Quality manual

The initial step in building the quality system is developing a quality manual. The quality manual addresses policies for each of the twelve QSEs. Policies are written to answer the question: What do we do in our organization? Each policy describes the organization’s intent and provides direction for the specific QSE. Policies for the personnel QSE often include the intent and direction for job descriptions and qualifications, orientation, training, competence assessment, and continuing education. Processes are described to transform the policy into action and answer the question: How does this happen in our pulmonary laboratory? Processes are generally a group of activities or procedures. There is a process to obtain reliable spirometry data that may result in a need for several procedures such as quality control, test performance, test result selection and reporting, selection of reference values, and interpretation of results (Table 12-1). Procedures answer the question: How do I do this activity? Each QSE incorporates policies, processes, and procedures to build a platform for the path of workflow to be followed for each test procedure.

Therapist/technologist Procedures for:

Therapist/technologist Procedures for:

 

Therapist/technologist Procedures for:

Therapist/technologist Procedures for:

Therapist/technologist
Support staff Procedures for:

Quality system essentials

Organization

The organizational commitment to quality is essential and may be a requirement to comply with governmental and accreditation standards. The organization is responsible for the development of the quality manual and integration into the system by committing time and personnel to the process. The department and pulmonary function laboratory organizational structure is available in the quality manual. A visible management commitment to the implementation of the quality policies, seeking customer feedback, and responding to quality reports is required for success. Risk assessment of the pulmonary laboratory’s work operations processes is included in the organization QSE. The pulmonary laboratory should set quality goals and objectives integrated with the organization’s quality planning.

Customer Focus

The pulmonary laboratory should identify their customers and expectations. Customers include both external and internal groups. Examples of external customers include accreditation and governmental agencies, patients, physicians, nurses, clinical support services, asthma educators, home care companies, and payers. All staff involved in the path of workflow are considered internal customers. An initial step is to evaluate the laboratory’s capability to meet the identified expectations. Physicians ordering tests likely will have an expectation of how long it should take to schedule a patient for testing and the length of time to receive an interpreted copy of the report. Surveys will identify expectations met and potential for continual improvement (Box 12-1). Complaints are recorded and managed according to the nonconforming event management QSE.

Box 12-1   Pulmonary Function Laboratory Customer Service Summary

How close to your scheduled appointment time were you called in for your test?

	Before appointment time
	On-time
	30 minutes late
	1 hour late
	Longer than 1 hour

How well did the technologist explain your test?

	Excellent
	Very good
	Good
	Fair
	Poor

How knowledgeable and technically skilled was the technologist performing your procedure?

	Excellent
	Very good
	Good
	Fair
	Poor

How courteous and professional was your technologist?

	Excellent
	Very good
	Good
	Fair
	Poor

Test performed: PFT ABG CPET O₂ titration

Comment:________________________________________________

Courtesy of Mayo Clinic PF Laboratory.

Facilities and Safety

The facility should be designed to support the workflow and accommodate the equipment to meet manufacturer’s and standards expectations. It should support efficient processes. The environment, processes, and safety procedures must be compliant with organizational and accreditation standards. Environmental conditions that may impact the quality of the test results are monitored and documented. Use of accurate room temperature, humidity, and barometric pressure monitors is a standard requirement.

Pulmonary function tests, including blood gas analysis, often involve patients with blood-borne or respiratory pathogens. Reasonable precautions applied to testing techniques and equipment handling can prevent cross-contamination among patients. Similar techniques can prevent infection of the technologist performing the tests.

Each laboratory should have written guidelines defining safety and infection control practices. The guidelines should be part of a policy and procedure manual (Box 12-2). Procedures should include, but not be limited to, handwashing techniques, use of protective equipment such as laboratory coats and gloves, and guidelines for equipment cleaning. The handling of contaminated materials (e.g., waste blood) should be clearly described. Policies and procedures should include education of technologists regarding proper handling of biologic hazards. Department policies and procedures should be consistent with those mandated by individual hospitals or institutions. Most accrediting agencies require written plans for safety, waste management, and chemical hygiene. In the United States, the Occupational Safety and Health Administration (OSHA) has published strict guidelines regarding handling of blood and other medical waste (see Evolve). Safety training should be documented and the training records maintained per institutional requirements, accreditation, and regulatory requirements in the records management system. Generally, some areas will require retraining on an annual basis.

Box 12-2   Pulmonary Function Procedure Manual

Items to be included in a typical procedure manual for a pulmonary function laboratory. For each procedure performed, the following should be present:

1. Description of each test performed in the laboratory and its purpose

2. Indications for ordering the test and contraindications, if any

3. Description of the general method(s) and any specific equipment required, including disposable supplies

4. Calibration of equipment required before testing (manufacturer’s documentation may be referenced)

5. Patient preparation for the test, if any (e.g., withholding medication), and patient assessment before beginning the test

6. Step-by-step procedure for measurement/calculation of results; how to perform the measurement or calculation manually is useful for quality monitoring

7. Quality control guidelines with acceptable limits of performance and corrective actions to be taken when control values are outside of their limits

8. Safety precautions related to the procedure (e.g., infection control, hazards) and alert values that require physician notification with read-back of critical results

9. Description of results reporting

10. References for all equations used for calculating results and for predicted normals, including a bibliography

11. Documentation of computer protocols for calculations and data storage; guidelines for computer downtime and software upgrades

12. Dated signatures of medical and technical directors (may be electronic)

Adapted from American Thoracic Society. Pulmonary function laboratory management manual. 2nd ed. New York: American Thoracic Society; 2005.

Pulmonary Function Tests

Pulmonary function testing does not present a significant risk of infection for patients or technologists. However, some potential hazards are involved. Most respiratory pathogens are spread by either direct contact with contaminated equipment or by an airborne route. Airborne organisms may be contained in droplet nuclei, on epithelial cells that have been shed, or in dust particles. The following guidelines can help reduce the possibility of cross-contamination or infection:

1. Disposable mouthpieces and noseclips should be used by the patient during spirometry. Reusable mouthpieces should be disinfected or sterilized after each use. Proper handwashing should be done immediately after direct contact with mouthpieces or valves. Gloves should be worn when handling potentially contaminated equipment. Hands should always be washed between patients and after removing gloves.

2. Tubing or valves through which subjects rebreathe should be changed after each test. Any equipment that shows visual condensation from expired gas should be disinfected before reuse. This is particularly important for maneuvers such as the FVC, where there is a potential for mucus, saliva, or droplet nuclei to contaminate the device. Breathing circuit components should be stored in sealed containers (e.g., plastic bags) after disinfection.

3. Spirometers should be cleaned according to the manufacturer’s recommendations. The frequency of cleaning should be appropriate for the number of tests performed. For open-circuit systems, only that part of the circuit through which air is rebreathed needs to be decontaminated between patients. Some flow-based systems offer pneumotachometers that can be changed between patients. Pneumotachometers not located proximal to the patient are less likely to be contaminated by mucus, saliva, or droplet nuclei.Disposable flow sensors should not be reused. Volume-displacement spirometers should be flushed using their full volume at least five times between patients. Flushing with room air helps clear droplet nuclei or similar airborne particulates. Water-sealed spirometers should be drained at least weekly and allowed to dry completely. They should be refilled with distilled water only. Bellows and rolling-seal spirometers should be disinfected on a routine basis. The spirometer may require recalibration after disinfecting.

4. Systems used for spirometry, lung volumes, and diffusing capacity tests often use breathing manifolds that are susceptible to contamination. Bacteria filters may be used to prevent contamination of these devices. Filters may impose increased resistance, affecting measurement of maximal flows as well as airway resistance or conductance. Some types of filters show increased resistance after continued use in expired gas. Spirometers with bacteria filters should be calibrated with the filter in line; the spirometer should meet the minimal recommendations in Table 12-2 with the filter in place. If filters are used for procedures such as lung volume determinations, their volumes must be included in the calculations.

Table 12-2

Minimal Recommendations for Spirometers
A 3-L calibration syringe is recommended for testing VC and FVC. Twenty-four standardized waveforms are available for validating FVC, FEV₁, and FEF_25%–75%. Twenty-six standard flow waveforms are available for validating PEF. Other flows require manufacturer’s proof of performance. A sine-wave pump is recommended for MVV validation.

Test	Range/Accuracy (BTPS)	Flow Range (L/sec)	Time (sec)	Resistance/Back-Pressure
VC	0.5–8 L/ ±3% of reading or ± 0.05 L, whichever is greater	0–14	30	N/A
FVC	0.5–8 L/ ±3% of reading or ± 0.05 L, whichever is greater	0–14	15	<1.5 cm H2O/L/sec (0.15 kPa/L/sec)
FEV1	0.5–8 L/ ±3% of reading or ± 0.05 L, whichever is greater	0–14	1	<1.5 cm H2O/L/sec (0.15 kPa/L/sec)
Time zero	Time point for calculating all FEVT values	N/A	N/A	Back extrapolation
PEF	Accuracy: ±10% of reading or ±0.3 L/sec (20 L/min), whichever is greater; repeatability: ±5% of reading or ±0.15 L/sec (10 L/min), whichever is greater	0–14	N/A	Mean resistance at 200, 400, 600 L/min (3.3, 6.7, 10 L/sec) <2.5 cm H2O/L/sec (0.25 kPa/L/sec)
(except PEF)	±5% of reading or ±0.2 L/sec, whichever is greater	0–14	N/A	<1.5 cm H2O/L/sec (0.15 kPa/L/sec)
FEF25%–75%	7.0 L/sec /±5% of reading or ±0.2 L/sec, whichever is greater	±14	15	<1.5 cm H2O/L/sec (0.15 kPa/L/sec)
MVV	250 L/min at VT of 2 L/ ±10% of reading or ±15 L/min, whichever is greater	±14 (±3%)	12–15	<1.5 cm H2O/L/sec (0.15 kPa/L/sec)

Adapted from Miller MR, Hankinson J, Brusasco V, et al. Standardisation of spirometry. Eur Resp J. 2005; 26:319-338.

5. Small-volume nebulizers, such as those used for bronchodilator administration or bronchial challenge, offer the greatest potential for cross-contamination. These devices, if reused, should be sterilized to destroy vegetative microorganisms, fungal spores, tubercle bacilli, and some viruses. Disposable single-use nebulizers are preferable but may not be practical for routines such as inhalation challenges. Metered-dose devices may be used for bronchodilator studies by using disposable mouthpieces or “spacers” to prevent colonization of the device. Common canister protocols for metered-dose inhaler usage have been evaluated in the laboratory.

6. Gloves or other barrier devices minimize the risk of infection for the technologist who must handle mouthpieces, tubing, or valves. Special precautions should be taken whenever there is evidence of blood on mouthpieces or tubing. There is a risk of acquiring infections such as tuberculosis or pneumonia caused by Pneumocystis carinii from infected patients. The technologist should wear a mask when testing subjects who have active tuberculosis or other diseases that can be transmitted by coughing.

7. Patients with respiratory diseases such as tuberculosis may warrant specially ventilated rooms, particularly if many individuals need testing. Risk of cross-contamination or infection can be greatly reduced by filtering and increasing the exchange rate of air in the testing room. Equipment can be reserved for testing infected patients only. An example may be using a spirometer for cepacia positive patients in a clinic patient room. Special patient organizations such as the Cystic Fibrosis Foundation may have additional requirements for infection control. Patients with known pathogens can also be tested in their own rooms or at the end of the day (to facilitate equipment decontamination).

8. Surveillance may include cultures of reusable components, such as mouthpieces, tubing, and valves, after disinfection.

Blood Gases

The Centers for Disease Control and Prevention (CDC) has established standard precautions that apply to personnel handling blood or other body fluids containing blood. Standard precautions apply to blood, semen, vaginal secretions, cerebrospinal fluid, synovial fluid, pleural fluid, pericardial fluid, and amniotic fluid. Some of these fluids (i.e., blood) are commonly encountered in the blood gas or pulmonary function laboratory. These fluids present a significant risk to the health care worker. Hepatitis B, HIV, and other blood-borne pathogens must be assumed to be present in these fluids.

PF Tip 12-1

The two most important practices for infection control in the pulmonary function laboratory are proper use of gloves and handwashing. Gloves should be worn any time blood is handled or drawn, including handling of blood-tinged mouthpieces. Handwashing is essential to prevent cross-contamination. Hands should be washed between patient contacts; any time mouthpieces, tubing, or nebulizers are handled; and when gloves are removed.

