Development of System for Healthcare Communication

Contents

EXECUTIVE SUMMARY

INTRODUCTION

Vision Statement

Mission Statement

Goals and Objectives

BACKGROUND

Current Challenges

Audit of current system

BUSINESS VALUES

MARKET ANALYSIS

PROPOSED SYSTEM

How to Use It

System Architecture

Technology

CHALLENGES

SWOT ANALYSIS

RISK ANALYSIS

WORK BREAKDOWN STRUCTURE

COST ESTIMATION

REVENUE MODEL

1. Traditional Revenue Model

2. Backup Model

THREE YEAR PROJECTION

APPENDIX A

APPENDIX B

APPENDIX C

EXECUTIVE SUMMARY

Protracted wait times at the emergency department (ED) are associated with increased frustration and mortality, and decreased patient satisfaction. Nearly 4 out of every 10 patients are exasperated even before their check-up begins. Reducing the waiting time at the ED can be quite challenging.

These wait times are mainly due to the triage process which is the sorting of and allocation of treatment to patients. This involves hospital admission, patient waiting for a bed, scarcity of medical staff and an abnormally high number of patients die to local accidents or disasters. As ED wait times often result in delayed treatment for patients who require immediate attention, hospitals must improve on increasing the efficiency of registering and prioritizing patients.

We are making use of technologies such as Machine Learning / Deep Learning in order to analyse, search and record medical data of patients, resource allocation and treatment recommendations. Text Recognition is used to standardize the existing health records of patients. Natural Language Processing (Add on) is used to transcribe doctor patient conversations for medical records.

Our system, Fidelity Healthcare will help to increase communication between departments and thereby improve efficiency. Our system consists of two parts – one in the ambulance which would send the patients’ information to the hospitals, thus reducing the time that goes on to collecting patient’s data and the second part in the hospitals which would help in allocating resources.

INTRODUCTION

Vision Statement

To help hospitals and patients decrease ED waiting time and improve resource allocation in hospitals by using Artificial Intelligence (AI) for the growing population.

Mission Statement

To solve problems to improve communication between different hospital departments and to improve patient satisfaction by upscaling the handling capacity of hospitals to provide immediate treatment to 90% of patients within 4 hours of reaching emergency department.

Goals and Objectives

Table 1 – Goals and Objectives

BACKGROUND

The current challenges faced by the patients and the hospitals are as described below:

Current Challenges

      Triage Time in emergency departments.

      Resource allocation problem in terms of staff and beds.

      Time that goes into collecting patient’s data.

      Staffing and resource allocation during the peak time.

      Communication waiting status between departments.

      Standardization of health records.

Audit of current system

When a patient is transported to the hospital in an ambulance or reaches the hospital in a private car, the patients are accessed and triaged after reaching the emergency department. Prioritising the patients and then treating them according to their medical condition takes time. (Betterhealth.vic.gov.au, 2019)

Using artificial intelligence in the emergency department automates the prioritising of patients which reduces the time of admissions. Also, details of patients reaching the hospital by ambulance are shared prior to reaching the hospital which helps the hospital triage the patient before the patient reaches the hospital. (Ambulance.nsw.gov.au, 2019)

BUSINESS VALUES

The proposed solution will improve the efficiency of the emergency departments of the hospital. This project will reduce the pressure from the employees due to the increase in the number of patients visiting the emergency department every year. This project also focuses on time offloading of patients from the ambulance and offloading emergency departments.

The solution focuses on improving the communication within the emergency department by the automatic allocation of resources. Improving the communication between the ambulance and hospital will improve the emergency response time by sending the patients information to the hospital in advance so that the hospital can make necessary arrangements in advance.

MARKET ANALYSIS

According to a survey by the NSW government, the number of patients visiting the emergency departments is increasing every year. About one in five adults and one in four children visit the hospital at least once a year. (NSW Government, 2019)

In Australia, there are 695 public hospitals and the initial target market which is NSW has a total of 305 hospitals out of which 222 are public hospitals, out of which about 143 are small hospitals (less than 50 beds), about 28 are big hospitals (more than 200 beds) and rest come in the middle bracket. (Australian Institute of Health and Welfare, 2019). Figure 1 shows the ratio of number of beds in the public hospitals in Australia.

