Time-Motion and Cost Comparison Study of Micro-ID, API

Transcription

1 JOURNAL OF CLINICAL MICROBIOLOGY, Dec 1981, p /81/ $02.00/0 Vol. 14, No. 6 Time-Motion and Cost Comparison Study of Micro-ID, API 20E, and Conventional Biochemical Testing in Identification of Enterobacteriaceae MARTHA J. BALE' AND JOHN M. MATSEN '2* Departments of Pathology2* and Pediatrics,' University of Utah School of Medicine, Salt Lake City, Utah Received 26 March 1981/Accepted 30 June 1981 A total of 730 Enterobacteriaceae strains isolated from 567 cultures were evaluated by a rapid kit method (Micro-ID; General Diagnostics, Morris Plains, N.J.; 4 h), an overnight incubation kit method (API 20E; Analytab Products, Plainview, N.Y.), and conventional biochemical test methodology (mostly overnight incubation and some rapid methods) to compare the amount of laboratory effort required, timing, and cost parameters. We assessed the amount of technologist time expended, the time sequence of culture reporting to physicians, the number of isolates requiring repeat testing or additional biochemical testing, the number of cultures held due to the need for identification of other organisms, the cost of total work-up, etc. Cultures evaluated included urines, respiratory cultures, wound cultures, body fluids, genital cultures, and cultures from miscellaneous categories. A total of 64% of the Enterobacteriaceae strains processed by the Micro-ID method could be identified within 24 h of receipt of the specimens in the clinical laboratories, in contrast to the need for an additional day required by the API or conventional biochemical methods. The Micro-ID method also required less technologist time (4.5 min) for set-up and interpretation than did either the API method (6 min) or conventional methods (7 min). Total direct costs (June 1981) per organism identified were: Micro-ID, $4.30; API 20E, $4.96; conventional biochemicals with commercially prepared media, $5.66. An estimate of 80% technologist time efficiency was made in all procedures. Bacterial kit identification systems have become the most widely used means by which present study details features of this laboratory negative organisms from culture specimens. The bacterial identification is carried out in the clinical microbiology laboratory (6). More labora- and clinical logistics comparison. tories use the API 20E (Analytab Products, MATERIALS AND METHODS Plainview, N.Y.) methodology (6) than any Specimens and organisms. Bacterial cultures other kit method currently available. The Micro- evaluated in this study were obtained from the Clinical ID system (General Diagnostics, Morris Plains, Microbiology Laboratory of the University of Utah N.J.) has a much more recent genesis, but also Medical Center, Salt Lake City. Specimens were divided by means of the general processing bench to is well proven (2, 3, 4). In addition, the Microwhich they were usually sent. Benches were assigned ID system offers an alternative identification either the Micro-ID kit, the API kit, or conventional time table; an identification is usually available biochemicals for the duration of the study, except for after 4 h of incubation, in contrast to the overnight incubation required with most kit methods employed during the period of this evaluation. weekends, when only conventional biochemicals were and conventional biochemical testing. Strains represented consecutive gram-negative isolates and included isolates from a variety of culture Both the API and Micro-ID kit methodologies, as well as routine biochemical testing, have specimens, with the exception of blood cultures, stool some distinct advantages inherent in laboratory cultures, and specimens processed by more than one use. Since all three methods have been shown identification system. Specimens were classified in six categories: urine, respiratory, genital, sterile body to be accurate, we decided to study each of them areas, wounds, and miscellaneous specimens. to assess time, motion, and cost on our routine Technologists. Technologists who processed specimens rotated duties so that each technologist had an bacteriology work benches. We were also interested in assessing the timing of reporting gram- opportunity to work with each of the three comparison 665