Body fluids to which the standard precautions do NOT apply include feces, nasal secretions, sputum, sweat, tears, urine, and vomitus, unless they contain visible blood. Some of these fluids may be encountered in the pulmonary function laboratory. These fluids present an extremely low or nonexistent risk for HIV or hepatitis B. However, they are potential sources for nosocomial infections from other non–blood-borne pathogens. Standard precautions do not apply to saliva, but infection control practices such as use of gloves and handwashing further minimize the risk involved in contact with mucous membranes of the mouth.

These standard precautions should be applied in the pulmonary function and/or blood gas laboratory:

1. Treat all blood and body fluid specimens as potentially contaminated.

2. Exercise care to prevent injuries from needles, scalpels, or other sharp instruments. Do not resheath used needles by hand. If a needle must be resheathed, use a one-handed technique or a device that holds the sheath. Do not remove used, unprotected needles from disposable syringes by hand. Do not bend, break, or otherwise manipulate used needles by hand. Use a rubber block or cork to obstruct used needles after arterial punctures. Use needle safety devices (now used in almost all blood gas kits) as described by the manufacturer. Place used syringes and needles, scalpel blades, and other sharp items in puncture-resistant containers. Locate the containers as close as possible to the area of use.

3. Use protective barriers to prevent exposure to blood, body fluids containing visible blood, and other fluids to which standard precautions apply. Examples of protective barriers include gloves, gowns, laboratory coats, masks, and protective eyewear. Gloves should be worn when drawing blood samples, whether from a needle puncture or an indwelling catheter. Gloves cannot prevent penetrating injuries caused by needles or sharp objects. Gloves are also indicated if the technologist has cuts, scratches, or other breaks in the skin. Protective barriers should be used in situations where contamination with blood may occur. These situations include obtaining blood samples from an uncooperative patient, performing finger-heel sticks on infants, and receiving training in blood drawing. Examination gloves should be worn for procedures involving contact with mucous membranes. Masks, gowns, and protective goggles may be indicated for procedures that present a possibility of blood splashing. Blood splashing may occur during arterial line placement or when drawing samples from arterial catheters.

4. Wear gloves while performing blood gas analysis. Laboratory coats or aprons that are resistant to liquids should also be worn. Protective eyewear may be necessary if there is risk of blood splashing during specimen handling. Maintenance of blood gas analyzers, such as repair of electrodes and emptying of waste containers, should be performed wearing similar protective gear. Laboratory coats or aprons should be left in the specimen handling area. Blood waste products (e.g., blood gas syringes) should be discarded in clearly marked biohazard containers.

5. Immediately and thoroughly wash hands and other skin surfaces that are contaminated with blood or other fluids to which the standard precautions apply. Hands should be washed after removing gloves. Blood spills should be cleaned up using a solution of 1 part 5% sodium hypochlorite (bleach) in 9 parts of water. Bleach should also be used to rinse sinks used for blood disposal.

Personnel

The personnel QSE is one of the most important elements in the pulmonary function laboratory quality system because of the impact of the technologist on the quality of the final product. Job qualifications and duties across the path of workflow are outlined and maintained in job descriptions for all personnel in the pulmonary laboratory. The American Association of Respiratory Care (AARC) Clinical Practice Guidelines (CPGs) and the American Thoracic Society − European Respiratory Society (ATS-ERS) standards outline job requirements and qualifications for individuals employed to perform pulmonary function tests. The ATS-ERS standards also suggest training and ongoing education for individuals in the laboratory.

An orientation/training plan and process is required for all new employees. An orientation manual and training guide is developed and updated as required. The training manual may include, but is not limited to, quality management, safety, computer systems, ethics, work processes, and procedures. Records documenting the orientation and training processes are maintained according to the document and records QSE (Table 12-3). Training occurs and documentation is maintained when processes or procedures are updated or added. Performance of quality control procedures has been identified as a primary training need in pulmonary function laboratories globally, regardless of professional credentials. Orientation also includes review of the quality manual and application to work duties in the laboratory.

Table 12-3

Components of a Complete Clinical Service Training Program

Type of Employment Training	Contents
Quality	• Organization’s quality system • Quality manual • Service’s path of workflow • Quality control program • Quality assurance program • Occurrence management program • Customer service program • Quality system responsibilities • [Good manufacturing practice (GMP): blood banking only] • [Proficiency testing: Screening and diagnostic laboratory testing only]

Computer

• Main organization’s system (e.g., hospital information system [HIS])

• Department’s system (e.g., laboratory information system [LIS])

• PC applications (e.g., e-mail, scheduling, word processing, spreadsheets, database)

• Intranet

• Online documents

• Other computer applications used in the job (e.g., documentation of training, continuing education, competence assessment)

Safety

• Accident reporting

• Emergency preparedness

• Hazardous waste disposal

• Chemical hygiene program

• Infection control (universal precautions, bioterrorism, etc.)

• Radiation safety, where needed

Work processes and procedures

• Processes in the path of workflow in which the employee works

• Procedures performed

Compliance (United States only)

• Medical necessity requirements

• Fraud and abuse reporting

• Health Insurance Portability and Accountability Act (HIPAA)

(From Blonshine S, Mottram CD, Berte LM, et al. Application of a quality management system model for respiratory services: Approved guidelines. 2nd ed. CLSI document HS4-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2006.)

Purchasing and Inventory

The purchasing QSE portion requires the laboratory to establish criteria for vendor qualification, selection, and evaluation. This is typically accomplished in partnership with the materials management team and identifying the supplies required for the work processes. It is good practice for the laboratory to maintain a copy of all purchase agreements, particularly for major equipment purchases.

The inventory QSE portion requires the laboratory to develop an inventory management system that identifies and maintains supplies for the work processes in a fiscally responsible environment. In order to meet regulatory requirements for consumables or critical supplies such as blood gas reagents, the laboratory must have processes and procedures documented. Records should be maintained with the date received, lot number, expiration dates, if acceptance criteria were met, date placed in service, expiration date, and disposition date if not used.

Equipment

The equipment QSE incorporates a management plan for selection and acquisition (SQ), installation qualification (IQ), identification, validation, re-verification (also referred to as performance qualification [PQ]), calibration, use, maintenance, service, repair, and decommissioning. The QSE also includes a process for operational qualification (OQ). Calibration is addressed with equipment quality control in the process control QSE. Computer system hardware, middleware, and software are also included in the equipment QSE but are not covered in depth in the chapter. Equipment files and records are developed and maintained for all processes and results. The CLSI document on equipment outlines multiple areas to consider and examples to develop a comprehensive equipment management plan. The CLSI guideline addresses all of the activities required from selection through the life of the equipment until it is decommissioned. The plan should include a master list of all equipment with identifying information, calibration and maintenance schedules, individuals responsible, and documentation of periodic review by management and medical director.

Equipment acquisition includes meeting the minimum equipment standards as outlined by the ATS-ERS standards as represented in Tables 12-2 and 12-4 for spirometers. Additional guidance may be found in the ATS-ERS standards for other types of equipment. The selection process may include a list of acceptable vendors, development of a product-evaluation matrix, equipment evaluation (on-site is preferred), written acceptable limits of accuracy and precision, information management options, computer standards, warranty and service agreements, and training options (Box 12-3). The initial selection may include an on-site evaluation with a comparison of old and new equipment.

Installation, Validation, and Verification

Selection for equipment location in the pulmonary laboratory includes consideration of environmental conditions that impact acceptable function. For example, equipment specifications outline temperatures appropriate for spirometry equipment. A plethysmograph is sensitive to pressure changes in the room, vibrations close to the device, or other significant changes such as a fan blowing on the device. The installation process should include a correlation study between old and new equipment. Correlation studies are required for blood gas equipment by CLIA’88 and are also prudent to complete with pulmonary function equipment. Studies have shown bias between vendors and equipment within a specific vendor. Equipment function is validated initially to determine compliance with manufacturer’s specifications, expected accuracy and precision, and quality control standards. The question to answer is: Have the requirements for the intended use or application been fulfilled? Verification means that the specified requirements are fulfilled such as calibration verification. All institutional, regulatory, or governmental requirements should be considered during the installation process (Box 12-4).

Box 12-4   Installation Process Verification Activities:

• Development of an installation manual

• Equipment validation performed by the manufacturer

• Equipment validation performed by the pulmonary laboratory staff

• Biomedical safety checks performed by an internal or external biomedical department (may be included in the preventive maintenance provided by the manufacturer)

• Validation of selected reference values by the laboratory under medical supervision (i.e., medical director)

(From Blonshine S, Mottram CD, Berte LM, et al. Application of a quality management system model for respiratory services: Approved guidelines. 2nd ed. CLSI document, HS4-A2. Wayne, PA: Clinical and Laboratory Standards Institute; 2006.)

Equipment Maintenance

An equipment maintenance plan should be developed and approved. If maintenance is provided by an outside contractor, it should be in compliance with the institutional plan. The type and complexity of instrumentation for a specific test determines the long-term and short-term maintenance that will be required. Daily maintenance includes replacing disposable items such as filters and gas conditioning devices. Preventive maintenance is scheduled in anticipation of equipment malfunction to reduce the possibility of equipment failure. Corrective maintenance or repair is unscheduled service that is required to correct equipment failure. These types of failures are often detected by QC procedures or unusual test results. Familiarity with the operating characteristics of spirometers, gas analyzers, Dlco systems, plethysmographs, metabolic systems, and application software requires manufacturer support and thorough documentation. Accurate records are essential to a comprehensive maintenance program. Documentation of procedures and repairs is required by most accrediting organizations. Upgrades to application software should be considered an essential component of equipment maintenance in the pulmonary function laboratory. Software upgrades generally require a re-verification process. Re-verification is often required after preventive maintenance and repairs (Table 12-5).

Table 12-5

Maintenance and Problem Log

Date	Time	Maintenance or Problem	Resolution	Technologist

(Courtesy Mayo Clinic, Rochester, MN)

When equipment is removed from service (decommissioning), the date should be documented along with the final disposition of the equipment. Special attention should be given to items such as equipment with biohazard material requiring decontamination or hard drives with patient information.

Process Management

The process management QSE includes multiple areas: analysis, design, and documentation of the pulmonary laboratory path of workflow, process validation or verification, process control or quality control, and a process to make changes to established processes. This should be a major component in the pulmonary laboratory quality plan, record keeping, and procedure manual because of the impact on the accuracy and precision of test results.

QC is essential in the operation of a pulmonary function laboratory to obtain valid and reproducible data. The type of equipment used (e.g., volume-based versus flow-based spirometer) often determines the specific procedures that are required for calibration and QC. For example, both flow-based and volume-based spirometers require calibration with a 3-L syringe, but only volumetric spirometers need to be checked for leaks. The number and complexity of the tests performed may also dictate which equipment and methods are used. Methods and equipment that have been validated in the scientific literature should be used whenever possible. QC is usually easier to perform when standardized techniques are used. The final QC plan includes the schedule of activities, QC methods, procedures for performing the QC methods, tolerance limits, and corrective action. It also must meet accreditation and regulatory requirements.

Control Methods: Mechanical and Biologic

A control is any known test signal for an instrument that can be used to determine its accuracy and precision. Controls or control materials must be available for spirometers, gas analyzers, Dlco systems, plethysmographs, blood gas analyzers, metabolic systems, and other instruments. Because many laboratories use computerized pulmonary function or blood gas analyzers, controls are required to ensure that both software and hardware are functioning within acceptable limits. In many instances, application software is used to record and evaluate control runs (e.g., automated blood gas analyzers). Control methods may vary from a mechanical control such as use of 3-L syringes for spirometers to commercially prepared materials for blood gas analyzers. Another control method is the use of biologic controls, which are test subjects for whom specific variables have been determined.

Quality Control Tools and Materials

Syringes

A large-volume syringe is the most common and frequently used mechanical QC tool used in pulmonary function testing. Syringes used for calibration should be accurate to within ±15 mL or ±0.5% of the stated volume (i.e., 15 mL for a 3-L syringe). Accuracy of calibration syringes should be verified annually. Several companies provide the service globally with the device in Figure 12-1, A and B. Syringes can be checked for leaks simply by occluding the port and trying to empty the syringe. Some laboratories use two syringes: one to calibrate and another to verify volume accuracy. This is highly recommended to exclude the syringe as a source of error. An alternative to have a 3-L and 7-L syringe is depicted in Figure 12-1.

