Quality control in the new environment: lab testing near the patient.
Evolving technology has wrought complex analytics systems that are simple to operate, relatively reliable, and inexpensive. These systems allow production of timely test results near the patient, facilitating diagnostic and medical management decisions.
Capillary glucose monitoring with reagent strips is widely used in hospital wards, office laboratories, and by patients to assess glucose control and to better guide therapeutic decisions. At one of the authors' institutions, Oregon Health Sciences University Hospital, fetal scalp pH is measured in the delivery suite, and glucose, potassium, and blood gas tests are performed in an operating room laboratory.
Soon coronary care units will be able to perform antiarrhythmic drug assays and improve their ability to rapidly adjust doses to appropriate levels. Office practices and urgent-care facilities currently can obtain timely serum potassium, theophylline, or anti-convulsant levels.
Operation of the new systems has been simplified to let individuals without formal laboratory training perform test analyses. Evidence indicates that the precision of these systems is consistent with today's standards--when medical technologists do the testing. However, the variability of capillary glucose determinations performed by nurses using reagent sticks on wards throughout the university hospital and in the nursery was significantly higher than that of glucose determinations performed by the central laboratory.
Similarly, glucose levels determined with one of the new desktop chemistry analyzers by health care workers lacking formal lab training were much more variable than the same analyses performed with an identical system by medical technologists in the hospital's laboratory. The source of the increased variance in this study could not be identified, but trained technologists were not able to induce erroneous or highly variable results with this system.
Trained technologists were, however, able to induce statistically and clinically significant erroneous results on another of the new generation of desktop chemistry analyzers without any indication from the system that a problem existed. This system's variability increased markedly when analyses were performed by individuals without formal lab training.
Microbiology testing has been found to be much less reliable in doctors' offices than in hospital or independent laboratories. For example, the office labs missed up to 70 per cent of the significant throat pathogen in one study, and their false-positive rates were as high as 25 per cent.
These findings bring to mind problems encountered in hospital and independent laboratories many years ago when they first acquired some of then-new automated laboratory analyzers. Today's high laboratory performance level resulted from several decades of design improvements in analytic systems, better reagent systems, the availability of a trained work force, and the evolution of protocols to insure the day-to-day reliability of reported data.
The advances in reagents and instruments and the protocols for insuring result reliability have effectively minimized systematic analytic error. This has been a sound strategy for improving laboratory quality, because the work force has been trained to minimize the frequency of random error. The increased variability of test information produced by office and ward staffs using the new simplified analytic systems can probably be attributed to an increase in the rate of random error.
It has been shown that office laboratory variability can be reduced and performance levels raised with relatively simple interventions allowing analyses to be performed by office staff without formal lab training. In the case of quantitative analyses, the improved performance has been demonstrated over a significant time period.
The state of Idaho, which has regulated office laboratories since 1977, bases its approach on a lab improvement program. Performance levels have been raised by required compliance with approved quality assurance guidelines and outcome evaluation. The QA guidelines cover such areas as preventive maintenance, record keeping, and day-to-day quality control; the outcome evaluation uses proficiency testing. It is not clear that such interventions insure the reliability of microbiology testing performed by a non-professional office-based lab staff.
Designing a program to insure the reliability of test results analyzed in office or ward laboratories requires an understanding of other factors associated with these settings. For example, Idaho has a high rate of turnover among staff members supervising office laboratories. Training new staff members can present major problems for the office-based lab director. Vendors rarely provide on-going training for replacement personnel--that cost has not been factored into the price of their products.
Consequently, without other incentives, educating newcomers will probably be left to the current staff or the departing employee. This becomes an important issue in any QA program for these sites, since office staff members who lack formal lab training tend to have a high rate of random error.
We know that medical technologists are professionals, and we understand that they require continuing education to maintain their skills. Physician lab directors may need significant incentives before they regard anyone in the office except themselves as a professional who requires such education.
The same turnover and training problems exist with bedside testing. However, the proximity of professional laboratory assistance, the availability of continuing education opportunities, and an appreciation of the importance of continuing education programs may minimize these problems.
The split of staff responsibility between laboratory and patient care activities is another key issue affecting the reliability of ward and office testing. The primary focus of attention and concern at these sites is the patient and the physician caring for the patient. The office lab staff may frequently have to interrupt test analysis to greet a patient, move someone into an examining room, or answer the phone. As most laboratorians know, such interruptions increase the likelihood of error.
With one of the new desktop chemistry analyzers, for example, delaying insertion of the reagent pack after the specimen has been applied markedly changes the glucose result. There's no built-in means of detecting the error or warning an operator of th e problem. The increased variance has been found to be statistically and clinically significant when this system is operated by staff without formal lab training.
Archiving of test results used for diagnostic or management decisions is a very important issue in office and ward laboratories. For example, the results of capillary glucose monitoring in the hospital are frequently used to modify therapy. These test results are seldom included in the laboratory section of the patient's chart. A fraction of the results never find their way onto the chart at all, while others may appear in the nursing notes, the progress notes, or in a metabolic flow chart, which is often discarded after the patient is discharged. The integrity of laboratory data becomes very important when professional responsibility for an adverse patient outcome must be evaluated.
