Medical Devices
Draft Guidance for Industry and FDA Staff - Total Product Life Cycle for Portable Invasive Blood Glucose Monitoring Systems
Draft Guidance - Not for Implementation
This guidance document is being distributed for comment purposes only.
Document issued on: October 24, 2006
Comments and suggestions regarding this draft document should be submitted within 90 days of publication in the Federal Register of the notice announcing the availability of the draft guidance. Submit written comments to the Division of Dockets Management (HFA-305), Food and Drug Administration, 5630 Fishers Lane, rm. 1061, Rockville, MD 20852. Alternatively, electronic comments may be submitted to Regulations.gov. Comments should be identified with the docket number 2006D-0353.
For questions regarding this document contact Carol Benson at 301-796-5459 or carol.benson@fda.hhs.gov
When final this draft document will replace "Review Criteria for Assessment of Portable Blood Glucose In Vitro Diagnostic Devices Using Glucose Oxidase, Dehydrogenase, or Hexokinase Methodology," issued February 14, 1996.
 |
U.S. Department of Health and Human Services Office of In Vitro Diagnostic Device Evaluation and Safety |
Contains Nonbinding Recommendations
Draft - Not for Implementation
Preface
Additional Copies
Additional copies are available from the Internet. You may also send an e-mail request to dsmica@fda.hhs.gov to receive an electronic copy of the guidance, or send a fax request to 301-847-8149 to receive a hard copy. Please use the document number 1603 to identify the guidance you are requesting.
Table of Contents
- INTRODUCTION AND SCOPE
- TERMINOLOGY
- CLINICAL USE OF BGMS DEVICES
- QUALITY SYSTEM REGULATION (QSR) AND RISK MANAGEMENT
- Quality System Regulation
- Risk Management of BGMS: Points to Consider
- Risk Assessment
- Risk Controls
- Human Factors Considerations in Design and Testing of BGMS Devices
- Validation of Risk Mitigation Measures Incorporated in the BGMS Device
- Description of Device Risk Management and Validation Results in the 510(k)
- PERFORMANCE EVALUATION AND CRITERIA FOR BGMS DEVICES
- Precision evaluation
- Method Comparison
- Linearity
- User Performance
- Interferences
- Safety and Reliability Testing
- System Traceability and Validation of Calibration and Control Materials.
- DESCRIPTION OF THE DEVICE
- SOFTWARE
- LABELING
APPENDIX 1. ALTERNATIVE SAMPLE SITE TESTING (AST)
APPENDIX 2. POTENTIAL SOURCES OF ERROR TO CONSIDER
APPENDIX 3. THIRD-PARTY STRIP MANUFACTURER CONSIDERATIONS
REFERENCES
Draft Guidance for Industry and FDA Staff
Total Product Life Cycle for Portable Invasive Blood Glucose Monitoring Systems
| This draft guidance, when finalized, will represent the Food and Drug Administration's (FDA's) current thinking on this topic. It does not create or confer any rights for or on any person and does not operate to bind FDA or the public. You can use an alternative approach if the approach satisfies the requirements of the applicable statutes and regulations. If you want to discuss an alternative approach, contact the FDA staff responsible for implementing this guidance. If you cannot identify the appropriate FDA staff, call the appropriate number listed on the title page of this guidance. |
Introduction and Scope
This guidance document provides FDA’s recommendations concerning portable invasive blood glucose monitoring system (abbreviated BGMS)1 devices. In addition to recommendations for preparation of premarket notifications (510(k)’s), this document discusses features of device design and risk management, including those relating to human factors that manufacturers should address during the product life cycle. The document is intended to complement ISO (International Standards Organization) standards on risk management for medical devices [1] and blood glucose monitoring systems, [2] as well as other FDA guidances cited below. The scope of this document includes BGMS devices for the quantitative measurement of glucose in blood by lay users at home, or by professionals in hospitals and other point of care settings, to manage carbohydrate metabolism disorders, including diabetes mellitus. Our intention is that when this guidance document is finalized, it will enable FDA to make more efficient and better-informed decisions based on consistent data, and contribute to marketing of more reliable, reproducible, and simple to use commercial devices.
BGMS devices are regulated under 21 CFR 862.1345 (Glucose Test System) as Class II devices. The following product codes for BGMS devices are within the scope of this guidance document:
NBW (blood glucose test system, over the counter)
CGA (glucose oxidase method)
CFR (hexokinase method)
LFR (glucose dehydrogenase method)
This document does not address devices used to screen and diagnose diabetes (such as clinical chemistry analyzers or semi-quantitative strips), implanted or continuous glucose sensors, non-invasive glucose devices, or devices for measurement of blood glucose in neonates2. The device types addressed in this document typically use capillary whole blood from fingersticks. Many aspects of this guidance might also apply to devices that test arterial or venous whole blood, or alternative anatomical sites. However, this guidance is not intended to fully address those other matrices or test sites. Appendix 1 briefly addresses alternative site testing (AST). We recommend that you contact the Chemistry and Toxicology Division in OIVD (Office of In Vitro Diagnostic Device Evaluation and Safety) if you have questions regarding these additional intended uses.
FDA's guidance documents, including this guidance, do not establish legally enforceable responsibilities. Instead, a guidance document describes the Agency's current thinking on a topic and should be viewed only as recommendations, unless specific regulatory or statutory requirements are cited. The use of the word should in an Agency guidance means that something is suggested or recommended, but not required.
The Least Burdensome Approach
This draft guidance document reflects our careful review of what we believe are the relevant issues for BGMS devices and what we believe would be the least burdensome way of addressing these issues. If you have comments on whether there is a less burdensome approach, however, please submit your comments as indicated on the cover of this document.
Abbreviated 510(k) Submissions
As explained in “The New 510(k) Paradigm - Alternate Approaches to Demonstrating Substantial Equivalence in Premarket Notifications”, a manufacturer may submit either a Traditional 510(k) or an Abbreviated 510(k). An Abbreviated 510(k) provides a means to streamline the review of data in a 510(k) through a reliance on FDA-recognized consensus standards, special controls, or FDA guidance documents. Guidance on the content and format for abbreviated and traditional 510(k)s is available at "Format for Traditional and Abbreviated 510(k)s". Also, see Section 514(c)(1)(B) of the Act and the FDA guidance, “Use of Standards in Substantial Equivalence Determinations” on use of standards in an abbreviated 510(k).