A syringe of at least 3-L volume (see Figure 12-1) should be used to generate a control signal for checking spirometers. A 3-L syringe can be used to verify volume-displacement spirometers and associated deflection of mechanical recorders. A large-volume syringe may also be used to check the volume accuracy of flow-based spirometers. Computerized systems often have the user inject (or withdraw) a 3-L volume to calibrate the spirometer, and then immediately perform additional injections to verify the calibration. Some portable flow-based spirometers (i.e., those using disposable flow sensors) do not provide for calibration but do allow checking or verification of a stored calibration. The calibration or verification should minimally be completed each day testing is completed.

Sine-Wave Rotary Pumps

Sine-wave rotary pumps produce a biphasic volume signal. A biphasic or sine-wave signal may be useful for checking volume and flow accuracy for both inspiration and expiration. A rotary-drive syringe may be useful for checking the frequency response of a spirometer or to evaluate a spirometer’s ability to adequately record tests such as the MVV. Sine-wave pumps are also commonly used in the calibration of body plethysmographs(Figure 12-2).

Computerized Syringes

Computerized syringes are also available for assessing the accuracy of commonly measured parameters such as forced expiratory volume (FEV₁) and FEF_25%–75%. These syringes use a built-in microprocessor to time the volume injection and calculate the flows. The microprocessor displays volume and flows for comparison with those produced by the spirometer. A computerized syringe provides a 3-L volume for calibration or volume checks and tests accuracy for commonly reported flows.

Computer-Driven Syringes

Computer-driven syringes incorporate large-volume syringes with a computer-controlled motor drive. Computerized syringes are usually used only by equipment manufacturers or for research applications. The ATS has developed a series of standard waveforms that may be used to drive a computer-controlled syringe. These waveforms are used to validate spirometers and peak flowmeters.

Explosive Decompression Devices

Explosive decompression devices simulate the exponential flow pattern of a forced expiratory maneuver. Such devices use compressed gas, such as carbon dioxide (CO₂), released through an orifice. The gas from the device is injected into a spirometer to generate a simulated forced exhalation. The primary advantage of these devices is that they allow flow and volume signals to be reproduced. When the control signal can be reproduced, both accuracy and precision can be assessed. Using a gas other than air may not work properly with certain types of flow-sensing spirometers.

Dlco Simulator

A Dlco simulator is a commercially available device, as depicted in Figure 12-3. This simulator uses precision gas mixtures to allow repeatable Dlco measurements at different levels (e.g., high Dlco, low Dlco) to simulate Dlco results over the range of expected results for patient testing. Two large-volume syringes are included; an adjustable 5-L syringe provides measured inspiratory volumes. A smaller second syringe is loaded with one of the precision gases; this gas is “exhaled” at the end of the breath-hold interval and sampled by the gas analyzers. Application software calculates the expected Dlco with the known gas concentrations (inspired and expired) along with the inspired volume, breath-hold time, and environmental conditions. The measured Dlco is then compared with the expected value and the percent error reported. By using different precision gases and varying the inspired volume, a range of expected Dlco values can be generated. This type of simulator is useful for all laboratories to identify the source of error in Dlco measurements and may also be used for performance qualification when installing new equipment. Use of the device is standard practice by manufacturers developing Dlco systems and evaluation of systems before shipment. It is also very useful for large laboratories with multiple Dlco systems, multicenter research applications in which accurate Dlco measurements are critical, and accreditation or regulatory programs. Studies have shown that the Dlco simulator data correlates with biologic data but may be more sensitive to identify the source of the problem or error.

Plethysmograph volumes can be verified with an isothermal bottle. This device is available commercially (Figure 12-4) or can be constructed. An isothermal volume analog can be constructed from a glass bottle or jar of 4–5 L (for adult-size plethysmographs). The jar is filled with metal wool, usually copper or steel. The metal wool acts as a heat sink (see Figure 12-4) so that small pressure changes within the jar can be measured with minimal temperature change. The mouth of the bottle is fitted with two connectors. One connector attaches to the mouth shutter (or the patient connection). The other is attached to a rubber bulb with a volume of 50–100 mL (e.g., the bulb from a blood pressure cuff). The actual volume of the lung analog can be determined by subtracting the volume of the metal wool from the volume of the bottle. The volume occupied by metal wool is calculated from its weight times its density. The gas volume of the empty bottle may be measured by filling it with water from a volumetric source. The volume of the connectors and rubber bulb should be added to the total volume.

Quality Control Concepts

Two concepts that are central to quality control are accuracy and precision. Accuracy may be defined as the extent to which measurement of a known quantity results in a value approximating that quantity. For most laboratory tests, repeated measurements of a control are made and the mean, or average, is calculated. If the mean value approximates the known value of the control, the instrument is considered accurate.

Precision may be defined as the extent to which repeated measurements of the same quantity can be reproduced. If a control is measured repeatedly and the results are similar, the instrument may be considered precise. Precision is often defined in terms of variability based on the standard deviation (SD) of a series of measurements (Box 12-5).

Box 12-5

The Mean and the Standard Deviation

Calculation of the mean and the standard deviation

The mean (< ?xml:namespace prefix = "mml" />X¯) and standard deviation (SD) are computed to determine the variability of a series of values. Performing multiple measurements of the same quantity (i.e., the “control”) allows the mean to be determined and precision to be expressed by the SD of the measurements. Assuming that all of the values sampled are normally distributed, 68.3% of them will be within ±1 SD of the mean, 95.5% will be within ±2 SD, and 99.7% within ±3 SD. When the mean and standard deviation have been determined for a series of measurements, subsequent values may be checked to see whether they are “in control.” Values between ±2 and ±3 SD from the mean should occur 5% of the time only (i.e., random error), and values more than ±3 SD from the mean should occur less than 1% of the time.

The mean (X¯) may be calculated as:

X¯=∑(X)N

where:

∑ = a symbol meaning “the sum of”

X = individual data values

N = number of items sampled

The SD is calculated as follows:

SD=∑(X)2N

where:

X² = deviations from the mean (X–X¯) squared

N = number of items sampled

If the SD is computed from a sample of 30 data points or less, N-1 is usually substituted for N.

Example calculation of the mean and SD for a series of Pco₂ values:

Sample	Pco2 (mm Hg)	Deviation from the Mean (X)	Deviation Squared (X2)
1	39	− 0.9	0.81
2	40	0.1	0.01
3	43	3.1	9.61
4	42	2.1	4.41
5	39	− 0.9	0.81
6	38	− 1.9	3.61
7	40	0.1	0.01
8	41	1.1	1.21
9	38	− 1.9	3.61
10	39	− 0.9	0.81
Total	399		24.90
Mean	39.9

SD=24.910−1 =2.77 =1.66

The range of Pco₂ values (in this example) within 2 SD of the mean is 39.9 ± (2 × 1.66), or from 36.6–43.2 mm Hg.

Accuracy and precision are desirable but may not always be present together in the same instrument. For example, a spirometer that consistently measures a 3-L test volume as 2.5 L is precise but not very accurate. A spirometer that evaluates a 3-L test volume as 2.5, 3.0, and 3.5 L on repeated maneuvers produces an accurate mean of 3.0 L, but the measurements are not precise. Determining both the accuracy and precision of instruments such as spirometers is critical because many pulmonary function variables are effort-dependent. The largest observed value, rather than the mean, is often reported as the best test (see Chapter 2). Reporting the largest result observed is based on the rationale that the subject cannot overshoot on a test that is effort-dependent. Other pulmonary function tests, such as the Dlco, are reported as an average of two or more acceptable maneuvers; in these instances, precision of the measuring devices (e.g., gas analyzers) needs to exceed the normal physiologic variability of the parameter being measured.

For instruments such as blood gas analyzers, accuracy is determined by measuring an unknown control and comparing the results with a large number of laboratories using similar equipment and methods. This is commonly referred to as proficiency testing (PT). Precision for blood gas analyzers is determined by checking the day-to-day variability of controls and expressing the variability in terms of the standard deviation. These general principles for assessing accuracy and precision can be applied to most types of pulmonary function equipment.

Calibration is the process in which the output signal from an instrument is adjusted to match a known input. This may be accomplished by one of several methods:

1. Adjustment of the analog output signal from the primary transducer (e.g., spirometer bell, flow sensor, gas analyzer)

2. Adjustment of the sensitivity of the recording device (specifically mechanical recorders)

3. Software correction or compensation

Calibration involves adjustment of the instrument (or its signal). It should not be confused with verification or QC. QC assesses function of the instrument after it has been calibrated. Most pulmonary function systems use software-based calibration but allow for adjustment of analog outputs as well.

Spirometry Calibration and Mechanical Quality Control

Spirometers that produce a voltage signal by means of a potentiometer (see Chapter 2) normally allow some form of “gain” adjustment so that the analog output can be matched to a known input of either volume or flow. For example, a 10-L volume-displacement spirometer may be equipped with a 10-V potentiometer. The potentiometer amplifier would be adjusted so that 0 V equals 0 L (zero), and 10 V equals 10 L (gain). The calibration could be verified by setting the spirometer at a specific volume and noting the analog signal (e.g., 5 L should equal 5 V).

A second technique is adjustment of the sensitivity of the recording device. This method is used for older spirometers equipped with mechanical recorders (e.g., kymographs, X-Y plotters, or strip chart recorders). In these devices, a known volume is injected into the spirometer and deflection of the recording device is adjusted to match the volume. For example, a strip chart recorder is turned on and has its pen adjusted to read 0 L when the spirometer is empty. A 3-L volume is then injected. The gain of the recorder is adjusted so that the tracing deflects to the 3-L mark on the graph paper. This method is appropriate when the recorded tracing is to be manually measured but is seldom used because most modern spirometers use computerized output (display or printer). The ability to evaluate a spirometer’s accuracy with a mechanical recorder may be useful for checking a computerized system.

Most spirometer systems are computerized. In computerized systems, the signal produced by the spirometer is often corrected by applying a software calibration factor. A known volume or flow is injected into the spirometer with a large-volume (usually 3-L) syringe. A correction (i.e., calibration) factor is calculated based on the measured versus expected values:

Correction factor=Expected volumeMeasured volume

The correction factor derived by this method is then stored in memory and applied to all subsequent volume measurements. For example, if a syringe with a volume of 3.00 L were injected into a spirometer and a volume of 2.97 L recorded, the correction factor would be as follows:

1.010=3.00L2.97L

The correction factor 1.010 would then be used to adjust subsequent measured volumes. This method assumes that the spirometer’s output is linear and that the same factor would be correct for any volume, large or small. Most automated spirometers allow the correction factor to be verified by re-injecting a known volume, usually 3 L. Taking into account the accuracy of the syringe (0.5%), after calibration the spirometer should display a volume of 3.0 L ±3.5%. Three and one-half percent of a standard 3-L volume means that the spirometer should read 3.00 ±0.105 L (range, 2.895–3.105 L).

Care should be taken that the gas in the syringe, which is at ambient temperature (ATPS) is not “temperature corrected” by the software. Many computerized spirometers provide software functions specifically for calibration and verification. This allows the use of a 3-L syringe without applying corrections that are necessary when patients are tested. Inaccurate temperature corrections produce an erroneous correction factor. Ambient temperature should be available from an accurate thermometer, both for calibration and for testing. If the ambient temperature changes significantly, the temperature used by the software should be updated, or recalibration may be needed. Many spirometers automatically measure ambient temperature. These devices should be checked regularly to verify that appropriate temperature corrections are being applied. The calibration syringe should be maintained at the same environmental conditions as the spirometer.

Other factors that may influence establishment of the software correction value include the accuracy of the large-volume syringe and the speed with which injections are performed. An inaccurate syringe or leaks in the connection to the spirometer may produce erroneous software corrections.

Some spirometers, particularly those that are flow-based, may require that the calibration volume be injected within certain flow limits. The ATS-ERS guidelines recommend a range from 0.5 to 12 L/sec with injection times of about 6 seconds and less than 0.5 seconds. Flow-based spirometers that measure both inspiratory and expiratory volumes require the syringe volume to be injected and withdrawn. This allows separate correction factors for inspired and expired gas to be generated. Many flow-sensing spirometers also require a “zero” before measuring exhaled volume. This means that the flow sensor must be held motionless (so there is no flow through it) while the software adjusts the output of the sensor to equal zero. Lack of a zero flow baseline is a potential cause of a falsely increased or decreased FVC caused by a zero drift. If an in-line bacteria filter will be used for testing, calibration should be performed with the device in place.

Spirometers that use disposable flow sensors may or may not allow calibration. Many disposable sensors are calibrated during manufacture and are coded so that the spirometer software applies appropriate correction factors (see Chapter 11). If these types of spirometers provide for user calibration, they should be calibrated at least daily. At a minimum, a daily calibration check (see the following paragraphs) should be performed with a sensor from the lot used for patient testing.

