Clinical Consequences of Specimen Rejection: A College of American Pathologists Q-Probes Analysis of 78 Clinical Laboratories

Participants prospectively reviewed consecutive blood and urine specimens submitted to the central chemistry and/or hematology (including coagulation) laboratory sections to identify specimens that were rejected for any reason. This process was conducted on all laboratory shifts until 80 rejected specimens that fit the study inclusion criteria were identified or until 6 weeks elapsed, whichever came first. Definitions of terms provided to participants are listed in Table 1.

Specimens included in the study were those received from inpatients and emergency department patients older than 6 months for on-site analysis. The following specimen types were excluded from the study: point-of-care, blood bank, bone marrow, specimens received in the form of pre-prepared smears, and specimens submitted for highly specialized testing (ie, flow cytometry, molecular DNA/RNA, cytogenetic analysis).

Participants provided the total number of chemistry and hematology accessions separately for the study period. For each rejected specimen included in the study, the following information was recorded: patient age, specimen type, laboratory testing section, test priority, time of specimen receipt and time of specimen rejection, rejection reason, detection method used for rejection due to improper labeling, whether or not a recollected or relabeled specimen was requested by the laboratory, time of recollected/relabeled request, time of recollected/relabeled specimen receipt, recollection method for blood and urine specimens, if applicable, and result time.

Participants were also asked to provide information regarding their laboratory's policies and practices in the following areas: presence of a detailed written policy regarding the detection and/or handling of rejected specimens, method(s) used to detect improperly labeled specimens, whether or not improperly labeled blood and/or urine specimens are allowed to be relabeled, and the nature of recommendation(s) made to providers when a specimen is rejected.

Five performance indicators were calculated for this study: specimen rejection rate, median processing delay, percentage of laboratory-abandoned specimens, percentage of provider-abandoned specimens (see Table 1), and percentage of relabeled specimens subsequently determined to not be from the intended patient (wrong specimen in container). Median processing delay was defined as the interval between the original and recollected/relabeled specimen receipt times.

Individual associations between the first 4 performance indicators with demographic and practice variables were tested by using Kruskal-Wallis tests for discrete-valued independent variables and regression analysis for the continuous independent variables. The 2 abandoned rates were skewed so a log transformation was used for the regression-based analyses.

Variables with significant associations for the individual associations (P < .10) were then included in a forward selection multivariate regression model. A significance level of .05 was used for the final multivariate models. All analyses were run by using SAS 9.1 (SAS Institute, Cary, North Carolina).

Table 2 lists the characteristics of the 78 institutions submitting data for this study. A total of 4794 rejected specimens were identified from 2 054 702 accessions. Most of the participating institutions (94%) are located in the United States, with the remaining in Saudi Arabia (2), Canada, Jordan, and Spain. Of the participating institutions, 39% are teaching hospitals and 27% have pathology residency training programs.

Table 3 illustrates the distribution of 4 of the 5 performance indicators among participants. The overall specimen rejection rate was 0.2%, with the 90th percentile at 0.1%, median at 0.3%, and 10th percentile at 1.5%. The overall median specimen processing delay due to specimen rejection was 65 minutes, with a median of 54 minutes for stat specimens and 88 minutes for nonstat specimens. Of the 4794 rejected specimens, 537 (11.2%) were abandoned by either the laboratory not requesting a recollected or relabeled specimen (laboratory abandoned) or by the provider or caregiver team not complying with the laboratory request (provider abandoned). The median laboratory abandonment rate was 1.3%. The median provider abandonment rate was 5%. Nonstat and urine specimens were significantly more likely to be abandoned, with approximately equal likelihood for laboratory- and provider-abandoned specimens.

Table 4 details information relating to the rejected specimens. A substantial majority (96.5%) of rejected specimens were blood, with essentially equal numbers of chemistry and hematology specimens. The reason for most rejected specimens (92.4%) was inappropriate/inadequate specimen. The remaining 7.6% were rejected owing to improper labeling. Regarding specimens found to be improperly labeled, most (73.6%) were detected by laboratory review (“lab check”), 9.9% by feedback from caregiver, and 9.1% by delta check (see Table 1). Most rejected specimens (87.7%) were ultimately recollected, 1.1% were relabeled/corrected, and 11.2% were abandoned (neither recollected nor relabeled/corrected).

