Hospitals Performance Statistic

Introduction 

Rationale for Analysis 

The reason for conducting the analysis was to investigate the level of performance of both non-profit and proprietary hospitals in the United States. Precisely, the analysis aims at determining whether hospitals in America have differences in the level of performance. The findings of the study will assist my organization in such a way that my affiliation will be able to use the output to implement possible financial, operational, or clinical improvements. 

The rate of Clostridium difficile infection is a key determinant of the level of performance of a give health care center. Miller, Polgreen, Cavanaugh & Polgreen (2016) uses the inpatient length of stay to justify hospitals quality. The study concluded hospitals with a high rate of Clostridium difficile infection are of low quality as compared those with few cases of Clostridium difficile infection. Findings from Miller et al. (2016) show that regions with high with a high number of proprietary hospitals depict a better performance as compared to regions with a high number of not-for-profit hospitals. 

I t is worth noting that hospital ownership has a significant impact towards the performance of the hospital. Birmingham & Oglesby (2018) brings up an important argument in their research; they argue that not-for-profit hospitals are performing better than the proprietary hospitals. Birmingham & Oglesby (2018) found out that the readmission rates for not-for-profit hospitals were higher since stakeholders under this type of hospitals are profit-based. Thus, increasing the number of readmissions would ensure profitability to the stakeholders. Downing et al. (2017) argue that regions with a high number of small hospitals depict a poor performance in the provision of their services, as they are associated with a high number of readmissions. These remarks provide the need for analyzing to determine whether regions in the United States show differences in the level of performance. 

Research Question 

Is there a difference in the level of performance of different hospitals in the United States? 

Research Hypothesis 

H o There is no difference in the level of hospitals performance in different regions of the United States. 

H 1 There is a difference in the level of performance in different hospitals in the United States. 

Methodology 

The dataset used for the analysis is the rate for all hospitals infection in the United States that are submitted to Medicare. The MySQL statistical tool was used for extracting the data needed for the analysis. The MySQL script used for extracting the data was: 

SELECT 

hospital_general_information.hospital_name, 

hospital_general_information.state, 

readmission_reduction.measure_name, 

readmission_reduction.number_of_discharges, 

readmission_reduction.number_of_readmissions, 

readmission_reduction.excess_readmission_ratio, 

regions.region_name 

FROM 

hospital_general_information 

JOIN 

readmission_reduction 

ON 

hospital_general_information.provider_id=readmission_reduction.provider_id 

JOIN 

regions 

ON 

hospital_general_information.state_code=regions.state_code 

WHERE 

readmission_reduction.number_of_discharges>0 

AND 

readmission_reduction.number_of_readmissions>0 

AND 

readmission_reduction.measure_name='READM-30-AMI-HRRP'; 

Definition of Data Fields 

hospital_general_information.hospital_name. This field contains the names of hospitals used for the study. 

hospital_general_information.state. This field entails the state in which the hospital originates. 

readmission_reduction.measure_name. This field summarizes the measures taken by hospitals to increase performance. 

readmission_reduction.number_of_discharges. This field contains the number of admissions that each hospital receives. 

readmission_reduction.number_of_readmissions. This field entails the number of readmissions that each hospital receives. 

readmission_reduction.excess_readmission_ratio. This field summarizes the ratio of readmissions in every hospital. 

regions.region_name. This field contains the regions under study. 

Data Preparation Method 

The data was prepared using RStudio to obtain the relevant information for conducting the analysis. Using R for data preparation was essential due to the benefits that the program presents in analyzing research data. First, using R was vital because it allows the researcher to identify a mistake quickly. Considering that the data set used for the study was large, having an R-script for data preparation would allow quick identification and debug of an error. Besides, the ability of R to create separate data sets without manipulating them was beneficial to the study since it reduced the time taken for preparing the data. Moreover, R was the required method for data preparation since the investigation involved thousands of hospitals in the United States; thus, the program enabled preparation of data without necessarily repeating the same procedure repeatedly. 

Data Summary 

After data preparation, a table was created, which summarized the data in accordance with every category listed under the column hospital ownership. Another summary table was also created that presented the count of each category listed under the column compared to national. Besides, a new data frame was constructed and limited to display data from hospitals that have ownership of either non-profit or proprietary. Finally, a cross-tabulation was then created that summarized the frequency of each category listed under the column hospital ownership by each type listed under compared to national. 

The rationale for Statistical Method 

ANOVA statistical technique was used to analyze the data for the study. The use of ANOVA is essential in providing an output that has a strong statistical power. Besides, the method reduces the number of random variabilities ( Gorecki & Smaga, 2015) . As a result, the analysis is exposed to minimal chances of obtaining an error in the results section. The method is relevant to the study since it provides an ideal estimation that will allow correct analysis of the results. 

