Journal of Emergencies, Trauma, and Shock
Home About us Editors Ahead of Print Current Issue Archives Search Instructions Subscribe Advertise Login 
Users online:56   Print this pageEmail this pageSmall font sizeDefault font sizeIncrease font size   


 
 Table of Contents    
ORIGINAL ARTICLE  
Year : 2011  |  Volume : 4  |  Issue : 2  |  Page : 178-183
Mortality rates following trauma: The difference is night and day


1 Department of Orthopaedics, The NYU Hospital for Joint Diseases, NY, New York, USA
2 New York University School of Medicine, NY, New York, USA
3 Department of Orthopaedics, Dartmouth Medical School, Hanover, NH; The Multidisciplinary Clinical Research Center, Lebanon, NH, USA
4 Department of Orthopaedics, Dartmouth Medical School, Hanover, NH; Dartmouth Hitchcock Medical Center, Lebanon, NH, USA

Click here for correspondence address and email

Date of Submission28-Mar-2010
Date of Acceptance17-Sep-2010
Date of Web Publication18-Jun-2011
 

   Abstract 

Background : Although most medical centers are equipped for 24-h care, some "middle of the night" services may not be as robust as they are during daylight hours. This would have potential impact upon certain outcome measurements in trauma patients. The purpose of this paper was to assess the effect of patient arrival time at hospital emergency departments on in-hospital survival following trauma. Materials and Methods : Data of patients, 18 years of age or older, with no evidence that they were transferred to or from that center were obtained from the National Trauma Data Bank Version 7.0. Patients meeting the above criteria were excluded if there was no valid mortality status, arrival time information, injury severity score, or trauma center designation. The primary analyses investigated the association of arrival time and trauma center level on mortality. Relative risks of mortality versus patient arrival time and trauma level were determined after controlling for age, gender, race, comorbidities, injury, region of the country, and year of admission. Results : In total, 601,388 or 71.7% of the 838,284 eligible patients were retained. The overall in-hospital mortality rate was 4.7%. The 6 p.m. to 6 a.m. time period had a significantly higher adjusted relative risk for in-hospital mortality than the 6 a.m. to 6 p.m. time frame (ARR=1.18, P<0.0001). This pattern held across trauma center levels, but was the weakest at Level I and the strongest at Level III/IV centers (Level I: ARR=1.10, Level II: ARR=1.14, and combined Level III/IV: ARR=1.32, all P<0.0001). Conclusion : Hospital arrival between midnight and 6 a.m. was associated with a higher mortality rate than other times of the day. This relationship held true across all trauma center levels. This information may warrant a redistribution of hospital resources across all time periods of the day.

Keywords: Mortality rates, time of day, trauma center level designation

How to cite this article:
Egol KA, Tolisano AM, Spratt KF, Koval KJ. Mortality rates following trauma: The difference is night and day. J Emerg Trauma Shock 2011;4:178-83

How to cite this URL:
Egol KA, Tolisano AM, Spratt KF, Koval KJ. Mortality rates following trauma: The difference is night and day. J Emerg Trauma Shock [serial online] 2011 [cited 2017 Jun 29];4:178-83. Available from: http://www.onlinejets.org/text.asp?2011/4/2/178/82202



   Introduction Top


According to the Centers for Disease Control, trauma-related injuries remain a leading cause of death each year in the United States. [1] It is imperative, therefore, to evaluate mortality rates due to trauma in a myriad of ways in order to understand trends and ultimately deliver the highest quality medical care to patients. Much of the research on trauma patient outcomes has focused on patient load or the American College of Surgeons (ACS) trauma center designation. In general, Level I trauma centers, as compared to Levels II, III, and IV trauma centers, have demonstrated better outcomes when treating patients of equally severe injuries. [2],[3] A number of studies have looked at survival patterns and outcomes in patients presenting to the hospital during "off-hours," defined as either weekends or nights. [4],[5],[6],[7],[8] These studies have demonstrated worse outcomes and generally higher mortality rates at night and/or during weekends in patient cohorts defined by selected medical conditions, including acute myocardial infarction, ischemic and hemorrhagic stroke, and aortic aneurysm.

