Discharge Patterns of Injury-related Hospitalizations with an Acute Wound in the United States

Chronic wounds affect over 6.5 million Americans per year or 1% of the US population.^1,2 The vast majority of these wounds are pressure ulcers, venous stasis ulcers, and diabetic foot ulcers.^1,3,4 These conditions have been well characterized epidemiologically and cause considerable morbidity, mortality, and financial burden.^5–12 Overall, chronic wounds account for 2% of all health care spending, with wound care products alone totaling more than $7 billion annually.^2,4,13 There is evidence to the significant clinical and economic benefit of having chronic and acute wounds managed by wound care specialists.^14–19 Multidisciplinary wound care teams have been shown to reduce the average wound healing time 2 to 4 times that of traditional medical-surgical treatment and save up to 93% of limbs that had been recommended for amputation due to nonhealing ulcers.^19–22
Previously published reports have provided recommendations on the ideal structure and multidisciplinary composition required for a successful wound care center;^14–19 however, no study has attempted to describe the acute wound patient population that could benefit from treatment by a wound care center. The majority of studies focused on specific chronic wound populations such as patients with diabetic foot ulcers, venous leg ulceration, or pressure ulcers.^{5,7–9,22–24}
The purpose of this study was to describe the demographic, medical care, and financial characteristics of injury-related hospitalizations with an acute wound. Patients with an injury diagnosis were selected because it is the antecedent event resulting in an acute wound. Inpatient hospitalization and referral to home health care were our primary interests because these patients typically have more extensive or complex wounds than routine discharges, thus would benefit from care by a multidisciplinary wound care team.^25,26 Identifying the populations most likely to have an acute complex wound, calculating the cost of these wounds, and describing wound care provider characteristics will inform the development of policy and prevention strategies.

Materials and Methods

Data source. Data from the Nationwide Inpatient Sample (NIS) from January 1, 2002 to December 31, 2002 were used. The NIS is a component of the Healthcare Cost and Utilization Project maintained by the Agency for Healthcare Research and Quality.²⁷ The NIS is one of the largest all-payer inpatient care databases in the US with more than 7 million hospitalizations in 2002. Since its implementation in 1988, the NIS has undergone only minor changes in the recorded variables, types of participating facilities, and number of hospitals sampled. In 2002, the NIS collected hospital discharge information on clinical treatment and resource utilization from 995 hospitals in 35 states, creating a 20% stratified sample of all US community hospitalizations.²⁷ The 2002 NIS included academic medical centers, public hospitals, children’s hospitals, and specialty hospitals such as obstetrics/gynecology, ear/nose/throat, and orthopedic institutions.²⁷ The 2002 NIS excluded short-term rehabilitation hospitals, long-term hospitals, psychiatric hospitals, and alcoholism/chemical dependency treatment centers. Hospitals were assigned to 5 strata for random selection based on geographical region, teaching status, ownership, number of beds, and rural/urban location. All data are discharge-level—individuals that were hospitalized multiple times will have multiple records in NIS. The 2002 NIS data were selected because 2002 was the most recent year with complete data at the time of the study.

Measures

Injury-related discharges. Injury was defined by ICD-9-CM codes 800–959 following the Centers for Disease Control and Prevention “Barell injury diagnosis matrix” guidelines.²⁸ A discharge was identified as “injury-related” if the primary or secondary diagnosis was an ICD-9-CM code between 800 and 959. Any patient “discharged to home health care” with a primary or secondary diagnosis of acute wound injury required additional health care resource utilization as a result of their injury compared to those “discharged routinely.” It is assumed that being “discharged to home healthcare” represents a step up in sophistication and cost of caring for these patients.
Type of wound. ICD-9-CM diagnosis codes were further used to categorize acute wounds into open wounds and burns. Researchers reviewed up to 15 diagnosis codes per discharge record to determine the presence and type of wound. Open wounds were identified by ICD-9-CM codes 870.0–897.9 for diagnosis of a cut, laceration, puncture wound, traumatic amputation, animal bite, or avulsion to any body part. These diagnosis codes exclude superficial injury, crushing, burn, puncture of internal organ, and open wounds incidental to a dislocation, fracture, internal injury, or intracranial injury. Burns had to cover 10% of total body surface or more and were identified by ICD-9-CM codes 948.10–948.99. Twenty-seven patients with an open wound and a burn were excluded from this study because this group was too small to analyze as a separate subgroup and multiple wound types could skew the analyses. Further, patients with a chronic wound were identified by ICD-9-CM codes 707.00–707.99 and excluded to isolate patients receiving care for an acute wound.
Comorbidity. The Agency for Healthcare Research and Quality coded 30 comorbidities in the Nationwide Inpatient Sample that identified coexisting medical conditions that were not directly related to the principal diagnosis and were likely to have originated prior to the hospital stay.^29,30 These comorbidities are determined by evaluating the diagnosis-related group (DRG) in effect at discharge, principal diagnosis, secondary diagnosis, and number of diagnoses for each record. This study investigated comorbidity information for diabetes without complications, diabetes with chronic complications, and obesity because these conditions are associated with an increased likelihood of chronic wound complications.^31–33 Hospitals in the state of Pennsylvania did not report comorbidity information with their hospital discharge records, thus 3,370 records in the authors’ dataset (5.8%) did not have comorbidity information.
Hospital characteristics. In the NIS, teaching hospital status was determined by the presence of an American Medical Association-approved residency program, membership in the Council of Teaching Hospitals, or a ratio of full-time equivalent interns and residents to beds of 0.25 or higher.³⁴ Bed-size categories are based on the number of short-term patient beds, geographical region of hospital, and teaching status (region is defined by the US Census Bureau). A hospital is defined as government- or privately-owned based on the ownership/control category reported to the American Hospital Association’s Annual Survey of Hospitals. A hospital is considered urban if it is located in a metropolitan statistical area and rural if it is located in a nonmetropolitan statistical area.
Statistical analysis. Data analyses were conducted using SAS (version 9.1, SAS Institute Inc., Cary, NC)³⁵ and SUDAAN (version 9.0.1, Research Triangle Institute, Research Triangle Park, NC)³⁶ statistical software to account for the weighting structures of the NIS. The SAS program was used to calculate all frequencies and percentages. The SUDAAN program was used to calculate all confidence intervals (CIs) for both national estimates and percentages. The NIS was designed as a stratified single-stage cluster of hospitals—all sampling probabilities are proportional to the number of hospitals in each stratum. Statisticians at the NIS provided discharge-level and hospital-level statistical weights that account for the complex survey design and sampling procedures to allow national estimation of all US community hospitals.³⁷ Researchers used discharge-level statistical weights to provide national estimates and to describe the patient and medical care characteristics of injury-related hospitalizations with an acute wound in the US. Hospital-level statistical weights were used to calculate percentages and confidence intervals for hospital characteristics.
The actual sample size is a statistically unweighted number and is specified when presented in the present study. All other frequencies and percentages are national estimates calculated using the statistical weights. National estimates and percentages of injury-related hospitalizations with a wound were calculated by age, gender, race, median household income for patient’s zip code, and type of wound. Injury rates per 10,000 were calculated for age, gender, race, and median household income using numerator-specific population estimates from the US Census Bureau.^38,39 Injury rates per 10,000 were calculated for wound types using the total population estimate as the denominator.³⁸ National estimates of routine and home health care discharges are calculated for each wound type. Percentages and 95% confidence intervals (95% CIs) are calculated for source of admission, length of stay, expected primary payer, and comorbidity variables for routine and home health care discharges by each wound type. Percentages and 95% CIs were calculated for hospital teaching status, ownership, rural/urban location, bed size, and region of US for routine and home health care discharges by wound type.
Multivariate logistic regression analyses were performed to model the relationship of several independent characteristics of medical care and of hospitals with the likelihood to be discharged to home health care compared to the likelihood to be discharged routinely, while controlling for confounding variables of age, gender, median household income. Specific measures that were significant in the univariate analyses and of particular interest to the study were explored, including medical care characteristics (primary payer and presence of a comorbidity) and hospital characteristics (teaching status, ownership, rural/urban location, bed size, and region of the US). Specifically, private insurance was used as a reference group for primary payer to reflect the medical standard of care. All other reference groups were selected because they had the lowest rates of discharge to home health care.
The Institutional Review Board of the Columbus Children’s Research Institute approved this study.

