Measuring the clinical impact of pathologist reviews of blood and body fluid smears: a laboratory outcome study.
Objective.--To assess the clinical usefulness of pathologist review of blood and body fluid smears.
Design.--Survey study. Pathology residents contacted the ordering physician after pathologist reviews were reported to assess their clinical impact.
Results.--Ninety-six pathologist reviews met criteria for study inclusion, and 64 ordering physicians were successfully contacted during the 2-month study period. Of the 64 cases, 19 reviews (30%) had been seen by the physician within 24 to 48 hours after the report was issued and 33 (51%) had not been seen; in 4 (6%) instances, physicians did not remember whether they had seen the review. Eight reviews (13%) were considered urgent enough to warrant immediate communication by the pathologist. Of the 27 reviews that were seen or directly communicated, 23 (85%) contributed to clinical diagnosis and/or patient management.
Conclusions.--This study demonstrates the contribution of pathologist reviews of blood and body fluids to clinical diagnosis and patient management. The results also highlight the problem of a lack of physician awareness of clinical pathology results.
(Arch Pathol Lab Med. 2007;131:468-472)
Physicians rely on the laboratory for accurate interpretations of peripheral blood and body fluid smears. The College of American Pathologists laboratory accreditation program requires that laboratories establish specific criteria for peripheral blood smear review by a pathologist or a technologist with qualifications in laboratory hematology. (1) However, there is no explicit requirement that the pathologists' review of blood smear findings be reported. Pathologist reviews (PRs) of blood and body fluid smears serve several purposes. First, they fulfill an important quality assurance function for the hematology laboratory by providing a secondary review of significant abnormal findings. Second, they offer interpretive information that may be clinically useful. Third, the PR process is an essential component of residency training in laboratory hematology.
Despite wide acceptance of the PR process, little is known about what impact this laboratory practice actually has on improving patient care. We designed a study to evaluate the clinical usefulness of PRs of peripheral blood and body fluid smears in an academic medical center. The hematology laboratory at the University Hospitals of Cleveland has had criteria for PR of abnormal blood and body fluid smears for more than 20 years. The PR comments are appended to the leukocyte differential count for complete blood count and leukocyte differential count (CBC/DIFF) reports and body fluid cell counts in the laboratory information system (LIS). On average, 220 blood smears and 170 body fluid smears are reviewed monthly, which provides a substantial test volume on which to evaluate the clinical usefulness of the test.
MATERIALS AND METHODS
Criteria for PRs are defined for quantitative and qualitative abnormalities of CBC/DIFF parameters, and include age-related criteria for anemia, mean cell volume, and absolute lymphocyte counts. Pathologist reviews are "reflex-ordered" by the LIS through a bidirectional interface with the automated hematology analyzer for those samples that meet numeric criteria based on analyzer results. Pathologist reviews are manually accessioned by laboratory technologists on smears that meet qualitative criteria based on actual smear review. The result field, "PATHOLOGIST REVIEW: pending," appears in the initial cell count report for each PR that is ordered. Each weekday morning, clinical pathology residents review the slides and corresponding reports for each pending PR and obtain relevant clinical information from the LIS and the anatomic pathology information system. The slide reviews are then "signed out" with an attending hematopathologist, and interpretive comments are entered into the LIS in the PATHOLOGIST REVIEW field and are verified electronically by the attending pathologist. For results that are deemed urgent, the ordering physician (OP) is contacted directly by pager. On weekends, the "on-call" pathology resident reviews urgent blood and body fluid smears and consults with an attending pathologist, as needed.
The study involved human subjects research, and therefore, institutional review board approval was obtained with a waiver of informed consent. We anticipated that the majority of OPs, who are the subjects of the research, would be housestaff physicians in internal medicine and pediatrics. After obtaining approval from the residency program directors for these 2 departments, the principal investigators met with the housestaff at their morning reports to explain the study and to request their participation. An information sheet describing the study was distributed at that time, with an explanation that subject and patient identities would be protected by deidentification of the data.
