The Effect of Removal of Point-of-Care Fecal Occult Blood Testing on Performance of Digital Rectal Examinations in the Emergency Department
Article Outline
Study objective
We determine whether removing point-of-care fecal occult blood testing from the emergency department (ED) is associated with a decrease in documented digital rectal examinations.
Methods
We performed a retrospective observational chart review study examining documented digital rectal examinations, before and after removal of a point-of-care fecal occult blood test, on all adult patients who presented to our ED with chief complaints that were likely to warrant a fecal occult blood test (intervention-sensitive). We studied the 6 months immediately before and after switching from bedside fecal occult blood testing to immunohistochemical laboratory fecal occult blood testing. We compared the results with those from a similar cohort of patients who presented during the same period, with chief complaints that would warrant a digital rectal examination for reasons other than fecal occult blood test (intervention-insensitive).
Results
A total of 4,981 and 5,557 patients met our inclusion criteria during the before and after intervention periods, respectively. We observed an overall reduction of 10% (95% confidence interval [CI] 8% to 12%) in digital rectal examinations in patients with intervention-sensitive chief complaints. The largest relative decreases in digital rectal examinations were observed in patients with chief complaints of abdominal pain, nausea/vomiting, and diarrhea. Smaller decreases were observed in gastrointestinal bleeding, constipation, and rectal problem. There was an overall reduction of 3% (95% CI 0% to 5%) in documented digital rectal examinations in intervention-insensitive chief complaints. After controlling for all covariates, digital rectal examinations decreased in the postintervention period for intervention-sensitive (odds ratio 0.44 [95% CI 0.39 to 0.50]) and, to a lesser extent, for intervention-insensitive (odds ratio 0.67 [95% CI 0.52 to 0.86]) conditions.
Conclusion
Removal of point-of-care fecal occult blood test from our ED was associated with a reduction in digital rectal examinations, particularly among chief complaints that may require fecal occult blood testing.
Introduction
Background
Use of point-of-care testing is becoming increasingly common in the emergency department (ED) setting as more bedside tests become available.1 How point-of-care testing affects diagnostic accuracy, length of stay, patient throughput, or costs remains unclear.2, 3, 4, 5, 6 Previous studies have demonstrated that introduction of point-of-care tests may lead to increased test ordering by physicians, but the effect of removing an existing point-of-care test is currently unknown.5
What is already known on this topic
Because of regulation, many hospitals no longer permit bedside fecal occult blood testing, instead requiring samples to be sent for laboratory analysis.
What question this study addressed
Does the added inconvenience and delay of laboratory fecal occult blood testing influence a physician's decision to perform digital rectal examination?
What this study adds to our knowledge
In this chart review of 6 months before and after prohibition of bedside fecal occult blood testing, the rate of digital rectal examinations decreased 10% for those for whom occult blood testing was likely the primary reason for the examination and 3% for those for whom it was likely not.
How this might change clinical practice
Replacing point-of-care testing with less convenient laboratory testing affects physician practice patterns. Regulations probably should be studied for unintended influence in actual clinical situations before implementation.
The fecal occult blood test is a point-of-care test used to detect occult gastrointestinal bleeding. It is a recommended part of the ED evaluation of various chief complaints and has been shown to be useful in making some diagnoses.7, 8 Digital rectal examinations are usually performed to obtain stool for the fecal occult blood test but are also performed to evaluate rectal tone, prostate condition, and the presence of abnormal masses.
There are numerous types of fecal occult blood tests, including the point-of-care guaiac test and various immunochemical methods.9, 10 Immunochemical tests appear to be more sensitive and specific than the guaiac-based point-of-care tests but are more expensive and time consuming.9, 10, 11
Importance
A previous study has shown a decrease in rates of digital rectal examinations and fecal occult blood tests when additional quality control measures were added to the standard guaiac point-of-care test.12 At our institution, the test for fecal occult blood was recently changed from a point-of-care guaiac test (done at the bedside by the examiner) to an immunochemical test performed by the laboratory. In the laboratory-based model, ordering a fecal occult blood test requires the provider (or patient) to collect a stool sample; place the sample in a vial; label, initial, and mark the sample with the time; send the vial to the laboratory; and enter a computer order for the test. It is unknown whether this change in policy and the additional time required to perform testing affected the providers' decisions to perform digital rectal examinations or order fecal occult blood tests.
