INFS 4018 Business Intelligence and Analytics-Health Data Dictionary

This research seeks to analyse the operations of the emergency department on behalf of the New South Wales Productivity Committee. The emergency department has come under scrutiny due to deteriorating quality of healthcare offered, resulting in negative outcomes.

The research proves that poor services at the ED, leads to loss of precious time which may lead to overcrowding of patients and with all this confusion deaths may occur (Anon,2017).Therefore there is need to reduce the long stay of patients in the ED and in the long run reduce mortality rate.  

Project Justification:

 All  These terms are defined in working terms in the following chapters.

The measurements of these variables will be taken and examined statistically, with the aim of finding out whether changes in one or more of the variables lead to longer or shorter hospital stays. To help in guiding the research, the scope is formulated as exploring ways through which the long stays in the emergency department can be reduced.

Material for analysis of data

The data used for this analysis was made through numerous observations (620999). This makes the data be seen as accurate since it is first hand.The variables(arrival and departure dates and time, triage category, type of visit and the mode of separation) will assist in achieving a stable definition and analysis of data collected. For instance, the departure date and time will be taken to mean the date and time when the admitted patient, having undergone preliminary treatment in. On the other hand, arrival date and time will denote when the patient is presented for treatment at the ED. 

The mode of separation is a term taken to mean the health status of the patient at the time of departure.  The mode of separation clarifies the way the separation occurred, as well as the condition the patient was in at the time.  An ED visit type refers to the reason for which a patient goes to the ED. It could be scheduled, unscheduled, among other specifications.

Analytic approach

The analytic approach has been organized in steps to enable more comfortable relation, both with the diagram that follows, as well as with the steps followed to conduct the research.

Steps

1: 620999 observations and the ensuing statistical and other analysis will be the main part of the study. The analysis will first aim to establish the outlier and extreme values, as well as the maximum and minimum values.

2: this step will involve removal of the outliers, and their replacement with maximum and minimum values. This will be considered as the cleaning up phase of the data collected.

3: having cleaned up the data, a general frequency analysis will be applied. This will be aimed at categorising the different patients under separate classes, including the separation mode, the triage and visit type.

4:  Analysis of Variance will be performed for the main variable under consideration - Stay Time Hours, and how it interacts with each of the variables which are independent, and they include separation mode, triage category and ED visit type

5: A fifth step will entail identification of the significant value, so that the mode of the study can then be derived.

 6: Make an assumption of the hypothesis: “Mean value of “Stay Time Hours” for each mode of separation is equal.

7: Make an assumption of the hypothesis: “Mean value of “Stay Time Hours” for each ED category visit type is equal.

8: Make an assumption of the hypothesis: “Mean value of “Stay Time Hours” for each triage category is equal.

9: Having calculated the difference between the means of groups as laid out in the preceding hypothesis, the last step will then be completed.

Findings

The data analysed revealed that 97.76% of the patients who visit the emergency room do so because of visit type is equals to 1. The result in this case implies that the ED should target them while to help bring down the length of stay.

Method of seperation 

 The results showed that 56% of emergency department visitors can be attributed to the mode of separation=4, while 28.9% of the visitors are of method of separation-1. As in the previous analysis, the study found that these people should also be targeted if the hospital is to reduce the time of stay. 

Further analysis of the data revealed that 42.7percent of people who visit the department dealing with emergency are in the category of triage is equals to 4, while 33 percent of people who visit the department dealing with emergency are in Triage category equals to 1.

The details of the workings and the reasons for these implications are contained in the appendix

The research ascertained that targeting these patients will be critical in helping reduce the length of stay.

The dependent variable’s mean value: Stay Time Hours is Equal for the visit types to the emergency department

For this study, the default value of alpha is 0.05. After applying ANOVA, the p-value generated was 0.0001. Given that the value is less than the value of alpha, the null hypothesis was rejected. It was further concluded that the mean values of the 12 groups, from 1 to 12 were not equal. This difference is taken by the researchers to mean that the patients who fall in the groups with a higher mean should also be the subject of drives to reduce the length of ED stay.

