Visual Representation of Data in Healthcare Research

Analysis of research paper and Visual representation of Data

SECTION A

References Gaskin, C. J., & Happell, B. (2013). Power of mental health nursing research: A statistical analysis of studies in the International Journal of Mental Health Nursing. International Journal Of Mental Health Nursing, 22(1), 69-75. doi:10.1111/j.1447-0349.2012.00845.x

The basis of the study was to investigate the differences in statistical power testing amongst studies published in the last two volumes of the IJMHN regarding their ability to detect small, medium and large effect sizes. A second aim was to detect the experiment wise type I error rate.  The power of these 23 studies analysed in this study was found to be 0.34, 0.79 and 0.94. Reviewing effect sizes and interpreting them is important as it will determine how significant the results of clinical studies are. It was also found that there were no reporting of adjustments made for experiment wise errors. Also only 17 % studies reported Apriori power analyses. Effect sizes were reported only for correlations and regressions but not for other tests such as c-tests, t-tests and ANOVA. The study also emphasises the importance of POWER analyses. .( Gaskin & Happell, 2013)

Identifying null hypothesis and alternative hypotheses

The null hypothesis was that there would be all statistical tests would have enough statistical power to detect small, medium and large effect sizes equally. The alternative hypothesis was that tests would differ in their ability to detect effects of different sizes and this would impact the findings of the experiment. (Gaskin & Happell, 2013)

Methods used for the study and type of data required

ANOVA can be used for both parametric (score based) data as well as non parametric data (ranking or order) based. Assumptions for ANOVA. One way ANOVA can be used to compare variables between different groups. One way repeated ANOVA can be used when you want to repeat the same test over the same group a number of times.  Two way ANOVA can be used for more complex studies when different set of groups and different effects of different variables on these groups are studied. The assumptions made for ANOVA are as follows: (“Analysis of Variance”, 2014)

1. the values for the groups should follow the normal distribution
2. the averages across different populations may be different
3. the Standard deviations across populations are same. (“Assumptions for one way Anova”, 2014)

The Dependent variable studied here was the statistical power of different tests to detect different effect sizes.

The Independent Variable here referred to the different kinds of statistical tests whose power was tested by checking the extent of effects.

The type of data used for the analysis which were the parametric score based data (mean values of power of tests) were acceptable and hence assumptions were met. (“Do your data violate”, 2014, “Anova assumptions”, 2014)

List of References:

Basic.northwestern.edu,. (2014). PROPHET StatGuide: Do your data violate one-way ANOVA assumptions?. Retrieved 25 April 2014, from http://www.basic.northwestern.edu/statguidefiles/oneway_anova_ass_viol.html

Csse.monash.edu.au,. (2014). Analysis of Variance (ANOVA). Retrieved 25 April 2014, from http://www.csse.monash.edu.au/~smarkham/resources/anova.htm

Math.colgate.edu,. (2014). The Assumptions for One-Way ANOVA. Retrieved 25 April 2014, from http://math.colgate.edu/math102/dlantz/examples/ANOVA/anovahyp.html

stat511.cwick.co.nz,. (2014). anova assumptions. Retrieved 23 April 2014, from http://stat511.cwick.co.nz/lectures/19-anova-assumptions.pdf

SECTION B

Data results are frequently plotted to give a visual representation of the relationship between data. For this question, provide an example of a relationship that you would expect to find in a health scenario [perhaps height and weight of diabetic patients]. Describe what you think the graph would look like [include whether you believe it would be linear or nonlinear] and discuss the meaning of the relationship as depicted in the graph.

Visual representations can be used to better convey the relationship between data obtained as a result of analysis. Visual representation of data involves the representation of data in the form of graphs or pictorial formats. (“Data visualization”, 2014). They help in easier and faster appreciation of the results of complex data obtained from experiments. Many research results involves the production of huge amounts of data, which when studied only in tabular form may appear too complex for many to obtain insights from them. Pictorial representations helps to understand and assimilate findings quickly, can be used to communicate findings and possible hypothesis and make predictions. This is because data when spread over charts and graphs can be better comprehended by the brain. Visualisations also help patterns to be spotted quickly even from amongst large amounts of data and data can be easily shared. (Wong, 2014)

Visual representations can be of great importance in heath research as they represent relationships between different variables under study in complex conditions in a way, which can make sense to the researcher, doctor, caregiver and patients. (“Using graphs and visual data in science”, 2014) Hence these are very important. Example of a situation would be to understand the relationship between the oxygen carrying capacity of blood and the amount of haemoglobin in the blood. This is important to health givers, given the preponderance of situations like anemia and disorders affecting the ability of the blood to transport oxygen. This relationship, if calculated and displayed in tables, would not convey much impact but when the data are plotted in the form of a graph, it would convey immediate sense to even the person with non-medical background. As for instance, in this case, initially there is an increase in the amount of oxygen absorbed by the haemoglobin in RBCs, after which following the stauraiton of haemoglobin, there should be a steady decline and plateauing. This can be better represented as a standard dissociation curve which would make better sense. This may resemble something like the graph below. In this case, the graph is considered to be sigmoidal in nature. That is an initial increase in the ability to take up oxygen followed by steadily declining ability and a point is reached at which, no matter how much the oxygen level is increased, there is no corresponding increase in the capacity of the haemoglobin. (Robertson & Guerroro, 2012)

List of References:

Roberson, R., & Bennett-Guerrero, E. (2012). Impact of red blood cell transfusion on global and regional measures of oxygenation. Mount Sinai Journal Of Medicine: A Journal Of Translational And Personalized Medicine, 79(1), 66–74.

Sas.com,. (2014). Data Visualization | SAS. Retrieved 25 April 2014, from http://www.sas.com/en_us/insights/big-data/data-visualization.html

Visionlearning,. (2014). Visionlearning.com. Retrieved 25 April 2014, from http://www.visionlearning.com/en/library/Process-of-Science/49/Using-Graphs-and-Visual-Data-in-Science/156

Wong, E. (2014). The Importance of Data Visualization. Bridgeable. Retrieved 25 April 2014, from http://bridgeable.com/the-importance-of-data-visualization/

Image from Wikimedia commons