Age in years at time of admission; n is fewer than 10.
Total charges per state (states that had greater than 10 admissions) with number of admissions listed above each respective state. The solid line indicates mean charges for an admission, $9600.
Acevedo JL, Lander L, Shah UK, Shah RK. Existence of Important Variations in the United States in the Treatment of Pediatric Mastoiditis. Arch Otolaryngol Head Neck Surg. 2009;135(1):28-32. doi:10.1001/archoto.2008.510
To determine national variations in resource utilization in the treatment of pediatric mastoiditis.
National pediatric inpatient database.
The Kids' Inpatient Database for 2003 was used to extract data for admissions for mastoiditis.
A total of 1049 patients (57% were male, and the mean age was 6.3 years) were identified. Median total charges for an admission were $9600; total charges were less than $28 604 in 90% of admissions. The mean length of stay (LOS) was 4.3 days (range, 0-87 days). A total of 792 procedures were performed; 50.0% of patients underwent tympanostomy tube placement and/or myringocentesis, and 21.6% underwent mastoidectomy. The LOS for nonsurgical patients was 3.7 days. The LOS for children undergoing tube placement was 4.6 days, with mean total charges of $15 713; for mastoidectomy, the LOS was 5.5 days, with mean total charges of $23 185. The primary payer was private insurance in 51.5% and Medicaid in 39.4%. Predictors of increased charges were treatment at teaching hospitals (P = .005), treatment at children's hospitals (P < .001), LOS (P < .001), the number of procedures (P < .001), and hospital region (P = .003). Wide geographic variation was noted with respect to the mean total charges per admission, which ranged from $5016 to $35 898.
In 2003, the median charge for a pediatric mastoiditis admission was $9600; 50% of patients underwent tympanostomy tube placement, and about 21.6% underwent a mastoidectomy. There was wide variation in total charges for admissions. Resource utilization was higher in teaching hospitals and in children’s' hospitals.
Mastoiditis remains the most common intratemporal complication of otitis media1 and has been estimated to necessitate mastoidectomy in roughly 20% of cases.2 As described by House,3 the introduction of antibiotics reduced the need to perform mastoidectomy by 80%. In his classic 1946 article, he demonstrated a reduction in the incidence of mastoidectomy for patients with mastoiditis from 43% in 1936 to 17.6% in 1943. Although some studies reported the incidence of mastoiditis to be stable, others found an increase in the incidence rate. This increase could be attributed to the delayed use of antibiotics for most children with acute otitis media and growing antibiotic resistance.1 Mastoiditis affects boys more often than girls, and patients are typically younger than 5 years.4 Treatment of mastoiditis almost universally involves antibiotics, but the necessity, timing, and extent of surgery remain controversial. Despite the prevalence and severity of this disease, and the controversies that exist in its treatment, to our knowledge there have been no large studies to date that elucidate the demographics or the health care and socioeconomic burdens of mastoiditis in children.
In addition to demographics and outcome measures, studies of resource utilization are critical.5 Today's practice environment demands cost-effective care, making the study of resource utilization especially pertinent. Resource utilization analysis assesses resource (money, hospital time, or length of stay [LOS]) use in the context of variables such as patient and hospital demographics. Such studies can be used as internal benchmarks within a practice or health care system to assess intrapractice variation, as well as to compare practices with national normative samples. Such data also can be used to identify factors associated with increased utilization, enabling early recognition of risk factors that will predispose to higher resource use and potentially allow an intervention. Hospital responsibility and control over health care resources demand that physicians take the lead in such analyses and formulate proactive treatment programs based on our findings.
A national database to allow such analysis was created by the Agency for Healthcare Research and Quality, Healthcare Cost and Utilization Project, called the Kids' Inpatient Database (KID).6 The KID6 is an all-payer pediatric database of hospital admissions of children (those ≤20 years)in 36 states for the year 2003. It includes 3428 hospitals and 2 984 129 pediatric discharges. Community, nonrehabilitation hospitals (including academic medical centers and specialty pediatric hospitals) are included in the KID.6 Other studies in otolaryngology have used the KID6 to establish data and trends and normative data for conditions such as subglottic stenosis and lymphatic malformations.5,7 Our goal was to describe the demographics of children treated for mastoiditis and the resource utilization associated with such treatment.
Institutional review board approval was obtained. The source of the data is the Healthcare Cost and Utilization Project, Agency for Healthcare Research and Quality, KID.6 The International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes 383.00 (acute mastoiditis without complications) and 383.01 (subperiosteal abscess of mastoid) were used as inclusion criteria to search the KID6 in all of the diagnosis fields. There were no a priori exclusion criteria. By using these ICD-9-CM codes, we excluded patients with some intracranial complications that can be found according to codes 383.02 (acute mastoiditis with other complications [Gradenigo syndrome]) and 383.2 (petrositis). Hospital-specific information obtained from the KID6 included hospital location, teaching or academic status, type of hospital (nonchildren’s, children’s, children's unit), and source of admission (emergency, outside hospital, routine). For the definition of patient location, the term micropolitan was defined as an area having an urban core of at least 10000 residents. In the KID,6 the American Hospital Association Annual Survey of Hospitals and the National Association of Children's Hospitals and Related Institutions guidelines8 are used as criteria to identify children's hospitals. The teaching status of a hospital was determined by having an American Medical Association–approved residency program, being a member of the Council of Teaching Hospitals, or having a ratio of full-time equivalent interns and residents to beds of at least 1 to 4.8
Total hospital charges accrued during hospitalization, available for almost all patients in the KID,6 were used as a surrogate for resource utilization.5 The rationale for such use has been detailed in previous studies.9
Univariate analysis was performed comparing pertinent variables to total charges (Table 1). The P value indicates the extent to which the variable is an independent predictor of increased total charges. Intragroup comparisons were performed to further identify the fields that affect total charges.
A total of 1049 patients (57% of whom were male) were identified using ICD-9-CM diagnosis codes 383.00 and 383.01. The mean age of patients in the sample was 6.3 years (range, <1-20 years) (Figure 1). Fifty-two percent of patients were 5 years or younger. The median total charges for all admissions was $9600, with total charges of less than $28 604 in 90.0% of admissions. The mean LOS was 4.3 days (range, 0-87 days). In the sampled cohort of patients, a total of 792 otologic procedures were performed; 50.0% underwent myringocentesis and/or tympanostomy tube placement, and 21.6% underwent mastoidectomy.
Total charges were summarized for each variable annotated in the KID6; univariate analysis was performed to determine whether differences between subgroups achieved statistical significance (Table 1). The primary payer was private insurance for 51.5% of patients and Medicaid for 39.4%. Patients treated at teaching hospitals had higher total charges than those treated at nonteaching hospitals (P = .005), with treatment at children's hospitals being more expensive than at nonchildren's hospitals (P = .01). Additional predictors of increased resource utilization were LOS (P < .001), number of procedures (P < .001), and hospitalization region (P = .003). Wide geographic variation was noted with regard to mean total charges per admission (Table 1 and Figure 2). The mean total charges of the 4 states with the lowest admissions was $5016. The mean (SD) number of admissions sampled from each state was 29.4 (31.6) (Figure 2).
The mean LOS for patients undergoing tympanostomy tube placement was 4.6 days, with total charges averaging $15 713; for patients undergoing mastoidectomy, the mean LOS was 5.5 days with mean total charges of $23 185 (these data are summarized in Tables 2, 3, and 4).
To our knowledge, this report is the first large-scale national analysis of the demographics and resource utilization of pediatric mastoiditis in the United States. This is surprising because mastoiditis is a disease process frequently encountered by otolaryngologists. Furthermore, the resource utilization of this disease process is presently unknown. A prior study6 has estimated that children younger than 2 years are the most frequently affected group and that males are affected more frequently than females.4 Other prior studies2,10 have estimated the percentage of patients failing antibiotic therapy and requiring a mastoidectomy at 20% to 25%. Our analysis of a large national data set is similar to those from relatively smaller studies.2- 4 In the analysis of sampled admissions from the KID in 2003, more than half of admissions were male, and the rate of mastoidectomy (21.6%) was similar to that previously published.2,10 The mean age of patients in our series was 6.3 years, and one-third (35%) of patients were younger than 2 years. It is of interest that over the past 60 years, the rate of mastoidectomy for mastoiditis remained amazingly constant being 17.6% in 19463 compared with 20.1% in the current study. This may imply that the advances in pharmacologic treatment of otitis media have been matched by compensatory or selective changes in the bacterial mechanism that obviate newer antibiotics and thereby maintain the virulence of these inciting organisms. The treatment for mastoiditis remains controversial, depending on patient and surgeon characteristics. Because the rates in our study are similar to historical means, it may be that physicians are trained in specific treatment algorithms and continue to treat patients in the manner in which they were trained. The only way to determine the optimal treatment for mastoiditis is with a properly constructed prospective trial; unfortunately, the KID6 does not allow an analysis to answer such sophisticated questions.
Khafif et al11 found that the mean LOS for patients with mastoiditis treated with antibiotics and a myringotomy and tympanostomy tube placement was 6.6 days. This was contrasted to a mean LOS of 13.8 days for patients having a mastoidectomy performed. In our analysis, the mean LOS for patients undergoing tympanostomy tube placement and mastoidectomy were statistically significantly shorter (see Table 2 for P values). Table 3 and Table 4 demonstrate that treatment at teaching institutions and children's specific centers results in an increased number of surgical procedures, which may account for the fact that teaching institutions have higher total charges.
