Utilization of fecal occult blood test in the acute hospital setting and its impact on clinical management and outcomes.

BACKGROUND: Despite lack of evidence supporting the use of fecal occult blood test (FOBT) in the acute hospital setting, FOBT is commonly used in the inpatient setting for reasons other than colorectal cancer (CRC) screening. AIMS: To evaluate practice trends in utilizing FOBT on inpatients and its impact on affecting clinical management and outcomes.
MATERIALS AND METHODS: A cross-sectional study of consecutive adult patients undergoing FOBT from January 1, 2011 to December 31, 2011 during an acute medical admission at a large urban safety net hospital was performed. Indications for FOBT, and the impact of FOBT results on endoscopic procedures performed and clinical outcomes were assessed. The number of inpatient endoscopic procedures before and after discontinuing inpatient FOBT was also assessed.
RESULTS: A total of 207 inpatient FOBTs were performed in 2011. The most common reason cited for FOBT was anemia (36%, n = 74) followed by gastrointestinal (GI) bleeding (27%, n = 55). Interestingly, 23% (n = 47) of the patients undergoing inpatient FOBT had overt GI bleeding. As expected, patients with positive FOBT were significantly more likely to undergo endoscopic examinations (P < 0.01). After discontinuing the availability of inpatient FOBT, patients were less likely to undergo endoscopic examinations [odds ratio (OR) 0.80, 95% confidence interval (CI) 0.75-0.85].
CONCLUSION: Inappropriate utilization of FOBT in the inpatient setting is common, even when the indication does not support its use. Setting limitations on inpatient FOBT may reduce the inappropriate utilization of inpatient FOBT. Quality improvement initiatives are needed to educate clinicians on the appropriate use of FOBT, which is primarily for average risk CRC screening.

Introduction

According to the American Gastroenterology Association's guidelines, fecal occult blood test (FOBT) should only be used in the context of colorectal cancer (CRC) screening.[1] Despite lack of evidence supporting the use of FOBT in the acute hospital setting, it is commonly used in this setting for reasons other than CRC screening.[2] This potential misuse has been questioned over concerns about its low specificity resulting in further unnecessary testing,[3] and low sensitivity that can delay appropriate diagnostic workup.[4] At this time, there are no clear guidelines recommending the use of FOBT beyond CRC screening.[1]

Previous studies describing the role of FOBT in the inpatient setting highlight the potential for FOBT to influence inappropriate clinical decision-making.[24567] It has also been suggested that implementation of FOBT in the inpatient setting contributes to increased hospital length of stay and increased medical costs.[7] Yet, FOBT is commonly used in the acute hospital setting and its influence on clinical decision-making, patient management, length of stay, and medical costs have not been fully elucidated. In addition, the impact of limiting the use of FOBT during inpatient settings has also not been well-studied.

To further elucidate the impact of inpatient FOBT, we retrospectively evaluated all adult patients at our large urban safety net medical center who underwent inpatient FOBT evaluation. A better understanding of FOBT practices in the inpatient setting can provide valuable information to guide quality improvement initiatives among trainees and nontrainee providers to improve utilization of FOBT.

Materials and Methods

We conducted a single-center cross-sectional study to evaluate all adult inpatients who underwent FOBT from January 1, 2011 to December 31, 2011 at a safety net hospital. FOBT ordered in the outpatient setting or in the emergency department among patients not admitted to the acute inpatient hospital were excluded. The samples were either collected by a nurse from a stool sample or a by physician during digital rectal example. All FOBTs were performed by guaiac-based Hemoccult® by CLI Awaived Incorporation (San Diego, California, USA) and the results were evaluated by in-house laboratories. Stool microscopy examination was not routinely performed for patients with FOBT positive results.