Goals of This Investigation
The objective of this study was to evaluate whether removal of point-of-care fecal occult blood testing was associated with a decrease in the number of digital rectal examinations in the ED. In addition, we sought to determine whether this change in digital rectal examination performance would also occur for patients who warranted a digital rectal examination for reasons other than fecal occult blood test.
Materials and Methods
Study Design and Setting
This study was a retrospective observational chart review study with a before-and-after design. The setting was an academic ED in a tertiary care university hospital with an annual census of approximately 50,000. The ED is staffed 24 hours per day by attending and resident emergency physicians and 15 hours per day by physician assistants. The study period was 6 months immediately before and after the change from point-of-care to laboratory fecal occult blood test testing. Inclusion dates for the preintervention group were April 1, 2008, to September 29, 2008, and October 1, 2008, to March 31, 2009, for the postintervention group. April 1, 2008, was selected as the starting date of the study because there were significant fluctuations in the ED facility and the number and composition of the staff before this date. These variables had stabilized by the start date. Because the change occurred in the middle of the day on September 30, 2008, we excluded data from this day. On the day of point-of-care fecal occult blood test removal, test media and developer were systematically removed from the ED and use of unauthorized fecal occult blood testing was highly discouraged. Other than the change in fecal occult blood testing, there were no other known changes during the study period such as education, training, or hospital protocols that may have affected digital rectal examination performance. Patients younger than 18 years and patients who left before a complete evaluation (eg, left against medical advice, left before visit complete, received only a medical screening examination, or triaged to the obstetrics unit) were excluded from the study. The study protocol was approved by our local institutional review board with a waiver of informed consent.
Data Collection and Processing
Registration and charting of all ED patient visits at our institution were documented with an electronic ED information system (PulseCheck; Picis, Waltham, MA). The electronic chart abstractor was a trained information technology specialist who was blind to the purpose of the study. We selected charts for electronic review according to the chief complaints that had potential for digital rectal examination during the ED clinical encounter. The set of chief complaints was based on a previous study of ED fecal occult blood testing.12 These chief complaints were further classified into “intervention-sensitive” and “intervention-insensitive,” determined a priori according to the potential to warrant fecal occult blood testing as part of the digital rectal examination. Patients who presented to our ED were assigned a chief complaint from a standardized list of complaints. Therefore, chart selection was unlikely to miss a significant number of patients with relevant chief complaints because of the inability to add free-text complaints. Intervention-sensitive chief complaints were abdominal pain, nausea and vomiting, diarrhea, constipation, rectal problem, and gastrointestinal bleeding. These complaints do not always merit a digital rectal examination, but when a digital rectal examination is performed, it is likely for obtaining fecal occult blood testing. Intervention-insensitive chief complaints were back pain and motor vehicle crash because, in most cases, the digital rectal examination is performed to assess rectal tone, prostate location, or gross blood, rather than for occult blood.
The primary outcome was rate of documented digital rectal examination, and the intervention was removal of point-of-care fecal occult blood testing. ED charts from all selected patient visits were searched electronically for the presence of the following words or phrases: “rectal examination,” “rectal tone,” “hemoccult,” “guaiac.” These phrases were chosen because they are included in the standardized physical examination template that is used by most physicians to document their findings. Providers have the option to free-text findings as well, but this is much less common and a brief review of charts demonstrated that these phrases were the most commonly used in free-text entries as well. Presence of one of these phrases in the physical examination or any free-text fields in the chart was assumed to indicate performance of a digital rectal examination. Additional covariates collected electronically were age, sex, race/ethnicity, triage acuity level, provider type, mode of arrival, payer source, ED disposition, and ED length of stay. A single abstractor, blind to the purpose of the study, was trained by one author (N.J.C.), using 20 practice charts and a discussion of coding rules. The abstractor performed a manual review of 215 charts randomly selected from eligible chief complaints to evaluate the reliability of the electronic chart abstraction. Performance of the chart abstractor was not actively monitored after the initial training. The interrater agreement of electronic and manual abstraction was measured by a κ statistic.
