Scenario & Questions

Scenario:

The Office of Data Analytics of the South Australian Government, whose services include data analysis, visualisation, and presentation, hires you as a consultant to use health data to inform the public.

You are instructed to work on dashboards, so visual representation of data, for the SA Department of Health. Dashboards visualising health data received a lot of attention during the Covid19 pandemic when many people sought to inform themselves moScenario & Questions

Scenario:

The Office of Data Analytics of the South Australian Government, whose services include data analysis, visualisation, and presentation, hires you as a consultant to use health data to inform the public.

You are instructed to work on dashboards, so visual representation of data, for the SA Department of Health. Dashboards visualising health data received a lot of attention during the Covid19 pandemic when many people sought to inform themselves more deeply about the pandemic from a global to a local scale. However, dashboards are used in other Health areas as well. As an important aspect of SA health services in hospitals and the utilisation of their emergency rooms, "Ramping", which is ambulances having to wait for empty space in the hospital emergency room, played an important political role in the last state election. Consequently, there is a dashboard about the Emergency Departments in the state.

As your boss wants to use your information and the way to get there for the future when you have moved on to a different job or department, she requires that you fully explain your answers, the appropriateness of any relevant choice and selection, and mention any relevant assumptions necessary for your results. When you make a visualisation, make sure that the carts are fully and correctly labelled and titled.

Directions

The dashboard below reports data for South Australia's seven Emergency Departments (ED).

FMC - Flinders Medical Centre

LMH - Lyell McEwin Hospital

MH - Modbury Hospital

NHS- Noarlunga Hospital

RAH - Royal Adelaide Hospital

TQEH - The Queen Elizabeth Hospital

WCH - Women's and Children's Hospital

YOU MUST USE THE DASHBOARD PROVIDED SO THAT EVERYONE REFERS TO THE SAME BASELINE.

Figure 1.1:The Emergency Department Dashboard of SA Health

Download linkDashboard 10Feb.png

(You can find the live version online on the SA Health website, but please use this version for the assignment) (The table is below the question)

Step 1: Add new visualisations

Firstly, draw your attention to theStatustable at the top left of the dashboard. In this step you will be asked to make new visualisations of the data in this table.

Five relevant columns provide information about various aspects of the emergency departments at the seven hospitals in the state. In particular, you will need to look at the following

EA (Expected Arrivals): Patients currently en route to the hospital for treatment

WTBS (Waiting to be seen): Patients in the ED not yet in treatment

Com Treat (Commenced Treatment): Patients in the ED currently in Treatment

CAP (Capacity of the ED): Treatment areas in the ED

Avg Wait (min): Average wait time between arrival and assessment (over the last two hours)

The major focus is to prepare and derive appropriate visualisations for different parts of the table:

The starting point is identifying the data you are working with to better understand what might be an appropriate way to approach it. The table contains information from three different data sets for the seven hospitals, with the first three columns based on the same data set with individual patients as the level of observation, and the latter two columns coming from two separate data sets with the hospitals as level of observation of each. Identify the data sets underlying the shown data and the type of data. (4 pts)

Now you are tasked to make visualisations for all three data sets.

make a pie chart to visualise the data in the CAP column. Explain all relevant design choices and what aspects your chart is intended to highlight. (5 pts)

You are now tasked to select and produce a visual representation of the Avg Wait (min) column. Select an appropriate chart type and make a chart showing the values for the different hospitals. Explain your choice of chart type and any other relevant design choice. (6 pts)

The data in the first three columns can be seen as a cross-classification table. make an appropriate chart showing the values of this table. Explain what aspects of the data your chart highlights.(8 pts)

Step 2: Analysing data numerically

The next task concerns what information you can read out of the provided numerical values in the dashboard. The focus is on what the numbers show about the waiting times at hospitals in South Australia

The Avg Wait (min) column contains the means of the underlying individual waiting times at each hospital. Discuss whether the mean is the most appropriate measure of central location for this data or whether the median or mode would be a better choice. (4 pts)

Calculate the average and standard deviation of the Avg Wait (min) column, sothe average and standard deviation of the waiting times of hospitals. Shortly explain the meaning of your results.(4 pts)

The column contains the average waiting time at each hospital. In 2. you have calculated the average of these hospital-based waiting times. Now you are interested in the average wait time for a patient overall (so regardless of the hospital the patient was treated). Is it possible to calculate the average wait time of all individual patients in the state rather than for each hospital?If yes, explain how and then do so. If no, explain why it is not possible to do so. (8 pts)

Step 3: Improve existing visualisations

The third step focuses on improving an existing visualisation within the dashboard, namely the RAH Waiting Times table, which is visualised in the chart right next to the table.

