Rosenfeld RM, Bhaya MH, Bower CM, Brookhouser PE, Casselbrant ML, Chan KH, Cunningham MJ, Derkay CS, Gray SD, Manning SC, Messner AH, Smith RJH. Impact of Tympanostomy Tubes on Child Quality of Life. Arch Otolaryngol Head Neck Surg. 2000;126(5):585-592. doi:10.1001/archotol.126.5.585
Copyright 2000 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2000
The objective benefits of tympanostomy tubes for otitis media are well established, but the subjective impact of surgery on child quality of life (QOL) has not been systematically studied.
To determine the subjective impact of tympanostomy tubes on child QOL, and to compare the variability in QOL before surgery with that observed after surgery.
Prospective, observational, before-and-after trial.
Fourteen referral-based pediatric otolaryngology practices in the United States.
Consecutive (64%) and convenience (36%) sample of 248 children (median age, 1.4 years) with otitis media scheduled for bilateral tympanostomy tube placement as an isolated surgical procedure.
Tympanostomy tubes were inserted as part of routine clinical care. Validated measures of QOL (OM-6 survey), satisfaction with health care decision (Satisfaction With Decision Scale), and satisfaction with office visit; surveys were completed at baseline (visit 1), at surgery (visit 2), and after surgery (visit 3).
Main Outcome Measures
Short-term changes in QOL before surgery (visit 1 to visit 2) and after surgery (visit 2 to visit 3).
Changes in QOL before surgery were mostly trivial, and were smaller than changes observed after surgery (P<.001). Large, moderate, and small improvements in QOL occurred after surgery in 56%, 15%, and 8% of children, respectively. Physical symptoms, caregiver concerns, emotional distress, and hearing loss were most improved, but significant changes were also seen for activity limitations and speech impairment. Trivial changes occurred in 17% of children, and 4% had poorer QOL. Predictors of poorer QOL were otorrhea 3 or more days (10% of variance) and decreased satisfaction with surgical decision (3% of variance). Hearing status, child age, type of otitis media (recurrent vs chronic), and office visit satisfaction were unrelated to outcome.
Tympanostomy tubes produce large short-term improvements in QOL for most children. The best outcomes occur when postoperative otorrhea is absent or minimal, and when parents are satisfied with their initial decision to have surgery. Further research is needed to document the long-term impact of tubes on child QOL.
MYRINGOTOMY WITH insertion of tube is the most common surgical procedure performed on children in the United States.1 In 1996 about 1 of every 110 children had tubes inserted, nearly double the rate of the next most frequent procedure, tonsillectomy. Although several randomized trials2- 7 have defined the objective benefits of tubes, information regarding the subjective impact is largely anecdotal.8,9 Reduced incidence and prevalence rates of otitis media may delight physicians, but children want to feel and function better, parents want to worry less about their child's health, and third-party payers want evidence that surgical interventions actually improve quality of life (QOL). Not surprisingly, the absence of systematic data regarding the subjective impact of tympanostomy tubes has fueled skepticism regarding the appropriateness of surgery.10
Quality of life is a subjective outcome that reflects the patient's perception of his or her health status.11 In the context of otitis media, it describes the net consequences of middle ear effusion and acute otitis media (AOM) on a child's daily activities, physical symptoms, social interactions, and emotional well-being. A valid and reliable measure of QOL for children with otitis media was recently developed to quantify clinically meaningful changes within patients over time.12 Preliminary data suggested moderate to large improvements in QOL after insertion of tympanostomy tubes for most children (66%), but only 37 children were studied. Furthermore, the results were difficult to interpret without a corresponding measure of the variability in QOL prior to surgery. Other QOL measures for otitis media have been reported,13,14 but have not been validated as measures of longitudinal change.
