Sepsis Alert Application

03/26/17

As I discussed in my previous blog post, sepsis is a huge point of concern for Dignity Health and anything they can do to reduce the damage of sepsis is important to them.

It is for this reason that they have begun several different projects to monitor and predict the occurrence of sepsis to hopefully stop or lessen the damage of the complication.

Currently, they have created an alert system that monitors each step of the sepsis treatment process and puts it on a timer to ensure it is all done within the time constraints set for the specific task in the process of treating sepsis. Each step is marked by a digital light that is illuminated green as completed, yellow as in progress, or red as needing to be done. This is an easy visual to guide physicians as they deal with many many patients, as well as a means for the backend systems to remind doctors that certain patients may need treatment immediately. Allowing a large server to handle all the current patient data so that it can help physicians manage their time.

As time goes on, they are developing systems that will be able to predict the incidence of sepsis before it has even officially begun. This leverages big data to track all the different cases of sepsis and correlate which combination of factors commonly lead to sepsis, thereby allowing the server to throw an error that a certain patient may be at high risk of sepsis. Yet another example of how we can use big data platforms to process massive amounts of information to accomplish an incredibly important result.

Sepsis

03/25/17

Today I will be explaining the idea of sepsis, as it is definitely a hot topic in the medical world.

Many of my readers may be wondering what sepsis is, or may have never even heard of it. Essentially it is a general term applied to when the body attempts to fight an infection but ends up damaging itself instead, which is a very prevalent and life threatening condition in the hospital.

Many different things can show sepsis is/has occurred such as: heightened heart rate, fast respiratory rate, or very high body temperature.

It has varying levels of severity ranging from standard sepsis to severe sepsis all the way to septic shock. As a sepsis develops, it can be very dangerous to the patient, largely due to the possibility of blood clots forming which impair blood-flow and cause tissue/organ death.

This is a very serious topic as it is not completely understood which factors exactly can lead to sepsis. Infections are highly commonplace in your average hospital, which is why sepsis is such a big issue at Dignity Health.

Leadership in the Workplace

03/19/17

During this most recent week, as a break from the coding and collaboration on big data projects, my adviser assigned me to work on modernizing a presentation from several years ago about how to lead effectively. This largely focused on the workplace, but I would like to share it as I believe it is very useful information even outside of that context.

The company strives on working in teams and strong leadership to empower all the members of the company, so this presentation describes the most effective ways to lead.

Presentation

Systems Management

03/18/17

This week I was given the task of charting out the overall run-times for a system that the hospital has running every day to discover possibly points where the system could be improved to run quicker. It was very new to me as I was handed a massive excel spreadsheet with start and stop times of many many different processes.

The specific system had to do with gathering different elements of patient data to feed into other systems, pulling records and clinical information.

At first I simply drew out a chart in excel to show the time it took for each process to complete over the time of day when it began. Some trends developed as expected, with most of them being consistent throughout the whole day, however there was a large peak for all applications around 6 am. I was informed this is because the systems go offline for a few hours then have to makeup the missed data when they restart at 6 am. I just then created a bit of a reusable sheet in excel to chart the same run times for any day whose data was inputted into the sheet. I used this to check over several days and noted the specific problem areas and reported it to my adviser.

It felt like I had really been able to help with a real issue and work to improve an application for the hospital. It also gave me an appreciation for what really goes on behind the scenes, thousands and thousands of different small processes that feed into one main objective.

Genomics Presentation

03/12/17

I have created an extensive powerpoint presentation around the idea of genomics and precision medicine.

It aims to explain genomics in a cursory sense and then explain how we can use big data to improve healthcare with integration of genetic information. This new approach to healthcare is being called "precision medicine" and essentially means that the treatment will be specific to each patient based on their genetic makeup rather than just a generic treatment for each disease. Big data will allow systems to analyze what treatments work best for patients of similar genetic makeup and recommend the best treatment plan for new patients.

Genomics Presentation

This is a work in progress but I will be updating the linked presentation as I make edits to it.

Power of Modern Websites

03/11/17

As I have became more and more experienced with angular java, I have begun to understand truly what is possible with modern web development tools.

Dignity Health builds nearly all of their applications online rather than as something they would download to each and every computer. Traditionally this would seem less powerful as websites are generally seen as much more limited than your traditional program on your computer, but in truth, they are now sometimes just as powerful as the downloaded normal programs.

Seeing as Dignity deals a lot with big data and servers, they host most of the processing on their large servers and merely use the web applications as a portal to access the server data or pass it to the system to process rather the individual computers. This is a hallmark of the modern technology and an extension of the idea of the "cloud". Transferring the processing to a server, rather than on a local machine, allows these web applications to be just as powerful while also being mobile and accessible anywhere since it is just a website in its most basic form.

The way they build their systems here really shows me what big data and that ideology can lead to, more versatile and centralized systems that can be accessed from nearly anywhere.

Recap of Week Four

03/04/17

This last week has really been a step towards working hands on with the projects that I am involved in. I was a part of an initial meeting for a project that involves creating a new algorithm for attributing a certain physician with the length of stay of patients. Currently it is the discharging physician at some of their hospitals that has that patients length of stay in the hospital, even though they may have not had much face time with them. The LOS information is useful for grading physicians and noting who is taking longer than usual or who has the quickest turn around for patients, which is compared with if they are readmitted to the hospital. The new algorithm is based on the doctors notes written about the patient, it uses them to attribute who actually tended to that patient with that patients length of stay compared to the old system which was frequently innacurate. It was very exciting to be a part of a project creating a system that is going to be actually piloted and used within the hospitals. I was able to see how they plan and begin to execute different tasks within this organization, all invaluable experience for the future.

In my other main projects, I have progressed with designing a Twitter user interface to have the program more user friendly in the future and hopefully I can implement that as time progresses. The genomics project has mostly consisted of refining my IT guide to genomics that hopefully I will be presenting soon to the department. I also was able to acquire more data in my prescription drug price project in order to make it more usable and concrete for our hospitals. That project is discussed in more depth in this week's other blog post. Thanks again for checking in.

Kyle Grace

Prescription Drug Prices

03/03/17

Hospitals buy drugs in bulk from manufactures and distributors for a wholesale price (buying from costco vs target) which theoretically should be a better price than someone going to a pharmacy since they buy in bulk. This was the main idea of the project I set out on in the area of perscription drug prices. I was given some purchasing data, the different prices and buy orders from some hospitals to compare what we paid vs the retail prices in local pharmacies to see if we really receive a good price. What we pay in the end is paid for by the patient whether by insurance or out of their pocket. I ran this analysis for the highest overall expense drugs (price x quantity) to see where we could have the biggest amount of savings. Overall I found there were plenty of fine prices for some drugs, however others were costing us more than it would cost a patient to buy at the local pharmacy. These were the drugs I cared about and will pass on to the hospital to ensure we are being charged a fair price.

I also received data across all facilities that Dignity Health operates for the highest cost drugs in the last year and I am in the process of analyzing that data as well. In the end, this information could save the company money and battle for better prices for our patients. Thanks for keeping updated with my progress.

Kyle Grace