This study used a large trauma database to determine whether trauma patients presenting during the night would have higher in-hospital mortality than similar patients arriving during the day, after controlling for multiple covariates including trauma center level. It was hypothesized that trauma patients presenting during the night would have higher in-hospital mortality rates than similar patients arriving during the day.


   Materials and Methods Top


Data source

Data were obtained from the National Trauma Data Bank (NTDB) Version 7.0, [9] which includes 1,860,758 patients admitted to emergency departments in the United States from 2002 to 2006. This databank is managed through the ACS.

Target population

The target population for this study were patients 18 years of age or older with no evidence that they were transferred to or from that center. Patients transferred into or from another hospital were excluded from the population of interest because of a primary intent to evaluate the effect of arrival time on mortality. Patients meeting the above criteria were excluded if there was no valid: (1) in-hospital mortality status, (2) arrival time information, (3) injury severity score (ISS), or (4) ACS-designation of trauma level.

Additional variables of interest included patient age, race (White, Black, Hispanic, Other), gender, Deyo-Charlson comorbidity index, [10],[11] hospital region (Northeast, Midwest, South, West), and year patient was admitted. Rather than exclude patients from the sample because of missing values for any of these variables, all missing values were coded as not specified and included in the analyses.

Study design

This retrospective study was designed to evaluate the association of hospital arrival time and trauma center level with in-hospital mortality. It was hypothesized that patients who arrived in the evening (6 p.m. to 12 a.m.) or late night (12 a.m. to 6 a.m.) would have higher in-hospital mortality rates than patients arriving during the day (6 a.m. to 6 p.m.). It was further hypothesized that this association would be stronger in lower level trauma centers compared to higher level trauma centers, since lower level trauma centers were considered less likely to have the same capabilities in terms of equipment and staff availability.

Analysis plan

Preliminary analyses were performed to: (1) describe the effects of applying the exclusion criteria on the study sample by comparing personal and health demographics from the target population who did or did not remain in the study sample due to exclusion criteria; (2) describe the sample population; and (3) explore the relationships between arrival time and trauma center level and the covariate set.

The primary analyses investigated the association of arrival time and trauma center level on in-hospital mortality within the generalized linear model analysis framework. Poisson regression techniques (a generalized linear model specifying a Poisson distribution with a log link) were used to determine the relative risk (RR) and adjusted relative risk (ARR) of in-hospital mortality associated with patient arrival time and trauma level. Adjusting factors included patient personal and health demographics, ISS, and general factors (region of the country and year of admission). Preplanned contrasts were specified to examine any significant arrival time main effects. These were used to interpret differences in mortality rates associated with arrival time and trauma center level, as well as mortality rate differences associated with arrival time within each trauma center level (i.e., simple effect associated with an arrival time by trauma center level interaction).


   Results Top


Sample

Of the 1,860,758 patients in the NTDB 7.0, 1,050,618(56.5%) were 18 years or older and not coded as having been transferred to or from the hospital. An examination of the ACS-designated trauma level coding indicated 838,284 of the 1,050,618 patients (79.8%) had a valid trauma level. Of these, 59.5% were Level I, 37.0% Level II, 3.3% Level III, and 0.2% Level IV trauma centers. Level III and Level IV trauma centers were thus grouped together due to the relatively small sizes of each individually. In total, 601,388 or 71.7% of the eligible 838,284 patients were retained after excluding patients who did not have valid: mortality data (8591; 1.0%), time of arrival data (208,736; 24.9%), or ISS (19,569; 2.3%). Overall, 42.8% of the sample was excluded due to exclusion criteria.