The purpose of this multicenter prospective study was to measure short-term changes in QOL for children after insertion of tympanostomy tubes, and to compare the magnitude of these changes with the variability in QOL observed before surgery. We also sought to describe the relationship, if any, of changes observed after surgery with baseline disease status and patient satisfaction. The objective benefits of tympanostomy tubes are already well-defined (eg, improved hearing, reduced prevalence of middle ear effusion, reduced incidence of AOM), and our purpose herein was not to duplicate existing research. Rather, we sought to quantify the subjective short-term impact of tubes, to facilitate evidence-based management decisions for children with chronic and recurrent otitis media.
Our study was conducted at 14 referral-based pediatric otolaryngology practices in the United States between September 1997 and June 1998. A pediatric otolaryngology practice was defined as one in which the primary physician was a member in good standing of the American Society of Pediatric Otolaryngology. Study sites were asked to recruit a consecutive sample of eligible patients. Informed consent for study participation was obtained for all patients in a manner consistent with the local institutional review board.
Specific inclusion criteria were (1) aged 6 months to 12 years, (2) scheduled for bilateral tympanostomy tube placement as an isolated surgical procedure, and (3) child accompanied by parent or primary caregiver. Specific exclusion criteria were (1) scheduled for unilateral tympanostomy tube placement only, (2) scheduled for concurrent surgery (eg, adenoidectomy), (3) middle ear pathological features other than otitis media (eg, cholesteatoma), and (4) parent or primary caregiver unable to read and understand the English language. Children with prior tympanostomy tubes were not excluded.
Data were collected at 3 consecutive times during the routine care of a child scheduled for insertion of tympanostomy tubes: the scheduling office visit (visit 1) at which a decision was made to schedule the child for bilateral tube placement, the day of surgery (visit 2), and the first postoperative visit (visit 3). The timing of the above visits was at the discretion of the study site physician, but the interval between consecutive visits was recorded for all children. We anticipated that data collection would occur over a 2- to 4-month period for most participants. At visit 2 and visit 3, the parent or caregiver who provided information at visit 1 was encouraged to provide subsequent information.
Changes in disease-specific QOL were assessed with the OM-6 survey,12 a valid and reliable measure of clinical change in 6 domains: physical suffering, hearing loss, speech impairment, emotional distress, activity limitations, and caregiver concerns. The child's proxy completes the survey and item responses are scored from 1 to 7 (higher scores indicate poorer QOL). A baseline score is calculated as the mean of the 6 item scores, and a follow-up score is similarly determined when the survey is readministered at a later date. Next, a change score is calculated by subtracting the follow-up score from the baseline score (positive values reflect improved QOL). A change score of less than 0.5 indicates trivial change, 0.5 to 0.9 indicates small change, 1.0 to 1.4 indicates moderate change, and 1.5 or greater indicates large change.
Effect sizes for QOL changes before and after surgery were determined by the standardized response mean defined as the mean change score divided by its SD.15 As a measure of change within patients, the standardized response mean permits meaningful comparisons to be made among different measures. A standardized response mean of 0.2 reflects small change; 0.5, moderate change; and 0.8 or more, large change. The standardized response mean and 95% confidence interval (CI) were calculated for individual OM-6 domains and for the overall survey score. Effect sizes after surgery (visit 2 to visit 3) were expected to be significantly greater than prior to surgery (visit 1 to visit 2) based on the demonstrated efficacy of tympanostomy tubes for chronic and recurrent otitis media.2- 7
To distinguish QOL changes related to tympanostomy tubes from other confounding factors, 2 techniques were used. First, each child served as his/her own control by comparing changes in QOL prior to surgery (visit 1 to visit 2) to the changes observed after surgery (visit 2 to visit 3). Second, the relationship between outcome and patient satisfaction was examined to ascertain that improved QOL did not simply reflect high satisfaction.16 Patient satisfaction was assessed at visit 1 using 2 independent components: satisfaction with office visit was measured using 4 items from the 9-item survey developed by the Medical Outcomes Trust,17 and satisfaction with decision to have surgery was measured using 5 items from the 6-item Satisfaction With Decision Scale18 (items were simplified from their original 11th-grade reading level to a 4th-grade level).
