Code 873.65 is the International Classification of Diseases, Ninth Revision, Clinical Modification code for open wound of palate, without mention of complication; and 873.75, open wound of palate, complicated. CPT indicates Current Procedural Terminology; ED, emergency department.
eTable. HCUP Demographic Characteristics for Those Diagnosed With Primary Palate Trauma by Year in the National Emergency Department Sample (NEDS)
Customize your JAMA Network experience by selecting one or more topics from the list below.
Smith A, Ray M, Chaiet S. Primary Palate Trauma in Patients Presenting to US Emergency Departments, 2006-2010. JAMA Otolaryngol Head Neck Surg. 2018;144(3):244–251. doi:10.1001/jamaoto.2017.3071
What is the current incidence of and practice pattern for palate trauma in the emergency department?
In this database analysis of the Nationwide Emergency Department Sample, among 22 094 patients with primary palate trauma, total hospital visits decreased from 1.58 per 100 000 people in 2006 to 1.26 per 100 000 people in 2010; most patients were aged 18 years or younger and discharged without undergoing head and neck imaging, while palate repair and mortality were rare. Factors associated with admission included complicated palate trauma, male sex, region, patient’s residence, and codiagnosis status; factors associated with imaging included patient’s residence, payment, and region.
Although it is often suggested in the otolaryngology literature to perform imaging, primary palate trauma usually results in routine discharge without imaging.
The sequelae of palate trauma vary from minimal discomfort to major neurovascular injury. Infrequency of palate trauma and clinician unfamiliarity with the disease process may lead to variation in evaluation, treatment, and disposition in the emergency department (ED).
To measure the incidence of primary palate trauma visits to US emergency departments with analysis of demographics, disposition, and repair and to determine frequency and factors associated with head and neck imaging.
Design, Setting, and Participants
A retrospective analysis using the Healthcare Cost and Utilization Project Nationwide Emergency Department Sample was performed of 22 094 patients presenting to US emergency departments from 2006 to 2010 with a primary diagnosis of palate trauma. Data analysis was conducted from March 29, 2016, to November 18, 2017.
Main Outcomes and Measures
National estimates of palate trauma were calculated from weights available within the database. Palate repair was identified by International Classification of Diseases, Ninth Revision, Clinical Modification procedure codes. Imaging was calculated from reliable Current Procedural Terminology coding facilities identified using a previously published method. Logistic models were calculated to identify clinical associations for admission, imaging, and palate repair.
A total of 22 094 patients (13 967 male and 8121 female patients, 6 missing data on sex; median age, 2.8 years [interquartile range, 1.1-6.1 years]) with primary palate trauma presented to US emergency departments during the study period. Total hospital visits decreased from 4715 (1.58 per 100 000 people) to 3915 (1.26 per 100 000 people) during the 5-year study period. A total of 19 819 patients (89.7%) had routine discharge from the hospital, while palate repair (965 [4.4%]) and mortality (34 [0.2%]) were rare. Complicated palate trauma (odds ratio [OR], 5.32; 95% CI, 3.10-9.15), male sex (OR, 1.57; 95% CI, 1.11-2.21), codiagnosis status (OR, 2.75; 95% CI, 1.84-4.12), and residence in the Northeast vs South (OR, 2.73; 95% CI, 1.11-6.71) increased the likelihood of admission, which was infrequent (1027 patients [4.6%]). After restriction to reliable Current Procedural Terminology coding facilities, head and neck imaging occurred in 823 of 6897 patients (11.9%). Factors associated with head and neck imaging included living in a medium vs large metropolitan area (OR 1.62; 95% CI, 1.04-2.55), while living in the Midwest vs South region (OR, 0.43; 95% CI, 0.25-0.74) had a negative association with imaging.
Conclusions and Relevance
Although it is often suggested in the otolaryngology literature to perform imaging, primary palate trauma usually results in a routine discharge home without imaging or repair. Imaging frequency should be noted since palate trauma could have life-threatening neurovascular sequelae, which presents an opportunity to define and promote optimal management for potential neurologic sequelae in the patients who were not imaged.