Description of sample

Overall, 4.7% of the sample died during hospitalization. [Table 1] summarizes the relationship of the primary variables of interest and the categorical covariates relative to the mortality outcome, as well as the relationships between the primary variables of interest and the covariates. Approximately, 65% of patients arriving between midnight and 6 a.m. were seen at Level I trauma centers, compared to 32% at Level II trauma centers and 3% at Level III/IV trauma centers.
Table 1: Comparison of arrival time and trauma center level with each other and selected categorical patient and hospital demographics


Click here to view


Primary variables and mortality

Differences in mortality across arrival time ranged from 4.6 to 4.9%. For Trauma Center Levels, the mortality rates were highest at Level 1(5.0%) and lowest at Levels III/IV combined (3.1%). The Levels I, II, and III/IV mortality levels for patients seen at night were 5.1%, 4.6%, and 3.4%, compared to day, which were 5.0%, 4.3%, and 2.9%, respectively.

Covariates with mortality

[Table 2] summarizes the relationship of mortality with the continuous covariates. The average age was in the mid 40s, with those who died older than those who survived. As expected, the ISS was significantly higher (27.5 vs. 8.9) and the Deyo-Charlson comorbidity index was slightly higher in the in-hospital mortality group compared to the survival group.
Table 2: Relationship of continuous covariates with mortality


Click here to view


Primary variables and continuous covariates

[Table 3] summarizes the relationship of arrival time and trauma center level with continuous covariates. Those who arrived at night were significantly younger than those who arrived at other times of the day. Day arrivals were an average 5 years older than those arriving in the evening and more than 10 years older than those arriving late at night. ISS and Deyo-Charlson comorbidity scores were relatively similar across the four time intervals.
Table 3: Relationship of continuous covariates with arrival time and trauma center level


Click here to view


Relative to trauma center level, Level I patients were: (1) younger than patients seen at other levels; (2) had a higher ISS compared to the Level II and Level III/IV centers (10.3 vs. 9.3 and 8.1, respectively); and (3) had lower Deyo-Charlson comorbidity scores, which might be expected given the generally younger age groups seen at Level I centers.

Mortality risk

[Table 4] summarizes the relative risks for mortality compared to arrival time and trauma level unadjusted for any other factors and adjusting for patient demographics, health status, injury severity, and general factors of hospital region and year of admission.
Table 4: Summary of unadjusted and adjusted relative risks for mortality related to arrival time and level of the treating trauma center


Click here to view


The effects of adjustment on the relative risks strengthened the relationships between arrival time and mortality and weakened the relationships between trauma level and mortality. The ARRs demonstrated substantially higher mortality risks for patients arriving late at night compared to morning (ARR=1.21, P<0.0001), afternoon (ARR=1.26, P<0.0001) and somewhat higher compared to arrivals during the evening (ARR=1.09, P<0.0001). Similarly, ARRs for mortality were also higher in the evening hours when referenced to the morning and afternoon hours (ARR=1.11, P<0.004 and ARR=1.16, P<0.0001, respectively). Comparing 6 p.m.-6 a.m. to 6 a.m.-6 p.m. demonstrated significantly higher mortality rates in the former (ARR=1.18, P<0.0001). Furthermore, after adjustment, relative mortality rates were all significantly higher from 6 p.m. to 6 a.m. referenced to 6 a.m. to 6 p.m. for Level I (ARR=1.10, P<0.0001), Level II (ARR=1.14, P<0.0001), and for combined Level III/IV (ARR=1.32, P<0.0001). Thus, although mortality rates were lowest for combined Level III/IV trauma centers, the mortality risk for patients arriving between 6 p.m. and 6 a.m. was 32.0% greater than for patients arriving between 6 a.m. and 6 p.m.


   Discussion Top


This study investigated the relationships of arrival time, trauma center level, and the interaction between these two factors on in-hospital mortality after adjusting for injury severity and additional patient and hospital factors.

As hypothesized, after adjusting for injury severity, patients' health, and other patient and hospital demography, in-hospital mortality was higher when patients arrived at night, with late night arrival demonstrating the greatest in-hospital mortality risk.