II. Terminology
Definitions below are for terms as used in this guidance document. They are based on the ISO harmonized database. For further information and context, see the harmonized terminology database on the CLSI website.
Accuracy (systematic error): Closeness of agreement of a test result with an accepted reference value.
Alternative sample site testing: Obtaining a capillary whole blood specimen from an anatomical site other than the fingertip.
Bias (trueness): Closeness of agreement between the average value of a large number of test results and reference value.
Blood Glucose Monitoring System (BGMS): Measuring system consisting of portable instrument and reagents used for the in vitro monitoring of glucose concentrations in blood. (In this guidance we are referring only to invasive BGMS devices.)
Control material: Substance, material, or article intended by its manufacturer to verify the performance characteristics of an in vitro diagnostic device.
Harm: Physical injury or damage to the health of people.
Hazard: Potential source of harm.
Precision : Closeness of agreement between independent measurement results obtained under stipulated conditions.
Residual risk: Risk remaining after protective measures have been taken.
Risk: Combination of the probability of occurrence of harm and the severity of that harm.
Risk analysis: Use of available information to identify hazards (potential sources of harm) and estimate their risk.
Risk assessment : Combination of risk analysis and risk evaluation.
Risk control: Process through which decisions are reached and protective measures are implemented for reducing risks to, or maintaining risks within, specified levels.
Risk evaluation: Judgment, based on risk analysis, of whether an acceptable (low) level of risk has been achieved.
Risk management: Systematic application of management policies, procedures, and practices to analyze, evaluate, and control risk.
III. Clinical Use of BGMS Devices
BGMS devices were introduced in the late 1970’s and are considered one of the most important medical advances in diabetes care. The American Diabetes Association (ADA) has issued several policy statements on the use of BGMS devices. An ADA Consensus Panel [3], comprised of physicians, diabetes educators, and laboratorians, recommended BGMS devices for the following patient groups:
- Women with pregnancies complicated by diabetes.
- Individuals with a propensity for severe ketosis or hypoglycemia.
- Individuals prone to hypoglycemia who may not experience the usual warning symptoms.
- Individuals on intensive insulin treatment programs, especially those using portable insulin infusion devices (insulin pumps) or multiple daily injections.
- Individuals with abnormal renal glucose thresholds.
The panel recommended that BGMS devices not be used to diagnose or screen for diabetes mellitus. Consistent with the views of the Consensus Panel, the intended use for FDA cleared or approved BGMS devices is to assist in the ongoing evaluation and management of individuals with diabetes. In the home, diabetic patients use BGMS devices to self-monitor fluctuations in blood glucose levels using capillary whole blood specimens (typically obtained by fingerstick, and more recently by alternative sampling sites in some devices). Healthcare professionals use BGMS devices, primarily with capillary whole blood, to quickly obtain blood glucose concentrations at the patient’s bedside.
IV. Quality System Regulation (QSR) and Risk Management
A. Quality System Regulation
This section is addressed to manufacturers of BGMS devices. The Quality System Regulation (QSR), 21 CFR Part 820, describes requirements intended to help ensure that medical devices meet standards of safety and effectiveness throughout the product life cycle. It requires that you establish specifications and controls for device design and manufacture, under a quality system; ensure that finished devices meet specifications and are correctly installed, checked, and serviced; ensure that quality data are analyzed to identify and correct quality problems; and ensure that complaints are processed. Your quality system is critical to help prevent and mitigate risks of hazards caused by either use errors or device errors. The FDA monitors device problem data and inspects the operations and records of manufacturers to determine compliance with the GMP requirements in the QSR. For guidance on complying with QSR, see the document on “Quality System Regulation”.
Requirements concerning device design controls are addressed in 21 CFR 820.30. Design verification and validation, including risk analysis (a component of risk management) and software validation, are components of design controls. We recommend that you refer to the FDA document “Design Control for Manufacturers” for a more comprehensive explanation of design controls.
B. Risk Management of BGMS: Points to Consider
The risk management process is concerned with identification, measurement, control, and minimization of risks to an acceptable level, on an ongoing basis, throughout the product life cycle. It involves a systematic application of management policies toward that end.
FDA receives numerous MDR (Medical Device Reporting) reports and other reports of injuries, including cases that required hospitalization or resulted in death, that were apparently related to the use of BGMS devices. When reporting adverse events with BGMS devices to FDA, manufacturers often report that the problem occurs because of a use error. Thus, as part of your quality system, it is important that the risk management plan you have in place addresses the known possible causes of use errors (human factors-related errors), as well as device system errors and failures. This section of the guidance addresses some aspects of risk management as they relate to BGMS devices. We recommend that you follow the standard ISO 14971: 2000, “Medical devices- Application of Risk Management to Medical Devices” when instituting a risk management plan for BGMS devices.
Two aspects of risk management include risk assessment and risk control [1]. These are discussed in sections IV.B.1 through IV.B.III, below. For devices that contain software, manufacturers should include a risk analysis in the 510(k). (See Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices”. ) Information concerning risk analysis that you should describe in your 510(k) is discussed in Section VII (Software) of this document.
Risk assessment includes identification of all possible causes (initiating events) of hazards, estimation of their risk, and determination of whether the risk is acceptable. Risks are determined by considering the severity of the consequences and the likelihood of occurrence of the hazard. Hazards can include erroneous results, no results, or results prone to misinterpretation. They can also include direct physical hazards, such as those caused by electrical malfunctions. As noted above, hazards could be caused by system failures, use errors, or a combination of both. Risk assessment is based on detailed knowledge of the device design itself, as well as knowledge of the intended use, and the skills and limitations of the intended use population. System failures and use errors can be identified using a failure mode and effects analysis (FMEA), or a similar tool during the design of the device. Risk assessment is typically an iterative process and is conducted at various stages during a product life cycle.
As part of your risk assessment you should consider preanalytic procedures (such as blood collection and sample application), analytic, and postanalytic procedures that may potentially cause hazards. The table in Appendix 2 lists examples of possible sources of error. These are meant as general examples, and you should consider and evaluate any other risks based on knowledge of your specific BGMS.
