Médecine et chirurgie du pied


Hallux Angles: A new medical software for smartphone for data archiving and angular measurements of foot operated for hallux valgus. Study and discussion on the software reliability Volume 30, issue 2, Juin 2014


This article was presented to the SFMCP congress (13th–14th December 2013)/Cet article a été présenté au congrès SFMCP 2013 (13–14 décembre 2013).


The purpose of this scientific research was to: 1) Verify the reliability of “Hallux Angles” (https://itunes.apple.com/gb/app/hallux-angles/id398338105?mt=8) in terms of archiving of sensitive data about the photos of the deformities of patients, and in terms of the correction index measured on radiographs before and after surgery; 2) Compare the indexes of angular deformities of patients measured with three methods: a) With an iPhone software, “Hallux Angles”; b) With a standard goniometer; c) With a PC software (Agfa CR) commonly used in radiology; 3) test the stability of the application over the time.

Materials and methods

Four observers filed 95 patients affected by hallux valgus, operated on in different clinics and hospitals from September 2012 to January 2013, with the application “Hallux Angles” adding personal data for each patient, such as the birthday, gender, and radiographic photos taken in the upright position and angle values measured (distal metatarsal articular angle (DMAA), hallux valgus angle (HVA), and intermetatarsal angle (IMA)) before and after surgery. The photos relative to the post-surgical procedure were taken at 40 days. With the data that emerged from the postoperative angle measurements, we obtained the corrective index. Measurements obtained with the software were then compared with those emerged using a standard goniometer and then verified thanks to a reliable PC software (Agfa CR), supervised by a radiologist, giving us the opportunity to test the reliability of the measurements collected. The variability of the obtained results from the two different methods was similar and superposable among the four observers. In particular, the deviation between the two methodologies (manual measurement with goniometer and measurement with the application) was 1.50 ± 1.2 for HVA, 1.2 ± 0.90 for DMAA, and 1.4 ± 1.1 for IMA. The final verification of the data through the measurement with radiology equipment by a radiologist produced a correlation coefficient according to the Bland-Altman method and was very satisfactory, and it was close to the statistical measurement of 95%, thus confirming an index correlation very reliable and with very small deviations.