4/12/2022

Client Information

  • Tara Nibhanupudy, Grad Student
  • Boston University School of Medicine
  • Department of Otolarynthology - Head & Neck surgery

Introduction

  • Meta Analysis of four studies
  • Effectiveness of intranasal corticosteroid sprays at alleviating the symptoms of Eustachian tube dysfunction (ETD)
  • ETD: a chronic blockage of the Eustachian tube
    • connects the middle ear to the back of the nose
  • Intranasal corticosteroid sprays
    • Sprayed directly into the nose
    • Use for treating ETD is currently not clinically advised
    • Often prescribed in an effort to avoid more invasive treatments

Eustachian TubeEustachian Tube

Eustachian Tube

Tympanometry

  • Alleviation of symptoms was determined by measuring tympanometry
  • Tympanometry: how the ear drum moves with sound stimuli and is the most objective method used to characterize resolution
  • There are three classifications
    • A denoting healthy levels of tympanometry
    • C being some abnormality
    • B being severe levels of tympanometry
  • Determine whether an ear has completely, partially, or not all resolved from symptoms of ETD

Example of a Tympanometry Graph

{width = 100%} ##

Differences between the Papers

  • How the data is reported
    • By ear vs By subject
  • Whether or not to differentiate level C1 and C2
    • B -> C1: Partial Resolution
    • B -> C2: Complete Resolution
  • length of the treatment.
    • Barati 2011: the duration of the treatment is 4 weeks,
    • Bhargava 2014: the the duration of treatment is 24 weeks.

Our Project

  • Earlier this semester: Provide Tara with visualization options for her poster presentation
  • Now: Create a meta-analysis with four of the studies
    • Barati 2011
    • Gluth 2011
    • Lildholdt 1982
    • Tracy 1998
  • Studies were chosen based on experimental methods and data collection
  • Outcome is comparing proportion of ears that are normalized post control/INCS

Revisiting the Data

  • Barati and Tracy might show intervention is helpful
  • Gluth difference between intervention and control
  • May predict that the meta analysis will not favor either condition

Meta Analysis Visualizations

  • Used to check for publication bias in meta-analysis
  • All four studies fall within the region
    • conclude that there is no evidence of publication biases present

  • The error bars around each of the studies cross the line of null effect
    • Studies are likely to have a p value >0.05
  • All of these studies have small sample sizes
    • Apparent from the size of the tails on the point estimates
  • Barati 2011 and Tracy 1998 are to the right of the line of null effect
    • Point estimates favor the control
  • Gluth 2011 and Lindholdt 1982 are to the left of the line of null effect - Point estimates favor the treatment
  • Neither the common nor mixed effects clearly favor either condition

Conclusion

  • Fail to reject the null hypothesis
    • that intranasal corticosteroid sprays have an effect in the treatment of Eustachian tube dysfunction
  • The studies do not provide sufficient evidence to conclude that an effect exists

Appendix

-Previous visualizations created for Tara

This figure shows the results from all of the studies. Results have been aggregated in order to be shown in one figure. This graph represents results by subject.

This figure shows the results from all of the studies. Results have been aggregated in order to be shown in one figure. This graph represents results by subject.

This figure shows the results from all of the studies but leaves off proportions of no change. Results have been aggregated in order to be shown in one figure. This graph represents results by subject.

This figure shows the results from all of the studies but leaves off proportions of no change. Results have been aggregated in order to be shown in one figure. This graph represents results by subject.

This figure shows the results from studies that reported the proportion of change in condition. All participants began at a type B and change was defined as moving to either type C or A. The results used in this figure are by subject.

This figure shows the results from studies that reported the proportion of change in condition. All participants began at a type B and change was defined as moving to either type C or A. The results used in this figure are by subject.

This figure shows the results from studies that reported the Patients' status with three levels: Complete resolution, Partial resolution and Unchanged

This figure shows the results from studies that reported the Patients’ status with three levels: Complete resolution, Partial resolution and Unchanged

This figure shows the results from three studies that reported the proportion of patients' status with two levels: Complete resolution and Partial resolution.

This figure shows the results from three studies that reported the proportion of patients’ status with two levels: Complete resolution and Partial resolution.