Martyn R. Smith

Design and Development of a Cardiopulmonary Resuscitation Mattress

This study introduces the design and construction of a mattress insert to produce more effective cardiopulmonary resuscitation (CPR). The mattress insert deflates, making the mattress insert a rigid surface. Using a device that administers a constant compression depth onto a manikin, we were able to show that our mattress insert more effectively directed the compressive force to the manikin compared to the current practice of using a headboard on top of a mattress. The mattress insert produced a statistically significant increase in the compression efficiency when compared to the current practice of using the headboard (81% vs. 53%). Because the mattress insert starts deflating immediately after the vacuum is turned on, 1 person is needed to initiate chest compressions. Compression begins while the mattress deflates. Running compression tests at incremental time periods, we found that our design reaches and surpasses the compression efficiency when using a headboard in 10 seconds.

Hyperglycemia as a predictor for mortality in veterans with pneumonia.

The elderly may have multiple problems including diabetes mellitus and pneumonia. This study was undertaken to evaluate the effect of hyperglycemia on the length of stay (LOS) and mortality of elderly veterans admitted for pneumonia. Forty-seven and forty-eight patients over the age of 60 admitted for the diagnosis of pneumonia at the Lexington, Kentucky and Iron Mountain, Michigan VAMCs respectively were evaluated. The mean age at Lexington was 72 and at Iron Mountain was 73.9. At Lexington the mean LOS was 39.5 days with a standard deviation of 9.2 for those with hyperglycemia (n = 18) and 16.5 days with a standard deviation of 2.5 for those without hyperglycemia (n = 29). At Iron Mountain the mean LOS was 21.5 days with a standard deviation of 3.3 for those with hyperglycemia (n = 10) and 13.6 days with a standard deviation of 1.2 for those without (n = 38). These differences were significant at p less than .01 using the 2-sample independent t-test. At Iron Mountain 5 of 38 (13%) of the patients with normoglycemia died during the hospital stay compared to 3 of 10 (30%) those with hyperglycemia. For Lexington the figures were 4 of 29 (13%) and 6 of 18 (33%) respectively. These differences were significant at p less than .05. Age did not have any significant effect on either mortality or length of stay. Hyperglycemia appears to be a significant predictor of a longer hospital course and increased mortality in elderly veterans hospitalized for pneumonia.

UTILIZATION OF THE LOCAL MEDIA IN AN INTRODUCTORY STATISTICS COURSE

Publisher Summary This chapter presents a supplemental technique used in teaching an introductory statistics course at Michigan Technological University. The class was assigned a group project. Their task was to design, carry out, and analyze an experiment to determine the number of articles in the local newspaper, the Daily Mining Gazett, that require knowledge of statistics to correctly interpret. The instructor adopted a hands off posture except to ask an occasional question. A study plan was developed. A great deal of time was spent debating the criteria for classifying the articles. The newspaper was monitored for 28 days. Part of the analysis of this data included testing to see if a Poisson model was reasonable for this data. Two procedures were proposed to test this, namely, a Chi-squared goodness-of fit and a variance test for Poisson distributions. These class projects helped significantly in having most of the students achieve all four goals in this course.