Abstract
Objective. As part of a long-term study of possible effects of early-life otitis media on speech, language, cognitive, and psychosocial development, we set out to delineate the occurrence and course of otitis media during the first 2 years of life in a sociodemographically diverse population of infants, and to identify related risk factors.
Methods. We enrolled healthy infants by age 2 months who presented for primary care at one of two urban hospitals or one of two small town/rural and four suburban private pediatric practices. We intensively monitored the infants' middle-ear status by pneumatic otoscopy, supplemented by tympanometry, throughout their first 2 years of life; we monitored the validity of the otoscopic observations on an ongoing basis; and we treated infants for otitis media according to specified guidelines.
Results. We followed 2253 infants until age 2 years. The proportions developing ≥1 episode of middle-ear effusion (MEE) between age 61 days (the starting point for data analysis) and ages 6, 12, and 24 months, respectively, were 47.8%, 78.9%, and 91.1%. Overall, the mean cumulative proportion of days with MEE was 20.4% in the first year of life and 16.6% in the second year of life. Tympanostomy-tube placement was performed on 1.8% and 4.2% of the infants during the first and second years of life, respectively. By every measure, the occurrence of MEE was highest among urban infants and lowest among suburban infants; these differences were greatest in the earliest months of life.
Overall, unadjusted mean cumulative proportions of days with MEE were higher among boys than girls, higher among black than white infants, and higher among Medicaid than private health insurance enrollees. Cumulative proportions of days with MEE varied directly with the number of smokers in the household and with the number of other children to whom infants were exposed, whether at home or in day care, and varied inversely with birth weight, maternal age, level of maternal education, a socioeconomic index, and duration of breastfeeding.
After adjustment, using multivariate analysis, the only variables that each remained independently and significantly related to the cumulative proportion of days with MEE were: during the first year of life, study site grouping, sex, the socioeconomic index, breastfeeding for ≥4 months, the number of smokers in the household, and an index rating the degree of exposure to other children at home or in day care; and during the second year of life, sex, the socioeconomic index, and the child exposure index. The duration of breastfeeding and the degree of exposure to tobacco smoke contributed little to the explained variance; most was attributable to differences in the socioeconomic index and the child exposure index.
Conclusions. Contrary to findings in many previous reports, the prevalence of otitis media during the first 2 years of life among lower-socioeconomic-status black infants appears to be as high as, if not higher than among lower-socioeconomic-status white infants, and certainly higher than among middle-class white infants. Among middle-class white infants the prevalence may also be higher than commonly assumed. The most important sociodemographic risk factors for otitis media appear to be low socioeconomic status and repeated exposure to large numbers of other children, whether at home or in day care.
- otitis media
- middle-ear effusion
- prevalence
- risk factors
- epidemiology
- sociodemographic factors
- critical review
Otitis media is, next to the common cold, the most commonly diagnosed and probably the most prevalent illness in United States children, with its peak incidence and prevalence during the first 2 years of life.1,2 Many studies of the epidemiology of otitis media in the first 2 years of life have been reported, contributing much to understanding of the disease, but all have had limitations in design, methodology, scope, or generalizability. We report here epidemiologic findings in 2253 Pittsburgh-area children, representing a broad sociodemographic spectrum, who were monitored closely for middle-ear disease during their first 2 years of life. The findings were obtained in the course of a prospective study currently in progress whose principal aims are to determine whether persistent otitis media with effusion (secretory otitis media; nonsuppurative otitis media) during the first 3 years of life results in lasting impairments of speech, language, cognitive, or psychosocial development, and if so, whether prompt tympanostomy-tube placement prevents or lessens the impairments.3 The present findings provide new information about the extent to which otitis media develops in different groups of infants, and they call into question commonly stated assumptions about the differential occurrence of otitis media in white and in black children. The findings also affirm the importance of certain risk factors for the development and persistence of otitis media that have been in dispute, while casting doubt on the importance of others.
SUBJECTS AND METHODS
Enrollment and Eligibility
During the period June 1991 to August 1992 we began enrolling infants who presented for primary care at Children's Hospital of Pittsburgh, The Mercy Hospital of Pittsburgh, or one of two small town/rural (Beaver and Kittanning) and four suburban (Brentwood, Gibsonia, Mt Lebanon, and Pleasant Hills) private pediatric practices in the Greater Pittsburgh area. The study was approved by the respective institutional review boards of the two hospitals. From the initiation of the study at each practice site, efforts were made to interview the parent(s) and solicit enrollment of every apparently normal infant whose first visit to the practice occurred within the first 2 months of life, and who met none of the following exclusion criteria: birth weight <5 lb (2268 g); small for gestational age; a history of neonatal asphyxia or other serious illness; a major congenital malformation or chronic illness; multiple birth; a sibling enrolled in the study; in foster care or adopted; mother dead, seriously ill, a known drug or alcohol abuser, or overwhelmingly limited socially or intellectually; mother aged <18 years; parents residing outside the area, or planning to move from the area within 5 years or to transfer care to a nonstudy practice; and English not the only household language. Certain infants were not enrolled because staff members were unable to interview the mother or because practice facilities were temporarily strained.
Follow-up and Monitoring of Middle-ear Status
All subjects are scheduled for evaluation at least monthly. At each visit a standardized interval history is obtained, including information on symptoms, feeding (breast milk vs formula), and day care attendance, and a study-team clinician examines each ear using a pneumatic otoscope with an air-tight lens assembly (model 20200, Welch Allyn Inc, Skaneateles Falls, NY). Cerumen is removed as necessary, and the clinician makes a forced-choice decision for each ear concerning the presence or absence of middle-ear effusion (MEE). When effusion is present, some of the clinicians use more stringent,4 and others, less stringent5 criteria in distinguishing acute (suppurative) otitis media from otitis media with effusion. At Children's Hospital, otoscopic examinations are performed by one of three specially trained pediatric nurse practitioners, and at the other study sites, by one of the primary care clinicians—currently a total of 32 pediatricians, five pediatric nurse practitioners, and one physician assistant—or one of three specially trained nurse otoscopists. (Formal training for each nurse otoscopist spanned approximately 3 months and entailed one-on-one, hands-on instruction in examination techniques and repeated one-on-one comparison of otoscopic findings and diagnoses with those of validated otoscopists until satisfactory agreement was achieved consistently). Decisions about doubtful otoscopic diagnoses are made in consultation with an on-site study-team pediatrician, or exceptionally, a study-team otolaryngologist using otomicroscopy. Data forms for each visit are verified independently to ensure accuracy of recording. Reimbursement is provided to parents at each visit.