Primary Data Analysis
We performed all analyses with Stata 10.1 (StataCorp, College Station, TX). Proportions with 95% confidence intervals (CIs) of performance of digital rectal examination before and after the change were calculated for each chief complaint and for intervention-sensitive and intervention-insensitive groups as a whole. Digital rectal examination performance was plotted monthly to assess temporal trends, in addition to the overall before-and-after comparison. Simple linear regression was used to assess trends for segments within this plot. We constructed multivariable models, stratified by intervention-sensitive and intervention-insensitive chief complaints and including all patient, provider, and visit characteristics, to evaluate the independent effect of the removal of point-of-care fecal occult blood testing on the outcome of digital rectal examination performance. Results are reported as odds ratios with 95% CIs. Finally, interrater reliability between our electronic data collection and manual data collection was assessed by calculating a κ statistic.
Results
A total of 26,102 patients were triaged in the preintervention period and 26,952 in the postintervention period. Of these, 6,012 and 6,048 had relevant chief complaints and 4,981 and 5,557 met our inclusion criteria in each of the periods, respectively (Figure 1). The most common reasons for exclusion were leaving before visit complete, triaged to obstetrics, and aged younger than 18 years.
Figure 1.
Included and excluded study subjects in the preintervention and postintervention periods. The figure shows the number of patients triaged in the preintervention and postintervention periods, respectively, along with a description of excluded subjects. PRE, Preintervention; POST, postintervention; CC, chief complaint; <18 y.o., patients younger than 18 years; AMA, left against medical advice; TXFR, transferred; LBVC, left before visit complete; TrOB, triaged to obstetrics floor.
The preintervention and postintervention groups were similar with regard to patient, provider, and visit characteristics (Table 1). Table 2 shows change in the rate of digital rectal examination performance by patient and visit characteristics. There was an overall reduction of 10% (95% CI 8% to 12%) in the performance of digital rectal examinations in patients with intervention-sensitive chief complaints. There was an overall reduction of 3% (95% CI 0% to 5%) in the performance of digital rectal examinations in intervention-insensitive chief complaints. After controlling for all covariates, digital rectal examination performance decreased in the postintervention period for intervention-sensitive (odds ratio 0.44 [95% CI 0.39 to 0.50]) and, to a lesser extent, for intervention-insensitive (odds ratio 0.67 [95% CI 0.52 to 0.86]) conditions. Interrater agreement between electronic and manual chart abstraction was excellent (κ=0.87).
Table 1. Comparison of patient characteristics before and after removal of point-of-care fecal occult blood test.
| Characteristics | Preintervention, n=4,981, No. (% [95% CI]) | Postintervention, n=5,557, No. (% [95% CI]) |
|---|---|---|
| Chief complaint | ||
| Intervention sensitive | ||
| Abdominal pain | 2,774 (56[54–57]) | 3,088(56[54–57]) |
| Nausea/vomiting | 708(14[13–15]) | 794(14[13–15]) |
| Diarrhea | 90(2[1–2]) | 81(1[1–2]) |
| GI bleeding | 117(2[2–3]) | 170(3[3–4]) |
| Constipation | 30(<1[0–1]) | 27(<1[0–1]) |
| Rectal problem | 88(2[1–2]) | 101(2[1–2]) |
| Intervention insensitive | ||
| Motor vehicle crash | 496(10[9–11]) | 499(9[8–10]) |
| Back pain | 678(14[13–15]) | 797(14[13–15]) |
| Characteristics | ||
| Age, y, median(IQR) | 37(26–51) | 36(26–50) |
| Female sex | 3,237(65[64–66]) | 3,579(64[63–66]) |
| Race/ethnicity | ||
| Non-Hispanic white | 2,398(48[47–50]) | 2,624(47[46–49]) |
| Non-Hispanic black | 1,222(25[23–26]) | 1,380(25[24–26]) |
| Hispanic | 1,125(23[21–24]) | 1,242(22[21–23]) |
| Other | 234(5[4–5]) | 302(5[5–6]) |
| Insurance | ||
| Private | 1,071(22[20–23]) | 1,113(20[19–21]) |
| Public | 1,094(22[21–23]) | 1,239(22[21–23]) |
| None | 2,379(48[46–49]) | 2,678(48[47–50]) |
| Missing | 437(9[8–10]) | 527(9[9–10]) |
| Arrival by EMS | 1,029(21[20–22]) | 1,218(22[21–23]) |
| Nonurgent triage acuity | 776(16[15–17]) | 844(15[14–16]) |
| Type of provider | ||
| Resident involved | 4,109(83[81–84]) | 4,704(85[84–86]) |
| Attending physician only | 413(8[8–9]) | 452(8[7–9]) |
| Physician assistant only | 452(9[8–10]) | 396(7[6–8]) |
| ED LOS, min, median(IQR) | 287(187–430) | 276(179–404) |
| Hospital admission | 949(19[18–20]) | 1,088(20[19–21]) |
Table 2. Comparison of digital rectal examination rates before and after removal of point-of-care fecal occult blood test.