Interpretations

From the data analysis it is evident that the system is not suitable. This was evident as the default value of alpha which was set as 0.05 yet after applying the ANOVA, the p-valued derived was 0.001, which is lower than the value of alpha. This error was found to be quite extreme as it  led to a rejection of the null hypothesis. 

1. It also meant that the values of the five groups under consideration were not equal, meaning that the patients falling under the groups with higher than average means should be the focus of quality improvement. From the analysis of the data, two of the five groups were in this category.
2. The findings also indicated that another factor was the lack of proper clarification of people. This is caused by not using the full data dictionary.

Recommendations

Having analysed the data, the results were further examined to identify the factors that could have led to the discrepancies. Among the main possibilities was erroneous coding, as well as incidents of patients arriving by an ambulance after the paramedics have already done triage.

In order to resolve this, the research proposes that the first step should be a clarification of the identity of the patient and reasons why the patient was not triaged. It is also recommended that there should be more extensive usage of the full data dictionary. This will help the department to have accurate and complete data of the patients they are dealing with. The recording of triage needs re-examination so that the department can do it more efficiently than was the case before. It must however be pointed out that the onus is on the decision-makers to examine these options using variables such as time, effort, expenses and arrive at the best decision.  

Data Dictionary:

The data disctionary is a collection of definitions for the puposes of forming a consistent working explanation of vaious terms seen as being significant to the study. These include the following.

• Departure time and date.

For the patient admitted, the departure date comes in two options.

1. When the patient is moved to the ward or a new unit which is separate.
2. When a patient moves from the emergency to another healthcare facility.

For the patient who is not admitted, time of departure should be taken to denote when the patient has already received preliminary treatment.

• Arrival date and time

This refers to when the patient is first present for health services at the ED.

6) Mode of separation

The mode of separation refers to the status, health wise, of the patient when the initial treatment at the ED has already been performed. The mode separation also gives details of the nature of separation, including where the patient then went, and the pace with which the separation was done where applicable.

 7) Emergency department visit type- The reason for a patient’s visits to the ED.

Visit types:

• Return visit – a visit which is planned is expected. It has been planned so as to follow up on treatment, or when there are test results which suggest that the patient should seek further treatment. Therefore, all planned visits involve previous patients to the ED
• Unplanned Return Visit for continuing condition – This refers to visits by patients who have previously been treated by the ED, but were not expected back for further treatment or for follow-up. This also means that the patient visits the ED seeking treatment for the same conditions for which they first visited.
• Presentation of the Outpatient – visit happening because of planned visit to a regular clinic.
• Admission which is pre-arranged:In this case, the patient and the ED have already arranged the admission. As such, there is no inspection or initial check up by ED staff. The patient may or may not be triaged
• Pre-arranged Admission: Includes ED check-up –While the admission has been prearranged, the patient does go through the usual admission procedures including triage and inspection by ED staff.
• Person to transit– The ED is responsible for administering treatment on the patient, but there is not expectation on their admission, since they are on transit elsewhere.
• Dead On Arrival– A patient who arrives at the Ed dead and receives no medical care by the ED has death on arrival as the documented mode of separation.
• Disaster –A visit due to a natural calamity or a significant man-made accident or deliberate attack, where a disaster plan is followed.
• Presentation of tele health –Here, the treatment of the patient is conducted through audio and visual equipment.

1. Summary of the steps taken during the exploration of Dataset (where individual results are consolidated), plus data quality.

In a set of data, one variable tasked with response “Stay Time Hours” is involved. It is derived by ascertaining the difference between arrival time and departure time.

To concur with the study’s objectives, “To reduce long stay of patients at the ED,”another variable is brought in to ensure analysis of Stay Time which obtained from current dataset obtained for review by identifying the difference between actual arrival and departure time.