Impalement injuries of the oral cavity and, more specifically, to the palate are an uncommon presenting symptom to the emergency department (ED). These injuries occur most commonly in the pediatric population, when children fall onto an object previously placed within the oral cavity; the most common objects noted within the literature are sticks, pencils, and cylindrical objects.1 The infrequent incidence of palate trauma may lead to clinician unfamiliarity and variation in evaluation, treatment, and disposition.
Furthermore, the effect of palate trauma can vary from minimal discomfort to major neurovascular injury.2-8 Patients with palate trauma should be evaluated for neurovascular injury with a complete neurologic examination.1,2,9-13 However, patients with penetrating palate trauma can display symptoms of neurovascular injury with delayed presentation up to 72 hours, which may not be diagnosed during overnight observation. The mechanism involves injury to the intima of the carotid artery with formation and propagation of a thrombus, which can take several days.2 Computed tomography (CT) with angiography (CTA) performed in the ED can help detect vascular injury but may not be necessary in all patients with palate trauma.12 These injuries often heal well without surgical repair; only 6% to 11% of patients with such injuries have historically undergone operative intervention, with size and complexity as the main factors for repair.1,14,15
To define and promote optimal management of palate trauma, it is necessary to first understand the current epidemiologic and practice patterns. This study was undertaken to measure the incidence of primary palate trauma visits to EDs in the United States, with analysis of demographics, disposition, and repair in the ED. Furthermore, as it is often suggested in the otolaryngology literature to obtain imaging of palate trauma,16-18 we also sought to determine the frequency of and factors associated with head and neck imaging.
We examined the years 2006 to 2010 of the Healthcare Cost and Utilization Project Nationwide Emergency Department Sample, which contains a 20% stratified sample of US ED discharges gathered and managed by the Agency for Healthcare Research and Quality and represents the largest all-payer ED data set in the United States. The Nationwide Emergency Department Sample varies slightly from year to year, containing 955 to 980 hospitals in 24 to 29 states in 2006-2010.19 University of Tennessee Health Science Center institutional review board exemption was obtained for this study, as it used information from a deidentified public database.
To assess the occurrence of primary palate trauma, we considered patients with a primary diagnosis of palate trauma, identified by International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) code 873.65 (open wound of palate, without mention of complication) or 873.75 (open wound of palate, complicated). The ICD-9-CM codes for palate trauma were also evaluated when coded as a secondary diagnosis. This process yielded very few secondary palate traumas (n = 836), which were not included in further analysis. Patient codiagnoses were defined as the 10 most frequent secondary diagnosis codes associated with primary palate trauma, excluding causes of palate trauma. Secondary diagnosis is defined as an additional diagnosis other than the primary diagnosis of palate trauma. Palate repair was identified by ICD-9-CM procedure codes 27.61 (suture of laceration of palate) or 27.69 (other plastic repair of palate). We further explored the rate of potential neurovascular sequelae based on their representative ICD-9-CM codes (433.xx, 434.xx, 436.xx, 443.xx, 444.xx, and 900.xx) appearing with primary palate trauma codes; as these codes were found to be cumulatively fewer than 10, they were not reportable.
Next, we evaluated the frequency of imaging in the ED. To overcome errors of omission in coding, we evaluated the quality of coding practices of individual EDs using a previously published method20: first, by excluding EDs without any Current Procedural Terminology (CPT) codes reported for any entry in the Nationwide Emergency Department Sample and, second, by excluding EDs with unreliable CPT coding procedures. Unreliable CPT coding procedures of an individual ED was defined using the proportion of patients with a CPT code for a forearm radiograph (CPT codes 73090, 73080, and 73110) who primarily received a diagnosis of a broken arm (ICD-9-CM code category 813.xx); unreliable EDs were excluded if this proportion fell below the median proportion of all EDs reporting CPT codes (Figure).20 After restricting the data by excluding EDs with these 2 criteria from the analysis, we evaluated the frequency of imaging of the head and neck with CPT codes (neck soft tissue, 70360; CT scans of the head and neck area, 70450, 70460, 70470, 70480, 70481, 70482, 70486, 70487, 70488, 70490, 70491, 70492, 70496, and 70498; and magnetic resonance imaging scans of the head and neck area, 70540, 70542, 70543, 70544, 70545, 70546, 70547, 70548, 70549, 70551, 70552, 70553, and 76380).