2 666 BALE ANI) MATSEN methods. A total of 11 technologists worked on this study at one time or another and followed strictly defined protocols. All data were recorded on uniform work sheets, and one of us (M.J.B.) reviewed all identification results to provide uniformnity. Final comparison data were extracted within 1 eek to ensure t imeliness. Individual system protocols. The methodologv used for each of the three comparison systems ha.s been previously published by oui research group (1). We used a six- or eight-tube biochemical identification set on all Entei-obcteriaceeae strains, including triple sugar iron agar, motility-indole-ornithine medium, Simmons Citrate agar, phenylalanine urea broth, lysine iron agar, and Voges-Proskauer broth, and performed deoxyribonuclease and cytochrome oxidase determinations on all non-lactose-fermenting organisms from MacConkey agar. The directions of the manufacturer were followed for both the API and Micro-ID systems. An average of seven tubes were used for cost comparison purposes since lactose fermenters required six tubes and non-lactose-fermenting organisms required eight tubes. IPuritv plates were not used for the conventional methodology, as a single discrete colonwas used for subculture to inoculate the triple sugar iron agar and for other tests. Subsequently, the triple sugar iron agar could be used as the inoculum for any further testing. The API and Micro-ID) methods both required a purity plate to have a pure growth of the organisms available for additional testing, if neces.sarv. In the case of the Micro-ID) (or a 5-h API) method, a purity plate was necessary to assess purity, since multiple colonies are necessary to obtain sufficient inoculum. Specimens for which equivocal results were obtained were tested on additional biochemical according to the manufacturer's directions that were specific for certain octal numbers obtained bv the kit used. For organisms evaluated by conventional biochemical testing, further determinations were carried out when procedural organism identification flow charts indicated. Time studies. One of us (M.J.B.) did time studies on technologists who performed identifications b- each method. Each step of the procedure was timed, an average time was accrued bv obtaining multiple (20) assessments, and the total time needed for processing specimens was calculated by summing the various component parts. We were careful to employ the times associated with various batbh quantities (one to seven organisms per batch) being processed bv the techbnologist Cost of materials and labor. We obtained contract kit costs for both the Micro-Il) and API methods according to the number of specimens processed in our laboratory. All conventional media used were prepared in our laboratory, with careful quality control carried out on each batch of media. However, for cost comparison purposes, we used the price of the media which we would have been charged had we used a commercial manufacturer-. We will report only direct costs in this publication. The cost for performing extra tests was figured, therefore, onlv on the basis of materials and technologies time required. An average was used to provide sonie guideline as to individual speciunen processing. However, labor costs for perforrnîng the additional hiochemical nmethods didlnot include the additional technologist time spent in referring to the reference I)ut)- lications, a steel) vhich is often necessary when dealing with unusual organisnis re(quiring sul)l)lenentarv testillg Ȯrganism reporting chronology. We determinled an elapse(d time as to when the identification was available on each grarn-negative organism processed in this studv. We also assessed hov manv organisrn reports were delayed because other organinsms in the same culture could not be reported due to timing or some other proce(lural reason. We further analyzed this on the basis of whether grarn-positive or gramnnegative organisms were contributing factors in delaved reporting, or as to why anv organisnms identified bl anv of the procedures could not be reported. A consideration which is difficult to put in perspective arises because some organisnis were reported via phone conversations or as prelirninarx results reported to physicians before a final report,was available. Other analyses. An assessment was made of the amount of incubator and storage space required for the materials necessary to perform each method in conjunction with the shelf life of the material utilized. Out-date considerations were analvzed as well as the amount of time, effort, and materials required to perform appropriate quality control. We also attempte(d to assess the ease of usage an(l the level of technologist acceptance of each of the methodologies. RESULTS Specimens and organisms. A total of 567 specimens with En terobactei iaceae isolates were processed in this evaluation. The study required 3 months to complete. There were 7,512 specimens in the five specimen categories received in the laboratory during this time period, but only those meeting the laboratory studv protocol were processed for identification and included in this study. Table 1 shows the five categories of specimens and the number processed by each of the three systems. There were 145 specimens processed by the API method, 150 specimens processed by the Micro-ID) System, and, because specimens processed on the weekends, 273 specimens processed by conven- TABLE 1. Types ot specimens wj. CIAIN. MIC'HOBIOL,. No. ( ') of spe-ciniens processed hb tollowitlg Specniren svstenc: Ail Micro-11) Conventional Urine 79 (55) 101 (67) 142 (52) Respiratory 43 (30) 22 (15) 59 (922) Genital 3 (2) 6 (4) 10 (4) Sterile bodv 5 (:3) :3 (2) 10 (4) areas Wound 15 (1t)) 18 (12) 52 (19)

