2. Recommendations for Follow-Up, Including Frequency of Testing According to the Latest

My Kidney Summary We propose to develop an application that can run on various devices (e.g., smart phones, tablets, personal computers) that will demystify kidney function for patients that have sustained acute kidney injury (AKI). Our user-friendly application focuses on three keys areas:

1. Interpretation for the user of his/her serum creatinine and estimated glomerular filtration rate (eGFR) based on his/her age, sex and race; 2. Recommendations for follow-up, including frequency of testing according to the latest practice guidelines; 3. Easy-to-use interface with information on medications that should be avoided or that require dosing adjustments based on the patient’s kidney function. Problem statement and motivation Fact 1: AKI is common (>0.2% annual population incidence, similar to myocardial infarction) and can lead to death or development of chronic kidney disease (CKD).1 Fact 2: Most drugs are excreted by the kidneys, and residual kidney dysfunction after AKI may cause drug accumulation, leading to toxicity—in some cases further injuring the kidney.

Fact 3: 70% of Americans take at least one prescription drug5 and approximately 50% take two or more prescription drugs.6 Note that (a) these estimates do not include over-the-counter drugs and (b) the number of prescriptions filled increased by 39% from 1999 to 2009.6 Experts have identified post-AKI management as a critical opportunity to improve outcomes.2 Patients sustaining AKI are frequently unaware and lack specific instructions as to what they should do in order to recover and stay healthy. Unlike cholesterol and blood pressure, normal values for serum creatinine vary by age, sex and race. Therefore, patients need help monitoring kidney function and with drug dosing adjustments to avoid adverse drug events. Solution description Our project will combine the expertise of physicians, pharmacists, patient educators, human factors experts, and computer scientists to design “My Kidney,” an intuitive, easy-to-use, free app that can run on smart phones, tablets and personal computers and that will empower patients recovering from AKI to take more control of their health. The app will allow patients to track their creatinine and interpret it based on their age, race and sex using the latest CKD-Epi equation and other simple to understand resources. The app will also provide easy access to information on medications in a non-medical way, and will allow the user to personalize her/his health monitoring experience after AKI. Understanding kidney function. Human beings constantly produce a waste product of our metabolism called creatinine that is removed from our body by our kidneys and excreted into the urine. Matching removal with production, we maintain a constant blood level of creatinine. Since creatinine production is relatively constant over the short-term for a given individual, we can use changes in serum creatinine to estimate the glomerular filtration rate (the amount kidneys can filter per unit time). However, creatinine production varies across individuals and in the same individuals over long periods of time. Thus, creatinine must be interpreted on a personalized basis based on age, sex, and race. Our solution will not take the place of expert medical advice but rather provide a tool for patients to facilitate conversations with their doctors and to track their kidney function over time. Post-AKI follow-up. Patients surviving AKI and not receiving dialysis have a variable clinical course.3 Recommendations for follow-up are available from the Kidney Disease Improving Global Outcomes (KDIGO) clinical practice guidelines.4 Through use of our application, these recommendations can be personalized to the patient. Medications and the kidney. There are no websites to assist people (using in laymen terms) about avoidance of nephrotoxins and drug dosing adjustments after an AKI. The limited websites available direct healthcare professionals with appropriate drug dosing,7 but are not user friendly for the general population: they use generic drug names and medical terminology such as “Q8h”. Ours will be the first information portal built to be understandable by the general public, to use brand as well as generic drug names, to include over-the-counter drugs, and to contain understandable directions. The website will be accessed through our mobile app. Technology. To carry out the development of this app, we will initially use an Android platform, given that Android covers about 75% of smartphone market,8 and offers a platform that is compatible with Java, JavaScript (JS) and HTML5, not requiring proprietary software to develop apps. Java, JS, and HTML5 are very portable, with the latter two being powerful and very efficient for web development (e.g., all the data can be easily displayed in text or tables). Portability is essential, enabling apps to be easily deployed to many platforms, and connect to laptops, tablets, and desktops, which the target population is likely to own. We’ll use a platform such as PhoneGap to facilitate and expedite app development, and also create the necessary interfaces with local datastores to start with and upload data to a more powerful device/ computer opportunistically. Native apps will not be developed because web apps (Java, JS, and HTML5) are already sufficiently fast and their portability is much higher (e.g., no need to port from Java to ObjectiveC when changing from Android to Apple devices). Human Factors. The overall design goal is to create a user friendly product that will help patients control their health. Our app will address aspects of human factors and usability during the design process. For example, since the targeted users tend to be older, we need to focus carefully on vision, dexterity, and memory issues. Features such as the size of the type, target size and placement of icons, and color contrast will be addressed. We will also add a voice interface. The interactions and effects must be simple and easily learned to minimize the user’s cognitive load. We will allow users to take pictures of medications instead of entering the information on the app. Providers will also be able download and enter information for users initially to increase adoption. This will result in a streamlined app that will accommodate older people, but at the same time yield a tool that can be embraced by a full range of patients. Computer Science Portion:

