Part I: Organisation and management
quality control, quality assurance and proficiency testing

Quality assurance with regard to tuberculosis bacteriology is a system designed to continuously improve the reliability, efficiency and use of tuberculosis laboratory services. The purpose of a quality assurance programme is to improve the efficiency and reliability of laboratory services. In order to achieve the required technical quality in laboratory diagnosis, a continuous system of quality assurance needs to be established. Intermediate laboratories should supervise the peripheral network, while the central or reference laboratory should supervise the intermediate network.

The components of a quality assurance programme are:

• quality control
• quality improvement
• proficiency testing

Quality control
Quality control is a process of effective and systematic monitoring of the performance of bench work in the tuberculosis laboratory against established limits of acceptable test performance. Quality control ensures that the information generated by the laboratory is accurate, reliable and reproducible and serves as a mechanism by which tuberculosis laboratories can validate the competency of their diagnostic services.

The specific aspects of quality control for microscopy and culture procedures are discussed extensively in the Technical Series on Microscopy and the Technical Series on Culture.

Quality improvement
Quality improvement is a process by which the components of tuberculosis laboratory services are analysed continuously to improve their reliability, efficiency and utilisation. It has been shown that the most effective and long-lasting improvements are achieved by anticipating and preventing problems rather than by identifying and correcting defects after they have occurred. Data collection, data analysis and creative problem-solving are the key components of this process. It involves continuous monitoring, identification of defects, followed by remedial action to prevent recurrence of problems. Often, problem-solving can be done efficiently only during on-site supervisory visits. These are the quickest and most effective form of quality improvement because of the personal contact and permits on the spot corrective action. Supervision should always be done by a more experienced laboratory technician. Activities during these visits should include the following:

• observing general laboratory hygiene and safety practices
• checking that written standard operating procedures for equipment and laboratory methods are in place and easily accessible
• checking that service and maintenance records of equipment are up to date
• evaluating the proportion of unsatisfactory specimens, including those where specimen containers have been reported to be broken or leaking. Should a particular health facility be identified as a source of the problem, the necessary consultation with health care  staff should take place
• checking that stains and reagents and culture media contain the necessary information on preparation and expiry dates and that expired stock is not in use
• checking that positive and negative controls are used as necessary during microscopy, culture or identification procedures
• analysing the monthly proportion of positive smear or culture results for deviations from the norm
• collecting a selection of positive and negative slides for re-checking. One of the  following procedures may be chosen by the national programme:
   
  • For every 100 slides examined, select all positives in the order in which they have been found. Next, select a collection of negative slides in random  fashion, eg.
    
    Suppose 10 positive slides (10%) were found, select 10% of negative slides by selecting every tenth slide (100/10 = 10). Suppose 25 positive slides (25%) were found, select 25% of negative slides by selecting every fourth slide (100/25 = 4).
     
  • Select all positives and the following negative slide.
     
  • Keep all the slides for two months. If requested by the supervising  laboratory, send all positive and negative slides. The selection of a proportion to be  checked will be done at the supervising laboratory, or by the supervisor during a visit.
     
• it is particularly important to collect slides from follow-up investigations since these are more likely to contain errors (false negative results).
• evaluating the monthly culture contamination rate in terms of the proportion of  specimens contaminated as well as the proportion of cultures (bottles or tubes) contaminated

A model laboratory evaluation form is presented in Annex 5.

It is essential that feedback be provided to the laboratory staff and that recommendations be made to immediately correct any deficiencies found. Training is usually very helpful in correcting deficiencies.

Proficiency testing
Although quality improvement is the quickest and most effective form of (external) quality assurance, it is often difficult to perform on a regular basis owing to limitations of time and travel. Indirect technical and administrative control through proficiency testing programmes should therefore become an essential component of quality assurance. Proficiency testing is highly desirable but not easy to achieve. In order for such programmes to be successful they must run in the form of a continuous assessment and they require skilled and dedicated staff. The feasibility and methodology of these programmes in developing countries is controversial. Nevertheless, the assurance of smear microscopy and/or culture quality is of utmost importance to National Tuberculosis Programmes.

Proficiency testing is a programme designed to allow participating laboratories to assess their capabilities by comparing their results to those obtained in higher level laboratories. For this purpose, material for testing is prepared by the central or reference laboratory and distributed to lower level laboratories. The recipients perform the necessary procedures and report their results to the central or reference laboratory which can then assess proficiency. Detection of deficiencies through this indirect system will then determine the need for quality improvement.

Participation in proficiency testing programmes may be compulsory, as in hierarchical laboratory organizations, where the central or reference laboratory is responsible for those at the lower levels. It may otherwise be voluntary, where a specific quality control laboratory (QCL) of the public health laboratory service provides a variety of material to any interested laboratory. Each laboratory has a unique identification code known only to itself and the QCL. The QCL sends out collective reports which enable each of the participating laboratories to compare its proficiency to those of the others.

Irrespective of whether proficiency testing programmes are compulsory or voluntary, the following (minimum) activities are recommended; preferably every six months:

Microscopy
An artificial set of standard smears with known results are sent to peripheral level laboratories. These should consist of two sets of stained and unstained smears. A minimum of five slides per set is required, covering the full range from negative to strongly positive, as follows:

Culture
An artificial set of sputum specimens with known results are sent to regional or central laboratories. One set of two specimens each should be inoculated with M. tuberculosis and a second set of two with M. fortuitum, which resembles M. tuberculosis in terms of colony morphology and reduces nitrate but is a rapid grower that is niacin negative. A third set of two specimens should be negative.

It should be realised that results from proficiency testing programmes may be biased, since laboratory staff are aware of the origin of proficiency smears and cultures and may dedicate more time and attention to their correct processing and examination. In addition, this method does not allow measuring the quality of smear preparation. One way to overcome this problem is to send specimens (rather than smears or cultures) to participating laboratories. These could be prepared artificially by adding 1 emulsified egg to 1 000ml of 1% (w/v) aqueous methylcellulose and distributing 3.5ml volumes into specimen containers. These resemble slightly purulent sputum and are then inoculated with standardised concentrations of M. tuberculosis and Nocardia asteroides (for microscopy) or M. tuberculosis and M. fortuitum (for culture). To check on decontamination and digestion procedures, specimens could be inoculated with Escherichia coli as a contaminant. Clinical specimens from patients (rather than artificial specimens) can also be used and inoculated as described before.

Results should be analysed in 2x2 tables as disagreement on negative readings (false positivity), disagreement on positive readings (false negativity) and disagreement as a function of bacillary concentration (1-10 AFB, 1+, 2+, 3+). Results should be reported to the central / reference laboratory within an agreed period of time.

Irrespective of the methods of proficiency testing, the most important aspect is regular feedback and corrective measures taken in a spirit of mutual trust and agreement.

In summary, well-designed and properly managed quality assurance programmes are an asset to any laboratory. Positive aspects of such programmes for tuberculosis bacteriology include the following:

• Potential problems in the isolation and identification of M. tuberculosis can be  greatly reduced by monitoring media and reagents before using them on clinical specimens
• Serious and costly breakdowns of equipment may be minimised by routine monitoring and  maintenance
• Laboratory reports can be more accurate and expeditious as the use of inadequate media, equipment and techniques is minimised
• The quality assurance programme can serve as a learning exercise, enabling the recognition and identification of problem areas that might otherwise have been overlooked
• A good quality assurance programme will enhance the credibility of the laboratory to outside clients