Data analysis was conducted from March 29, 2016, to November 18, 2017. Distributions of demographic and hospital characteristics for each cohort (the original cohort and the restricted cohort) were assessed with χ2 tests. Comparisons between complicated and uncomplicated palate traumas were compared with χ2 tests and the Mood median test.21 Effect sizes with corresponding 95% CIs were calculated as difference in medians for continuous variables and using the Cramer V for categorical variables.22 Univariable and multivariable logistic regressions were applied to identify significant associations of hospital admission, palate repair, or imaging. Adjusted odds ratios (ORs) selected from an optimal logistic model using backward selection were formulated. Variables with significant univariable associations were considered for selection. Rejection criteria were based on insignificant regression coefficients (adjusted OR). All statistical analyses were performed in SAS software, version 9.4 (SAS Institute Inc) using appropriate Taylor linearization procedures to adjust for study design, with the exception of the 95% CIs for difference in medians, which were calculated in R Gui (R Core Team). All statistical tests were assessed at the α = .05 level.
Authors completed the online Agency for Healthcare Research and Quality data use agreement. As per Agency for Healthcare Research and Quality limitations, columns in the data set with fewer than 10 discharges were omitted.
A total of 22 094 patients with primary diagnoses of palate trauma, open wound of palate without complication or complicated, presented to US EDs in 2006-2010 (4715 patients in 2006, 4560 patients in 2007, 4630 patients in 2008, 4274 patients in 2009, and 3915 patients in 2010). These numbers corresponded with a decreasing incidence of 1.58 per 100 000 people in 2006 to 1.26 per 100 000 people in 2010. Most patient demographics and hospital characteristics remained constant across the years. However, the rate of patients with Medicaid as the primary payer increased from 37.9% (1789 of 4715) in 2006 to 46.4% (1815 of 3915) in 2010, and the rate of patients with private insurance decreased from 44.3% (2091 of 4715) in 2006 to 37.6% (1473 of 3915) in 2010 (eTable in the Supplement). Median total charges for services also increased from $317 in 2006 (interquartile range [IQR], $192-$632) to $509 in 2010 (IQR, $293-$1068). Rates of repair varied from 2.8% (109 of 3915) in 2010 to 6.0% (275 of 4560) in 2007, without a trend.
When all unrestricted years were combined into a single cohort (2006-2010), the population was primarily male (13 967 [63.2%]); living in large, central metropolitan areas (5302 [24.0%]), with a median household income of $38 000-$46 999 (6146 [27.8%]); and paying with private insurance (9646 [43.7%]) (Table 1 and eTable in the Supplement). The population was mainly 18 years of age or younger (19 568 [88.6%]), with a median age of 2.8 years (IQR, 1.1-6.1) (Table 1). Most ED visits resulted in a routine discharge to home (19 819 [89.7%]). A total of 34 deaths (0.2%) were reported. The top 10 most frequent codiagnoses (1929 members of the cohort [8.7%] had at least 1 codiagnosis) in this cohort were asthma; hypertension; contusion of the face, scalp, and neck; tobacco use; type 1 and 2 diabetes; abrasion to the head; long-term anticoagulant use; face and neck injury; oral soft-tissue disease; and attention-deficit disorder with hyperactivity.