Risk control refers to implementation of protective measures (also called mitigation measures) to reduce risks or maintain them within specified levels [1]. Once you identify a potential cause of a hazard, you should incorporate into the device features to mitigate the risk of the hazard. An effective risk management program reduces risks until the residual risk does not affect the safety or effectiveness of a device beyond a tolerable allowance that will meet the needs of the user. Risk mitigation measures can include features to enhance the robustness3 of reagents and instruments. They can include internal device checks, external control material, and implementation of other quality assurance measures, as discussed in this section. Risk control can also include measures to enhance ease of use for the intended use population based on consideration of human factors. Human factors are discussed below in Section IV.B.3. We do not generally consider labeling measures alone to offer sufficient risk control for BGMS devices.
Hazards (e.g., erroneous results) are more likely to cause harm (e.g., incorrect treatment) when they, or the events that might cause them (e.g., incorrect strip placement, expired reagents), are not immediately identified by the user. Thus, mechanisms that alert the user to conditions that may cause inaccurate results are important types of mitigation measures. In addition, it is important that users be able to check the correct functioning of the system (e.g., reagents, meter) and the correct performance of procedural steps. External control materials and internal checks are examples of protective measures that can address these issues.
External control material
One mechanism used to alert the user to inappropriate testing conditions is with external control material. We recommend that you incorporate convenient and easy to use external control materials into your device system. These materials are used to assess the operator’s technique and the performance of the entire test system. To increase the likelihood that users will run control materials, we recommend that you include these materials in the packaging with the reagents. When you do not include these external control materials with the reagents, you should recommend the use of specific control materials suitable for your device. You should provide or recommend at least two levels of control materials. You should also indicate to users an appropriate frequency for testing these materials. Examples include each time a user opens a new vial of reagent or if the user suspects that the meter is not working properly. The control material you provide (or recommend) should mimic patient samples as closely as possible. In the device labeling you should provide instructions concerning what the user should do if results they obtain with control materials are not within your allowable range.
Internal checks
We recommend that you incorporate internal checks and warnings that will alert the user when error conditions exist. We encourage you to incorporate lockout functions into the device for this purpose when feasible. Checks with warning or lockout functions can be used to indicate problems, such as the following:
- Improperly performed quality control, calibration, or other operational steps.
- Environmental conditions outside of the operating ranges of the meter (e.g., temperature, altitude).
- Battery integrity failure.
- Detection system not functioning properly.
- Reagent integrity problems.
- Improper sample volume or sample flow.
- Use of reagents that are not compatible with the meter.
- Inadequately calibrated device.
Other quality assurance measures
Other quality assurance measures that you should include in your system are:
- Easy-to-use maintenance instructions.
- A toll-free U.S. telephone number that users can call and consult with a trained professional about problems with quality assurance.
We recommend that in the labeling you advise users that they should periodically compare the test system to another test system known to be well maintained and monitored by a healthcare provider.
3. Human Factors Considerations in Design and Testing of BGMS Devices
This section supplements the guidance “Medical Device Use-Safety: Incorporating Human Factors Engineering into Risk Management, Guidance for Industry and FDA Premarket and Design Control Reviewers” on the FDA web site. We recommend that you consult that document and references within it concerning human factors. Because of the potential for use errors in point of care and home use devices, it is especially important for manufacturers to adequately evaluate human factors when designing BGMS devices; this is an important component of risk management. User characteristics, device-user interface, and use environment, described below, are all central to human factors considerations.
User characteristics of BGMS
User characteristicsthat affect performance include a variety of behavioral, psychological, perceptual, and physical traits and abilities, as well as knowledge about the device. Examples include manual coordination, memory, and sensory perception. Diabetes can deteriorate eyesight, and age may contribute to hearing loss in older Type 2 diabetics.
Device-user interface characteristics for BGMS
Device-user interface characteristics are features of the device that communicate to the user, or features that the user manipulates to control the device. These include visual and auditory error messages, vibrations, flashing lights, alarms, or beeps, menus, push buttons, labeling or markings on the device, and user instructions. You should consider user characteristics like vision or hearing loss when you design the device and determine whether the device interface is adequate. Users with a limited range of abilities should be able to understand any error messages, alarms, or other information and be able to easily manipulate the push buttons. You should consider requirements placed on the user when performing all tasks involved in the operation and maintenance of the meter. Tasks include actions taken by the user and observation tasks such as reading and interpreting meter output. In general, you should take into consideration device features such as the following:
- Factors relating to ease of operation (e.g., easy to insert strips correctly, easy to read instructions)
- Factors relating to ease of maintenance (e.g., easy to change battery, easy to recognize when battery needs to be changed)
- Readability of the result (e.g. clear displays, appropriate font size)
- Unambiguous messages to the user (e.g., clear error codes)
Human factors evaluations should address whether any aspects of user interaction with the device are particularly prone to causing users to make errors.
Use environment of a BGMS
Environmental conditions vary between the home, office, restaurant, automobile, outdoors, and sporting events. Lighting, glare from meter surfaces, distractions, time pressure, noise, etc., may affect the user’s ability to effectively complete a measurement. To help minimize the impact of environmental factors, you should make devices easy to use with plain displays. Environmental factors that affect the BGMS include high or low temperature, physical abuse, vibrations, relative humidity, altitude, and electromagnetic interference. One example of a mitigation measure is that the device may display an error message or automatically shut off to alert users that it detects a temperature above that which is appropriate for use of the device. You should consider device features that will protect the device from typical “wear and tear.”
Some examples of features that have been incorporated in recent devices to enhance ease of use include smaller blood sample volume requirements; warnings (failure alerts) that notify users when they have not followed certain procedures correctly; and elimination of steps, such as not requiring the user to remove excess blood (by wiping, blotting, or washing).
In the context of risk management, potential use errors are a potential cause of a hazard. As with other hazards, we recommend that you prioritize these errors in terms of the risk of harm they present to users. High priorities should be assigned to those use-related hazards that involve adverse impact on treatment, such as obtaining inaccurate results, misreading or misinterpreting test results, or incorrectly storing test results (for meters that have this capability). You should correct problems identified in human factors evaluations through changes in the device design, or (if sufficient) through changes in the label. We recommend that you perform human factors testing early in product development. Some use errors may not be detected during user validation studies (such as described in Section V-D.) especially if they occur at low frequencies, or only under certain conditions (e.g., high temperature, high altitude). Methods to evaluate such errors include: testing under conditions that stress the system to mimic foreseeable use errors; obtaining subjective feedback from representative intended users regarding any aspects of the procedure they found difficult or confusing; and (non-intervening) observation of device use by a trained technician.
