Tag Archives: research

EMS Research: Using t Tests

When considering the emergency medical services, there has been much discussion regarding the utility of advanced life support and its effectiveness within the emergency medical services (Stiell et al., 2005; Stiell et al., 2003; Stiell et al., 2002; Stiell et al., 1999). One of the most basic skills that paramedics use exclusively is intravenous cannulation and the subsequent delivery of isotonic intravenous fluid. Intravenous cannulation is one of the first advanced skills that paramedics utilize within the course of treatment as it allows to correct for shock, provides a means for administering parenteral medications, and provides a means for drawing blood for testing either in the field or upon arrival at the receiving emergency department. As the body’s stress increases when dehydration is present, it is imperative to correct dehydration during the course of treating most ailments; otherwise, the body’s own compensatory mechanisms can fail despite otherwise adequate treatment (Wakefield, Mentes, Holman, & Culp, 2008). Additionally, dehydration can mask some critical tests, such as other blood values and radiological findings (Hash, Stephens, Laurens, & Vogel, 2000).

Though the research is limited, it is also important to note that judicious use, or overuse, of intravenous fluids can be detrimental in some cases (Rotstein et al., 2008). In order to test the effectiveness of paramedic treatment of co-morbid dehydration, we can observe for fluid status before and after treatment as well as between those patients transported by paramedic ambulance as compared to patients who present to the emergency department by other means (e.g. basic life support ambulance, walk-in); however, it is first important to understand if those patients who present to the emergency department are, indeed, dehydrated.

In order to study if paramedics have an impact in treating co-morbid dehydration, there has to be an assumption that a) most people are not dehydrated and b) people who present to the emergency department (the dependent variable) are more dehydrated (independent variable) than most of the population. As we can never be sure of the hydration status of the entire population at any given time or the standard deviation of the entire population, we can use the normal mean blood urea nitrogen value of 10 mmol/L and assume a normal distribution (Hash et al., 2000).

H0:μ=10: Patients who present to the emergency department are not dehydrated (BUN = 10 mmol/L)
Ha:μ>10: Patients who present to the emergency department are dehydrated (BUN > 10 mmol/L)

Once the random sample of BUN values have been obtained, I can use the t-distribution to find the value of the t-test statistic:

t = (x̄ - μ) / (s / √n)

Next, I would compute the degrees of freedom (it is important to note that the sample size [n] must be greater than 30 as the standard deviation of the population is not known):

DOF = n - 1

As this test is one-tailed (specifically, right-tailed), and I am concerned with a 95% CI, I would compare the t-value with the t-table row indicated by the DOF. If the t-value is greater than the t-value corresponding with the DOF, then I will be able to reject the null hypothesis; otherwise, if the computed t-value is less than the table value, I will not be able to reject the null hypothesis.

References

Hash, R. B., Stephens, J. L., Laurens, M. B., & Vogel, R. L. (2000). The relationship between volume status, hydration, and radiographic findings in the diagnosis of community-acquired pneumonia. Journal of Family Practice, 49(9), 833-837.

Rotstein, C., Evans, G., Born, A., Grossman, R., Light, R. B., Magder, S., … & Zhanel, G. G. (2008). Clinical practice guidelines for hospital-acquired pneumonia and ventilator-associated pneumonia in adults. Canadian Journal of Infectious Diseases & Medical Microbiology, 19(1), 19–53.

Stiell, I. G., Nesbitt, L., Pickett, W., Brisson, D., Banek, J., Field, B, … & Wells, G., for the OPALS Study Group. (2005). OPALS Major Trauma Study: impact of advanced life support on survival and morbidity. Academy of Emergency Medicine, 12(5), 7.

Stiell, I. G., Nesbitt, L., Wells, G. A., Beaudoin, T., Spaite, D. W., Brisson, D., … & Cousineau, D., for the OPALS Study Group. (2003). Multicenter controlled trial to evaluate the impact of ALS on out-of-hospital chest pain patients. Academy of Emergency Medicine, 10(5), 501.