The tablet/smartphone application (app) and web interface (web-app) will provide access to data. We will use a MySQL database, standard in the industry, easily accessed by apps and web-apps alike. The back-end server will reside in a secure network, and access to the data will be protected (password, fingerprints, FaceLock, etc), and HIPPA compliant. The app user interfaces will be decided in months 1-2 to expedite implementation, which will use PhoneGap or Android Studio, to facilitate the multi-pronged approach (app, web-app, and phone, tablet, browser). A few user interfaces will be implemented, to facilitate the use of the apps: (a) automatic drug recognition from the vial/label (take a picture, immediate recognition); (b) web pages/app pages allow users to fill out surveys and provide feedback about usage, ease of use, and efficiency of the system; (c) interaction thru (color coded) alerts, through blinking, SMS, push notifications, and/or email; (d) optionally, with doctor agreement, the systems will generate a digest of alerts to doctors, to facilitate monitoring; and (e) optionally, study the feasibility of using text-to-speech interface, with standard libraries or with human voices due to very limited vocabulary, so that caregivers do not need to look at device (while driving, etc). We will test the app using unit and regression tests, on 3 different platforms: a laptop using 3 different browsers, a tablet, and a phone (different brand from the tablet). Efficiency of app is the response time, that is, the time from starting the app until the response is received, including accessing the server for drug information, calculating the needed metrics, generating the alerts. Deliverables:  Trial app and website with patients;  Storage (database?) schema for server-side and client-side  Form to collect user feedback to update website and app;  Carry out more extensive testing using Testology.com or utest.com;  Study feasibility of voice recognition (server-side);  Automatically download patient information to the app upon discharge.  Team introduction John A. Kellum, MD (Critical Care Medicine); Sandra Kane-Gill, PharmD (School of Pharmacy); Hoda Kaldas, MD and Jose Bernardo, MD (Renal-Electrolyte Division); Daniel Mosse, PhD (Dept Computer Science); Michelle Fowler, RN, BSN (A Renal Patient Educator at UPMC); Gayle Hess, MDes (independent contractor, Human Factors Expert). Scale up feasibility statement Using a user-centered design approach, the product design will be tested at the end of each of phase by several volunteers from the targeted user group. The knowledge gained from these brief tests will help improve the design.

Description of a customer (sustainability) and user for the free app The target app users are patients who have recently been discharged after sustaining AKI. The app will sustain itself by (a) designing it with little maintenance from the start; (b) selling ads or being acquired by device manufacturers; (c) getting contributions from studies and NIH funded projects; (d) potentially selling aggregated data (not individual data). References: 1. Kellum JA, Bellomo R, Ronco C. Kidney attack. JAMA : the journal of the American Medical Association. 2012;307(21):2265–2266. doi:10.1001/jama.2012.4315. 2. Goldstein SL, Jaber BL, Faubel S, Chawla LS, Acute Kidney Injury Advisory Group of American Society of Nephrology. AKI transition of care: a potential opportunity to detect and prevent CKD. Clin J Am Soc Nephrol. 2013;8(3):476–483. doi:10.2215/CJN.12101112. 3. Chawla LS, Eggers PW, Star RA, Kimmel PL. Acute kidney injury and chronic kidney disease as interconnected syndromes. N Engl J Med. 2014;371(1):58–66. 4. KDIGO AKI Work Group. Kidney Disease: Improving Global Outcomes (KDIGO) Clinical Practice Guideline for Acute Kidney Injury. Kidney inter, Suppl 2012; 2: 1–138. 5. Zhong W Maradit-Kremers H, St. Sauver JL et al. Age and sex patterns of drug prescribing in a defined American population. Mayo Clin Proc. 2013;88(7):697-707. 6. The Kaiser Family Foundation. Prescription drug trends. http://kaiserfamilyfoundation. files.wordpress.com/2013/01/3057-08.pdf (accessed Oct. 1, 2014) 7. McAuley D. Renal dosing- database. h t t p : // w w w . g lobal r ph . c o m / i nde x _ r ena l . h t m (accessed Oct. 1, 2014) 8. h tt p : / / ww w . c ne t. c o m / ne w s / and r oid - b e a t s - i o s- 5 - t o - 1 - i n - q 3 - s m a r t phon e -m a r k e t -s h a re/ (accessed Oct 1, 2014)