Until January 1995, tympanometry, using a GSI 33 Middle Ear Analyzer (Grason-Stadler Inc, Milford, NH), was performed routinely at each study visit (except when otorrhea was present). Since then, because of its limited positive predictive value,6-8 tympanometry has been performed only when MEE is diagnosed, suspected, or newly resolved, or when the presence or absence of MEE is critical in a protocol-related decision, or on request by a study-team clinician. Tympanograms are classified and interpreted according to criteria modified from those reported previously.7-10 When otoscopic and tympanometric findings appear discrepant, examinations are repeated and the final otoscopic diagnosis is used to classify middle-ear status. More detailed otoscopic-tympanometric relationships in study subjects than those described previously7,8 will be reported separately. Tympanometers are recalibrated annually.
Assessment of Interobserver Reliability
For assessing the accuracy of study clinicians' otoscopic diagnoses, concurrent independent diagnosis by one of the authors (J.L.P.) is used as the criterion standard, since, in other studies, his otoscopic diagnoses of the presence or absence of MEE have conformed closely to independent findings at myringotomy.4,11 At Children's Hospital, comparisons of diagnoses are recorded when, for any reason, a subject is examined independently by both a study-team nurse practitioner and J.L.P (most commonly, when an otoscopic diagnosis is problematic). At each of the other study sites, J.L.P. conducts approximately monthly to bimonthly half-day interobserver exercises on-site with study-team clinicians, using as subjects unselected nonstudy as well as study patients who present to the practices for care.
At Children's Hospital, in 489 comparisons of independent diagnoses concerning the presence or absence of MEE in individual ears, agreement between study-team nurse practitioners and J.L.P. was found in 92.6% (κ=.85). At the other seven study sites, in 2099 such comparisons, the overall degree of agreement between study-team clinicians and J.L.P. was 91.0% (κ=.76), and the range at the respective sites was 85.0% (κ=.58) to 96.5% (κ=.89). Based on J.L.P.'s diagnoses as the criterion standard, clinicians' diagnostic sensitivity at the eight sites, respectively, ranged from 74.4% to 93.8%, and their specificity, from 87.8% to 97.8%. Overall, sensitivity was 88.8%, specificity was 92.4%, positive predictive value was 83.7%, and negative predictive value was 95.0%. No consistent or suggestive relationship was apparent between degrees of diagnostic agreement at the respective study sites and cumulative proportions of days with MEE at those sites.
Medical Treatment of Otitis Media
Medical treatment of otitis media generally follows specified guidelines agreed to by all collaborating pediatricians, but all individual treatment decisions are made by or with the assent of subjects' primary care clinicians. Originally the guidelines called for antimicrobial treatment routinely for new episodes of either acute otitis media or otitis media with effusion, and for antimicrobial prophylaxis routinely for recurrent acute otitis media (three episodes of acute otitis media within 6 months or four episodes within 1 year). However, in September 1994, because of the increasing prevalence nationwide of infections due to multiply resistant Streptococcus pneumoniae and the heightened risk of such infections in patients receiving antimicrobial treatment,12 the guidelines were revised. Under the revision, antimicrobial treatment for otitis media with effusion is limited to instances in which persistence of effusion is prompting consideration of tympanostomy-tube placement, and use of antimicrobial prophylaxis for recurrent acute otitis media is kept sharply curtailed.13 In most instances antimicrobials are provided to parents without cost.
Estimating Cumulative Proportions of Days With MEE
As used in the present report, the term middle-ear effusion (MEE) encompasses all types of otitis media in which effusion is present, ie, acute otitis media with or without otorrhea, otitis media with effusion, and otorrhea through a tympanostomy tube. For each child we estimate the cumulative proportions of days on which unilateral and bilateral MEE, respectively, are present, based on diagnoses at individual visits and interpolations for intervals between visits provided that the intervals do not exceed 90 days. Interpolation rules are as follows: if on two successive visits an ear is effusion-free, the ear is assumed to have been effusion-free throughout the interval between the visits. Similarly, if on two successive visits an ear is observed to contain effusion, the ear is assumed to have contained effusion throughout the interval. If an ear is effusion-free on one visit but contains effusion on the next visit, or if the reverse sequence obtains, the ear is assumed to have had the status of the first visit throughout the first half of the between-visit interval, and the status of the second visit throughout the second half. If an interval between any two visits by a subject exceeds 90 days, interpolation is not performed, and the subject's data set for the particular year of life is considered incomplete.
Tympanostomy-tube Placement
Subjects who sooner or later during the first 3 years of life develop MEE that appears substantial in degree and that persists, despite antimicrobial treatment, for extended periods of specified duration (eg, bilateral MEE for 90 days or unilateral MEE for 135 days continuously, or bilateral or unilateral MEE intermittently for specified proportions of longer periods) are assigned randomly, subject to parental consent, to undergo tympanostomy-tube placement either promptly or after a defined extended period if MEE remains present.3 (Further details concerning the respective durations are available from the authors). Parents who withhold consent for randomization may elect or decline tube placement. Tube placement also is performed in certain children not meeting randomization criteria, because of inordinately recurrent acute otitis media.
Developmental Assessment
Language development, behavior, and level of parental stress are assessed in all subjects through parent questionnaires at ages 1 and 2 years; preliminary results of these assessments have recently been reported.14-17 Formal tests of speech, language, cognition, and psychosocial development will be administered at ages 3, 4, and 6 years to all subjects who have met randomization criteria and to a representative sample of other subjects who collectively exhibit a spectrum of MEE experience from none to much.
Statistical Analysis
In the component of the study reported on here, the main objectives were to document the occurrence of MEE during the first 2 years of life using various measures, and to investigate relationships between various putative risk factors and cumulative proportions of days with MEE. To categorize socioeconomic status we constructed an index consisting of health insurance status (Medicaid vs private) as the primary element and maternal education (less than high school, high school, and college graduate or higher) as the secondary element. At follow-up visits, infants were considered breastfed if their diet at the time included any breast milk. The number of other children in the household was based on data collected on admission to the study. Day care attendance was classified according to both number of hours per week (<10, 10 to 20, or >20) and number of other children in the group (none, 1 to 4, or ≥5). Because for many infants these values did not remain constant, we categorized infants using the modal (most commonly occurring) value in the course of a year for number of hours, and the associated modal value for number of children. (Categorization using maximal rather than modal values gave similar results in relation to cumulative proportions of days with otitis media). To categorize overall exposure to other children, we constructed a child exposure index that took into account the number of other children in the household and, for those in day care, also the modal value for the number of children in that setting. For the index, we classified infants into one of three groups, without regard to weekly time in day care: (1) at home with no other children, or in day care with no other children; (2) at home with ≥1 other child, or in day care with <5 other children; and (3) in day care with ≥5 other children, irrespective of the number of other children at home.
