Assessment of Outer and Middle Ear Pathologies in Lilongwe, Malawi

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

This article on middle and outer ear disorders in low-income nations like Malawi has piqued my curiosity. The study presents a retrospective analysis of every patient who visited a hearing center in a particular area of Malawi and who was chosen as a candidate for treatment for middle and outer ear disorders.  Interesting findings were the high percentage of patients with cerumen, the low percentage of children with draining ears who visited an ENT test because it was considered normal, and the huge number of patients who did not attend the follow-up. The paper's weakness, in my opinion, is the absence of a pure audiometry evaluation. Although this assumption isn't always accurate, the authors have assumed that all patients with middle ear and external ear disorders have conductive hearing loss. When using pure tone audiometry, many patients with minor ear drum perforations have normal hearing. It depends on factors like placements and size. Furthermore, ear wax accumulation is not always indicative of hearing loss. These factors need to be brought up during the discussion and perhaps in the title. Furthermore viewing a comprehensive table including all the diseases encountered and their percentage would be really appreciated. In this method, the authors' goal of drawing attention to health issues in developing nations can be achieved without sacrificing the paper's scientific value.

Author Response

Thank you for your feedback. We absolutely agree with your comment. Due to resource limitations, PTA are not available for many of these patients and thus surrogate markers were chosen and we had more directly addressed this limitation in the limitations section. Additionally, we have changed the title of the paper to better reflect the main take-away of the paper- being the prevalence of outer and middle ear pathologies. We also changed the wording throughout the paper to better the guide the reader in understanding that some of the pathologies stated (perforations, CSOM, OME, negative pressure) are sometimes associated with conductive hearing loss (not always). We hope that these changes address the concerns raised and improve the clarity of the manuscript

Reviewer 2 Report

Comments and Suggestions for Authors

The weakest point of this paper is that it is based solely on tympanic membrane (TM) findings and there are too small cases that evaluated hearing.

The idea of the study is interesting, though you cannnot tell conductive hearing loss only by the TM findings.

I'm afraid the article is not worth to publish in the present form.

Comments on the Quality of English Language

4.2. Limitations to the study→Limitations of the study

Author Response

Thank you for your feedback. We acknowledge the limitation of the lack of PTA but given the limitations in access to care and services, these clinical findings were used as surrogate markers and better clarified the limitations in the limitations section.  We have changed the title of the paper to better reflect the main take-away of the paper- being the prevalence of outer and middle ear pathologies. We also changed the wording throughout the paper to better the guide the reader in understanding that some of the pathologies stated (perforations, CSOM, OME, negative pressure) are sometimes associated with conductive hearing loss (not always). We hope these changes are suitable.

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

This manuscript describes prevalence and clinical characteristics of outer and middle ear pathologies in Lilongwe, Malawi. This work is critically important for highlighting the inequity in audiological care across the globe and identifying needs in low-income countries. The authors should be commended for their work in this important area. The significance of this work is somewhat undersold in the current manuscript, and additional opportunities to highlight the importance of the findings are detailed below, along with other suggestions for improving the quality of the manuscript.

 

Major suggestions:

Throughout the results section and in all the data figures, prevalence rates for right and left ears are calculated separately and displayed side-by-side, as if to imply comparison. There is no reason to expect right and left ears to differ in prevalence rates, and the data clearly indicate roughly the same results for the two ears. The separation of right and left ears becomes quite distracting, particularly in the text of the results section. Furthermore, separate treatment of right and left ears obscures the relative number of unilateral vs. bilateral cases in the dataset. The comparison of prevalence rates for unilateral and bilateral outer/middle ear pathologies is considerably more interesting than comparing right and left ear rates, as there are important clinical implications for bilateral vs. unilateral pathologies. While many outer/middle ear pathologies are unilateral, bilateral pathologies carry a much heavier burden for communication. I suggest reanalyzing these data with an agnostic stance towards right vs. left ear, and instead, group cases based on unilateral vs. bilateral presentation. Presenting the data this way makes for a more clinically meaningful comparison and may reveal interesting trends in the subgroup analyses for age group and payment group.

