Healthcare Analytics – EMR Systems Need Help

Electronic Medical Record (EMR) systems have proven their worth within the healthcare system by improving patient care, reducing costs, and augmenting patient safety. These systems contain a wealth of information that is currently under-utilized and not analyzed effectively. With the correct analytics technology, a great deal of insight can be garnered from such data to help improve the operations of the healthcare provider.

Within the last several years, many EMR vendors have caught wind of this and have upgraded their core product with additional data analytic capabilities. In theory, this sounds like an excellent combination. However, in practice, these solutions present many challenges.

The following are challenges met by many EMR systems:

  • Numerous EMR vendors haven chosen to build their own add-on analytic software, rather than utilizing an established commercially available platform. This proves to be a difficult undertaking for the vendor, steering them off course from their core offering and producing a rudimentary analytics solution subpar to those currently available.
  • Data integration with other external systems such as financial and operational is complex and difficult to achieve by built-in EMR analytics. This prohibits the healthcare provider from gaining insight across domains thereby eliminating the possibility of comprehensive reporting or analysis.
  • Data visualization is often not user friendly and designed more for a data analyst rather than a healthcare provider or business user. Proper visualization of EMR data requires domain expertise in the healthcare space and a high degree of design knowledge to present data simply and effectively. EMR vendors may often have expertise in one, but not both, of these arenas.
  • Healthcare systems often require customization of analytics software in order to meet specific requirements; whether it be to analyze their patient data or achieve clinical metrics. Many current built-in EMR analytics systems are unable to achieve this degree of customization, often leading to erroneous results which could potentially impact the healthcare system’s financial and patient care outcomes.

These challenges often force healthcare systems to develop their own analytics capabilities, a daunting feat which requires time and resources often not available. What is truly necessary is an analytics software provider that not only knows healthcare well, but is also able to utilize the best-in-breed analytics software. This, coupled with the domain expertise within the healthcare space, will lead to the creation of an integrated solution that meets the requirements of the healthcare provider.

Mobile Analytics – To App or Not To App?

“App,” a word previously only recognized by individuals with a computer degree, became universally known after Apple released the App Store for its iPhone mobile devices in 2008. Since then there have been over 1 million apps created with more than 60 billion downloads. Google soon caught on to this wave and launched its own app store, as did Microsoft. Apple set forth a trend defining how people interact with software on their mobile devices.

Analytics software on a mobile device should follow this same paradigm. If you want to analyze emergency wait time usage, readmission rates, or clinical coding, the phrase “there’s an app for that” should come to mind. The user experience should be comparable, if not superior, to that provided by other popular apps like Angry Birds, or Google Maps.

In a study completed by Flurry which analyzed time spent on mobile devices in the US, the firm concluded that consumers spent approximately 2 hours 42 minutes each day in 2014. The time spent was broken down into 2 categories: Mobile-Web and Apps. Unsurprisingly, US consumers spent 86% of their time interacting with their mobile device via Apps and only 14% via the mobile web browser.

Based on this evidence, there are several key components to making a successful analytics application on a mobile device:

  • Native gestures supported by the device such as swipe, pull down, pinch and expand should be utilized as much as possible when interacting with the visualizations
  • The user interface should refactor itself automatically and take advantage of different screen sizes to offer the best user experience (tablet vs. smartphone vs. watch)
  • Interactivity with the visualizations and data should be instantaneous. For example, filtering on a field such as year or region should have almost no delay
  • To accommodate slow or spotty network connections, the data should be cached on the device as much as possible or completely provided in an offline mode and synchronized
  • The app should utilize and integrate with an organization’s mobile device management (MDM) infrastructure
  • Security features such as local data encryption on the device, expiration of content, application pin and others should be provided and made configurable to the organization’s standards

Analytics vendors who follow these principles will see much better user adoption rates than those who try to provide a “one-size-fits-all offering” (defined as a single solution for desktops, tablets, and phones typically a web browser interface). The latter option is undoubtedly a more attractive choice for the analytics vendor given the ease in reusability, lower development costs and less maintenance. However, such an offering in the long run will suffer a much higher cost – lower user adoption.