose 90 days after the initial violence-related trauma in an effort to focus on interventions related to treatment failure rather than recidivism, Cunningham et al.
and colleagues reported a high early re-assault rate amount violently injured youth immediately after discharge due to issues of retaliation or unresolved conflicts. As a database level investigation, this study was not able to distinguish between patients who returned to the ED with complications related to their index trauma and those who returned with a new violent injury or completely un-related complaints. Detailed chart review is ongoing to accurately classify this information, which will be crucial for informing future interventions. Preliminary data suggests the majority of visits were related to complications, and that wound care and pain control are responsible for the overwhelming majority of low cost revisits related to the index violent injury. While unrelated ED revisits do occur, this classification is not straightforward. Are revisits for exacerbation of a chronic medical issue shortly after discharge truly unrelated, or do they represent issues exacerbated by the index injury/pain, or gaps in receiving daily medications due to travel times, time intensive ED workups, unclear medical records, or a myriad of other social, structural, and economic factors?
Hospital cost for the purposes of our analysis represented an efficient way to assess patterns of visit complexity. While hospital cost is more consistent than hospital or professional charges, ultimately this number is set by the hospital system. Costs of even basic items such as certain medications may vary widely across different facilities. For this reason, only hospital costs within our single institution were compared. Despite similar social determinants being present, the findings of cost analysis will likely be unique in different hospital systems depending on the availability of individual community resources.
While some papers have standardized costs to “Medicare” costs, our patient population was generally too young to apply this method. Additionally, hospital cost does not provide information regarding reimbursement. While reimbursement data would have provided insight into whether these revisits resulted in a loss of revenue to the hospital system, this information was not obtainable.
The cost trends associated with ED revisits among subgroups of young Black men who are victims of penetrating trauma vs. older white men who are victims of blunt assault raise questions regarding the potential contributing factors driving the observed costs trends. These questions will facilitate a subsequent hypothesis-driven investigation. Further speculation would go beyond the limitations of the data generated in this study. Further study will also include analysis of revisit type with an associated cost analysis to further categorize reasons for “high” and “low” cost revisits within various sociodemographic groups and determine opportunities for intervention.
5. Conclusion
Hospital costs of return ED visits after index violent injury demonstrate patterns based on frequency, sex/gender, age, race/ethnicity, and mechanism of the initial injury. While cost analysis was a starting point to identify areas of intervention, this analysis is not expected to substitute as a community needs assessment. An important aspect of any follow-up studies will be seeking to understand the driving forces behind these cost trends with the intention of developing interventions for longitudinal evaluation. Identification of groups within a health care system as having high frequency or high-cost visits allows teams caring for these populations to provide financial justification to pilot interventions. Ultimately, the goal of identifying subgroups using a financial metric needs further study but has demonstrated feasibility as a starting point for directing limited resources.
Author contribution statement
Kara T Kleber, Ian Brown, Christy Adams contributed to the study design, interpretation, writing and critical revision. Angela Sardo and Jeffery Hoch were responsible for data collection, analysis, and interpretation. Nicole Kravitz-Wirtz and Shani Buggs contributed to data interpretation and critical revision.
Conflict of interest and source funding
No authors have any conflict of interest. The primary author was awarded a UC Davis ORG Resident Research Grant which was obtained for financial analysis.
References
Lee J.
Quraishi S.A.
Bhatnagar S.
Zafonte R.D.
Masiakos P.T.
The economic cost of firearm-related injuries in the United States from 2006 to 2010.
00450-0&id=gr1.jpg){kind=link}
00450-0&id=gr2.jpg){kind=link}
00450-0&id=gr3.jpg){kind=link}