3 VOL. 14, 1981 tional means only. This selection process resulted in 52 to 67% of all specimens studied being urine specimens, whereas for the same period of time approximately 35% of the specimens received in the clinical laboratory were urine specimens. Respiratory specimens varied from 15 to 30% between each of the three systems. The only other specimen type which seemed to vary much among the systems used for evaluation was that of wounds, where 10% were processed by the API method, 12% by the Micro-ID System, and 19% by conventional biochemical testing Ṫable 2 shows the organisms identified by each system; there was relative homogeneity in the distribution of organisms among the three systems. Escherichia coli (391) represented 53 to 55% of the total isolates, with Proteus (111) being the second most frequent and Klebsiella (103) the third most frequently isolated genus. Technologists. All technologists had considerable experience with both the conventional and API methodologies, but some had only limited experience with the Micro-ID system for routine use in the clinical laboratory. Time study results obtained early in the course of the evaluation with technologists unfamiliar with the Micro-ID method differed from results obtained after these same technologists became acquainted with the procedure. Data reported here reflect the timing evaluations of each of the three systems for technologists proficient with the individual systems. Table 3 shows the comparative processing parameter results for the three identification TABLE 2. Organisms identified by each system No. (%') of organisms identified bh following system: Organism _ API Micro-ID Conventional E. coli 109 (55) 99 (53) 183 (53) Klebsiella 40 (20) 19 (10) 44 (13) Enterobacter, 10 (5) 20 (11) 36 (10) Hafnia Proteus 26 (13) 31 (17) 54 (16) Providencia 4 (2) 4 (2) 5 (1) Citrobacter 8 (4) 6 (3) 11 (3) Serratia 2 (1) 8 (4) Il (3) COST COMI>ARISON ()F ENTERIC IDENTIFICATION 667 svstems. There were a total of 730 Enterobacteriaceae isolates identified. In the course of processing the specimens, 17%`e of those processed by the API method required subculture either to appropriately isolate the organism for subsequent biochemical testing or to provide a sufficient inoculum for API testing. This number increased somewhat with the Micro-ID system to 25%7e, a figure very comparable to the percentage (23%k) of isolates which required subculture for conventional biochemical processing. A total of 5%,l of the isolates processed with the API system required biochemical testing in addition to the API strip. A total of 12%ï of the isolates processed by the Micro-ID system and 16'7% of those processed by conventional means required additional tests. Table 4 shows an overview of the timing of final culture results for each of the three systems. A total of 44 7x of the organisms identified by the Micro-ID system were reported on the first day after the specimens were received in the laboratory. A total of 67%7r of the organisms were identified on the first day, but a portion of these were reported to the physician via a preliminary culture report because of the presence of other organisms, gram-negative bacilli or otherwise, which were present in the specimens. On the second day, 68%ï of the specimens processed by the API system could be reported to the physician, and, by that time, 83'. of the specimens processed by the Micro-ID system could be reported finally. On the second day, only 60%,- of the specimens processed by conventional means could be reported. Specimens cumulatively reported to the physician by days 3, 4, and 5 were rather similar between the three methods. The mean time for final report distribution was 2.44 days for the API systems, 1.79 days for the Micro-ID system, and 2.55 days for conventional biochemical processing. Final reports for three times as many specimens were delayed with the Micro-ID method due to organisms other than Enterobacteriaceae than were delayed with either the API system or conventional testing (Table 5). However, preliminary culture reports were distributed which included the identification of organisms rapidly defined by the Micro-ID method. Thus, the TABLE 3. Comparative processing parameter results for the three identification sstemx No. ('.) of isolates No. of.n. of isolates requiring additional h nulte tet multiple wth culttlres.blochemlcal testifig isolates identified plates or repeat testing eateries S Svstem stem No.tore Enterohoclteriacee requiring subculture.io i API (17) 10 (5) :32 (22) Micro-ID (25) 23 (12) :31 (21) Conventional (23) 56 (16) 45 (16)