All analyses exclude findings within the first 2 months of life because of infants' variable study entry dates before age 2 months and because otoscopic findings before age 2 months are often equivocal. Analyses were performed separately for the first and second years of life and the 2 years combined. All statistical tests were two-tailed. χ2 tests were used to test for differences between proportions; tests involving fourfold tables incorporated the Yates correction. Analysis of variance was used to test for differences between means, and a Kruskal-Wallis analysis was used for differences between medians when data were not distributed normally. Relationships of individual sociodemographic variables to cumulative proportions of days with MEE were tested using analysis of variance, after applying an arcsine transformation to the proportions to obtain a distribution that better approximated a normal distribution. A modification18was used to test relationships involving linear trends. A linear regression model was used to control for potentially confounding variables and to test for interactions, ie, whether effects of individual variables depended on the presence, absence, or level of other variables. For multivariate analyses, inclusion in models was restricted to variables significant at P ≤ .05. For these analyses, when outcomes were similar for subgroups with differing levels of the variable in question, or when the numbers of subjects in cells were small, adjacent categories for individual variables were grouped.
RESULTS
Enrollment and Subjects' Characteristics
Of 6111 infants born during the period April 1, 1991 to December 31, 1993 and presenting for care at one of the eight study sites within the first 2 months of life, we obtained enrollment interviews on 5991 (98.0%), of whom 4655 (77.7%) were found eligible for enrollment. The most common reasons for ineligibility, in order of frequency, were birth weight <5 lb, a sibling already enrolled in the study, mother aged <18 years, and a foreign language spoken in the home. Of the 4655 eligible infants, 3663 (78.7%) were enrolled—81.5% at the two hospital sites, 85.4% at the two small town/rural sites, and 68.7% at the four suburban sites. As of December 31, 1995, 2253 (61.5%) of the 3663 enrolled children had complete data sets (defined as no intervals longer than 90 days between follow-up visits) for their first 2 years of life, 98 (2.7%) had data sets for their first 2 years of life that were incomplete in some degree, 133 (3.6%) had been discharged before age 2 years because of having been found to meet an exclusion criterion, 8 (0.2%) had died before age 2 years of causes seemingly unrelated to otitis media, and 1171 (32.0%) had been lost to follow-up before age 2 years—38.9% at the urban hospital sites, 19.6% at the small town/rural sites, and 23.5% at the suburban sites. The 2253 children with complete data sets constitute the subjects of the present report.
Comparability of Subjects With Others
In the 120 infants whose parents were not interviewed, distributions according to sex, race, health insurance status, birth weight, maternal age, and birth order did not differ significantly from distributions in infants whose parents were interviewed and who were eligible for enrollment (all P values ≥.10). Of the 992 eligible infants whose parents declined enrollment, larger proportions were white (P < .0001), had private health insurance (P < .0001), had mothers aged ≥30 years (P < .0001), and had mothers who were college graduates (P = .004) than was the case in enrolled infants. In the 98 infants still enrolled at age 2 years but with incomplete data sets, and in the 1171 infants lost to follow-up before age 2 years, larger proportions were black (P < .0001), were Medicaid enrollees (P < .0001), had mothers aged ≤24 years (P < .0001), and had mothers who had not graduated from high school (P < .01) than was the case in infants with complete data sets through age 2 years.
Subjects' Characteristics in Relation to Study Site Groupings
Based on geographic location and a judgment about associated general socioeconomic level, and on distributions of subjects according to health insurance status, race, and maternal education, the eight study sites appeared to fall logically into three discrete groupings–urban, small town/rural, and suburban. Sociodemographic characteristics of the enrolled subjects in relation to their study site groupings are summarized in Table 1. At the urban sites, most infants were black and a large majority were Medicaid beneficiaries; at the small town/rural sites, a large majority were white and about two thirds were covered by private health insurance and one third by Medicaid; and at the suburban sites, virtually all infants were white and a large majority were covered by private health insurance. Birth weights, mothers' ages, levels of maternal education, the prevalence and duration of breastfeeding, and the use of day care all tended to be lowest at the urban sites and highest at the suburban sites, whereas the numbers of smokers in the household tended to be highest at the urban sites and lowest at the suburban sites. The numbers of other children in households (and birth order, not shown) differed little across study site groupings.
Distribution of Sociodemographic Characteristics of Subjects According to Study Site Grouping
Occurrence and Treatment of Otitis Media
Subjects Followed to Age 2 Years
Table 2 shows data on selected measures of the occurrence and treatment of otitis media during the first and second years of life, respectively, in relation to subjects' study site groupings. The mean annual numbers of visits were slightly lower among urban than among small town/rural and suburban infants. Overall, 91.1% of the infants developed ≥1 episode of MEE between age 61 days (the starting point for data analysis) and age 24 months. The mean cumulative proportion of days with MEE was 20.4% in the first year of life and 16.6% in the second year of life. The median total number of days of antimicrobial treatment for otitis media was 28 in both the first and the second year of life; most of the antimicrobial treatment received by study infants was for otitis media. Tympanostomy-tube placement was performed on 1.8% and 4.2% of the infants during the first and second years of life, respectively. Fig 1shows distributions of infants in the three study site groupings in relation to mean cumulative proportions of days with MEE during the first and second years of life, and Fig 2 shows the proportions of infants with MEE in each study site grouping, by month of life. By each of the measures shown in the table and the figures, the occurrence of MEE was highest among urban infants and lowest among suburban infants, and virtually all of the differences between the study-site groupings were statistically significant. Correspondingly, both antimicrobial treatment and tube placement were carried out most in urban infants and least in suburban infants. As shown in Fig 2, differences in the occurrence of otitis media between infants in the respective study site groupings were greatest in the earliest months of life. Finally, for individual infants, the extent of MEE during their second year of life tended to reflect their first-year experiences: overall, Pearson's correlation coefficient between the mean cumulative proportions of days with MEE in the first and second years of life, respectively, was .48; values at the individual study sites ranged from .35 to .59.
