598 September 2002 Family Medicine

Otitis Media in Children: Use of Diagnostic Tools by Family Practice Residents

Jonathan E. MacClements, MD; Michael Parchman, MD; Cindy Passmore, MA

Background and Objectives: Although media is a common problem in primary care, little is known about the use of diagnostic tools such as pneumatic otoscopy, , acoustic reflectometry, or tympanocentesis by family physicians in training. Methods: This was a self-reported observational study of family practice residents’ use of diagnostic tools. Twenty-three family practice programs in Texas and Oklahoma were surveyed during November and December 2000. Residents were asked about their use of diagnostic tools, and, if tools were not used, they were asked the reason for not using them. Residents were also asked about the criteria they used to diagnose otitis media, and their responses were compared to criteria recommended by national guideline panels. Results: The response rate was 61% (n=316). The percentage of residents using pneumatic otoscopy was 66%, tympanometry 29%, acoustic reflectometry 2%, and tympanocentesis 0%. The most common reasons cited for not using tools were lack of training or unavailability of equipment. Fifty-five percent of family practice residents do not report pneumatic otoscopy for diagnosing otitis media and thus did not use recommended criteria. Faculty train- ing of residents in the use of diagnostic tools was associated with a higher rate of using these tools. Discussion: Few residents believed that diagnostic tools had no value in the diagnosis of otitis media, but lack of training or equipment problems were reported as contributing to their not using these tools. Half of family practice residents may be inadequately diagnosing middle ear problems since they did not re- port that pneumatic otoscopy was necessary for diagnosing otitis media. Because training was associated with higher rates of using appropriate diagnostic tools, family medicine faculty can play a significant role in improving the residents’ diagnostic skills.

(Fam Med 2002;34(8):598-603.)

Middle ear problems are common in children. Acute The diagnosis of otitis media is not easy,4 and there otitis media affects at least 80% of children younger is significant variability in the ability of primary care than age 2 years1 and results in at least 3 million visits physicians to diagnosis it.8 Symptoms of acute otitis to US family physicians per year.2 Middle ear effusion media are neither sensitive nor specific, and most chil- following acute otitis media is also a common problem dren with otitis media with effusion are asymptomatic.9 that causes in a substantial proportion of The physical examination is also potentially inaccu- children and is a frequent reason why primary care pro- rate since cerumen may impair visualization of the tym- viders refer children to specialists.3 panic membrane, or children may be uncooperative with Middle ear problems of acute otitis media and otitis the examination. Further, subjective impressions of the media with effusion are overdiagnosed.4 Overdiagno- appearance of the tympanic membrane are difficult to sis leads to unnecessary and inappropriate antibiotic quantify and grade on a purely clinical basis. Finally, treatment and subsequent development of antibiotic equipment may not work, and time constraints make it resistance.5-7 difficult to devote the time necessary to achieve an ad- equate examination. Numerous tools have been developed to improve di- From the Department of Family Practice, University of Texas Health Cen- agnostic accuracy. A review of the literature shows that ter at Tyler (Dr MacClements); the Department of Family Practice, Univer- a pneumatic improves accuracy and should sity of Texas Health Science Center at San Antonio (Dr Parchman); and the Department of Research, Family Practice Faculty Development Center of be used to assess the mobility of the tympanic mem- Texas, Waco, Tex (Ms Passmore). brane in diagnosing otitis media.1,4,9-11 Bluestone and Residency Education Vol. 34, No. 8 599

