Assessing diabetic foot ulcers using mobile phone images is not reliable
An important new study investigating the validity and reliability of using mobile phone images to diagnose diabetic foot ulcer characteristics has recently been published in the journal Scientific Reports.
Mobile phone images are used for pretty much everything these days, including assessing and recording the progress of diabetic foot ulcers. In particular, clinicians with a challenging diabetic foot ulcer often seek expert advice from more experienced colleagues based on the mobile phone images they send. Up until now we just assumed these images were beneficial for assessing diabetic foot ulcers. That was until Queensland researchers Jaap van Netten, Damien Clark and colleagues decided to test the validity and reliability of using mobile phone images to assess 50 diabetic foot ulcer patients by 5 clinicians with different experience levels. And what they found was not what they expected!
The authors specifically tested the validity and reliability of diagnosing 12 important diabetic foot ulcer items: 9 clinical characteristics (such as identifying granulation tissue and infection) and 3 treatment decisions (such as “when should the patient be seen next in the clinic?”). As per any good scientific work, they did this by following a standardized research protocol. However, they also cleverly attempted to replicate real-world practice within this protocol to ensure it gave real-world findings.
First, they got an expert clinician to perform an in-person clinical assessment of the patient with a diabetic foot ulcer using the validated Queensland High Risk Foot Form. This expert completed questions on the 12 study items. Second, a research assistant took mobile phone images of the foot ulcer from 4 standard angles and distances. Third, a ‘remote’ clinician who had never seen the patient before viewed the images and some baseline Queensland High Risk Foot Form information, and completed the same questions on the 12 study items. Last, the same remote clinician repeated this process a minimum of two weeks later. They did all this with 5 different remote clinicians assessing 50 different patients twice; hundreds of assessments in total. They then compared the expert in-person assessments to the remote mobile phone assessments to see if they agreed.
What they found – after using some complex statistics that we’ll leave out of today’s latest research or we’d be here all day – was that only one (of the 12) item was valid and reliable (the treatment decision “is peri-wound debridement needed?”), regardless of the experience levels of the clinicians. All other items were both not valid (i.e.: the live assessment of the expert in the clinic differed from the remote assessment) nor reliable (i.e.: the remote assessors gave different assessment on two days). They concluded, “mobile phone images had low validity and reliability for remote assessment of diabetic foot ulcers and should not be used as a stand-alone diagnostic instrument”.
This was a reasonably robust study: it investigated a large cohort of typical patients with diabetic foot ulcers, clinicians with different levels of experience, a protocol that reflected typical daily clinical practice and some clever statistics. Yet, there were some limitations: the image quality was not controlled (although the authors report an overall high quality for images taken) and while the remote clinicians received some background medical and foot disease history, they didn’t receive any ulcer assessment history which probably would happen in typical clinical practice.
In summary, this was a high quality study that sheds important light on an area of diabetic foot ulcer management that has been assumed to be beneficial for many years now. The author’s main conclusion was that “mobile phone images …. should NOT be used as a stand-alone diagnostic instrument”. So, we hear you ask: how does this new research marry with the recommendations to use remote telehealth for people with diabetic foot ulcers in the Australian NHMRC diabetic foot disease guidelines and recently launched Australian diabetes-related foot disease strategy 2018-2022?
Well as the authors go on to conclude, “clinicians who use mobile phone images in clinical practice should obtain as much additional information as possible when making treatment decisions based on these images, and be cautious of the low diagnostic accuracy”. Importantly, they recommend this additional information includes: extensive communication platforms between the remote and expert clinician, previous foot ulcer assessment history, radiological or microbial information and potentially infra-red temperature measurements of both feet amongst others.
In conclusion, we suggest this is not the end of using mobile phone images for diabetic foot ulcer telehealth, rather it’s just the beginning. In the near future, we need to develop suites of diabetic foot ulcer diagnostic information that are readily available on mobile phone devices. But for now, do not trust a mobile phone image by itself, as you may miss-diagnose rather easily. Watch this space!