Formative usability testing – Frequently Asked Questions?

Formative usability testing is not a regulatory requirement, but it is necessary if you want to successfully develop medical devices.

What is the difference between formative and summative usability testing?

“Formative” tests are any usability tests that you perform during the development process, while “summative” testing is the final usability testing you perform to validate that your chosen user interface is effective. Many design teams perform formative testing of one kind or another without even realizing that is what they are doing. Unfortunately, design teams often forget to document the testing they performed during prototyping and product development. Formative usability testing probably always existed as part of product development, but not everyone recognizes the term and identifies the work they have done as “formative.” The most important reason for documenting formative usability testing is to identify which user interface designs failed and why so that future design teams can learn from your failures.

Why don’t more companies do usability testing?

Everyone likes to believe they can skip steps in the learning process, but some lessons can only be learned the hard way. When a medical device design team is developing a user interface for a new product, they need to learn which designs will fail and why before they can fully understand how to design the best user interface for the device. Therefore, most design and development teams will select a user interface that they are familiar with or they see used by a competitor product. The team will not always test the proposed design solution, because they have no reason to believe that the chosen interface will fail. Unfortunately, this can lead to failure later in the design process. Then the team will need to backtrack and repeat the evaluation of various interface designs.

What is the best approach?

“Fail small and fail fast” is the best advice for anyone performing formative usability testing. Instead of writing a lengthy protocol and recruiting 10 subjects to evaluate your proposed user interface, you might consider building a couple of different prototypes and asking two or three people which prototype they prefer and why? Another simple question is, “Tell me what you think of this design?” Iterative formative testing over time with different users is better than one single testing session with a lot of users. It is also better to start collecting formative usability testing data as early in the development process as possible. Gathering data earlier in the process will ensure that users direct the development of your device instead of the design team developing a new device in a direction that is not preferred by users.

When during the design process should formative testing be planned?

Formative testing should be planned during the development phase of the design process. During this phase, medical device manufacturers evaluate multiple design solutions as risk controls for their devices. Use-related risks should be included in this, and the formative usability testing is intended to identify which user interface will do the best job of eliminating the use errors. It is important to evaluate these potential user interfaces and to verify that there are no use errors that the design team overlooked during this phase of the design. This is also the phase of design when the instructions for use are developed and user training is developed. All of this formative usability testing should be completed prior to your design freeze and the start of the verification and validation testing.

What are the different types of formative testing?

Formative usability testing can be used as a pilot for your summative usability testing protocol prior to scheduling the final testing. However, there are many other types of formative testing. The most common reason for doing this testing is to identify any potential use errors that were not originally identified in your user-related risk analysis (URRA). Another type of testing is to simulate use of the device to make sure that every user task is identified in the instructions for use. Finally, design teams will conduct formative usability testing to develop training materials for training new users on how to properly use your medical device.

Which types of formative tests are the most useful?

Use-related risks are difficult to identify unless you conduct simulated use testing with your device. Therefore, you need to get your device in the hands of your intended users, in the intended use environment, and ask them to simulate the use of the device. It is not critical to evaluate a specific number of users. Two or three users might be enough, but simulated use by intended users in the intended use environment is essential to give you the information you need regarding potential use errors. It is also important to avoid “leading” the users. Instead of asking users to perform a specific task, ask users to show you how they would use the device. Ask them what they like about the device, and ask them what they don’t like about the device. Ask users what they think about the device, and ask them how it compares to other devices they are already using.

Who should you recruit for your formative usability testing?

You should start your human factors process by defining the intended user of your device and by defining if there is more than one user group. You then should recruit subjects that are within this user group(s). You can use employees or friends to help you with initial feedback about the usability of your device’s user interface. However, what seems intuitive to one person may be the opposite for other people with different experiences. Even the sequence of steps in which users perform the same tasks can impact usability. Therefore, be very cautious about relying upon data collected only from subjects that are outside your intended user group. Most companies disregard this advice because they are unsure of how to recruit their intended users. However, if your company has difficulty identifying intended users for testing, you will also have difficulty marketing and selling your device later. This struggle may be an indicator that you need to involve marketing and salespeople that can get your prototypes in the hands of the intended users.

How should you document formative studies?

When you are performing summative usability testing you already know exactly what your use-related risks are and you have a list of critical tasks that you are trying to verify users can perform without use errors. Because these tasks are clearly defined, it is easier to write a protocol and it is easier to design data collection forms for study moderators to use. In contrast, when you are conducting formative usability testing you are trying to identify use errors that you are not already aware of. Therefore, it is much harder to write a detailed protocol and design a data collection form. For this reason, it is critical to capture the data with video recordings. This is a safety measure you are taking to ensure that you will not miss valuable use errors or use tasks that you had not already identified. The use of video to record data allows the moderator to focus on observation and interviewing users with open-ended questions. This will generate the most value for your design team during the development process.

Where is testing performed?