Start with creating a different visualisation of just one category. Draw a CDF for the "waiting for bed" category. If the table does not provide enough information to make a definitive or correct design choice, make a choice that appears reasonable to you and provide an explanation and justification. (10 pts)

make an alternative visualisation of the whole table. Draw a different visualisation of the table, which you consider an improvement over the stacked bars. Explain your approach's advantages and disadvantages compared to the shown component bar chart. (Note: if there is no information provided for a category, such as Radiology, you can drop it)(15 pts)

Step 4:

This final step looks at the information you can deduce from the data about the relationships between different events. This will focus on the "RAH Streams" table.

Upon arrival at the ED, patients are assessed and assigned to a stream of care appropriate to their needs (Ambulatory, Assessment, Mental Health, Resuscitation, Unclassified). This component of the dashboard shows the breakdown for the selected hospital of the number of patients in each stage of care (e.g. 'waiting to be seen' or 'being treated') for each stream.

For the purpose of this step, you will focus on three columns:

"Wait to See - Tot"is the total number of patients at the time that has been assigned to a stream and currently waiting to see a relevant clinician.

"Treat - Tot", which is the total number of patients that have commenced treatment

"Wait for Bed - Tot",which is the total number of patients where treatment has led to a decision to admit to the regular hospital and are currently waiting to be moved onto an inpatient ward

Taking the streams and the column categories together, there are 8 events: 5 Streams (Amb, Ass, MH, Res, Unc) and 3 categories (WTS, TRE, WFB).

You are asked to start with providing some basic information about these events:

Build a contingency table with relative frequencies for the 5 streams and 3 categories(8 pts)

Determine the marginal probabilities for the 8 events. Show your calculations(4 pts)

A senior colleague passes on some probabilities involving two events that she things your supervisor might be interested in, though also indicates that your supervisor doesn't know how to read probability notation.

Calculate the following probabilities and then explain each in a sentence in normal language. P(Ass U Res), P(MH|WFB), P(Unc Amb), P(WTSI Amb), P(TRE - Res) (15 pts)

Your supervisor follows up and asks you to provide answers to a few more questions about the patients in the ED that he is interested in. As he wants you to send your answers to your senior colleague as well, you are also expected to include probability notation in your answer.

Calculate the respective probability and translate each question into the relevant probability notation. (9 pts)

What is the probability that a patient classified into the ambulatory stream is currently undergoing treatment?

What is the probability that a patient is classified into the resuscitation or the mental health stream?

What is the probability that a patient that is not a mental health case is currently waiting for a bed?re deeply about the pandemic from a global to a local scale. However, dashboards are used in other Health areas as well. As an important aspect of SA health services in hospitals and the utilisation of their emergency rooms, "Ramping", which is ambulances having to wait for empty space in the hospital emergency room, played an important political role in the last state election. Consequently, there is a dashboard about the Emergency Departments in the state.

As your boss wants to use your information and the way to get there for the future when you have moved on to a different job or department, she requires that you fully explain your answers, the appropriateness of any relevant choice and selection, and mention any relevant assumptions necessary for your results. When you make a visualisation, make sure that the carts are fully and correctly labelled and titled.

Directions

The dashboard below reports data for South Australia's seven Emergency Departments (ED).

FMC - Flinders Medical Centre

LMH - Lyell McEwin Hospital

MH - Modbury Hospital

NHS- Noarlunga Hospital

RAH - Royal Adelaide Hospital

TQEH - The Queen Elizabeth Hospital

WCH - Women's and Children's Hospital

YOU MUST USE THE DASHBOARD PROVIDED SO THAT EVERYONE REFERS TO THE SAME BASELINE.

Figure 1.1:The Emergency Department Dashboard of SA Health

Download linkDashboard 10Feb.png

(You can find the live version online on the SA Health website, but please use this version for the assignment)

Step 1: Add new visualisations

Firstly, draw your attention to theStatustable at the top left of the dashboard. In this step you will be asked to make new visualisations of the data in this table.