Selected Measures of the Occurrence and Treatment of Otitis Media During the First Two Years of Life According to Study Site Grouping
Distributions of infants in the three study site groupings in relation to mean cumulative proportions of days with middle-ear effusion during the first year of life, the second year of life, and the first and second years combined.
Proportions of infants with middle-ear effusion in each study site grouping, by month of life.
First-Year Outcomes in Subjects Lost to Follow-up During the Second Year of Life
Of the 1171 subjects lost to follow-up before age 2 years, 597 had complete data sets for the first year of life. In these subjects the mean cumulative proportion of days with MEE during the first year of life was 20.6%, compared with 20.4% in the infants with 2 full years of data (Table 2).
Cumulative Proportions of Days With MEE in Relation to Sociodemographic Characteristics
Total Days With MEE
Tables 3 and 4 show, respectively, mean cumulative proportions of days with MEE (days with unilateral MEE and with bilateral MEE totaled) during the first and second years of life in relation to selected sociodemographic characteristics. Overall, proportions were higher among boys than girls, higher among black than white infants, highest among infants born during summer and lowest among infants born during winter, and higher among Medicaid than private health insurance enrollees. In most instances the differences between subgroups were statistically significant. A number of statistically significant linear trends also were evident: cumulative proportions of days with MEE varied directly with the number of smokers in the household and with the number of other children to whom infants were exposed, whether at home or in day care, and varied inversely with birth weight, maternal age, level of maternal education, socioeconomic index, and duration of breastfeeding. The trend regarding maternal education appeared more pronounced among Medicaid than private insurance enrollees, and the trend regarding breastfeeding appeared most pronounced among urban and least pronounced among suburban infants, but tests for interactions found these differences not statistically significant.
Mean Cumulative Percent of Days With Middle-ear Effusion During First Year of Life* According to Study Site Grouping and Selected Sociodemographic Characteristics
Mean Cumulative Percent of Days With Middle-ear Effusion During Second Year of Life According to Study Site Grouping and Selected Sociodemographic Characteristics
Day care relationships were more complex. Overall, in both the first and second years of life, infants in day care had higher cumulative proportions of days with MEE than did infants at home, and overall, cumulative proportions of days with MEE increased as time in day care (<10, 10 to 20, >20 hours per week) increased and as the number of other children in the day care setting (none, <5, ≥5) increased. However, the trend regarding numbers of other children in the day care setting was much more pronounced than the trend regarding weekly time in day care. Moreover, irrespective of weekly time in day care, cumulative proportions of days with MEE were substantially lower in infants in day care with no other children, than in infants at home. Because of these relationships, and because of the strong positive relationship noted previously between days with MEE and number of other children in the household, we constructed the child exposure index described previously. As shown in Tables 3 and 4, in infants in each of the study site groupings classified according to this index, cumulative proportions of days with MEE showed a strong positive relationship to the degree of exposure to other children.
Finally, within individual categories or levels of the various sociodemographic characteristics, cumulative proportions of days with MEE were, in virtually all instances in the first year of life and in most instances in the second year of life, highest among urban infants and lowest among suburban infants.
Days With Bilateral MEE
Overall, days with bilateral MEE comprised 55.9% of total days with MEE in the first year of life, 48.8% in the second year of life, and 52.5% in the 2 years combined. From subgroup to subgroup, the fraction of total MEE days comprised by days with bilateral MEE varied relatively little, and appeared unrelated to whether the cumulative proportion of total days with MEE in a subgroup was relatively high or relatively low in comparison with the proportions in other subgroups.
Multivariate Relationships
Of the sociodemographic variables that, during the first 2 years of life, showed statistically significant correlations with the cumulative proportion of days with MEE, the variable most highly correlated was study site (r = .23;P < .001). Given that finding, we first investigated all possible bivariate models that each contained, respectively, study site and one of the other statistically significant variables. In those models, we found that race, birth weight, maternal age, and season of birth were no longer significantly related (P < .05) to the cumulative proportion of days with MEE, but that study site remained significantly related to the cumulative proportion in each model. We then proceeded to a multivariate model that included study site grouping and all of the remaining significant individual variables (the socioeconomic and child exposure indexes were used in place of their component elements). In this model, during the first year of life, study site grouping, sex, socioeconomic index, breastfeeding for ≥4 months, number of smokers in the household, and child exposure index each remained independently and significantly (P < .05) related to cumulative proportion of days with MEE. During the second year of life, only the relationships with sex, socioeconomic index, and child exposure index remained significant, although the relationship with study site grouping approached significance (P = .12). TheR2 for the model was .143 during the first year of life and .078 during the second year of life. The duration of breastfeeding and the degree of exposure to tobacco smoke contributed little to the explained variance; most was attributable to differences in the socioeconomic index and the child exposure index.
DISCUSSION
Study Design and Methodologic Issues
Previous Studies
Studies of the epidemiology of otitis media that include the first 2 years of life have variously had cross-sectional,19-29retrospective cohort,30-42 and prospective cohort43-69 designs. The cross-sectional studies have provided information that was, of necessity, limited temporally. In the retrospective cohort studies, a major inherent limitation has been questionable ascertainment of both the presence and the absence of otitis media over time. The prospective cohort studies, many of which, like the present study, were designed to address specific otitis-related questions rather than the epidemiology of otitis media per se, also have had limitations of various kinds. Some of the studies involved particular ethnic and/or socioeconomic groups largely or ex-clusively.24,44,47,49,51-5558-61,63,66,68,69 Some involved limited numbers of subjects.44-47,50,52,55,58,59,64,66,69 Some based conclusions on observation periods of less than 1 year,43,46 or on observations made less often than every 3 months53 or with variable46,60 or unspecified49,51,54,55,59 frequency. Some based diagnoses of otitis media on a finding of otorrhea or persistent ear pain only,43,45,62 or on otoscopic findings whose validity was either not tested or tested with unspecified stringency, regularity, and results,48,49,51,53,55,57,59,60 or largely on findings of tympanometry or reflectometry,47,67 or solely on information obtained from parents.63,68 Some measured the occurrence of otitis media by considering only episodes of acute otitis media,43,45,49,51,53-55,59,62,68 or only the presence of MEE,47,64,67 or only the number of tympanostomy-tube operations.56 Some investigated no putative risk factors,52,61,69 or only one or two such factors.44-4749-51,57,63-65 Finally, some studies, in implicating certain sociodemographic characteristics as risk factors, failed to include analyses for possible confounding.43,48,51,55
The Present Study
In the present study, in order to investigate possible relationships between early-life otitis media and later developmental impairment, we enrolled a large, sociodemographically diverse population, we intensively monitored their middle-ear status otoscopically throughout the first 2 years of life, we monitored the validity of the otoscopic observations on an ongoing basis, we treated infants according to specified guidelines, we used a number of measures of the occurrence of otitis media, we considered a range of presumed or putative risk factors (both for otitis media and for developmental impairment), and we performed analyses to identify interactions and to avoid confounding. In its inclusion of all of these elements, we believe the present study to be unique.