The ARR of in-hospital mortality was directly related to trauma center level, with the greatest risk of in-hospital mortality at Level I trauma centers and the lowest RR at the Level III/IV trauma centers. These results were expected for unadjusted RR estimates, but were unexpected with risk adjustment. Although risk adjustments did reduce these values, suggesting little difference in mortality rates between Level I and II centers, in-hospital mortality risks were still substantially higher at Level I and II centers compared with Levels III/IV. This pattern of results may be due, in part, to the much lower overall mortality rates observed at Level III/IV trauma centers (3.1%) compared to Level I and II trauma centers (5.0% and 4.4%, respectively). Although these rates are not large in absolute difference, as a ratio, 5.04/3.11 and 4.11/3.11 represent rates that are 62% and 43% higher at Level I and Level II centers compared with Level III/IV centers. The case mix adjustment variables available may not have been sufficient to adequately "level the playing field" when comparing across trauma center levels without regard to a particular patient cohort. This is a potential limitation of this study as the literature has generally reported that Level I trauma centers have superior in-hospital survival rates compared to hospitals with lower level trauma center designations. [2],[3]

Furthermore, limitations associated with the use of the NTDB have been previously published. [12] In general, these limitations include the fact that this was a retrospective database study with all of the problems inherent with this methodology. Thus,

  1. The NTDB version 7.0 may not contain a number of important data elements that might better adjust for case mix across the Trauma Center Levels.
  2. Level III and IV trauma centers may be underrepresented in the database.
  3. Some of the variables of particular interest were notable not by their absence, but by their lack of response. For example, in this study a large number of patients were excluded because the data base did not include the hospital trauma center level and/or patient arrival time.
  4. The NTDB does not provide information about day of the week, which limited our ability to further evaluate in-hospital mortality based on weekday or weekend status.
  5. Data entry is voluntary at participating hospitals and, therefore, may not reflect the actual number of trauma patients seen at an institution. Further, the NTDB provides no weighting information that would allow the user to estimate national incidence rates.


The comparison between arrival time and trauma center level was evaluated by comparing night and day in-hospital mortality separately within each trauma center level. Here the results were, again, as expected. The adjustment factors strengthened the risk estimates, with nighttime in-hospital mortalities increased from level I (10%) to Level II (14%) to Level III/IV (32%).

It is important to note that patients presenting to the hospital throughout the entire 24-h period sustained relatively equally severe injuries. Although both the ISS and Deyo-Charlson comorbidity index scores demonstrated statistical significance across the four time blocks, the observed mean differences across the four time blocks (10.3, 9.7, 9.6, and 9.9) and (0.12, 0.15, 0.14, and 0.14), respectively, were not considered clinically important, especially since the variances associated with these means were also quite similar. This finding is similar to the two studies performed by Guly et al. and Carmody et al. in which statistically, but not clinically significant higher ISS correlated with weekend and nighttime admissions. [13],[14]

One potential cause for the disparity in mortality rates between night and day could be that physicians and staff are more tired at night and perhaps more likely to make an error. Disruption of circadian rhythms and its effect on performance have been reported in shift workers, with performance declines observed at night relative to day. [15] Yet, an independent study on emergency room physicians did not find any decline in functional outcomes for patients when nightshifts were compared to dayshifts, although a small increase in early mortality was detected. [16]

Another possibility may be that hospitals are simply understaffed at night. As discussed in the Carmody study, many hospitals have linearly arranged shifts, with equal numbers of house staff working at night as compared to day, when in fact more patients are admitted at night. [13] A study performed by Vaziri et al. demonstrated a sine wave appearance to admissions at hospital emergency departments, showing that there were fewest admissions around 6 a.m. and the most admissions around 7 p.m. [17] Appropriate allocation of resources should reflect these patterns.

As indicated throughout this paper, the large sample sizes afforded studies using large databases makes it important to distinguish between statistically significant and clinically important results. Although all of the variables in this study demonstrated statistically significant relationships with in-hospital mortality, many of the observed differences were relatively small.