 

The data figures should be redesigned with more descriptive figure captions that explain the pattern of results than can be seen in the figure. The y-axis in all figures is percentage values that are unlabeled, making it unclear what percentage is being shown (i.e., are these percentage of all cases, or percentage of a subgroup of cases, like in Figure 4). Suggestions for each figure are described below:

- Figure 1. This figure really highlights the issue raised above regarding the comparison of right and left ears. Both bars show essentially the same result, which is that ~78% of case had abnormal otoscopic findings. A more compelling figure would compare the percentage of cases with otoscopic abnormalities noted unilaterally vs. bilaterally. It would be even better if the “abnormal otoscopic findings” could be broken down a bit more into specific otoscopic indications. The text indicates 51% was occlusive wax but does not provide any further diagnostic information about the other half of the cases. An ideal figure would show prevalence rates for unliteral and bilateral otoscopic abnormalities, then further break down these two rates into proportions associated with wax occlusion, otitis externa, foreign objects, etc.

- Figure 2. Same as Figure 1, separating right and left ears is not meaningful, but indicating whether type B tymps were unilateral or bilateral has important clinical indications. From the current figure, it is impossible to tell whether 18% of cases had bilateral type B tympanograms, or if 36% of cases had unilateral type B tympanograms (each ear represented in equal numbers). The truth is almost certainly somewhere in the middle, but it would be clearer to present the data in this way. If there were cases with type C tympanograms (which it seems like there were from line 121), they should also be presented in Figure 2.

- Figure 3. Again, indicating unilateral vs. bilateral would be helpful here, particularly in the cases of referral for medication. Wax removal vs. wax softener could be collapsed into a single category for cerumen management, as the distinction between removal and softening does not carry through to the rest of the manuscript. Any other interventions that were observed in the data should be reported here instead. If there are a considerable number of cases with other interventions that were less common (e.g., 4% foreign object removal, 1.5% water irrigation, 0.75% monitoring skin lesions, etc.) these could be grouped into an “other” category with examples described in the text.

- Figure 4. This figure is confusing because the nested percentages are all displayed on the same axis scale. The first bar indicates that 66% of cases included a recommendation for follow up. The next bar seems to indicate that 46.2% of cases included a completed follow up. Based on the text (132-133) it sounds like 46.2% *of the 66% with f/u recommendations* completed that follow-up appointment. If that is correct, the relative size of the bars in Figure 4 should be adjusted to show this more clearly. A better figure design in this case would be a stacked pie chart, where the base shows that 66% of cases recommended follow up, then the next level shows a slice of that 66% that completed the follow up, then the next level shows a slice of those cases that showed improvement (in either ear), then a final level showing recommendations for further follow up. If the authors prefer to stick with a bar graph, that is fine, but it should be plotted on a consistent axis that is labeled as “percentage of cases” rather than just percentages values without a reference to what the percentage represents.

 

Sections on subgroup analyses for private vs. community patients and adult vs. pediatric patients would benefit from tables to summarize the data. Currently, the text is very dense with percentages and p values, many of which are not significant. A clearer way to present these data would be in a table that indicates the percentages for private/community and adult/peds cases for each of the dimensions described on lines 143-169 and 171-199. Again, presenting data on monaural vs. bilateral presentation is preferable to right ear and left ear. You could indicate significant differences between subgroups in the table with asterisks. That would free up a lot of space in the text to build a more readable narrative about the observed differences between subgroups. As a related note, I couldn’t find the percentage of adult vs. pediatric cases in the sample anywhere in the text (maybe I just missed it).

 

Minor suggestions:

Line 48-51: rewrite this sentence.

Line 60-63: rewrite this sentence.

