Diabetes is a major global health problem and is characterized as one of the largest epidemics. It is important for patient to strictly follow the treatment. However, there is commonly difficult to record and presentation of the data of treatment progress. It was found that regularly record the GLU and diet and exercise can help patient monitor and predicate the health status, so that it can benefit the health and convenience the treatment. SMART on FHIR has been widely used in the Australian healthcare system. It improves the efficiency of clinical workflows interoperability and data security, and achieve better medical outcomes. Therefore, the aim of this project is to build an application with SMART on FHIR to provide a useful tool for type 1 diabetes treatment.

SMART is a standards-based health technology platform, composed of source tools, publicly accessible libraries, and open standards, allowing innovators to create applications that can run safely in the healthcare system. The SMART technology is significant to Australian healthcare system as it is highly secure and can provide seamless user experience when run across any healthcare system (Pais, Parry & Huang, 2017).

FHIR is used for resources expansion and adaptation, and no matter how the system is developed, any system can interpret it. FHIR stands for "Fast Healthcare Interoperability Resource". HL7 developed FHIR as a standard in response to the growing use of EHR. FHIR is an interoperability specification for the exchange of healthcare information electronically, as the healthcare industry develops, data interoperability has always become a critical issue for care providers to attract patients, improve the efficiency of clinical workflows, and achieve better medical outcomes. The aim of using FHIR is to try to solve the increasing problem of digitalisation in the healthcare industry and simplify patient medical records to make them easier to access, understand and discover (Jánki & Bilicki, 2018).

Therefore, SMART is related to technology of application part, and FHIR is related to data. These two concepts combined together to create a framework for building a medical application without rely on a specific supplier. SMART on FHIR has been wildly used in Australia healthcare system. The SMART Universal Health Application Platform can help integrate FHIR clinical data into one data store through an unlimited user interface. It also can help healthcare providers deploy applications faster and also help users to phase out outdated applications without losing data (Radović, 2019).

Diabetes is a chronic disease. It occurs when the pancreas cannot produce enough insulin to control blood sugar levels or the body develops insulin resistance. Diabetes is the biggest challenge to the Australian health system in the 21st century. Diabetes is the fastest growing chronic disease in Australia. Compared with other chronic diseases such as heart disease and cancer, it grows faster (Rowlands, 2020).

In Australia, about 1.7 million people have diabetes. Diabetes also could cause many complications to patients such as Blindness, Heart disease,, and Kidney disease, etc(Rowlands, 2020). There are two types of diabetes, Type 1 and Type2. This app will focus on Type 2 diabetes.

Type 1 diabetes was formerly known as insulin-dependent diabetes or juvenile diabetes. The pancreas produces almost no insulin for patients. Various factors can cause type 1 diabetes, including genetic factors and certain viruses. Even though type 1 diabetes appears more in adolescence or childhood, it still can develop in adults. There is still a cure for type 1 diabetes. The focus of treatment is to prevent complications by control blood sugar levels through insulin, lifestyle, and diet. As long as patients can control blood glucose levels (GLU) as close to the normal level as possible, they can prevent both long-term and short-term complications (Mayo, 2020). 

The diabetes treatment is managed by delivery of insulin through a pump, while regularly monitoring of blood glucose levels, taking regular exercise, and following an eating plan and a healthy diet (Mayo, 2020).

This app focuses on make better diabetes treatment process. The app can monitor and record the blood glucose changes it measures, check device status, and record diet and activities.

In Australia, about 1.7 million people have diabetes. For every person diagnosed with diabetes, they usually need family members or caregivers to support and take care of them. This means that about 2.4 million Australians suffer from diabetes every day(Mayo, 2020).

Therefore, the main stakeholder of this application will be patients with diabetes and their families, caregivers, as well as healthcare workers, specifically doctors.