Klein9 have documented that the use of diagnostic aids The survey included questions about the residents’ in addition to pneumatic otoscopy, such as age, gender, and year of training. The number of weeks tympanometry,12,1 3 acoustic reflectometry, 14- 17 and of ENT training in medical school and residency was tympanocentesis18 further improve diagnostic accuracy. also included. To assess the use of adjunctive tools, the Little information is available, however, as to whether residents were asked to report their actual practices and primary care physicians use these adjunctive diagnos- not what they believed to be the academically correct tic tools. answers. This study surveyed the practice of family practice A 5-point Likert scale (with response options of residents in Texas and Oklahoma in the diagnosis of never, seldom, occasionally, frequently, and almost al- acute otitis media/otitis media with effusion. The spe- ways) was used to evaluate the frequency of the resi- cific aims of the survey were to (1) determine which dents’ use of each of the following adjunctive tools: tools, if any, are used by family practice residents in pneumatic otoscopy, audiometry, Rinne/Weber test with the diagnosis of otitis media, (2) identify why these a 256 Hz tuning fork, tympanometry, acoustic reflec- adjunctive tools may not be used, (3) determine the tometry, and tympanocentesis. In the survey, pneumatic minimal criteria used by family practice residents to otoscopy was defined as air insufflation with a bulb in make the diagnosis of otitis media, (4) determine the addition to routine otoscopy. If the resident’s reported residents’ perceptions regarding the adequacy of their frequency of use of an adjunctive tool was never or sel- training in this field, and (5) determine the amount of dom, then a checklist to determine the resident’s rea- ear, nose, and throat (ENT) training (in weeks) received sons for not using this tool was completed (more than during medical school and residency. one item could be checked). The items on the checklist are shown in Table 1, Part 1. Residents selected one Methods item from a checklist shown in Table 1, Part 2 to deter- Subjects mine their minimal criteria for diagnosis of acute otitis To maximize response rates, we selected a purpo- media/otitis media with effusion. sive sample of family practice residents in Oklahoma A 6-point Likert scale of don’t know, never, seldom, and Texas. It was felt that this sample would improve occasionally, frequently, and almost always was fur- response rates because of the familiarity of these pro- ther used to evaluate the residents’ perception of the grams with the Texas Faculty Development Fellowship use and teaching of adjunctive tools by family medi- Program, where this research was conducted. The 1999 cine faculty and other specialties. Residents were also National Resident Matching Program Directory was asked to use a 5-point Likert scale (1=strongly disagree used to identify all 28 family practice programs in Texas and two family practice programs in Oklahoma, all of which had faculty who previously attended the Texas faculty development fellowship. An individual faculty Table 1 member of each residency program was chosen based on the following criteria: (1) present or past fellow of Checklists Used in Survey Instrument the Texas faculty development fellowship, (2) behav- ioral health faculty member of the Behavioral Science Part 1—Checklist of reasons for not using a diagnostic tool (1) Do not know what this adjunctive tool is Consortium of Texas, or (3) program director. (2) Not available or not functioning Surveys were mailed on November 22 to these indi- (3) Lack of training viduals. An accompanying cover letter requested that (4) No benefit/does not change the diagnosis (5) Refer patient for further evaluation the surveys be distributed to that residency’s residents (6) Other and collected and mailed back in an enclosed, self- addressed envelope by December 20, 2000. An e-mail/ Part 2—Checklist of criteria for diagnosing acute otitis media/otitis media with effusion telephone call was also made to the individual faculty (1) Symptoms only to ensure that he/she received the survey. A telephone (2) Symptoms and otoscopic tympanic membrane signs (without call to determine the number of residents enrolled in insufflation) (3) Symptoms and pneumatic otoscopic tympanic membrane signs (with the program at the time of the survey was made during insufflation) February 2001 to each of the residencies that chose to (4) Symptoms and pneumatic otoscopic tympanic membrane signs (with participate. insufflation) and additional adjunctive tools Part 3—Checklist of management of cerumen obstructing the tympanic Instruments membrane prior to diagnosis A survey was developed after reviewing the litera- (1) Leave cerumen alone and diagnose (2) Remove cerumen and diagnose ture, discussing the use of adjunctive tools with ENT (3) Refer patient for further evaluation specialists and consulting with faculty development (4) Other fellows and research consultants. 600 September 2002 Family Medicine to 5=strongly agree) to rate whether their own training in the diagnosis of otitis media was adequate. In addi- Table 2 tion, the residents were given the opportunity to pro- vide suggestions on improvements of the training of Demographics* primary health care providers in the diagnosis of otitis media. Finally, residents were asked to select an op- Age (in years)—mean=31.5, range=25–55 tion from a checklist shown in Table 1, Part 3, describ- ing their management of cerumen obstructing the tym- Gender—male=59%, female=41% panic membrane prior to diagnosis. Number of residents in each year—first year=101 (32%), second year=100 (32%), third year=115 (36%) Data Analysis * n=316 Descriptive statistics, histograms, and a pie chart were used to present the data. Those respondents who failed to complete the Likert scale for frequency of use of a diagnostic tool but who went on to list their overall response rate of 61%. The demographics of the reasons for not using the tool were included in the re- respondents are shown in Table 2. sults. A chi-square test was used to determine if there was Frequency of Diagnostic Tool Use a relationship between faculty training and minimal The frequency with which residents used diagnostic criteria used by the residents, to determine the rela- tools for diagnosing acute otitis media and otitis media tionship between program year and minimal diagnos- with effusion is shown in Table 3. Pneumatic otoscopy tic criteria, to determine if the number of weeks of ENT was the most used diagnostic aid; 66% of residents used training during medical school or residency was re- it occasionally, frequently, or always. Tympanometry lated to the frequency of pneumatic otoscopy or the was the next most commonly used diagnostic tool. Less residents’ minimal criteria for diagnosing otitis media, than 2% of the residents used an acoustic reflectome- and to compare the residents’ perception of which fac- ter. For those respondents who reported that they never ulty (ie, faculty from which specialty) were teaching or seldom used a diagnostic tool, the reason for not us- these diagnostic tools. Student’s t test was used to de- ing each tool is shown in Figure 1. Of those who never termine the residents’ perception of adequacy of their or seldom used pneumatic otoscopy, 42% stated it was training in the group using adequate diagnostic crite- not available or not working, and 39% stated that their ria compared to the group not using adequate diag- reason was lack of training. Similar results were ob- nostic criteria. tained for tympanometry.