Open palate wounds without complication (ICD-9-CM code 873.65) accounted for 96.9% of primary palate trauma (n = 21 405) and open palate wounds with complication (ICD-9-CM code 873.75) accounted for 3.1% of primary palate trauma (n = 689) (Table 1). When comparing patients with palate trauma without complication and those with complication, there were significant differences in age, location of patient’s residence, primary payment method, total charges, disposition of patient from the ED, and mortality. Specifically, patients with palate trauma without complication were younger (median age, 2.7 vs 8.8 years; difference, −6.1 years [IQR, −13.2 to −3.0 years]) compared with those with a complicated palate injury, who were more often older than 18 years (275 [39.9%] vs 2251 [10.2%]; Cramer V, 0.083 [IQR, 0.070-0.096]). Although the distribution of primary payer was significantly different between the types of palate trauma, both groups were primarily paying with private insurance with a health maintenance organization (9646 [43.7%]). Patients with palate trauma without complications underwent routine discharge more frequently (19 317 [90.2%] vs 502 [72.9%]) and were less often admitted (908 [4.2%] vs 119 [17.3%]; Cramer V, 0.148 [IQR, 0.135-0.162]). Complicated palate traumas had higher median total charges for ED services than palate trauma without complications ($910 vs $406; difference in medians, −$504 [IQR, −$169 to −$14]), higher mortality rate (24 [3.5%] vs 10 [0.05%]; Cramer V, 0.068 [0.055-0.082]), and more common palate repair (59 [8.6%] vs 906 [4.2%]; Cramer V, 0.012 [0.007-0.026]).
Factors crudely associated with palate repair, performed in 965 patients (4.4%) overall, include patient’s residence (large fringe metropolitan area [OR, 1.99; 95% CI, 1.17-3.39], medium metropolitan area [OR, 1.98; 95% CI, 1.01-3.87], small metropolitan area [OR, 4.66; 95% CI, 2.32-9.36], or micropolitan area [OR, 2.32; 95% CI, 1.07-5.00]; all vs large central metropolitan area), primary payer (Medicaid vs private insurance: OR, 0.62; 95% CI, 0.41-0.94), and hospital region (Northeast vs South: OR, 0.36; 95% CI, 0.19-0.69; and West vs South: OR, 0.36; 95% CI, 0.19-0.69). On multivariable analysis, these significant factors revealed that patients living in a small metropolitan area (vs large central metropolitan; OR, 3.97; 95% CI, 1.84-8.59) were more likely to receive palate repair, while patients paying with Medicaid (vs private insurance; OR, 0.56; 95% CI, 0.37-0.84) and those living in the Northeast (vs South; OR, 0.32; 95% CI, 0.16-0.63) or West (vs South; OR, 0.49; 95% CI, 0.24-0.99) were less likely to receive palate repair.
A total of 1027 patients (4.6%) with primary palate trauma were admitted. Factors univariately associated with admission included type of palate trauma, age, sex, location of patient’s residence, payment method, hospital region, and codiagnosis (yes or no; having at least 1 top 10 codiagnosis). Factors considered for hospital admission during multivariable analysis were diagnosis coding of complicated palate trauma vs without complication (OR, 5.32; 95% CI, 3.10-9.15), males vs females (OR, 1.57; 95% CI, 1.11-2.21), and having a defined codiagnosis (OR, 2.75; 95% CI, 1.84-4.12) (Table 2). Hospitals in the Northeast had greater odds of admitting patients than hospitals in the South (OR, 2.73; 95% CI, 1.11-6.71), while those living in a medium metropolitan area (OR 0.54; 95% CI, 0.31-0.94), small metropolitan area (OR, 0.37; 95% CI, 0.20-0.67), micropolitan area (OR, 0.31; 95% CI, 0.15-0.62), or any other area (OR, 0.13; 95% CI, 0.05-0.35) were negatively associated with admission compared with a large, central metropolitan area.
For patients with palate trauma as a secondary diagnosis code, we analyzed associated primary diagnosis codes and determined that 193 secondary palate injuries occurred due to malar or maxilla fractures and 49 due to closed base of skull injuries. The remainder had primary diagnosis coding of a coinjury, such as injury to the lip, oral cavity, or pharynx (ICD-9-CM code 405) or contusion or general injury to the face (ICD-9-CM code 189). Based on the coding sequence of diagnosis codes, it is extremely rare to have a palate injury secondary to skeletal trauma. These 836 patients with secondary palate traumas were not included in further analysis to keep patients with primary palate trauma isolated.