­­­­

Figure 1 : Ratio of beds in public hospitals

PROPOSED SYSTEM

To reduce the turnaround time in the emergency departments, an AI system will be implemented in the ambulance which will send the patients data to the hospital and suggest diagnosis based on the patient’s data. The system installed in the hospital will allocate resources according to the inventory available in the hospital.

The existing medical records of all the people will be entered in the system so that AI can propose the diagnosis in a more efficient way by reading the medical history of the patient.

The system implemented in the Ambulance will read the patient’s vitals and send the details to the hospital and the necessary requirements are arranged before the patient reaches the hospital.

The system implemented the hospital will display the proposed diagnosis for the patient based on the information received from the ambulance. The system will notify the doctor and the nurses and keep them on standby for the timely treatment of the patient. The system will also automate the hospital inventory by notifying if any emergency beds can be offloaded.

Howto Use It

For using the system, the hospital will be required to enter the details of the hospital like hospital staff, number of beds in the hospital, Doctor’s information available at the hospital, etc. The hospital staff will be trained to use the new system. The ambulance respondent would be trained to use the system to send the details to the hospital using the new system.

System Architecture

Figure 2 : System architecture
 

The proposed system, as shown in Figure 2, will send real-time information of the patient from the Ambulance to the Hospital. The system in the Hospital will then suggest immediate treatment for the patient to the Ambulance. It will also allocate resources (doctors, nurses, operation or laboratory rooms) for the incoming emergency and notify them regarding the same.

Technology

Figure 3: System technologies
 

Coming to the technology used in this project, the hardware components are basic and are readily available in the market. Figure 3 shows the technology flow in the system.

A computer and keyboard would be required in the Hospital, which is already present. The hardware components used in the ambulance along with their functions are summarized in the following table.

Table 2: Hardware Technologies

The technologies used for building the software is as follows:

Table 3: Software Technologies

CHALLENGES

Figure 4 : Challenges

SWOT ANALYSIS

Figure 5 : SWOT Analysis

RISK ANALYSIS

1. Client-Side Risks:

Figure 6 : Client-Side Risks

2. Business-Side Risks:

Figure 7: Business-Side Risks
 

WORK BREAKDOWN STRUCTURE

Figure 8: Work Breakdown Structure

The three components of the proposed solution are- Ambulance App, Hospital App and Hospital Admin app. Although all the components are essential, but the nucleus of the solution will be Hospital App.  Entire development period of the proposed AI powered healthcare tool, as per existing scope, is 86 days. Further bifurcation of this period with respect to corresponding modules is given below-

1. Ambulance App- 17 days
2. Hospital App- 46 days
3. Hospital Admin- 7 days
4. Quality Assurance- 16 days

Total= 86 days

COST ESTIMATION

It is often said that saving a life cannot be quantified but the development cost of the software tools required to do so can be. Considering the present IT industry standards, the estimated cost of development of the proposed product is AUD 400,000 which will need to be disbursed in first quarter post sign-offs. The estimated amount comprises of the salary of the resources required for product development along with the initial set up cost. (Payscale.com, 2019)
 

REVENUE MODEL

Core revenue model of the company will be based on subscriptions. However, due to unforeseen challenges a backup revenue model is also developed for financial sustenance of the company.
 

1.     Traditional Revenue Model

In traditional revenue model, hospitals will be prime customers of the company and users of the proposed solution. Each hospital will be categorised as small, medium and large based on the number of beds it provides to patients. (Australian Institute of Health and Welfare, 2019) For every category, there will be an exclusive subscription offer as shown in Figure 9.

Figure 9: Subscription Model

Each subscription will have a term of 1 year and each new subscriber will get the benefit of first month free trial.

2.     Backup Model

In this model, customer segment of the company will shift from hospitals to IT companies already providing healthcare solutions. Plan of action in this scenario will comprise of targeting IT healthcare companies and offering the proposed product as plug-in services. This will enable the customers, IT companies, to improvise their existing products and instigate user innovation. On the other side, this will facilitate creation of alternate revenue channels for the provider.
 