Diabetes is a chronic disease that requires long-term treatment. A large number of patients spend a lot of medical resources to treat diabetes. In addition, diabetes requires patients to take medication continuously and regularly, measure blood sugar, control their diet, and exercise (Mayo, 2020). When using this application, users can clearly and accurately know the changes of their GLU, control their diet to avoid the increase of GLU, record activities, and most importantly check the device status and link status. Therefore, the application can help patients, their families and caregivers have an easier treatment process. Also, as the application can accurately record-a long period of GLU changes, thus, it can provide complete and accurate data to healthcare workers and doctors, they can have a better analysis of these data, easier diagnosis, and make further treatment suggestions.

4.1 structure and flow
4.2 Iteration
According to feedback from paper prototype testing, we made a prototype on XD and modified some small parts of design, such as the sort order of device status list, give more information about what user should input when they adding new diet or personal training.

The second round testing used XD prototype vision 1. We iterate the prototype on XD based on user feedback. In this iteration, we changed the design of Diet and PT page. The Diet page used to use column diagram to present diet intake and Carbohydrate. However, most of users didn't understand this diagram. Therefore, we change this part to circle diagram to clearly describe total intake, Carbohydrate on total intake, and remaining diet that they can intake today.

We also change the PT page. This page used to use 24 cube as 24 hours and different colours as different activities. However, majority of users think this is unnecessarily complex, and they even need to count "cube" to know the time. Therefore, in the XD prototype, we also use a cycle diagram to present how much exercise they have done this week and how much they still need to do. We also put activity name, intensity and time in a plan list, so that user can know more detail about their personal training. The "Adding" button used to be in each page, however, we delete some of them and only keep it on Diet page and PUT page. Because the "Adding" is only used for add new diet and new personal training, so it could misleading users if have it on every page, because some of users thought the "adding" button on the device status page is used to add new device.

4.3 Fianl prototype vision description

GLU Dashboard is the homepage of this application. The device can measure patient's GLU and record it, and generate a curve. This curve can save data up to 1 week and the curve representation period could filter as 24h, 48h, 72h or 1 week. At the bottom is some time period, a part of curve is red colour and a circle on it is in order to point out the real time GLU to user, and also a number on the top to give specific value, the arrow beside the number represent if the GLU increase or decrease than before. The green curve after the red curve part is the predication of future GLU. Moreover, user can see the diet and personal training intake value and time on the graph.

There is also a diet page on the dashboard. This page shows users how many carbohydrates you have intake today and how much are left. Also shows use the insulin and carbs amount. User can add new diet by click "adding" button, It has the carbs value of each food, and count as part of user intake today. At the top of this page is the insulin value sent by the pump, and the carbohydrate value will be automatically calculated based on your input intake.

The last page on dashboard is PT. User can make activity list and record their daily training time. As exercise is a kind of diabetes treatment, thus, it is important to regularly do trainings and reach the required training time. This page can show user how long they have done on training per week and how long they still need to do. It also can check the training history of other date. The user can also add a new personal training by clicking the "Add" button, and the user can select the type and time of exercise they do for this training.

The device status page aggregate all devices and their status. It shows the remaining age or remaining battery of each device. If the progress bar is green, it means that the battery or device age can still be used for a long time. If is red, it means it is nearly out of power and need be charged soon. The activity bar chart shows the connection time and status of pump and Openaps.

The "Notification" page can notify the user when the device out of power or its service life expires.

5.1Card sorting
We conducted a card sorting test with 10 participants, we used the open card method, ask them to sort the card. This testing aims to confirm the flow and navigation of the prototype. According to their sorting result, the majority of them put "equipment connection" and "accessories age" into the "Setting" page, and aggregate "Intake", "Daily exercise", "Real-time GLU" and "Blood sugar trend" into the Main page. Then they put "Reminder setting" into the "Device" page. Therefore, we discuss the sorting result and confirm the use of this as the flow and navigation of the prototype. **5.2 Paper prototype and testing **

Task
1.check the status of the pump battery.
2.Add new diet and personal training.
3.Check the PT that you have done today
4.check the calories of carbohydrates
5.check the GLU curve within 24 hours

Observation 1 Most of the task can be done easily. Took a long time on think about the input of an additional new diet and personal training. Suggestions: Add more specific details about what should input when adding a new diet.