Results

Demographics. The NIS collected data on 7,853,982 hospitalizations between January 1, 2002 and December 31, 2002. Of these, 55,795 hospitalizations were directly related to an injury and had an acute wound diagnosis (Table 1). Based on these data, there were an estimated 272,268 (95% CI = 272,201–272,335) injury-related hospitalizations with an acute wound in 2002. This estimate is derived from statistical weighting of the 55,795 discharges. These injury-related hospitalizations with an acute wound represented 14.8% of the 1,835,654 total estimated injury-related hospitalizations. The subject selection process is summarized in Figure 1.
The distribution of patients’ age, gender, race, median household income for patient’s zip code, and type of wound is shown in Table 1. The number of injury-related hospitalizations with an acute wound was the highest in young adulthood (n =10,683 for persons aged 20–29 years). Yet, the rate of such hospitalizations showed a bimodal peak among young adults and all persons age 70 years and above (up to 45.2 injury-related hospitalizations per 10,000 persons age 90 years and older). Men accounted for more than twice the number and rate of injury-related hospitalizations with a wound as women did (n =183,994 versus n = 88,241; 13.0 versus 6.0 injury-related hospitalizations per 10,000 persons, respectively). For persons with recorded race (72.7%), white patients comprised 61.3% of injury-related hospitalizations with a wound. However, minority populations had much higher rates of such hospitalizations (23.9 per 10,000 non-black minority persons, 8.9 per 10,000 black persons, and 5.3 per 10,000 white persons). By income, patients with a median household income for their zip code of $45,000 or greater represented almost half (42.7%) of all injury-related hospitalizations with an acute wound. Yet, the rate of injury-related hospitalization with an acute wound was highest among persons with a median household income of $25,000–$44,999 for their zip code (17.9 per 10,000 persons from $25,000–$34,999 median household income for their zip code, and 23.5 per 10,000 for $35,000–$44,999).
Characteristics of medical care. For each type of wound, the total number of hospitalizations, source of admission, length of stay, expected primary payer, and comorbidity information is presented for routine discharges and for home health care discharges in Table 2. Open wounds comprised 96.9% of all wounds and 5.8% were discharged to home health care. Burns comprised 3.1% of all wounds and 17.8% of burn patients were discharged to home health care. For open wounds and burns, patients discharged to home health care were significantly more likely to have Medicare as their expected primary payer and to be admitted from another hospital or facility than patients discharged routinely were. Overall, government health care programs and self-pay/other insurance types were the most common payers for patients with an acute wound.
Patients with an open wound that were discharged to home health care were significantly more likely to have uncomplicated diabetes, diabetes with chronic complications, and/or obesity compared to those patients discharged routinely. For open wounds, 8.3% of patients discharged to home health care had uncomplicated diabetes compared to 4.1% of routine discharges; 2.2% of these patients had diabetes with chronic complications compared to 0.5% of routine discharges; and 2.2% of these patients were obese compared to 1.2% of routine discharges. Patients with an open wound who were discharged to home health care were significantly more likely to have an inpatient stay of 8 days or longer than patients who were discharged routinely (29.4% of discharges to home health care compared to 9.2% of routine discharges). Further, a significantly higher percentage of patients with diabetes were discharged to home health care compared to patients with no diabetes, for both open wounds and burns.
Characteristics of hospitals. Hospital teaching status, ownership, rural/urban location, bed size, and geographic region by type of wound and type of discharge are shown in Table 3. Significantly more burn patients were discharged to home health care at teaching and government-owned hospitals compared to nonteaching and privately owned hospitals (18.7% versus 12.4% of patents were discharged to home health care at teaching versus nonteaching hospitals; 18.6% versus 10.2% of patients at government-owned versus privately-owned hospitals). For burn patients, the percentage of patients discharged to home health care increased as bed size increased (7.7% of patients discharged to home health care at small hospitals, 14.0% at medium hospitals, and 20.0% at large hospitals). Hospitals in the Western US discharged significantly less burn and open wound patients to home health care than hospitals in all other regions of the US (4.3% of burn patients were discharged to home health care, compared to 24.0% of burn patients in the Northeast; 4.7% of open wound patients, compared to 6.8% in the Northeast). Further, hospitals in the Northeast and South discharged significantly more burn patients to home health care than hospitals in the Midwest (24.0% and 22.4% of burn patients respectively, compared to 14.4% of burn patients hospitalized in the Midwest). There were no other statistically significant trends with regard to hospital characteristics and wound types.
Results of logistic regression analyses. Multivariate logistic regression was used to control for the confounding effects of demographic variables while further exploring risk factors for discharges to home health care compared to routine discharges (Table 4). For open wound types, Medicare patients were significantly more likely to be discharged to home health care than privately insured patients were (odds ratio [OR] = 1.35, 95% CI = 1.17–1.56 with private insurance as the reference group). Diabetes with (OR = 2.81, 95% CI = 2.26–3.32) or without (OR = 1.26, 95% CI = 1.07–1.48) complications significantly increased the likelihood to be discharged to home health care for patients with an open wound. Obesity also significantly increased the likelihood to be discharged to home health care for patients with open wounds (OR = 1.41, 95% CI = 1.03–1.93). The likelihood of being discharged to home health care was also significantly increased with increasing age and if the patient was a woman.
By hospital characteristics, open wound patients treated at urban hospitals were significantly more likely to be discharged to home health care than those treated at rural hospitals (OR = 1.43). Open wound patients treated at hospitals in the West were significantly less likely to be discharged to home health care than those treated at any hospital in the Northeast, South, or Midwest (OR = 1.47, 1.36, and 1.18, respectively). For burn patients, the likelihood to be discharged to home health care increased with increasing age (OR = up to 8.29 for patients age 80–89 years) and with treatment at a hospital in the Midwest or South (OR = 2.63 and 5.59, respectively).

