How to make medical devices safely and efficiently using process controls

In the film Wolf of Wall Street Leonardo DiCaprio challenged his team to “sell me this pen,” implying that a good salesperson could sell anything. In a previous article I shared how to design anything using Design Controls, and the article you’re reading now is how to “make this pen” using Manufacturing Controls.

Design Controls should be slightly ambiguous to allow room for innovation, but Process Controls should be capable of making the same design indefinitely. Effective companies learn to efficiently transition from “design a pen” to “make this pen,” and medical device companies must do this while also complying with government regulations.
This article will teach how to comply with medical device manufacturing process controls. To keep it easy to understand I’ll use the example of how to make the pen we designed in the first article.

But first I’ll share some background on why regulations are required for medical devices. If you’d like, start by watching Leonardo DiCaprio sell a pen as he portrays a real-world Wall Street broker who went to prison for violating finance regulations. Unlike that character, most medical device companies are ethical; their challenge is ensuring employees understand healthcare regulations and apply them efficiently.
This article can take between 7 and 10 minutes to read, depending on if you skip the examples.
Background
Governments require that medical devices be manufactured following documented procedures. This is to protect patients, and the requirements are clear and concise: a manufacturing process must consistently produce the same product and be monitored throughout the lifetime of a product according to a documented plan.
These requirements are described in FDA 21 CFR 820.30, ISO 13485, and the European Union Medical Device Requirements. Most companies add additional requirements based on best-practices or guidance documents such as the FDA guidance for medical device process validation,

(maintained by the International Medical Device Regulators Forum), or the FDA guidance for pharmaceutical process validation.

Medical device guidance is usually based on lessons learned from mistakes. For example, recent lessons in Europe led to the EU-MDR after 400,000 people received toxic implants. In the United States 9,000 people received toxic hip implants caused by a small change in the manufacturing process that wasn’t validated, leaving machining oil inside of micro-pores in the metal. 3,500 people had their bones dissolve, requiring a second surgery and forever affecting their ability to walk.

It’s worth re-emphasizing that guidance documents are suggestions to help companies comply with requirements, and the requirements are simply to have a validated process that is monitored for effectiveness according to a plan.
Step 1: Plan

Effective planning is difficult to describe because it comes from experience and wisdom. Try to start with the end in mind and include multiple check-points for team members to contribute new information and iterate the plan.
“We recommend an integrated team approach to process validation that includes expertise from a variety of disciplines (e.g., process engineering, industrial pharmacy, analytical chemistry, microbiology, statistics, manufacturing, and quality assurance). Project plans, along with the full support of senior management, are essential elements for success.”
A thorough manufacturing plan would include monitoring the process over time and include a method to change the plan based on new information. For this article we’ll assume that a thorough plan was created and documented as as #PenProcessValidationPlan.doc for all team members to easily access.
#PenProcessValidationPlan.doc
This is for a simple example; skip it if you’d like
100 pens shall be made in normal operating conditions as described in the Performance Qualification, #Pen-PQ-Protocol.doc.
A sample of 15 pens shall be used for process validation statistics. The samples shall be weighed and compared to design output specifications of 2.08 ounces +/- 0.05 ounces; if any samples are outside of specifications the manufacturing team shall re-assess the manufacturing process or, if applicable in a higher-level policy, the design specifications.
The pens shall be statistically analyzed using company policy #ProcessStatistics.doc, calculating the average, standard deviation, and the Cpk for normally distributed data.
A Cpk of 1.33 is the minimum acceptable level unless this plan is revised with justification.
The process shall be repeated with other randomly selected groups of 10 pens for a total of five groups; all groups shall have a Cpk of 1.33 or higher.
This plan shall link to process monitoring plan to randomly sample future manufacturing lots and compare against this validation according to #Pen-Process-Monitoring-Plan.doc. Any deviation from statistics in this plan shall be addressed in a Corrective Action – Preventive Action (CAPA).
Any changes to this plan must be agreed upon by the manufacturing team with justification for the change.
In other words, a plan includes a plan to improve the plan. The initial plan should be team-driven and start with what is known or expected and improved based on what is learned. Any changes to the plan should include a diverse team, and changes should be documented for government auditors to review.
Step 2: Design Transfer

Design Transfer is a required step between Design Controls and Process Controls, between “design a pen” and “make this pen,” ensuring that all drawings, software, etc. created in the design phase are sufficient for a manufacturing process to consistently make the same product.
For new designs, products made for process validation are also used for design validation. Efficient companies know this and start their design process with the end in mind using concurrent design, simultaneously improving both the design and the manufacturing process. All validation testing would be conducted throughout concurrent design and the final tests would be when there are no more changes to either the design or manufacturing process.
Concurrent design is difficult to explain in an article because it’s more about communication and teamwork than regulations. A common trait of effective teams is a planning process that includes multiple check-points for team members to contribute new information and iterate the plan.