While the design team is developing the list of design inputs for your new device, the team must create a definition for the intended users and the intended use environment. The formative usability testing and summative testing should be conducted in the intended use environment or you will need to simulate that use environment. If you are struggling to figure out how to simulate the intended use environment, you should systematically identify the characteristics of the intended use environment. These characteristics include temperature, humidity, ambient noise, other equipment that is present, the number of people present, and the dimensions of the space. If you have a room available with temperature and humidity control, you can add ambient noise by recording the intended use environment. You can rent equipment, or you can place objects of the same size in the space. You can also identify the workspace restrictions by taping the floor to establish boundaries for the simulation. By adding these characteristics to a simulated environment, you open the possibilities for additional places that can be used for formative usability testing.

What will happen if you skip formative testing?

If you skip formative usability testing, you will increase the possibility of failing your summative usability testing. If this happens, then your summative testing becomes your formative usability testing. After you fail, you will need to revise your testing protocol and repeat the study. Another possibility is that you will fail to identify a potential use error. If the FDA identifies this use error you will need to repeat your testing. If the use error is never identified, then you may end up with complaints or medical device reporting of use errors. In extreme cases, this could result in serious injuries or death.

Is monitoring every procedure required?

Process monitoring is required but do you know whether monitoring every procedure is required by the FDA QSR or ISO 13485?

One of the elements that Medical Device Academy has incorporated into each procedure we created in our turnkey quality system is a section titled, “monitoring and measurement.” The purpose of this section is to force each process owner to identify a process metric for monitoring every procedure. In some cases, we suggest a metric that would be appropriate for most companies establishing a new quality system. In other procedures, we use the following default text:

“Enter a quality metric that you want to track for this process in accordance with ISO 13485:2016, Clause 8.2.5 and the procedure for Monitoring, Measurement, and Analysis (SYS-017).”

Where are the requirements for process monitoring in 21 CFR 820?

Some of the companies that have purchased our turnkey quality system have asked, “Is it required to monitor and measure something in every procedure?” In general, it is not a specific requirement to have a metric specified in each procedure. In fact, if your quality system is not ISO 13485 certified, there are actually only a few places where the US FDA requires monitoring. The FDA does not have a section specific to monitoring and measurement of processes, but there is a section of the regulations specific to statistical techniques (i.e. 21 CFR 820.250). However, it does not state in the QSR that statistical analysis is required for all processes. In fact, there are only six instances where the word “statistical” is used:

• 21 CFR 820.100(a)(1) – “Analyzing processes, work operations, concessions, quality audit reports, quality records, service records, complaints, returned product, and other sources of quality data to identify existing and potential causes of nonconforming product, or other quality problems. Appropriate statistical methodology shall be employed where necessary to detect recurring quality problems;”
• 21 CFR 820.200(b) – “Each manufacturer shall analyze service reports with appropriate statistical methodology in accordance with § 820.100.”
• 21 CFR 820.250 – “(a) Where appropriate, each manufacturer shall establish and maintain procedures for identifying valid statistical techniques required for establishing, controlling, and verifying the acceptability of process capability and product characteristics. (b) Sampling plans, when used, shall be written and based on a valid statistical rationale. Each manufacturer shall establish and maintain procedures to ensure that sampling methods are adequate for their intended use and to ensure that when changes occur the sampling plans are reviewed. These activities shall be documented.” Note: the other two instances are the title of 21 CFR 820.250.

The word “monitoring” is equally rare (i.e. 4x) in the QSR:

• 21 CFR 820.70(a) – “Each manufacturer shall develop, conduct, control, and monitor production processes to ensure that a device conforms to its specifications…Where process controls are needed…(2) Monitoring and control of process parameters and component and device characteristics during production.”
• 21 CFR 820.75(b) – “Each manufacturer shall establish and maintain procedures for monitoring and control of process parameters for validated processes to ensure that the specified requirements continue to be met…(2) For validated processes, the monitoring and control methods and data, the date performed, and, where appropriate, the individual(s) performing the process or the major equipment used shall be documented.”

Where are the requirements for process monitoring in ISO 13485:2016?

ISO 13485:2016 has a section specific to monitoring and measurement of processes (i.e. Clause 8.2.5). In addition, the word “monitoring” occurs 52 times in the standard and there are 60 incidents of some variant or the exact word. , but there is a section of the regulations specific to statistical techniques (i.e. 21 CFR 820.250). However, it does not state in the QSR that statistical analysis is required for all processes. In fact, there are only six instances where the word “statistical” is used. There are four Clause headings that actually include the word monitoring:

• Clause 7.6, Control of monitoring and measuring equipment
• Clause 8.2, Monitoring and measurement
• Clause 8.2.5, Monitoring and measurement of processes
• Clause 8.2.6, Monitoring and measurement of product

In Clause 1, Scope, and Clause 4.1.5, the Standard states that any outsourced processes remain the responsibility of the company and must be accounted for in the quality system by monitoring, maintaining, and controlling the processes.

Monitoring of risk is included in the definition of “risk management” in the Standard (i.e. Clause 3.18).

