Non-Parametric or T-tests
Assessment of Non-Parametric
This author will provide an assessment of non-parametric analog. Non-parametric methods don’t demand the researcher to formulate distributional assumptions concerning data. Secondly, nonparametric data uses qualitative methods as opposed to quantitative methods. These are the identified and defined procedures of statistics, which are used to test the hypothesis. They, therefore, don’tneed a normal distribution since they are ranks and count-based (Plichta& Kelvin, 2013). The two non-parametric tests used in the article are the Pearson’s Chi squared test for the nominal/ordinal data and the non-parametric Mann-Whitney’s U-test for continuous variables. Unlike the parametric t-test, non-parametric test Mann-Whitney’s U-test makes no assumptions about the distribution of the data (for example, normality). The Mann-Whitney, U-test can be alternated to replace the Independent Group’s T-Test, where the assumptionof normalityor the variance’s equality has not been achieved. Descriptive statistics were used to summarize the data in this article. Using the Mann-Whitney U- test analysis I will, analyze the continuous variables to see if the assumption were met. The aim of a Chi-squared test will examine the relationship between the variables. In this article, the author will determine if the new pain protocol did or did not have an effect on the time to pain assessment or the time to analgesia.
The Study
This article focused on management of pain in the emergency departments (EDs). A study was completed to test the nurse initiated pain protocol’s effect intervention. In the hypothesis of the study; the nurse initiated pain protocol intervention had no effect on either the time or pain assessment of the time to analgesia.
Pain is not only one of the most frequently used nursing diagnoses, but also evident in every aspect of nursing. Nurses should know and record the language that is used to indicate pain and behaviors describing the same. At the same time, they need to know the behavior that is used by the patients in order to communicate and cope with the pain that comes with it (Breen, 2002, p. 57). Treating pain in the emergency department remains a problem; the study compares before and after in an ED in order to carry out a test on the effect of this nurse initiated pain protocol (NIPP) strategy. In order to carry out this experiment, 889 adults patients; who included 144 peoplein one group (control group), while another group of 745 from other groups (intervention) was selected (Finn, Rae, Gibson, Swift, Watters & Jacobs, 2012). Nurse initiated analgesia is understood as the introduction of analgesia to patients by nursing staff, by use of a predefined protocol, before the patient is attended to, by a medical specialist or officer (Kelly et al., 2005, p. 151).
One of the essential outcomes from this research article is the fact that even with the nurse analyzing the pain; ED nurses fail to consistently administer analgesic drugs up to the time when the patient is analyzed by a medical doctor, or officer. As determined by the ED’s workload, a considerable delay might take place, between the patients’ presentation at the ED a doctor attending to the patient. At the same time, it might take longer before pain relief is finally administered (Finn, Rae, Gibson, Swift, Watters & Jacobs, 2012). The study was done in the ED of an urban teaching hospital in Australia. The ED is medically staffed by a mix of consultant emergency physicians, senior registrars, and resident medical officers and intern. The ED nurses are predominately registered nurses, the triage shift is covered by senior registered nurses.
This research, carried out and conducted in the University of Western Australia, has brought a breakthrough towards understanding nurse-initiated analgesia to patients, and the impacts it has on the amount of pain that the patients experience. This research experiment, conducted through The Before and After Study, has shown the necessity of the initiation of analgesia, since its impacts have been seen through the outcome of the amount of pain that the patients experience, before and after the initiation. While treating pain as an emergency has still remained to be a problem, nurse-initiated analgesia provides an alternative, whereby the nurses can contain the patients’ pain, even before the patient is attended to by the medical officers.
Medical officers have, consequently, advocated for this administration by the ED nurses, based on a number of reasons. Firstly, it is necessary to note the fact that in some cases, medical officers might take time before attending to patients. In such a case, analgesia will assist the patients, by reducing the pain (Finn, Rae, Gibson, Swift, Watters & Jacobs, 2012). However, nurses have been advised and encouraged to adopt the non-pharmacological means to relief pain, and these methods include cold packs, support or splinting.