Quality control of spirometers is closely related to calibration, and the two are sometimes confused. An important distinction is that calibration (i.e., adjustment) may or may not be needed, but QC should be performed on a routine basis. Calibration, whether it includes the output of the spirometer, recorder sensitivity, or generation of a software correction factor, involves adjustment of the device to perform within certain limits. QC is a test performed to determine the accuracy, precision, or both of the device with a known standard or signal. Various control methods (e.g., signal generators) are available for spirometers.

Calibration for spirometer volume measurements should be performed at least once each day that the device is to be used. For field studies, accuracy may need to be checked more often. Frequent checks are recommended for industrial applications or epidemiologic research, especially if the spirometer is moved or used for a large number of tests or if the ambient temperature and humidity change rapidly. The accuracy of any spirometer can be calculated as follows:

% Error=Expected Volume−Measured VolumeExpected Volume×100

where:

Expected volume = known syringe volume (usually 3 L)

Measured volume = volume recorded for the test

The maximum acceptable error for spirometers, according to ATS-ERS recommendations, is ±3.5% or ±65 mL, whichever is larger. Table 12-2 lists the minimal requirements for spirometers (excluding the accuracy of the 3-L syringe). If the error exceeds these limits, careful examination of the spirometer, recording device, software, most recent calibration, and testing technique should be performed (Box 12-6).

Flow-based spirometers should have their accuracy checked (calibration performed) using at least three different flows ranging from 0.5–12.0 L/sec. This can be accomplished easily by injecting the 3-L volume over intervals of less than or equal to 0.5 seconds up to about 6 seconds. At each flow, the volume accuracy of ±3.5% should be maintained.

Linearity of volume-based spirometers should be verified at least quarterly. Volume-displacement spirometers should be checked in 1-L increments across their volume range (i.e., 0–8 L). A 3-L syringe injection performed when the spirometer is nearly empty or nearly full should yield accurate results. The linearity of flow-sensing spirometers should be tested weekly by injecting a series of 3-L volumes at low, moderate, and high flows. Different flows can be generated by varying the speed at which the syringe is emptied. Applying different flows and measuring the resulting volumes may indicate if the spirometer (and its software) is accurate across the range of flows. For example, three different injection times, 0.5–1.0 seconds, 1.0–1.5 seconds, and 5.0–6.0 seconds, may be used with a 3-L syringe to simulate a wide range of flows. The variance between volumes at low and high should not exceed 90 mL or 3% (see Figure 12-5).

In addition to checking the volume and flow accuracy of spirometers, several other important aspects of QC require routine evaluation. Leak checks should also be performed on volume-based spirometers daily before assessing volume accuracy. The spirometer should be filled with air to approximately half of its capacity and 3 cm H₂O pressure applied with the breathing port occluded. The pressure may be generated by a weight or spring. Any volume loss greater than 30 mL/min is a significant leak and should be corrected before calibration or patient testing.

To assess flow resistance, the back-pressure from a spirometer should be less than 1.5 cm H₂O up to flows of 14 L/sec. Resistance to flow is measured by placing an accurate manometer or pressure transducer at the patient connection and applying a known flow. This is easily accomplished with flow-sensing devices but somewhat difficult with volume-displacement devices. Measurement of flow resistance is normally performed only when there is some reason to suspect that the spirometer is causing undue resistance. The total resistance requirement must be met with all tubing, valves, and filters in place.

Frequency response refers to the spirometer’s ability to produce accurate volume and flow measurements across a wide range of frequencies. Frequency response is most critical for peak expiratory flow (PEF) and maximal voluntary ventilation (MVV) maneuvers. Frequency response is usually evaluated by means of a sine-wave pump or computer-driven syringe. It should be measured as part of the manufacturer’s validation and rechecked if the spirometer is suspect.

Flow-sensing spirometers directly measure flow and indirectly calculate volume by integration or counting volume pulses. It may be necessary to assess the flow accuracy of such devices. Inaccurate measurement of flow usually results in inaccurate volume determinations. A rotameter (a large calibrated flow-metering device) may be used in conjunction with an adjustable compressed gas source to supply a gas at a known flow to the device. A weighted volume-displacement spirometer, such as a water-seal type, can also be used to generate a known flow. Most commercial flow-sensing spirometers use a volume signal (e.g., a 3-L syringe) to perform software calibration/verification as previously described. It may be useful to check the flow signal from the spirometer at different known flows if the volume accuracy is observed to vary with flow.

Recorders and displays should meet the ATS-ERS recommendations in Table 12-4. Printed records or computer-generated displays of spirometry signals are required for diagnostic functions, validation, or when waveforms are to be measured manually. Printed copies of volume-time or flow-volume graphs should be available for diagnostic spirometry. Flow-volume curves should be plotted with expired flow in the positive direction on the vertical (Y) axis and expired volume from left to right on the horizontal (X) axis. A flow-to-volume aspect ratio of 2:1 should be maintained (i.e., 2 L/sec flow for each 1 L of volume). Accurate recorder speed and volume sensitivity are particularly important if FEV₁ or other flows are calculated manually. Recorder accuracy should be checked at least quarterly, and the accuracy of the timer should be within 2% of the stated value. Paper speed of strip chart recorders can be easily checked with a stopwatch. Mechanical recording devices (e.g., kymographs, strip-chart recorders) may require repair or replacement of drive motors if paper speed is determined to be inaccurate. Most pulmonary function systems are computerized, and printed tracings are generated by ink-jet, thermal, or laser printers. The output of these devices should adhere to the recommended scale factors but often do not. In effect, it may be difficult or impossible to check the timing during forced spirometry using computer-generated tracings.

QC for spirometers should be performed as if a patient were being tested. The 3-L syringe should be connected to the patient port, with the circuitry used for the actual test. Spirometer temperature correction may need to be set to 37° C (i.e., no correction applied). Most computerized spirometers provide a specific routine for volume checks or calibration that disables temperature corrections. In some systems, temperature correction cannot be disabled. In these spirometers, injection of 3 L at ATPS results in a reading greater than 3 L because the system attempts to “correct” the volume to body temperature (BTPS). For water-seal spirometers, the syringe should be filled and emptied several times to allow equilibration with the humidified air in the device. Some flow-sensing spirometers require a length of tubing between the flow sensor and syringe to reduce artifact caused by turbulent flow in the syringe. If an in-line bacteria filter is used, volume verification should be performed with it in place. The action to be taken if controls exceed specified limits should be documented in the procedure manual.

Gas Analyzers and Dlco Systems

Accurate analysis of inspired and expired gases is required to measure lung volumes by dilution methods, Dlco, and gas exchange during exercise or metabolic testing. The validity of these tests depends on accuracy of both the spirometer and gas analyzers used. Various types of gas analyzers are commonly used in pulmonary function testing (see Chapter 11). Calibration refers to the process of adjusting analyzer output to match the input of a known concentration of gas. QC refers to a method for routinely checking the accuracy and/or precision of the gas analyzer. Important factors related to calibration techniques for gas analyzers include the following.

Physiologic Range

Many gas analyzers are not linear or exhibit poor accuracy over a wide range of gas concentrations. Analyzers should be calibrated to match the physiologic range over which measurements will be made. For example, an oxygen (O₂) analyzer may be used to measure fractional concentrations from 0.00–1.00, representing a wide physiologic range. If the O₂ analyzer is to be used for exercise tests in subjects who are breathing air, an appropriate calibration range might be from 0.12–0.21. This narrow interval represents the physiologic range of expired O₂ likely to be encountered during an exercise test in a patient breathing air. Reducing the physiologic range of an analyzer generally allows better accuracy and precision. Some types of analyzers provide range adjustments for this purpose. Calibration gases should represent the extremes of the physiologic range. In the example of the O₂ analyzer described, air and a gas containing 12% O₂ would be appropriate.

Sampling Conditions

Gas analyzers should be calibrated under the same conditions that will be encountered during the test. Analyzers that are sensitive to the partial pressure of gas may be affected by the sample flow rate. For some tests, gas is sampled continuously from the breathing circuit using a pump (e.g., breath-by-breath gas analysis during exercise). Sample flow through the pump must be adjusted before calibration and then left unchanged during sampling. If gas flow stops before analysis is actually performed (e.g., some types of Dlco systems), sample flow is usually not critical. A gas analyzer in this type of system should be calibrated under conditions of zero flow. Measurement errors may occur if an analyzer is calibrated and then configuration of the sampling circuit is changed. This may happen if tubing, valves, or stopcocks are added or changed. Any gas conditioning devices, such as those used for CO₂, H₂O vapor, or dust, should be in place during calibration as well. If a CO₂ or H₂O absorber is changed, calibration should be repeated. Gas conditioning devices should be changed at the frequency recommended by the manufacturer.

Two-Point Calibration

The most common technique for analyzer calibration, two-point calibration, involves introducing two known gases. One gas is typically used to “zero” or adjust the low end of the range, whereas the second gas is used to “span” or adjust the high end of the range. Adjusting the span is actually setting the gain of the analyzer so that a known input produces a known output. For some pulmonary function tests, the gas to be analyzed is not normally present in expired air (e.g., helium [He], CO, or neon [Ne]). For such tests, air may be used to zero the analyzer. Calibration gas (cal gas) representing the other end of the physiologic range may be used to span the analyzer. He dilution FRC and Dlcosb are examples of such tests. He and CO analyzers are zeroed by drawing ambient air into the measuring chambers. He and CO are assumed to be absent from the atmosphere. Compressed air from a cylinder or other source may instead be used if there is concern about trace amounts of other gases (e.g., CO₂, CO, argon) in the ambient air. Then calibration gas containing a known concentration of the gas to be analyzed is introduced. The analyzer gain is adjusted to match the known concentration. The calibration gas approximates the concentration to be analyzed during the test. The analyzer may then be re-zeroed and the entire process repeated to verify the calibration. A similar technique may be used with two gases of known concentration if the expirate normally contains varying concentrations of the gas. For example, air and 12% O₂ may be used to perform a two-point calibration of an O₂ analyzer for exercise testing. Depending on the stability of the analyzer, calibration may need to be repeated before (and after) each test or measurement. Gas analyzers should be calibrated before each patient for gas dilution lung volumes, Dlco, exercise tests, and metabolic studies. Gas analyzers used for monitoring (e.g., capnographs) should be calibrated on a schedule appropriate for the extent of use. Calibration should be performed according to the manufacturer’s recommendations. The accuracy of the calibration gas should reflect the necessary accuracy of the measurements involved. For exercise or metabolic studies, calibration gases should be accurate to at least two decimal places (i.e., hundredth of a percent). Calibration gases may require verification by an independent method.

Multiple-Point (Linearity) Calibration

An assumption made by a two-point calibration is that analyzer output is linear between the points used. To verify linearity or to determine the pattern of nonlinearity, three or more calibration points must be determined (see Figure 12-5). A multiple-point calibration is performed in a manner similar to the two-point calibration except that concentrations of known gases across the range to be analyzed are checked and plotted. If multiple points are determined, regression analysis may be used to determine the slope (or type of curve) relating the measured gas concentrations to the expected gas concentrations. A spreadsheet or graphing calculator can be used to analyze the data points. If the analyzer is linear, the points plotted will approximate a straight line. A minimum of three points (i.e., gases) is required to demonstrate linearity. If the analyzer is nonlinear, a calibration curve must be constructed to correct the results. In most instances, an equation describing a nonlinear curve can be generated. This equation can then be used either manually or by software to correct analyzer readings. Computerized systems often use an equation or a table of points representing a calibration curve. This allows an analyzer to be calibrated using two points only. Most nonlinear analyzers incorporate electronics that linearize their output. Linearity of analyzers used for Dlco, lung volumes, exercise, and metabolic studies should be assessed at least quarterly.

PF Tip 12-3

The 3-L calibration syringe can often be used as a lung model to evaluate the accuracy of gas analyzers used for lung volumes or Dlco tests. For example, a Dlco test can be simulated by setting the calibration syringe to a volume of 1 L and then connecting it to the patient port of the pulmonary function system. The syringe is then withdrawn to “inspire” 2 L of test gas. After a 10-second breath hold, the syringe is emptied. This simulates a lung with a total volume (TLC) of 3 L and an inspired volume (VC) of 2 L. With this technique, the Dlco should be close to zero because no diffusion occurred in the syringe.