Prevalence of Otitis Media
Most studies of the prevalence of otitis media that used methods similar to ours have involved infants from low-socioeconomic-status families largely or exclusively,50,52,57,58,64,66 and the setting for several of the studies52,64,68,69 was a group day care program. Findings in these studies were generally similar to those in our 634 urban infants. Thus, Marchant et al50found that at least one episode of otitis media developed in 73% of infants by age 6 months, compared with 61.0% for urban infants in the present study, and in 77% by age 1 year, compared with 84.5% in the present study. The values found by age 2 years were 77% by Wright et al,57 91.9% by Casselbrant et al,66 and 99.0% by Zeisel et al,69 compared with 93.4% in our urban infants. In two studies by Roberts and colleagues,52,64 the median cumulative proportions of days with MEE during the first year of life were 12.1% and 26.5%, respectively, compared with 23.4% in our urban infants, and during the second year of life, 14.8% and 18.4%, respectively, compared with 15.2% in our urban infants.
To our knowledge, the occurrence of otitis media in middle-class infants has been described in only two large-scale studies that used methods similar to ours.60,61 In the study reported by Roland et al,61 483 infants in a private pediatric practice in Dallas were monitored during the age period 6 to 18 months. The infants' mean cumulative proportion of days with MEE during that age period was 19.7%, compared with values, for the same age period in the present study, of 19.7% in small town/rural infants and 16.9% in suburban infants.
In the study reported by Teele et al,60 pediatricians at two centers in Greater Boston followed 698 infants for primary care from age <3 months to age ≥3 years. More than 95% of the infants were white; 62% were enrolled in a mainly higher-socioeconomic-status suburban private practice and 38% in a mainly lower-socioeconomic-status urban health center. (In comparison, of our small town/rural infants, 97.6% were white and 64.6% had private health insurance; of our suburban infants, 99.4% were white and 95.8% had private health insurance). The proportion of the Boston infants who developed ≥1 episode of otitis media by age 6 months was approximately 36%; by age 1 year, 63%; and by age 2 years, 80%. Corresponding values in the present study were 46%, 76.4%, and 90.4%, respectively, in the small town/rural infants, and 37.9%, 77.1%, and 90.0% respectively, in the suburban infants. In the Boston infants, the mean cumulative proportions of days with MEE during the first and second years of life were 8.1% and 7.4%, respectively, and the median proportions, 4.5% and 6.2%, respectively. In the present study, in the small town/rural infants, the mean proportions during the first and second years of life were 19.0% and 15.6%, respectively, and the median proportions, 14.8% and 11.2%, respectively. And in the present study in the suburban infants, the corresponding mean proportions were 15.4% and 13.9%, respectively, and the median proportions, 12.5% and 11.2%, respectively.
In summary, findings in the present study confirm the high prevalence of otitis media described previously in infants from economically disadvantaged families, and suggest that the prevalence may be higher than described previously in infants from middle-class families. The reasons for the reported differences in prevalence between studies remains speculative; the differences may have been real, perhaps due, in turn, to populational or secular differences, or they may have been attributable to differences in either the intensity of surveillance, the accuracy of disease detection, the methods for estimating the duration of MEE during intervals between visits, or combinations of these factors.
Sociodemographic Factors in Relation to Prevalence
Sex
Some previous studies32,38,41,48,51,53,54,56,60 found the occurrence of otitis media greater in boys than in girls, whereas others23,24,29,35,47,57,58,59,66,67,70,71 found no sex-related differences. The latter studies had, on average, much smaller sample sizes, and many relied on tympanometric24,47,67,71 or questionnaire29assessment of the presence or absence of otitis media. In the present study, male infants consistently had higher mean cumulative proportions of days with MEE than female infants, even after controlling for confounding variables, but the differences were not large.
Race
Of the 11 studies,23,25,35,36,46,48,50,57,58,66,67 to our knowledge, in which the relative occurrence of otitis media in white and in black children was addressed, all but four35,58,66,67 found the occurrence of otitis media to be greater in white children (one67 found the occurrence greater in black children). In the present study, at the two urban, largely low-socioeconomic-status practice sites that included substantial numbers of both black infants and white infants, the cumulative proportion of days with MEE during the first year of life was higher in black infants, but during the second year of life the values in black and in white infants were identical. Findings in previous studies suggesting higher rates in white than in black children seem likely to have been due to differences between black and white children in rates of observation of the children, or in accuracy of detection of otitis media, or in both.
Birth Weight and Maternal Age
Based on the analyses described previously, the inverse relationship we found between cumulative days with MEE, and birth weight and maternal age, respectively, appears attributable in each instance to commonly underlying socioeconomic circumstances.
Season of Birth
Only two studies, to our knowledge, have investigated relationships between the occurrence of otitis media and the season of birth. One study35 found the occurrence greatest in children born in late summer or fall; the other study60found no relationship. Overall in our study, the cumulative proportion of days with MEE during the first year of life was highest in infants born during summer and lowest in infants born during winter, but the difference was small, and after adjustment for site, the association was no longer significant. One must conclude that season of birth has little if any bearing on the occurrence of otitis media.
Socioeconomic Status
Poverty has long been considered an important risk factor for both the development and the severity of otitis media,19 but more recent studies are divided on whether associations exist. A number of studies19,22,32,39,43,53,54,62,64 have reported an inverse relationship between the prevalence of otitis and socioeconomic status, a number of studies23,24,28,34,44,55,59,60,68,70,72 have found no relationship, and one study67 found the prevalence at ages 12 to 18 months actually higher in infants with higher socioeconomic status. In the present study we found an inverse and strong relationship, during both the first and second years of life, between the cumulative proportion of days with MEE and a socioeconomic index based on the type of health insurance and the level of maternal education.
In addition, within individual strata defined by the socioeconomic index, consistent differences between study site groupings in the cumulative proportion of days with MEE during the first year of life (Table 3) indicated that site-related differences beyond those indicated by race, education, and type of insurance were also operative. Those differences were likely socioecomic in nature.
