Abstract
Introduction
This study aimed to determine the correlation between trauma patient co-morbidities, insurance status and final disposition.
Methods
We evaluated the impact of co-morbid conditions and insurance status on trauma patient outcomes utilising the National Trauma Data Bank. Paired T-tests were used to determine significance (P < 0.05).
Results
Patients who were discharged to home had the highest rate of private insurance (27%, P < 0.01), while those sent to a facility had the highest rate of public insurance such as Medicare or Medicaid (50%, P < 0.01) and the lowest rate of no insurance (5%, P < 0.01). Patients who died had the lowest rate of private insurance (17%, P < 0.01) and 15% had no insurance (P < 0.01). Complications and co-morbidities were significantly more common in patients who died compared to those sent to another facility or home.
Conclusion
Improving access to private insurance is associated with improved trauma-related morbidity and mortality.
Keywords
Introduction
Trauma is the leading cause of death in patients aged 1–44 years and is responsible for more than 5.8 million deaths worldwide annually (CDC 2012). The World Health Organisation (WHO) projects that by 2020, injury will surpass infectious diseases as the leading cause of death worldwide (NIH 2012). For continual improvement in trauma care, it is vital to assess patient outcomes in a variety of contexts to ascertain factors associated with optimal outcome in trauma patients (Haider et al., 2008; Hollis et al., 2006; McGwin et al., 2004; Salim et al., 2010; Shoko et al., 2010; Rosen et al., 2009a; Wutzler et al., 2009).
Pre-existing medical conditions as well as insurance status have been reported to affect morbidity and mortality rates in both the general and trauma populations in single institution studies (Greene et al., 2010; Haider et al., 2008; Hollis et al., 2006; McGwin et al., 2004; Salim et al., 2010; Shoko et al., 2010; Wutzler et al., 2009).
The National Trauma Data Bank (NTDB) is the largest collection of US trauma-related statistical data in the world and includes detailed information on demographics, trauma-related primary outcomes, procedures, and complications. While a number of studies have been done to evaluate the role of various co-morbid conditions such as alcohol, heart failure, renal dysfunction and age in the trauma population, there is a lack of information regarding the role of insurance on trauma patient outcomes (Hollis et al., 2006; McGwin et al., 2004; Shoko et al., 2010; Wutzler et al., 2009).
This study aimed to determine the correlation between trauma patient co-morbidities, insurance status and final disposition.
Methods
The 2008 NTDB was used with permission from the American College of Surgeons. The NTDB identifies 24 comorbid conditions, 26 complications, 10 insurance types, and nine dispositions. We evaluated the impact of each of the 24 comorbid conditions, 26 complications, and insurance status (public, private, none) on patient disposition (home, other facility, death).
A total of 140,333 incidents were compiled from 90 Level 1 and 2 trauma centers throughout the United States. For this subset of patients, comprehensive horizontal data were recorded including demographics, vital signs, injury severity score (ISS), mechanism of injury, diagnosis, procedures, comorbidities, complications, length of stay and disposition. The mechanism of injury is reported using the ICD-9-CM code and the data transformed into blunt and penetrating injuries by using the CDC injury matrix. The entire 2008 NTDB was used for analysis, and all patients were automatically included. Inherent within the NTDB is a selection bias in that patients who die prior to arriving at the hospital are not included; further, as the data are compiled on a voluntary basis by hospitals, not all patients at the 90 Level 1 and 2 trauma centers around the United States may be included. To account for missing data, data analysis was done with and without imputation for missing values; no statistically significant difference was identified.
The categories of co-morbidity used by the National Trauma Data Bank.
The categories of complications utilised by the National Trauma Data Bank.
ICD-9 procedure codes were identified imaging tests, vascular surgery procedures, and non-vascular procedures. Patient disposition is grouped into nine categories in the NTDB; these data were grouped into patients discharged to home (with or without home health, and patients who left against medical advice), patients discharged to another facility (including acute care hospitals, intermediate care facilities, skilled nursing facilities, long-term care facilities, and hospice), and patients who died.
Data analysis was completed using a combination of univariate and multivariate statistical tests. Unpaired T-tests were used to determine significance (P < 0.05).
Results
Patients who were discharged to home had the highest rate of private insurance (27%, P < 0.01), while those sent to a facility had the highest rate of public insurance such as Medicare or Medicaid (50%, P < 0.01) and the lowest rate of no insurance (5%, P < 0.01). Patients who died had the lowest rate of private insurance (17%, P < 0.01) and 15% had no insurance (P < 0.01).
Comorbidities that were significantly more common in patients who died compared to those discharged to another facility or to home included a history of cerebrovascular accident (CVA) (odds ratio (OR) 3.1), hypertension (OR 1.2), and impaired sensorium (OR 1.4). Patients who were discharged to a facility instead of home were significantly more likely to have bleeding disorders, cardiovascular disease, diabetes, obesity, respiratory disease, and steroid use.