Discussion

Major findings. Overall, there were an estimated 282,471 injury-related hospitalizations with an acute wound in the US in 2002. The highest rate of such hospitalizations occurred among the elderly. Patients with an open wound represented the vast majority of these hospitalizations, but only a small percentage (5.8%) were discharged to home health care while almost one fifth of burn patients were discharged to home health care. The present study is one of the first to describe that obese patients have an increased likelihood to be discharged to home health care after an injury-related hospitalization with an acute wound compared to non-obese persons.^32,33 Discharges to home health care were of primary interest because these patients typically have more extensive or complex wounds than routine discharges, and thus would benefit from care by a multidisciplinary wound care team.^25,26 Analysis of discharge trends by patient, medical care, and hospital characteristics revealed important findings that can inform prevention and intervention efforts to reduce medical costs and improve patient care of acute wounds.
By patient characteristics, more than half of all injury-related hospitalizations with a wound occurred among men and persons age 10–39 years. However, the rate of such hospitalizations showed a bimodal peak in young adulthood and in old age. The first peak at 20–39 years, and the overall predominance of men, is likely related to well-documented higher rates of injury among young adults and men.^40,41 The second peak of injury-related hospitalizations with an acute wound rapidly escalates after age 70. Further, the likelihood of patients to be discharged to home health care also increased with increased age, even after controlling for demographics, comorbidities, and hospital characteristics. The increased rate of acute wounds in older adults hospitalized for an injury is significant because research has shown complications from an injury are a frequent cause of health decline and change in social situation among patients 65 years or older.^42,43 Acute wound, subsequent infection, or delayed healing may be important complications from injury in the elderly population. This finding is even more significant because the NIS likely underreports acute wounds after injury in older adults because the CDC-recommended ICD-9 CM codes used in this study specifically exclude open wounds secondary to a dislocation, fracture, internal injury, or intracranial injury. This age group has a high prevalence of fractures and other trauma. Age-related declines in the rate of healing and in defenses against microorganisms make the elderly patient an ideal candidate for treatment by a multidisciplinary wound care team.⁴² This is one of the first studies to document an increased rate of acute wounds after an injury in the elderly population.
By medical care characteristics, all patients discharged to home health care had higher percentages of uncomplicated diabetes, diabetes with chronic complications, and obesity than those discharged routinely. Specifically, patients with an open wound and chronic complications from diabetes were almost 3 times more likely to be discharged to home health care than routinely discharged. Patients with an open wound and either uncomplicated diabetes or obesity were almost 1.5 times more likely to be discharged to home health care than discharged routinely. All of these relationships were statistically significant. These findings are consistent with previous research that has shown patients with diabetes have an increased risk for wound complications and an increased likelihood that acute wounds will become chronic.⁶ Thus, the data in the present study appear valid and representative. Studies documenting a higher incidence of wound complications among obese persons are limited to case reports or to data from a single hospital.^32,33 This study is one of the first to describe the increased risk of acute wound complications due to obesity in a nationally representative inpatient sample.
By hospital characteristics, patients with an acute wound were significantly more likely to be discharged to home health care from urban hospitals and from hospitals outside the Western US even after controlling for demographics and comorbidities. There were no statistically significant differences by bed size, teaching status, or ownership for patients with an open wound. As the size of hospital increased, the percentages of burn patients being discharged to home health care also increased. Only 42% of routine discharges and 40.6% of home health care referrals had private insurance. Consistent with a higher rate of home health care discharges among elderly, Medicare was significantly more likely to be a payer for home health care discharges than private insurance. Internationally, most multidisciplinary wound centers are affiliated with large teaching hospitals.^14,16,21,44 These data describe the distribution of home health care discharges and can inform future planning for wound care services.
Identifying those persons most likely to require additional health care resources to manage their wounds helped indicate which group may benefit from a more aggressive approach to close a wound as soon as possible. For those patients most likely to require home health care (eg, obese, patients with diabetes, elderly, or prolonged hospital stay), early consideration may be given to surgical versus nonsurgical wound management, to avoid the consequences of missed work and the expense of home health care treatments while waiting for these wounds to heal. Since most injuries occur in working age men, it is critical to get them recovered from their injuries and back into the work force as quickly as possible. For any person with an acute wounds complex enough to require home health care, wound care centers may represent the best medical management option to achieve the goal of rapid recovery. Their efficacy in managing chronic wounds has been well documented and, given the evidence in the medical literature supporting the success of high volume centers in managing specific disease conditions, it is not unreasonable to expect efficacy in managing acute wounds. Additional research will need to be done to assess the cost effectiveness of this strategy, but shortening the duration of wound management by even a few days may be sufficient to justify this approach, given the high costs of home health care and the opportunity costs (losses due to time off work).
This study also provides important information about the cost of caring for these patients, which has important policy ramifications. Acute open wounds that require additional health care resource utilization via referral to home health care impose a significant economic burden on society and the healthcare system in the US. Only 40.6 % of patients with open wounds referred to home health care had private insurance and for burn patients only 34.9% of home health care referrals had private insurance. Taxpayer subsidized government insurance paid for 48.9% of acute open wounds and 56.5% of burn wound patients referred to home health care. These patients were significantly more likely to receive their care in a public teaching institution typically in a large urban setting and for burn patients to have been treated in a burn center. The increased complexity of care that must be provided to these patients and the poor payor mix of this population put a severe economic strain on both the institutions and physicians. When only 40% have insurance, the burden of providing care to the remaining 60% that provide little to no reimbursement is not sustainable and could reach critical proportions if Medicare reduces reimbursement even further.

