Satish Misra, MD contributed to this post

Patients with diabetes are increasingly turning to their smartphones for help in managing this chronic disease. They now have more choices than ever in regards to apps that help maintain their blood glucose, provide ideas for meals, and help maintain a healthy weight and lifestyle. At quick glance, there are over 100 apps for patients with diabetes in the app store.

Created by SAYDT (Sistema Informatizado Ayuda Diagnostico Tratamiento), the app is for patients with Type 2 Diabetes who use insulin. According to its iTunes description, the user enters their current insulin regimen and a few other parameters, and the app delivers recommendations on insulin doses “automatically adjusted to get optimal glycemic control.”

Given the burden that diabetes places on our patients and our healthcare system, its important that healthcare providers take the lead on helping patients figure out which resources are useful. Here we look at InsulinSmart, both the app itself and some of the important questions it raises about patient safety in this as-yet unregulated environment.

The Potential Danger

Defining an insulin regimen is tricky business – the benefits of success are long-term and the consequences for error are severe and quick. It can often be a relatively time-intensive process for a patient involving frequent blood glucose checks, multiple trips to a physicians office, and so on.

InsulinSmart claims to go beyond the usual blood glucose diary by allowing users to formulate an “optimal” insulin regimen. Other apps – for example Insulin Calc – apply user defined parameters to calculating nutritional or correctional mealtime doses of insulin.

There are several important issues to consider with these apps, taking InsulinSmart and Insulin Calc as examples. First, neither app clearly defines and supports how their calculations are made. Second, neither app has any sort of defined safety mechanism – either pre-market safety controls or post-market adverse event reporting. Now, granted, we may be asking a bit much from developers who don’t have websites (check out the developer sites for InsulinSmart and Insulin Calc), though I suspect few of us would take a medication from a pharmaceutical company that doesn’t have a website.

To be fair, these apps – and others like them – could be very well-designed and effective. However, the lack of the critical pieces of information to show that is, in and of itself, enough of a reason to consider them a potential danger. One feature of the mHealth industry is that many folks from non-healthcare industries have wandered into the medical world. For us, a bad product can mean lost lives and injured people – hence a naturally risk-adverse nature that is sure to be foreign to individuals in the IT industry. So when it comes to an app that calculates insulin doses, I want to know that it has been rigorously tested before I recommend it to a patient and the absence of that knowledge makes these kinds of apps a potential danger

These kinds of apps are also sure to garner the attention of FDA regulators. Based on the draft guidance on mobile medical apps, these apps would all qualify as clinical decision support tools and be subject to regulation. And the absence of a learned intermediary between the app and the clinical action (the injection of insulin) would likely lead to these apps being considered higher risk – imagine any number of simple errors (a decimal point off) that could lead to a fatal outcome.

They could also garner the attention of the FTC, particularly when claims of improving glucose control are made. As we saw with the recent FTC action against developers of acne “treatment” apps, claims of clinical efficacy are taken quite seriously and the inability of the developer to deliver on the promised outcome can have significant consequences.