For practical reasons, it was not possible to obtain clinical follow-up on all PRs. Therefore, we narrowed the PR criteria for study inclusion to focus on those that were most likely to have clinical significance (Table 1). Reactive neutrophilias, joint fluid crystal exams, newborn blood smears, cytopenias on known cancer chemotherapy patients, and negative body fluid smears were excluded. Pathologist reviews on patients with previously diagnosed hematologic diseases or previously described hematologic abnormalities were also excluded. During the 2-month study period (September 7 through October 30, 2005), cases that met the criteria for study inclusion were identified at daily sign-out. Pathology residents contacted the OP by alpha-numeric pager the day after the PR was reported. A maximum of 3 attempts to reach the OP were made on successive days, up to 48 hours after the PR was issued. For PRs performed on Fridays, calls were made the following Monday. It was not always possible to reach OPs from the Emergency Department (ED) the next day. In those cases, follow-up was attempted with the inpatient housestaff for those patients who were admitted to the hospital. Similarly, if a housestaff physician could not be reached due to "post-call" hours limitations, another housestaff physician on the same clinical service, or the attending physician, was paged. The interview consisted of the following 5 questions:
1. Did you see the PR on this patient sample?
2. Did the PR contribute to the clinical diagnosis?
3. Did the PR affect patient management?
4. Did the PR lead to the ordering of additional laboratory tests? If yes, what tests were ordered?
5. Did the PR result in the consultation of a subspecialty service? If yes, what service was consulted?
If the OP had not seen the PR, then the interview was stopped and the physician was informed of the content of the PR. Pathologist reviews that were considered urgent for patient care (eg, suspicious for leukemia or other serious blood disorder) were called directly to the OP at the time of sign-out, and the interview was conducted during the same telephone call.
There were 603 total PRs during the study period; 96 (16%) of these met the criteria for study inclusion (Figure 1). The distribution of PRs by diagnosis category is shown in Figure 2 and the distribution of PRs by physician type, patient type, and patient location is shown in Table 2.
[FIGURE 1 OMITTED]
Contact with an OP was made in 64 (67%) of 96 cases. Nineteen PRs (30%) had been read by an OP prior to being contacted by a pathology resident, 8 PRs (13%) were considered urgent enough to warrant immediate communication, 33 PRs (51%) had not seen by an OP at the time of contact, and in 4 instances (6%), the physicians did not remember whether they had seen the PR (Figure 3).
For the remaining 32 cases (33%), no contact was made with a physician. These instances involved cases where the physician did not return the pages after 3 attempts (10), a hematology/oncology consultation had already been requested before the PR was issued (8), the physician was on vacation (2), no OP was listed in the LIS (10), or the physician listed in the LIS as the OP denied knowledge of the patient (2).
The distribution of PRs that were seen by the OPs is shown in Table 3. Since the total number of cases is small, statistical significance was not evaluated. The highest proportions of PRs that were seen by physicians were for cytopenias (83%) and body fluids with suspicious or malignant cells (50%). A somewhat higher proportion of inpatient PRs was seen compared to outpatient PRs (39% vs 28%). There were 12 PRs performed on patients from the ED; 5 of these were from patients who were discharged from the ED and 7 were from patients who were admitted to the hospital. None of the PRs for the discharged patients was seen by a physician, whereas 4 of 7 of the PRs on ED patients who were admitted to the hospital were seen by a physician, generally from the inpatient service.
The 27 PRs that were read by clinicians (19) or directly called to them (8) could be evaluated for their clinical effect on patient care (Table 4). In 23 (85%) of these 27 cases, the OP responded that the PR contributed to the clinical diagnosis. The diagnoses for these patients included 7 new cases of acute leukemia, 6 cases of viral or bacterial meningitis, 3 cases of iron-deficiency anemia, 2 cases of microangiopathic hemolytic anemia, 1 malignant effusion, 1 case of bacterial sepsis, and 3 cases for which no specific diagnosis was made. Direct effects on patient management prompted by the PR were evident from 6 cases in which additional laboratory tests including serum iron studies and bone marrow examination were performed and 3 cases in which consultation by the hematology/oncology service was obtained. Of the 9 cases in which additional laboratory tests or hematology/oncology consultation were obtained, 5 involved PRs that were considered urgent and were called directly to the OP.
Although PRs of peripheral blood and body fluid smears are a well-established and widely accepted practice in hospital laboratories, few references in the published literature address the clinical usefulness of the test. (2-4) Javidian et al2 stated that the PR is an important aspect of quality assurance and of training pathology residents. However, another study demonstrated that the review of blood smears for anemia did not correlate with diagnostic accuracy or the number of tests that were subsequently ordered. (4) At an April 2005 conference, the Institute for Quality in Laboratory Medicine proposed evaluating clinician follow-up of laboratory test results as a quality indicator for clinical laboratories. This recommendation was the motivation for this laboratory quality outcome study.