Dlco Systems

Gas analyzer accuracy and linearity is particularly important in pulmonary function systems that measure Dlco. For measurement of Dlco by the single-breath method, analyzer linearity is more critical than the absolute measurement of gas concentrations. Small errors in which the analyzer outputs for CO and tracer gas are not linear with respect to one another can result in significant errors in the calculation of Dlco. The nonlinearity of each of the analyzers should be 0.5% or less of the full scale. Analyzers used for Dlco tests also need to be stable, with minimal drift (<±0.5% of full scale) between calibration and testing. To detect drift or similar problems during testing, the actual readings of the analyzers should be displayed. Removal of water vapor and CO₂ is usually accomplished by chemical absorbers or related devices (see Chapter 11). Some analyzers use software corrections for the effects of water vapor and CO₂ rather than physically removing or altering the interfering gases.

Quality control of gas analyzers can be performed by submitting known concentrations of gases to the analyzer, by testing a lung analog, or by using biologic controls. Several gases with concentrations that span the range of the analyzer can be maintained. This can be a costly means of QC for pulmonary function laboratories. A simpler technique is to prepare serial dilutions of a known gas using a large-volume syringe. The syringe may be the type used for volume calibration. For example, 100 mL of He and 900 mL of air may be mixed in a syringe to produce a 10% He mixture which is then injected into the analyzer. Subsequently, 100 mL of He may be diluted in 1000 mL, then 1100 mL, and so on, with the expected concentrations calculated as follows:

Expected test gas=Volume of test gasTotal volume of gas

where:

Total volume of gas = test gas + added air + syringe dead space

As each dilution is analyzed, the meter reading is recorded and plotted against the expected percentage (see Figure 12-5). This method is simple and available in most laboratories. Care must be taken when preparing samples so that air does not leak into the syringe, further diluting the test gas. The volume of air in the syringe connectors (i.e., dead space) must be included when calculating the dilution of the test gas. Some calibrated syringes include their dead space volume.

A second method of verifying analyzer performance involves simulating either lung volume or Dlco tests. This may be accomplished using a lung analog. A lung analog is simply an airtight container of known volume. The lung volume simulator is attached at the patient connection with the system set up for a lung volume or Dlco test. A large-volume syringe is used to “ventilate” the lung analog, mimicking the patient’s breathing. The resulting lung volume (e.g., FRC) is compared with the known volume of the analog system. A calibration syringe may also be used by itself as the lung analog. Many calibration syringes feature a locking collar that can be adjusted so that only a portion of the syringe’s volume can be emptied. With a known volume of air in the syringe, the test is performed by filling and emptying the syringe to the starting volume.

Simulation of the Dlcosb maneuver with a calibration syringe can be used to check analyzer linearity. Both the tracer gas (He, Ne, CH₄) and CO are diluted equally in the syringe, and their relative concentrations should be identical if the two analyzers used are linear with respect to one another. This causes the calculated Dlcosb to be near zero (approximately ±0.3 mL CO/min/mmHg with a 3-L syringe). If the two analyzers are not linear in relation to one another, the ratio of tracer gas to CO will not equal 1. Calculated Dlcosb may be either slightly above or below zero. This method tests not only the gas analyzers, but also the volume transducer, breathing circuit, and software. Temperature or gas corrections should be disabled. A linearity check at different dilutions can be performed by varying the volume of the calibration syringe. For each different syringe volume, however, the calculated Dlcosb should be close to zero.

A Dlco simulator is commercially available (see Figure 12-3). This simulator uses precision gas mixtures to allow repeatable Dlco measurements at different levels (e.g., high Dlco, low Dlco). Two large-volume syringes are included; an adjustable 5-L syringe provides measured inspiratory volumes. A smaller second syringe is loaded with one of the precision gases; this gas is “exhaled” at the end of the breath-hold interval and sampled by the gas analyzers. Application software calculates the expected Dlco with the known gas concentrations (inspired and expired) along with the inspired volume, breath-hold time, and environmental conditions. The measured Dlco is then compared with the expected value and the percent error reported. By using different precision gases and varying the inspired volume, a range of expected Dlco values can be generated. This type of simulator is useful for large laboratories with multiple Dlco systems, or for multicenter research applications in which accurate Dlco measurements are critical.

Some computerized pulmonary function systems may make lung simulators more challenging to use, although it has become common practice for manufacturers to develop Dlco systems with a lung simulator. The device may also be used to verify the function of new systems before shipment and in the laboratory during installation. When using a Dlco simulation device, the laboratory staff must understand how calculations are completed both in the testing software of the PF equipment and the Dlco simulation calculation software to avoid errors. The software may be designed to make all necessary corrections for human subjects, giving erroneous results when a simulator is used. However, if the software reports gas analyzer values, the accuracy and linearity of various dilutions can usually be checked (Box 12-7).

Body Plethysmographs

The calibration techniques described here apply primarily to variable-pressure, constant-volume plethysmographs. Flow-based plethysmographs may require slightly different calibration procedures for the box transducer. Mouth pressure and pneumotachometer calibrations are similar for both types of plethysmographs.

Mouth Pressure Transducer

Calibration is done by connecting the pressure transducer to a water manometer or a similar device that can generate an accurate pressure. The manometer is a fluid-filled, U-shaped tube with a calibration scale that allows very accurate pressures to be generated (Figure 12-6). Some plethysmographs use a weighted piston to produce a calibration pressure signal. The mouth pressure transducer should be able to accurately record pressures greater than ±50 cm H₂O at frequencies of 8 Hz or more. The actual mouth pressures encountered are often less than this level. Air is injected into one port of the U-tube manometer. For example, a small volume of air may be introduced to cause a deflection of 5 cm. In effect, this creates a difference of 10 cm between the two columns of the manometer. The gain of the mouth pressure amplifier is then adjusted so that its signal display deflects by an amount equivalent to 10 cm H₂O/cm. The display device is most commonly a computer screen. The deflection then equals the calibration factor for the mouth pressure transducer. In the previous example, if a pressure change of 10 cm H₂O resulted in a 1-cm deflection on the display, the calibration “factor” would be 10 cm H₂O/cm. Because almost all plethysmograph systems are computerized, the analog output of the transducers is measured and a software correction factor is determined. The correction (or calibration) factor is calculated in a manner similar to that used for spirometer output (see section on spirometers). The correction factor is then applied by the software as the mouth pressure signals are acquired.

Box Pressure Transducer

Calibration of the box pressure transducer is accomplished by closing the door of the plethysmograph and applying a volume signal comparable to what occurs during patient testing. In a plethysmograph of 500–700 L, a volume signal of 25–50 mL is typical. The box pressure transducer should be capable of accurately measuring pressures as small as ±0.2 cm H₂O. The box pressure transducer typically requires a range of up to 5–10 cm H₂O to accommodate large changes in box pressure (e.g., thermal drift) An adjustable sine-wave pump connected to a small syringe is ideal for box calibration. A small volume is pumped into and out of the box (see Figure 12-6). With the pump operating, the gain of the box pressure transducer is adjusted so that volume change in the box causes a specific pressure change. For example, if the pressure signal generated by a 30-mL volume change is adjusted to cause a 2-cm deflection on the display, the box pressure calibration factor would be 15 mL/cm. For computerized systems (most plethysmographs), a software calibration factor is derived rather than an actual adjustment of the displayed signal. In other words, no actual adjustment of the output of the box pressure transducer is necessary; the software correction is simply applied to all signals generated during measurements. The calibration procedure may be repeated by adjusting the pump speed from 0.5–8.0 cycles/sec (Hz). Varying the frequency allows the frequency response of the box and transducer to be checked. The volume deflection or calibration factor should not change at frequencies up to 8.0 Hz. Flow-based plethysmographs may be calibrated similarly. The output of the box flow transducer is adjusted (instead of a pressure transducer) to correspond to a known volume change within the plethysmograph chamber. Flow boxes can be used as pressure boxes simply by occluding the flow sensor in the box wall. For plethysmographs that use this method, a box pressure transducer may need calibration as well.

Plethysmographs are normally calibrated empty. A volume correction for the displacement of gas by the patient can be calculated from the patient’s weight. This correction is then applied in the calculation of results (http://evolve.elsevier.com/Mottram/Ruppel/).

Flow Transducer

The pneumotachometer (flow sensor) may be calibrated by applying either a known flow or a known volume. A precise flow may be generated with a rotameter or similar calibrated flow meter. Most systems, however, calibrate the pneumotachometer with a 3-L syringe. The flow is integrated, and the gain of the flow signal is then adjusted until the output of the integrator matches the 3-L volume, just as is done for most flow-based spirometers. The flow sensor used in the plethysmograph should meet the volume range and accuracy requirements for spirometers (see Table 12-2). Just as for the box and mouth pressure transducers, a software calibration factor may be computed rather than physically adjusting the output of the flow sensor. Pressure-differential, heated-wire, or Pitot tube flow sensors may be used (see Chapter 11). Once the flow sensor is calibrated, both volumes and flows (as needed for measurement of Raw) can be measured.

QC of body plethysmographs may be accomplished with one or more of the following: an isothermal lung analog, fixed resistors, biologic controls, or comparison with gas dilution or radiologic lung volumes.

The accuracy check is performed with an assistant seated in the sealed plethysmograph. The isothermal volume device is connected to the mouthpiece. The mouth shutter is then closed. While the assistant holds his or her breath, the bulb is squeezed at a rate of 1–2 times per second. A P_MOUTH/P_BOX tangent is recorded just as would be done when testing a patient. Thoracic gas volume (V_TG) is calculated as usual, except that PH₂O is not subtracted. The V_TG calculated should equal the volume of the isothermal lung analog (as determined previously) within 50 mL or ±3%, whichever is greater. Correction should be made for the assistant’s volume (based on body weight) plus the known volume of the isothermal lung analog. The procedure may be repeated by squeezing the bulb at 0.5–5.0 cycles/sec to check the frequency response of the box. If the box’s frequency response is “flat,” tangents should not change when the bulb is squeezed at different rates. The lung analog must contain a sufficient mass of metal wool to act as a heat sink (i.e., isothermal). The metal wool “absorbs” changes in temperature that would result from the compression and decompression of gas in the bottle. If there is not enough metal wool, small temperature changes may affect the volume determination. Some manufacturers provide an automated isothermal lung analog that can be placed in the plethysmograph and operated under software control to provide measurements across the range of expected results for both pediatrics and adults (see Figure 12-4).

The accuracy of the box for measuring airway resistance (Raw) can be assessed with known resistances. A resistor can be made using a plug with a small-diameter orifice. Alternatively, a resistor can be constructed from capillary tubes arranged lengthwise in a flow tube. In either case, the pressure drop across the resistor must be measured at a known flow rate. Some manufacturers supply resistors with known resistances. The resistor is then inserted between the pneumotachometer/mouth shutter assembly and a test subject. The subject, whose Raw has been previously measured, then has Raw measured with the resistor in place. The increase in measured Raw should approximate that of the resistor (Box 12-8).

Box 12-8   Common Plethysmograph Problems

Some problems detected by routine quality control of body plethysmographs include the following:

• Leaks in door seals or connectors (pressure boxes)

• Improperly calibrated mouth or box pressure transducers

• Damaged flow sensors (pneumotachometers) or mouth shutter mechanisms

• Excessive thermal drift

• Poor frequency response

• Inappropriate software calibration factors

• Procedural errors (e.g., testing before thermal equilibrium is reached)

A third method of checking plethysmograph accuracy is to compare the V_TG with FRC determined by gas dilution. Correlations greater than 0.90 have been demonstrated between gas dilution and plethysmograph lung volumes in healthy individuals. Differences greater than 10% (in healthy individuals) for volumes measured by plethysmograph and gas dilution are not specific but may indicate equipment malfunction. This method, as well as use of biologic controls, is based on measurements of healthy individuals with normal day-to-day variability. It is important that control subjects perform the breathing maneuvers correctly (see Chapter 3).

Biologic Controls

Biologic controls are healthy subjects who are available for repeated tests. These controls can be laboratory personnel or other individuals who can be tested repeatedly. Using biologic controls does not eliminate other control devices such as large-volume syringes. Although a 3-L syringe can verify volume and flow accuracy of a spirometer, biologic controls can evaluate an entire system, including spirometers, gas analyzers, plethysmographs, and software. A disadvantage of using biologic controls is that pulmonary function varies from day to day. However, by establishing means and measures of variability from repeated tests, real problems with most pulmonary function equipment can be identified (How to Box 12-2).