Office staff members who lack formal lab training don't fully appreciate that one patient's specimen cannot be differentiated from others unless it is properly identified. This can also be a problem with bedside testing, but it is particularly critical in a physician's office where patients with the same surname--commonly, brothers and sisters brought in by their parents--may require lab testing during the same visit.
Insuring the reliability of test results in an office laboratory requires the development of practice standards. There will probably always be a discrepancy in precision, accuracy, and reliability between results produced in an office laboratory and those produced in a hospital or independent lab. This dual standard of practice may be acceptable because timely results can improve not only the technical but also the personal aspects of the health care process (providing the patient with diagnostic information before he or she leaves the doctor's office). It is essential, however, to keep the discrepancy as small as possible and to minimize the variability of reported results so that they provide a reliable basis for patient care decisions.
The standards of practice should include guidelines for specimen handling, specimen identification throughout the analytic process, and appropriate record keeping to document these and other aspects of office laboratory work. The standards should also include guidelines for maintaining analytic systems to minimize breakdowns or malfunctions that might lead to errorneous results. Finally, the standards should provide for checking on the functionality of each analytic systems as a whole on a day-to-day basis and for checking the office laboratory's performance level.
Contemporary quality control programs in hospital and independent laboratories are designed to detect systematic error before malfunction or breakdown but are only insensitive detectors of random error. That is operationally acceptable in settings where professional technologists perform the analyses because they have been trained to process tests in a manner that keeps the random error rate to a minimum. Office and hospital ward staffs performing analyses generally have not been trained in this manner.
Studies of the desktop chemistry analyzers indicate that quality control protocols used with highly automated analysis are not enough to insure reliable results. It may be necessary to develop new approaches to validating the reliability of day-to-day operation, with a more sensitive detector of random error. For example, instrument manufacturers may want to build a positive and negative process control into each analytic system to assess the analyst as a variable. Methods are also needed to positively identify each specimen from collection through reporting, preferably avoiding the logs currently used in hospital and independent labs.
Proficiency survey protocols that employ unknown specimens to document performance levels in office and ward laboratories are probably adequate. However, the type of specimens now used may be inappropriate for the office lab. These specimens, as well as many of those used for daily quality control, are generally lyophilized and require reconstitution. It may be too much to expect that office staff members can accurately measure and reconstitute these materials.
Although many laboratories use liquid control materials for day-to-day quality control, ethylene glycol in these products disrupts or destroys the integrity of reagents in some of the desktop chemistry systems. For example, dry reagents in a supporting matrix of gelatin or cellulose acetate are damaged by ethylene glycol.
Controls and unknown samples used to evaluate performance should effectively be delivered in liquid form to office and ward laboratories. This is essential to minimize the variability attributed to reconstitution of the material and allow an unambiguous assessment of the system's reliability (in the case of day-to-day controls) and the lab's performance level.
Current approaches to validation of microbiology testing are not adequate to insure the reliability of results generated by office laboratories. Day-to-day quality assurance protocols in microbiology depend on component control. This validates the media's ability to support and inhibit growth appropriately and demonstrates that the biochemical indicators, including those in the media, function properly. Conventional microbiology, however, requires a significant degree of analyst skill, and the methods do not include a check on the correctness of the analyst's observations or interpretations.
It is possible to improve the reliability of microbiology results by developing systems that minimize operator interpretation and include positive and negative controls for pathogens commonly found in an office practice. New microbiology methods amenable to "process control," such as immunodiagnostic methods for identifying group A beta hemolytic streptococci, may also improve the reliability of microbiology results produced in the office lab.
As we have noted, performing test analyses near the patient provides timely results, which can facilitate diagnosis and treatment. At the same time, dispersal of lab technology into doctors' offices and wards introduces problems concerning reliability. Some of these problems were recognized when newly automated instruments were first brought into centralized labs. Other problems are uniqe to labs that perform tests near the patient. Applying fundamental practices used in hospital and independent labs will deal with some but not all of these problems.
The nature of office and ward staffs demands new daily quality control protocols for quantitative analyses--protocols that monitor the analyst as part of the analytic system. The nature of these staffs also requires development of control materials that delivered to the user in liquid form and stabilized by means other than ethylene glycol.
Finally, quality assurance for microbiology in the office laboratory requires development of systems with built-in process control or semiquantitative methods amenable to process control that can be used to identify potentially pathogenic bacteria.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Part 7; testing in physician's offices and wards needs new QC protocols|
|Author:||Belsey, Richard; Baer, Daniel M.|
|Publication:||Medical Laboratory Observer|
|Date:||Mar 1, 1987|
|Previous Article:||Monitoring drug therapy via clinical laboratory tests.|
|Next Article:||Sharpening your communication skills.|