Clause 4.1.3 states that the organization shall, “b) ensure the availability of resources and information necessary to support the operation and monitoring of these processes…d) monitor, measure as appropriate, and analyze these processes.”

Clause 4.2.3 states that the contents of the Medical Device File (i.e. MDR or TF), shall include, “d) procedures for measuring and monitoring.”

Monitoring and measurement of processes and product are required inputs to the Management Review in Clauses 5.6.2e) and f).

Clause 6.4.1 requires a procedure for monitoring the work environment if it can have an effect on product quality.

Clause 7.1 requires the company to consider including monitoring in product realization planning.

Clause 7.4.1 requires a plan for monitoring suppliers.

Clause 7.5.1 requires monitoring production and service, including the monitoring of process parameters and product characteristics.

Clause 7.5.6 requires monitoring of validated process parameters.

Clause 7.5.8 requires identification of status with regard to product monitoring and measurement (i.e. inspection status).

Clause 7.6 requires monitoring and measurement of calibrated devices and validation of any computer software used to monitor calibrated devices.

Clause 8.1 states that companies shall plan and implement monitoring and measurement of processes.

Clause 8.2 is titled, “Monitoring and measurement.”

Clause 8.2.1 requires monitoring of customer feedback.

Clause 8.2.5 requires monitoring of processes to ensure planned results are achieved.

Clause 8.2.6 requires monitoring of products to ensure product requirements have been met.

Clause 8.4 requires data analysis of monitoring data from at least six different processes:

1. Feedback
2. Conformity to product requirements
3. Characteristics and trends of processes and products, including opportunities for improvement
4. Suppliers
5. Audits
6. Service reports, as appropriate

In summary, while not every single clause that requires a procedure includes a requirement for monitoring, there are a number of processes where the requirement to monitor the process is explicitly stated.

Why do all of our procedures include the requirement for metrics?

Medical Device Academy expanded the requirement for monitoring to all procedures for five reasons:

1. Quality objectives must be “established at relevant functions and levels within the organization.” Therefore, establishing monitoring requirements for each procedure ensures that top management has metrics for every process and a lack of data is never an excuse for not establishing a new quality objective when improvement is needed.
2. If every procedure has a requirement for monitoring, then employees don’t have to remember which processes require monitoring and which processes do not explicitly require monitoring.
3. The process approach to auditing includes metrics of the process as one of the seven items that are included in every process turtle diagram, and therefore, including metrics for each procedure facilitates the process approach to auditing.
4. If a company does not have a process metric already established, it is often difficult to perform an investigation of the root cause of quality issues. If a metric is already being monitored for the process, this facilitates the investigation of the root cause and you can use the baseline monitoring data to help you establish effectiveness criteria for the corrective action.
5. Finally, most companies struggle to identify preventive actions as required by Clause 8.5.3, and we have found that data analysis of monitoring data is the best source of identifying new preventive actions.

What are the disadvantages when you monitor and measure something in every procedure?

The primary reason for resistance to identifying a metric for monitoring in every procedure is that it will increase the workload for the employees responsible for that process. However, monitoring of data does not always increase workload. In fact, when process data is recorded in real-time on a run chart it is often possible to identify a trend much earlier than when data is simply recorded and subjected to monitoring.

• Example #1: The automatic tracking of toner in a printer tells HP when to ship you a new toner cartridge before you need it. This ensures that there is no loss in productivity because you never run out of ink or the ability to print documents.
• Example #2: Companies will use project management software (e.g. Asana) to monitor labor utilization. This will help identify when a specific resource is nearing capacity. When this occurs, the project manager can add time buffers to prerequisite steps and adjust the starting date of the resource-limited tasks to an earlier starting date. This ensures that more time is available to finish the task or to take advantage of resource availability at an earlier date.
• Example #3: Monitoring the revision date for procedures helps the document control process owner identify procedures that should be evaluated for the need to be revised and updated. Often this is articulated as a quality objective of reviewing and updating all procedures within 2 years. This also ensures that procedures remain current and compliant with regulatory requirements.

What are the advantages of monitoring every procedure?

The phrase “what gets measured gets managed” is a popular business philosophy that implies measuring employee activity increases the likelihood that employees will complete a task or perform it well. In contrast, if a process is not monitored, employees may assume that it is not important and the tasks may be skipped or completely forgotten. Setting quantitative goals is also sometimes integrated with economic incentives or bonuses that are granted to individuals and teams.

FDA transition from QSR to ISO 13485

The US FDA is planning its transition from 21 CFR 820 to ISO 13485 as the quality system criteria. This will force companies to make adjustments to their quality systems and increase the amount of process monitoring performed. My general advice is to work with employees that are performing tasks to identify streamlined methods for monitoring those tasks without being overly burdensome. Then you and the employees you manage can analyze the data together and identify opportunities for improvement. When you do this, experiment with manual methods using whiteboards and paper charts that are visible in public areas first. Only implement automated solutions after you have optimized the data being collected and the frequency of data collection, and remember that not every process will benefit from automated statistical process control. Sometimes the simple approach is best.