NIT3171 Business Analytics: 1199016

Introduction 

This report aims at presenting descriptive, prescriptive and descriptive analyses performed on the clinics data given by the selected medical organisation business in Australia. Based on the findings, it will thus be easy to decide on whether to establish small pharmacy inside or open up another clinic branch in another area. The report also shows the current business finding, strength, weakness, and the possible development approaches as discussed on sub sections.

Descriptive

This part of the report presents the descriptive analysis as discussed in sub sections below.

1. The statistical details of data, such as patients’ cities, ethnic background, gender and age.

i. Patient’s cities

By considering the cities where the patients originate from, the figure below shows the distribution of their origin. The majority of the patients originate from Hinchinbrook (about 14%, 9795 patients) whereas the second highest city where the patients came from was Hoxton Park (about 9.1%, 6430 patients). Also, about 8.8% (6245 patients) of the patients came from Middleton Grange city.

Fig 1: Distribution of top 15 cities patients’ cities

By closing the top 15 cities, about 2.3%, (1608 patients) of the patients originate from Greenfield Park.

ii. Ethnic Background

By taking ethnic background of the patients into consideration, results showed that majority of the patients were of Australian origin (about 29.7%, 21121 patients). Also, about 27.5% (19439 patients) were of Iraqi origin while the third highest tribe was the Torres Strait Islander (about 15.7%, 11150 patients). The fourth highest tribe was the Assyrian (about 7.1%, 4994 patients).

Figure 2: Distribution of major ethnic groups

iii. Gender of patients

By considering gender distribution, majority of the patients were females (about 66%, 46854 patients) while males patients were 24624 (34%). The 3D pie chart below shows the distribution of gender of the patients.

Fig3: Gender distribution

iv. Age of patients

The distribution of age was also analysed and it was found that the average age of the patients was approximately 38.4 with a standard deviation of 22. The oldest patient was 98 years old and the youngest patient was aged less than 1 year.

2. The statistical details of data, such as the highest 10 descriptions, drugs and reasons 

i. Highest 10 descriptions

By analyzing the top 10 most description, the highest was surgery consultation of level B (about 37.4%, 26605 patients) which was closely followed by direct-billing incentive (about 30.0%, 21435 patients). The third mots description was Surgery consultation of the level C (represented by about 11.8%, 8382 patients).

Figure 4: Highest 10 descriptions

ii. Drugs

By analyzing the drugs, the table below gives the count and the average percent of each drug within the clinic. As per the table, the most administered drug was Augmentin Duo Forte (875mg; 125mg Tablet)

Drug NameCountPercent
Augmentin Duo Forte 875mg;125mg Tablet27803.9%
Keflex 500mg Capsule26793.7%
Amoxycillin 500mg Capsule11711.6%
Redipred 5mg/mL Solution11351.6%
Ventolin CFC-Free 100mcg/dose Inhaler10671.5%
Panadeine Forte 500mg;30mg Tablet9891.4%
Mobic 7.5mg Tablet8771.2%
Amoxycillin 250mg/5mL Syrup7901.1%
Roxithromycin 150mg Tablet7851.1%
Amoxil 500mg Capsule7771.1%

iii. Reasons

By considering the reasons which was noted down, the ‘Recall” stood out greatly (3.7%, count = 2592). Reminder was about 2.7% (count = 1646). Also, URTI, bacteria was also represented by 2.7% (count = 1847). The last in the top 10 category was the care plan that was represented by 1.8% (count = 1278). The figure below shows the distribution and obtained via the analysis.

Fig 5: Reasons

3. Estimated percentage of each type (antibiotic, chronic and pain relief)

By dividing the medicine into three categories, the chronic drugs forms main percentage (41.1%) while pain relief drugs were the least estimated (24.1%). The table below shows the descriptive analysis in regard to this.

Medicine Estimated Percentage
  
i. Keflex  – antibiotic drug  33.8%
ii. Diaformin – Chronic drug  41.1%
 iii. Buscopan – Pain relief drug24.1%

Figure 6: Estimated percentage of each type

4. Graph showing the highest cities and years

By performing the analysis based on the highest cities and years, the results of the number of patients from the various cities kept on increasing from 2015 through to 2018. Most cities however experienced a decline in the number of patients in the year 2019.The figure below shows the distribution on the same.

Figure 6: Highest cities verses years

5. Percentage of patients who use private health insurances

By taking into consideration the number of patients who use private health insurance, only about 5.4% (3796 patients) used private health insurances. Majority of patients (about 94.6%, count = 67634) of the patients used Medicare. The table below shows the results as obtained through the analysis of the data.

PayerCountPercent
Private Medical insurance37965.4%
Medicare6763494.6%
Total71430100.0%

6. Average of diabetic patients and the binomial distribution 

The analysis about diabetic patients of both type I and type II was carried out and the results are as shown in the table below.

Patient typeCount (n)Average percentage (%)
Diabetic (Type I)71665.3%
Diabetic (Type II)61434.7%
Total1430100.0%

Figure 7: Binomial distribution

7. Cities (suburbs) that host less than 20% of diabetics 

The following cities host less than 20% of diabetic patients as obtained.

CityCount (n)Percent of diabetes
Fairfield West9615.3%
Colyton63214.7%
Kemps Creek43010.0%
Prairiewood 1237.91%
Lakemba 13.6513.65

Predictive

8. Predict the annual income for 2020

To predict annual income for 2020, a regression analysis was conducted to try and predict the annual income for 2020. The regression coefficients is presented as shown in the table below.