The results of our study demonstrate the "added value" of PRs of abnormal blood and body fluid smears for patient care. A limitation of the study is the small sample size. Clinical pathology outcome studies are difficult to design and perform. Physicians are a notoriously difficult group to survey, and chart reviews are extremely labor intensive and time consuming. To minimize the imposition on clinicians and to maximize response rate, the study focused on a subset of PRs, and the study period was limited to 2 months in duration. In spite of these limitations, the value of the PRs is clear, as 85% of the PRs that were seen by a clinician were considered clinically useful. The number of PRs that were useful to clinicians during this time period might actually have been higher than our results suggest because we only followed up on a fraction (16%) of the total PRs performed during the study period. The large number of PRs that were not included in the study consisted of severe normocytic anemias, reactive neutrophilias, cytopenias on oncology chemotherapy patients, neonatal blood smears, joint fluids, negative body fluids, and repeat PRs. Many of these PRs might also have been clinically useful to the primary clinical team and consultants. In some cases, the value of the PR was immediately evident, such as the 7 new cases of acute leukemia and 1 case in which carcinoma cells were first detected in the pleural fluid cytospin. However, some PRs might be clinically useful even when they do not lead directly to a specific diagnosis because they help to exclude other possibilities. Although PRs can contribute to more prompt diagnoses and treatments for patients, it is not possible to determine from this study whether hospital length of stay was affected.
The results suggest that PRs for body fluids, unexplained cytopenias, and findings suggestive of myeloproliferative or myelodysplastic disorders may be more closely followed by clinicians than PRs performed for the other indications. This behavior makes sense because anemia is a fairly common finding in hospitalized patients, whereas leukopenia and thrombocytopenia are more worrisome findings that may signal previously unsuspected bone marrow disease or iatrogenic complications. Likewise, the presence of unsuspected blasts on a leukocyte differential count is a cause for great concern and is likely to alert a clinician to follow-up on the pending PR. When invasive procedures such as thoracentesis or spinal tap are done for the purpose of collecting diagnostic samples, physicians would be expected to follow-up on the final results. Also, when the differential diagnosis is contingent on specific blood or body fluid findings, the clinicians may look for the PR to confirm or exclude diagnoses.
It is a matter for concern that as many as 51% of PRs were not seen by an OP up to 2 days after the report was available in the hospital information system. It is possible that our results might underestimate the percentage of PRs that were actually seen by clinicians, as a member of the clinical team other than the OP might have seen some PRs. Several reasons for the OP not seeing the PRs were offered: (1) physician went off service, (2) "a nurse checks the labs," (3) the attending physician only checks laboratory results 2 days a week, (4) the person listed as the OP denied knowledge of the patient, and (5) ED patients are not followed up on by the ED physician. Some of these reasons are especially worrisome because they suggest the possibility that other laboratory data on patients might also be missed. These patterns of physician behavior highlight a major issue in pathology and laboratory medicine, namely a lack of clinician awareness of laboratory and pathology results, and emphasize the importance of calling in critical laboratory results.
The process of pathologist review contributes to training pathology residents in laboratory hematology, hematopathology, and cytopathology. Daily review of abnormal blood smears involves residents in the operations of the laboratory and requires them to integrate blood smear findings with automated analyzer outputs, technologist interpretations, and clinical information. A solid foundation in blood smear morphology is a prerequisite for bone marrow morphology and is essential to hematopathologic diagnosis. Many anatomic pathology-trained cytopathologists are not comfortable reviewing air-dried, Wright-stained cytospin preparations from the hematology laboratory. Pathology residents benefit from correlating cytospin smears prepared in the hematology laboratory with Papanicolaou-stained cytopathology smears. Our PR system also provides an opportunity for graded responsibility for residents, as required by the Accreditation Council for Graduate Medical Education. Finally, the PR system improves communication with clinicians about laboratory results and encourages medical staff to come to the laboratory to review smears with the hematopathologists, which they do regularly.
(1.) Sarewitz S, ed. Hematology and Coagulation Checklist. Northfield, Ill: College of American Pathologists; October 6, 2005:40.
(2.) Javidian P, Garshelis L, Peterson P. Pathologist review of the peripheral smear: a mandatory quality assurance activity? Clin Lab Med. 1993;13:853-861.
(3.) Peterson P, Blomberg DJ, Rabinovitch A, Cornbleet PJ, for the Hematology and Clinical Microscopy Resource Committee of the College of American Pathologists. Physician review of the peripheral blood smear: when and why--an opinion. Lab Hematol. 2001;7:175-179.