At Level I trauma centers, patients arriving at night had a 5.1% in-hospital mortality rate and a RR of 1.1. This means that patients arriving at night had a 10% higher mortality rate than patients who arrived during the day. If there were no increased risk of in-hospital mortality at night, then per 1000 patients who arrive at night, the number of in-hospital deaths would be expected to decrease by 5 patients. By these same calculations, for Level II trauma centers-where night mortality rates were 4.6-the number of in-hospital deaths per 1000 patients seen at night would be expected to decrease by 6 patients. For Level III/IV trauma centers-where night mortality rates were 3.4%-the number of in-hospital deaths per 1000 patients seen at night would be expected to decrease by 8 patients. Some might argue that these different relative risks across trauma level centers suggest that higher-level trauma centers may allocate resources throughout the day more appropriately. Others might say that the consequences of 5 vs. 6 vs. 8 additional lives saved per 1000 patients seen at night is not a large enough benefit to consider making changes in-hospital staffing among the trauma centers to maximize allocation of resources throughout the 24-h day.

Strengths of this study include large patient numbers, the ability to limit the cohort to patients who were seen at the original trauma center where they arrived, and a large number of patient personal, health, and injury demographics useful for adjusting for case mix across trauma center levels.


   Conclusion Top


As hypothesized, in-hospital mortality rates were related to arrival time, with night arrivals having higher in-hospital mortality rates compared to day arrivals. However, in this patient population, defined simply as all adults 18 years or older who were treated at the hospital where they were originally seen, the magnitude of the higher risks for in-hospital mortality at night did not translate into substantially higher numbers of lives lost. This result is due, in part, to the generally low overall in-hospital mortality rates for this general patient population at all trauma centers. Further studies evaluating differential in-hospital mortality risk after traumas based on hospital arrival time could focus on more specific patient cohorts.

 
   References Top

1.Heron M, Hoyert DL, Murphy SL, Xu J, Kochanek KD, Tejada-Vera B. National Vital Statistics Reports. Atlanta (GA): Centers for Disease Control and Prevention; 2009. Available from: http://www.cdc.gov/nchs/data/nvsr/nvsr57/nvsr57_14.pdf [last cited on 2009 Sep 19].   Back to cited text no. 1
    
2.Haas B, Jurkovich GJ, Wang J, Rivara FP, Mackenzie EJ, Nathens AB. Survival advantage in trauma centers: Expeditious intervention or experience? J Am Coll Surg 2009;208:28-36.   Back to cited text no. 2
[PUBMED]  [FULLTEXT]  
3.Demetriades D, Martin M, Salim A, Rhee P, Brown C, Chan L. The effect of trauma center designation and trauma volume on outcome in specific severe injuries. Ann Surg 2005;242:512-9.  Back to cited text no. 3
[PUBMED]  [FULLTEXT]  
4.Magid DJ, Wang Y, Herrin J, McNamara RL, Bradley EH, Curtis JP, et al. Relationship between time of day, day of week, timeliness of reperfusion, and in-hospital mortality for patients with acute ST-segment elevation myocardial infarction. JAMA 2005;294:803-12.  Back to cited text no. 4
[PUBMED]  [FULLTEXT]  
5.Saposnik G, Baibergenova A, Bayer N, Hachinski V. Weekends: A dangerous time for having a stroke? Stroke 2007;38:1211-5.  Back to cited text no. 5
[PUBMED]  [FULLTEXT]  
6.Bell CM, Redelmeier DA. Mortality among patients admitted to hospitals on weekends as compared with weekdays. N Engl J Med 2001;345:663-8.  Back to cited text no. 6
[PUBMED]  [FULLTEXT]  
7.Peberdy MA, Ornato JP, Larkin GL, Braithwaite RS, Kashner TM, Carey SM, et al. Survival from in-hospital cardiac arrest during nights and weekends. JAMA 2008;299:785-92.  Back to cited text no. 7
[PUBMED]  [FULLTEXT]  
8.Reeves MJ, Smith E, Fonarow G, Hernandez A, Pan W, Schwamm LH. Off-hour admission and in-hospital stroke case fatality in the get with the guidelines-stroke program. Stroke 2009;40:569-76.  Back to cited text no. 8
    
9.National Trauma Data Bank Version 7.0. Chicago (IL): American College of Surgeons (US); 2007. Available from: http://www.ntdbdatacenter.com/Default.aspx [last cited 2007 Jul].  Back to cited text no. 9
    