After applying CPT coding quality restrictions, 15 196 patients were removed, resulting in 6897 patients (Figure). The population shifted to a larger proportion living in a large, fringe metropolitan area (full population, 4702 [21.3%]; restricted cohort, 2057 [29.8%]) and less often living in a large, central metropolitan area (full population, 5302 [24.0%]; restricted cohort, 1347 [19.5%]) (Table 3). In addition, the hospital region shifted away from the Midwest (full population, 5910 [26.8%]; restricted cohort, 1144 [16.6%]) and West (full population, 5670 [25.7%]; restricted cohort, 957 [13.9%]) and toward the Northeast (full population, 3312 [15.0%]; restricted cohort, 2113 [30.6%]) and South (full population, 7202 [32.6%]; restricted cohort, 2683 [38.9%]).
Before applying the CPT coding restrictions, we found that 1321 patients (6.0%) received any type of head and neck imaging. After applying coding exclusions, the use of any head and neck imaging increased to 823 patients (11.9%) (Table 3), who most commonly underwent radiography of the soft tissue of the neck (396 of 6897 [5.7%]). With regards to CT scan imaging, 440 patients (2.0%) received a CT scan before restriction, as did 235 (3.4%) within the restricted cohort. Factors univariately associated with any head and neck imaging included location of the patient’s residence, payment method, and hospital region. Two characteristics were associated with increased adjusted odds of any head and neck imaging in this restricted cohort: patients living in medium vs large, central metropolitan areas (OR, 1.62; 95% CI, 1.04-2.55) and patients paying with Medicaid vs private insurance (OR, 3.33; 95% CI, 1.00-10.91]) (Table 4). Living in small metropolitan areas vs large, central metropolitan areas decreased the odds of any head and neck imaging (OR, 0.40; 95% CI, 0.19-0.84), as did living in the Midwest vs South hospital region (OR, 0.43; 95% CI, 0.25-0.74).
Based on this national representative sample in 2006-2010, we found that primary palate trauma occurs in 1.26 to 1.58 patients per 100 000 people, primary palate trauma occurs mostly in children (88.6%; median age, 2.8 years), a small percentage of patients are admitted to the hospital (4.6%), and associations of admission included male sex, increased number of codiagnoses, Northeast region (reference, South), and complexity of the injury. We compared the regional distribution of palate trauma to the US census, averaged from 2006 to 2010, and found an overrepresentation of palate trauma in the Northeast and South, with underrepresentation in the Midwest and West, relative to the US population in those regions.
A total of 823 patients (11.9%) underwent imaging of the head and neck after reliable CPT coding restrictions were applied, most commonly radiography of the soft tissue of the neck (396 [5.7%]). Computed tomographic imaging of the head and neck is performed infrequently (234 of 6897 [3.4%]), and there does not seem to be a consistent modality used. Aside from geographic factors (increased rates of imaging in the South vs Midwest and variations within city areas), hospital location, and payment method, we could not identify any additional associations for undergoing imaging in the ED. A total of 965 patients (4.4%) underwent palate repair; those living in a small metropolitan area (vs large, central metropolitan area) were more likely to receive repair, while patients paying with Medicaid (vs private insurance) and living in the Northeast or West (vs the South) were less likely to receive repair.
In 1997, a study of 26 patients with oropharyngeal trauma concluded with the recommendation to perform radiography of the lateral soft tissue to determine if free retropharyngeal air is present. If present, the recommendation was to undergo CT scan.15 Findings on CTA are more diagnostic of neurovascular injury and include thrombus, hematoma, intimal tear, or edema around the carotid artery, which would be difficult to identify on radiographs of the soft tissue of the neck. Therefore, several larger studies have been conducted, with the most thorough evaluation in 2005 that included a systematic review and algorithm for management of pediatric oropharyngeal trauma.16 Patients with nonfocal results of examinations should undergo CT with contrast; if neurologic complication is expected or concerning clinical findings are present (expanding hematoma, diminished pulses, ongoing bleeding, or lateral palate location), the authors recommended CTA as the imaging study of choice. Furthermore, otolaryngology texts have advocated CT imaging, especially for large midline injuries, lateral injuries, or suspected vascular injuries.17,18 Our findings are contrary to these more current teachings because radiography of the soft tissue of the neck was the most common imaging modality performed, with rare use of CT imaging.
