To be of good quality, a diagnostic test must be clinically relevant, i.e., it must help in the prevention or treatment of disease. Other measures of quality in a diagnostic test are:
• Reliability: Is the result correct?
• Reproducibility: Is the same result obtained when the test is repeated?
• Speed: Is the test rapid enough to be of use to the doctor in prescribing treatment?
• Cost - benefit ratio: Is the cost of the test reasonable in relation to the benefit to the patient and the community?
Factors that affect the reliability and reproducibility of laboratory results
Sources of error may include the following:
• Personnel. The performance of the laboratory worker or technician is directly related to the quality of education and training received, the person’s experience, and the conditions of employment.
• Environmental factors. Inadequate working space, lighting, or ventilation, extreme temperatures, excessive noise levels, or unsafe working conditions may affect results.
• Specimens. The method and time of sampling and the source of the specimen are often outside the direct control of the laboratory, but have a direct bearing on the ability of the laboratory to achieve reliable results. Other factors that the laboratory can control and that affect quality are the transport, identification, storage, and preparation (processing) of specimens. The laboratory therefore has a role in educating those taking and transporting specimens. Written instructions should be made available and regularly reviewed with the clinical and nursing staff.
• Laboratory materials. The quality of reagents, chemicals, glassware, stains, culture media, and laboratory animals all influence the reliability of test results.
• Test method. Some methods are more reliable than others.
• Equipment. Lack of equipment or the use of substandard or poorly maintained instruments will give unreliable results.
• Examination and reading. Hurried reading of results, or failure to examine a sufficient number of microscope fields, can cause errors.
• Reporting. Transcription errors, or incomplete reports, cause problems.
Quality of interpretation of test results
Interpretation is of particular importance in microbiology. At each stage in the examination of a specimen, the results should be interpreted in order to select the optimum test, in terms of speed and reliability, for the next stage of the examination.
Quality assurance in the microbiology laboratory
Quality assurance is the sum of all those activities in which the laboratory is engaged to ensure that test results are of good quality. It must be:
• comprehensive, to cover every step in the cycle from collection of the specimen to sending of the report to the doctor (see Fig. 1);
• rational, to concentrate on the most critical steps in the cycle;
• regular, to provide continuous monitoring of test procedures;
• frequent, to detect and correct errors as they occur.
GOOD QUALITY LABORATORY SERVICES MEAN
GOOD QUALITY MEDICINE
Quality assurance helps to ensure that expensive tests are used as economically as possible; it also determines whether new tests are valid or worthless, improves the performance of clinical and public health laboratories, and may help to make the results obtained in different laboratories comparable.
Types of quality assurance
There are two types of quality assurance:
• Internal: this is called QUALITY CONTROL. This means that each laboratory has a programme to check the quality of its own tests.
Fig. 1. Steps in laboratory investigation of an infected patient
Internal quality control involves, ideally:
- continuous monitoring of test quality;
- comprehensive checking of all steps, from collection of the specimen (whenever possible) to issue of the report.
Laboratories have an ethical responsibility to the patient to produce accurate, meaningful results.
INTERNAL QUALITY CONTROL IS ABSOLUTELY ESSENTIAL FOR GOOD
• External: this is called QUALITY ASSESSMENT. This means that the laboratory performance is controlled by an external agency. In some countries, participation is mandatory (regulated by the government) and required for licensure.
External quality assessment involves:
- periodic monitoring of test quality;
- spot checking of identification tests, and sometimes of isolation techniques.
Quality criteria in microbiology
An important criterion of quality for a microbiological test is how much it contributes to the prevention or cure of infectious diseases; this is called its clinical relevance. Clinical relevance can only be ensured when there is good communication between the clinician and the laboratory.
To illustrate clinical relevance, here are some examples:
1. If a few colonies of Gram-negative rods are isolated from the sputum or throat swab of a hospitalized patient, further identification and an antibiogram are of no clinical relevance, since neither procedure will have any effect on treatment of the patient.
2. If Streptococcus pyogenes is isolated, a full antibiogram has no clinical relevance, since benzylpenicillin is the drug of choice, and this is always active in vitro.
3. If Escherichia coli is isolated from a sporadic case of diarrhoea, identification of the serotype is of no clinical relevance, since there is no clearly established correlation between serotype and pathogenicity.
4. If a Gram-stained smear shows “mixed anaerobic flora”, routine identification of the anaerobes is of no clinical relevance. It would be costly in time and materials, and would not affect treatment of the patient.
5. If a yeast is isolated from a respiratory tract specimen, an identification test for Cryptococcus should be done. Further identification tests have no clinical relevance, since they would have no effect on patient management.
In summary, a test of good quality is one that is accurate and gives useful results for the prevention or cure of infection. It is not necessary to isolate and identify all the different types of organism in the sample.
For tests that give quantitative results, reliability is measured by how close the results are to the true value. Some examples of tests of this kind are:
• antibiotic assay in serum;
• measurement of minimal inhibitory concentration (MIC) values of antibiotics in vitro;
• serum antibody titrations.
For tests that give qualitative results, reliability is measured by whether the result is correct. Examples of tests of this kind are:
• identification of pathogens;
• antibiotic susceptibility testing of isolates by the disc method.
Standard terminology for microorganisms is essential to reliability. Internationally recognized nomenclature should always be used. For example: Staphylococcus aureus, NOT “pathogenic staphylococcus”; Streptococcus pyogenes, NOT “haemolytic streptococcus”.
Use of uniform, approved methods is essential. For example, disc susceptibility tests should be performed with an internationally recognized technique, such as the modified Kirby-Bauer test
The reproducibility or precision of a microbiological test is reduced by two things:
1. Lack of homogeneity. A single sample from a patient may contain more than one organism. Repeat culturing may therefore isolate different organisms.
2. Lack of stability. As time passes, the microorganisms in a specimen multiply or die at different rates. Repeat culturing may therefore isolate different organisms. To improve precision, therefore, specimens should be tested as soon as possible after collection.
The efficiency of a microbiological test is its ability to give the correct diagnosis of a pathogen or a pathological condition. This is measured by two criteria:
Sensitivity = total no. of positive results/total no. of infected patients.
The greater the sensitivity of a test, the fewer the number of false-negative results.
For example, the sensitivity of MacConkey agar is poor for the isolation of Salmonella typhi from stool. This important enteric pathogen is often missed because of overgrowth by nonpathogenic intestinal bacteria.
Specificity = total no. of negative results/total no. of uninfected patients
The greater the Specificity of a test, the fewer the number of false-positive results.
• Ziehl-Neelsen staining of sputum is highly specific for diagnosing tuberculosis, because it gives only a few false-positive results.
• Ziehl-Neelsen staining of urine is much less specific, because it gives many false-positive results (as a result of atypical mycobacteria).
• The Widal test has a low specificity for the diagnosis of typhoid fever, because cross-agglutinating antibodies remaining from past infections with related salmonella serotypes give false-positive results.
The sensitivity and specificity of a test are interrelated. By changing the criteria of limitation, the sensitivity of a test can be increased at the cost of reducing its specificity, and vice versa.