Baseline data regarding demographics and disease status were collected at visit 1 and included the child's sex, primary indication for tympanostomy tubes (chronic otitis media with effusion [OME] or recurrent AOM), whether the child was otitis prone (4 or more episodes of AOM in the past 12 months with at least 1 episode in the past 6 months), history of tympanostomy tubes, current middle ear status (unilateral, bilateral, or no middle ear effusion), and the presence or absence of hearing loss (pure tone average at 500, 1000, and 2000 Hz of 20-dB hearing level or poorer in soundfield or for the better-hearing ear). Additional data obtained at visit 3 were the amount of otorrhea since surgery (none, 1-2 days, 3-7 days, or ≥8 days) and the current status of the tympanostomy tubes (one, both, or neither intact, patent, and dry).
Data entry was performed using a microcomputer database and verified for 100% accuracy. All statistical analyses and sample size calculations were performed using True Epistat software.19 A sample size of 120 subjects was needed to estimate the postoperative change score with a precision of ±0.2 units, assuming an anticipated mean change score of 1.5 to 2.0, α of .05, and β of 0.20. Change scores before and after surgery were compared using the matched samples t test with an α level of .05. Univariate analysis of factors influencing postoperative change was accomplished using Pearson correlation for numerical data, Spearman rank correlation for ordinal data, and Kruskal-Wallis analysis of variance. Multivariate analysis was performed using stepwise multiple linear regression. Ninety-five percent CIs were calculated for all main results.
The initial sample consisted of 248 children from 14 pediatric otolaryngology practices. Nine practices (64%) recruited a consecutive sample and 5 (36%) recruited a convenience sample. One hundred eleven children (45%) were from the central United States, 89 (36%) from the eastern United States, and 48 (19%) from the western United States. The median age at entry was 1.4 years (range, 0.5-9.9 years), with upper and lower quartiles of 1.0 and 2.3 years. Most children (60%) were boys. Few children had prior tympanostomy tubes (13%), but many (83%) were otitis prone. Audiometry results were available for 183 children (74%), of whom 112 (61%) had hearing loss.
The primary indication for tube insertion was recurrent AOM (56%), followed by OME (42%) and retraction-type ear disease (2%). At visit 1, most children had bilateral OME (68%), with the remainder having unilateral OME (16%) or no OME (16%). Tympanostomy tubes were inserted in the operating room by all study physicians, although 2 reported performing at least some of the procedures in a freestanding surgical center. All physicians used short-acting grommet-type tubes, designed to extrude spontaneously within 6 to 18 months. Seven physicians routinely used antibiotic-containing ear drops in the perioperative period (55% of children in the study sample) and 7 physicians used drops on an individualized basis (45% of children in the study sample).
OM-6 baseline survey responses are shown in Table 1 and Table 2. Caregiver concerns, physical suffering, and emotional distress were the domains of greatest impact, followed by activity limitations, hearing loss, and speech problems. Eighty-eight percent of caregivers were worried or concerned about their child's ear infections or fluid at least some of the time, with 42% spending most or all of their time preoccupied. Similarly, physical suffering was at least somewhat of a problem for 85% of children and at least quite a bit of a problem for 56%. Emotional distress was at least a moderate problem for nearly two thirds of children, and activity limitations were at least a moderate problem for about half. Hearing loss was perceived to be at least somewhat of a problem for 45% of children, and speech impairment was at least somewhat of a problem for 34%.
Nearly all parents were satisfied with their decision to schedule tympanostomy tube placement (Table 3), although a few (5 parents) believed they were not told enough to make an informed decision. All parents expected to follow through with tube insertion, with 81% expressing a strong commitment. Satisfaction with the overall office visit was also high (Table 4), and was rated excellent by 70% of parents. About 85% of parents rated the physician's personal manner and technical skills as excellent.