For this article, I’ll assume that the pen design is complete rather than being designed concurrently so that we can focus on manufacturing guidance.
Step 3: Installation Qualification
Installation Qualification, IQ, is recommended by guidance documents but is not required by regulations. The IQ lists everything necessary to make a product, from equipment to drawings to supplies.
You’ve seen real-world IQ’s if you’ve watched a cooking show. The chef is efficient because they had a crew ensuring that the ingredients were prepared ahead of time, knives were sharpened, the oven meets specifications, etc. An IQ is simply a detailed list of everything someone would need to replicate your manufacturing equipment and preparation.

For our pen, let’s assume a simple machine makes the plastic parts, another machine injects the ink, and a test station performs some form of verification and validation testing of critical design features. The IQ would be protocol to follow, usually in the form of a checklist. The completed checklist would be another document demonstrating that everything was installed according to plan.
#Pen-IQ-protocol.doc
This is for a simple example; skip it if you’d like
Each check box must be initialed by an operator trained to company quality systems and the completed protocol must be signed off by their manager and a representative from quality-control.
Ensure facility is capable:
__ 10X power outlets with 120V 60Hz electricity
__ Ventilation per government requirements, OSHA paragraph xyz(example)
__ Training records for company quality system of all personnel signing this protocol
Ensure all is available:
__ Plastic injection machine, Company ABC, Model #1234
__ Instruction manual, Plastic injection machine, Company ABC, Model #1234
__ Ink injection machine, custom designed by our company, part #Ink6789.dwg
__ Instruction manual, Ink injection machine, custom designed by our company, part #Ink6789.dwg
__ Pen assembly shake-test machine, custom designed by our company, part #PenTest5678.dwg
__ Instruction manual, Pen assembly shake-test machine, custom designed by our company, part #PenTest5678.dwg
Check for functionality:
__ Plastic injection machine, Company ABC, Model #1234, turns on and completes setup
__ Ink injection machine, custom designed by our company, part #Ink6789.dwg, turns on and performs self-check
__ Pen assembly shake-test machine, custom designed by our company, part #PenTest5678.dwg, turns on and performs self-check
Check for availability:
__ Design Outputs (high-level document listing all outputs)
__ Purchased materials are ready according to #PenProcessPlan.doc
__ Training requirement documents for this manufacturing process
Document this completed protocol:
__ #Pen-IQ-protocol-results.doc
Step 4: Operational Qualification
Operational Qualification, OQ, is recommended by guidance documents but is not required by regulations. Operational Qualification begins after a completed IQ protocol and ensures all equipment is tested in worse-case scenarios, and that a product can be made with worse-case specifications.
For the cooking show example, if we needed cheese for a recipe and specified a moisture content of 65-78%, worse-case testing would include the extremes of both 65% and 78% moisture, and we’d cook on the driest day of the year and the most humid day of the year.
For our pen example, an OQ protocol would include turning on all equipment simultaneously to ensure electrical outlets can handle full capacity, perhaps testing worse-case of other equipment operating simultaneously. The high and low tolerances of raw materials would be tested to ensure pens could be made at all extremes of specifications.
The details of worse-case vary for every product, which is why an OQ protocol should be team-driven and continuously improved based on real-world experience to ensure new lessons are included in all documents. Like an IQ, an OQ usually has a protocol and a result documented, #Pen-OQ-protocol.doc and #Pen-OQ-protocol-results.doc.
Step 4: Performance Qualification
Performance Qualification, PQ, is recommended by guidance documents but is not required by regulations. The PQ is simply making products in normal operating conditions. It should have a protocol and results, #Pen-PQ-Protocol.doc and #Pen-PQ-Protocol-Results.doc.
Combined, the IQ, OQ, and PQ are the foundation of manufacturing processes. But before making products for the public the process must be repeatable. In other words, you must test a few of the products and demonstrate that there’s no difference from the few products you test and the products you manufacture on a larger scale. Doing this requires statistical analysis and monitoring.
Statistics

A complete course in statistics is out of scope of this article, but I’ll emphasize the most common and important concepts using references from the United States National Institute of Standards and Technology Engineering Statistics Handbook, which is similar to examples in guidance documents on process controls.
Let’s assume that the pen design specifications include that each each pen weighs 2.08 ounces +/- 0.05 ounces, which means the lower limit would be 2.03 ounces and the upper limit would be 2.13 ounces. Our pen’s validation plan already told us how to do this.
Statistical analysis from #PenProcessValidationPlan.doc
This is for a simple example; skip it if you’d like
First, weigh all pens
Weights (ounces) for 15 pens is documented #PenWeightTestResults.doc
100% of pens must be within the upper and lower limits to proceed.
Second, calculate average (mean) and standard deviation

Mean = 2.08 (calculated from our example)
Standard Deviation = 0.01 (calculated from our example)
Third, ensure your data fits a standard, normalized bell curve.