 CoefficientsStandard Errort StatP-valueLower 95%Upper 95%
Intercept-15910445291.37-3.512910.000444-247875-70333.4
Year81.0356222.452393.609220.00030737.029125.0422

Table 7: Regression analysis.

From the above, the regression equation is given as follows;

The annual income for 2020 would be;

Thus the annual income for 2020 is predicted to be $467,763.5.

9. Predicting the number of cold and flu

Figure 8: Predicting number of cold flu

According to the figure above, it is evident that the most ideal non-linear model that helps developers predict the future number of cold and Flu cases in each month is a 2-period moving average. The non-linear formula is given as follows (Bertsimas and Kallus, 2019)

10. Association between diseases

Hypothesis 1

In order to determine whether there is any significant association between obesity and heart diseases, the following hypothesis was tested;

Null hypothesis (H0): There is no association between heart diseases and obesity

Alternative hypothesis (HA): There is association between obesity and heart diseases

The findings suggested that a significant association existed between obesity and heart diseases.

Hypothesis 2

We also tested whether there is any significant association between diabetes and hypertension. The following hypothesis was tested;

Null hypothesis (H0): There is no association between diabetes and hypertension.

Alternative hypothesis (HA): There is association between diabetes and hypertension.

The findings suggested that a significant association exists between diabetes and hypertension.

Hypothesis 3

The last association we sought to test was whether there is significant association between hypertension and obesity. The following hypothesis was tested;

Null hypothesis (H0): There is no association between hypertension and obesity.

Alternative hypothesis (HA): There is association between hypertension and obesity.

The finding suggested that a significant association exists between hypertension and obesity.

Prescriptive

11. Managerial and procedural suggestions to improve the annual revenue.

Based on the results of the analysis, it was evident that majority of the patients came from Hinchinbrook (13.7%, 9794 patients). The second highest city where the patients came from was Hoxton Park (9.5%, 6432 patients). 8.7% (6245 patients) of the patients came from Middleton Grange. This information can help the management in terms of finding out reasons behind these cities having more cases. The reason for the high numbers could be frequent outbreaks which makes the numbers rise (Hojja, 2014).

In terms of ethnicity, the management of the hospital needs to consider hiring more nurses and physicians of the Australian origin. The results of the study showed that majority of the patients were of Australian origin.  The second highest ethnicity was the Iraqi origin that was represented by 27.4% (count = 19439) were of Iraqi origin while the third highest tribe was the Torres Strait Islander (15.7%, count = 11151). This information gives the management a clear view of how they need to do their hiring based on ethnicity for balancing purposes (Elliot et al., 2016).

Gender showed that majority of the patients were female, it would be important that the management considers employing more female nurses to take care of the female patients as this will help boost the hospital revenues (Waller, 2013).

There is need to open other branches in cities such as Hinchinbrook, Hoxton Park and Middleton Grange which had the highest number of patients. The number of patients coming from these three cities was so high that the management needs to consider opening branches in these cities (Daniel, 2017).

12. Attributes that may increase patient’s regularity

To analyse this, we first defined regularity based on the following aspects

  • Regular – having attended 5 visits or more
    • Irregular – having attended between 2 and 4 visits
    • Rare – having attended less than or equal to 1 visit

The finding suggested that some of the attributes that may increase the patient’s regularity is ethnicity. It was established that some ethnic groups are more regular than others.

13. Probability

If the average visit for {Regular} visitors was ψ per year, the probability of having (ψ – 10)/year visits is determined as follows;

Thus, the probability is approximately = 0.15

Conclusion 

In regard to decision the decision of whether to establish small pharmacy inside or open up another clinic branch in another area by a medical organisation, it was necessary to carry out various analysis on the provided data to make inferences. This report thus presents the findings and determination of the right decision.

Based on the descriptive, prescriptive and predictive analysis performed there is much evince to suggest that there is need to establish a new clinic in other areas.

 

References

Bertsimas, D., & Kallus, N. (2019). From predictive to prescriptive analytics. Management Science.

Carlos, J. E., and (2015). Descriptive statistics, graphics and quality assessment. Bioinformatics21(16), 3445-3447.

Daniel, C. E., and Using Descriptive Statistics Bartz, A. E. (2018). Basic statistical concepts. New York: Macmillan. Devore, J., and Peck.

Eckerson, W. W. (2017). Predictive analytics. Extending the Value of Your Data Warehousing Investment. TDWI Best Practices Report1, 1-36.

Elliott, J., Dennis, K., Dyer, P. R., Kesteloot, H., Liu, O. … and Zhou, B. F. (2016). INTERMAP: background, aims, design, methods, and descriptive statistics. Journal of human hypertension17(9), 591-608.

Hojja, H. (2014). Descriptive statistics for multivariate distributions. Statistics & Probability Letters1(6), 327-332.

Shmueli, G., & Koppius, O. R. (2011). Predictive analytics in information systems research. MIS quarterly, 553-572.

Song, S. K., Kim, D. J., Hwang, M., Kim, J., Jeong, D. H., Lee, S. … & Sung, W. (2013, December). Prescriptive analytics system for improving research power. In 2013 IEEE 16th International Conference on Computational Science and Engineering (pp. 1144-1145). IEEE.

Waller, M. A., & Fawcett, S. E. (2013). Data science, predictive analytics, and big data: a revolution that will transform supply chain design and management. Journal of Business Logistics34(2), 77-84.