1.gif (1892 bytes)

Letters to the Editor

Indian Pediatrics 2004; 41:1280-1282

Clinical Significance and Type II Errors under the Magnifying Lens


The article by McGlone, et al.(1) attempting to simplify the insertion of umbilical artery catheters (UACs) by use of a magnifying lens kindled our interest. We pen this letter in an effort to prevent a good idea being discarded due to inappropriate interpretation of the p value.

Unquestionably both the time for insertion of a UAC as well as success rate will depend on the experience of the clinician(s) performing the procedure. Depending on the institution, this group may be as diverse as pediatric house-staff interns, neonatal-perinatal fellows and registered nurse practitioners. The average time taken to insert a UAC as noted in the pilot study undertaken by the authors(1) as well as previously(2) is about 310-330 seconds if employs the conventional method. Squire, et al.(2) found that use of the side-entry technique greatly decreased the average UAC insertion time to 143 seconds. In the current study(1), the authors report that in half their total study population, UACs were inserted in 88 seconds or less with or without the use of the magnifying lens. Although it is possible that the registrars who participated in the study were extremely skilled, as the authors note, results so highly divergent from the pilot data raise the possibility that the study was contaminated by attention bias (the so-called Hawthorne effect)(3); we suspect that the greatly diminished power of the study (to 6%) is a result of the compounded effect of both these variables. Given that all clinicians attempting to insert UACs will not be as skilled as the group of registrars that participated in the study, the use of a magnifying lens may still prove valuable. Indeed, in training situations, with inexperienced clinicians, the use of a magnifying device may be invaluable. In the paper, the authors have done an excellent job of describing in detail the methods employed and the results obtained; while they make appropriate comments as part of their discussion to ponder their results, they fail to correctly state their conclusion or the key message for readers. Their conclusions that there was no difference between the two study groups fall into the realm of possible Type II errors(4).

A more appropriate method of presenting the same results would be to begin, as is traditionally done(5), with the null hypothesis that use of a magnifying lens would not decrease the time of insertion of UACs, and then conclude that given the limitations of power (due to overestimation of effect, underestimation of variation in operator skills and the Hawthorne effect), the null hypothesis could not be rejected. When analyzing data, it is vital to distinguish between results that demonstrate a treatment had no effect from results that are inconclusive. A non-significant difference should always be accompanied by a confidence interval (CI). If a 95% CI for the difference contains no clinically important values, then "no difference" has been demonstrated. If the interval contains values that would be considered clinically important, then the result is inconclusive. The authors did not provide such an interval and doing so requires access to the raw data. However, enough information was provided to estimate an approximate 95% CI for the individual medians since, for the sample sizes used (23 and 21), a 95% CI is approximated by the inter-quartile range. Thus, approximate 95% CI’s are 50-192 for the 88 sec median and 55-222 for the 70 sec median. These intervals are very wide and they emphasize the fact that, even though the difference is not statistically significant, the true difference might be clinically important.

In closing, we would like to remind the readers and ourselves that one should always make a distinction between clinical significance (which may vary on a case by case basis) and statistical significance that may be difficult to prove in the clinical setting. We submit that this report of a ‘negative’ trial still has clinically useful information. Inasmuch as pediatric and microvascular surgeons routinely employ magnifying devices for their procedures, perinatal clinicians may find a relatively inexpensive, yet potentially useful piece of equipment such as a magnifying lens a worthwhile addition to their armamentarium.

Caraciolo J. Fernandes,
Department of Pediatrics,
Texas Children’s Hospital &
Baylor College of Medicine,
Houston, TX 77030.
E-mail: [email protected]

E. O’Brian Smith,

Department of Pediatrics-Nutrition,
Children’s Nutrition Research Center,
Baylor College of Medicine,
Houston, TX 77030

 

References

 

1. McGlone L, Devlin L, Doherty DA, Patole S. Use of magnifying lens to aid neonatal umbilical arterial catheter insertion. Indian Pediatr 2004; 41: 250-254.

2. Squire SJ, Hornung TL, Kirchhoff KT. Comparing two methods of umbilical artery catheter placement. Am J Perinatol 1990; 7: 8-12.

3. Hartman JM, Forsen JW, Jr., Wallace MS, Neely JG. Tutorials in clinical research: part IV: recognizing and controlling bias. Laryngoscope 2002; 112: 23-31.

4. Neely JG, Hartman JM, Forsen JW, Jr., Wallace MS. Tutorials in clinical research: VII. Understanding comparative statistics (contrast) - part B: application of T-test, Mann-Whitney U, and chi-square. Laryngoscope 2003; 113: 1719-1725.

5. Neely JG, Hartman JM, Forsen JW, Jr., Wallace MS. Tutorials in clinical research: part VII. Understanding comparative statistics (contrast) - part A: general concepts of statistical significance. Laryngoscope 2003; 113: 1534-1540.

 

Home

Past Issue

About IP

About IAP

Feedback

Links

 Author Info.

  Subscription