Final disposition of patients.
ISS, injury severity score; ICU, intensive care unit; LOS, length of stay; SD, standard deviation.
Discussion
Numerous studies have been conducted to evaluate the effect of comorbidities, insurance status, and complications in the trauma population (CDC 2012; Greene et al., 2010; Haider et al., 2008; Hollis et al., 2006; McGwin et al., 2004; Salim et al., 2010; Shoko et al., 2010; Wutzler et al., 2009).
A 1999 comprehensive review on trauma and co-morbidity studies reported that increasing age, male sex, age >40, evidence of co-morbidities, alcohol use, and smoking status were all significant factors in increasing morbidity and mortality in the trauma patient (Wardle, 1999). All of these factors were also found to be contributors toward increased length of stay, ICU stay, and time on ventilator.
In the older population, co-morbidities only had an effect on mortality in patients aged 50–64 years who sustained minimal trauma (ISS < 16), while patients more than 65 years of age saw no difference in mortality, regardless of pre-existing medical conditions (McGwin et al., 2004). Co-morbidities tended to have the greatest impact on minimally injured patients when compared to the severely injured. There was no difference in the incidence of co-morbid conditions as a function of insurance status.
Large trauma database studies have supported published single institutional data with regard to patient outcomes relative to comorbidities (Hollis et al., 2006; Shoko et al., 2010; Wutzler et al., 2009). An analysis of 65,743 patients from the United Kingdom Trauma Network showed that with the presence of pre-existing medical conditions, the odds ratio for mortality in the trauma population was 2.0 for moderately severe injuries, and 5.9 for mildly severe injuries (Hollis et al., 2006). Another analysis of 20,257 patients from the Japan Trauma Data Bank provided additional insight, specifically identifying cirrhosis, active cancer, COPD, haematological disorders, use of anticoagulation, dementia, and mental retardation as factors associated with an increase in hospital mortality in trauma patients (Shoko et al., 2010).
Our study identified CVA, hypertension, and impaired sensorium to be most commonly reported in patients who died while in-patients discharged to a facility instead of home were noted to have bleeding disorders, cardiovascular disease, diabetes, obesity, respiratory disease and steroid use.
Insurance status is another factor that has been reportedly correlated with patient disposition (Greene et al., 2010; Haider et al., 2008; Salim et al., 2010). A 1993 national cohort study analysing 4694 patients utilising the first National Health and Nutrition Examination Survey found that patients with no health insurance have an increased mortality risk across all sociodemographic groups (Franks et al., 1993). The 2009 third National Health and Nutrition Examination Survey supported these findings with a similar strength of association, Wilper et al., 2009).
In the trauma population, a single institution study at a Level 1 trauma center reported that despite being younger and minimally injured, uninsured patients had a higher mortality rate than other trauma patients (Salim et al., 2010). Database studies with large numbers of trauma patients have supported this finding (Greene et al., 2010; Haider et al., 2008; Salim et al., 2010). Discrepancies in outcomes of trauma patients in the context of insurance in the paediatric population level have yielded similar results (Rosen et al., 2009b).
Our results support published results regarding the association between uninsured patients and disposition (Greene et al., 2010; Haider et al., 2008; Salim et al., 2010). In our study, patients who were discharged to home had the highest rate of private insurance while those sent to a facility had the highest rate of public insurance such as Medicare or Medicaid. Patients with no insurance had the highest mortality rates. Reasons for this association could include preferential treatment, the fact that uninsured patients may seek care later, receive care initially at an outside hospital and then are transferred, all resulting in a delay in definitive care (Rosen et al., 2009a). Institutions that treat the uninsured population may have fewer resources and may not be as technologically advanced (Rosen et al., 2009a).
Limitations
This study is limited by its utilisation of a national database where under-reporting and incomplete ICD-9 coding may underestimate the overall incidence of trauma.
Conclusion
Previous studies from patient populations suffering from trauma have shown that co-morbidities and insurance status can have significant effects on outcomes in these patients. Assessing what these effects are in the context of discharge (home, facility, death) provides insight into the discrepancy between mortality rates and may lead to the development of improved clinical assessment tools and trauma patient management. Patients at highest risk of death after trauma are those with significant comorbidities such as hypertension or cerebrovascular disease, and lack of private insurance or no insurance at all. Patients with significant cardiovascular, respiratory, or endocrine disease tend to have a longer LOS and ultimately be discharged to another facility. As expected, patients with the most complications tend to have the poorest outcomes. The findings that high blood pressure and cerebrovascular disease being related to mortality from trauma implies that better control of these parameters may improve short-term outcome. Improving access to insurance for all patients may also significantly improve trauma-related mortality.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.