4. Validation of Risk Mitigation Measures Incorporated in the BGMS Device
As part of your risk analysis you should evaluate the effectiveness of the risk mitigation measures incorporated into your device. You should validate that your BGMS does not generate results when the user performs a critical step in an incorrect manner or when an instrument error that compromises the test result occurs. This typically involves testing replicates of a sample while incrementally varying common variables, one at a time to challenge, or stress the system. Examples of variables may include sample volume, strip placement, and battery strength. The table below illustrates two examples of scenarios that generate incorrect results, associated mitigation measures, and validation testing that stresses the system:
| Examples of potential use errors | Example of a Mitigation measure | Example Validation Study |
|
Results of testing indicate accurate results are obtained across a range of 50-80 ° F. Strips stored outside a temperature range of 50-80 ° F yield inaccurate results. |
A temperature monitor added to the outside of the test strip vial indicates when strips have been exposed to <50 ° or >80 ° F. | Studies evaluate whether test strips change color at temperatures ranging from 0-50 ° and 80 ° -100 ° F, and whether users understand the meaning of the color change. |
|
Improper strip placement (outside 2 mm of the end of the guide) leads to inaccurate results.
|
Example 1: An alert is incorporated in the device so that an error message is triggered if the strip is inserted incorrectly. |
Studies evaluate the device response when the strip is inserted at 2 mm, 3 mm etc, or when the strip is otherwise inserted incorrectly (e.g., backwards). |
| Example 2: Strips are designed so they are sufficiently centered in the guide and cannot be moved in any direction. |
Studies evaluate whether the strips could still be inserted improperly (e.g., outside 2mm in this case). |
5. Description of Device Risk Management and Validation Results in the 510(k)
As noted in the guidance document “ Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices”, you should include a risk analysis in a 510(k) for devices that include software. See Section V. below concerning information to include concerning your risk analysis.
For any device features you have incorporated as mitigation measures, especially those that you describe in your labeling, you should summarize testing methods and results to demonstrate performance of these features under conditions that stress the system.
Devices
Sections A-G below indicate the types of information we recommend you include in the 510(k) concerning the performance characteristics of your BGMS. We recommend that you follow the study designs, statistical evaluations and presentations outlined in ISO 15197 [2] for precision, accuracy, and user performance evaluations. Relevant CLSI documents and ISO document sections are listed in the table below. If you followed all recommendations specified in the ISO document without deviation, you may indicate this and provide a summary of your procedures, results and conclusions, as recommended below. If you modified any procedures recommended in the standard or followed other procedures, you should describe further specifics of your testing procedure.
| Performance Characteristic | ISO 15197 Section |
| Precision | 7.2 |
| Accuracy - Evaluation | 7.3 |
| Criteria for Acceptable accuracy | 7.4 |
| User Performance | 8 |
| Linearity | CLSI/NCCLS EP-6A |
| Interfering Agents | CLSI/NCCLS EP-7A |
| Safety and reliability (including environmental factors) |
6 |
A. Precision evaluation
We recommend that you follow ISO 15197, Section 7.2 for your precision evaluation. In your 510(k), we recommend that you include the following concerning your evaluation:
- Glucose concentrations tested.
- Sample types tested (matrix; origin, e.g., control material, patient samples; preparation; anticoagulants).
- Parameters varied (including lots, operators, sites, calibrations).
- Number of samples tested.
- Number of replicate measurements under each condition (e.g., each day, user, reagent system).
- Number of meters and reagent system units used.
- Number of operators.
- Time span of the evaluation.
- Statistical methods used.
- Results (mean, SD, and CV for each condition evaluated).
- Identification and explanation of outliers you excluded (if any).
B. Method Comparison
You should compare results obtained with your device, in a setting representative of your intended use, to reference glucose concentration values obtained by a legally marketed device that has been well-validated for precision and trueness (e.g., Yellow Springs Instrument). We recommend that you follow ISO 15197, Section 7.3, for this study. We recommend that you describe the following in your 510(k):
- Study setting (e.g., description of the type of location and operators, and how the study conditions simulate intended use conditions).
- Type of sample collected (e.g., anatomical collection site, matrix).
- Samples that were altered (if any) in order to obtain sufficient numbers at extreme concentrations.
- Description (including traceability if appropriate) of the comparator method.
- Number of patients.
- Number of operators and sites.
- Patient selection and exclusion criteria (these should be based on the intended use population you describe in the labeling).
- Patient demographics (age, education, disease state).
- The range of glucose concentrations tested (See table 3 in ISO 15197).
- Number of reagent systems, lots, meters used in testing.
- Any relevant environmental conditions (temperature, altitude, other).
- Statistical methods used, including identification of any outliers.
In the description of your results, we recommend that you include:
- A plot of results of the comparator device versus your device with all data points shown, and each point representing a single measurement.
- Regression equations with confidence intervals.
- A plot of the difference in results (between the reference and candidate device) versus glucose concentration.
C. Linearity
You should evaluate your device for linearity across the reportable range. For example, see CLSI EP6-A [5].
D. User Performance
We recommend that you follow guidelines in ISO 15197, Section 8, for your user performance study. You should design this study to evaluate whether intended users can operate and obtain correct results with the device, using only the instructions that you plan to routinely provide with the device when you market it. Thus, you should provide users with only these routine instructions in English. Studies using labeling in other languages are not sufficient to support comprehensibility of the English labeling. You should try to simulate intended use (i.e., home use) conditions as much as possible. If the device is intended for lay users, you should exclude individuals that have laboratory testing experience or training from participation. You should compare results obtained by lay users performing self-tests to results obtained by professionals using a validated method (or using your new method, if it has already been validated by a method comparison study, such as the one described above, in Section B).
We recommend that you include the following in your 510(k):
- Specific procedures performed by lay users and professionals.
- Description of users, including the number of users, their age range, their education level, their native language, and their work experience.
- Instructions provided to users in the study.
- Summary of any user feedback (e.g., discussions, questionnaires) evaluated as part of the user study.
- Descriptions of the types of sites, and their locations.
You should include calibration and quality control procedures as part of user performance studies since this is an important aspect of device use with which lay users may have less experience.