Observation 2: Didn't understand what the meaning of the cube, time? Suggestions: Add descriptions to GLU, DIET and PT

Observation 3: Had a question about the add new diet, the user said what if she doesn't know the total calories of her diet? Suggestion: Add a calorie calculator, so that the user can clearly know the total calories.

Observation 4: Most of the task can be done easily. suggestion: The device status list should be re-ranged, grouped by reminding day and percentage, don't mix them.

Observation 5: Confuse about the colour on the PT page, why there are deep green and light green? Are they represent the same activities? Suggestion: Add more explanation about the colour on the PT page.

5.3 XD prototype and USER TESTING 1.0

Link: https://xd.adobe.com/view/e3521436-0089-4b53-823d-1b8ca2a8eb1b-b601/

Task
1.check the connection status of the pump.
2.Add diet after eat an apple.
3.Check the Personal Trainings that you have done yesterday.
4.Check how many carbohydrates you can intake today

78 female Most of the tasks can be done easily, however, the “Adding” button misled her, she thought only the button on the diet page can add a new diet.

1. The best features of this tool were to check the personal Training.

2. The worst features of this tool was the "adding" she thought the Add button on the device status page is used to add a new device rather than add new diet and personal training.

3. I suggest the following improvements for this tool is that make the “adding” button more navigational.

76 Male Most of the tasks can be done easily, but the remaining intake percentage of carbohydrates make him feel inconvenient.

1. The best features of this tool was check the status of devices.

2. The worst features of this tool was Check how many carbohydrates you can intake today.

3. I suggest the following improvements for this tool is that not use percentage to show the remaining carbohydrates, because users don’t know the total carbohydrates that they can intake, should give them a more specific number.

Male, 22 years old, major in sports science In addition to not finding the entrance to add functions, because I am a young man, I have a lot of experience in using the app and it is very smooth to use.

1. Reset the position of the icon to be added. You can't know what is added before you click in.

2. Since the training standard is a week as a unit, the training page can be directly turned into weekly

3. According to Australian health instructions, there is a difference in intensity for 150 minutes a week. For example, in exercises with a heart rate greater than 80%, the calculation time can be multiplied by two.

Female, 52 years old, mild diabetes The abbreviations for blood sugar and personal training on the main page are not clear, which leads to a long time spent on understanding. I was very hesitant when I was looking for additional functions. I clicked on the food interface before I clicked the add icon.

1. The icons of equipment and settings cannot be understood at once and can be changed to bells and the like

2. You can add some explanations for each device component

3. Add function can only add diet and training, you can not add the function in the device and setting interface

Male, 25 years old Use and complete the task very smoothly

1. Add some intensity differences in the personal training section, or you can enter the heart rate when adding

2. For personal training, time is more important than calorie consumption. You can replace each indicator with time.

3. Eating records can replace the percentage of carbohydrates with the total grams of carbohydrate content in each food.

SUS result 65.5

5.4 Real app testing

In google doc

Jánki, Z. R., & Bilicki, V. (2018, June). Full-stack FHIR-based MBaaS with Server-and Client- side Caching Capable WebDAO. In THE 11TH CONFERENCE OF PHD STUDENTS IN COMPUTER SCIENCE (p. 170).

Mayo, C. (2020). Type 1 diabetes - Symptoms and causes. Mayo Clinic. Retrieved 23 October 2020, from https://www.mayoclinic.org/diseases-conditions/type-1-diabetes/symptoms-causes/syc-20353011.

Pais, S., Parry, D., & Huang, Y. (2017, January). Suitability of fast healthcare interoperability resources (FHIR) for wellness data. In Proceedings of the 50th Hawaii International Conference on System Sciences.

Radović, B. (2019). Integration between electronic prescribing systems and pharmacy information systems using the FHIR standard (Doctoral dissertation, Univerza v Ljubljani, Fakulteta za elektrotehniko).

Rowlands, H. (2020). Diabetes | Type 1 Diabetes | Type 2 Diabetes | MedlinePlus. Medlineplus.gov. Retrieved 23 October 2020, from https://medlineplus.gov/diabetes.html.