Limitations

Originally, this study intended to identify how many patients with acute wounds require complex care so that it could be combined with the known information on chronic wounds and therefore provide an indication of the overall need for wound care management. Adhering to the CDC definition of injuries, patients with acute wounds secondary to complications of care and adverse effects were excluded. This population may also benefit from wound care expertise and the size of this population is likely quite large because it includes patients with open wounds from surgical procedures. Also, since the NIS tracks only hospitalization data, the authors’ findings may not reflect patterns of home health care referral in outpatient physician offices, emergency departments, self-referral, or other sources. Thus, the results underreport the true incidence of acute complex wounds. Further, the 2002 NIS collected data from a network of 995 hospitals and has potential data quality issues as discussed by the AHRQ and by independent researchers.^45,46 Despite these limitations, the NIS is one of the largest all-payer inpatient databases and allows estimation of the national incidence and characteristics of injury-related hospitalizations with an acute wound that are discharged to home health care. This is one of the first studies to document an increased rate of acute wounds with injury-related hospitalizations among the elderly and an increased likelihood to require home health care after injury among obese persons.

Conclusion
This study is one of the first to use to nationally representative hospital discharge data to describe the demographic, medical care, and hospital characteristics of injury-related hospitalizations with an acute wound. These data describe an increased rate of acute wounds after injury among the elderly and an increased health care resource allocation for the treatment of acute wounds in obese population.

This manuscript was supported in part by an Alpha Omega Alpha Student Research Fellowship.
Dr. Huiyun Xiang was supported in part by a research grant from the Centers for Disease Control and Prevention (CDC), the United States Department of Health and Human Services (Grant #: R49CE00241-01, PI: Huiyun Xiang). The views expressed here are those of the authors and do not necessarily reflect the official views of the CDC.