We encourage you to obtain user feedback through discussions, trained observers, questionnaires, or other methods. However, you should take measures to ensure that the lay user performance results are not themselves influenced by feedback to the lay users from these activities. Thus, you should conduct these types of activities after the lay user volunteers do the testing and provide their results.
The large majority of BGMS devices cleared by FDA have been intended for use by lay users. If your device is only for use by health care professionals, you should conduct studies in which multiple operators (health care professionals) at 3 or more intended use sites test your BGMS device and compare results to those from the comparator device.
When representing results in your 510(k), you should present regression equations with confidence intervals and plots showing all data points, as described above in the method comparison section. You should include comparison of user results, professional results, and reference results, as appropriate. Discrepant user test results should trigger human factors investigations.
E. Interferences
You should evaluate the effect of potentially interfering conditions on device performance including haemolysis, icterus, lipemia, and varying haematocrit levels, as well as the effect of common over the counter (OTC) substances and frequently administered drugs for diabetes. If your system uses venous specimens, you should test anticoagulants as well. We recommend that you follow CLSI EP-7A [6] in the design of these evaluations. You should evaluate interferents at the highest levels at which they might occur in patient samples. You should test interferents in samples containing glucose concentrations that represent clinical decision points.
If you find analytical interference with more than one substance, you should evaluate whether these substances, when combined, have an additive effect on the accuracy of the BGMS.
In your 510(k) you should describe the following concerning your interference evaluations:
- Sample types evaluated (matrix, preparation, or origin).
- Concentrations of interfering substance and of glucose in the test sample.
- Calculations/Statistical methods used, number of replicates tested.
- Results (e.g., percent recoveries in the presence of interferent and any effects on precision) .
You should describe your method and results for testing the effect of incorrect (insufficient or excessive) sample volumes.
F. Safety and Reliability Testing
We recommend that you follow ISO 15197 concerning safety and reliability testing.
In your 510(k), you should include a brief summary of testing and acceptance criteria for performance across the range of your recommended testing conditions for temperature, humidity, altitude, and storage. You should also include certification for electromagnetic compatibility (EMC) testing. Many BGMS safety issues are covered by an FDA-recognized standard, International Electrotechnical Commission (IEC) 60601-1: Medical Electrical Equipment - Part 1: General Requirements for Safety, 3rd edition. This standard covers a broad range of device hazards, including electrical safety, mechanical, excessive temperature, fire, liquid ingress, spillage, and cleaning. If you rely on conformance to this standard, you should also conform to the collateral standards in the 60601 series, unless your submission explicitly states otherwise. The relevant collateral standards include:
- IEC 60601-1-2: Medical Electrical Equipment - Part 1: General Requirements for Safety; Electromagnetic Compatibility – Requirements and Tests.
- IEC 60601-1-6: Medical Electrical Equipment: Collateral Standard: Usability, analysis, test and validation of human factors compatibility.
G. System Traceability and Validation of Calibration and Control Materials.
You should include the following information concerning calibrator and control material in your 510(k):
- A description of the traceability of the system’s calibration to a higher order, e.g., internationally recognized reference material and /or method. (ISO 14971, Section 4.3; ISO 17511 [4] ; and ISO 15197, Appendix B.
- A description of stability evaluations and results for calibrators and controls.
- A description of how you established the control material limits (allowable ranges). Ranges that are too broad may be incapable of reliably detecting unacceptable levels of imprecision or bias.
- A description of controls and calibrator matrices and results of testing for matrix effects. For example, this could include testing calibrator and control materials in parallel with patient samples of similar known values and comparing the bias and precision you observe.
VI. Description of the Device
We recommend that you include information, such as the following, in the description of your device in the 510(k). Typically much of this information is included in the User Manual. However, if any of the information is not appropriate for the intended user (e.g., highly technical explanations), you can include it in the 510(k) only.
General description:
- Physical components of the system (including diagrams where appropriate).
- Manufacturer’s performance specifications.
- Description and explanation of the chemical reactions.
- Description of the composition and levels of control material.
- User maintenance needs (e.g., batteries).
- Features of the device designed to enhance robustness, including ease of use.
Description of features controlled by the software:
- Displays and user messages: This includes how the system determines and displays the glucose concentration; messages or displays that appear while a user is taking a measurement; and features such as how a user can retrieve past results from storage in the device.
- Error messages: This includes any error messages that the BGMS displays. Examples include displays or messages that the user sees when a strip is inserted incorrectly or removed prematurely or when damaged, incorrect, or deteriorated strips are used. You should also describe the error tolerance for user actions, such as these, that are inconsistent with device operation.
- User prompts: You should describe prompts that the device provides to the user, expected user responses, and timing issues (e.g., how quickly does the user need to respond, what happens if they respond after the allowed time). Examples of a user prompt are messages to the user to add specimen to the strip, insert it in the meter, calibrate the meter, or store a result.
- Alarms and other feedback: You should describe how the system responds to errors in user action, user inaction, or system status, e.g., low batteries or high ambient temperature. You should describe how the system detects and alerts the user when glucose levels are outside of the linear range of the system. You should describe any self-diagnostic routines that the system performs.
You should also identify the expected responses by the user to messages. This includes whether and how the user should input information or press certain buttons to correctly set up the meter or to respond to a message.
In prescription use, the health care professional may input patient information when performing a test. You should describe how the system recognizes and stores such information and how the health care provider downloads that information. You should describe features, such as maintaining a calendar and clock to record events. You should provide instructions for any features of the device that a user needs to set and correct.
You should clarify how the system recognizes and distinguishes calibration or control materials from patient specimens. You should also explain how the system compensates for differences between strip lots or strip types.
VII. Software
For software documentation of glucose meters, components, and accessories, you should follow the “Documentation in a Premarket Submission” section in the “Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices”. Definitions for terminology used below can be found in “Glossary of Computerized System and Software Development Terminology”. The following sections address specific issues for BGMS devices and complement the software guidance.
Level of Concern
FDA considers glucose meters to be a moderate level of concern. Glucose results displayed by the meter will be the basis for treatment, including determination of insulin dosage by the patient or health care provider. Incorrect glucose results or failure of the software to detect an error could result in improper diabetes management.
Very simple accessory software programs intended for tracking glucose measurement might in some cases have a minor level of concern. In contrast, advanced programs intended to make treatment recommendations, such as insulin dosage, may have a major level of concern.