Breastfeeding
Over the past 45 years a large number of studies in normal infants20,21,26,30,32,37,38,41,43,45,48,49,53,54,60,67,68,71,73and one study in infants with cleft palate74 have reported a protective effect of breast milk feeding against otitis media. A lesser number of studies27-29,34,39,55,59,62 have not found such an effect. In the present study, although our univariate analysis showed an inverse relationship between the duration of breastfeeding and the occurrence of otitis media during both the first and second years of life, after adjusting for other statistically significant variables a significant relationship remained only during the first year of life. Although our measure of the extent of breastfeeding was not precise, our findings indicate that the protective effect of breast milk is, at best, limited. Nonetheless, the findings may be viewed as further supporting the undisputed desirability of breast milk feeding.
Exposure to Tobacco Smoke
A positive relationship between the occurrence of otitis media and household exposure to tobacco smoke has been found in some studies27,41,54,55,62,65,67,73,75-77 but not in others.28,34,59,60,66,68,71,78 In these studies generally, in addition to the methodologic limitations noted previously, measures used to estimate the degree of exposure have been problematic. One cross-sectional study in 7-year-olds77 used a single salivary cotinine assay to estimate each subject's degree of exposure, and one prospective study65 used a single serum cotinine assay, near the first birthday, to estimate the degree of exposure in relation to the cumulative occurrence of otitis media over the first 3 years of life. In the present study, after adjusting for other statistically significant variables, we found a positive relationship between the number of smokers in the household and the occurrence of otitis media during the first, but not the second, year of life. As is well known and as we also found, household smoking correlates inversely with socioeconomic status. Thus, confounding by socioeconomic status makes difficult an accurate assessment of the true state of relationships, particularly given the lack of precision in the various estimates involved. Again, our findings may be viewed as further supporting an undisputed desirability–that of avoiding exposure of infants to cigarette smoke.
Exposure to Other Children, at Home and in Day Care
Many studies32,41,43,53-56,66,71,73 have reported a positive relationship between the occurrence of otitis media and the number of older siblings or other children in the household, whereas other studies29,34,39,44,59,60 have failed to find such a relationship. In contrast, except for one study59 in which findings were equivocal, all studies,24,29,31,34,39,41,49,51,53-56,71,73 to our knowledge, that explored a positive relationship between the occurrence of otitis media and group day care attendance found such a relationship. As described previously, by combining the two loci of exposure into a single index, we found a strong, positive relationship between the cumulative proportion of days with MEE and the extent of exposure to other children. What appears to matter most is notwhere but how many.
Study Limitations
Study Population
Categories of infants not available to us for study in substantial numbers were middle-class black infants, as well as Native American, Hispanic, and Asian infants of any socioeconomic category. Extrapolation of study findings to those population groups may therefore not be warranted.
Subject Selection
Parents who declined to enroll their infants in this study tended to be of higher socioeconomic status overall than parents who enrolled their infants, a circumstance that likely skewed our findings in the direction of greater prevalence of otitis media. On the other hand, our findings likely were skewed in the opposite direction by two circumstances that kept certain infants with risk factors for otitis media from being included: first, many of the infants not enrolled in the study were not enrolled because they were, or soon would be, in day care (and had working parents unable or unwilling to meet the required schedule of monthly visits for 3 years); and second, enrolled infants from lower-socioeconomic-status families were the ones most likely to have been lost to follow-up, or to have had incomplete data sets, so that their data could not be included in the full-year findings. In any case, we believe that the study included large enough numbers of children in each of the socioeconomic categories to permit confidence in the findings regarding each category.
Estimates of MEE Duration
Our study relies on estimates of MEE duration. These estimates are derived from interpolations concerning infants' middle-ear status during the intervals between examinations, and, in turn, on determinations of the presence or absence of MEE on examination by some 35 clinicians at eight study sites. Clearly there is room for error in such estimates. Nonetheless, we believe that more intensive surveillance of such a large number of infants would not be practicable, that our data on interobserver reliability provide assurance concerning the accuracy of diagnoses generally, and that our interpolation scheme is reasonable. Of course, ethical considerations precluded our withholding treatment that we considered appropriate, but the overall pattern of disease occurrence would have been little influenced by the relatively small number of tube placements that were carried out.
Type of Otitis Media
In the present report we have not distinguished between episodes of acute otitis media and episodes of otitis media with effusion because our study-team clinicians do not all use the same criteria in making that distinction, and because the distinction is likely not germane to the main objective of the parent study, which is to analyze relationships, if any, between cumulative duration of MEE and later developmental outcome. The distinction is nonetheless important clinically, we believe, in everyday care, in deciding on the advisability of antimicrobial treatment.4,13
Putative Risk Factors and the Multivariate Model
Because we did not plan the present study to investigate the epidemiology of otitis media per se, we did not obtain information about certain putative risk factors—eg, a family history of otitis media or of allergy—often considered important, nor did we attempt to quantify, beyond the measures we used, the extent of breastfeeding and of exposure to tobacco smoke. Other measures that we used—eg, our socioeconomic index—may also not have been sensitive enough to detect subtle but important additional differences. The multivariate model we used may have been limited accordingly, and may have been further limited by our having not included other, as yet unidentified variables associated with the development of otitis media. Such limitations are to some extent inherent in epidemiologic studies that attempt to link disease outcomes to putative risk factors.79Notwithstanding the limitations of the multivariate model, the fact that certain variables (race, birth weight, maternal age, and season of birth) were found to be no longer significantly related to the cumulative duration of MEE after adjustment for study site—a factor encompassing important socioeconomic differences—underscores the potential for confounding that exists in attempting to relate disease outcomes to specific sociodemographic variables.
Clinical Importance and Generalizability
Continuing controversy concerning associations between the occurrence of otitis media and certain putative risk factors can be readily ascribed to the various limitations of the studies discussed previously. Because we designed and carried out the present study specifically to avoid many of those limitations, and because the number of infants we studied was large and diverse, we believe that many of our findings are both conclusive and generalizable. Four of our findings seem the most important. First, although the prevalence of otitis media among white middle-class infants is well known, our findings suggest that it may be even greater than commonly assumed. Second, in contrast with findings in most previous studies of the question, we found the prevalence of otitis media in black lower-socioeconomic-status infants to be as high, if not higher, than in white lower-socioeconomic-status infants, and certainly higher than in white middle-class infants. Third, we affirmed a strong inverse relationship between the occurrence of otitis media and socioeconomic status. And fourth, we affirmed a strong positive relationship between the occurrence of otitis media and the number of other children to whom an infant is habitually exposed, whether in day care or at home.
