Trends in healthcare utilization and costs associated with acute otitis media in the United States during 2008–2014

Every year, an estimated 709 million episodes of acute otitis media (AOM) occur globally, half of which are in children < 5 years of age [1]. By 3 years of age, at least 80% of children are expected to have had at least one episode [2]. This makes AOM the most common pediatric infectious disease and most common reason for medical visits, antibiotic prescriptions, and hearing impairment in children [2].

Streptococcus pneumoniae, non-typable Haemophilus influenzae, and Moraxella catarrhalis are the most common causes of AOM [3]. Over the last several decades, a variety of vaccines have been introduced in the US to control S. pneumoniae, including a 14-valuent pneumococcal polysaccharide vaccine in 1977, the current 23-valuent pneumococcal polysaccharide vaccine in 1983, a 7-valent pneumococcal conjugate vaccine (PCV7) in 2000, and a 13-valent pneumococcal conjugate vaccine (PCV13) in 2010 [4]. Vaccines against non-typeable H. influenzae and M. catarrhalis, however, are not currently available in the US. PCV7 greatly reduced the incidence of invasive pneumococcal disease and S. pneumoniae carriage [5], but the extent to which it affected AOM rates remains unclear [6], and even after PCV13 was introduced in 2010, S. pneumoniae remained one of the most frequent causes of AOM, accounting for 26–36% of cases in the US [7].

Further data on the real burden of AOM and its impact on the healthcare system are needed to assess the impact of recent interventions, especially PCV13, on AOM and to inform the development and use of new vaccines and other preventive measures. The aim of the current study was therefore to investigate changes in the incidence of AOM from 2008, just before PCV13 was introduced, to 2014, the most recent year for which data were available, based on US insurance records in the Truven MarketScan database. The study also examined the costs associated with index AOM events in the US during the two most recent years for which data were available (2013–2014).

This study showed that AOM-associated healthcare utilization and costs in the US remain substantial and changed little between 2008 and 2014. Children < 5 years of age had the highest rates of AOM. More than 90% of the AOM cases were seen in the outpatient setting and, other than rare hospitalizations, the rest were seen in ED/UC settings.

Randomized clinical trials have shown 0%–9% efficacy of PCV7 vs. all-cause AOM, suggesting that it would reduce AOM incidence, but database studies have not consistently indicated a decrease in AOM incidence since PCV7 was introduced in 2000 [6]. Between 2008 and 2014, incidence rates decreased slightly for children < 1 year of age but increased for adults, especially those older than 65 years. We also observed a small increase in the proportion of cases seen in ED and UC settings and slight change in the regional distribution of cases. Overall, however, the total incidence of AOM changed little and remained primarily an illness of young children. This suggests that PCV13, introduced in 2010, has not provided direct or indirect benefits beyond those already provided by PCV7.

Like S. pneumoniae, non-typable H. influenzae is a principal cause of AOM [3]. H. influenzae is normally a commensal microorganism present in the respiratory tract. Since H. influenzae type b vaccine was introduced in the late 1980s, non-typeable H. influenzae has become dominant [11]. Data collected between 2008 and 2010 in Rochester, New York, indicated that non-typable H. influenzae accounted for approximately the same proportion of AOM (28–34%) as S. pneumoniae (26–36%) [7]. The combined non-typeable H. influenzae protein D-pneumococcal polysaccharide vaccine is not available in the US, and the H. influenzae type b vaccines routinely used in the US do not prevent illness caused by non-typeable strains [12]. Protein-based non-typeable H. influenzae vaccines are being developed and are hoped to reduce AOM, but they may not be available for some years [11, 13]. Influenza is another vaccine-preventable illness that can cause AOM, although influenza viruses are minor contributors to the overall burden of AOM, so vaccination against influenza results in only a small reduction in AOM incidence [14].

AOM is one the illnesses for which antibiotics are most commonly prescribed [15, 16], and antibiotic prescription rates per consultation for AOM remain high in most countries (~ 80% of all consultations), especially in children < 2 years of age [17]. Other analyses of MarketScan data revealed that antibiotics were prescribed for 58% of AOM visits in 1997–2004 [16] and 52%–66% of ambulatory AOM visits in 2000–2011 [18]. As shown elsewhere [19], the current analysis showed that AOM episodes in the US peaks during the winter months, likely due to increased activity of respiratory pathogens. Thus, antibiotic prescriptions for AOM can help account for the increase in antibiotic prescriptions and antibiotic resistance of S. pneumoniae during winter months [20‐22]. Reducing AOM may therefore not only reduce healthcare utilization and costs but also the spread of antibiotic resistance.

The estimated cost per AOM episode during 2013–2014 was similar to past studies. In 1997, Kaplan et al. reported that the average total cost of treating an episode of AOM was $115.8 in the US [23]. Similarly, a 2012 analysis of data from a large pharmacy network in the US found that, for members < 20 years of age, the unadjusted average cost per AOM episode ranged from $102 to $255 depending on whether the case was seen in a retail clinic or office visit [24]. These costs, however, do not capture indirect economic consequences of AOM such as out-of-pocket costs or lost productivity.

The results of this retrospective study are strengthened by the multiple years covered and the large size and representativeness of the MarketScan database [9], which includes nearly 40 million individuals and more than 2.5 million cases of AOM each year. In addition, the results reflect real-world outpatient and inpatient settings. We therefore expect that the estimates of AOM healthcare resources use and costs are reliable.

Our study has several limitations. Accuracy of the estimates could have been affected by measurement errors or misclassifications. For example, patients could have been misclassified because of misdiagnosis or miscoding. Miscoding could have occurred if a provider mistakenly submitted the wrong code, used a less descriptive ICD-9-CM billing code, misclassified a condition, up-coded to maximize reimbursement, or omitted a diagnosed condition on the billing forms. Also, data for some individuals could have been missing if they did not use their insurance for medical or pharmacy encounters. Another limitation is that we were unable to compare healthcare utilization by vaccination status because the information is not routinely collected by the MarketScan database. Further, regions were defined according to US Census categories, so we were unable to obtain for alternative regions or sub-region. Finally, the study was limited to individuals insured commercially or through the Medicare program. Care should be taken when attempting to generalize these results to other AOM patient populations (other private insurance, other public insurance, or no insurance) who may have different rates and types of resource use due to different coverage.

AOM-associated healthcare utilization and costs in the US remain substantial and changed little between 2008 and 2014. This suggests that AOM caused by non-vaccine strains of S. pneumoniae, along with non-typeable H. influenzae and other infections continue to be a significant burden, especially for young children. This burden of AOM translates to frequent antibiotic use and probably also to bacterial resistance to antibiotics.