The first 4 performance indicators were tested for associations with institutional demographic and laboratory policy variables. A P value of less than .05 was considered statistically significant. Institutional and/or laboratory policy relationships were found with 3 of the first 4 performance indicators. No relationships were found with the rate of specimen rejection. The significant associations, detailed in Table 5, were as follows: longer specimen processing delays tended to occur in institutions with higher occupied bed size; higher rates of laboratory-abandoned specimens tended to occur in laboratories that communicate to the patient's caregiver that the specimen is rejected without providing any direction for specimen replacement; and higher rates of provider-abandoned specimens tended to occur in institutions that train pathology residents/fellows or that do not allow improperly labeled chemistry or hematology blood specimens to be relabeled without requiring the specimen to be recollected.

Table 6 summarizes responses received from participants regarding their policies and practices related to detection and resolution of rejected specimens. Of the participating laboratories, 45% reported that they allow improperly labeled blood specimens (see Table 1) to be relabeled (among these, 85% when incompletely labeled, 59% when mislabeled, and 38% when unlabeled). Likewise, 37% allow improperly labeled urine specimens to be relabeled (among these, 89% when incompletely labeled, 57% when mislabeled, and 36% when unlabeled). Regarding the laboratories that allow relabeling of specimens and the relabeled specimens included in this study, the following were found: (1) These laboratories reported no better performance regarding specimen processing delay, but a slightly lower rate of provider-abandoned specimens (see Table 5); (2) of the 51 improperly labeled specimens that were relabeled, 20 specimens (39%) were subsequently discovered to not be from the intended patient (wrong specimen in container). Of the 6 laboratories that reported the 20 relabeled specimens with “wrong specimen in container,” 4 allowed relabeling of unlabeled blood and/or urine specimens, and 2 stated that they did not allow relabeling of any improperly labeled specimens (ie, there was an apparent breach of policy). Among the 51 improperly labeled specimens allowed to be relabeled, there was no statistically significant difference in the rate of relabeling between stat and routine test priorities.

Defects in laboratory specimens that result in specimen rejection are a cause of inconvenience and discomfort to patients, lead to a delay in the availability of often critical laboratory results, and, if the methods used to detect specimen defects are not effective, may also represent a major compromise of patient safety. To address the potential harm to patients from misidentification, including mislabeling of laboratory specimens, the Joint Commission⁵  has for the past several years designated accurate patient identification as one of its national patient safety goals.

This study clearly documents the consequences of laboratory specimen defects and resulting specimen rejection. Prospective analysis performed by 78 institutions of a total of 2 054 702 specimen accessions revealed an overall specimen rejection rate of 0.2%. This is generally in line with the rate reported in past studies of specimen rejection,^1–4  which have typically shown rates of less than 0.3% and up to 0.75%, with lower rejection rates reported in more recent studies.^3,4 

The first and most direct consequence of specimen rejection is the need to collect a new specimen from the patient. In this study, 86.8% of rejected blood specimens led to repeated phlebotomy. Likewise, 13.8% of rejected urine specimens required recatheterization of the patient to collect a new urine sample. Both of these repeated procedures are associated with inconvenience and discomfort for the patient, and both introduce the potential for patient complications.

Another well-known consequence of specimen rejection is a delay in the performance and reporting of the results of the ordered tests. In this study, the median specimen processing delay was 65 minutes, with a range of 41 minutes (90th percentile) to 93 minutes (10th percentile). The median processing delay was longer for tests with routine priority (88 minutes) but was still 54 minutes for stat tests. This indicates that delays in availability of urgent and sometimes critical test results of nearly a full hour are commonly seen when stat specimens must be rejected. An interesting association with even longer specimen processing delay (median = 91 minutes) was hospital occupied bed size, specifically with bed size of greater than 450 beds. No definite explanation for this association was apparent; however, this may be a function of the greater complexity of very large hospitals, which could contribute to less effective communication among patient care staff members and slower response time to special requests, such as for recollection of a rejected laboratory specimen. Although not specifically assessed in this study, longer delays in test result availability in the largest hospitals, where the clinical acuity of patients is often higher, may have an even greater negative impact on patient safety and on the delivery of care to a larger number of critically ill patients.