If your data does not approximate a bell curve, collect more samples or apply mathematical techniques such as a Fischer Transform or Box-Cox Transform to convert the data to a normalized curve. Otherwise the following steps are invalid and other statistical methods should be used.
Fourth, calculate Cpk using your upper and lower limits

Upper limit = 2.13 ounces (given by design requirements)
Lower limit = 2.03 ounces (given by design requirements)
Average = 2.08 oz (calculated in our example)
Standard deviation = 0.01 oz (calculated in our example)
Cpk = 1.67 (calculated)
Note:

Cpk is sometimes confused with Cp, even in online tutorials or with highly qualified consultants. You should use Cpk, which should be >= 1.00. The minimum acceptable Cpk will depend on your product, your process, and your acceptable risk, and should be clearly stated in your plan as a minimum acceptable level.
Finally, ensure that your Cpk is greater than your minimum requirement and accurately represents your manufacturing process. If so, that sample size can be used for process validation.
For our pens, I’ll assume that we made many more than we tested, randomly selecting groups matching our sample size, repeating our statistics, and obtained the same or better Cpk values. In other words, our assumption of sample size was valid therefore our process is validated as being in control statistically.
If Cpk were worse for any group then our either sample size was too small or our process is unstable and can’t be validated; we’d have to improve our plan or our process.
A detailed plan would also include how to monitor manufacturing and continuously improve. It would also include guidelines for when to reevaluate the process or when to treat errors as single events. Single events may be acceptable if other steps in the process prevent future single events from impacting patients. In other words, subsequent steps in your process prevent faulty products from reaching patients.
People often misapply the rules when they don’t understand the concepts. These misapplications cause inefficient companies and risks to patients. I believe that 5 out of 4 manufacturing managers make mistakes using statistics (ha!) which is why I suggest studying guidance documents on process controls.

Software
All software used in manufacturing or quality control shall be validated, and all changes to software shall be re-validated. In other words, if you use software, spreadsheets, or automated process then those must be validated independently.
Software validation should be included in your plan. That’s not as challenging as it seems, but is beyond the scope of this article. The FDA has guidance for validating software used in manufacturing, and you may find examples on the internet or my blog.
Final Report

A final report would document all results and ensure the plan was completed, and any changes to the plan were documented with justifications. I suggest making the plan have a checklist of all documents necessary for process validation and using that checklist for the final report.
Monitor, improve, & re-validate
An ongoing program to collect and analyze product and process data that relate to product quality must be established (§ 211.180(e)). The data collected should include relevant process trends and quality of incoming materials or components, in-process material, and finished products. The data should be statistically trended and reviewed by trained personnel. The information collected should verify that the quality attributes are being appropriately controlled throughout the process.
All manufacturing processes eventually become less efficient due to unforeseen changes in materials from suppliers, inconsistent human factors such as operator techniques, etc. Or, the process validation didn’t have enough of a sample size to represent large-scale manufacturing so the process should be improved.
Efficient companies set “action limits” that are within control limits so that manufacturing trends can be adjusted before becoming a risk to patients. Guidance documents suggest using control charts such as X-bar and R charts.

In addition to control charts you can look for other indications of process drift through high scrap rates, re-works, or other last-minute solutions to keep a manufacturing line operating. A process monitoring plan would address all of these and continuously improve based on what each company learns.
#Pen-Process-Monitoring-Plan.doc.
This is a simple example; skip it if you’d like
This plan shall link to

#PenProcessValidationPlan.doc to ensure that the pen manufacturing process maintains validated according to the validation plan. Any drifting of the Cpk below validation levels shall be addressed

in a Corrective Action – Preventive Action (CAPA). Trends in Cpk , trends of pens rejected due to falling out of upper or lower limits, and numbers of reworked pens shall be discussed in annual senior management meetings and closed out with a documented justification.

Any changes to this plan must be agreed upon by a senior management team with justification for the change.
Efficient companies update their IQ, OQ, and PQ’s to accurately represent their real-world manufacturing. Many people feel they don’t have time to update process documents, saying, “I’m so busy chopping wood that I don’t have time to sharpen the axe.” The human tendency to work harder rather than sharpen their axe has been recognized for thousands of years, which is why effective companies have continuous improvement built into their quality system.

