4 min reading time

As originally asked by Jose Luis Maglione.

Hello Group!!
I would like to share with you a doubt about how to assess how often a calibration is needed and what international Regulation/Standards should be tested for devices like signal generators, Patient simulator, etc. Where they are not in contact with patients, but they are used for testing Patient monitor accuracy, for instace.
As you know, these devices do not have mechanical pieces (in fact, most of the time they are built with microprocessors, resistors, capacitors, etc ) , so that, is there a formula / guideline to apply when I want to set how often a device has to be calibrated. What Standards are applicable?
Thank you all of you in advance!

---

Dan O’Leary
President at Ombu Enterprises, LLC
The question of calibration interval depends on the stability of the equipment. You need to specify the precision and accuracy of the measurement. The calibration should verify that the equipment continues to meet the precision and accuracy specifications. The calibration interval is set so the calibration occurs before the equipment goes out of specification.

There seems to be two parts to this discussion. The first is setting the calibration recall intervals for existing equipment in use, while the second seems to be setting the manufacturer’s recommended recall interval for a newly designed piece of equipment.

For the first case, most locations follow the manufacturer’s recommendations. There are also statistical methods, time series, to estimate drift and predict when the instrument will go out of calibration. There is a simple method called stair-step. Start with, say, a 12 month recall interval. If the equipment is in specification at the next calibration, extend the interval by one month. If the equipment is out of calibration (or nearly so) shorten the interval by one month.

For the second case, you do not need real time tests. You could do a stability analysis on the circuit. You could do a time series analysis, taking readings frequently to predict drift. You could combine the time series with an acceleration factor, such as temperature, to predict drift.

If you are setting calibration recall intervals for a new medical device, don’t forget to include it in your design validation under 820.30.

Regards,
Dan

Anand Paropkari
Head, Business Quality, Agile Business Transformation Coach, SAFe SPC
Unlike mechanical instruments electronic instrumnets may not have wear & tear. So any measurement error that may arise, is due to external factors affecting characteristics of circuit (like hi/lo temp, humidity, vibrations etc) or specs of circuit components (accuracy of value, leakege currents of semiconductors, failure rate etc.). Thus 1 year is a very much agreed upon frequency for general applications. Special cases where very high accuracy of measurement is needed e.g. +/- 0.0001% the calibration freq. would be more.

Stephen Glassic
Available: Biomedical Equipment Technician, Field Service Engineer, Electronic/Electromechanical Technician
Also Jose, I’m sure there are other guidelines used in the process of bringing a piece of test equipment to market and it would depend on what market you are targeting. You might be able to get more input from Metrology oriented groups here on LimkedIn.
Here is one such group. Metrology and Test Measurement; [https://www.linkedin.com/groups?home=&gid=1792150&trk=anet_ug_hm|leo://plh/https%3A*3*3www%2Elinkedin%2Ecom*3groups%3Fhome%3D%26gid%3D1792150%26trk%3Danet_ug_hm/ztA_?_t=tracking_disc]

Here is another group. Biomedical Test & Measurement Equipment; [https://www.linkedin.com/groups?home=&gid=2464195&trk=anet_ug_hm|leo://plh/https%3A*3*3www%2Elinkedin%2Ecom*3groups%3Fhome%3D%26gid%3D2464195%26trk%3Danet_ug_hm/d8eO?_t=tracking_disc]

Stephen Glassic
Available: Biomedical Equipment Technician, Field Service Engineer, Electronic/Electromechanical Technician
Jose, One year is the what is mandated by the JCAHO (Joint Commission on the Accreditation of Hospital Organizations) and the FDA here in the US. I believe it is also considered the golden rule in most other industries here in the US.

Jose Luis Maglione
Sr. Project Manager . Clinical Trials – Client Services – BioMedical Engineer.
Thank you All!
It is true, the golden rule is one year. But i thought that there were some statistical test or any other process to support this conclusion. I am building an electronic device like patient simulator but i do not have one year to keep it running and detect if it has a deviation.
So that i tried to found some criteria, formula, model to assess the ageing and any other adverse event on electronic components.
Thank you very much for yours inputs !!

Srinivasan Sundar
Manager at Parkway Pantai Pte Ltd (Mount Elizabeth Novena Hospital)
Thanks forthe comments. Since these Biomedical TEst equipments are microprocessor based ad electronic devices can the intyerval to calibrate them be extended beyond the one year norm (say two years) as the cost is very exhorbitant and Fluke monopolises all these test equipment.

Stephen Glassic
Available: Biomedical Equipment Technician, Field Service Engineer, Electronic/Electromechanical Technician
In the US the general guideline is to have the biomedical equipment test equipment calibrated yearly or any time there is suspicion that it is out of calibration. We would generally send the equipment to the OEM for calibration. There are also other 3rd party calibration labs that perform calibrations. The NIST (National Institute of Standards and Technology) is an agency of the US Government that oversees weights and measurements standards within the US therefore the equipment must be calibrated using equipment that has been calibrated and traceable back to the standard which is maintained by the NIST. The equipment used to perform the calibration must have a 10% higher accuracy than the equipment being calibrated. I am not sure how often that equipment has to be calibrated and exactly how the NIST insures the accuracy of their standards.

I know there is also the International Bureau of Weights and Measures. I don’t know if every individual country has their own counterpart.

Obviously some equipment, especially any high power measurement equipment, would be more critical and some may even require a shorter calibration interval but here in the US the general guideline is one year. This goes for ECG simulators, Defibrillator Analyzers, Pulmonary Function analyzers, ESU and NIBP testers Ultrasound Watt Meters and so on. With some equipment such as signal generators of various types it is probably rare that they would be out of calibration but they are still required to be checked. Back in the old days when a lot of discrete components and manual calibration potentiometers were used it was a different story and calibration intervals were shorter as well. Nowadays a lot of test equipment as well as medical equipment goes through its own self calibration routine but periodic PM and Calibration testing is currently still required.