Change in QOL was measured using the difference in scores for consecutive OM-6 surveys (change scores). The mean time between OM-6 surveys at visits 1 and 2 was 15 days (SD, 9.9), and at visits 2 and 3 was 32 days (SD, 21.9). Although complete data were obtained for all patients at all 3 visits, the change scores were considered valid only when the minimum time interval exceeded 14 days. Of the initial 248 patients, 224 (90%) had valid change scores after surgery, with a mean interval of 34 days between visits 2 and 3. Valid changes scores before and after surgery were available for 115 patients (46%), with a mean interval of 22 days between visits 1 and 2 and 34 days between visits 2 and 3.
Most children had a trivial or small change in QOL before surgery and a moderate to large change after surgery (Table 5). Large improvements occurred for 50% of children after surgery, but were observed in only 12% from the initial office visit to the day of tube insertion (visit 1-2). Before surgery the mean change score was 0.33 (SD, 1.0; range, −2.3 to 3.3). In contrast, after surgery the mean change score was 1.40 (SD, 1.3; range, −2.2 to 4.8). The difference between preoperative and postoperative change scores was 1.07, which was highly significant both clinically (95% CI, 0.71 to 1.43) and statistically (matched samples t test, t114=5.86, P<.001).
Similar improvements after surgery are noted when results are analyzed by effect size (Table 6) instead of change scores. Moderate to large effect sizes were obtained for the OM-6 survey score and all individual domains after surgery, but for none of the domains before surgery. The 95% CIs for the preoperative and postoperative effect sizes in Table 6 do not overlap (except for activity limitations), indicating significant differences at P<.05. The largest degrees of clinical improvement occurred with caregiver concerns, physical suffering, and emotional distress. Moderate improvements were reported for hearing loss, speech impairment, and activity limitations.
Table 7, Table 8, and Table 9 show the impact of tympanostomy tubes on QOL for all children with valid change scores after surgery (n=224). The findings are similar to those in Table 5 and Table 6, but precision is increased by the larger sample size. Caregiver concerns, physical suffering, and emotional distress again showed large improvements, with corresponding changes noted in hearing loss and activity limitations. Speech impairment remained the domain for which parents perceived the least short-term change. Only 4% to 13% of children had poorer QOL in any given domain after surgery (Table 8). Of the 8 children (4%) described in Table 9 who had poorer overall QOL, the degree of worsening was small for 4 children and large for the remainder.
Multivariate analysis of factors associated with change in QOL after surgery (Table 10) demonstrated the most favorable outcomes for children who were otitis prone, had parents who were satisfied with their initial decision to place tubes, and had less than 3 days of postoperative otorrhea. The degree of otorrhea explained 9.5% of the variance in observed outcomes, and satisfaction with decision explained an additional 3.4%. At the first postoperative office visit 157 children (70%) had no otorrhea, 24 (11%) had 1 to 2 days, 22 (10%) had 3 to 7 days, and 21 (9%) had 8 or more days. The incidence of otorrhea was unrelated to whether antibiotic ear drops were used routinely or sporadically (31% vs 28%). Otitis-prone status accounted for 1.4% of the variance, but the effect was not statistically significant (P=.06).
Factors that were not associated with outcome included child age, indication for surgery, history of prior tubes, hearing loss, middle ear status (eg, effusion presence and laterality), satisfaction with office visit, and the time between surgery and outcome assessment. Although tube status at the first postoperative office visit was associated with outcome in the univariate analysis, the relationship was no longer significant when adjusted for otorrhea. When first examined after surgery 190 children (85%) had both tubes functioning (eg, dry and patent lumen), 25 (11%) had only one tube functioning, and 9 (4%) had neither tube functioning.