Minimal Criteria Criteria for Diagnosing Otitis Media Minimal criteria for diagnosing otitis media (Table The residents’ criteria for diagnosing acute otitis 1, part 2) were defined as adequate criteria (those us- media/otitis media with effusion is shown in Figure 2. ing a pneumatic otoscope: selections 3 or 4) and inad- Based on accepted criteria for proper diagnosis of otitis equate criteria (those not using a pneumatic otoscope: selections 1 or 2). This definition was derived from the US Department of Table 3 Health and Human Services 10 Clinical Practice Guideline, Residents’ Frequency of Use of Diagnostic Tools which recommends that a sus- pected otitis media with effusion Tool be evaluated with a pneumatic (n= total # Almost of respondents) Never Seldom Occasionally Frequently Always otoscope and that otoscopy alone Pneumatic otoscopy (n=305) 42 (14%) 61 (20%) 88 (28%) 70 (23%) 44 (15%) without insufflation is not ad- equate. Tympanometry (n=305) 146 (47%) 72 (24%) 66 (22%) 18 (6%) 3 (1%) Acoustic reflectometry (n=309) 290 (94%) 13 (4%) 4 (1%) 2 (.6%) 0 Results Of the 30 residencies that were Tympanocentesis (n=305) 288 (94%) 15 (5%) 1 (.5%) 1 (.5%) 0 recruited, 23 chose to participate Audiometry (n=306) 235 (77%) 44 (14%) 23 (8%) 4 (1%) 0 in the study. A total of 316 sur- veys were returned from 521 resi- Tuning fork (n=304) 199 (65%) 68 (22%) 32 (11%) 3 (1%) 2 (1%) dents in these programs for an Residency Education Vol. 34, No. 8 601

Figure 1

Residents’ Reasons for Never/Seldom Using a Diagnostic Tool

Do not Not Lack of No benefit/ Refer Other know what available or training does not patient this is not change the functioning diagnosis

media, 55% of family practice residents were not us- ing sufficient diagnostic criteria. Ninety-six percent re- ported removing cerumen prior to diagnosis. Figure 2

Training Family Practice Residents’ Criteria for Diagnosing There was a significant association between resi- Acute Otitis Media/Otitis Media With Effustion dents’ correct use of diagnostic criteria and their per- ception of how often they were taught to use adjunc- tive tools by family medicine or other faculty (χ2=24.32, P=.0002). There was no statistical difference in the resi- dents’ perception of adequacy of training between those residents using a pneumatic otoscope (adequate) com- pared to those not using a pneumatic otoscope (inad- equate). There was no association between year of train- ing and diagnostic criteria. There was also no associa- tion between the weeks of ENT training during medi- cal school or residency related to the frequency of pneu- matic otoscopy or the minimal criteria identified by residents for diagnosing otitis media. This study showed no statistical difference between the residents’ perception of which faculty was doing the teaching of these diagnostic skills (family practice faculty mean=2.31 and other faculty mean of teach- ing=2.09). Surprisingly, the residents’ perception of their adequacy of ENT training did not improve with the residents’ year of training. The median total number of weeks training in ENT during medical school was 2 weeks. The third-year resi- dents showed a mean of 2.2 weeks ENT training dur- ing their residency, compared with a mean of .8 weeks for the first-year residents. Twenty-seven percent of residents provided commentary that included requests for further ENT training. 602 September 2002 Family Medicine