Changes to the IQ, OQ, or PQ may require re-validation of all or part of the manufacturing process. Not all changes require a complete re-validation, especially if the process is continuously monitored. But that decision should never be left up to just a few people; good people make unwise decisions when pressured by job timeline or when they are unable to foresee the consequences of actions. In the example of Sulzer’s hip stem recall, thousands of patients suffered because a manufacturing process wasn’t re-validated after what people assumed was a small change. Effective quality systems that are based on team-driven plans help medical device companies make wise decisions.

Next step
Product performance should be monitored in real-world use through post-market surveillance. This ensures real-word data is fed back into your design, manufacturing, and distribution. This is the intention behind international quality control standards such as ISO 13495:2016 and the EU-MDR that emphasize the process-approach to continuous improvement.
Keep in touch
I help companies produce products that benefit society, training and consulting on international regulations, innovation, & purpose-driven workplaces where problems are proactively solved as teams.

Contact meto discuss if I could help train your company, connect on Linkedin, where I infrequently post articles, or subscribe to this blog for yearly updates.
Post-Script
I sometimes use blogs in workshops, so I’m copying the entire text of FDA process control requirements here. There’s no need to read them unless you’d like to see how simple they are. I highlighted key concepts from this article.
Sec. 820.70 Production and process controls.
(a) General. Each manufacturer shall develop, conduct, control, and monitor production processes to ensure that a device conforms to its specifications. Where deviations from device specifications could occur as a result of the manufacturing process, the manufacturer shall establish and maintain process control procedures that describe any process controls necessary to ensure conformance to specifications. Where process controls are needed they shall include:
(1) Documented instructions, standard operating procedures (SOP’s), and methods that define and control the manner of production;
(2) Monitoring and control of process parameters and component and device characteristics during production;
(3) Compliance with specified reference standards or codes;
(4) The approval of processes and process equipment; and
(5) Criteria for workmanship which shall be expressed in documented standards or by means of identified and approved representative samples.
(b) Production and process changes. Each manufacturer shall establish and maintain procedures for changes to a specification, method, process, or procedure. Such changes shall be verified or where appropriate validated according to 820.75, before implementation and these activities shall be documented. Changes shall be approved in accordance with 820.40.
(c) Environmental control. Where environmental conditions could reasonably be expected to have an adverse effect on product quality, the manufacturer shall establish and maintain procedures to adequately control these environmental conditions. Environmental control system(s) shall be periodically inspected to verify that the system, including necessary equipment, is adequate and functioning properly. These activities shall be documented and reviewed.
(d) Personnel. Each manufacturer shall establish and maintain requirements for the health, cleanliness, personal practices, and clothing of personnel if contact between such personnel and product or environment could reasonably be expected to have an adverse effect on product quality. The manufacturer shall ensure that maintenance and other personnel who are required to work temporarily under special environmental conditions are appropriately trained or supervised by a trained individual.
(e) Contamination control. Each manufacturer shall establish and maintain procedures to prevent contamination of equipment or product by substances that could reasonably be expected to have an adverse effect on product quality.
(f) Buildings. Buildings shall be of suitable design and contain sufficient space to perform necessary operations, prevent mixups, and assure orderly handling.
(g) Equipment. Each manufacturer shall ensure that all equipment used in the manufacturing process meets specified requirements and is appropriately designed, constructed, placed, and installed to facilitate maintenance, adjustment, cleaning, and use.
(1) Maintenance schedule. Each manufacturer shall establish and maintain schedules for the adjustment, cleaning, and other maintenance of equipment to ensure that manufacturing specifications are met. Maintenance activities, including the date and individual(s) performing the maintenance activities, shall be documented.
(2) Inspection. Each manufacturer shall conduct periodic inspections in accordance with established procedures to ensure adherence to applicable equipment maintenance schedules. The inspections, including the date and individual(s) conducting the inspections, shall be documented.
(3) Adjustment. Each manufacturer shall ensure that any inherent limitations or allowable tolerances are visibly posted on or near equipment requiring periodic adjustments or are readily available to personnel performing these adjustments.
(h) Manufacturing material. Where a manufacturing material could reasonably be expected to have an adverse effect on product quality, the manufacturer shall establish and maintain procedures for the use and removal of such manufacturing material to ensure that it is removed or limited to an amount that does not adversely affect the device’s quality. The removal or reduction of such manufacturing material shall be documented.
(i) Automated processes. When computers or automated data processing systems are used as part of production or the quality system, the manufacturer shall validate computer software for its intended use according to an established protocol. All software changes shall be validated before approval and issuance. These validation activities and results shall be documented.