Stiell, I. G., Wells, G. A., Spaite, D. W., Nichol, G., Nesbitt, L., De Maio, V. J., … & Cousineau, D., for the OPALS Study Group. (2002). Multicenter controlled clinical trial to evaluate the impact of advanced life support on out-of-hospital respiratory distress patients. Academy of Emergency Medicine, 9(5), 357.

Stiell, I. G., Wells, G. A., Spaite, D. W., Nichol, G., O’Brien, B., Munkley, D. P., … & Anderson, S., for the OPALS Study Group. (1999). The Ontario Prehospital Advanced Life Support (OPALS) Study Part II: Rationale and methodology for trauma and respiratory distress patients. Annals of Emergency Medicine, 34, 256-262.

Wakefield, B. J., Mentes, J., Holman, J. E., & Culp, K. (2008). Risk factors and outcomes associated with hospital admission for dehydration. Rehabilitation Nursing, 33(6), 233-241. doi:10.1002/j.2048-7940.2008.tb00234.x

Scrutinizing the Literature of EMR

 As I scrutinize Dimitropoulos and Rizk (2009) for possible inclusion in a literature review for my research, I find it both promising and troubling. The article appears to be pertinent to my research question of how various laws and practices might adversely affect shared access of electronic health records; however, it is important to understand if this article is a documentation of primary research or a review of existing research, and as I describe below, this is unclear. This lack of clarity obscures other facets of the article that important to a researcher. These are also described below.

Initially, the work of Dimitropoulos and Rizk appears to be pertinent to my research based on the title and the publication in which it appears. Health Affairs is a respected journal within the realm of public health research, practice, and instruction, and it is ranked seventh of all health policy and service journals by Journal Ranking (http://www.journal-ranking.com). Publication within Health Affairs does not degrade the reputation of the authors and serves only to promote their work to their peers. As my research is within the realm of public health, Health Affairs is an obvious avenue to pursue for relevant work, and as this article by Dimitropoulos and Rizk appears to reflect a specific focus on the relationship between privacy laws and the ability, or lack thereof, to share health information, it appears to have relevance.

According to the abstract, Dimitropoulos and Rizk (2009) examine how variations is state (and, territorial) privacy laws might inhibit sharing health information via an central exchange, or repository. Though it would seem plausible for Dimitropoulos and Rizk to conduct their own research, the abstract seems to imply that they are merely reporting on the findings of a committee charged with examining such irregularities in privacy laws amongst the states and territories, presumably, of Canada. After reading the report, though, I find a disconnect between the abstract and the article. In the abstract, it appears more as if the authors are detached reporters, but within the body of the article, it seems as though they appear to take ownership of the primary research. This is confusing as it was plainly stated that the research was conducted by a large consortium of state officials: “the project initially engaged organizations in thirty-four … and later … forty-two jurisdictions. This collaborative work is commonly referred to as the Health Information Security and Privacy Collaboration (HISPC)” (p. 429).

This report is confusing to read as the perspective shifts frequently between first- and third-person. Additionally, the authors describe opinions formed and emotions felt during the primary research (opinions and emotions that only the primary researchers could know), yet it is unclear if these were transmitted through other writing or if the authors formed and felt these themselves. It is unclear whether the authors, Dimitropoulos and Rizk (2009), were participating researchers or merely reporters.

Both authors are noted to work for RTI International’s Survey Research Division, yet this corporation is not credited with any of the original research (Dimitropoulos & Rizk, 2009). I would have to conduct further research into the authors, their employer, and the project, itself, in order to make a final determination of the credibility of this article. This research would, hopefully, give the authors’ words better context, also. Complicating this is the absence of clearly delineated references, although a few appear within the Notes section that appear to be worth investigating.