It is of interest that in this national sample, only approximately 7% of patients were discharged to their homes with services (Table 1). Patients who went home with services also had mean charges approximately $3500 higher per admission than those of routine patients. Indeed, it may be that these patients were sicker and hence their admissions more complicated, leading to increased charges.
Several interesting trends were noted in the analysis of the sampled admissions in the KID6 (Table 1). Certain variables, such as LOS and increased number of procedures, intuitively can be understood to lead to increased resource utilization, and analysis of KID6 confirms these relationships for patients with mastoiditis.
Other variables associated with increased resource utilization, such as treatment at teaching hospitals and treatment at children's hospitals, are not as intuitive. The higher charges at teaching institutions and at children's hospitals may be due to these facilities being forced to handle the burden of sicker children; that is, it may be that a selection bias exists and that children with more complicated presentations or comorbidities are selectively referred to such centers. Perhaps higher charges may be attributed to those incumbent in being treated at a training institution. These increased charges may support the case for changing reimbursement mechanisms to account for increased resource needs. Further study of theses question is being undertaken.
Also of interest is the greater than 7-fold increase in resource utilization seen in Minnesota (the highest utilization) compared with the 4 states with the lowest utilization. Furthermore, when stratifying the country into regions, Minnesota is located in the region with the lowest mean total charges (P < .049); this reveals the ability of users of statistical databases to purposefully mitigate outliers by using certain selection criteria. It is beyond the scope of this study and the limitations of the KID6 to interpret why such variations occur. It is hard to believe that children present with such variable degrees of disease severity among these states to the extent that this would explain the 7-fold difference in charges. Perhaps a systems approach and analysis may explain the difference. Comparative analysis of payer mix and resource utilization among states and analyses with other nations' data may be fruitful in formulating cost-effective treatment strategies.
The primary limitation of our study is that for the purposes of data extraction, our inclusion criteria were based according to the ICD-9-CM diagnosis code. The bias is that the data in the KID6 are only as good as the reporting institutions, and the veracity of those mechanisms cannot be corroborated. It is nearly impossible to accurately determine comorbid conditions of the patients identified in the KID,6 and this is a disadvantage of using large national databases. Furthermore, it may be that some patients were missed by our search strategy and that another database or search criteria may have provided a higher yield.
Mastoiditis remains a clinically significant entity, with 1049 sampled admissions in 2003. Boys younger than 2 years are the most commonly affected. Wide variations exist in the resource utilization of this condition, with a 7-fold difference across the United States in the 36 states included in the KID.6
Median total charges for all admissions were noted to be $9600, with total charges of less than $28 604 in 90% of admissions. The mean LOS was 4.3 days; for patients undergoing myringocentesis and/or tympanostomy tube placement, the mean LOS was 4.6 days, with mean total charges of $15 713; and for patients undergoing mastoidectomy, the mean LOS was 5.5 days, with mean total charges of $23 185. Fifty percent of patients underwent myringocentesis and/or tympanostomy tube placement, 21.6% underwent mastoidectomy. Increased LOS (P <.001, increased number of procedures (P <.001), region of hospitalization (P = .003), treatment at a teaching facility (P = .005), and treatment at a pediatric hospital (P <.001) were all associated with statistically significant increases in resource utilization. Further study is warranted to elucidate these trends and offer more efficient and cost-effective health care for children.
Correspondence: Rahul K. Shah, MD, Division of Otolaryngology, Children's National Medical Center, 111 Michigan Ave NW, Washington, DC 20010 (email@example.com).
Submitted for Publication: January 18, 2008; final revision received May 27, 2008; accepted May 28, 2008.
Author Contributions: Drs Lander and Shah 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: Lander, U. K. Shah, and R. K. Shah. Acquisition of data: R. K. Shah. Analysis and interpretation of data: Acevedo, Lander, U. K. Shah, and R. K. Shah. Drafting of the manuscript: Acevedo, U. K. Shah, and R. K. Shah. Critical revision of the manuscript for important intellectual content: Lander, U. K. Shah, and R. K. Shah. Statistical analysis: Lander and R. K. Shah. Administrative, technical, and material support: Acevedo, U. K. Shah, and R. K. Shah. Study supervision: U. K. Shah and R. K. Shah.
Financial Disclosure: None reported.
Previous Presentation: This study was presented as a poster at the 2008 American Society of Pediatric Otolaryngology Scientific Program; May 2-4, 2008; Orlando, Florida.