Date Published: April 17, 2019
Publisher: Public Library of Science
Author(s): Saeedeh Moayedi-Nia, Leila Barss, Olivia Oxlade, Chantal Valiquette, Mei-Xin Ly, Jonathon R. Campbell, Zhiyi Lan, Placide Nsengiyumva, Federica Fregonese, Mayara Lisboa Bastos, Danielle Sampath, Nicholas Winters, Dick Menzies, Walter Dehority.
The Tuberculin Skin Test (TST) is a relatively simple test for detecting latent tuberculosis infection (LTBI) but requires regular quality assurance to ensure proper technique for administration and reading. The objective of this study was to estimate the accuracy and reproducibility of an mhealth approach (the mTST) to measure the size of swelling immediately following TST administration (TST injection bleb) and after 48–72 hours (TST induration).
Five non-clinical and one clinical reviewer measured the size of TST injection blebs, and TST indurations using smartphone acquired photos of sites of TST administration and readings in patients, or saline injections in volunteers. The reference standard was the onsite measurement (measured by an experienced TB nurse) of the actual TST injection bleb, or induration. Agreement of reviewers’ measurements with the reference standard, as well as agreement within and between reviewers, was estimated using Cohen’s kappa coefficient.
Using the mTST method to assess bleb size in 64 photos of different TST injections, agreement between reviewers, and the reference standard was very good to excellent (κ ranged from 0.75 to 0.87), and within-reviewer reproducibility of readings was excellent (κ ranged from 0.86 to 0.96). Using the mTST method to assess TST induration in 72 photos, reviewers were able to detect no induration (<5mm) and induration of 15mm or greater with accuracy of 95% and 92% respectively, but accuracy was only 20% and 77% for reactions of 5-9mm and 10-14mm respectively. The mTST approach appears to be a reliable tool to assess TST administration. The mTST approach was accurate to read indurations of 0-4mm or 15+mm, but less accurate for reactions of 5-14mm. We believe the mTST approach could be useful for training and quality assurance in locations where on-site supervision is not possible.
Based on recent estimates, approximately 1.7 billion people worldwide have latent tuberculosis infection (LTBI). Among this vast population, an estimated 10% (170 million people) will develop active tuberculosis (TB) over their lifetime. Testing and treatment of people with LTBI has been shown to significantly reduce the risk of progression to active TB. Testing for LTBI is an important element of LTBI management. Either a tuberculin skin test (TST) or interferon gamma release assay (IGRA) may be used for diagnosis. The TST is often the only method available, particularly in low and middle-income countries (LMIC) where the incidence of active and latent TB is highest[4,5]. However, incorrect administration technique has been cited as a potential cause for false negative TST results, and poor reading technique may also result in incorrect results. While the techniques for TST administration and reading are relatively simple, training and supervision are needed to establish and maintain proficiency, as for any diagnostic test.
Ethical approval for this study was obtained from the research ethics board of the McGill University Health Centre Research Institute (MUHC-REB# 2018–3787). Patients provided verbal agreement to allow the nurse to take a photo of the TST site (administration or reading). The verbal consent procedure used English or French verbal scripts, which were approved by the MUHC-REB. As no patient identifying information was collected, written consent was not deemed to be necessary. A note was placed in the patient’s chart if verbal consent was given to participate in the research study. To maintain confidentiality, no personal or identifying information was collected, and photos were not taken if there were any potentially identifiable information or marks (e.g. tattoos, birthmark) in the area of TST injection. When photos were stored electronically they were labelled only with the date of the photo, and the nurse’s measurements.
In this study we have shown that reviewers using the mTST approach had good accuracy in identifying TST injections of correct and incorrect amounts of test material, and that the interpretation of these photos is highly reproducible, based on between and within-reviewer agreement. Using the mTST approach, reviewers could identify TST indurations less than 5mm and larger than 14mm accurately, compared to the reference standard of on-site measurements by an experienced TB nurse. However, this approach was less accurate for indurations of 5–14 mm. Agreement within and between-reviewers was very good–indicating that reading indurations with the mTST method provided reproducible results.
In conclusion, the mTST approach appears to be reliable for assessment of technique of TST injection. Although this method is less accurate to detect small TST indurations in the 5-9mm range, it is much more reliable for larger indurations, particularly indurations of 15mm or larger. We believe the mTST is a simple and easily implemented method that could serve as a useful training and quality control tool in many settings.