Guaiac Interpretation by non-Certified Clinicians

Abstract

Objective

To assess the ability of non-certified clinicians to interpret guaiac test results.

Methods

50 clinicians were shown photographs of 20 guaiac test cards. They were asked to classify these as positive, negative, or uncertain.

Results

31 nurses, 4 clinical assistants, 7 emergency attending MDs, 2 resident MDs, and 6 medical students comprised the study group. Mean correct percentages for 15 clearly positive or negative guaiac test cards ranged from 87% to 96% in each of the above groups. Correct results were lower for five of the borderline positive guaiac cards.

Conclusion

This study demonstrates that all clinician groups correctly interpreted the guaiac result most of the time, but there was a moderate error rate. This data supports the Joint Commission standard for this particular point of care test; clinicians interpreting guaiac testing should be trained with demonstrated competency.

Introduction

Hospital accrediting organizations require clinicians to be certified (documentation that includes training and competency testing) to perform a point of care test (also known as a bedside test) under the provisions known as waived testing.¹ These include glucose measurements, urine pregnancy tests, urine dipsticks, stool guaiac tests, and other tests. The consequence of this is the restriction and reduction of the spectrum of practice by clinicians.² The guaiac test is a test for blood in the stool or gastric contents. It is very sensitive and it is commonly used for gastrointestinal cancer screening for occult blood in primary care practices. In emergency departments (ED), it is commonly used to confirm the presence of blood in the stool. It appears to be very simple to interpret as shown in figures 1 and 2. Most or nearly all physicians are aware of how this test is performed and interpreted. Yet regulations restrict the interpretation of guaiac results to personnel who have completed training and are demonstrated as competent to perform this test.³ While the actual training is not difficult, there are additional administrative burdens to maintain log books with test device serial numbers and the annual demonstration of competency for all the clinicians performing this test. The stat lab will typically provide a guaiac result to ED patients in about one hour. The bedside test result provides a result in about 1 minute. This shorter time could speed patient throughput and reduce costs. The ED at Kapi‘olani Medical Center For Women & Children (KMCWC) currently does not perform bedside stool guaiacs. The lab at KMCWC performs approximately 20 stool guaiac tests per month. We hypothesize that because the guaiac test is simple to interpret, clinicians do not need special training or certification to interpret this point of care test.

Figure 1.

Guaiac test card interpretation. Two spots of stool sample are smeared on the opposite side of the white paper window labeled as 1 and 2. The card is then reversed and a liquid developer is applied to to the paper window. There is a control strip on the right labeled as “performance control area.” The control strip should turn blue when the developer is applied to confirm that the card and the developer are both working. The stool spot in area 1 is clearly positive with a definite blue result. The stool spot in area 2 is borderline positive with a hint of blue within the spot.

Figure 2.

Negative test. The control strip is blue, but the sample has no blue color.

The purpose of this study is to assess the ability of non-certified clinicians to interpret guaiac test results.

Methods

A medical technologist who is certified to interpret guaiac results photographed a convenience sample of twenty guaiac cards (Seracult, Propper Manufacturing Company Inc., Long Island City, NY). These were actual lab specimens and the study sample was selected to get some positive, negative, and borderline results. The medical technologist obtained 8 positive results, 5 borderline positive results, and 7 negative results. These 20 photographs were cropped and printed with a photographic printer so that they were roughly the actual size of the original guaiac cards. No identifying information was included in any of the samples.

Potential study subjects were approached by the study investigator to ask if they might be interested in participating in this study. If they were not currently certified to read guaiac cards (had not completed a competency training and testing for this) and if they were willing to participate in the study, the study was described to the potential study subject and he/she was given a study information sheet as part of a verbal consent process. This study was approved by the institutional review board (IRB).

The convenience sample of fifty non-certified medical personnel consisted of nurses, clinical assistants, attending physicians, resident physicians, and medical students. They were briefly instructed by the study investigator on how to interpret a guaiac card result. They were asked to interpret the guaiac card photographs as positive, negative, or uncertain. The results were then compared for accuracy with the results of the certified medical technologist. These card photographs all included the control test strip that clearly tested positive “blue” on all 20 sample cards.

Results

“Correct” guaiac result interpretations were defined as a result by the study subject that agrees with the interpretation in the result key. To examine the basic interpretation skill of the study group, the results summarized in table 1 excluded the 5 “borderline positive” guaiac cards so that the 15 remaining guaiac card results were clearly positive or negative.