10.Li B, Evans D, Faris P, Dean S, Quan H. Risk adjustment performance of Charlson and Elixhauser comorbidities in ICD-9 and ICD-10 administrative databases. BMC Health Serv Res 2008;8:12.  Back to cited text no. 10
[PUBMED]  [FULLTEXT]  
11.Lash TL, Mor V, Wieland D, Ferrucci L, Satariano W, Silliman RA. Methodology, Design, and analytic techniques to address measurement of comorbid disease. J Gerontol A Biol Sci Med Sci 2007;62:281-5.  Back to cited text no. 11
[PUBMED]  [FULLTEXT]  
12.Koval KJ, Tingey CW, Spratt KF. Are patients being transferred to level-I trauma centers for reasons other than medical necessity? J Bone Joint Surg Am 2006;88:2124-32.  Back to cited text no. 12
[PUBMED]    
13.Carmody IC, Romero J, Velmahos GC. Day for night: Should we staff a trauma center like a nightclub? Am Surg 2002;68:1048-51.  Back to cited text no. 13
[PUBMED]    
14.Guly HR, Leighton G, Woodford M, Bouamra O, Lecky F. The effect of working hours on outcome from major trauma. Emerg Med J 2006;23:276-80.  Back to cited text no. 14
[PUBMED]  [FULLTEXT]  
15.Santhi N, Horowitz TS, Duffy JF, Czeisler CA. Acute sleep deprivation and circadian misalignment associated with transition onto the first night of work impairs visual selective attention. PLoS ONE 2007;2:1233.  Back to cited text no. 15
    
16.Silbergleit R, Kronick SL, Philpott S, Lowell MJ, Wagner C. Quality of emergency care on the night shift. Acad Emerg Med 2006;13:325-30.  Back to cited text no. 16
[PUBMED]  [FULLTEXT]  
17.Vaziri K, Roland JC, Robinson L, Fakhry SM. Optimizing physician staffing and resource allocation: Sine-wave variation in hourly trauma admission volume. J Trauma 2007;62:610-4.  Back to cited text no. 17
[PUBMED]    

Top
Correspondence Address:
Anthony M Tolisano
New York University School of Medicine, NY, New York
USA
Login to access the Email id

Source of Support: None, Conflict of Interest: None


DOI: 10.4103/0974-2700.82202

Rights and Permissions



 
 
    Tables

  [Table 1], [Table 2], [Table 3], [Table 4]

This article has been cited by
1 Trauma care inside and outside business hours: comparison of process quality and outcome indicators in a German level-1 trauma center
Wolfgang Parsch,Markus Loibl,Uli Schmucker,Franz Hilber,Michael Nerlich,Antonio Ernstberger
Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine. 2014; 22(1)
[Pubmed] | [DOI]
2 External factors and the incidence of severe trauma: Time, date, season and moon
Carolina I.A. Pape-Köhler,Christian Simanski,Ulrike Nienaber,Rolf Lefering
Injury. 2014; 45: S93
[Pubmed] | [DOI]
3 Off-hours effect: does time of day or day of week influence mortality in patients with acute myocardial infarction?
Atsushi Sorita,Mohammad Hassan Murad,Henry H Ting
Interventional Cardiology. 2014; 6(4): 349
[Pubmed] | [DOI]
4 Effects of time of hospital admission on outcomes after severe traumatic brain injury in Austria
Walter Mauritz,Alexandra Brazinova,Marek Majdan,Veronika Rehorcikova,Johannes Leitgeb
Wiener klinische Wochenschrift. 2014; 126(9-10): 278
[Pubmed] | [DOI]



 

Top
  
 
  Search
 
  
    Similar in PUBMED
   Search Pubmed for
   Search in Google Scholar for
 Related articles
    Email Alert *
    Add to My List *
* Registration required (free)  


    Abstract
    Introduction
    Materials and Me...
    Results
    Discussion
    Conclusion
    References
    Article Tables

 Article Access Statistics
    Viewed1878    
    Printed131    
    Emailed1    
    PDF Downloaded10    
    Comments [Add]    
    Cited by others 4    

Recommend this journal