ANOVA statistical method will be used to obtain the p-value and level of significance that will be used to answer the research question. The F-value will be used to estimate the validity of the null hypothesis. Besides, if the F-value will be less than the level of significance, the null hypothesis will be rejected. The limitation for the data used for the analysis was that the data was that the data used was large; hence, being complex for conducting an analysis. 

Results 

Df Sum Sq Mean Sq F value Pr(>F) 

region_name 3 0.413 0.13778 48.85 <2e-16 *** 

Residuals 1755 4.950 0.00282 

Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1 

diff lwr upr p adj 

Northeast Region-Midwest Region 0.036923272 0.027135380 0.0467111631 0.0000000 

South Region-Midwest Region 0.012372386 0.003982091 0.0207626822 0.0008886 

West Region-Midwest Region -0.009488795 -0.019825116 0.0008475269 0.0852001 

South Region-Northeast Region -0.024550885 -0.033509531 -0.0155922398 0.0000000 

West Region-Northeast Region -0.046412066 -0.057214835 -0.0356092978 0.0000000 

West Region-South Region -0.021861181 -0.031415973 -0.0123063890 0.0000000 

Boxplot for Hospitals Performance in the United States 

From the ANOVA analysis, the level of significance is less than 0.5; thus, there is a statistical significance between the performances of hospitals in the United States. Besides, the F distribution is 2e-16, which follows an F(3, 1755). The low F value indicates that the null hypothesis for the study cannot be rejected; thus, it is considered to be valid. From the boxplot, it is visible that the means for different regions in the United States have no significant difference from each other. In that case, it is concluded that their no big disparity between the level of performance of hospitals in the US despite their locality. 

Conclusion 

Conclusively, the research aimed at investigating whether non-profit hospitals in the United States regions presented a low performance as compared to the proprietary hospitals in the United States. The results obtained from the analysis answered the research question in that it presented the mean difference in performance in hospitals from different regions of the United States. The hypothesis for the study concerned the disparities in hospital performance from different areas of the United States. In that case, it is worth pointing out that the results are related to the hypothesis for the study since they test the level of performance of hospitals in the United States. 

The main limitation of the study is that only six regions were used for the research despite having many regions in the United States. Thus, the population under study was not adequate to derive an ideal conclusion for the whole country. Further studies should try to incorporate more regions to draw a better outcome that represents the whole of America. 

The findings from the investigation are essential in formulating measures that will see better performance of hospitals in the United States. Regions with a high level of readmissions will have to strategize ways in which they can use to mitigate the situation of experiencing poor performance. Thus, the affected hospitals will use the findings to make financial, operational or clinical improvements to reduce the level of readmissions; thus, showing better performance. 

References 

Birmingham, L. E., & Oglesby, W. H. (2018). Readmission rates in not-for-profit vs. proprietary hospitals before and after the hospital readmission reduction program implementation. BMC health services research , 18 (1), 31. 

Downing, N. S., Cloninger, A., Venkatesh, A. K., Hsieh, A., Drye, E. E., Coifman, R. R., & Krumholz, H. M. (2017). Describing the performance of U.S. hospitals by applying big data analytics. PLoS ONE , 12 (6), e0179603. http://doi.org/10.1371/journal.pone.0179603 

Gorecki, T., & Smaga, Ł. (2015). A comparison of tests for the one-way ANOVA problem for functional data. Computational Statistics , 30 (4), 987-1010. 

Miller, A. C., Polgreen, L. A., Cavanaugh, J. E., & Polgreen, P. M. (2016). Hospital Clostridium difficile Infection Rates and Prediction of Length of Stay in Patients Without C. difficile Infection. Infection Control and Hospital Epidemiology , 37 (4), 404–410. http://doi.org/10.1017/ice.2015.340 

Appendix 

MySQL Script 

SELECT 

hospital_general_information.hospital_name, 

hospital_general_information.state, 

readmission_reduction.measure_name, 

readmission_reduction.number_of_discharges, 

readmission_reduction.number_of_readmissions, 

readmission_reduction.excess_readmission_ratio, 

regions.region_name 

FROM 

hospital_general_information 

JOIN 

readmission_reduction 

ON 

hospital_general_information.provider_id=readmission_reduction.provider_id 

JOIN 

regions 

ON 

hospital_general_information.state_code=regions.state_code 

WHERE 

readmission_reduction.number_of_discharges>0 

AND 

readmission_reduction.number_of_readmissions>0 

AND 

readmission_reduction.measure_name='READM-30-AMI-HRRP'; 

R Script 

attach(AMI) 

table(region_name) 

AMIrate<-number_of_readmissions/number_of_discharges 

mean(AMIrate) 

tapply(AMIrate,region_name,mean) 

Myanova<-aov(AMIrate~region_name) 

summary(Myanova) 

TukeyHSD(Myanova) 

boxplot(AMIrate~region_name)