An additional significant consequence of laboratory specimen rejection is abandonment of the ordered test(s). Such abandonment may occur when the laboratory fails to request recollection or relabeling/correction of a defective specimen (laboratory abandonment) or when the provider or patient care team fails to comply with such a request (provider abandonment). The overall specimen abandonment rate in this study was 11.2%, meaning that these patients received no result(s) for the test(s) originally ordered. The laboratory abandonment rate was lower (median rate = 1.3%) than when the provider abandoned the specimen (median rate = 5%), but still significant. Specimen abandonment rates were significantly higher for all nonstat tests and for urine specimens. The association with nonstat priority may be a simple function of lower clinical urgency and a decision to wait for the next routine phlebotomy session or next urine collection opportunity to recollect a rejected specimen. Higher overall abandonment of urine specimens is likely a reflection of the difficulty in having patients produce urine samples on demand or in recollecting catheter urine specimens. Another association with higher specimen abandonment, this time by the provider, was institutions that train pathology residents/fellows. Similar to the longer specimen processing delay seen in the largest hospitals, the higher rate of provider specimen abandonment may be related to the complexity and communication challenges typical in academic medical centers.

A troubling finding in this study was the very high rate of mislabeled specimens that were the result of allowing providers and other patient care team members to relabel specimens that were received in the laboratory improperly labeled. Of the participating institutions, 45% allow relabeling of blood specimens and 37% allow relabeling of urine specimens. Among the institutions allowing specimen relabeling, this was reported to be allowed at higher rates for incompletely labeled specimens, but 59% and 38% of participating institutions allow correction of mislabeled or unlabeled blood specimens, respectively. Among these institutions, the percentage allowing correction of mislabeled or unlabeled urine specimens was similar (57% and 36%, respectively). When asked to indicate how many relabeled specimens ultimately turned out to not be from the intended patient (ie, wrong specimen in container), respondents reported that 39% of such specimens had this outcome. Though this may represent a slight overestimate of the true risk of a “wrong specimen in container” collection event (since multiple blood tubes may have been collected at the same time, but given different accession numbers and thus considered separate events), the incidence of “wrong blood in container” in this study is clearly concerning. An interesting sidelight to this finding was the fact that those institutions that allow specimen relabeling reported no significant shortening of the specimen processing delay resulting from specimen rejection, although they did report a slightly lower rate of specimen abandonment by the provider. Thus, allowing specimen relabeling resulted in little measurable advantage, yet this practice clearly introduced an increased risk to patient safety. These findings represent a strong indictment of the practice of routinely allowing relabeling of improperly labeled specimens. This practice has been increasingly discouraged in recent years,^6–8  a recommendation convincingly reinforced by this study.

In summary, specimen rejection leads to a high rate of recollection of specimens, including repeated phlebotomy and recatheterization for urine, causing discomfort and potential complications in the affected patients. A major consequence of specimen rejection is a significant delay in availability of test results, of particular importance with stat tests and a situation potentially made worse in large and/or teaching hospitals, where a higher number of critically ill patients typically receive care. Specimen rejection also leads to frequent abandonment of the test(s) originally ordered, more so with nonstat tests and urine specimens. Finally, the practice of allowing relabeling of improperly labeled specimens is associated with little clinical benefit but introduces a significant likelihood of specimen mislabeling and potential harm to the patient. If laboratories choose to allow relabeling of blood and urine specimens, they should construct their policy thoughtfully, considering the relative risks and benefits, and perhaps restrict relabeling to correction of only the simplest labeling defects. The policy should also address relabeling of specimens that either cannot be recollected (eg, surgical pathology specimens), or that cannot be recollected without potential harm to the patient (eg, cerebrospinal fluid or other body cavity fluid specimens, bone marrow aspirate/biopsy specimens).