A detailed chart review of both paper charts and electronic medical records was performed, and the data collected included: age, sex, race/ethnicity, admitting diagnosis, admitting service, presence of overt gastrointestinal (GI) bleeding (i.e., hematemesis, “coffee-ground” emesis, melena or black tarry stool, hematochezia, bloody diarrhea, bright red blood hematemesis, bright red blood per rectum), indication for ordering the test, and test result. In addition, among patients who underwent endoscopic evaluation, endoscopic assessment of findings as well as histopathology reports of specimens were also reviewed. An endoscopic procedure was considered to be clinically significant if the findings contributed to a change in clinical management or the findings justified the indication for endoscopy on retrospective analysis. This determination was made by an expert opinion in a manner that was blinded to the FOBT result after reviewing each patient's clinical record.

The availability of inpatient FOBT was discontinued at our hospital on May 1, 2012. To identify the impact of inpatient FOBT discontinuation, the number of inpatient endoscopies performed from May 1, 2010 to May 1, 2012 and April 30, 2012 to April 30, 2014 were evaluated. Trends in the number of endoscopic procedures performed were adjusted for prevalence of anemia by International Statistical Classification of Diseases (ICD)-9 codes (280, 280.9, 285.1, 285.9), age, sex, and overall length of stay in all admitted patients.

Descriptive analyses of continuous variables were presented as mean and standard deviation for normally distributed variables and median and range for nonnormally distributed variables. Categorical variables were presented as proportions and frequencies. FOBT results (positive versus negative) were stratified by indication for FOBT, whether endoscopy was performed, and the resulting clinical significance if endoscopy was performed. Comparisons between categorical variables were analyzed using the Fisher's exact test or chi-square test as appropriate. Comparisons between continuous variables utilized Student's t-test. Statistical significance was met with a two-sided P < 0.05. All analyses and data transformations were performed using SAS (SAS Institute, Cary, NC). This study was approved by the hospital's institutional review board.

Results

A total of 1,007 FOBTs were performed in the calendar year 2011. Seven hundred and sixty-seven tests were ordered by outpatient clinics. Twenty-nine tests were ordered in the emergency room that did not result in a hospital admission, and four in-patients with FOBT had missing medical records that precluded data review. Our final study cohort included 207 patients who had inpatient FOBTs. The description of patients with an inpatient FOBT is presented in Table 1.

Table 1

Description of patients with an inpatient FOBT

Age		Years
	Mean±SD	56±12.3
	Range	18-90
Sex		N (%)
	Male	134 (65)
Race
	Black	79 (38)
	White	59 (29)
	Native American/Eskimo	37 (16)
	Asian-Pacific Islander	30 (14)
	Unknown	2 (1)
Ethnicity
	Non-Hispanic	159 (77)
	Hispanic	38 (18)
	Unknown	10 (5)
Service
	Medicine	197 (95)
	General surgery	2 (1)
	Orthopedics	2 (1)
	Neurosurgery	1 (0.5)
	Trauma	3 (1.5)
	OB/GYN	1 (0.5)
	Podiatry	1 (0.5)
Indication
	Anemia	74 (36)
	GI bleed	55 (27)
	GI bleed+Anemia	8 (3.9)
	Unknown	70 (34)
Overt blood
	No	160 (77)
	Yes	47 (23)
	Melena	18 (38)
	Black/Dark stool	6 (13)
	Hematemesis	9 (19)
	Bloody diarrhea	6 (13)
	Hematochezia	9 (19)
	Other	3 (6)
	Unknown	1 (2)
Test result
	Positive	70 (34)
	Negative	131 (63)
	Indeterminate	6 (3)

Of the 207 tests ordered 74 (36%) were ordered for anemia workup, 55 (27%) with a concern for GI bleed, and 8 (3.9%) for GI bleed and anemia [Table 2]. The indication for FOBT could not be determined in 70 (34%) patients. Overt GI bleed was present in 47 (23%) the patients on whom the test was performed. Despite the observation that 23% of inpatient FOBTs was performed in patients who had overt GI bleeding, of more concern is the observation that 29 (53%) of the tests performed for overt GI bleeding were positive while a majority of the tests ordered for “anemia” were negative (n = 55, 74%). Tests ordered for “GI bleed” workup were more likely to be positive compared to tests ordered for anemia workup (53% versus 24%, P = 0.001).

