Statistical Process Control (SPC) is a method of monitoring a manufacturing process using statistical tools, most often control charts, to distinguish normal random variation from variation caused by an assignable fault. In medical device production, SPC provides objective evidence that a validated process stays in control between formal revalidations.
What is Statistical Process Control (SPC)?
Statistical Process Control (SPC) treats a production process as a source of measurable variation. Samples are taken at set intervals, a critical characteristic is measured (weld strength, solder paste height, molding shot weight), and results are plotted against statistically derived control limits. Those limits come from the process, not the drawing tolerance.
Two kinds of variation matter. Common cause variation is the inherent noise of a stable process. Special cause variation is a signal: a tool wearing out, a new resin lot, an operator deviating from the work instruction. SPC separates the two. In the device lifecycle, it begins at process characterization, is formalized at validation, and runs through routine production.
Why Statistical Process Control (SPC) matters in medical device development
A device process that drifts does not announce itself. It produces conforming units until it does not, and by then, several lots may be in distribution. SPC shortens the gap between a process change and its detection.
Under the FDA Quality Management System Regulation (QMSR), 21 CFR Part 820, effective February 2, 2026, ISO 13485:2016 is incorporated by reference. Clause 8.2.5 requires monitoring and measurement of processes, and clause 8.4 requires documented procedures for analyzing data on process characteristics and trends. The legacy statistical techniques section, 820.250, no longer stands as separate regulatory text, but the expectation for statistically valid methods carries through ISO 13485. EU MDR 2017/745 imposes an equivalent burden via Annex IX.
Auditors read SPC records as evidence that a validated state is maintained. Where a process is validated under ISO 13485 clause 7.5.6 because output cannot be verified downstream (sterile barrier sealing, adhesive cure, laser welding), SPC data is often the only continuous proof that validation holds.
How Statistical Process Control (SPC) works
SPC is a loop, not a report. A typical implementation runs as follows:
- Select the characteristic. Work from the process FMEA and design outputs. Chart what drives patient risk, not what is easy to measure.
- Qualify the measurement system. Run a gauge R&R study first. A chart built on a noisy gauge measures noise.
- Establish control limits. Collect data from a stable period, typically 20 to 25 subgroups, and set limits at plus or minus three sigma. Control limits describe the voice of the process. Specification limits describe the voice of the customer. Never plot one as the other.
- Choose the chart type. Variable data uses X-bar and R or X-bar and S charts, individuals, and moving range for low-volume work. Attribute data uses p, np, c, or u charts. ISO 7870-2 covers Shewhart chart construction.
- Apply detection rules. A point outside the limits is the obvious signal. Runs, trends, and hugging patterns catch drift sooner.
- Act on signals. Out-of-control points route to a documented investigation and to CAPA when systemic.
- Assess capability. Cp, Cpk, Pp, and Ppk quantify how much tolerance the process consumes and mean nothing until the process is in control.
Capability and control answer different questions. A process can be stable and still produce nonconforming parts.
Common challenges and best practices
The most frequent failure is charting everything. Teams instrument dozens of parameters, generate charts nobody reads, and lose the signal. Chart the few characteristics traceable to a hazard in the ISO 14971:2019 risk file.
The second failure is putting specification limits on a control chart, which hides drift: a process can wander badly and still sit inside a wide tolerance.
Also avoid setting limits from an unstable baseline, recalculating limits after every process tweak until the chart tracks drift instead of detecting it, and treating an out-of-control point as a paperwork exercise. Reaction plans belong in the work instruction, written before the signal appears.
Good practice looks unglamorous. Operators own the chart at the machine. Subgroup size and sampling frequency are justified, not inherited. Limits are recalculated only under change control, and chart data feeds the management review inputs of ISO 13485 clause 5.6.
How SJML helps with Statistical Process Control (SPC)
SJML applies process monitoring across its manufacturing lines, including medical PCBA on high-speed SMT with SPI, AOI, and X-ray inspection, precision metal and medical-grade plastics, and cleanroom assembly. Process validation (IQ/OQ/PQ), PFMEA, and PPAP are part of new product introduction, and SAP-integrated MES supports traceability of process and product data. The QARA team connects trend data to CAPA, ISO 14971 risk files, and regulatory sustenance across FDA, EU MDR, and MDSAP frameworks.
Talk to SJML’s manufacturing team →
Frequently asked questions
Process validation establishes documented evidence that a process consistently produces conforming output, usually through IQ, OQ, and PQ. Statistical Process Control (SPC) runs afterward, monitoring the process during routine production and signaling when output moves away from the state validated. Validation is an event. SPC is continuous surveillance.
The FDA does not mandate control charts by name. Under the QMSR, 21 CFR Part 820, ISO 13485:2016 is incorporated by reference, and clauses 8.2.5 and 8.4 require monitoring of process characteristics and analysis of trends using appropriate methods. Statistical techniques, once chosen, must be documented and justified.
Cpk estimates capability from within-subgroup variation and describes what the process could achieve if it stayed in control. Ppk uses total observed variation and reflects actual performance. A large gap between the two usually means the process is unstable. Both indices assume a normal distribution.
Use an individual and moving range chart (I-MR) when parts are produced one at a time or in small batches, common for Class III implants and capital equipment subassemblies. For attribute data such as pass or fail visual inspection, a p-chart handles varying lot sizes. Chart choice follows data type, not volume.
No. SPC reduces reliance on inspection but does not eliminate it. Under ISO 13485 clause 8.2.6, monitoring and measurement of product remains a requirement, and release criteria must be met. Mature SPC on a capable process can support a reduced sampling plan if that reduction is justified and documented.
Related terms
- Process Validation
- Design Verification
- Process FMEA (PFMEA)
- CAPA
- ISO 13485