Discussion Conclusions Although experts recommend use of pneumatic otos- If this study reflects national practices, then half of copy for the diagnosis of otitis media,9-11,18-20 only two family practice residents may have inadequate training thirds of family practice residents in Texas and Okla- in the use of appropriate diagnostic criteria to evaluate homa used this tool as an aid to diagnosis, and only middle ear problems. Although our results suggest that 15% used it consistently. The tympanometer was used a lack of training and unavailability of diagnostic tools by less than one third of the residents, and the acoustic are to blame, further research is needed to confirm these reflectometer was used by less than 2% of the residents. findings. These results suggest the need for more fam- Few of these residents believed that these tools had no ily medicine faculty development in the correct use of value in the diagnosis of otitis media. Rather, a lack of these diagnostic tools and application of appropriate training or equipment problems were reported as con- criteria for diagnosing otitis media. Proposed training11 tributing to not using these tools. Consistent with the in diagnostic tools for family practice clinicians could reported frequency of use of these diagnostic tools, as include several or more of the following: (1) core lec- many as half of residents may be using insufficient cri- tures, (2) instructional videos,20 (3) supervised practice teria in diagnosing otitis media. of pneumatic otoscopy, (4) visualization of tympanic The year of training did not influence the family prac- membranes via video monitoring, (5) performance of tice residents’ use of diagnostic tools, but, as expected, tympanometry and acoustic reflectometry, and (6) ex- the more training they received in using these tools, by perience in the operating room with an otolaryngolo- either family medicine or other specialty faculty, did gist evaluating the tympanic membrane before and af- increase the likelihood that they used them. Family ter a myringotomy. The literature confirms that train- medicine faculty may be assuming that the diagnosis ing interventions do improve the diagnosis of otitis of otitis media is being taught by other departments, media. Once trained, faculty can also ensure the avail- but the amount of time spent teaching and demonstrat- ability and working order of these diagnostic tools. New ing otitis media diagnostic tools was reported to be the tools and treatments are being developed to improve same for both family medicine faculty and other spe- the management of otitis media, and physicians will cialty faculty. need to stay informed of current technology.26 The residents perceived that their training in otitis media was slightly above average. This perception did Acknowledgment: This study was presented at the fellowship graduation to not improve with the residents’ year of training. Since regional family practice faculty at McLennan Medical Education and Re- search Foundation, April 2001. there was no difference in perception of adequacy of training between the residents who do and do not use a Corresponding Author: Address correspondence to Dr MacClements, Uni- versity of Texas Health Center at Tyler, Department of Family Practice, pneumatic otoscope (adequate versus inadequate crite- 11937 US Highway 271, Tyler, TX 75708-3154. 903-877-7206. Fax: 903- ria), the residents were not likely to recognize their own 877-7778. [email protected]. deficiencies in diagnosing middle ear problems. The residents were more likely to use appropriate diagnostic criteria if their faculty taught the use of di- REFERENCES agnostic tools. Several studies show that practitioners 1. Schloss M. Otitis media: to treat or not to treat? Can Respir J 1999;6 trained in pneumatic otoscopy and the use of suppl A:51A-53A. tympanometry or acoustic reflectometry have a statis- 2. US Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention statistics, 1997 data. tically significant improvement in their ability to diag- www.aafp.org/facts/table18.html. 11,17,20-23 nose otitis media. A study performed in Pitts- 3. Forrest CB, Glade GB, Baker AE, Bocian AB, Kang M, Starfield B. burgh at the Children’s Hospital showed that the sensi- The pediatric primary-specialty care interface: how pediatricians refer children and adolescents to specialty care. Arch Pediatr Adolesc Med tivity and specificity of clinicians’ otoscopic diagnoses, 1999;153(7):705-14. when compared to myringotomy, improved signifi- 4. Pichichero ME. Acute otitis media: part 1. Improving diagnostic accu- cantly after they learned to use a pneumatic otoscope.24 racy. Am Fam Physician 2000;61(7):2051-6. 5. Block SL. Management of acute otitis media in the 1990s: the decade of resistant pneumococcus. Pediatric Drugs 1999;1(1):31-50. Limitations 6. Pichichero ME, Reiner SA, Brook I, et al. Controversies in the medical This was a self-reported observational study of fam- management of persistent and recurrent acute otitis media. Recommen- dations of a clinical advisory committee. Ann Otol Rhinol Laryngol ily practice residents only and was limited by the resi- Suppl 2000;183:1-12. dents’ perception of their diagnostic skills. The study 7. Hooton TM, Levy SB. Confronting the antibiotic resistance crisis: was also limited geographically to residencies in Texas making appropriate therapeutic decisions in community medical prac- tice. Alliance for Prudent Use of Antibiotics (APUA) antibiotic resis- and Oklahoma. Further, only 23 of 30 residencies chose tance monograph: guidelines from the Summit on Antimicrobial Re- to participate in the study, although we believe that the sistance. Boston: APUA,1999:5-6. residencies were sufficiently heterogeneous not to limit 8. Lyon JL, Ashton A, Turner B, Magill M. Variation in the diagnosis of upper respiratory tract infections and otitis media in an urgent care set- the results. The response rate of 61% is less than opti- ting. Arch Fam Med 1998;7(3):249-54. mal but higher than that reported in some similar aca- 9. Bluestone CD, Klein JO. Otitis media in infants and children. New demic surveys.25 York: W.B. Saunders Company, 1995. Residency Education Vol. 34, No. 8 603