We believe that the information provided by this study will be useful to clinicians caring for children, and may also be useful to health care planners in assessing and forecasting health care utilization and costs. Ultimately, what is needed is better understanding of the long-term impact, if any, of early-life otitis media, a question that we are currently addressing in the parent study.3
ACKNOWLEDGMENTS
This work was supported by grant HD26026 from the National Institute for Child Health and Human Development and the Agency for Health Care Policy and Research, and by a gift from SmithKline Beecham Laboratories.
We are indebted to the following pediatricians who made the decisions, participated in the planning, and assisted in the efforts to incorporate this study into their practices, and who, at no small inconvenience and cost, have provided unflagging support for study activities–at Beaver: David J. Cahill, MD, James Scibilia, MD, and Julius A. Vogel Jr, MD; at Brentwood, Mark Diamond, MD, and Thomas D. Skelly, MD; at Gibsonia: Amelia V. Agustin, MD, and Eva A. Vogeley, MD; at Kittanning: Harold A. Altman, MD, James K. Greenbaum, MD, Kenneth R. Keppel, MD, and Donald J. Vigliotti, MD; at Mt Lebanon: Scott L. Tyson, MD, and Celeste J. Welkon, MD; at Pleasant Hills: K. Gopalkrishna Pai, MD, and Harvey M. Rubin, MD, and at The Mercy Hospital of Pittsburgh: Bradley J. Bradford, MD.
In addition to those above, and those named as authors, the following persons served as study-team clinicians–at Children's Hospital of Pittsburgh: Irene Fabian, CRNP, Nancy J. Guerra, CRNP, and Alejandro Hoberman, MD; at Beaver: Allen H. Chamovitz, MD, Sharon N. Cowden, MD, Valentina E. DiCenzo, MD, Verda S. Graf, BS, PA, S. Nasrin Ghorbanian, MD, George R. Haddad, MD, Janet D. Liljestrand, MD, and Jennifer J. Momen, MD; at Brentwood: Norman L. Cohen, MD, Kristin L. Frederick, MD, Joan Schiebel, RN, and Brenda E. Watkins, MD; at Kittanning, Tracy Balentine, RN, Shirley Baum, CRNP, Lawrence J. Butler, MD, Thomas G. Lynch, MD, and JoAnn Nickleach, MD; at Mt Lebanon: Barbara J. Bahl, CRNP, Barbara Braman, CRNP, M. Bridgetta Devlin, CRNP, Holly A. Frost, MD, Lisa M. Hakos-Zoffel, CRNP, Thelma L. Herlich, MD, and Elizabeth H. Michael, CRNP; at Pleasant Hills: Todd H. Wolynn, MD; and at The Mercy Hospital of Pittsburgh: Barbara L. Ayars, MD, Kimberly Brown, MD, Michael J. Daly, MD, Karla Falcon, MD, Pamela Heald, CRNP, Cynthia M. Hoess, MD, Barbara L. McNulty, CRNP, Yolanda Moore-Forbes, MD, Charles A. Pohl, MD, Sharon M. Roncevich, MD, Sherrill J. Rudy, CRNP, Evelyn J. Schmidt, RN, Sarah H. Springer, MD, and Karen S. Vargo, MD. We are also grateful to the many Children's Hospital of Pittsburgh and Mercy Hospital of Pittsburgh pediatric house officers who served as primary care clinicians for study subjects and whose collaboration was essential for the successful conduct of the study.
In addition to the authors, the following persons participated integrally in the planning and assisted in the implementation of this study: Charles D. Bluestone, MD; Thomas F. Campbell, PhD; Christine A. Dollaghan, PhD; Heidi M. Feldman, PhD, MD; Robert J. Nozza, PhD; and Diane L. Sabo, PhD.
The following people also assisted in the study: Ida Smith (clinic support coordinator); Miki Rakay (volunteer patient liaison specialist); Robin L. Lavelle and Valerie S. Quickley (schedulers); Jennifer A. Aliberti, Sandra Barnett, Nanci Barrett, Cindy Brown, Nancy Ciaburri, Christin E. Costella, Cyndi Getty, SueEllen Hall, Jerome Hill, Anthony Heard, Karen Horox, Isabel Hunter, Beverly Joyce, Deborah Klemm, Judy Lazzeri, Jean Martin, Sue Musser, Karen M. Noto, Shirley Petrie, Deborah M. Pettibon, Dawn M. Rone, Brenda Shaffer, Sunitha Somanath, Dana Valasek, and Sheila Vasbinder (study technicians); Sharon M. Caputo, BS; Ron F. Hollis, BS; and Stephen Sefcik, BS (programmers); Kathleen A. Cecotti, BA; Sharon A. DiBridge, BS; Charlotte Heller, Toni L. McKeever, and Karen S. Pourboghrat, BA (research/data assistants); Jennifer S. Dietrich, L. Annabelle Kyle, and Robin E. Rice, BS (administration); Dayna L. Pitcairn, MA (developmental test coordinator); and Margaretha L. Casselbrant, MD, PhD; Kenny H. Chan, MD; Joseph E. Dohar, MD; Margaret A. Kenna, MD; J. Christopher Post, MD; Sylvan E. Stool, MD; and Robert F. Yellon, MD (otolaryngologist consultants).
Finally, we are grateful to SmithKline Beecham Pharmaceuticals and to Abbott Laboratories, Pfizer Inc, and The Upjohn Company for supplying antimicrobial drugs for study subjects who developed otitis media, and to Grason-Stadler Inc, a Welch Allyn Company, for supplying special paper for recording tympanograms.
Footnotes
- Received April 1, 1996.
- Accepted May 24, 1996.
Reprint requests to (J.L.P.) Children's Hospital of Pittsburgh, One Children's Place, Pittsburgh, PA 15213–2583.
Presented in part at the combined annual meeting of the American Pediatric Society and the Society for Pediatric Research, San Diego, CA, May 7–11, 1995.
- MEE =
- middle-ear effusion
REFERENCES
- Copyright © 1997 American Academy of Pediatrics
In this issue
Jump to section
Related Articles
Cited By...
- Extended high-frequency hearing enhances speech perception in noise
- How Young Children Learn Language and Speech
- Treatment of otitis media by transtympanic delivery of antibiotics
- Autoinflation reduces middle ear effusion in children with otitis media with effusion
- Early Psychosocial Exposures, Hair Cortisol Levels, and Disease Risk
- Race, Otitis Media, and Antibiotic Selection
- Variation in Antimicrobial Prescribing for Otitis Media by Race: Different Wrinkle in Disparity?
- Xylitol Syrup for the Prevention of Acute Otitis Media
- Fifteen-minute consultation: the infant with frequent infections
- The Impoverished Brain: Disparities in Maternal Education Affect the Neural Response to Sound
- Role of the Zinc Uptake ABC Transporter of Moraxella catarrhalis in Persistence in the Respiratory Tract
- Racial Differences in Antibiotic Prescribing by Primary Care Pediatricians
- The Diagnosis and Management of Acute Otitis Media
- Pediatric Office Pain: Crying for Attention
- Quality Measures for the Care of Children With Otitis Media With Effusion
- Construction and Immunogenicity of Recombinant Adenovirus Vaccines Expressing the HMW1, HMW2, or Hia Adhesion Protein of Nontypeable Haemophilus influenzae
- Indirect Pathogenicity of Haemophilus influenzae and Moraxella catarrhalis in Polymicrobial Otitis Media Occurs via Interspecies Quorum Signaling
- Prolonged and Exclusive Breastfeeding Reduces the Risk of Infectious Diseases in Infancy
- LuxS Promotes Biofilm Maturation and Persistence of Nontypeable Haemophilus influenzae In Vivo via Modulation of Lipooligosaccharides on the Bacterial Surface
- Antibodies Specific for the Hia Adhesion Proteins of Nontypeable Haemophilus influenzae Mediate Opsonophagocytic Activity
- New Vaccines Against Otitis Media: Projected Benefits and Cost-effectiveness
- Overuse of tympanostomy tubes in New York metropolitan area: evidence from five hospital cohort
- Acute otitis media in the first two months of life: characteristics and diagnostic difficulties
- Incidence of Acute Otitis Media and Sinusitis Complicating Upper Respiratory Tract Infection: The Effect of Age
- Reduction of Frequent Otitis Media and Pressure-Equalizing Tube Insertions in Children After Introduction of Pneumococcal Conjugate Vaccine
- Is Breast Really Best? Risk and Total Motherhood in the National Breastfeeding Awareness Campaign
- Cost-Effectiveness Analysis of Treatment Options for Acute Otitis Media
- Antibodies Specific for the High-Molecular-Weight Adhesion Proteins of Nontypeable Haemophilus influenzae Are Opsonophagocytic for both Homologous and Heterologous Strains
- Tympanometric Findings and the Probability of Middle-Ear Effusion in 3686 Infants and Young Children
- Chinchilla and Murine Models of Upper Respiratory Tract Infections with Respiratory Syncytial Virus
- Understanding the dimensions of socioeconomic status that influence toddlers' health: unique impact of lack of money for basic needs in Quebec's birth cohort
- Tympanic Membrane Abnormalities and Hearing Levels at the Ages of 5 and 6 Years in Relation to Persistent Otitis Media and Tympanostomy Tube Insertion in the First 3 Years of Life: A Prospective Study Incorporating a Randomized Clinical Trial
- Diagnosis and Management of Acute Otitis Media
- Otitis Media With Effusion
- Human Antibodies Specific for the High-Molecular-Weight Adhesion Proteins of Nontypeable Haemophilus influenzae Mediate Opsonophagocytic Activity
- Trends in Otitis Media Among Children in the United States
- Otitis Media and Tympanostomy Tube Insertion During the First Three Years of Life: Developmental Outcomes at the Age of Four Years
- Effects of low income on infant health
- Does Adenoidectomy Prevent Recurrent Otitis in Young Children?
- Promoting the Appropriate Use of Oral Antibiotics: There Is Some Very Good News
- Diagnostic Accuracy, Tympanocentesis Training Performance, and Antibiotic Selection by Pediatric Residents in Management of Otitis Media
- Black-White Differences in Health Care Utilization Among US Children With Frequent Ear Infections
- Outpatient and Hospital Visits Associated With Otitis Media Among American Indian and Alaska Native Children Younger Than 5 Years
- Pediatric Resident Training in the Diagnosis and Treatment of Acute Otitis Media
- Positional Otitis Media and Otorrhea After Tympanostomy-Tube Placement
- Respiratory Infections in Infants: Interaction of Parental Allergy, Child Care, and Siblings---The PIAMA Study
- Otorrhea in Young Children After Tympanostomy-Tube Placement for Persistent Middle-Ear Effusion: Prevalence, Incidence, and Duration
- Amoxicillin for otitis media in general practice
- Language, Speech Sound Production, and Cognition in Three-Year-Old Children in Relation to Otitis Media in Their First Three Years of Life
- Parental Stress and Parent-rated Child Behavior in Relation to Otitis Media in the First Three Years of Life
- Parent-reported Language and Communication Skills at One and Two Years of Age in Relation to Otitis Media in the First Two Years of Life
- Day Care Attendance in the First Year of Life and Illnesses of the Upper and Lower Respiratory Tract in Children With a Familial History of Atopy
- Epidemiology of Otitis Media Onset by Six Months of Age
- Universal Newborn Hearing Screening: Should We Leap Before We Look?
- A Longitudinal Study of Otitis Media With Effusion Among 2- to 5-Year-Old African-American Children in Child Care
- Otitis Media, the Caregiving Environment, and Language and Cognitive Outcomes at 2 Years
- The Journal in 1947 and 1997: A Dramatic Change
- The Rational Use of Antimicrobials in Acute Otitis Media: A Bacteriologic Investigation of Otitis Media in Infancy, by E. A. Mortimer Jr, and R. L. Watterson Jr, Pediatrics, 1956;17:359-366; Otitis Media in the Practice of Pediatrics: Bacteriological and Clinical Observations, by J. D. Coffey Jr,Pediatrics, 1966;38:25-32; Acute Otitis Media: Treatment Results in Relation to Bacterial Etiology, by B. W. Nilson, et al, Pediatrics, 1969;43:351-358; The "In Vivo Sensitivity Test"--Bacteriology of Middle Ear Exudate During Antimicrobial Therapy in Otitis Media, by V. M. Howie and J. H. Ploussard, Pediatrics, 1969;44:940-944; andOtitis Media: A Clinical and Bacteriological Correlation, by V. M. Howie, et al, Pediatrics, 1970;45:29-35