PACKAGE "ACCURACY"

     Introduction: 

 ✔ Diagnostic tests play a very crucial role in contemporary medical practice as it allows the medical practitioner to rule out the disease and diagnose disease in patients who have an illness. Clinicians need insight into the possibility that an individual has a condition to arrive at therapeutic choices and give personalized treatment. This requires integrating knowledge of preliminary chance with diagnostic tests.

 ✔In order to decide on possibilities for therapy, medical diagnostics are frequently used for medical systems; moreover, numerous of these strategies need to be more accurate. Documentation must be utilized to inform the use of evaluations for diagnosis in healthcare situations. But regrettably, several clinicians conduct diagnostic tests without incorporating the supporting data.

 ✔ Sensitivity and specificity are very crucial indicators for the accuracy of the diagnostic test and to know the validity of the diagnostic test. In addition to these indicators, other indicators like Positive Predictive Value (PPV), Negative Predictive Value(NPV), Likelihood ratios(LRs), Precision and Recall will assure about the outcome of the diagnostic tests which will guide the clinician to choose the type of diagnostic test needed according to the patient.

✔Package "Accuracy" 

It includes code of functions Test_Ability(), lr_plus(), lr_minus(), youden(), precision(), recall() and pie_chart(). Data includes a few diagnostic tests with true positives, true negatives, false positives and false negatives. The description gives basic information about the package and author. It also includes a namespace generated by roxygen2. 

✔Test_Ability() function enables the medical practitioner to find the Sensitivity, Specificity, Positive Predictive Value(PPV) and Negative Predictive Value(NPV) of a particular diagnostic test upon input of the values(True Positives, False Positives, True Negatives and False Negatives).

Sensitivity: It is the capacity of the diagnostic test to identify true positives or a percentage of the population with a disease condition who tested positive.

Sensitivity= True Positives/(True positives+False Negatives)

Specificity:  It refers to the capacity of the diagnostic test to identify genuine negatives or the percentage of individuals who are free of diseased conditions.

Specificity= True Negatives/(True Negatives+False Positives)

Positive Predictive Value:

It determines the number of True positives among all outcomes which resulted in positive.

PPV=True Positives/ True Positive +False Positives

Negative Predictive Value:

It determines the number of true negatives among all outcomes which resulted in negative.

NPV=True Negatives/(True Negatives+False Negatives)

✔Likelihood Ratio (LR): The likelihood Ratio shows how valuable the diagnostic procedure is for improving the chances of getting the correct diagnosis. It is the proportion of the likelihood that test results will be positive in patients with the condition compared to patients without the illness.

lr_plus()function computes the positive likelihood ratio of the diagnostic test when values of sensitivity and Specificity are given as input.

Positive LR(LR+):  It is the possibility of receiving a positive test outcome relative to patients who do not have the condition in question.

LR+  = Sensitivity/ (1-Specificity)

lr_minus()function validates the negative likelihood ratio of the diagnostic test when values of sensitivity and Specificity are given as input.

Negative LR(LR-): It is the proportion between patients with an illness and individuals lacking the condition in terms of the likelihood of receiving a negative outcome from the diagnostic test.

LR-  = (1-Sensitivity)/Specificity

✔youden() function gives the Youden's Index when sensitivity and Specificity are given as input.

Youden’s Index (YI): It is obtained by subtracting 1 out of the diagnostic procedure's total range of sensitivity to specificity, which is presented as a fraction rather than a percentage.

YI =(Sensitivity+Specificity) -1

✔precision()permits to find of accurate forecasts along with recall

Precision: It estimates how many accurate forecasts were made with diagnostic tests.

Precision= True Positives/(True Positives+False Positives)

✔recall() function when implemented gives a positive outcome proportion compared to all possible positive cases

Recall: It calculates the proportion of successful positive forecasts provided compared to all possible positive forecasts.

Recall= True Positives/ (True Positives+ False Negatives)

Both Precision and Recall metrics can be used for the Precision-Recall curve which is mainly focused on the positive outcome of the diagnostic test.

✔pie_chart()allows finding which is the maximum of the input values which constitute a larger portion of all values and permits to come to a conclusion about the diagnostic test based on the requirements.

URL to git repo: https://github.com/VedaVangala/Test_Accuracy

  References:

Wickham, H. (2015). R Packages. (Chapter 7)

Matloff. N. (2011). The Artof R programming. (Chapter 13)

The R repository https://cran.r-project.org/web/packages

Dhamnetiya, D., Jha, R. P., Shalini, S., & Bhattacharyya, K. (2022). How to Analyze the Diagnostic Performance of a New Test? Explained with Illustrations. Journal of Laboratory Physicians, 14(1), 90-98. https://doi.org/10.1055/s-0041-1734019

Shreffler J, Huecker MR. Diagnostic Testing Accuracy: Sensitivity, Specificity, Predictive Values and Likelihood Ratios. [Updated 2023 Mar 6]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2023 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK557491/