However, there are benefits to avoiding routine imaging. In otherwise healthy pediatric patients without concomitant symptoms, the use of additional imaging modalities likely delivers more harm than benefit owing to risk of radiation-induced malignant neoplasm. The risk of pediatric malignant neoplasm after CT is estimated between 0.2% and 0.7%.23,24 Using a decision tree analysis of CTA for every patient vs no CTA performed, the point at which the detection of stroke outweighs the development of malignant neoplasm was at a stroke rate of 4.5%.12 Clinically, the incidence of stroke after palate trauma has been as high as 2.5%,5 indicating that a management algorithm in which all healthy patients with palate trauma undergo CTA would be creating more malignant neoplasms than detecting possible strokes. Supporting this algorithm, we found all numbers of neurovascular sequelae codes reported during the acute ED encounter to be less than the reporting limit for this data set (<10); however, this number could be underestimated in our study since 72-hour follow-up is not recorded and limits our study’s ability to assess any prognostic factors for delayed stroke. Although this evidence further supports the management of palate trauma without CT imaging unless a clinical sign or symptom of neurovascular compromise is apparent, clinicians must still attempt to identify patients at risk for delayed stroke. Our study shows that this management algorithm appears to be the most common approach, contrary to some otolaryngology literature.
This study uses a previously reported algorithm20 to determine frequency of imaging, restricting the cohort to only hospitals that were adequately reporting CPT codes. In the prior article, the authors found that trauma centers were less likely to perform CT than were nonteaching hospitals.20 Our current research replicates the feasibility of this method and provides future investigators with a tool for identifying events associated with a CPT code that occurs during the ED visit.
In our analysis of the algorithm, we compared demographics to ensure that the algorithm did not alter the cohort (ie, hospital CPT reporting was random). After applying CPT coding quality restrictions, the population shifted to a larger proportion living in a large, fringe metropolitan area and the region shifted away from the Midwest and West toward the Northeast and South. Although a demographic shift occurred, this could be owing in part to the change in the study design weights within the database when poor-quality EDs were removed.
A limitation of this study are variables after the initial evaluation, as this data set only includes the events that occurred within the ED. Information regarding hospital stay, delayed neurovascular complications, delayed repair, and other outcome variables, such as readmission, are not reported. This database is based on patient discharges; therefore, 1 patient presenting to the ED multiple times for palate trauma appears as individual unlinked encounters; however, we believe this circumstance to be infrequent and not affecting the overall cohort. Use of antibiotics is also lacking from this data set.
Primary palate trauma encompasses a small burden on US EDs. Although it is often suggested in the otolaryngology literature to perform imaging, primary palate trauma usually results in routine discharge home without imaging. Furthermore, most of these patients are male and do not receive immediate repair to the palate injury. Imaging infrequency and lack of standardization should be noted since palate trauma could have life-threatening neurovascular sequelae, which may occur up to 72 hours after injury, a point that should be emphasized to patients and caregivers going home after palate trauma. Follow-up with the patient’s primary care clinician within this time window is necessary. An in-depth conversation with the patient and/or caregiver is needed to describe symptoms of neurologic sequelae, with emphasis on the expedient need for care if this rare event should occur.
Accepted for Publication: November 21, 2017.
Corresponding Author: Aaron Smith, MD, Department of Otolaryngology–Head and Neck Surgery, University of Tennessee Health Science Center, 910 Madison Ave, Ste 430, Memphis, TN 38163 (email@example.com).
Published Online: February 1, 2018. doi:10.1001/jamaoto.2017.3071
Author Contributions: Drs Smith and Ray had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: All authors.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Smith, Ray.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Smith, Ray.
Administrative, technical, or material support: Smith.
Study supervision: Chaiet.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.
Meeting Presentation: This study was presented at the Combined Otolaryngology Spring Meeting of the American Academy of Facial Plastic Surgery; May 20, 2016; Chicago, Illinois.
Create a personal account or sign in to: