Clinical Wisdom and the Evidence Base Otitis Media with Effusion

In this issue of the Journal, van Balen and de Melker¹ provide new insights into the natural history of otitis media with effusion (OME). Since family physicians often manage the early stages of illness, we are interested in both diagnosis and prognosis, with prognosis frequently being more important. OME is diagnosable in a large portion of children. It is diagnosable at some point during each year in up to 97% of young children in daycare, for example, and is present for the average child approximately 20% of the time.^2,3

This suggests that rather than viewing OME as distinct pathology, it should be considered part of the natural course of events: a physiologic phenomenon that reflects eustachian tube obstruction with secondary fluid accumulation in the middle ear, which commonly follows upper respiratory tract infections (URIs) or acute otitis media. >From a primary care clinical perspective, what distinguishes OME as a physiologic process from OME as a pathologic process-a disease-is not the anatomic characteristics of the ear or the nature of the fluid, but the prognosis.

The importance of prognosis

The natural history of OME includes spontaneous clearing of the fluid in 60% to 65% of children every 3 months, with an even higher rate of clearing if the fluid occurs following acute otitis media.⁴ Among those children with persistent effusion, only a minority have significant hearing deficits bilaterally, resulting in only a few requiring intervention, possibly less than 5% of those with OME.⁵ We only need to be concerned about the child who is destined to develop an adverse outcome of consequence. That is the significance of the search for prognostic discriminators, the focus of the work of van Balen and de Melker.

First, we need to consider what the possible adverse outcomes are; persistent fluid by itself may be of little consequence. The ultimate outcomes that suggest that OME is clinically important are persistent hearing deficits and the conjectured (yet to be supported by strong evidence) long-term impairment of the development of speech and language resulting from such deficits. The Agency for Health Care Policy and Research panel on OME⁶ recommended that children have their hearing tested before they are treated for persistent OME, and only those with bilateral hearing deficits should be considered for the insertion of tympanostomy tubes. However, because of the difficulty in testing hearing in young children, the identification of those with persistent OME has evolved as a surrogate predictor of persistent hearing deficits to limit the number of children requiring a hearing evaluation.

Thus the real prognostic question of interest is: Which young children are likely to have a persistent hearing deficit leading to impaired language development? If the ultimate outcome of language development is the focus, the presence of the ameliorating behaviors of the family and other caregivers, such as daily reading with the child, becomes as important as the presence of fluid.⁷

Predictors of persistent ome

Within this context, the van Balen and de Melker study provides 2 major advances in our understanding: one based on their clinical wisdom and the other based on the results of their data analyses. Perhaps their more important contribution is the former. In designing their study-possibly to decrease the workload of the physicians involved-they sought to select a group of children likely to have bilateral OME during routine office visits to family physicians. Their clinical experience and review of the literature led them to use a selection strategy that included children with hearing loss (including subjective), language and speech problems, mouth breathing and snoring, a history of recurrent URI (6 or more episodes in 12 months), a family history of otitis media, and acute otitis 6 weeks previously. Children with these characteristics represent a very different group of children from those with asymptomatic OME discovered by routine screening or as a chance observation during 1 of the numerous preventive health care visits that occur during the first years of life.

An examination of these inclusion criteria suggests that most are likely to be indicators of significant persistent eustachian tube obstruction, leading to persistent OME. Thus, at study entry the investigators had already taken a major step toward identifying a group of children with the prognosis of persistent OME. The value of their inclusion criteria as diagnostic and prognostic indicators is verified by their data. Eighty-four percent had bilateral OME at the initial visit; the expected rate is 20% to 30% among all young children visiting family physicians.³ Seventy-six percent had persistent OME at a 3-month follow-up visit, with three quarters of these having bilateral OME. Thus, the inclusion criteria were powerful predictors of persistent OME.

Given the prognostic power of the set of variables included in their selection strategy, it is not surprising that a very limited number of additional potential prognostic factors contributed to prognosis. The identification of these additional factors is the evidence-based contribution of their study. Recurrent URIs, and in their absence intact adenoids, summer-fall presentation, and a history of otitis during the first year of life all marginally increase the likelihood of persistent effusion, although the discriminant power of these factors is low. However, these items and the inclusion criteria may be important and possibly discriminating factors in the general population of children visiting family physicians and pediatricians.

Future research

What remains for future investigation is the evaluation of the discriminating capacity and screening characteristics—both sensitivity and specificity—of each of the identification factors when applied to general primary care populations. For now, the most important take-home message might be that the inclusion criteria coupled with URI at follow-up are highly predictive of persistent OME and deserving of further patient follow-up and a hearing evaluation.

In this issue of the Journal, van Balen and de Melker¹ provide new insights into the natural history of otitis media with effusion (OME). Since family physicians often manage the early stages of illness, we are interested in both diagnosis and prognosis, with prognosis frequently being more important. OME is diagnosable in a large portion of children. It is diagnosable at some point during each year in up to 97% of young children in daycare, for example, and is present for the average child approximately 20% of the time.^2,3

This suggests that rather than viewing OME as distinct pathology, it should be considered part of the natural course of events: a physiologic phenomenon that reflects eustachian tube obstruction with secondary fluid accumulation in the middle ear, which commonly follows upper respiratory tract infections (URIs) or acute otitis media. >From a primary care clinical perspective, what distinguishes OME as a physiologic process from OME as a pathologic process-a disease-is not the anatomic characteristics of the ear or the nature of the fluid, but the prognosis.

The importance of prognosis

The natural history of OME includes spontaneous clearing of the fluid in 60% to 65% of children every 3 months, with an even higher rate of clearing if the fluid occurs following acute otitis media.⁴ Among those children with persistent effusion, only a minority have significant hearing deficits bilaterally, resulting in only a few requiring intervention, possibly less than 5% of those with OME.⁵ We only need to be concerned about the child who is destined to develop an adverse outcome of consequence. That is the significance of the search for prognostic discriminators, the focus of the work of van Balen and de Melker.

First, we need to consider what the possible adverse outcomes are; persistent fluid by itself may be of little consequence. The ultimate outcomes that suggest that OME is clinically important are persistent hearing deficits and the conjectured (yet to be supported by strong evidence) long-term impairment of the development of speech and language resulting from such deficits. The Agency for Health Care Policy and Research panel on OME⁶ recommended that children have their hearing tested before they are treated for persistent OME, and only those with bilateral hearing deficits should be considered for the insertion of tympanostomy tubes. However, because of the difficulty in testing hearing in young children, the identification of those with persistent OME has evolved as a surrogate predictor of persistent hearing deficits to limit the number of children requiring a hearing evaluation.

Thus the real prognostic question of interest is: Which young children are likely to have a persistent hearing deficit leading to impaired language development? If the ultimate outcome of language development is the focus, the presence of the ameliorating behaviors of the family and other caregivers, such as daily reading with the child, becomes as important as the presence of fluid.⁷

Predictors of persistent ome

Within this context, the van Balen and de Melker study provides 2 major advances in our understanding: one based on their clinical wisdom and the other based on the results of their data analyses. Perhaps their more important contribution is the former. In designing their study-possibly to decrease the workload of the physicians involved-they sought to select a group of children likely to have bilateral OME during routine office visits to family physicians. Their clinical experience and review of the literature led them to use a selection strategy that included children with hearing loss (including subjective), language and speech problems, mouth breathing and snoring, a history of recurrent URI (6 or more episodes in 12 months), a family history of otitis media, and acute otitis 6 weeks previously. Children with these characteristics represent a very different group of children from those with asymptomatic OME discovered by routine screening or as a chance observation during 1 of the numerous preventive health care visits that occur during the first years of life.

An examination of these inclusion criteria suggests that most are likely to be indicators of significant persistent eustachian tube obstruction, leading to persistent OME. Thus, at study entry the investigators had already taken a major step toward identifying a group of children with the prognosis of persistent OME. The value of their inclusion criteria as diagnostic and prognostic indicators is verified by their data. Eighty-four percent had bilateral OME at the initial visit; the expected rate is 20% to 30% among all young children visiting family physicians.³ Seventy-six percent had persistent OME at a 3-month follow-up visit, with three quarters of these having bilateral OME. Thus, the inclusion criteria were powerful predictors of persistent OME.

Given the prognostic power of the set of variables included in their selection strategy, it is not surprising that a very limited number of additional potential prognostic factors contributed to prognosis. The identification of these additional factors is the evidence-based contribution of their study. Recurrent URIs, and in their absence intact adenoids, summer-fall presentation, and a history of otitis during the first year of life all marginally increase the likelihood of persistent effusion, although the discriminant power of these factors is low. However, these items and the inclusion criteria may be important and possibly discriminating factors in the general population of children visiting family physicians and pediatricians.

Future research

What remains for future investigation is the evaluation of the discriminating capacity and screening characteristics—both sensitivity and specificity—of each of the identification factors when applied to general primary care populations. For now, the most important take-home message might be that the inclusion criteria coupled with URI at follow-up are highly predictive of persistent OME and deserving of further patient follow-up and a hearing evaluation.

