Trends in Acute Otitis Media-Related Health Care Utilization by Privately Insured Young Children in the United States, 1997–2004

METHODS

RESULTS

AOM-Related Ambulatory Visits

Annual ambulatory visit rates for AOM from 1997 to 2004 are shown in Fig 1. Comparing the rate in 2004 with the average rate for the baseline period (1997–1999), we observed 929 fewer ambulatory visits (95% CI: 917–940 visits) for AOM per 1000 person-years in children <2 years of age, that is, a 42.7% decrease from 2173 visits per 1000 person-years to 1244 visits per 1000 person-years (P < .001) (Table 2). On average, 2.1% of these ambulatory AOM visits were associated with tympanostomy and related procedures; this proportion did not change substantially over time. Annual ambulatory visit rates for other acute respiratory infections also decreased from 1997 to 2004 (Fig 1).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1

Rates of ambulatory visits and antibiotic prescriptions for AOM and non-AOM acute respiratory infections, 1997 to 2004. ARI indicates acute respiratory infection.

View this table:

• View inline
• View popup

TABLE 2

Rates of AOM-Related Ambulatory Visits According to Demographic Characteristics, in 1997 to 1999 and 2004

To assess whether reductions in AOM-related ambulatory visits might have been associated with secular trends in health care utilization during the study period, we calculated the proportions of total ambulatory visits that were AOM-related. We hypothesized that major secular trends would likely have affected all ambulatory visits. On average, AOM-related ambulatory visits constituted 15.7% of all annual ambulatory visits made by enrollees in the database during 1997 to 1999, compared with 10.6% in 2004 (a decrease of 32.4%).

Rates of AOM-related ambulatory visits decreased significantly among enrollees in both managed care and non–managed care plans (Fig 2). Among children enrolled in non–managed care plans, the rate of ambulatory visits decreased from an average of 2429 visits per 1000 person-years in the prevaccination years to 1231 visits per 1000 person-years in 2004 (a 49.3% reduction) (Table 2). For children enrolled in managed care plans, the corresponding decrease was from 2032 visits per 1000 person-years to 1244 visits per 1000 person-years (a 38.8% reduction). In 2004, there was no statistically significant difference in AOM visit rates between children enrolled in managed care plans and those enrolled in non–managed care plans.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2

Rates of ambulatory visits and antibiotic prescriptions for AOM according to insurance plan type.

Statistically significant decreases in the rates of AOM-related ambulatory visits were seen among enrollees of both genders. Among boys, the rate of ambulatory visits decreased from 2310 visits per 1000 person-years in the prevaccination years to 1348 visits per 1000 person-years in 2004 (a 41.6% reduction). For girls, the corresponding decrease was from 2028 visits per 1000 person-years to 1134 visits per 1000 person-years (a 44.1% reduction).

Controlling for the covariates and interaction terms in the multivariate model, the AOM-related ambulatory visit rate decreased significantly over time (P < .0.001). All of the covariates (gender and insurance plan type) and interactions (gender-time and insurance-time) were significant in the model. One of the significant interactions, the insurance-time interaction, suggested that the decreases among children enrolled in non–managed care plans occurred earlier than the decreases among children enrolled in managed care plans.

DISCUSSION

We conducted a comprehensive evaluation of AOM-related health care utilization by young children, including trends in ambulatory visits, antibiotic prescriptions, and related expenditures, over a period that spanned the introduction and maturation of the pneumococcal conjugate vaccination program in the United States. We found that, in a privately insured population, rates of ambulatory visits, antibiotic prescriptions, and resulting medical expenditures for AOM had decreased markedly by 2004, compared with prevaccination years. These results add to the growing evidence base of benefits of PCV7 vaccination and suggest that the vaccine may have contributed substantially to reducing the burden of AOM and associated medical costs.

AOM is a significant cause of outpatient visits and antibiotic prescriptions for children, and estimations of the burden of vaccine-preventable AOM and other pneumococcal disease-related syndromes have important implications for immunization policy. Although several studies have documented reductions in the rates of invasive pneumococcal disease (including drug-resistant infections) with routine PCV7 use,^14,15,16 population-level data on the immunization program's effect on AOM are currently limited.²² We found that, between 1997 to 1999 and 2004, rates of ambulatory visits and antibiotic prescriptions attributable to AOM decreased by 43% and 42%, respectively. Although other factors, such as clinical practice guidelines to reduce inappropriate antibiotic use, also might have contributed to the observed trend, this finding is consistent with studies that suggested that S pneumoniae was a major contributor to the burden of AOM-related ambulatory visits and antibiotic prescriptions in children <2 years of age.¹ The observed decrease in AOM rates after the introduction of PCV7 was gradual and consistent with increasing vaccination coverage during the study period.^10–13