(4.) Simmons JO, Noel GL, Diehl LF. Does review of peripheral blood smears help in the initial workup of common anemias? J Gen Intern Med. 1989;4:473- 481.
Accepted for publication August 3, 2006.
From the Department of Pathology, University Hospitals of Cleveland, Cleveland, Ohio.
The authors have no relevant financial interest in the products or companies described in this article.
Reprints: Linda M. Sandhaus, MD, University Hospitals of Cleveland, Department of Pathology, 11100 Euclid Ave, Cleveland, OH 44106 (email: email@example.com).
Linda M. Sandhaus, MD; David N. Wald, MD, PhD; Kenan J. Sauder, MD; Erica L. Steele, DO; Howard J. Meyerson, MD
Table 1. Pathologist Review Criteria for Study Inclusion * 1. Anemia: microcytic anemia (MCV _ 70), macrocytic anemia (MCV > 110), or anemia with morphologic features of hemolysis 2. Lymphocytosis suggestive of viral infection or lymphoproliferative disorder 3. Findings suggestive of previously undiagnosed myeloproliferative disorder, myelodysplastic syndrome, or acute leukemia 4. Thrombocytopenia, neutropenia, or combined cytopenias in a nononcology patient 5. Cerebrospinal fluid or other body fluid with cells suspicious for malignancy 6. Cerebrospinal fluid with leukocytosis consistent with meningitis * MCV indicates mean cell volume. Table 2. Distribution of 96 Pathologist Reviews by Physician and Patient Type Percentage Physician type Housestaff 50 Attending 48 Nurse practitioner 2 Patient type Adult 72 Pediatric 28 Patient location Inpatient 55 Outpatient 45 Table 3. Distribution of 64 Pathologist Reviews (PRs) for Which Contact Was Made With an Ordering Physician * No. Seen/Total No. Called PRs (%) Directly Total (%) Adult 14/40 (35) 6 20/46 (44) Pediatrics 5/16 (31) 2 7/18 (39) Attending 8/25 (32) 6 14/31 (45) Housestaff 10/30 (33) 2 12/32 (38) NP 1/1 (100) 0 1/1 (100) A 4/24 (17) 1 5/25 (20) B 6/12 (50) 1 7/13 (54) C 5/6 (83) 2 7/8 (88) L 2/9 (22) 1 3/10 (30) M 2/5 (40) 3 5/8 (63) Outpatient 7/25 (28) 4 11/29 (38) Inpatient 12/31 (39) 16/35 (46) Total 19/56 (34) 8 27/64 (42) * NP indicates nurse practitioner; A, anemia; B, body fluid; C, cyto- penias; L, lymphocytosis; and M, myelproliferative disorder/myelodys- plastic syndrome. Table 4. Clinical Effects of 27 Pathologist Reviews Additional Pathologist Affected Tests Clinical Review Type Diagnosis Ordered Consultation Anemia 3/5 2 0 Body fluid 7/7 1 1 MPD/MDS * 5/5 3 2 Cytopenias 6/7 0 0 Lymphocytosis 2/3 0 0 Total 23/27 6 3 * MPD/MDS indicates myelproliferative disorder/myelodysplastic syndrome. Figure 2. Distribution of pathologist reviews by diagnosis category, N = 96. MPD/MDS, myeloproliferative disorder/myelodysplastic syndrome. MPD/MDS 14% Lymphocytosis and MPD/MDS 1% Lymphocytosis 18% Cytopenia(s) 15% Body fluid 19% Anemia 33% Figure 3. Distribution of 64 pathologist reviews for which contact was made with an ordering physician. In 13% of the cases, the pathologist reviews were considered urgent and were called directly to the ordering physician. Do Not Remember 6% Did Not See PR 51% Saw PR 30% PR Called by Pathologist 13%
|Printer friendly Cite/link Email Feedback|
|Author:||Sandhaus, Linda M.; Wald, David N.; Sauder, Kenan J.; Steele, Erica L.; Meyerson, Howard J.|
|Publication:||Archives of Pathology & Laboratory Medicine|
|Article Type:||Clinical report|
|Date:||Mar 1, 2007|
|Previous Article:||Comparison of ataxia-telangiectasia mutated protein expression in diffuse large B-cell lymphomas of primary central nervous system and non-central...|
|Next Article:||The impact of amino-terminal pro-brain natriuretic peptide testing on hospital length of stay and morbidity in patients with acute decompensated...|