You should include information on your analysis and determination of the level of concern for your device. The level of concern determines the degree of software documentation that you should include in your 510(k), as described in “Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices.”
Software Description
You should describe the features controlled by the software. Typically, a glucose meter interfaces with the user and with other components of the device through the software. You should describe how the software interacts with and controls the features referred to in the Description of the Device (Section VI, above).
Risk Analysis
As noted above in Section IV.B.5, you should conduct a risk analysis to identify potential causes of system failures. You should use this as a basis for rating associated risks prior to mitigations, identifying appropriate mitigations, and determining residual risks to characterize the operational limits of the device.
You should include the following concerning your risk analysis:
- Identification of your risk analysis methods (e.g., FMEA, fault tree analysis). You should describe the method you used in sufficient detail, so that FDA can readily interpret the results of your analysis during review.
- Potential sources of errors, or causes of hazards, including human factors, that you considered in the evaluation. (For examples, see Appendix 2.)
- The probability of the causes of the hazards you considered and the severity of the hazards.
- Your criteria for an acceptable level of risk and confirmation that this has been achieved.
- Mitigation measures you incorporated into the device, including measures to prevent or accommodate use errors, measures to address user characteristics (particularly vision and manual dexterity), and measures to mitigate device failure. This can be included in the device description section of your 510(k).
- A summary of validation testing and results for the mitigation measures you incorporated into your device, which were performed under conditions that challenge the system.
You should describe any malfunctions due to environmental stress on the meter, test strips, calibrators, quality control materials, and on the measurement process itself. Appendix 2 contains a table of examples of potential sources of error.
As noted above (Section IV.B.), you should validate that your BGMS does not generate results when the user performs a critical step in an incorrect manner or when an instrument error that compromises the test result occurs. This typically involves testing replicates of a sample while incrementally varying common variables, one at a time.
We recommend that you present the risk analysis in a tabular format, such as that discussed in the "Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices.”
Software Requirement Specifications
You should describe the specifications for the software in your BGMS. These specifications should be based on the role of the software or firmware on your risk analysis and the requirements you impose based on your description and indications for use of the device. It is extremely important that the particular clinical needs of the intended users drive these specifications. The intended users that you should keep in mind, when developing specifications, may include lay diabetics or health care professionals, as appropriate for your device. For example, you should consider whether the software interface with the user will be appropriate for someone experiencing diabetes related eyesight loss. The displays should be appropriate, and signals, including error messages, should be sufficiently audible and/or visible. In your 510(k), you should discuss software specifications that control for these types of concerns.
The manufacturer’s software specifications provide a baseline for software development and should be traceable to Verification Validation and Testing (VV&T) of the software. You should evaluate each functional requirement of the software and link it to VV&T activities in the submission.
Verification, Validation, and Testing
You should describe VV&T activities at the unit, integration, and system level in your QS records. You should include error codes, result flag validation, system level test protocols, pass/fail criteria, and testing results (linked to the risk analysis and software functional requirements). You should discuss your results and how you tested each of the items listed below:
- Fault, alarm, and hazard testing
- Error, range checking, and boundary value testing
- Timing analysis and testing
- Special algorithms and interpretation testing and analysis
- Stress testing
- Device options, accessories, and configurations testing
- Communications testing
- Memory utilization testing
- Qualification of off-the-shelf software
- Acceptance and beta site testing
- Regression testing
You should provide a signed statement stating that you have developed and validated all versions according to the described software development, verification, and validation plan. In the signed statement, you should include language to the effect that testing meets predetermined acceptance criteria and that additional versions will follow these activities and meet acceptance criteria before release.
Complying with an FDA recognized software consensus standard (e.g., AAMI SW68)4 may reduce the amount of VV&T documentation you should submit. However, even in this case, you should identify the following for hazards not reduced to a “broadly acceptable” level in your risk analysis: mitigations, VV&T activities, and any impact on safety and effectiveness for the glucose meter and its accessories.
Other Software Information
Other documentation listed in FDA’s “Guidance for the Content of Premarket Submissions for Software Contained in Medical Devices” include the Architecture Design Chart, Design Specifications, traceability analysis, summary of software development, Revision level history, unresolved anomalies, and release version number.
When to submit a premarket submission for modifications to device software or software components
A premarket notification is sometimes appropriate when modifications to software and/or software components of devices affect device performance. See “Deciding When to Submit a 510(k) for a Change to an Existing Device” for guidance. You should contact the Chemistry and Toxicology Division in the Office of In Vitro Diagnostic Device Evaluation and Safety (OIVD) when in doubt about whether a modification may need a traditional or a special 510(k).
VIII. Labeling
The labeling of a BGMS includes a user manual, separate package inserts for test strips and controls, and box and container labels for the meter, test strips, and control materials.
The labeling must address all applicable items in 21 CFR 809.10. We recommend that you also address all items outlined in ISO 15197, Section 5 [2]. You may use international symbols for professional-use BGMS devices (see FDA guidance titled, “Use of Symbols on Labels and in Labeling of In Vitro Diagnostic Devices Intended for Professional Use”). However, if the labeling is for over-the- counter devices, international symbols should not be used. The meter label should include applicable safety certification symbols, e.g., UL.
The package inserts for test strips and controls and the user manual should be simple, concise, and easy to understand. Graphics such as line drawings, illustrations, symbols, icons, photographs, tables, and graphs are very useful tools. You should use the same terms throughout the labeling to identify the device and its parts, avoiding synonyms or alternate phrases. You should assess the readability level of the package insert and identify your evaluation method and results in the 510(k). Information for developing home use labeling can be obtained from the following documents:
- “Guidance on Medical Device Patient Labeling; Final Guidance for Industry and FDA”
- "Write it Right”
- CLSI GP-14, “Labeling of Home-Use In Vitro Testing Products; Approved Guideline,” 1996
- “Assessing the Safety/Effectiveness of Home-use In Vitro Diagnostic Devices (IVDs); Draft Points to Consider Regarding Labeling and Premarket Submissions”.
In cases where technical information, required by 21 CFR 809.10(b), may not be understandable or useful for a lay user (e.g., chemical details of test principle or statistical analyses), this information may be presented in a separate section for healthcare professionals or followed by clarifying statements appropriate for lay users. Separate manuals and package inserts may be written for over-the-counter and professional use.