Five relevant columns provide information about various aspects of the emergency departments at the seven hospitals in the state. In particular, you will need to look at the following

EA (Expected Arrivals): Patients currently en route to the hospital for treatment

WTBS (Waiting to be seen): Patients in the ED not yet in treatment

Com Treat (Commenced Treatment): Patients in the ED currently in Treatment

CAP (Capacity of the ED): Treatment areas in the ED

Avg Wait (min): Average wait time between arrival and assessment (over the last two hours)

The major focus is to prepare and derive appropriate visualisations for different parts of the table:

The starting point is identifying the data you are working with to better understand what might be an appropriate way to approach it. The table contains information from three different data sets for the seven hospitals, with the first three columns based on the same data set with individual patients as the level of observation, and the latter two columns coming from two separate data sets with the hospitals as level of observation of each. Identify the data sets underlying the shown data and the type of data. (4 pts)

Now you are tasked to make visualisations for all three data sets.

make a pie chart to visualise the data in the CAP column. Explain all relevant design choices and what aspects your chart is intended to highlight. (5 pts)

You are now tasked to select and produce a visual representation of the Avg Wait (min) column. Select an appropriate chart type and make a chart showing the values for the different hospitals. Explain your choice of chart type and any other relevant design choice. (6 pts)

The data in the first three columns can be seen as a cross-classification table. make an appropriate chart showing the values of this table. Explain what aspects of the data your chart highlights.(8 pts)

Step 2: Analysing data numerically

The next task concerns what information you can read out of the provided numerical values in the dashboard. The focus is on what the numbers show about the waiting times at hospitals in South Australia

The Avg Wait (min) column contains the means of the underlying individual waiting times at each hospital. Discuss whether the mean is the most appropriate measure of central location for this data or whether the median or mode would be a better choice. (4 pts)

Calculate the average and standard deviation of the Avg Wait (min) column, sothe average and standard deviation of the waiting times of hospitals. Shortly explain the meaning of your results.(4 pts)

The column contains the average waiting time at each hospital. In 2. you have calculated the average of these hospital-based waiting times. Now you are interested in the average wait time for a patient overall (so regardless of the hospital the patient was treated). Is it possible to calculate the average wait time of all individual patients in the state rather than for each hospital?If yes, explain how and then do so. If no, explain why it is not possible to do so. (8 pts)

Step 3: Improve existing visualisations

The third step focuses on improving an existing visualisation within the dashboard, namely the RAH Waiting Times table, which is visualised in the chart right next to the table.

Start with creating a different visualisation of just one category. Draw a CDF for the "waiting for bed" category. If the table does not provide enough information to make a definitive or correct design choice, make a choice that appears reasonable to you and provide an explanation and justification. (10 pts)

make an alternative visualisation of the whole table. Draw a different visualisation of the table, which you consider an improvement over the stacked bars. Explain your approach's advantages and disadvantages compared to the shown component bar chart. (Note: if there is no information provided for a category, such as Radiology, you can drop it)(15 pts)

Step 4:

This final step looks at the information you can deduce from the data about the relationships between different events. This will focus on the "RAH Streams" table.

Upon arrival at the ED, patients are assessed and assigned to a stream of care appropriate to their needs (Ambulatory, Assessment, Mental Health, Resuscitation, Unclassified). This component of the dashboard shows the breakdown for the selected hospital of the number of patients in each stage of care (e.g. 'waiting to be seen' or 'being treated') for each stream.

For the purpose of this step, you will focus on three columns:

"Wait to See - Tot"is the total number of patients at the time that has been assigned to a stream and currently waiting to see a relevant clinician.

"Treat - Tot", which is the total number of patients that have commenced treatment

"Wait for Bed - Tot",which is the total number of patients where treatment has led to a decision to admit to the regular hospital and are currently waiting to be moved onto an inpatient ward

Taking the streams and the column categories together, there are 8 events: 5 Streams (Amb, Ass, MH, Res, Unc) and 3 categories (WTS, TRE, WFB).

You are asked to start with providing some basic information about these events:

Build a contingency table with relative frequencies for the 5 streams and 3 categories(8 pts)

Determine the marginal probabilities for the 8 events. Show your calculations(4 pts)

A senior colleague passes on some probabilities involving two events that she things your supervisor might be interested in, though also indicates that your supervisor doesn't know how to read probability notation.

Calculate the following probabilities and then explain each in a sentence in normal language. P(Ass U Res), P(MH|WFB), P(Unc Amb), P(WTSI Amb), P(TRE - Res) (15 pts)

Your supervisor follows up and asks you to provide answers to a few more questions about the patients in the ED that he is interested in. As he wants you to send your answers to your senior colleague as well, you are also expected to include probability notation in your answer.

Calculate the respective probability and translate each question into the relevant probability notation. (9 pts)

What is the probability that a patient classified into the ambulatory stream is currently undergoing treatment?

What is the probability that a patient is classified into the resuscitation or the mental health stream?

What is the probability that a patient that is not a mental health case is currently waiting for a bed?