The formula is:

 (Actual time of departure) – (actual time of arrival) = Stay Time Hours

This is the total number of hours by the patient at the ED. For example time duration for which the patient remains at the ED.

Being a derivable of the analysis, focus has centered on four variables outlined below:

 Results:

Result Number 1-The number of patients who fall under the categories which include triage category, ED visit and the mode of separation.

• Analysis 1 – exactly 97.76 percent of patient who visit the ED are visit type equals to 1. They should be the focus of efforts to bring a reduction to the stay time.

• Analysis 2 –the separation mode for 56 percent of those people who arrived at the ED was = 4visit emergency department are identified as the mode of separation =4. 28.9% of the ED patients had the mode of separation as =1. Likewise, the research proposes their targeting as a step towards reducing the time of stay.
• Analysis 3 – Further data analysis revealed that 42.7% of ED patients were categorised under triage category = 4, while another 33% were grouped under Triage category = 1. This should be another target group in efforts to make a reduction to the stay time.

 

 Result Number Two: Outlier, maximum and minimum values in a database which is existent according to the details: 

Outlier values are 3679.50000 and -8759.93333

Minimum Value: 0.2

Maximum Value 21.6

Mean Value: 4.8

Result Number Three: Maximum and minimum values in cleaned sets of data.

After data is cleaned by getting rid of outliers below are Minimum, Maximum and Mean values details:

Result Number Four: Null Hypothesis: Mean value of Dependent Variable: Stay Time Hours is Equal for all 5 groups of triage category.

After application of the ANOVA test to compare mean values of all groups of triage category, the result is: 

Alpha’s default value has been denoted as 0.05. Since the p-value derived from the ANOVA test is 0.001, the Null hypothesis is rejected. The conclusion here is that the mean values of the five groups are not identical. This further implies that those patients in groups with a higher mean should be target in efforts to make a reduction of stay time for that specific group.

Result Number Fi: Null Hypothesis: Mean value of Dependent Variable: Stay Time Hours is Equal for all 13 groups of mode of separation

After applying ANOVA test to compare mean values of all 13 groups of mode of separation, the following result is arrived at:

The alpha value of 0.05 is found to be higher than the p=-value derived after the application of the ANOVA test. Therefore, the null hypothesis is rejected. Additionally the men values of the 13 groups under consideration are found to be different. For the two groups whose mean is higher than the rest, a renewed focus on the patients in them should be employed so as to reduce the duration of stay time.

Result Number Six: Null Hypothesis: Dependent Variable’s mean value: Hours of Stay Time is Equal for all 12 groups of Emergency Department visit type. 

Alpha’s default value is equals to 0.05 and p-value that has been derived from calculation of the ANOVAs is .0001. The lesser total of p-value, in relation to alpha, leads to a rejection of the Null hypothesis. In conclusion, Group 1 to Group 12’s mean values are not equal meaning that the mean value of a minimum of 2 groups out of 12 groups are not equal and are not similar to each other. This therefore means that, patients who come to form the group whose mean value is very much higher will make up the focus group so as to bring about a reduction in the time duration of stay for that individual group. 

Member’s contribution

 Reference:

Anon., n.d. NSW Health Admitted Patient Data Collection. [Online] Available at   HYPERLINK "https://internal.health.nsw.gov.au/im/ims/ap/index.html "  https://internal.health.nsw.gov.au/im/ims/ap/index.html.

Anon., n.d. NSW Health Data Dictionary. [Online] Available at   HYPERLINK "https://internal.health.nsw.gov.au/im/ims/standards/nsw-health-data-dictionary.html  "   https://internal.health.nsw.gov.au/im/ims/standards/nsw-health-data-dictionary.html.

Anon., n.d. NSW Health Emergency Department Data Collection. [Online] Available at   HYPERLINK "https://internal.health.nsw.gov.au/data/collections/edc/  "   https://internal.health.nsw.gov.au/data/collections/edc/.

 (Anon., 2017)www.sas.com  (Anon., n.d.).