Dimitropoulos and Rizk (2009) describe an effort to create a cohesive environment that will enhance the ability to share health information throughout a number of jurisdictions. As such, there is no scientific inquiry and it follows that adherence to the scientific method would be inappropriate. Again, however, it is unclear if this research is original or not.

In closing, it appears that Dimitropoulos and Rizk (2009) are credible in their writing; however, as each article must be able to stand on its own, and the article is lacking in form and perspective, I question the origination, application, and utility of this article, at least as it pertains to my original research question. Privacy in computing has been a major concern in the past two decades (Johnson, 2004). I feel that I could find more pertinent literature by expanding my search beyond this article.

References

Dimitropoulos, L. & Rizk, S. (2009).A state-based approach to privacy and security for interoperable health information exchange.Health Affairs, 28(2), 428-434. doi:10.1377/hlthaff.28.2.428

Johnson, D. G. (2004). Computer ethics. In L. Floridi (Ed.), The Blackwell guide to the philosophy of computing and information (pp. 65-75). Malden, MA: Blackwell.

Flawed Conclusions in Literature Review

For this week’s discussion, I have chosen to analyze an article (Sakr et al., 2006) that attempts to outline the efficacy and potential dangers of certain drugs used to treat shock. As a critical care paramedic, the discussion surrounding this article can provide insight to choosing alternative therapies when caring for my patients, but it is important for me to understand the potential biases and limitations of such a study that could lead to flawed conclusions (Gluud, 2006).

Sakr et al. (2006) collected data on ICU admissions over a two week period to further understand how dopamine effects mortality and morbidity when administered in response to hemodynamic compromise. Also, other administered vasoactive drugs were included in the analysis whether administered concomitantly with dopamine or instead of dopamine. The researchers did not distinguish between etiologies except to delineate between septic shock and non-septic shock. Patients who presented with shock or suffered a shock state within the first 24 hours of admission were included in the analysis. Patients admitted to the ICU mainly for 24 hour surgical observation where not included.

Shock is defined as “a state of inadequate cellular sustenance associated with inadequate or inappropriate tissue perfusion resulting in abnormal cellular metabolism” (Hillman & Bishop, 2004, p. 121). There are many etiologies of shock, including sepsis, anaphylactic, neurogenic, hypovolemic, cardiogenic, and others, which respond differently to various therapies. This confounder creates an information bias, as this variable is not identified in the data collection and cannot be scrutinized. Simply identifying the etiology of each shock state would limit this bias. The researchers, however, acknowledge this limitation and others.

Another confounding variables is the time constraint of the data. In regards to septic shock, this variable becomes evident. Many pathogens spread predictively during certain times of the year. The concomitant treatment of these infections could predispose patients to suffer a prolonged state of shock (in cases where the pathogen might not be immediately recognized) or provide for an ideal treatment pathway when the pathogen and the antibiotic regimen are fully understood and effective. This selection bias could be controlled by choosing patients who present throughout the year.

As Gluud (2006) points out:

When intervention effects are moderate or small, the human processing of data, unsystematic data collection, and the human capacity to overcome illnesses spontaneously limit the value of uncontrolled observations. Experimental models are essential for estimation of toxicity and pathophysiology.
(p. 494)

References

Gluud, L. L. (2006). Bias in Clinical Intervention Research. American Journal of Epidemiology, 163(6), 493–501. doi:10.1093/aje/kwj069

Hillman, K. & Bishop, G. (2004). Clinical Intensive Care and Acute Medicine. West Nyack, N.Y.: Cambridge University Press.

Sakr, Y., Reinhart, K., Vincent, J., Sprung, C. L., Moreno, R., Ranieri, V. M., De Backer, D., & Payen, D. (2006). Does Dopamine Administration in Shock Influence Outcome? Results of the Sepsis Occurrence in Acutely Ill Patients (SOAP) Study. Critical Care Medicine, 34(3), 589-597. doi:10.1097/01.CCM.0000201896.45809.E3