4 668 BALE AND MATSEN TABLE 4. Timing of final culture results for each system Final report is- No. ( ) of organisms identified by followsued at follow- ing systern: ing no. of days - _- after receipt API Micro-ID Conventional 1 66 (44) 2 98 (68) 59 (39) 164 (60) 3 31 (22) 18 (12) 76 (28) 4 12 (8) 5 (3) 26 (10) >5 3 (2) 2 (1) 7 (3) Mean TABLE 5. Organism Reports delayed due to organisms other than the Enterobacteriaceae No. (%) of reports delayed in following system: API Micro-ID" Conventional Staphylococcus 7 (5) Enterococci 2 (1) 8 (5) 22 (15) 6 (2) 6 (2) Haemophilus Pseudomonas aerugi- 3 (2) 1 (1) 6 (4) 7 (5) 4 (1) 5 (2) no<sa Other non-enter-obac- 1 (1) 1 (1) 6 (2) teriaceae gram-negative rods " The Micro-ID delay, in general, should be viewed as a delay in large part offset by the ability to report the results of the Micro-ID identifications on a preliminary culture report. delay which would be otherwise expected and which related to those organisms requiring overnight incubation for appropriate biochemical testing was at least partially obviated. Staphylococci, which can be tested by means of rapid coagulase and catalase testing, were not a problem in this regard. Susceptibility testing results accompanied organism identification reports in almost every instance where a susceptibility test would appropriately be performed. This was possible in our situation, owing to the routine use of the Autobac instrument, which provides a 3- to 5-h susceptibility result. The evaluation of cost (Table 6) was an important part of this study. The overall costs were calculated on the basis of direct costs, recognizing the differences among laboratories for indirect cost formulae in charging for specimens. As previously indicated, all media used in the study were prepared in our laboratory. For purposes of the cost comparison, however, we used a price for media which was a composite of the prices quoted to us by two commercial manufacturers. Had we taken only our own labor and direct costs into consideration, the cost of performing conventional biochemical methods in our labo- TABLE 6. Expense Costper identification Cost ($) J. CLIN. MICROBIOI. per identification by following system: Conventional API Micro-ID Media or strip " Labor (80v efficiency)l' Ancillary Saline tube Purity sheep blood agar plate Pipette Reagents Additional biochemi cals Subculture plates 0.09 t) Labor cost for additional biochemicals and subcultures Total direct cost per iso late "'In the conventional system, 7 tubes were used. 'The time required to perform an identification was 6 min in the API system, 4.5 min in the Micro-ID system, and 7 min in the conventional system. ratory would have fallen between the figure for the API method ($4.96) and that for the Micro- ID method ($4.30) (Table 6). Results of the assessment of incubator space requirements showed that the Micro-ID and API systems were very similar in their requirements, with the conventional system necessitating considerably more space for routine processing. The shelf storage life for conventional media, again, was not as optimal as those for the Micro-ID and API systems, and the shelf outdates for these systems are listed as: API, 18 months from the date of manufacture; Micro- ID, 15 months from the date of manufacture; conventional media, 3 to 4 weeks snap cap, 3 to 6 months screw cap (5). The shelf life for the conventional biochemical media used rarely was a problem, as the number of specimens processed allowed for a rapid use of the media prepared in our laboratory. Technologists who were unaccustomed to the use of Micro-ID method before this evaluation were somewhat reluctant to use it in the beginning. They felt very comfortable with the combination of conventional media and the API system which we had used routinely in our laboratory. Therefore, acceptance of the new system was not readily admitted by technologists in the beginning of the study. Subsequently, it was shown to be equally as popular as the other