Our study provides the first systematic evidence regarding the impact of tympanostomy tubes on child QOL. Nearly 4 of 5 children improved within several weeks of surgery (Table 9), with more than half experiencing a large benefit. In contrast, most children had only small or trivial changes in their QOL status in the weeks preceding surgery (Table 5). These data are not offered to support tube efficacy, but rather to show that the objective benefits in randomized trials2- 7 are accompanied by subjective improvements. Outcomes research and randomized trials are synergistic; there is a need for both the realism of experimental studies and the empiricism of personal experience. As noted by Alvan Barach, we must "Remember to cure the patient as well as the disease."20
In contrast to randomized trials our selection criteria were broad, consistent with the philosophy that outcomes research should apply to a diverse target population.21 Children were recruited and observed during routine clinical care provided at 14 different practice sites. The common feature was a need for bilateral tympanostomy tubes without concurrent pharyngeal or other surgery. Indications for surgery were at the discretion of individual physicians, but tubes were generally inserted for (1) bilateral OME lasting at least 3 to 4 months, especially when a hearing loss was present, or (2) recurrent AOM (minimum of 3 episodes in the prior 6 months), especially with coexisting OME (unilateral or bilateral) or breakthrough episodes of infection despite antimicrobial prophylaxis.
The beneficial impact of tympanostomy tubes on child QOL was relatively independent of patient satisfaction. Satisfaction with the physician and office visit did not correlate with outcome. The overall experience was rated as excellent by 70% of patients (Table 4), compared with similar ratings for only 49% of visits to health maintenance organizations and 64% of visits with solo practitioners.17 Parent satisfaction with the decision to have surgery did correlate with outcomes (Table 10), but explained only 3.4% of the observed variations. Nonetheless, this relationship underscores the need to educate families who elect tympanostomy tubes so they are comfortable with their decision. The personal significance of treatments to patients must be considered along with the clinical or statistical significance demonstrated in research studies.22
Change in health status is best measured by assessing patients before and after an intervention (as was done in this study), rather than by asking patients how much better or worse they feel after the intervention is performed.23 The latter method is inaccurate because people simply do not remember how they were at the beginning, and because retrospective estimates of initial state are typically highly correlated with the present state (not the initial state). The OM-6 survey was an ideal outcome measure for our study because the difference between consecutive assessments generates a change score that readily classifies the direction of change as positive (improvement) or negative (worsening), and the magnitude of change as trivial, small, moderate, or large.12 Further, the OM-6 is easy to complete and readily integrated into routine clinical care.
The improved QOL after tympanostomy tube insertion was significant at the individual and group levels. Change scores reflect health status variations within individuals,24,25 and for most children showed moderate to large improvements (Table 8 and Table 9) in overall and domain-specific QOL. Effect sizes reflect health status variations within groups,26 and showed large improvements (Table 7) in overall QOL and in most domains. For example, the effect size for the overall OM-6 survey score following tube placement was 1.20 (95% CI, 1.07 to 1.33), which greatly exceeds the 0.80 threshold for a large impact. What constitutes a worthwhile effect size for otitis media treatments is controversial, but an effect size of 0.30 is the generally accepted minimum.27
The association of otorrhea with poorer outcomes suggests that prevention be considered. Unfortunately, equivalent rates of early postoperative otorrhea occur (10%-20%) regardless of surgical technique,28 use of prophylactic antimicrobial ear drops, or the degree of water precautions (if any) prescribed.29 A meta-analysis of 11 randomized trials of antimicrobial ear drops given during or shortly after tube placement showed an absolute decrease in otorrhea rates of 7% by child (95% CI, −4% to 17%)30 and 6% by ear (95% CI, 4% to 9%). The results by child are not statistically significant, and the results by ear beg the issue of clinical importance (need to treat 17 children to prevent otorrhea in 1 child). Similarly, a meta-analysis of 5 prospective studies showed no difference in otorrhea rates for swimmers vs nonswimmers.31 Water precautions do not reduce the incidence of bathing-related otorrhea.32