Table 2A

FOBT result by ordering indication, performance of procedure, and clinical significance

Variables	Positive FOBT		Negative FOBT		P value
	N (%)	Confidence interval	N (%)	Confidence interval
Indication					0.001
GI Bleed	29 (53)	0.39-0.66	23 (43)	0.29-0.56
Anemia	18 (24)	0.15-0.36	55 (74)	0.63-0.84
GI Bleed+Anemia	4 (50)	0.16-0.84	4 (50)	0.16-0.84
Unknown	19 (27)	0.17-0.39	49 (70)	0.58-0.80
Procedure performed					0.0001
Yes	24 (34)	0.23-0.47	14 (11)	0.06-0.17
No	46 (66)	0.53-0.77	117 (89)	0.83-0.94
Clinical significance					0.77
Significant	20 (83)	0.63-0.96	8 (57)	0.29-0.829
Not significant	4 (17)	0.05-0.37	6 (43)	0.18-0.71

Percentages may not total to 100% because indeterminate tests were not presented

Of the FOBTs sent, 70 (34%) had at least one positive test result, 131(63%) had negative results, and 6 (0.03%) were indeterminate. Among all patients that underwent inpatient FOBT, 39 (19%) underwent endoscopic examinations [17—esophagogastroduodenoscopy (EGD), 18—EGD plus colonoscopy, 3—flexible sigmoidoscopy, 1—EGD plus flexible sigmoidoscopy]. Patients with positive FOBT were more likely to have endoscopic examinations compared to those with negative FOBT (34% versus 11%, P = 0.0001). Among patients with positive FOBT, 2.9% were found to have a malignancy (gastric adenocarcinoma or leiomyoma) and 2.9% were found to have adenomas. Among patients with negative FOBT who underwent an endoscopic procedure, malignancy (carcinoid) and adenomas were detected in 0.8% and 0.2%, respectively. Overall, patients with positive FOBT had a higher number of clinically significant endoscopic findings (i.e., either altered the management or supported the indication for endoscopy) compared to patients with negative FOBT; however, the difference was not statistically significant (83% versus 57%, P = 0.77).

A total of 47,223 admissions took place at Highland Hospital between May 1, 2010 and April 30 2014 (23,162 admissions before and 24,061 after FOBT was discontinued). The characteristics of inpatient pre- and post-FOBT are presented in Table 3. No significant difference in length of hospital stay was observed in the eras before and after the discontinuation of FOBT. However, there were more patients with anemia in the group prior to discontinuing FOBT (2,435 versus 2,294, P = 0.0004); these patients were also younger (41.8 versus 40.8, P < 0.0001). After adjusting for baseline anemia, age, and gender between the two groups, there was a significant decrease in the number of inpatient procedures (P < 0.0001). In the era following discontinuation of FOBT, inpatients had a lower odds of undergoing endoscopic examination [odds ratio (OR) 0.80, 95% confidence interval (CI) 0.75-0.85]. Table 2 also provides data on the diagnosis stratified by FOBT status and whether endoscopic procedures were performed.

Table 3

Description of groups before and after discontinuation of FOBT

Variables	Before	SD	After	SD	P value
Number of admissions	23,162		24,061
Average age (years)	41.8	21.6	42.8	21.4	<0.0001
Males	11,503 (49.7%)		12,273 (51%)		0.0035
Incidence of anemia	2,435 (10.5%)		2,294 (9.5%)		0.0004
Average length of stay (days)	4.0	9.8	4.09	9.0	0.4
Number of procedures/admission	2710 (11.7%)		2311 (9.6%)		<0.0001*

*P value based on logistic regression adjusting for age, gender, and baseline anemia, SD = Standard deviation

Discussion

While FOBT is primarily indicated for CRC screening in average-risk individuals, our current cross-sectional study highlights that FOBT is still commonly used in the acute inpatient setting, and is often used for inappropriate indications. As expected, patients with positive FOBT results were more likely to undergo endoscopic examinations. However, the discontinuation of FOBT did not lead to significant changes in the clinical outcome.

Inpatient FOBT continues to be inappropriately used for reasons other than CRC screening in symptomatic individuals, with anemia being the most common indication. In a study from the Netherlands, a majority of inpatient FOBTs was ordered for anemia (41%) while suspicion of rectal bleeding (17%), abdominal pain (14%), even changes in bowel habits (10%) were among other reasons.[4] Narula et al. also found anemia to be the leading indication for FOBT (51%) while only one patient underwent FOBT for CRC screening.[7] However, in an Australian study 57% of the inpatient tests were ordered to evaluate GI bleeding while only 10% were ordered for anemia.[2] In their cohort, only 3% of the FOBTs were ordered for CRC screening.[2] These studies support Sharma et al. who in 2001 studied the implementation of FOBT in hospitalized patients by reviewing 1,000 randomly selected patients after discharge and discovered that 71% of FOBTs were considered inappropriate due to the patient's age, active GI bleed, or use on medication such as aspirin or other nonsteroidal anti-inflammatory drugs.[5]

Our study revealed that 23% of the FOBTs may have been ordered in patients with overt GI bleeding, which is significantly less than the 89% reported by Sharma et al. in 2001 and much higher than the 5% reported by Friedman et al. Our finding supports previous reports of FOBT misuse in the presence of overt GI bleeding. Our review of the medical chart did not clearly identify why FOBT was ordered despite overt GI bleeding in 23% of the patients. Highlighting the inappropriate use and reasons for inappropriate use of FOBT provide opportunities for targeted interventions for quality improvement.

Among patients with positive FOBT in the inpatient setting, only about one-third appear to undergo an endoscopic procedure (32%,[2] 38%,[4] 34%,[6] 34% in our study). These findings demonstrate that differences in the proportion of positive inpatient FOBTs do not necessarily translate into differences in the endoscopic procedures performed. However, a recent study of 637 anemic patients over 3-month period at two hospitals found that the only significant predictor of having an endoscopic evaluation of anemic patients, without acute GI bleeding, was a positive FOBT (OR 5.2,95% CI 1.7-16.2).[8] Our data also demonstrate that patients with positive FOBT were more likely to get a procedure. Moreover, the decrease in the number of endoscopies after discontinuation of FOBT also suggests that FOBT results may influence the need for endoscopy.

The impact of discontinuing FOBT in the acute hospital setting has not been well-studied. After correcting for multiple confounders, the discontinuation of inpatient FOBT at our hospital appears to have decreased the number of endoscopic procedures performed among inpatients. In the era following discontinuation of inpatient FOBT, inpatients were 20% less likely to undergo an inpatient endoscopy. However, the reduced number of procedures is likely affected by multiple confounders both measured and unmeasured, and it is unlikely that discontinuation of inpatient FOBT alone contributed to these trends observed. Furthermore, despite the results of FOBT, elderly patients with iron deficiency anemia warrant endoscopic evaluations, given the higher age-related risk of malignancies and the utility of FOBT to avoid endoscopic evaluation may be more relevant in lower-risk populations including premenopausal women.[9]

The strengths of this study include the evaluation of inpatient FOBT in an ethnically diverse patient population at an urban safety net hospital. The evaluation of FOBT discontinuation in the acute hospital setting, albeit with significant limitations, has not been well-studied. Our study had several limitations that should be acknowledged. The retrospective cross-sectional nature of our study limited the ability to evaluate longitudinal outcomes, and adequately assess and capture all potential confounders that could have impacted our outcomes of endoscopic procedures performed and length of hospital stay. The data included in our analyses were limited by the availability and accessibility of data from paper charts and electronic health records, which often lacked adequate granular details needed to perform more sophisticated analyses. Also, while the indication for FOBT was evaluated an indication was not identified in one-third of the patients, the results of whom could have significantly impacted our study outcomes. Furthermore, it was not clearly identified why FOBT was ordered in the 23% of the patients who had overt signs of GI bleeding. Since an inpatient gastroenterology consultant evaluated all patients undergoing endoscopy, it was difficult to assess for potential bidirectional biases resulting from the FOBT test results. In addition, while our patient population is unique, this also potentially limits the generalizability of our findings to a larger population cohort. Specifically, the utilization of FOBT in practice settings outside of the US may have a role in noninvasive evaluation or as an adjunctive evaluation tool for etiologies such as parasitic infections.

In summary, our study demonstrates that in a large urban safety net hospital, inpatient FOBT for reasons other than CRC occurs and the inappropriate use of inpatient FOBT is not uncommon. Implementing limitations on the inpatient use of FOBT may reduce the inappropriate utilization of this tool that was designed for CRC screening in average-risk adults. These findings highlight the need for educational interventions and quality improvement initiatives to improve the appropriate utilization of FOBT.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.

Table 2B

FOBT result stratified by medical diagnoses

Diagnosis	Positive FOBT	Number of endoscopy	Negative FOBT	Number of endoscopy
Abdominal pain	1	—	2	1
Alcohol related	3	2	7	1
Anemia	2	—	9	—
Biliary disease	0	—	2	—
Cardiac disease	8	1	17	—
Cellulitis	3	1	3	—
Diabetic complications	4	2	1	—
Diarrhea, nausea and or vomiting	4	—	3	1
Diverticulitis	0	—	2	—
Drug toxicity	0	—	1	—
DVT or PE	1	1	3	2
Failure to thrive	1	—	2	1
GI bleed	9	7	3	1*
GI Malignancy	2	1	0	—
Inflammatory bowel disease	1	1	1	1
Kidney or electrolyte abnormality	2	—	5	—
Liver failure- or cirrhosis-related	7	4	9	1
Neurological	3	1	15	—
Non-GI cancer	1	—	4	—
Other	3	1	7	—
Pancreatitis	0	—	1	—
Peptic ulcer	0	—	1	—
Perforated ulcer	0	—	1	—
Pulmonary or pneumonia	6	—	17	3
Rheumatologic	0	—	2	1
Sepsis	8	1	6	—
Small bowel obstruction	0	—	1	—
Urinary tract infection	1	1	3	1
Volume overload	0	—	3	—

DVT = Deep vein thrombosis, PE = Pulmonary embolism, Abdominal pain — Abdominal pain of unclear etiology; kidney or electrolyte — acute kidney injury, chronic kidney disease, electrolyte abnormality; alcohol related — alcohol withdrawal, alcohol intoxication, alcoholic hepatitis, alcoholic pancreatitis; rheumatologic — lupus, rheumatoid arthritis, scleroderma; cardiac disease — congestive heart failure, cardiac arrhythmia, acute coronary syndrome; biliary disease — cholecystitis, obstructive jaundice; diarrhea, nausea, and/or vomiting — diarrhea, nausea, and vomiting, nonspecific gastroenteritis; failure to thrive — failure to thrive of unclear etiology; GI bleed — occult gastrointestinal bleeding of undetermined etiology; pancreatitis — pancreatitis, nonalcohol-related; perforated ulcer — perforated gastric ulcer; pulmonary — pneumonia, acute respiratory distress syndrome; sepsis — severe sepsis, septic shock; neurologic — syncope, stroke, seizure; urinary tract infection — urinary tract infection, pyelonephritis; volume overload — volume overload of multifactorial etiologies