| Characteristics | Preintervention, No. (%) | Postintervention, No. (%) | Preintervention-Postintervention, Δ% (95% CI) |
|---|---|---|---|
| Chief complaint | |||
| Intervention sensitive | 927(24) | 609(14) | 10(8to12) |
| Abdominal pain | 590(21) | 341(11) | 10(8to12) |
| Nausea/vomiting | 122(17) | 62(8) | 9(6to12) |
| Diarrhea | 42(47) | 18(22) | 24(11to38) |
| GI bleeding | 87(74) | 109(64) | 10(0to21) |
| Constipation | 18(60) | 11(41) | 19(−6to45) |
| Rectal problem | 68(77) | 68(67) | 10(−3to23) |
| Intervention insensitive | 170(14) | 155(12) | 3(0to5) |
| Motor vehicle crash | 49(10) | 50(10) | 0(−4to4) |
| Back pain | 121(18) | 105(13) | 5(1to8) |
| Characteristics | |||
| Age, y, median (IQR)⁎ | 45(30–57) | 45(32–57) | 0 |
| Sex | |||
| Female | 569(18) | 344(10) | 8(6to10) |
| Male | 528(30) | 420(21) | 9(6to12) |
| Race/ethnicity | |||
| Non-Hispanic white | 562(23) | 401(15) | 8(6to10) |
| Non-Hispanic black | 244(20) | 160(12) | 8(6to11) |
| Hispanic | 233(21) | 157(13) | 8(5to11) |
| Other | 58(25) | 45(15) | 10(3to17) |
| Insurance | |||
| Private | 221(21) | 153(14) | 7(4to10) |
| Public | 281(26) | 180(15) | 11(8to14) |
| None | 496(21) | 349(13) | 8(6to10) |
| Missing | 99(23) | 82(16) | 7(2to12) |
| Mode of arrival | |||
| EMS | 235(23) | 199(16) | 6(3to10) |
| Non-EMS | 859(22) | 565(13) | 9(7to10) |
| Triage acuity | |||
| Nonurgent | 103(13) | 84(10) | 3(0to6) |
| Urgent | 994(24) | 680(14) | 9(8to11) |
| Type of provider | |||
| Resident involved | 1,011(25) | 706(15) | 10(8to11) |
| Attending physician only | 42(10) | 23(5) | 5(2to9) |
| Physician assistant only | 44(10) | 35(9) | 1(−3to5) |
| ED LOS, min, median (IQR)⁎ | 336(222–479) | 296(201–436) | 40 |
| Disposition | |||
| Hospital admission | 296(31) | 205(19) | 12(9to16) |
| Not admitted | 801(20) | 559(13) | 7(6to9) |
⁎Refers only to patients who received digital rectal examination. |
Figure 2 shows a trend towards decreased rates of digital rectal examinations in the postintervention period, especially among patients with intervention-sensitive complaints. We noted a decrease in the rates of digital rectal examinations in month 4 of the study period in the intervention-sensitive group. Month 4 corresponded to a new academic year (July 2008), and therefore this decrease could represent the “July phenomenon.”13 Because of this possibility, we performed a post hoc analysis that excluded months 1 to 3 from the data set. We found that if July to September only was used as the baseline, there was still a 7% (95% CI 5% to 9%) reduction in digital rectal examination rates for intervention-sensitive chief complaints, but no change (2% [95% CI −1% to 5%]) for intervention-insensitive chief complaints.
