Statistical Analysis Of Medical Data Essay.

The variable ‘anticlot’ indicates the type of anti-clotting drug prescribed to the patients.Produce the relevant graph and table to summarise the ‘anticlot’ variable and write a report describing this variable in the style presented in the course materials.

The age of patients in years is provided in the data file. Produce the relevant graph and tables to summarise the ‘age’ variable and write a report describing this variable in the style presented in the course materials.

Current research indicates that the number of everyday life stressors has risen in the last decade. Hospital records for 2014 show that 74% of Australians admitted to hospital for suspected myocardial infarction had normal blood pressure at the time of admission. The researchers have suggested, that due to increasing stress, the percentage of Australians admitted to hospital for suspected myocardial infarction who have normal blood pressure at the time of admission is now lower than this.

Conduct a Binomial Test using the ‘bp ’ variable to test the researchers’ hypothesis. Produce the relevant output and write a Binomial test report based on your output in the style presented in the course materials.

Due to federal cuts to the Government Health System, the researchers predicted that the average length of stay in hospital for patients admitted for suspected myocardial infarction is now lower than the average of 5.5 days recorded in 2014.

Conduct a One-Sample t-test using the ‘los’ variable to test this prediction. Produce the relevant output and write a One-sample t-test report based on your output in the style presented in the course materials.

The researchers have also claimed that the average length of stay in hospital for patients who have a history of diabetes is higher than the average for those who do not. Conduct an Independent samples t-test using the ‘los’ and ‘diabetes’ variables to investigate this claim. Produce the relevant output and write an Independent samples t-test report based on your output in the style presented in the course materials.

Question 1: Summary of Anti-clotting Drugs Prescribed

The variable ‘anticlot’ indicates the type of anti-clotting drug prescribed to the patients.

Produce the relevant graph and table to summarise the ‘anticlot’ variable and write a report describing this variable in the style presented in the course materials.

In this question, the study sought to understand the most frequent type of anti-clotting drug prescribed to the patients. Table 1 below gives the frequency distribution of the various types of anti-clotting drugs prescribed to the patients. It can be seen that the most prescribed drug is Aspirin (42%, n = 844) while the least prescribed drug is Warfarin (7%, n =148).  22% (n = 442) of the patients did not receive any drug prescription while 27% (n = 544) of the patients were prescribed for Heparin.

		Frequency	Percent	Valid Percent	Cumulative Percent
Valid	Aspirin	844	42.2	42.7	42.7
	Heparin	544	27.2	27.5	70.2
	Warfarin	148	7.4	7.5	77.7
	None	442	22.1	22.3	100.0
	Total	1978	98.9	100.0
Missing	System	22	1.1
Total		2000	100.0

The above results can also be visualized in the bar chart presented in figure 1 below;

Figure 1: Bar chart of anti-clotting drugs prescribed to the patients

The age of patients in years is provided in the data file.

Produce the relevant graph and tables to summarise the ‘age’ variable and write a report describing this variable in the style presented in the course materials.

The average age of the patients in years was found to be 63.45 years with oldest patient being 95 years old while the youngest patient was 45 years old.

		Age in years
N	Statistic	1995
Range	Statistic	50
Minimum	Statistic	45
Maximum	Statistic	95
Mean	Statistic	63.45
Std. Deviation	Statistic	8.216
Variance	Statistic	67.507
Skewness	Statistic	.243
	Std. Error	.055
Kurtosis	Statistic	-.401
	Std. Error	.110

The normality test for the variable “Age in years” showed that the variable does not come from a normally distributed dataset (p < 0.001).

Tests of Normality
	Kolmogorov-Smirnova			Shapiro-Wilk
	Statistic	df	Sig.	Statistic	df	Sig.
Age in years	.055	1995	.000	.989	1995	.000
a. Lilliefors Significance Correction

A boxplot was drawn to see the distribution of the dataset. A number of outliers were observed in the dataset further suggesting that the dataset is not symmetrical.

Figure 2: Boxplot for age in years

However, the histogram presented below suggests that the dataset seems to be normally distributed since the graph is almost symmetrical and in the shape of a bell-shaped curve.

Figure 3: Histogram for age in years

Current research indicates that the number of everyday life stressors has risen in the last decade. Hospital records for 2014 show that 74% of Australians admitted to hospital for suspected myocardial infarction had normal blood pressure at the time of admission. The researchers have suggested, that due to increasing stress, the percentage of Australians admitted to hospital for suspected myocardial infarction who have normal blood pressure at the time of admission is now lower than this.