Line 67: End this paragraph with one more strong sentence highlighting the significance of this work. Estimating prevalence is necessary for documenting the need for these services in the area, and identifying the specific services and populations are that are most needed.

Lines 70-76: These two sections can probably be combined, along with information on lines 85-87. This section should also include a brief description of the difference in meaning between the types of payment visits (i.e., what are community patients vs. private pay patients). It would also be helpful to indicate the age cutoff used to distinguish adult vs. pediatric cases. This is another spot where a table of values might be beneficial to summarize the total number of cases, number (and percentage) of community vs. private pay cases, number (and percentage) of adult vs. pediatric cases.

Lines 123-125 and 179-181: Otoacoustic emission screening results are described on these lines, but it is unclear why these results are included. The primary focus of the paper is on outer and middle ear pathologies and associated conductive hearing loss. OAEs are contraindicated in most of these cases, because the measurement is highly susceptible to distortion by conductive pathologies. I suggest removing this information to maintain consistency with the focus of the paper.

Lines 253-260: Consider noting that in countries with more advanced pediatric care, the importance of consistent access to auditory information for speech and language development is stressed to parents and clinicians. In developing nations like Malawi, this may not be a high priority. If treatment of recurrent middle ear infections to prevent speech and language delays is not be prioritized by parents and pediatricians, it would lead to an under-estimate of the prevalence of transient middle ear pathologies in children.

Lines 273-281: This is another opportunity to highlight the importance of the work and potential implications of these data. The authors note that treatment of outer and middle ear pathologies can be achieved with “low-cost clinical interventions rather than expensive hearing amplification devices needed for sensorineural hearing loss.” Consider also noting that if these outer/middle ear pathologies are effectively managed early in the disease progression, it can prevent permanent conductive hearing loss caused by damage to middle ear structures. This would eliminate the need for long-term amplification for some patients. For this reason, effective and accessible treatment of outer/middle ear pathologies is not just cost effective, it is likely to *save* money for low-income nations in the long run, as it will reduce the incidence of permanent conductive hearing loss.  

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The authors responded satisfactorily to the criticisms posed and remedied them. Paper is now worthy of publication.

Author Response

Thank you to the reviewer for their time and consideration

Reviewer 2 Report

Comments and Suggestions for Authors

The points which suggested is improved, and the manuscript is acceptable for publication.

Author Response

Thank you to the reviewer for their time and consideration

Reviewer 3 Report

Comments and Suggestions for Authors

This revised manuscript has improved considerably since the original submission. I have only minor suggestions for further improvement. 

First, the authors' use the word 'prevalence' in describing their results, which is not entirely appropriate here. In epidemiological literature, prevalence is defined as the proportion of the *total population* that presents with a particular condition at a given point in time. Results in this study are reported in terms of the percentage of the *sample* that presented with a particular condition. For example, on line 17 of the abstract, it is reported that the 'prevalence' of abnormal findings was 98.2%. I suggest replacing prevalence with 'proportion of abnormal cases' to avoid confusion with the standard definition of prevalence. Alternatively, the authors could use an estimate of the total population of Lilongwe to determine the actual prevalence of outer and middle ear pathologies, since they report that this clinic is the main auditory service clinic in the region (line 75). Providing an actual prevalence estimate would strengthen the impact of the work and provide a meaningful benchmark for a difficult-to-quantify metric in this part of the world. In addressing this issue, be sure to also adjust wording on lines 116, 129, 189, and anywhere else in the document that prevalence is used to refer to the proportion of cases in this sample.