Education
Education was also necessary before the nurses had the full mandate to administer analgesia, whereby they were educated concerning the NIPP, as this was necessary knowledge that they needed, prior to the administration of the same. This education consisted of hard copies summarizing the protocol guidelines, as well as the complete protocol details. All this information was to be filed in the nursing protocol guidelines, at the clinical nurses’ educator’s office. At the same time, this information was provided in various areas, such as the ED offices and the transit area and triage desk.
Interventions
With the help of the clinical, medical, pharmacy and the nursing staff, the ED NIPP was developed, whereby a chart was drawn, indicating the protocol medication, which included the drugs to be administered to the patients as well as the specific dosage. At the same time, this flow chart included the indications, potential adverse effects that might come with using the drugs, as well as contraindications. However, this flowchart required medical approval, prior to its administration.
Outcomes
After this research study was carried out and experimented upon, there are several outcomes that they research team recorded, and these outcomes had various impacts on the medical field, especially in the administration of the ED NIPP. For instance, the team recorded the time the first pain was experienced, compared to the first administration of analgesia. This was called the time interval difference, and was used to analyze the difference between analgesia’s administration and induction, and the absence of the same.
Data
The data used, in this case, was also recorded, as well as the data that the research and experiment team collected. The team acquired the data that they used from a clinical record, and had age, gender, ED presentation time and triage category as the demographics. At the same time, the clinical records analyzed self-administration of analgesia that took place, prior to the presentation of ED. The ED diagnosis was manually categorized.
Another essential aspect to note is the fact that prior to the NIPP implementation, the sample size was calculated, and summed up to 144, represented as N, with a standard deviation of 56.3 minutes, and a 50% relative difference. However, considering the fact that the intention of the research team was to run for 3 months but ended up taking longer, this means that the number of cases might have gone higher than 144, which was the initial number of cases.
Limitations of the Study
The results of the study cannot be used to predict or generalize other EDS characterized by different staffing profiles as the study used a single center based on a sample of patients who were non consecutive and convenient. The patients with a pain score of zero were incorrectly entered into the study. The patients (11%) recorded with no initial pain score made it not possible to conclude with certainty whether this constituted to pain not recorded or not assessed. The data used for comparison between the groups brought up some differences. This is because part of the data used for comparison purposes was collected a year before the intervention for audit purposes. The data should have been treated as pre-intervention (historical cohort) for the process of auditing only.
The treatment of this data in the comparison gave rise to confounding the results risk. There was a small number of about three point nine percent of the patients with zero pain score. This was against the inclusion criteria of the study which was to include those patients who represented themselves to ED with pain. Their reason for giving zero pain score could still be associated with some pain for example no pain at rest but some pain on movement. The fact that the study was single centered a case of mix cannot be predicted using the results of this study but would require a multi centered study.
Statistical Analysis
Comparisons between the data collected prior to the study and those collected after the study were compared using two statistical tools in this study: Chi square for nominal and ordinal data, and Mann Whitney test for continuous data. The use of these two tools is, however, limiting since they both deal with comparison of categorical data. The researcher should have opted for a non-categorical tool such as r-test for variables such as age and initial pain score (Stephens, 2004). The use of multivariate regressions was a commendable decision so as to establish the contribution of each variable to the effect being measured. Odds Ratio was also used in this study to determine the likelihood of occurrence of various scenarios (Hatcher, 2013). It is imperative to highlight that the researcher did not conduct precision tests or tests for accuracy. The researcher should have used the q-test to clean the data so as to ensure that the large values are outliers at the stated confidence level.
F-test
This tool is used to test for precision of the study conducted by comparing the standard deviations.
Ho : There is no significant difference between the standard deviation before and after the groups were made at a confidence level of 95%.
Condition: If Fcal is less than Ftab, we accept the Ho
Fcal = SD12/ SD22
NB: The bigger SD always becomes the numerator, hence:
Fcal = 202/ 19.72 = 400/ 388.09
= 1.0307
Ftab at (n1-1), (n2-1)
= (144-1), (144-1)
= 143, 143
Ftab (143,143) = 1.34
Since, Fcal, 1.0307, is less than Ftab, 1.34, we accept the null hypothesis that there is no significant difference between the standard deviation before and after the groups were made at a confidence level of 95%.