Figure 2.
ED patients with documented digital rectal examination before and after removal of point-of-care fecal occult blood testing. The figure shows the rates of digital rectal examinations each month during the entire study period. Intervention-sensitive chief complaints are represented by grey squares; intervention-insensitive chief complaints are represented by grey triangles; all chief complaints combined are represented by black diamonds with associated 95% CI bars.
Limitations
This study has several potential limitations. The biggest possible source of error is in data collection. The categorization of chief complaints as intervention sensitive and intervention insensitive is not all-inclusive and is subject to misclassification. We relied on an electronic chart review for data extraction. Although our manual review of a random sample of charts demonstrated a good κ statistic for agreement, it is possible that our automated extraction has missed some digital rectal examination documentation or produced false-positive digital rectal examination documentation. As mentioned previously, most providers use “check boxes” for documentation of rectal examinations, but free-text documentation is also available, and our search strategy may have missed some of these cases. For example, if a provider typed “heme neg stool” in the free-text section of the physical examination section, this would not have been identified for our study. In addition, after this switch to laboratory testing, providers may have relied on the laboratory report to serve as documentation of an implied digital rectal examination. Furthermore, fecal occult blood tests can be performed without a digital rectal examination if stool is collected from a diaper or bedpan, for instance. We suspect this occurs relatively infrequently, however. The study spans 2 academic years and as such includes a different set of providers in the preintervention and postintervention periods. We performed a second analysis with 3-month preintervention and postintervention periods to limit the July phenomenon. The multivariable analysis also adjusted for provider type, and the change could not be explained by this variable. Finally, this study was conducted at a single ED in an academic setting and may not generalize to other EDs.
Discussion
We found that removal of point-of-care fecal occult blood tests was associated with a decrease in the rates of digital rectal examination in the ED among patients with chief complaints that may warrant a fecal occult blood test. We found a smaller decrease in the rates of digital rectal examinations in patients who are likely to have a rectal examination for reasons other than fecal occult blood testing, which suggests that removal of the point-of-care fecal occult blood test has changed the practice habits of providers in our ED. Our observation is similar to the that in the study by Adams et al,12 who found through formal review of written charts that adding quality assurance steps to the fecal occult blood test also reduced performance of fecal occult blood tests and digital rectal examinations. Our study is different, however, because we used electronic medical record abstraction to study the effect of removing a point-of-care test entirely. Adams et al12 specifically chose before-and-after periods separated by 1 year to eliminate the July phenomenon, whereas our study examined periods immediately before and after the change in testing to capture data on behavior immediately before the change. The July phenomenon remains as a potential confounder, but we believe the immediate and sustained change in documented digital rectal examinations suggests a real difference.
There has been much debate about the use, misuse, advantages, and disadvantages of point-of-care testing in the ED.14, 15, 16, 17 Many of the previous studies addressing ED point-of-care testing focus on what happens when a new point-of-care test is introduced to an ED. Our study adds additional information to this debate by demonstrating that provider behaviors are also changed when an established point-of-care test is removed.
Some have speculated that the convenience of point-of-care testing may increase the number of tests that are ordered by providers.17 Reducing test convenience has been shown to reduce test ordering in some settings.18 The utility of the digital rectal examination was not necessarily reduced, but the convenience of obtaining fecal occult blood testing was dramatically reduced by the inability to obtain results immediately at the bedside. We assume that the purpose of digital rectal examination in the intervention-sensitive group was to obtain stool for fecal occult blood testing. If our results simply reflect the effect of convenience, then the higher rates of digital rectal examinations before removal of point-of-care testing may reflect unneeded testing. Others have argued that barriers to testing cause a decrease in necessary tests.12 In addition, the characteristics of the test itself may influence the degree to which convenience influences test ordering. Some consider the fecal occult blood test to be of little value in the ED setting. More essential tests may be less sensitive to the removal of point-of-care testing. Whatever the cause, in our ED, the additional steps required to order the new fecal occult blood test seem to have discouraged clinicians from performing digital rectal examinations. Hence, the removal of the point-of-care fecal occult blood test at our institution has affected the way we practice. The clinical effect of this change cannot be assessed by the current study.