Second, the data in Figure 2 and described in the text (lines 127-139) are difficult to reconcile as percentages. My first thought in reading line 130 was that a sample containing 79% unilateral type A tymps and 49% bilateral type A tymps is… 128%? I then started adding up percentages in Figure 2 and found that 84.9% of the sample had bilateral (presumably symmetrical) tympanograms of various types, and 118.6% had unilateral tympanograms of various types. Does a single participant with asymmetrical tympanograms (Type A in one ear, Type B in the other) count twice (once as unilateral type A, and once as unilateral type B)? Does a symmetrical patient only count once (bilateral Type A)? I can’t figure out how to make these percentages add up to 100%. I suggest revising this section/figure or adding text to explain the coding process.

 

Finally, Figure 5 and 6 are excellent additions to the manuscript, and I suggest adding short paragraphs around lines 185 and 252 to explain them. The statistical findings are described in detail below in these sections, but the utility of the figures is minimized by not referring to them until the very end of the two sections.

There are also two minor figure issues that should be fixed. Percentages in Figure 3 overlap in a couple places, making it hard to read. Figure 5 is shifted to the right on the page and cutoff.

Author Response

First, the authors' use the word 'prevalence' in describing their results, which is not entirely appropriate here. In epidemiological literature, prevalence is defined as the proportion of the *total population* that presents with a particular condition at a given point in time. Results in this study are reported in terms of the percentage of the *sample* that presented with a particular condition. For example, on line 17 of the abstract, it is reported that the 'prevalence' of abnormal findings was 98.2%. I suggest replacing prevalence with 'proportion of abnormal cases' to avoid confusion with the standard definition of prevalence. Alternatively, the authors could use an estimate of the total population of Lilongwe to determine the actual prevalence of outer and middle ear pathologies, since they report that this clinic is the main auditory service clinic in the region (line 75). Providing an actual prevalence estimate would strengthen the impact of the work and provide a meaningful benchmark for a difficult-to-quantify metric in this part of the world. In addressing this issue, be sure to also adjust wording on lines 116, 129, 189, and anywhere else in the document that prevalence is used to refer to the proportion of cases in this sample.

Thank you to the reviewer for this feedback. Their point is well taken and we have updated as requested.

Second, the data in Figure 2 and described in the text (lines 127-139) are difficult to reconcile as percentages. My first thought in reading line 130 was that a sample containing 79% unilateral type A tymps and 49% bilateral type A tymps is… 128%? I then started adding up percentages in Figure 2 and found that 84.9% of the sample had bilateral (presumably symmetrical) tympanograms of various types, and 118.6% had unilateral tympanograms of various types. Does a single participant with asymmetrical tympanograms (Type A in one ear, Type B in the other) count twice (once as unilateral type A, and once as unilateral type B)? Does a symmetrical patient only count once (bilateral Type A)? I can’t figure out how to make these percentages add up to 100%. I suggest revising this section/figure or adding text to explain the coding process.

Thank you for making an excellent point. We agree with the comment, and accordingly, we have added the following section under methods and added a clarification to Figure 2 now, which reads:

Under Outcomes, we have now added “In addition to classifying each abnormal finding for the left and right ear, we also classified abnormal findings as unilateral and bilateral. When the abnormal finding was present in both the left and right ear, the abnormality was counted twice for the summary percentage and therefore may exceed one hundred percent”.  

We have added the following explanation to Figure 2 legend: Tympanometry Results broken down by unilateral vs bilateral findings. The percentage for each bar represents Type of tympanometry finding and may overlap with other tympanometry finding due to bilateral presence of different Type of abnormal tympanometry findings.

Finally, Figure 5 and 6 are excellent additions to the manuscript, and I suggest adding short paragraphs around lines 185 and 252 to explain them. The statistical findings are described in detail below in these sections, but the utility of the figures is minimized by not referring to them until the very end of the two sections.

Thank you for the suggestion; a reference to the new figures has been added at the beginning of each section.

There are also two minor figure issues that should be fixed. Percentages in Figure 3 overlap in a couple places, making it hard to read. Figure 5 is shifted to the right on the page and cutoff.

Thank you for this feedback -- an error seems to have occured with pasting in the figure and they have been fixed as requested