Control subjects should have normal lung function (i.e., no asthma or other respiratory symptoms) and, ideally, span a range of values. For example, a 64-inch-tall female and a 72-inch-tall male will provide a wide range of values for most pulmonary function parameters. Pulmonary function studies on controls should be performed on a regular basis (weekly or monthly). All tests should use the same protocols applied to the patient population. Control measurements should meet all criteria for acceptability and repeatability. Tests should be performed at the same time of the day to minimize the effects of diurnal variation. If the laboratory has multiple pulmonary function systems, controls may be tested on each instrument on the same day to provide a check of inter-instrument bias.

To provide useful statistics, 20 sets of measurements should be recorded. Pulmonary function variables that are not derived from other measurements should be recorded. These include FVC, FEV₁, FRC, and Dlco. Calculated values such as TLC or DL/V_A can also be used; however, if subsequent control tests show significant differences, it may be unclear which component test is at fault. Typical values selected from metabolic studies may include o₂, CO₂, tidal volume, _E, and respiratory rate at different work rates. See chapter 7 for specifics on metabolic QC performance. A calculator or a computer spreadsheet may be used to perform the simple statistics required (Table 12-6). Most spreadsheets have built-in functions to calculate mean and SD and to allow data to be graphed. The coefficient of variation (CV) may be calculated by dividing the SD by the mean. The coefficient of repeatability (CR) may also be calculated. Separate statistics should be calculated for each control and for separate instruments. Data more than 1 or 2 years old should be evaluated with more recentmeasurements to account for small normal changes in pulmonary function that occur over time (Figure 12-7).

Testing biologic controls is also a means of evaluating gas analyzers. This method may not detect small changes in analyzer performance because of day-to-day variability of lung volumes, Dlco, or resting energy expenditure. Despite variability as high as 10% for Dlco or exercise parameters in healthy patients, gas analyzer malfunctions may be detected. Current research has shown that the variability in biologic controls for Dlco may be as low as 6% if the equipment is well maintained. Biologic controls may be the simplest means of checking automated exercise/metabolic systems that depend on accurate gas analysis. Abnormal results from biologic controls can suggest which component of the gas analyzer may be faulty (see chapter 7 for completed explanation of biologic QC with metabolic systems).

A simple but effective means of checking plethysmograph function is to measure TLC, V_TG, and Raw from biologic control subjects. A series of 20 box measurements when QC is in control (usually done over a period of days or weeks) provides an appropriate mean value for comparison with subsequent results. Day-to-day variability in trained subjects is usually less than 10% for FRC_pleth, so changes in FRC_pleth that are greater than 10% suggest a problem with the box. TLC variability is generally less than 5% with linked maneuvers. This method allows checking of the box, the transducers, the recording devices (if any), as well as the software. A discrepancy between the established mean and an individual QC measurement will detect a problem but may not indicate which component is causing the problem. For example, a control with an established FRC_pleth of 3 L is measured again, and the FRC_pleth is calculated to be 2 L. The biologic control establishes that there is a problem, but the cause of the discrepancy requires further investigation. In this example, either incorrect calibration of box or mouth pressure transducers or a leaky door seal may be the cause. Biologic control data from other methods for lung volume determinations is also helpful in troubleshooting the cause of the discrepancy.

Calibration and Quality Control of Blood Gas Analyzers

Modern blood gas analyzers rely on a microprocessor or computer to control functions such as calibration. The user selects a “calibration schedule” appropriate for the complexity and number of tests performed. For example, in a laboratory that performs many blood gas analyses, automated one-point calibration may be performed every 30 minutes with two-point calibration every 2 hours. Government agencies and some voluntary credentialing organizations have specific schedules for the frequency and type of calibrations that must be performed. Different calibration schedules may be required for different types of a blood gas analyzer (e.g., point-of-care devices versus laboratory instruments).

For traditional blood gas electrodes, calibration involves exposing the gas electrodes (i.e., Po₂, Pco₂) to one or two gases (or liquids) with known partial pressures of O₂ and CO₂. If only one gas is used, the calibration is termed a one-point calibration; if two gases are used, it is called a two-point calibration. Similarly, one or two known buffers may be used to calibrate the pH electrode. Calibration gases spanning the physiologic ranges of the Po₂ and Pco₂ electrodes are used, just as for gas analyzers. Typical combinations include one calibration gas with a fractional O₂ concentration of 0.20 (20%) and a fractional CO₂ concentration of 0.05 (5%). A second calibration gas may have a CO₂ concentration of 0.10 (10%) with an O₂ concentration close to zero.

As for gas analyzers, a “low” gas (or buffer solution) is used to zero or balance each electrode. A “high” gas (or buffer) is used to adjust the gain (also called the slope) of the electrode’s amplifier. “Electronic” zeroing may be done without exposing the electrode to a gas with a partial pressure of zero. Some blood gas analyzers use this method to zero the Po₂ electrode. Computerized systems allow the user to select how frequently and what type of calibration is performed on the gas and pH electrodes. The Po₂ electrode is usually calibrated over a range of 0–150 mm Hg. The Pco₂ electrode is typically calibrated for the range of 40–80 mm Hg. The pH electrode is usually calibrated using buffers with pH values of 6.840 (low) and 7.384 (high).

Most blood gas analyzers use precision gases to calibrate the gas electrodes, even though gas tensions are measured in liquid (i.e., blood). Some difference may exist when partial pressure of gas is analyzed in gaseous versus liquid medium, especially for O₂. The reduction of O₂ at the tip of the polarographic electrode occurs more rapidly in a gaseous medium than in a liquid. If the electrode is calibrated with a gas, its response when measuring a liquid will be to read slightly lower. This difference is termed the gas-liquid factor. Gas-liquid corrections may be clinically important when measuring high partial pressures of O₂, particularly above 400 mm Hg.

Computerized calibration brings calibration gases or buffers into contact with the electrodes. Electrode responses to the calibration gas or buffer are then stored. The microprocessor compares the measured responses to expected calibration values. The computer then “corrects” the zero and gain (for a two-point calibration) so that measured and expected values match. Most computerized blood gas analyzers compare the current calibration results with the previous calibration. The difference between calibrations is termed drift and indicates an electrode’s stability.

Automatic calibrations can be programmed to occur at predetermined intervals. Adjustments are performed automatically, based on the response of the electrodes. Because of this, all conditions for an acceptable calibration must be met before the procedure actually begins. Automated blood gas analyzers check most conditions that may affect accurate results, such as temperature of the measuring chamber. During automatic calibration, inadequate buffer or the wrong calibration gas may cause the microprocessor to correct inappropriately for a properly functioning electrode. A similar problem arises if protein contaminates the electrode tip, altering its sensitivity. The microprocessor adjusts the electrode’s output in an attempt to bring it into range. This process works well for minor changes in electrode sensitivity. However, electrodes cannot be properly calibrated if they are contaminated with protein, if membranes are damaged, or if the electrolyte is depleted. The user must maintain reagents, calibration gases, and electrodes so that automatic calibration can occur successfully. Systematic errors can sometimes be masked by automatic calibration. Contamination of the calibration gases or buffers is a common example. If the microprocessor adjusts electrodes to match a contaminated calibration standard, the calibrations appear normal but analysis of control samples will show differences. Detection of these errors usually requires appropriate QC and proficiency testing (described later in this section). Automated blood gas analyzers reduce variability by controlling calibration as well as sample analysis but require careful attention to function appropriately.

Many blood gas analyzers (such as point-of-care devices) use electronic checks of the function of the various sensors (e.g., Po₂, Pco₂). Some analyzers use a combination of electronic checks and traditional calibration methods. Electronic checks are typically performed immediately before sample processing, and some checks are performed during analysis to detect bubbles, clots, and so forth. Electronic checks are sufficient for ensuring proper functioning of sensors such as optodes, spectrophotometers, or fluorescence quenching devices. The adequacy of electronic sensor checks and traditional calibration methods must be demonstrated by appropriate quality controls and calibration verification.

Two methods of QC for blood gas analysis may be used: tonometry of whole blood and commercially prepared controls. Although tonometry is no longer commonly used in blood gas laboratories, the method is the basis for commercially prepared controls. Interpretation of blood gas QC is the same for either method.

Tonometry

A tonometer allows precision gas mixtures to be equilibrated with either whole blood or a buffer solution. One type of tonometer creates a thin film of blood or buffer by spinning it in a chamber flooded with precision gas. A second type bubbles gas through the blood or buffer; the bubbles create a large surface for gas exchange. In both types, the tonometer is maintained at 37° C and the gas is humidified. The equilibration time is determined by gas flow and volume of control material. A portion of the blood or buffer is then injected into the blood gas analyzer. The expected gas tensions are calculated from fractional concentrations of the precision gas. If whole blood is used as the control material, only Po₂ and Pco₂ can be checked because pH cannot be easily calculated. Blood is ideal for QC of the gas electrodes because its viscosity and gas exchange properties are the same as those of patient samples. For the most precise control of the Po₂ electrode, tonometry is the method of choice. However, pH cannot be accurately calculated for whole blood because its buffering capacity is usually unknown. Tonometry of a bicarbonate-based buffer using a known fractional concentration of CO₂ allows both gas and pH electrodes to be quality controlled. The gas exchange characteristics of this type of buffer make it less useful than whole blood for QC of Po₂ and Pco₂.

Tonometry can be performed inexpensively with pooled waste blood and small amounts of precision gas. Using pooled blood requires special care. All blood specimens must be handled using standard precautions (see the section on infection control and safety section). QC of the pH electrode requires additional tonometry of a buffer. Three levels of control materials spanning the measuring range of the electrode are recommended. Three precision gas mixtures are therefore required.

Accuracy of tonometry is highly dependent on a standardized technique. Sampling syringes should be lubricated and then flushed with the precision gas. Careful attention to the preparation and sampling from the tonometer is required to obtain reproducible results. The values obtained using tonometry may depend on individual technique. Problems that occur with tonometry include contamination of the precision gas resulting from leaky connections, improper temperature control of the chamber, or inadequate gas flow to achieve equilibrium. Because of the complexity required for multilevel controls for all three electrodes, tonometry is no longer widely used.

Commercially Prepared Controls

There are two types of commercially prepared controls: aqueous and fluorocarbon-based emulsions. The aqueous material is usually a bicarbonate buffer. The fluorocarbon-based control material is a perfluorinated compound that has enhanced O₂ dissolving characteristics. Multiple levels (e.g., acidosis, alkalosis, normal) of these materials provide control over the range of blood gases seen clinically. Aqueous and fluorocarbon-based controls usually contain dyes that generate known absorptions when injected into hemoximeters designed for blood. These dyes allow the materials to be used for QC of blood gases and hemoximetry at the same time.

Both types of controls are packaged in sealed glass ampoules of 2–3 mL volume. They require minimum preparation for use. Aqueous and fluorocarbon-based controls can be stored under refrigeration for long periods or at room temperature for day-to-day use. Most aqueous and fluorocarbon-based controls have shelf lives of 1 year or longer. Each requires agitation for 10–15 seconds before use to ensure equilibration with the gas sealed in the ampoule. Care must be taken when handling the glass ampoules because temperature changes (from the hands) can affect the amount of gas dissolved in the liquid, particularly for O₂. If the ampoules are stored at a temperature significantly different from 25°C, the control values for Po₂ may need to be adjusted. Commercially prepared controls may cost more than samples prepared with tonometry. They are convenient to use, however, and may be less susceptible to handling errors than tonometered materials.

One problem with aqueous controls (and to a lesser extent with fluorocarbon solutions) is poor precision of Po₂. The O₂ carrying capacity of these materials is much lower than that of whole blood. Consequently, the Po₂ in the control material changes rapidly on exposure to air. Controls with low Po₂ values (50–60 mm Hg) become quickly contaminated after opening. Aqueous or fluorocarbon controls may produce a wide range of “expected” Po₂ values, limiting their usefulness in detecting an out-of-control device. Some of these difficulties may be overcome by careful statistical evaluation of Po₂ control data as described in this section.

Some analyzers provide automatic measurement of QC materials. In these devices, control materials are contained in cartridges, similar to cartridge-based reagents. Controls are then run on a predetermined schedule with little operator intervention. In most instruments, the auto-QC software can be set to “lock out” the analyzer for those analytes (e.g., pH, Po₂) that fail QC. This prevents the analyzer from being used to report values that might be inaccurate.