Nine subjects were previously certified (completed a competency course and test) to interpret guaiac cards but their certification period had elapsed at the time of study participation. Out of the 15 clearly positive or negative cards, the previously certified subjects had a mean of 87% correct, while the subjects who were never previously certified had a mean of 90% correct. Since all the previously certified subjects were nurses (ED and ICU), this can also be compared to the 23 never previously certified nurses who had a mean of 87% correct. None of these are significantly different.

Linear regression determined there to be no relationship between years of experience and the number of correct guaiac interpretations.

The medical group (ED attending MD, resident MD, and medical students) combined to have a mean 95% correct which was higher than the mean 87% correct by the combined nursing group (all RN + clinical assistants) (P = .01). While this difference of 7% might seem large, it actually only represents 1 more correct response out of the 15 guaiac results.

In examining the five “borderline positive results”, the results are summarized in table 2. The percentage of correctly interpreted positive results for these five borderline positive guaiacs was lower than the overall percentage of correctly interpreted results for the 15 guaiacs that were more clearly positive or negative.

Table 2.

Guaiac card interpretations for 5 borderline positive guaiac cards

Study Group	n	Mean Number and Percent of Guaiacs Correctly Marked Positive	Percentage of Subjects Who Correctly Marked All 5 as Positive	Mean Number of Guaiacs Incorrectly Marked Negative	Mean Number of Guaiacs Marked as uncertain
ED RN	20	3.9 (77%)	50%	0.8	0.4
ICU RN	11	3.4 (67%)	36%	1.4	0.3
All RNs	31	3.7 (74%)	45%	1.0	0.3
CA	4	3.8 (75%)	25%	1.0	0.3
All RN + CA	35	3.7 (74%)	43%	1.0	0.3
ED MD	7	2.9 (57%)	0%	1.7	0.4
Resident MD	2	5.0 (100%)	100%	0	0
Medical Students	6	2.8 (57%)	0%	2.0	0.2
All MDs + MS	15	3.1 (63%)	13%	1.6	0.3

Following the above results, we asked 13 certified medical technologists to interpret all 20 guaiac cards. They used the same guaiac card images. All 13 medical technologists interpreted all 20 guaiac cards correctly (100%). This included the correct interpretation of the borderline positive guaiac cards.

Discussion

This study demonstrates that all clinician groups correctly interpreted the guaiac result most of the time, but there was a moderate error rate. These clinicians were not currently “certified” to interpret guaiac cards. In comparison, certified medical technologists were 100% correct in interpreting the guaiac cards. Previously certified clinicians in the study group did not perform significantly better than the other clinicians. Physicians performed slightly better than non-physicians.

Although the interpretation of guaiac cards appears to be objective and straightforward, this study demonstrates that consistent and perfect interpretation is not easily achieved. Therefore the certification requirement for clinicians interpreting guaiac testing is supported by this study. If this test is to be useful for clinical decision-making, its interpretation needs to be accurate. While most of the interpretations were accurate, some clinicians had accuracy rates as low as 60%. This potential for error is unacceptable.

This study is not able to address the question of whether the completion of a competency course and test would improve the accuracy rate of the clinicians, but it is clear that amongst this convenience sample of non-certified clinicians, the accuracy rate is not perfect.

In conclusion, this data supports the Joint Commission standard for this particular point of care test; clinicians interpreting guaiac testing should be trained with demonstrated competency.

Figure 3.

This is a borderline positive result. Both spots have a slight hint of blue color.

Table 1.

Percentage of 15 clearly positive or negative guaiac cards interpreted correctly among study subject groups.

Study Group	n	Mean Correct Percentage	Correct Percentage Range	Number of Subjects with All 15 Correct
ED RN	20	88%	60% to 100%	5
ICU RN	11	84%	60% to 100%	1
All RNs	31	87%	60% to 100%	6
Clinical assistants (CA)	4	88%	73% to 100%	0
All RN + CA	35	87%	60% to 100%	6
ED attending MD	7	96%	87% to 100%	4
Resident MD	2	93%	87% to 100%	1
Medical Students (MS)	6	94%	80% to 100%	3
All MDs + MS	15	95%	60% to 100%	8

Conflict of Interest

The authors have no financial conflicts of interest.