The decreases in rates of AOM-related visits observed in our study exceeded vaccine efficacy estimates (6%–9%) reported in prelicensure clinical trials^19,20 but were consistent with results of other postmarketing national database studies evaluating the population impact of the PCV7 program. A national study showed that, already by 2002 to 2003, rates of otitis media-related outpatient medical care visits by children <2 years of age had decreased by 20%, compared with rates before PCV7 introduction.²¹ A clinic-based study found a ∼35% reduction in S pneumoniae isolates among bacterial isolates causing AOM, which suggests that the microbiologic features of AOM also might have changed since PCV7 introduction.²⁹ It is possible that the herd effects that have been demonstrated for invasive pneumococcal disease and pneumonia¹⁷ extend to other noninvasive pneumococcus-related diseases and benefit unvaccinated and partially vaccinated children through reduced carriage and transmission of vaccine serotypes.³⁰ By preventing the first episode of AOM, PCV7 may protect children against later susceptibility to frequent otitis media, complications, and ambulatory visits.^31,32 Several other factors, however, also might have influenced the observed reductions in visits. First, although we were unable to measure it in our study, uptake of PCV7 in this privately insured population might have been higher than in the general population.³³ Second, the trend of decreasing antibiotic use for children in the past decade may be continuing.^6,7,34 Third, the definition of AOM we used (AOM-related visits) was based on ICD-9-coded diagnoses of AOM and was different from case definitions used in clinical trials (episodes of AOM).^19,35

It is also possible that awareness of PCV7 vaccination status might have had an indirect effect on physician practices of diagnosing and coding AOM and other acute respiratory infections in vaccinated children. However, we found no compensatory increase in visit rates for non–AOM-related acute respiratory infections during the decrease in AOM visits, making it unlikely that a shift in diagnostic and coding practices would explain the observed decrease. The impact of educational efforts to promote judicious antibiotic prescribing is another important, potential, contributing factor.⁶ However, rates of antibiotic prescriptions for non-AOM visits did not change during the study period, which suggests that changes in practice patterns of prescribing antibiotics might not explain the observed decreases in AOM visits and related antibiotic prescriptions. Finally, because the largest absolute reductions in rates of AOM visits and prescriptions in the MarketScan databases had occurred already by 2002, it is unlikely that the 2004 American Academy of Pediatrics clinical practice guideline³ to consider deferring antibiotic treatment for some children with AOM had an effect on the trends observed in our study.

Pneumococcal illnesses are associated with considerable economic burden in the United States and in other countries. One US study estimated that diseases associated with streptococcus pneumoniae caused $2.5 billion in direct medical costs and $3 billion in work-loss and productivity costs for each US birth cohort.³⁶ Our data suggested that, compared with the prevaccination period, direct medical costs for AOM were reduced by an estimated $460 million annually. Approximately one fifth of direct AOM-related health care expenditures were attributable to antibiotic prescriptions. Our findings of benefits exceeded those of a prevaccine study that projected that implementation of pneumococcal conjugate vaccination in infants and young children could reduce annual direct medical costs of pneumococcus-associated diseases by $342 million.³⁶ Because our analysis of the costs of AOM included complete outpatient visit and antibiotic prescription data for the patients with a diagnosis of AOM from the same data set, the estimates for direct medical costs attributable to these outcomes are likely to be reasonable and can be used to update cost-effectiveness studies for PCV7. The total net savings from the PCV7 program could be even greater if indirect costs such as costs of parents' loss of work time because of the child's illness were included. In our study, the overall prevaccine rates of AOM visits were similar to those reported in clinical studies²⁰ and, as expected, were severalfold higher than the incidence of invasive pneumococcal disease (1.88 and 0.34 invasive cases per 1000 persons in <2-year-old children for the prevaccination period and 2004,³⁷ respectively). If our findings are confirmed by other studies, then use of PCV7 to prevent pneumococcal disease is likely to be more cost-effective (or even cost-beneficial) than estimated previously.^38,39

Several limitations to our analysis should be considered. Our estimates are based on data for privately insured persons, predominantly in managed-care health plans. Nationally, although most of the population is privately insured, utilization and medical expenditures per case of AOM may be greater for persons with Medicaid or without insurance,⁴⁰ which makes these estimates conservative. However, because the proportion of the US population with private insurance did not change during the study period,⁴¹ any bias in our analysis is likely to have been consistent over time, and our estimates should reflect the degree of reduction in AOM-related health care utilization and expenditures among persons with employee-sponsored health insurance, predominantly in a managed care setting.

Our study might have missed some AOM-related health care utilization because of incompleteness of the databases, incomplete coding, or multiple insurance plans for some enrollees. In general, these later instances likely would be rare. Another limitation is that the MarketScan population almost tripled during our study period and trends for AOM-related outcomes could be affected by secular changes in the MarketScan population or in insurance or associated utilization patterns. However, the proportion of total ambulatory visits that were AOM-related among all MarketScan enrollees decreased by 32.4%, without a compensatory increase in other acute respiratory infection-related visits.⁴² Therefore, this analysis should control at least partially for potential changes in practice standards for ambulatory visits that might have occurred over time.

The multivariate model results also suggested that the decreases were unlikely to be artifacts of secular changes in insurance coverage. Decreases in ambulatory visits occurred among both managed care and non–managed care insurance plans, although earlier in non–managed care plans. In addition, although managed care organizations became more widespread during the course of the study, our finding that the AOM reduction was greater in non–managed care plans would have biased our results to be conservative. Additional study is needed to understand why changes for non–managed care plans were larger than those for managed care plans.

MarketScan obtains its data from employer-based insurance plans, particularly from larger employers. The study population is therefore somewhat homogeneous (eg, all of these children had private insurance). Nonetheless, the MarketScan population is large, draws from numerous medical settings, and represents a wide variety of geographic, social, and economic strata; the data complement and enhance available surveillance data.^14,43 In conclusion, our data suggest that AOM-related health care utilization in children <2 years of age has been markedly reduced since the introduction of PCV7, highlighting the health and economic benefits of PCV7 vaccination in young children in the United States.