The premarket notification should include labeling in sufficient detail to satisfy the requirements of 21 CFR 807.87(e). We recommend that you include the following in your labeling. These recommendations are aimed at assisting you in preparing labeling that satisfies the requirements of 21 CFR 807.87(e) and 21 CFR 809.10.
- Operational steps.
- Description of testing conditions that may cause clinically significant errors with your device (e.g., specific drugs, oxygen therapy, testing with venous, arterial, or neonatal blood, high altitude, or EMC interference). You should indicate the most extreme conditions (e.g., the highest altitude) at which device performance was shown to be acceptable during testing. If your meter does not differentiate between glucose and other sugars such as maltose, you should include a warning that the meter should not be used for patients treated with any therapy containing these substances. See "FDA Reminders For Falsely Elevated Glucose Readings From Use of Inappropriate Test Method".
- Descriptions of clinical situations in which the BGMS performance may not be acceptable. You should clarify that critically ill patients (e.g., those with severe hypotension or shock, hyperglycemic-hyperosmolar state , hypoxia, severe dehydration, diabetic ketoacidosis) should not be tested with blood glucose meters because inaccurate results may occur.
- Limitations against alternative site use or neonatal use (unless appropriate studies are performed and included in the 510(k)). (See FDA document entitled “Points to Consider for Portable Blood Glucose Monitoring Devices Intended for Bedside Use in the Neonate Nursery.” http://www.fda.gov/cdrh/ode/122.pdf) When alternative sampling sites are indicated you should clarify that results from these sites may lag behind finger stick during periods of glucose change, or reduced peripheral circulation (e.g., shock).
- Optimal operating conditions of the system and appropriate storage conditions of the meter, strips, and controls.
- A section on the use of control and calibration steps to ensure optimal performance of the system. This includes recommendations for how and when to perform quality control checks and instructions for what to do if the control material values are not within the manufacturer’s allowable range. As part of the quality control information in your labeling, we recommend that you advise users that they should periodically compare the test system to another test system known to be well maintained and monitored by a healthcare provider.
- Performance data summaries from in-house and user studies. You should describe all study description summaries and results in the package inserts for professional use. For presentation of accuracy in particular, see Figures 1 and 2, below.
- A list or summary of error messages, descriptions of what those error messages mean, and appropriate troubleshooting procedures for those error messages.
- A U.S. telephone number for user assistance, including hours of operation, in the manual and in the package inserts. If user assistance is not provided 24 hours/7 days a week/365 days a year, provide instructions for what measures the user should take when user assistance is not available.
- Statements instructing users to contact their healthcare provider, if they obtain results that are not consistent with the way they feel, and to not change their medication regimen without approval from a healthcare provider
- Information regarding whether glucose measurements are reported as whole blood or plasma equivalents.
We provide the following data presentations as examples for capturing the results of your accuracy studies. We recommend using the presentation shown in Figure 1 for professional use devices. We recommend using the presentation shown in Figure 2 for lay use devices.
|
Figure 1: The list-specific meter and specific reagent strips blood glucose meter system was tested on 100 capillary blood samples, and the results were compared to the XYZ laboratory method (specify matrix if different). The tables below show how well the two methods compared: The tables show differences in glucose values between the “new device” and the “comparator.” The first table represents samples for glucose results lower than 75 mg/dL. The second table represents samples for glucose results greater than 75 mg/dL.
*Acceptance criteria in ISO15197 are that 95% of all differences in glucose values (i.e., comparator glucose value minus new device glucose value) should be within 15 mg/dL of for glucose values less than 75 mg/dL, and within 20% for glucose values greater than 75 mg/dL. Note: When glucose meter results are compared to the laboratory results, difference values below 75 mg/dL are expressed in mg/dL, while those above 75 mg/dL are compared in percent. |
|
Figure 2: Diabetes experts have suggested that glucose meters should agree within 15 mg/dL of a laboratory method when the glucose concentration is lower than 75 mg/dL, and within 20% of a laboratory method when the glucose concentration is 75 mg/dL or higher. The chart below displays how often the [Product Name] meter achieves this goal. The chart is based on a study done on 100 patients to see how well the [Product Name] compared to laboratory results.
|
Appendix 1. Alternative Sample Site Testing (AST)
Sampling from anatomical sites other than the fingertip, i.e., forearm, upper arm, thigh, calf, palm, may be indicated for some BGMS devices. Some users may prefer obtaining blood from alternative sampling sites because of less pain or greater choice in puncture sites. However, studies have shown that during times of rapidly changing glucose (i.e., after meals, medication, or exercise), the glucose level in blood from the alternative site may be significantly different from the glucose level from the finger. Additionally, glucose levels may not rise as high or fall as low as levels in the fingertip. This can result in delayed or undetected hypoglycemia when glucose is measured in alternative sites during non-fasting times.
If you have not presented studies that demonstrate equivalence between AST and fingerstick testing, you should include a prominent warning in the labeling against use of the device for AST. Alternatively, if you want clearance for AST, you should provide in your 510(k) a description of studies and results performed with your device to evaluate whether results at alternative sites are equivalent to fingertip. In your indications for use, you should evaluate performance separately for each alternative site. As with fingertip testing, validation for BGMS devices intended for lay use in AST should be performed with lay users in a simulated intended use setting.
If the AST studies you conduct do not include any challenges that increase or decrease glucose levels, you should include the following limitations in your package insert:
- Alternative site results may be different from fingertip results when glucose levels are changing rapidly (e.g., after a meal, after taking insulin, or during or after exercise).
- Do not rely on test results at an alternative sampling site, but use samples taken from the fingertip, if any of the following applies:
- you think your blood sugar is low
- you are not aware of symptoms when you become hypoglycemic
- the site results do not agree with the way you feel
- after a meal
- after exercise
- during illness
- during times of stress
If your labeling does not include the above limitations, you should present in your 510(k), a description and the results of time course studies that include “challenges” that increase and decrease glucose concentrations. The studies should demonstrate that alternative site measurements safely and effectively mimic fingertip measurements over time relative to the challenge time point. An appropriate study would include testing of paired samples at various time points using both the alternative site and the fingertip. You should design the study to demonstrate that (1) there is not a significant time delay in detecting rising and falling glucose levels and that (2) the AST captures the magnitude of hypoglycemic and hyperglycemic values shown with the fingertip testing.