At our institution, the implementation of immunochemical fecal occult blood testing was a system-wide change that affected outpatient clinics, inpatient wards, and the ED. The main impetus for this change was the increased sensitivity of the laboratory test. Although this method may be important for early colon cancer screening and detection, we doubt the importance of this increased sensitivity for acute care in the ED setting. In addition, the immunochemical test is more expensive than the bedside guaiac test, and this increased cost may not result in a true benefit.
In summary, the results of this study are hypothesis generating and require confirmation in larger prospective studies. The removal of a point-of-care fecal occult blood test from our ED was associated with a decreased rate of digital rectal examination performance, particularly among chief complaints that may require fecal occult blood testing. If these results are confirmed, administrators should consider the potential clinical impact before implementing or removing point-of-care testing.
References
- . Point of care testing in the emergency department. J Emerg Med. 2002;22:393–404
- Decreased patient charges following implementation of point-of-care cardiac troponin monitoring in acute coronary syndrome patients in a community hospital cardiology unit. Clin Chim Acta. 2006;370:191–195
- A randomized trial to assess the efficacy of point-of-care testing in decreasing length of stay in a pediatric emergency department. Pediatr Emerg Care. 2007;23:457–462
- Implementation of a rapid whole blood D-dimer test in the emergency department of an urban academic medical center: impact on ED length of stay and ancillary test utilization. Am J Clin Pathol. 2009;132:326–331
- Impact of point-of-care testing on patients' length of stay in a large emergency department. Clin Chem. 1996;42:711–717
- Point-of-care versus central laboratory testing: an economic analysis in an academic medical center. Clin Ther. 1994;16:898–910
- . Rosen's Emergency Medicine: Concepts and Clinical Practice. 6th ed.. Philadelphia, PA: Elsevier Health Sciences; 2006;
- Screening of patients with acute infectious diarrhea: evaluation of clinical features, fecal microscopy, and faecal occult blood testing. Scand J Gastroenterol. 2000;35:54–60
- Comparison of conventional guaiac to four immunochemical methods for fecal occult blood testing: implications for clinical practice in hospital and outpatient settings. Clin Chim Acta. 2009;400:120–122
- . Fecal occult blood tests in occult gastrointestinal bleeding. Semin Gastrointest Dis. 1999;10:48–52
- Screening for colorectal neoplasms with new fecal occult blood tests: update on performance characteristics. J Natl Cancer Inst. 2007;99:1462–1470
- The law of unintended consequences: the Joint Commission regulations and the digital rectal examination. Ann Emerg Med. 2008;51:197–201
- Indirect costs for medical education (Is there a July phenomenon?). Arch Intern Med. 1989;149:765–768
- . Point-of-care testing—can we move from anecdote to evidence?. Clin Chem Lab Med. 2006;44:1049–1051
- . Near-patient testing—point-of-care or point of costs and convenience?. Br J Anaesth. 2003;90:425–427
- . Point-of-care testing: no pain, no gain. QJM. 2001;94:571–573
- . Point-of-care blood testing and cost-effective emergency medicine. Acad Emerg Med. 1995;2:163–164
- A cost-effective method for reducing the volume of laboratory tests in a university-associated teaching hospital. Mt Sinai J Med. 2006;73:787–794
Supervising editor: Steven M. Green, MD
Author contributions: NJC and AAG conceived the study, designed the trial, and obtained institutional review board approval. NJC and MY performed electronic chart abstraction and data collection. AAG provided statistical advice and the primary data analysis. NJC authored the article, and all authors contributed substantially to its revision. NJC takes responsibility for the paper as a whole.
Funding and support: By Annals policy, all authors are required to disclose any and all commercial, financial, and other relationships in any way related to the subject of this article that might create any potential conflict of interest. The authors have stated that no such relationships exist. See the Manuscript Submission Agreement in this issue for examples of specific conflicts covered by this statement.
Publication date: Available online January 8, 2010.
Reprints not available from the authors.
Please see page 136 for the Editor's Capsule Summary of this article.
PII: S0196-0644(09)01842-3
doi:10.1016/j.annemergmed.2009.12.021
© 2009 American College of Emergency Physicians. Published by Elsevier Inc. All rights reserved.