A sound statistical method of interpreting “control runs” is necessary to detect blood gas analyzer malfunctions (Box 12-9). The most commonly used method for detecting out-of-control situations is calculating the control mean ±2 SD. A series of runs of the same control material is performed. Twenty to 30 runs provide an adequate base for calculation of the mean and SD (see Evolve http://evolve.elsevier.com/Mottram/Ruppel/). One SD on either side of the mean in a normal distribution includes approximately 67% of the data points. Two SD include 95% of the data points in a normal distribution. Three SD include 99% of the data points in a normal series. A QC value that falls within ±2 SD of the mean is usually considered to be “in control.” If the control value falls between 2 and 3 SD from the mean, there is only a 5% chance that the run is in control. The normal variability that occurs when multiple measurements are performed is called random error. One of 20 control runs (i.e., 5%) can be expected to produce a result in the 2–3 SD range and still be acceptable. In practice, if a control run shows a value that is more than 2 SD above or below the mean, the control is usually repeated. If the second run shows a value within 2 SD of the mean, the first value was probably a random error. Conversely, if the second run produces a result that is similar to the first run (>2 SD on the same side of the mean), the instrument is probably “out of control.” This simple approach works very well for detecting most types of errors that occur in analytic instruments like blood gas analyzers. The same method has also been applied to pulmonary function equipment.

More complex sets of rules have been developed to distinguish true out-of-control situations from random errors. A widely used set of rules is that proposed by Westgard (see Selected Bibliography). The rules are selected to provide the greatest probability for detecting real errors and rejecting false errors. This approach to QC is termed the multiple-rule method. An example of the multiple-rule method may be applied as follows:

These are just some of the rules that may be applied; not all rules have to be used at all times. Rules 1 and 2 detect marked changes in electrode performance, sometimes called a shift, by examining how far from the mean a single control value falls. Rule 3 looks for a shift by comparing two consecutive control runs. Rule 4 looks for shifts in electrode performance by noting excessive variability between consecutive runs. Rules 5 and 6 look for “trends” in instrument response by evaluating an unexpected pattern (i.e., multiple consecutive values on the same side of the mean) in the recent history of control runs.

The multiple-rule method can also be applied when two or more levels of controls are evaluated on the same measurement device (electrode). The rules may be applied by linking multiple levels (e.g., high, normal, low) of control material. For example, if three levels of controls all show values greater than 2 SD above their respective means, it is likely that the electrode (sensor) is out of control.

One problem with a strict statistical approach is that if outliers (i.e., values more than 2 SD from the mean) are sometimes excluded, the SD becomes smaller with repeated calculations. Eventually, valid control data may be rejected. This situation can be managed by including data in the calculations that are clinically acceptable, although they may be more than 2 SD above or below the mean.

When using the multiple-rule method, it is necessary not only to evaluate the mean and SD of the current control run, but also to keep a control history. This is often accomplished by means of a control chart (Figure 12-8), also called a Levey-Jennings plot. A graph for each control is created with the mean ±2 SD on the Y-axis and control run number (or time) on the X-axis. Individual controls are then plotted as they are run to track electrode performance.

To provide adequate QC for a blood gas analyzer, three levels of control materials are normally used. Three levels of control for each of the three electrodes (i.e., pH, Pco₂, and Po₂) require that 9 means and 9 SD must be calculated for each instrument. Controls may also be used for blood oximeters, and this adds more control histories to be managed. When controls are run several times daily, tracking multiple runs can become complex. To simplify this task, computerized QC programs are often used. Such programs are usually included in the software for automated blood gas analyzers. Many laboratory computer systems also support statistical databases for control data. The chief advantages of computerized QC are simplified data storage and maintenance of necessary statistics. Multiple rules can be applied easily to each new control run to detect problems. Computerized records and control charts can be printed. These types of routine QC records are required by many accrediting agencies. (See Evolve for a list of regulatory agencies.)

QC of blood gas analyzers should be performed on a schedule appropriate for the number of specimens analyzed. In most laboratories, controls must be performed daily or more often. In busy laboratories, multiple levels of controls may be required on each shift. QC is also usually required after any significant maintenance is performed in order to verify proper function of the instrument.

Quality control charts, or Levey-Jennings plots, are used to assess the QC. Figure 12-8 illustrates three examples of Levey-Jennings charts for pH, Pco₂, and Po₂. The mean for each control material is plotted as a solid line, and the ±2 SD lines are dashed. The left Y-axis on each graph is labeled with the actual mean and ±2 SD values. Consecutive control runs are plotted on the X-axis. On the pH control chart (top) all values vary about the mean in a regular fashion; the electrode appears in control (all values within ±2 SD) for the 13 measurements plotted. The Pco₂ chart (middle) shows somewhat more variability. Control run 8 shows a value outside of the ±2 SD range. This may be considered a random error because subsequent controls show normal variability about the mean. The Po₂ chart (bottom) shows a trend of decreasing control values. Runs 6 and 7 both produce values of more than 2 SD below the mean. This pattern suggests that the electrode or sensor is malfunctioning and needs to be serviced. By applying multiple rules (see text) to the interpretation of consecutive control runs, with or without charts, most out-of-control situations can be detected.

In addition to providing regular checks of acceptable instrument performance, routine QC establishes the precision (variability) of each measurement. Instrument precision must be determined so that blood gas interpretation can be related to a range of values. For example, the variability of the Po₂ measurement for a specific analyzer may be determined to be ±6 mm Hg (i.e., 2 SD) around a mean of 50 mm Hg. Each Po₂ result (near 50 mm Hg) can then be interpreted with some certainty that it is within 6 mm Hg of the reported value.

Several other techniques related to QC of blood gas analyzers are commonly used. Inter-laboratory comparison of control results can compare the performance of similar instruments for measuring the same lot of control materials. Manufacturers of control materials and some voluntary credentialing agencies provide such inter-laboratory databases. Each laboratory submits its control values for a specific period, usually once per month. The laboratory then receives a report showing its performance related to all participating laboratories. This type of comparison is useful for detecting systematic errors that may go unnoticed when running daily controls. Accreditation standards may also require intra-laboratory comparison for specific analytes.

Documents and Records

The Documents and Records QSE is central to a quality system. It encompasses a management system for document and record creation, control, and retention. Records are generated, based on the documents and activities that are performed in the pulmonary function laboratory. Forms are created to record data or results from all procedures in the path of workflow and each QSE. Specific recommended documents and records are addressed with each QSE. They may be electronic or paper.

The system for creating, reviewing, and approving documents is an example of a process document (see Table 12-1). The document for maintaining a record of procedures, effective dates, and location is an example of a form and could be created in an electronic database, a paper form, or as a spreadsheet. A procedure manual that outlines the steps for each test performed is a required document. Multiple forms may also be required to document other processes and procedures such as quality control procedures and results, equipment maintenance, quality improvement activities, equipment validation procedures, ordering forms, and pre-test instructions. Box 12-2 outlines a typical technical procedure manual for the pulmonary function laboratory.

Information Management

The Information Management QSE provides the guidance for managing information generated in the pulmonary laboratory in either paper or electronic formats and dissemination of the information in adherence to organizational, accreditation, and regulatory requirements. An example of a required procedure is instructions that define the flow of information. The process for meeting confidentiality requirements such as HIPAA should be available to staff members. Security for data access both paper and electronic along with levels of access is determined. This may include assigning one individual with authority to change predicted author sets or primary user configurable files that impact data integrity. Record-keeping processes should also be identified. A process for verifying data integrity during equipment installation, repairs and after software changes is a critical component in the laboratory. Software upgrades can cause unanticipated changes in formulas or how data is obtained and processed. Repeating all QC is recommended as a part of the process to ensure data integrity after any change in hardware of software. Data integrity may also be impacted when transferring data between electronic systems such as from the pulmonary laboratory system to the electronic medical record. This includes agreement of terms and/or abbreviations used in each system. When electronic systems are used to collect and disseminate data, a downtime procedure is required. Systems used for interpretation of the results also require specific process and procedures for managing the data, maintaining standards, and tracking availability of the data for clinical use within specified timeframes. Planning for information management may involve both staff and users of the data to increase satisfaction. Billing processes related to tests performed and interpretations should be outlined to avoid errors and maintain compliance with regulations.

Nonconforming Event Management

The Nonconforming Event Management QSE provides guidance to detect and document nonconforming events. This may also be known as occurrence management. This QSE provides the framework for managing equipment or events that do not meet specified requirements, classifying events, and analyzing the data in order to correct problems. This allows the laboratory to identify and correct system problems. QC results which are unacceptable are documented, investigated and the root cause evaluated to prevent further problems. This improves troubleshooting efficiency and often prevention of further events. Unacceptable blood gas PT results must be investigated and the probable cause and resolution documented. The nonconforming QSE is also linked to the organization’s risk management.

Assessments

The Assessment QSE addresses the use of external and internal monitoring (Box 12-10). The purpose is to determine whether the defined process meets the requirements and evaluate how well the processes are functioning. Examples of external assessments include the proficiency testing program for the blood gas laboratory. Internal assessments are quality indicators such as monitoring turnaround time for pulmonary function results from performance to release the interpretation results.

Using comparative data between pulmonary laboratories or clinics within a health care organization is another method for completing an internal assessment. The percentage of tests meeting acceptable and repeatable criteria for spirometry for each technologist, laboratory, or clinic could be used as an internal assessment.

Inter-laboratory proficiency testing (PT) consists of comparing unknown control specimens from a single source in multiple laboratories. This allows an individual laboratory to compare its results with other laboratories using similar methods. Results from laboratories that used different methods (e.g., analyzers) may also be compared. Results of proficiency testing are usually reported as means and SD for each instrument participating in the program. Proficiency testing does not measure day-to-day precision, as does daily QC. However, it provides a measure of the absolute accuracy of the individual laboratory. An analyzer may have acceptable precision as determined by daily QC but be inaccurate when compared with analyzers from other laboratories. Proficiency testing often detects systematic errors that occur because of improper calibration, contaminated reagents, or procedural problems. Multiple levels of unknowns (for PT) are usually provided to check the range of values seen in clinical practice. Proficiency testing programs are available from professional organizations such as the College of American Pathologists, as well as from commercial vendors. Satisfactory performance on inter-laboratory proficiency testing has been mandated by the U.S. Department of Health and Human Services under the Clinical Laboratory Improvement Amendment of 1988 (see Evolve).

Continual Improvement

The Continual Improvement QSE identifies opportunities for improvement from multiple sources such as customer surveys, nonconforming events, evidence-based practice, ATS-ERS standards, PT results, internal assessments, external inspections or evaluations, and quality indicators. Improvement opportunities may be within a specific QSE or the path of workflow. A defined strategy for continual improvement should be used. A quality report is submitted at least annually to upper management (Box 12-11). One common effective strategy in pulmonary laboratories is the provision of a technologist’s feedback loop.

Technologist’s Feedback

A well-trained and highly motivated technologist is a key component for obtaining valid data, particularly in tests that require patient instruction and encouragement. As a component of the continual improvement QSE, a program based on established criteria for acceptability and repeatability can be used to provide feedback on test performance to individuals conducting the tests. This may also be included in performance appraisals and as a competence tool during training.

Routine review of tests performed by each technologist is recommended. If criteria for acceptability and repeatability have been recorded (as described in the path of workflow), these can be used to grade the performance of the technologist. This information forms the basis for reinforcing superior performance or correcting identified problems. Feedback should include the type and extent of unacceptable or nonrepeatable tests. Feedback should also include what corrective action can be taken to improve performance. Feedback needs to be ongoing to maintain a high level of proficiency. For research applications or in clinical trials, review of test data and performance, along with feedback to the personnel conducting the tests, may be necessary to provide the highest quality results.

Path of workflow

The path of workflow was briefly described in Chapter 1. This section will illustrate and describe key processes in the path of workflow for the pulmonary function laboratory from the time a patient is assessed for testing clinical correlation of the results. The pre-test processes are shown in Figure 12-9.

Pre-Test Process

Patient assessment is the entry point to a therapist-driven protocol (TDP) or ordering of pulmonary function tests. TDPs are developed with current scientific evidence in partnership and approval of the medical director. The assessment may include, but is not limited to, the clinical history, signs and symptoms, and other abnormal tests, which will also lead to the clinical indications outlined in the ATS-ERS statements. A standardized ordering format can assist with the test request process.

The test request key process includes generation of the order, pre-test instructions, and scheduling the tests. Test requests may be generated electronically or by paper. The request should meet all institutional, accreditation, and regulatory requirements. The requestor and location to send the final report are also required. Other important areas to consider are the patient’s physical and mental status, medications that may impact test results or that need to be held before testing, special preparation instructions, and the clinical indication for testing. Ordering instructions and guidance are needed for the authorized requestor to facilitate efficiency, accurate orders, and prevention of medical errors. These instructions answer questions related to tests available, patient consent requirements, how to complete the requisition, medications that need to be held for test validity, how to schedule the test with contact information, and any other special considerations. Specific preparation instructions for tests such as bronchial provocation testing may be provided both in a verbal and written format to the patient.