Conduct a Binomial Test using the ‘bp ’ variable to test the researchers’ hypothesis. Produce the relevant output and write a Binomial test report based on your output in the style presented in the course materials.

In this question, we conducted a Binomial test using the ‘bp’ variable to test the researchers’ hypothesis that 74% of Australians admitted to hospital for suspected myocardial infarction had normal blood pressure at the time of admission (Howell, 2007). The hypothesis set to be tested was;

Descriptive Statistics
	N	Mean	Std. Deviation	Minimum	Maximum
Blood pressure	1996	.79	.411	0	1

Binomial Test
		Category	N	Observed Prop.	Test Prop.	Exact Sig. (1-tailed)
Blood pressure	Group 1	Normal	1568	.79	.74	.000
	Group 2	Non-normal	428	.21
	Total		1996	1.00

From table 5 above, we can clearly see that the proportion of Australians admitted to hospital for suspected myocardial infarction had normal blood pressure at the time of admission is 79% for the sample data provided. Binomial test that was conducted revealed that there is significant evidence that the proportion of Australians admitted to hospital for suspected myocardial infarction had normal blood pressure at the time of admission is greater than 74%. The claim by the researchers that the proportion is lower than 74% is therefore not significantly valid.

Due to federal cuts to the Government Health System, the researchers predicted that the average length of stay in hospital for patients admitted for suspected myocardial infarction is now lower than the average of 5.5 days recorded in 2014.

Conduct a One-Sample t-test using the ‘los’ variable to test this prediction.

Produce the relevant output and write a One-sample t-test report based on your output in the style presented in the course materials.

The aim of this question was to test the claim that the average length of stay in hospital for patients admitted for suspected myocardial infarction is now lower than the average of 5.5 days recorded in 2014.

Using one-sample t-test, we tested the above claim at 5% level of significance (John , 2006).

Table 6: One-Sample Statistics
	N	Mean	Std. Deviation	Std. Error Mean
Length of stay	1995	5.54	1.439	.032

Table 7: One-Sample Test
	Test Value = 5.5
	t	df	Sig. (2-tailed)	Mean Difference	95% Confidence Interval of the Difference
					Lower	Upper
Length of stay	1.112	1994	.266	.036	-.03	.10

The mean length of stay was found to be 5.54. Results from the one-sample t-test revealed that there is no evidence predicted that the average length of stay in hospital for patients admitted for suspected myocardial infarction is now lower than the average of 5.5 days recorded in 2014 (p > 0.05).

The researchers have also claimed that the average length of stay in hospital for patients who have a history of diabetes is higher than the average for those who do not.

Conduct an Independent samples t-test using the ‘los’ and ‘diabetes’ variables to investigate this claim. Produce the relevant output and write an Independent samples t-test report based on your output in the style presented in the course materials.

In this question we tested the claim that the average length of stay in hospital for patients who have a history of diabetes is higher than the average for those who do not. The hypothesis is;

Where  is the average length of stay in hospital for patients who have a history of diabetes while  is the average length of stay in hospital for patients who do not have a history of diabetes.

To test this claim, independent t-test was conducted at 5% level of significance. Results are given the tables below.

Group Statistics
	History of diabetes	N	Mean	Std. Deviation	Std. Error Mean
Length of stay	No	1780	5.52	1.374	.033
	Yes	214	5.68	1.899	.130

Table 9: Independent Samples Test
		Levene's Test for Equality of Variances		t-test for Equality of Means
		F	Sig.	t	df	Sig. (2-tailed)	Mean Difference	Std. Error Difference	95% Confidence Interval of the Difference
									Lower	Upper
Length of stay	Equal variances assumed	32.045	.000	-1.572	1992	.116	-.164	.104	-.368	.040
	Equal variances not assumed			-1.223	240.51	.223	-.164	.134	-.427	.100

An independent samples t-test was performed to compare the average length of stay in hospital for patients who have a history of diabetes is higher than the average for those who do not (Zimmerman, 2007). There was no significant difference in the length of stay for the two groups of patients. The length of stay for those with history of diabetes (M = 5.68, SD = 1.90, N = 214) was not significantly different to those with no history of diabetes (M = 5.52, SD = 1.37, N = 1780), t(1992) = -1.57, p > .001, one-tailed.

References

Howell, D. C. (2007). Statistical methods for psychology.

John , R. A. (2006). Mathematical Statistics and Data Analysis.

Zimmerman, D. W. (2007). A Note on Interpretation of the Paired-Samples t Test. Journal of Educational and Behavioral Statistics, 22(3), 349–360.