Development, Implementation, and Impact of Acceptability Criteria for Serologic Tests for Infectious Diseases

Abstract

Serologic testing is essential for the diagnosis of some infectious diseases and yet is fraught with potential pitfalls. All parts of the diagnostic process must be optimized to ensure that serologic tests perform adequately. Recognizing that a lack of clinical data and correctly timed, paired sera frequently led to uninterpretable serology results at our laboratory, we developed and implemented simple acceptability criteria for serologic tests. We assessed the impact of these criteria by comparing submissions and results for the year before and the year after implementation of the criteria. The number of serologic tests performed declined by 25% after implementation of the acceptability criteria, despite an increase in requests for serologic tests. Inappropriate testing of acute-phase sera alone fell from 49 to 0% (P < 0.001) for the tests monitored. Appropriate submission of paired sera rose from 9 to 19% (P = 0.006). The proportion of results classified as interpretable rose from 52 to 100% (P < 0.001). We recommend that acceptability criteria be developed and applied to samples submitted to clinical microbiology laboratories for serologic testing.

Serologic testing is central to the diagnosis of some acute, recent, or chronic infectious diseases. Sometimes serologic testing must be relied upon as the only diagnostic test that is practical, because the suspected etiologic agent is impossible, difficult, or dangerous to grow in cultures in a routine diagnostic laboratory or because colonization and disease cannot readily be distinguished by classical microbiological methods. However, failure to optimize all parts of the diagnostic process can render the results of serologic testing difficult or impossible to interpret.

Criteria that guide the decision to test various specimens submitted to clinical microbiology laboratories have been developed and applied in several areas of bacteriology and parasitology, including those for processing cerebrospinal fluid (1, 7), stool (2, 5), and sputum (4, 6). These have been associated with time and cost savings (3). Such measures also improve the utility of data provided to physicians by the clinical laboratory.

We observed that samples submitted to our laboratory for serologic testing for infectious diseases frequently lacked adequate clinical data for the interpretation of results and that correctly timed, paired samples were often not supplied. To address this, we developed, implemented, and assessed the impact of acceptability criteria for serologic tests for infectious diseases. We studied all serum samples submitted for serologic tests conducted on behalf of Duke University Medical Center patients by the North Carolina State Laboratory of Public Health and the Centers for Disease Control and Prevention (Table 1) between 1 December 1998 and 30 November 2000.

TABLE 1.

Seriologic tests studied

Test category	Specific infections	Testing of acutephase sample alone acceptable?	Testing of convalescent- phase sample alone acceptable?	Are paired sera the “gold standard”?
Bacterial agglutinins panel	Brucella abortus, Francisella tularensis	No	Yes	Yes
Rickettsial panel	Ehrlichia chaffeensis, Anaplasma phagocytophilum, Rickettsia riskettsii, Rickettsia typhi	No	Yes	Yes
Arboviral panel	Eastern equine encephalitis, La Crosse virus, St. Louis encephalitis, Western equine encephalitis	No	Yes	Yes
Bacterial respiratory panel	Chlamydia psittaci, Coxiella burnetti, Legionella, Mycoplasma pneumoniae	No	Yes	Yes
Viral respiratory panel	Adenovirus group, influenza virus A/B, parainfluenza virus 1-3, respiratory syncytial virus	No	Yes	Yes
Parasitic	Babesia, Echinococcus, Entamoeba histolytica, Leishmania, Paragonimus westermanii, Plasmodium, Taenia solium, Toxocara, Trichinella spiralis, Trypanosoma cruzi, Schistosoma, Strongyloides	No	Yes	Noa
Others	Bartonella, Chlamydia trachomatis, dengue, hantavirus, Japanese encephalitis, Leptospira, lymphocytic choriomeningitis, measles, mumps	Nob	Yes	Yes

^aPaired sera recommended for individuals from areas of endemicity.

^bTesting acute sample alone acceptable for measles.

The findings described in this work were presented in part at the 41st Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, Ill., abstract D-1400, 18 December 2001.

Simple acceptability criteria for these tests were developed on the basis of the recommendations of the manufacturer and the performing laboratory. According to those recommendations, seriologic assays are performed only when the following conditions are satisfied: (i) the date of onset is provided and (ii) a correctly timed acute-phase sample (in cases in which a reliable immunoglobulin M assay is available), or acute- and convalescent-phase serum samples (in cases in which they are needed or recommended), or a convalescent-phase sample alone (in cases in which acute-phase sera are not available and results can be interpreted in the clinical context on the basis of the date of onset) is submitted.