Examples of challenges to increase glucose concentrations may include defined meal tests or oral glucose tolerance tests. Examples of challenges to decrease glucose concentrations may include giving rapid-acting insulin or oral hypoglycemic agents or vigorous exercise. Challenges should achieve peak rates of glucose increase and decrease of at least 2 mg/dL per minute forat least 30 minutes (>60 mg/dL change over 30 minutes) in order to provide an adequate challenge to evaluate BGMS performance in a non-steady state situation.
The appropriate number of study subjects and time points to sample within the specified time interval depends upon use instructions and performance claims for the BGMS (e.g., accuracy, precision, and interference). In any case, the study should be sufficiently powered to allow the BGMS measurement error to be distinguished from differences due to AST. Your data analysis method should address glucose rates of change, the clinical significance of the values (error grid analysis or its equivalent), and plots of individual patient data over time (as opposed to averages of all patients over time). You should present results to demonstrate that (1) there is not a significant time lag time (or lead time) in detection of rising and falling glucose levels by the AST relative to fingertip testing, and (2) the AST captures both the maximum and minimum values observed with the fingertip testing. We recommend that you contact OIVD to discuss appropriate study designs for BGMS devices intended for use with samples from alternative sampling sites during times of changing glucose levels.
The same accuracy criteria used to assess differences between glucose meter and laboratory results obtained from fingertip capillary blood samples apply to other anatomic sampling sites. Regardless of the anatomic site used for glucose meter testing, the reference sample should be obtained at the fingertip, since this is the value commonly used for therapeutic decisions.
Appendix 2. Potential sources of error to consider
The following table lists potential sources of error associated with the design, production, and use of BGMS devices. We do not intend for this to be a complete list. You should consider all sources of error based on your knowledge of your specific device. Documents such as CLSI EP-18A [7] and ISO 14971 [1] also provide lists of preanalytical, analytical, and post-analytical errors to consider.
| Category |
Source of error or failure |
| Operator |
Failure to follow procedure correctly, for example:
|
| Reagent |
|
| Environmental |
|
| Software |
|
| Hardware |
|
| System |
|
| Clinical |
|
Appendix 3. Third-Party Strip Manufacturer Considerations
This section is addressed to third-party strip manufacturers. Third-party glucose strips (also known as generic glucose reagent strips, independent component glucose strips, and off-brand component glucose strips) are strips manufactured for a meter by a non-contract manufacturer other than the original meter manufacturer. You (third party manufacturer) should address all sections in this guidance document regarding risk management (analytical performance evaluation and user performance) using the meter that you recommend for use with your strip. You should include testing of lot-to-lot variability in your accuracy and precision studies. Similarly to the original manufacturers, you should test your strip under all conditions for which the meter is intended, including (as appropriate) alternative anatomical sites and matrices and provide this information in your 510(k).
You should ensure that the software is fully functional with your strip. In the 510(k), you should provide results demonstrating that your strip generates all the same messages and error codes that are generated by use of the original strip.
Control material tested with your strip should also generate the same results as are generated with the original strips. Similarly, the allowable ranges with that material should be the same.
Minor variations between different lots of strips often occur. To minimize result variability between strip lots, manufacturers typically assign a unique code corresponding to a specific algorithm programmed into the meters for each lot. Users enter these codes into the meter when a new lot of strips is started. A third-party manufacturer is not likely to have access to the specific algorithms. Consequently, you will not be able to provide the same codes with strips as the original manufacturer. Thus, in the 510(k), you should provide a description of the steps you are taking to assure that the third-party strips correspond to the correct meter algorithms that were provided in the 510(k).
Incorrect third-party test strip dimensional tolerances (e.g., size and shape) can stress and/or damage a meter's electrical connectors, making further measurements impossible. Because third party strips are made independently of the meter and the original meter manufacturer may have modified the meter or created other limitations without the third-party strip manufacturer’s knowledge, it is important that you demonstrate that each element of the interface between the strip and meter is the same as the original strip manufactured for use on that meter. In your 510(k), you should compare your strip to the original strip used with the meter in terms of dimensions, reagents, matrix and other features of the strip meter interface.
You should provide labeling that is similar to that of the original meter. Customer service should be available even after business hours to help users troubleshoot performance problems at critical times. You should indicate in your labeling the telephone number and hours for customer service. You should include the specific meters that have been validated for use with the third-party strips. You should indicate the date of the last validation of the strips with those meters and clarify that performance characteristics have not been determined for meters manufactured after that date.
References
[1] ISO 14971. Medical Devices – Application of risk management to medical devices. International Organization for Standardization, Geneva Switzerland; 2003.
[2] ISO 15197. In vitro diagnostic test systems – Requirements for in vitro blood glucose monitoring systems for self-testing in managing diabetes mellitus. International Organization for Standardization, Geneva Switzerland; 2003.
[3] Consensus statement on self-monitoring of blood glucose. American Diabetes Association. Diabetes Care, 10 (1), 1987, pp 95-99.
[4] ISO 17511. In vitro diagnostic medical devices – Measurement of quantities in biological samples – Metrological traceability of values assigned to calibrators and control materials. International Organization for Standardization, Geneva, Switzerland; 2003.
[5] CLSI EP6-A. Evaluation of the linearity of quantitative measurement procedures: A statistical approach; Approved Guideline Clinical and Laboratory Standards Institute, Wayne, Pennsylvania 19087-1898 USA, 2003.
[6] CLSI EP7-A. Interference Testing in Clinical Chemistry; Approved Guideline. Clinical and Laboratory Standards Institute, Wayne, Pennsylvania 19087-1898 USA, 2003.
[7] CLSI EP18-A. Quality Management for unit use testing; Approved Guideline, Clinical and Laboratory Standards Institute, Wayne, Pennsylvania 19087-1898 USA, 2003.
1 Sometimes referred to in the literature as SMBG (Self-monitoring blood glucose device).
2 See “Points to Consider for Portable Blood Glucose Monitoring Devices Intended for Bedside Use in the Neonate Nursery” for guidance on blood glucose monitoring for neonates. See “In Vitro Diagnostic Glucose Test System” for guidance on laboratory glucose analyzers.
3 We use the term robustness here to refer to failure-resistance of the device under challenging conditions such as long-term storage or foreseeable environmental extremes.
4 AAMI SW68:2001; Medical Device Software – Software Life Cycle Processes.