10. Stool SE, Berg AO, Berman S, Carney CJ, Cooley JR. Otitis media 18. Block SL, Blumer JL, O’Rourke EJ, Seidel J, Schulman ST. Acute oti- with effusion in young children. Clinical practice guideline. Washing- tis media, advanced therapies challenge standard protocol. Medical ton, DC: US Department of Health and Human Services, 1994. Crossfire 2000. 11. Pelton SO. Otoscopy for the diagnosis of otitis media. Pediatr Infect 19. Kaleida PH, Fireman P. Diagnostic assessment of otitis media. Clin Dis J 1998;17:540-3. Allergy Immunol 2000;15:247-62. 12. Brookhouser PE. Use of tympanometry in office practice for diagnosis 20. Hoberman A, McMillan JA, Kaleida PH, Paradise JL, Butz AM. A view of otitis media. Pediatr Infect Dis J 1998;17:544-51. through the otoscope, distinguishing acute otitis media from otitis me- 13. Palmu A, Puhakka H, RahkoT, Takala AK. Diagnostic value of dia with effusion. (Monograph) Baltimore: The Johns Hopkins Univer- tympanometry in infants in clinical practice. Int J Pediatr sity School of Medicine, 2000:1-12. Otorhinolaryngol 1999;49:3:207-13. 21. Van Balen FA, Aarts AM, De Melker RA. Tympanometry by general 14. Barnett ED, Klein JO, Hawkins KA, Cabral HJ, Kenna M, Healy G. practitioners: reliable? Int J Pediatr Otorhinolaryngol 1999;48(2):117- Comparison of spectral gradient acoustic reflectometry and other diag- 23. nostic techniques for detection of middle ear effusion in children with 22. Hemlin C, Hassler E, Hulcrantz M, Papatziamos G, Krakau I. Aspects middle ear disease. Pediatr Infect Dis J 1998;17:556-9. of diagnosis of acute otitis media. Fam Pract 1998;15(2):133-7. 15. Kemaloglu YK, Sener T, Beder L, Bayazit Y, Göksu N. Predictive value 23. Jensen PM, Lous J. Criteria, performance, and diagnostic problems in of acoustic reflectometry (angle and reflectivity) and tympanometry. diagnosing acute otitis media. Fam Pract 1999;16(3):262-8. Int J Pediatr Otorhinolaryngol 1999;48(2):137-42. 24. Kaleida PH, Stool SE. Assessment of otoscopist’s accuracy regarding 16. Block SL, Mandel E, McLinn S, et al. Spectral gradient acoustic reflec- middle ear effusion. AJDC 1992;146:433-5. tometry for the detection of middle ear effusion by pediatricians and 25. Baruch Y. Response rate in academic studies— a comparative analysis. parents. Pediatr Infect Dis J 1998;17:560-4. Human Relations 1999;52:4. 17. Kimball S. Acoustic reflectometry: spectral gradient analysis for im- 26. Klein JO. Management of otitis media: 2000 and beyond. Pediatr In- proved detection of middle ear effusion in children. Pediatr Infect Dis J fect Dis J 2000;19(4):383-7. 1998;17:552-5.