The acceptability criteria were introduced on 1 December 1999 in conjunction with notification and education of medical center staff in the monthly clinical microbiology newsletter. Each sample submitted during this second year of the study was reviewed for compliance with the acceptability criteria. For each serum sample, the clinical microbiology laboratory issued an interim report to the clinician indicating that an onset date and/or a convalescent-phase serum sample was needed before the serologic test would be done. Exceptions to the acceptability criteria were considered only after consultation with, and approval by, the Medical Microbiology Fellow, Clinical Microbiology Faculty, Serology Section Head, or Laboratory Director. Such exceptions were tracked in a logbook. Samples not meeting acceptability criteria were not forwarded for testing but were held for 3 months before being discarded.

To assess the impact of the acceptability criteria, we compared the number of test submissions and the proportions of rejected, inappropriately tested, positive, and interpretable tests during the year prior to the intervention compared to the year after the intervention. Statistical analyses were done using Epi Info 6.04 software (Centers for Disease Control and Prevention, Atlanta, Ga.).

We rejected 41% of tests submitted during the postintervention year, because they did not meet acceptability criteria (Table 2). Requests for rickettsial, Legionella, Bartonella, and arbovirus serology were most frequently rejected (Table 3). No exceptions to the acceptability criteria were required during the year after implementation. The number of serologic tests performed declined by 25% after implementation of the acceptability criteria, despite an increase in requests for serologic tests. Inappropriate testing of acute-phase sera alone fell from 49 to 0% (P < 0.001). Appropriate submission of paired sera rose from 9 to 19% (P = 0.006). The proportion of results classified as interpretable rose from 52 to 100% (P < 0.001).

TABLE 2.

Impact of acceptability criteria

Time period	No. of sera submitted	No. (%) of sera rejecteda	No. (%) of acute-phase sera inappropriately testeda	No. (%) of paired sera submittedb	No. (%) of positive resultsc	No. (%) of inter- pretable resultsa
Preintervention	204	0 (0)	100 (49)	18 (9)	16 (8)	106 (52)
Postintervention	262	108 (41)	0 (0)	29 (19)	17 (11)	154 (100)

^aP, <0.001.

^bP, 0.006.

^cP, not significant.

TABLE 3.

Serologic test types rejected

Assay	No. of sera rejected/total no. of sera (%)
Adenovirus group	2/2 (100)
Arboviral panel	5/15 (33)
Bartonella	8/23 (35)
Brucella	1/23 (4)
Chlamydia trachomatis	4/8 (50)
Coxiella burnetti	2/6 (33)
Ehrlichia chaffeensis	12/24 (50)
Francisella tularensis	2/8 (25)
Legionella	9/13 (69)
Leptospira	2/11 (18)
Rocky Mountain spotted fever	54/95 (57)
Strongyloides	4/7 (57)
Toxocara	2/7 (29)
Trichinella spiralis	1/3 (33)

It is our impression that samples submitted for serologic testing might not receive the scrutiny given other types of specimens received by the clinical microbiology laboratory. Consequently, opportunities to improve the value of the test to the clinician and patient are lost. Furthermore, inappropriate testing of samples may result in diagnostic errors and waste of healthcare resources, including taxpayers' money, for state and federal laboratories. After documenting that a lack of clinical data and correctly timed, paired sera frequently led to uninterpretable serology results at our laboratory, we developed and implemented simple acceptability criteria for serologic tests for infectious diseases. By comparing submissions and results for the year before and the year after implementation of the criteria, we showed improved correct submission of specimens and improved interpretability of results. These changes promote improved patient care and judicious use of healthcare resources. This study was not designed to determine whether diagnoses were missed because samples were rejected. However, considerable efforts were made to educate healthcare providers and to encourage consultation with laboratory staff about specific diagnostic questions in individual patients.

Although serology for establishing immune status respecting an infectious agent is relatively straightforward, its use for the diagnosis of acute, recent, or chronic infections is fraught with pitfalls. Consequently, all parts of the diagnostic process must be optimized to ensure that the serological test performs adequately. In each case an assessment needs to be made of the value of a serologic test compared with other available diagnostic methods. If a serologic test is found to be both indicated and useful, adequate clinical data need to be supplied. The onset date is vital for interpretation of a result, appropriate handling of the sample in the laboratory, and selection of the best assay and testing strategy. Other clinical data are necessary to determine pretest probability of disease (e.g., travel and exposure history) and causes of potentially misleading results (e.g., vaccination and past medical history). A frequent deficiency is the failure to submit both acute- and convalescent-phase sera when they are needed.

We recommend that acceptability criteria should be developed and applied to samples submitted to clinical microbiology laboratories for serologic testing. Our acceptability criteria can be easily implemented when done in conjunction with education of health providers and clear lines of communication for consultation about individual patients. Criteria should be reviewed regularly to adapt them to inevitable changes in serologic assays, particularly those conducted at reference centers. Reference centers should inform the referring laboratories of changes in the tests that they provide and indicate the optimal sample requirements for each assay. Future refinements might include closer tailoring of criteria to individual assays and requirements for more detailed clinical information that would allow thorough evaluation of pretest probability before each assay is conducted.