THREE YEAR PROJECTION

After launching our solution in the market in the second quarter of the first year we will approach the hospitals in NSW initially. Then, based on the progress and successful implementation, we will move our solution to other states such as Victoria, Queensland, South Australia and so on, covering all the public hospitals in the continent.

With the salaries, office, server and other expenditure we will become profitable in the third quarter of the second year. Taking the acceptable take up rate as 35% and a quiet decent drop rate of 20%, we will have about 213 hospitals using our solution by the end of the third year with a team of about 30 people including developers, AI specialist, marketing agents and support agents as well. (Salaries and Salary, 2019) (Payscale.com, 2019) (Datamation.com, 2019)
 

CONCLUSION

This project would be applying a new way to reduce waiting hours in emergency departments in NSW which would in turn help hospitals to serve patients better by providing health care attention. It’s crucial to reduce wait times to provide better care to patients and also to improve communication within hospital departments. ED mortality rates and departmental crowding explain that crowding should be considered as a public health concern.

Collecting patient’s health record in the ambulance and sending to the hospitals prior to the admission can save time that goes onto collecting the details in the hospitals and assignment of resources, be it beds, or doctors, will be much more simplified. Investments in the right health care solutions can provide a viable future with measurable impact, within a generation. More flexible EDs can get, the better value they will provide.

REFERENCES

• Australian Institute of Health and Welfare. (2019). Hospital resources 2016–17: Australian hospital statistics, Report editions – Australian Institute of Health and Welfare. [online] Available at: https://www.aihw.gov.au/reports/hospitals/ahs-2016-17-hospital-resources/report-editions [Accessed 10 May 2019].
• NSW Government. (2019). Improving service levels in hospitals. [online] Available at: https://www.nsw.gov.au/improving-nsw/premiers-priorities/improving-service-levels-hospitals/ [Accessed 10 May 2019].
• Google Docs. (2019). Cost Estimation 3 year model. [online] Available at: https://docs.google.com/spreadsheets/d/1TBjf3x63Wza5mtbiiCdCbu_rUHeUF9tPJmBsyASP2F0/edit?usp=sharing [Accessed 11 May 2019].
• Payscale.com. (2019). Front End Developer / Engineer Salary (Australia) | PayScale. [online] Available at: https://www.payscale.com/research/AU/Job=Front_End_Developer_%2F_Engineer/Salary [Accessed 12 May 2019].
• Datamation.com. (2019). Artificial Intelligence Salaries: Paychecks Heading Skyward. [online] Available at: https://www.datamation.com/artificial-intelligence/ai-salaries.html [Accessed 12 May 2019].
• Salaries, A. and Salary, M. (2019). Marketing Agent Annual Salary ($40,057 Avg | May 2019) – ZipRecruiter. [online] ZipRecruiter. Available at: https://www.ziprecruiter.com/Salaries/Marketing-Agent-Salary [Accessed 12 May 2019].
• Payscale.com. (2019). Customer Support Representative Salary (Australia) | PayScale. [online] Available at: https://www.payscale.com/research/AU/Job=Customer_Support_Representative/Salary [Accessed 12 May 2019].
• Payscale.com. (2019). Quality Assurance Analyst Salary (Australia) | PayScale. [online] Available at: https://www.payscale.com/research/AU/Job=Quality_Assurance_Analyst/Salary [Accessed 12 May 2019].
• Betterhealth.vic.gov.au. (2019). What to expect when you arrive at hospital. [online] Available at: https://www.betterhealth.vic.gov.au/health/servicesandsupport/hospital-admission-what-to-expect [Accessed 12 May 2019].
• Ambulance.nsw.gov.au. (2019). Frequently Asked Questions – NSW Ambulance. [online] Available at: http://www.ambulance.nsw.gov.au/Calling-an-Ambulance/Frequently-Asked-Questions.html [Accessed 12 May 2019].

APPENDIX A

Project 3 Years Projection:

APPENDIX B

Project Cost Estimation:

(Cost Estimation 3-year model. Google Docs, 2019)

APPENDIX C

Project Schedule – Gantt Chart: