From the Division of Infectious Diseases, Department of Medicine (Messrs Reiss and Labowitz and Dr Wormser), and the Department of Ophthalmology (Dr Forman), New York Medical College, Valhalla.
Color blindness is a common hereditary X-linked disorder.
To investigate whether color blindness affects the ability to detect the presence of blood in body fluids.
Ten color-blind subjects and 20 sex- and age-matched control subjects were shown 94 photographs of stool, urine, or sputum. Frank blood was present in 57 (61%) of the photographs. Surveys were done to determine if board-certified internists had ever considered whether color blindness would affect detection of blood and whether an inquiry on color blindness was included in their standard medical interview.
Color-blind subjects were significantly less able to identify correctly whether pictures of body fluids showed blood compared with non–color-blind controls (P = .001); the lowest rate of correct identifications occurred with pictures of stool (median of 26 [70%] of 37 for color-blind subjects vs 36.5 [99%] of 37 for controls; P<.001). The more severely color-blind subjects were significantly less accurate than those with less severe color deficiency (P = .009). Only 2 (10%) of the 21 physicians had ever considered the possibility that color blindness might affect the ability of patients to detect blood, and none routinely asked their patients about color blindness.
Color blindness impairs recognition of blood in body fluids. Color-blind individuals and their health care providers need to be made aware of this limitation.
COLOR BLINDNESS (dyschromatopsia) is a common hereditary X-linked disorder, with a prevalence of approximately 8% in white men.1,2 In addition, up to 0.7% of white women are color blind.2 Most individuals with hereditary color deficiency have defective perception of red and green.2 Color vision deficiency may also be acquired as a result of eye diseases such as optic neuritis.3
In 1996, Stiff and colleagues4 described 3 color-blind patients who presented with hematochezia due to colorectal disease. These researchers suggested that there had been a delay in seeking medical attention because the patients did not realize that they were bleeding because of their color blindness. In each of these cases, it was the patient's spouse who first noticed the blood.
Because of the medical significance of recognizing bleeding from the gastrointestinal tract in a timely fashion, and from other sites such as the genitourinary and respiratory tracts, we undertook a study to determine if color blindness does in fact impair the ability to detect blood in excretions and secretions (hereafter referred to as "body fluids"). In addition, we asked color-blind subjects if their health care providers routinely inquired about whether they were color blind or if they routinely volunteered this information during medical interviews. We also surveyed board-certified internists and gastroenterologists to determine if they had ever considered whether color blindness would affect detection of blood in body fluids and if they routinely inquired of their patients if they were color blind.
Color-blind subjects and non–color-blind control subjects were recruited by fliers and by e-mail postings at the Valhalla campus of New York Medical College. Color-blind subjects and controls were first tested using the Hardy, Rand, and Rittler pseudoisochromatic plates.5,6 Subjects were further tested with the Ishihara 14-plate test.7,8Color blindness on the Ishihara scale was defined as 2 or more mistakes on the first 11 plates, in accordance with the test's instructions. For inclusion into this study, color-blind volunteers had to show color deficiency on the Hardy, Rand, and Rittler and the Ishihara plates. To maintain consistency when viewing the photographs described later, color vision testing was done using a lamp with a 60-W soft white incandescent bulb. Color-blind subjects were classified as to red-green or blue-yellow color deficiency, and among those who were color deficient, as to severity based on the suggested criteria for interpretation of the Hardy, Rand, and Rittler diagnostic series of plates.5 Red-green color–deficient subjects were classified as having very mild, mild, or strong deficiency, and blue-yellow color–deficient subjects were classified as having medium or strong deficiency. Subjects with reproducible mistakes on the red-green or blue-yellow Hardy, Rand, and Rittler screening series of plates without errors on the diagnostic series were designated as having very mild red-green or blue-yellow deficiency, respectively.
Each color-blind subject was matched by age (within 5 years) and sex with 2 healthy control volunteers who were not color blind based on the testing described previously. Each color-blind and non–color-blind participant was shown the same series of 12.7 × 17.8-cm glossy photographs (N = 94). These photographs, arranged in random order, showed stool, urine, or sputum specimens, with and without the presence of fresh blood. The number of specimens that contained blood is as follows:
The specimens were photographed within collection containers (n = 57), on white tissue paper (n = 20), or for some stool or urine samples directly within a toilet bowel (n = 17). Generally, the blood was introduced into the samples after it was obtained from venipuncture of one of the investigators.
The subjects and control volunteers were informed that they would be shown a series of photographs of body fluids, some of which contained blood. The pictures were presented sequentially, and the participants were asked whether blood was visualized. The participants were given 30 seconds to respond. The number of correct and incorrect identifications was tabulated; a nonresponse was scored as an incorrect response. Testing was performed by either of 2 of the investigators (M.J.R. or D.A.L.) using a lamp with a 60-W soft white incandescent bulb (to simulate usual domestic lighting). The 94 pictures used were selected from a group of approximately 400 photographs. The photographs included were regarded as unambiguous as to the presence or absence of blood (Figure 1), based on a determination by the investigators (M.J.R., D.A.L., S.F., and G.P.W.) (who were not color blind) and on the results of a pilot study of 4 non–color-blind persons.
Illustrative photograph depicting a stool sample with frank blood. Four (40%) of the 10 color-blind subjects misidentified this photograph, compared with none of the 20 control subjects.
Each of the color-blind subjects was asked the following questions, and their answers were tabulated.
During the course of an interview with a health care provider (other than your eye doctor), have you been asked whether you are color blind? Possible answers included (a) all of the time, (b) occasionally, and (c) not that I can recall.
Before the testing done in this study, if you were asked by a health care provider about the presence of blood in stool, you would have (a) routinely mentioned that you are color blind or (b) mentioned that you are color blind only if specifically asked.
A separate written questionnaire was administered to all physician members of the general medicine and gastroenterology divisions of New York Medical College who were in attendance at a divisional staff meeting. All of the physicians present agreed to participate, but results were tabulated for only those who were board certified. The questions asked included the following:
Before this questionnaire, did you have an opinion as to whether color-blind individuals would be able to detect the presence of blood in stool reliably? Possible answers included (a) never considered the issue, (b) assumed that they would be able to detect the presence of blood most of the time, (c) assumed that they would not be able to detect blood reliably, or (d) other.
During the course of medical interviews, do you inquire about color blindness? Possible answers included (a) routinely, (b) occasionally, or (c) almost never.
During medical interviews, have patients mentioned to you that they are color blind? Possible answers included (a) quite often, (b) possibly once or twice, (c) not that I recall, or (d) other.
When you ask a patient about the presence of blood in stool and a negative response is given, (a) do you routinely ask if the patient is color blind? (yes or no) and (b) have any patients mentioned spontaneously that they are color blind? (yes or no).
If yes to 4b, how often has this occurred?
Continuous variables that were normally distributed were assessed with the t test or a 1-way analysis of variance. Skewed variables were assessed with the Mann-Whitney test or the Kruskal-Wallis test. Categorical variables were compared with the Pearson χ2 or the Fisher exact test. All tests were 2-tailed, and P<.05 was considered statistically significant. Statistical software (Statistical Product and Service Solutions, version 9.0; SPSS Inc, Chicago, Ill) was used for these analyses. We also performed an additional analysis that took into account the matching and found the results to be similar to those of the unmatched analysis. For simplicity, only the results of the unmatched analysis are presented.
Based on data from a pilot study, it was calculated that 20 controls and 10 color-blind participants would be required to have 90% statistical power. It was assumed that the control group would correctly identify an average of 90% and that the color-blind participants would identify an average of less than 80% of the photographs. The common SD was estimated to be 10%. The α level was .05, and this was a 1-tailed test. Sample size calculations were performed with a software package.9
Ten subjects who believed that they were color blind participated in this study. All 10 were confirmed to be color blind based on our testing, including 9 men and 1 woman. Of these 10 individuals, 5 were physicians, medical students, or other health care workers. For each subject with a color vision deficiency, 2 control volunteers were recruited who were of the same sex and within 5 years of the same age. Fourteen (70%) of the 20 control volunteers were physicians, medical students, or other health care personnel (P = .43 for the comparison with color-blind subjects of the proportion who were health care workers). All color-blind and control participants were white adults (age range, 22-75 years) (Table 1).
The control group was significantly (P = .001) more likely to identify correctly whether the pictures showed blood compared with the color-blind subjects (Table 2). The median percentage correct in the control group was 98.9% (95% confidence interval [CI], 97.7%-99.4%) compared with 77.1% (95% CI, 73.2%-88.8%) for the color-blind subjects (P<.001). The control group was able to identify correctly 98.4% (1122/1140) (95% CI, 97.5%-99.1%) of the photographs in which blood was depicted and 98.6% (730/740) (95% CI, 97.5%-99.3%) of the photographs in which blood was not depicted, compared with 79.8% (455/570) (95% CI, 76.5%-83.1%) and 82.7% (306/370) (95% CI, 78.8%-86.6%), respectively, for the color-blind volunteers (P<.001).
For both groups, the lowest rate of correctly identified photographs occurred with pictures showing stool samples compared with other body fluids (median percentage correct in the control group was 99% with stool vs 100% with either urine or sputum [P<.01]; and median percentage correct in the color-blind group was 70% with stool compared with 90% with urine and 78% with sputum [P<.02]). The control group was significantly more accurate than the color-blind subjects regardless of whether the background in the photograph was a collection container, white tissue paper, or a toilet bowl (P<.001).
Six of the color-blind subjects were classified as having very mild red-green color blindness (one of whom also had blue-yellow deficiency), one was mildly deficient, and 3 were regarded as strongly color blind (Table 3). Subjects with even very mild red-green color deficiency were significantly less accurate than controls in identifying the photographs correctly (P = .002). Moreover, the more severely color-blind volunteers were significantly less likely to identify the pictures correctly compared with those classified as having very mild deficiency (P = .009).
Based on the written survey responses of a convenience sample of 16 board-certified general internists, 0% (95% CI, 0%-20.6%) routinely inquire about color blindness during medical interviews despite negative responses to questions on whether the patient had observed blood in body fluids. Similarly, among a convenience sample of 5 board-certified practicing gastroenterologists, all indicated that they almost never ask patients about color blindness. Based on the survey, only 2 (10%) (95% CI, 1.2%-30.4%) of the 21 physicians had ever considered the possibility that color blindness might affect the ability of patients to detect blood in body fluids. Furthermore, both groups of physicians indicated that patients rarely volunteer that they are color blind. These responses were corroborated by the color-blind subjects. Of the 10 color-blind subjects, 0% (95% CI, 0%-30.8%) ever recalled a health care provider (other than an ophthalmologist) inquiring about color blindness during a medical interview. When answering questions about the presence of blood in stool, only 2 of the 10 had ever mentioned that they were color blind without being specifically asked.
Questions concerning the presence of blood in body fluids are routinely included in the "review of systems" section of the medical history. These questions are important because bleeding may be the first sign of an otherwise asymptomatic but serious medical condition. In one study10 of 103 patients aged 45 years and older with hematochezia, subsequent evaluation revealed a colorectal neoplasm in 41%, including carcinoma in 4%. Similarly, important medical diagnoses, such as genitourinary tract malignant neoplasms, may be present in asymptomatic patients with hematuria.11 Among the leading causes of hemoptysis is bronchogenic carcinoma.12
Color blindness is a common hereditary disorder in men. Most color-blind persons (up to 98%) have red-green color deficiency of varying degrees.2 This study has demonstrated that red-green color-blind persons are significantly less able to identify whether photographs of stool, urine, or sputum depict blood compared with non–color-blind age- and sex-matched controls. For example, of 37 photographs of stool (20 of which showed blood), the median percentage of incorrect responses among the color-blind subjects was approximately 30%, compared with 1% for the controls (P<.001) (Table 2). The color-blind subjects were significantly better able to discriminate among pictures of urine (median, 10% incorrect responses; P = .005) and sputum (median, 22% incorrect responses; P = .01), presumably because of the greater contrast of blood in these body fluids compared with stool.
Compared with other studies,13 a greater proportion of our color-blind subjects (50%) were classified as having only very mild red-green color deficiency. This might be attributed in part to our use of tungsten illumination, which could lead to an underestimation of the degree of color vision deficiency.14 Our study does not permit any conclusion on the effect of blue-yellow color blindness on detection of blood in body fluids, since the single subject who had this deficiency also had a red-green deficiency.
Additional limitations of this study are that photographs rather than actual specimens were used and that blood was artificially introduced into the specimens. Furthermore, more than 50% of the participants were health care workers who may be better able to recognize blood in body fluids than other individuals.
The exact frequency with which color-blind individuals (or controls) would fail to recognize blood in body fluids under natural conditions cannot be determined from this study. It is likely, however, that they would recognize blood less well outside of a standardized testing environment. In our study, certain advantages were afforded the volunteers: they were informed that the photographs may contain blood, the pictures were explicit, fresh red (rather than old) blood was added to the samples of body fluids, and the photographs were reviewed under good lighting.
A survey of a convenience sample of 21 board-certified internists and gastroenterologists at our center revealed that only 2 (10%) (95% CI, 1.2%-30.4%) had ever even considered the possible impact of color blindness on the ability of patients to recognize blood in body fluids, and they rarely asked patients during medical interviews if they were color blind. Furthermore, color-blind patients rarely volunteer that they are color blind during medical interviews unless specifically asked. These findings are likely to be similar at other medical centers since questions on color blindness are not recommended by many leading textbooks of physical diagnosis.15- 17
The results of this study suggest that inquiries about color blindness (and possibly testing for color blindness) should be added to the standard medical interview. For color-blind patients, the medical history is liable to be inaccurate regarding the presence of blood in body fluids. Color-blind patients, particularly those who have a bleeding diathesis due to illness or medications or those recognized to have a gastrointestinal, genitourinary, or pulmonary disease potentially associated with bleeding, should be made aware that color blindness may impair the ability to recognize blood. Regular laboratory testing of pertinent body fluids for blood should be considered for such individuals.
Accepted for publication July 28, 2000.
We thank Dan Byrne, Robert B. Nadelman, MD, Stephen Peterson, MD, Edward Lebovics, MD, Eleanor Bramesco, and Lisa Giarratano for their assistance.
Corresponding author and reprints: Gary P. Wormser, MD, Division of Infectious Diseases, Department of Medicine, Room 209SE, Macy Pavilion, Westchester Medical Center, New York Medical College, Valhalla, NY 10595.
Thank you for submitting a comment on this article. It will be reviewed by JAMA Internal Medicine editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Register and get free email Table of Contents alerts, saved searches, PowerPoint downloads, CME quizzes, and more
Subscribe for full-text access to content from 1998 forward and a host of useful features
Activate your current subscription (AMA members and current subscribers)
Purchase Online Access to this article for 24 hours
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 7
Customize your page view by dragging & repositioning the boxes below.
and access these and other features:
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.