Recognizing that otorrhea cannot be prevented, the next best management strategy lies in parental education regarding the generally benign and self-limited nature of the problem. The relative risk of otorrhea is higher for children with inflamed middle ear mucosa at surgery, mucoid effusion, and those with bacterial pathogens in the ear canal or middle ear effusion.33 These factors relate to baseline disease status, and usually resolve after ventilation to the middle ear space is established. In contrast, delayed tympanostomy tube otorrhea is usually secondary to upper respiratory tract infection32 and for most patients is brief, painless, and does not recur. Only 9% of children with tubes in randomized trials have 3 or more discrete bouts of otorrhea, and 5% experience chronic otorrhea at some point in the course of their intubation.34 Most tube-related otorrhea can be managed without systemic antibiotics.35
A limitation of our study is that QOL was assessed after a mean intubation period of only 34 days. Whereas the large magnitude of improvement reported is impressive, it excludes the impact of subsequent adverse events or potential tube-related complications on QOL. Evidence from randomized trials, however, suggests that tubes remain efficacious as long as middle ear ventilation is maintained (usually 6-18 months). Complications are rare, particularly when short-acting (grommet-type) tubes are used.29 Incidence rates of localized retraction or perforation of the tympanic membrane are about 0.5% to 2.0% and rates of cholesteatoma or medial tube displacement are less than 0.2% to 0.5%. Sequelae not requiring intervention are more common (tympanosclerosis or focal atrophy), but are generally more cosmetic than functional. The hearing loss caused by tympanosclerosis is insignificant (<0.5 dB), which cannot serve as an argument against judicious tube placement.36
Our results are not intended to justify liberal use of tympanostomy tubes. All children studied had reached the level of subspeciality referral, and most had bilateral OME with hearing loss and recurrent AOM. Children with otitis media of lesser severity often do well without surgery. About 50% of OME resolves within 1 month, increasing to 60% by 3 months, and 75% by 6 months; the chance of spontaneous resolution diminishes greatly after 3 to 6 months of persistence.37 Similarly, children with recurrent AOM (and no baseline OME) are likely to have 1.5 to 2.5 less episodes the next year, with 50% having no AOM over a median period of 6 months and more than 85% having 2 or fewer episodes.37 Antibiotic therapy improves short-term resolution of OME by 16% and prophylaxis reduces incidence of AOM by 0.12 episodes per patient-month.38 These benefits, however, are modest and argue for restrictive use given growing concerns over antibiotic-induced bacterial resistance.
Medical or surgical interventions cannot be endorsed simply because an outcomes study shows improved QOL. From a methodological perspective, observational research offers a level of evidence that is vastly inferior to individual or pooled randomized controlled trials.39 Whereas we can state that nearly 80% of children improved after tube insertion (Table 9), we cannot state with certainty that all benefits observed were because of tube insertion. The smaller changes seen in the immediate preoperative period are reassuring (Table 5 and Table 6), but do not preclude some impact of halo and placebo effects on final outcome. Further, the relatively brief duration of observation (several months) may be influenced by natural history and regression to a mean symptom level.
Several large-scale randomized studies2- 7 have demonstrated that tympanostomy tubes significantly reduce the incidence and prevalence of otitis media, and promote optimal hearing while they remain patent and functional. Our investigation shows that these objective benefits are perceived and appreciated by parents, and manifest with greatly improved quality of life for most children within several weeks of tube placement. Ongoing research is required to document the long-term impact of tubes on QOL and to identify which factors correlate most with perceived outcomes. Better appreciation of these factors should help children feel and function best while outgrowing otitis media.
Accepted for publication November 17, 1999.
This study was funded, in part, by a grant from the American Society of Pediatric Otolaryngology.
Presented in part at the annual meeting of the American Society of Pediatric Otolaryngology, Palm Desert, Calif, April 28, 1999.
We thank Glenn Isaacson, MD (Temple University, Philadelphia, Pa), and Anthony Magit, MD (University of California at San Diego), for their help in recruiting patients. We also thank the following individuals who served as study coordinators at the individual data collection sites: Ellen Bair, Diane Burke, Mary Carol Cryer, Pat Fall, Jennifer Johnson, and Ann Tirrel.
Corresponding author: Richard M. Rosenfeld, MD, MPH, Department of Otolaryngology, 339 Hicks St, Brooklyn, NY 11201 (e-mail: email@example.com).