5 VOL. 14, 1981 two methodologies once the technologists became proficient with the system. DISCUSSION The accuracy of all three of the identification systems analyzed in this particular study has been well demonstrated (1, 2, 3, 4). In our own analysis of these systems, we feel there are certain advantages and disadvantages for each (Table 7). This study has demonstrated that the Micro- ID system was more cost efficient and timely than either of the other two methods studied and that 44% of ail organisms identified by the Micro-ID system could be reported out on the first day after specimens were received in the laboratory. The results with the Micro-ID system demonstrated that, in actuality, 67% of the organisms were identified on the first day. On many of these organisms, we could not issue a final report to the physician because of the presence of other organisms which required overnight incubation to appropriately identify them. However identification by the Micro-ID system could be conveyed by means of the preliminary COST COMPARISON OF ENTERIC IDENTIFICATION 669 culture report distributed for all specimens on which a final report could not be sent. We found that we subcultured more frequently with Micro-ID kits than we did with API kits, primarily because, as a rapid test, we required a large inoçulum to appropriately load the identification system. The API system now has a 5-h adaptation, using a new profile index (data base) computation. The new API adaptation also requires a heavier inoculum and has the same liability in this regard. On the whole, results were more quickly derived by means of the Micro-ID system, as is indicated above, than with either of the overnight methods. Both kit methodologies required essentially the same amount of storage and incubator space. Both had comparable shelf storage lives, and the considerations of their usage were primarily based on technologist acceptance, test validity, laboratory cost considerations, and the timeliness of organism reports. In summary, there are advantages for each of the three systems, and the considerations listed in our evaluation are items which each laboratory will have to evaluate independently. The TABLE 7. Comparison of advantages and disadvantages of the three systems Advantages Disadvantages API method Micro-ID Conventional API method Micro-ID Conventional system testing system testing 1. Has more bio- 1. More rapid re- 1. Classic method- 1. Nonrapid, re- 1. Requires a 1. Short storage chemical tests sults ology quires 18 h for larger inocu- time incubation lum; therefore, all isolates cannot be processed immediately 2. Requires fewer 2. Least costly 2. Format can be 2. Slightly more h incubasubcultures of altered to in- expensive tion required individual iso- clude short sets lates concomitant with morphology criteria to actually be cheaper than any kit method for most gramnegative isolates 3. Provides fewer 3. Least amount 3. Greater flexibil- :3. New 5-h test re- 3. Gives highest incomplete re- of technologist ity in biochemi- quires a larger percentage of sults with ini- time required cal test usage inoculum, with incomplete retial testing necessary delays sults for certain isolates 4. Can be used 4. Requires more for non-glucose- incubator and fermenting refrigerator organisms space 5. Newly devised 5. No computermanual allows based probafor 5-h identi- bility fication

6 670 BALE ANI) MATSEN pressure for more rapid reporting and the considerations of cost factors have led us to utilize the Micro-ID system as the primary kit methodology employed in our laboratories. ACKNOWLEDGMENTS We vish to acknowledge the part icipation of the personnel of the Clinical Microbiology Iaboratory at the utniversitv, of Utah Hospital. We also gratefully acknowledge the excellent secretarial assistance of Constance Staples. LITERATURE CITEI) 1. Aldridge, K. E., B. B. Gardner, S. J. Clark and J. M. Matsen Coniparison of Micro-Il). API-20E, and conventional media svstenms in identification of Enterobaoceriaceae. J. Clin. Microhiol. 7: Barry, A. L., and R. E. Badal. Rapid identification of Entetrobactetiaceoe with the Micro-ID systemnersus J. CIIN. MICROBIOL. A1>I-2(tE and convent onal media J. Clin. Microl)iol. 10: ;3. Blazevie, I). J., 1). L. Mackay, and N. M. Warwood. 197,9. Comiparison of Micro-ID and API-20E svstenis for identification of Enteroh(cteî-iceoe. J. Clin. Micrlo- 1)iol. 9:605-;( Edberg, S. C., B. Atkinson, C. Chambers, M. H. Moore, L. Palumbo, C. F. Zorgen, and J. M. Singer Clinical evaluation of the Micro-Il), API-2(IE, and (conventional media svstens for identification of Ente-oha(oferiaceae. J. Clin. Microbiol. 10: ê. 5. Inhorn, S. (ed.) Quality assurance practices for health lai)oratories, ) American P>ublic Health Association, Washington, D.C. 6. Marymont, J. H., III, J. H. Marymont, Jr., and T. J. Gavan Performance of'entei-obacteriaeae identification svstemns: an analysis of College of Anercali Pathologist survey data. Am. J. Clin. Pathol. (suppl.) 70:5' Downloaded from on January 16, 2019 by guest