In this issue of the Journal, van Balen and de Melker¹ provide new insights into the natural history of otitis media with effusion (OME). Since family physicians often manage the early stages of illness, we are interested in both diagnosis and prognosis, with prognosis frequently being more important. OME is diagnosable in a large portion of children. It is diagnosable at some point during each year in up to 97% of young children in daycare, for example, and is present for the average child approximately 20% of the time.^2,3

This suggests that rather than viewing OME as distinct pathology, it should be considered part of the natural course of events: a physiologic phenomenon that reflects eustachian tube obstruction with secondary fluid accumulation in the middle ear, which commonly follows upper respiratory tract infections (URIs) or acute otitis media. >From a primary care clinical perspective, what distinguishes OME as a physiologic process from OME as a pathologic process-a disease-is not the anatomic characteristics of the ear or the nature of the fluid, but the prognosis.

The importance of prognosis

The natural history of OME includes spontaneous clearing of the fluid in 60% to 65% of children every 3 months, with an even higher rate of clearing if the fluid occurs following acute otitis media.⁴ Among those children with persistent effusion, only a minority have significant hearing deficits bilaterally, resulting in only a few requiring intervention, possibly less than 5% of those with OME.⁵ We only need to be concerned about the child who is destined to develop an adverse outcome of consequence. That is the significance of the search for prognostic discriminators, the focus of the work of van Balen and de Melker.

First, we need to consider what the possible adverse outcomes are; persistent fluid by itself may be of little consequence. The ultimate outcomes that suggest that OME is clinically important are persistent hearing deficits and the conjectured (yet to be supported by strong evidence) long-term impairment of the development of speech and language resulting from such deficits. The Agency for Health Care Policy and Research panel on OME⁶ recommended that children have their hearing tested before they are treated for persistent OME, and only those with bilateral hearing deficits should be considered for the insertion of tympanostomy tubes. However, because of the difficulty in testing hearing in young children, the identification of those with persistent OME has evolved as a surrogate predictor of persistent hearing deficits to limit the number of children requiring a hearing evaluation.

Thus the real prognostic question of interest is: Which young children are likely to have a persistent hearing deficit leading to impaired language development? If the ultimate outcome of language development is the focus, the presence of the ameliorating behaviors of the family and other caregivers, such as daily reading with the child, becomes as important as the presence of fluid.⁷

Predictors of persistent ome

Within this context, the van Balen and de Melker study provides 2 major advances in our understanding: one based on their clinical wisdom and the other based on the results of their data analyses. Perhaps their more important contribution is the former. In designing their study-possibly to decrease the workload of the physicians involved-they sought to select a group of children likely to have bilateral OME during routine office visits to family physicians. Their clinical experience and review of the literature led them to use a selection strategy that included children with hearing loss (including subjective), language and speech problems, mouth breathing and snoring, a history of recurrent URI (6 or more episodes in 12 months), a family history of otitis media, and acute otitis 6 weeks previously. Children with these characteristics represent a very different group of children from those with asymptomatic OME discovered by routine screening or as a chance observation during 1 of the numerous preventive health care visits that occur during the first years of life.

An examination of these inclusion criteria suggests that most are likely to be indicators of significant persistent eustachian tube obstruction, leading to persistent OME. Thus, at study entry the investigators had already taken a major step toward identifying a group of children with the prognosis of persistent OME. The value of their inclusion criteria as diagnostic and prognostic indicators is verified by their data. Eighty-four percent had bilateral OME at the initial visit; the expected rate is 20% to 30% among all young children visiting family physicians.³ Seventy-six percent had persistent OME at a 3-month follow-up visit, with three quarters of these having bilateral OME. Thus, the inclusion criteria were powerful predictors of persistent OME.

Given the prognostic power of the set of variables included in their selection strategy, it is not surprising that a very limited number of additional potential prognostic factors contributed to prognosis. The identification of these additional factors is the evidence-based contribution of their study. Recurrent URIs, and in their absence intact adenoids, summer-fall presentation, and a history of otitis during the first year of life all marginally increase the likelihood of persistent effusion, although the discriminant power of these factors is low. However, these items and the inclusion criteria may be important and possibly discriminating factors in the general population of children visiting family physicians and pediatricians.

Future research

What remains for future investigation is the evaluation of the discriminating capacity and screening characteristics—both sensitivity and specificity—of each of the identification factors when applied to general primary care populations. For now, the most important take-home message might be that the inclusion criteria coupled with URI at follow-up are highly predictive of persistent OME and deserving of further patient follow-up and a hearing evaluation.