Ho : There is no significant difference between the standard deviation before and after the groups were made at a confidence level of 95%.
Condition: If Fcal is less than Ftab, we accept the Ho
Fcal = SD12/ SD22
NB: The bigger SD always becomes the numerator, hence:
Fcal = 50.52/ 39.62 = 2550.25/ 1568.16
Fcal = 1.6263
Ftab at (n1-1), (n2-1)
= (53-1), (443-1)
= 52, 442
Ftab (52, 442) = 1.38
Since, Fcal, 1.6263, is more than Ftab, 1.38, we reject the null hypothesis that there is no significant difference between the standard deviation before and after the groups were made at a confidence level of 95%.
Since there is a significant difference between the two studies, the following formula shall be used to determine whether there is a significant difference between the means of the studies.
Ho There is no significant difference between the means of the period before and after the groups were formed at a confidence level of 95%.
Condition: If Tcal is less than Ttab we accept the Ho
Tcal= (x1-x2)/ √[s12/n1 + s22/ n2]
= ( 55.6 – 20.9 ) / √[50.52 /53 + 39.62 / 443 ]
= 34.7/ √[2550.25/53 + 1568.16/ 443]
= 34.7/ √[48.1179 + 3.5399]
= 34.7/ √51.6578
= 34.7/ 7.1873
Tcal = 4.828
df= √[ (s12/n1 + s22/ n2 )/ { s12/n1/ n1-1} + { s22/ n2/ n2-1}
df= √[51.6578 / 48.1179/ 52 + 3.5399 / 442 ]
df= √[51.6578/ 0.9253 + 0.008 ]
df = √[ 51.6578/ 0.9333]
df= √55.3496
df= 7.44
(Round off to the nearest who number)
df= 7
Ttab at 7 = 2.365
Since Tcal, 4.828, is more than Ttab, 2.365, we reject the hull hypothesis.
Summary
The f-test results from the two tables (Table 1 and 2) indicate that the researcher did not consistently maintain precision throughout the study. This is because the f-test results from the second table indicate that the standard deviations are not precise whereas the results from the first table indicate precision in the data collection. It is, however, essential to highlight that lack of precision in the two tables does not automatically result in lack of accuracy. This is because accuracy involves the closeness of a value to the true value whereas precision indicates how the variables agree amongst themselves.
The lack of precision may be attributed to the fact that there was a one year break between the study before and the study after the groups were formed. This prolonged period may have resulted in other variables not studied in this research interfering with the sample population. This is critical to the study because the sample constituted of human beings who are affected significantly by both physical and emotional variables resulting in an alteration in body metabolism. It is proposed that if the researcher was to conduct another study as a sequel to this one, that they should reduce the period of time between the first study and the second one to maintain precision. As previously stated, the significant difference between the standard deviations of the two studies does not mean that the data is inaccurate; however, the agreeability of the variables is what is at question (Urdan, 2010). The lack in precision in the second table is noted to have affected the means of the data from the t-test calculated on the two means.
References
Arber, A. (2004). ‘Is pain what the patient says it is? Interpreting an account of pain.’International Journal of Palliative Nursing, 10(10), 491-494. Retrieved from http://libsys.uah.edu.kaplan on March, 31, 2013
Breen, J. (2002). ‘Transitions in the Concept of Chronic Pain.’Advanced Nursing Science, 24(4), 48-59. Retrieved from http://libsys.uah.edu.kaplan on March 31, 2013.
Finn, J., Rae, A., Gibson, N., Swift, R., Watters, T., & Jacobs, I. (2012).“Reducing time to analgesia in the emergency department using a nurse-initiated pain protocol: A before-and-after study.”Contemporary Nurse, 43(1), 29-37.
Hatcher, L. (2013). Advanced statistics in research. Chicago: Shadow Finch Media LLC
Stephens, L. (2004). Advanced statistics demystified. New York: Routledge
Urdan, T. (2010). Statistics in Plain English, 3rd Edition. New York: Routledge