Patient preparation is a critical key process in obtaining reliable data. The demographic information provides the data to calculate the predicated reference values. An accurate height is required to obtain reliable reference values. A stadiometer provides the most accurate height measurements. All height measurements are made with shoes removed. In those cases where a standing height cannot be obtained, calculations are available in most manufacturer’s software programs to use the arm span measured from fingertip to fingertip. Deviations from standard practice should be included in the technologist’s comments. Calibrated scales provide an accurate weight.

The clinical indication for testing should be verified again before testing each individual because the clinical presentation can change between ordering and testing. The absence of contraindications needs to be confirmed before testing. Patient adherence to any pre-test instructions such as holding medications before a test should be confirmed and documented. Patient preparation is the point where a final assessment of age-specific considerations that may impact test results is completed.

Equipment preparation includes the final preparation of the equipment just before testing that is required to obtain reliable results. Quality control is completed at the frequencies predetermined for the laboratory and typically would be completed before this step unless a probable out-of-control situation is defined after the patient has started the testing process. Equipment that is found out of control should not be used for patient testing. Specific instructions on how to handle out-of-control situations are included in the procedure manual. Equipment calibration is performed, and all necessary supplies are ready for testing. Reference values may be selected if there are multiple options based on specific patient populations. The reference sets selected are based on current evidence and the patient population to be tested (Chapter 13). For example, the CF foundation currently requires PF laboratories to use defined reference authors.

Testing

Test Method Selection

The testing method is dependent on the patient population, equipment available, tests ordered, and, ultimately, the patient’s needs (Figure 12-10). In order to decrease unwanted variations that impact test reliability, the laboratory can validate its testing processes, equipment, and software. An initial validation is completed when new equipment is installed. Re-verification is completed when processes are changed, software is changed, or after repairs and preventive maintenance.

Patient training is essential to obtaining reliable results. The patient’s anxiety level and understanding of why they are being tested are important considerations to obtaining maximum cooperation. Communication strategies may need to be altered based on the age of the patient. Chapter 8 addresses training and coaching pediatric patients. Regardless of the test performed, key steps to performing the test are explained and demonstrated to the patient. If needed, an interpreter should be available. Demonstration of the procedure should always be done to achieve maximum understanding of the test procedure. The technologist training the patient needs to display the same effort required by the patient to achieve reliable tests.

Test Performance

A primary means of ensuring data quality is to rigidly control the procedures by which data are obtained. For many pulmonary function tests, how the data are obtained depends on the technologist’s ability to train the patient adequately, conduct the procedure, and to elicit subject cooperation in the test maneuvers. Technologist and subject performance, as well as proper equipment function, must be evaluated for each test completed. This may be accomplished by using appropriate criteria to judge the acceptability of results. Recommendations for testing procedures published by the ATS-ERS are widely recognized as standards for test performance.

Each pulmonary function laboratory should have a written technical procedure manual (Documents and Records QSE) that includes the following:

• Methods used for specific tests

• Specific guidelines as to how tests are to be performed

• Limitations of each procedure (if any)

Providing a quality output in the pulmonary function laboratory requires not only appropriate calibration, verification, and QC, but also careful attention to how the data are obtained (i.e., testing techniques). Testing technique may be compared with “sampling” technique in other laboratory sciences. In pulmonary function testing, sampling refers to procedures used to obtain patient data. How the data are obtained becomes extremely important because many of the tests performed are effort-dependent. Eliciting maximal effort and cooperation from the patient is often just as critical as correct performance of the equipment. Applying objective criteria to determine the validity of data is one means of providing high-quality results.

Using Criteria for Acceptability/Repeatability

Criteria for assessing the validity of various tests have been described in Chapters 2 through 9. Standards for pulmonary function testing have been published by the ATS in conjunction with the ERS. Criteria for acceptability have three primary uses:

1. To provide a basis for decision making during testing. Standards or guidelines can be used to decide whether equipment is functioning properly, whether the patient is giving maximal effort, or whether testing should be continued or repeated. Standardized criteria also help characterize the types of problems known to occur during specific tests (e.g., poor effort during spirometry).

2. To evaluate validity of pulmonary function data from an individual patient. Criteria may be applied by the technologist performing the test, by computer software, or by the clinician responsible for interpretation. This may consist of assigning a letter grade or a code to individual efforts or tests.

3. To score or evaluate the performance of the technologist. Many pulmonary function tests, especially spirometry, depend on the interaction between the technologist and the patient. Criteria for acceptability can be used to gauge the performance of individual technologists and to provide objective feedback.

Implementation of a quality system in the pulmonary function laboratory should use acceptability and repeatability criteria during and after testing (results review process), both to evaluate individual tests and to provide feedback for technologists. For each of these, certain procedures will be similar.

Examine Printed Tracings or Displayed Graphics Whenever Available

Compare the observed tracing with the characteristics of an acceptable curve or pattern. Computer graphic displays make this particularly easy and can usually display tracings in real time. Graphics may be superimposed or displayed side-by-side to assess patient effort and cooperation. Similarly, expected values can be displayed graphically along with each individual effort. The user should be able to modify the graphic display (e.g., change graphing scale) to allow for extremes such as very low flows or volumes. During testing, graphs of multiple efforts should be available. Storage of graphic data (all acceptable maneuvers) may be useful for assessing data quality after testing has been completed. Some portable (office) spirometers may not display graphics of volume-time or flow-volume curves during testing but are able to print graphics. In these cases, the printed graphs may be used to assess test quality. Real-time displays for both flow-volume and volume-time displays during testing are particularly helpful in assessing test variation during the testing process rather than post-testing.

Look at Numeric Data

Are the largest values from multiple efforts within the accepted range of repeatability? The decision to perform additional maneuvers is usually based on repeatability. Most manufacturers provide software that applies current standard guidelines for repeatability. Data from multiple efforts should be maintained during testing to allow selection of appropriate results for the final report. Storage of all efforts may be necessary for subsequent review or editing. Data that are not repeatable may be valid (i.e., usable) but need to be identified as such.

Evaluate Key Indicators

Most pulmonary function tests have one or two features that determine whether the test was performed acceptably. For spirometry, the start-of-test and duration of effort are key indicators. For gas dilution lung volumes, absence of leaks and test duration are key indicators. For Dlcosb, inspired volumes and breath-hold times are important. Key indicators vary with the method used for specific tests. In each instance, the important indicator should be assessed in relation to an accepted standard. During testing, these indicators help determine whether additional patient instruction or tests are needed. Key indicators are also useful in assessing what factors might influence the interpretation of the test (e.g., the patient was unable to blow out for 6 seconds).

After determining the acceptability and repeatability of maneuvers for each test, the technologist makes a decision when an adequate number of maneuvers have been completed to ensure reliable results or the patient is unable to continue testing.

Reviewing the results again post-testing and selecting the final data for the final report follow the same standards used during the testing process and the pre-established laboratory standards. A checklist can be helpful at each workstation to assist in the process.

The results of various tests should be consistent with the clinical history and presentation of the patient. Spirometry, lung volumes, Dlco, and blood gas values should all suggest a similar interpretation for a specific diagnosis. Discrepancies among tests may indicate a technical problem rather than a clinical condition. Comparing results from similar measurements is helpful with an understanding of the expected relationships.

PF Tip 12-5

1. SVC (lung volumes) > FVC (spirometry)

2. TLC (lung volumes) > VA (Dlco)

3. FRCPL > FRCn₂ or FRC_He in obstructive patterns

4. IVC (Dlco) < SVC (lung volumes)

Patient assessment for further testing occurs when a therapist-driven protocol has been implemented or the correlation with the patient’s clinical condition suggests further testing is indicated.

Scoring or grading the quality of a patient’s test is an important component of quality assurance for pulmonary function testing. The technologist administering the test can accomplish this by adding notes. Some automated spirometers use software that grades test performance or that includes codes denoting problems with the test. Evaluation of spirometry, lung volumes, Dlco, and any other tests performed should be included.

The technologist’s comments or notes can usually be added to the test results. The commentary should be based on standardized criteria. If a particular test meets all criteria, that fact should be stated. Failure to meet any of the laboratory’s criteria should be documented as well. The reason the patient was unable to perform the test acceptably should be explained whenever possible. Failure to meet criteria for acceptability does not necessarily invalidate a test. For some patients, their best performance may fail one or more of the criteria. Table 12-7 lists examples of statements that may be used to document test quality.

Table 12-7

Technologist’s Comments (Examples)

Test	Comments
Spirometry	Meets all ATS-ERS recommendations.
	Poor start of test or patient effort.
	Expiration did not last 6 seconds, or no obvious plateau for age 10 years old or greater (3 seconds for <10).
	Back-extrapolated volume was >5% of FVC.
	Patient was unable to continue to exhale because of _____.
	Two best acceptable FVC maneuvers were not within 150 mL.
	Two best acceptable FEV1 maneuvers were not within 150 mL.
	MVV does not correlate with FEV1.
Lung volumes (gas dilution)	Lung volumes by (method)—meet ATS-ERS recommendations.
	Lung volumes reported were the average of (n) FRC determinations.
	Slow VC was (greater/less) than FVC (_____%).
	Lung volumes were unacceptable because of a leak.
	Equilibration not reached within 7 minutes—He dilution.
	Alveolar N2 >1.5% after 7 minutes—N2 washout.
Plethysmography	All plethysmographic measurements met ATS-ERS recommendations.
	FRCpleth measurements were variable.
	Raw measurements were variable.
	Patient was unable to pant at the correct frequency.
Dlcosb	Meets all ATS-ERS recommendations.
	Dlco reported is average of (n) maneuvers.
	Predicted Dlco corrected for an Hb of _____.
	Inspired volume <85% of best VC (_____%).
	Breath-hold time not within 8–12 seconds (_____ seconds).
	Dlco values not within 3 mL CO/min/mm Hg or 10%.
	Dlco not corrected for Hb or COHb.

*Values in parentheses may be filled in with appropriate values from the patient’s data.

The technologist’s comments can be included in the final report. Many automated systems provide for “free text” comments to be included with tabular data. Some software supports “canned text” functions that allow predetermined statements (see Table 12-7) to be entered with a single keystroke. The technologist’s name or initials should be included. The technologist’s comments should be clearly identified to avoid confusion with the physician’s interpretation.

Some computerized spirometer systems automatically score FVC maneuvers. The score may be indicated by a letter or numeric code that is attached to each maneuver. For example, an FVC maneuver that meets all criteria (e.g., start-of-test) might be scored with an “A.” Other systems allow the technologist to select a user-defined code to attach to individual maneuvers. Both techniques can be used to provide feedback that enhances quality assurance.

Post-Testing

Results reporting occurs after the quality review is completed and the final data have been selected for the report (Figure 12-11). A system for a secondary review of results as a routine or random evaluation leads to continual improvement. It also provides a forum to discuss methods to improve testing processes. The report format needs to provide numeric and graphic results for the interpretation process. If preliminary and final reports are used, both must be accurately labeled in the patient chart and the process defined when a preliminary is replaced by a final report. Turnaround times for report generation can be a quality indicator for the laboratory. If a report is found to be an erroneous post release, a system for the correction and replacement of the original report is required. This includes a method to verify that the health care provider is notified. The laboratory should establish acceptable limits for each test. Notification procedures for “alert” results and a documentation process should be available in the procedure manual and immediately available for all technologists.

Interpretation of the results includes providing a process for standardization based on current standards and evidence. Some systems provide a template for interpretation and potential interpretation statements. To establish a standardized process for interpretation, objective criteria for separating normal and abnormal results are required. Additional information in the medical record or history such as smoking, occupational exposure, recent illnesses, and medications is useful when reviewing data. Frequently, the data also are compared to previous data, which should be easily identified or available in the medical record. The selection of reference values is a critical element in reliable interpretation systems.

The turnaround time for interpretation is an important quality indicator. In an ATS survey to evaluate average turnaround time for interpretation of the test results, the authors found the following results: <1 day (15%), 1-2 days (30%), 3-4 days (27%), 5-6 days (15%), >7 days (3%). This highlights the variability in the process between institutions. A delayed interpretation has the potential to delay the clinical consultation and important therapeutic interventions in patient care. Applying the test results to the patient is the last step in the path of workflow.

Understanding and establishing the path of workflow processes leads to the needed procedures to develop an effective system for the delivery of patient care in the pulmonary laboratory.

Summary

Case Studies