The use of mobile communication devices is rapidly increasing. Over 27 million mobile handsets are used in the UK alone . Most hospitals have a mobile phone policy, which is based on the concern that wireless communications could interfere with medical equipment, as well as the precept that phones should not be used in certain areas due to their intrinsic ‘annoyance’ factor.
Some medical equipment used in the intensive care unit (ICU) environment communicate via wireless technology and these devices are screened to ensure they produce low levels of radio frequency emissions. The risk of radio frequency interference caused by devices not designed for the ICU environment is of some concern. This risk is two-fold: firstly, there is an increase in medical devices used on ICU and secondly, there is an increase in the number of electromagnetic interference (EMI) sources in the environment . Staff communicate by mobile phone or radio handset and data are transferred between staff by wireless technology. This is becoming common with the increasing availability of Bluetooth® compatible applications. As health care professionals find these devices becoming more integral to their working lives, the assumption that they should not be used in a hospital environment is being challenged .
The objective of this study was to look at a number of ICU ventilators whose performance had not previously been tested for EMI. The study aimed to generate EMI produced by radio handsets, mobile phones and Bluetooth enabled devices, and to monitor the effect these devices had on both displayed and actual ventilator performance.
The research was conducted in an unused bay in the critical care area. A matrix of 10 cm × 10 cm squares was marked on the floor to provide a reliable method of measuring distances during the test. A wheeled stand with 10 cm markings provided measurement in the vertical axis. The EMI device itself was held in a cradle, suspended from the stand (Fig. 1).
We tested a number of transport and ICU ventilators exposed to EMI from different radio frequency-generating communication devices (Table 1). The ventilator under test was placed in the middle of the 100 cm matrix grid. The centre of the device on its normal stand was designated as the arbitrary zero point in the vertical plane, and the markings on the drip stand were adjusted to show 100 cm above the device and 100 cm below. All the cables from the ventilator being tested exited to the rear of the device, and were held in a straight line approximately 40 cm above the ground.
All ventilators were connected to a VT-1 Ventilator tester (Bio-Tek Instruments, Winooski, Vermont, USA). The VT-1 was programmed for ambient conditions to simulate a healthy adult lung (compliance 0.05 L cmH2O−1, atmospheric pressure 760 mmHg, relative humidity 50%, temperature 25°C). The VT-1 was programmed to measure the following parameters: respiratory rate, inspiratory: expiratory ratio, tidal volume, minute volume, inspiratory time, inspiratory hold, expiratory time, expiratory hold, cycle time, peak airway pressure, peak lung pressure, end expiratory pressure, mean airway pressure, inspiratory flow and expiratory flow.
Each ventilator was set to provide a standard output (volume assist mode, tidal volume 400 mL, rate 15, inspiratory: expiratory ratio 1: 2, FiO2 0.21, positive end-expiratory pressure (PEEP) 5 cmH2O). The VISION ventilator was set in the pressure assist/control mode; peak pressure 20 cmH2O, PEEP 5 cmH2O, FiO2 0.21, inspiratory time 1 s, rate 15 min−1. Prior to testing, each ventilator underwent a maintenance check to ensure it was functioning correctly.
We ensured the VT-1 was not affected by EMI from the communication devices by placing it in the centre of the matrix, and connecting it to a Drager Oxylog (Dragerwërk AG, Lübeck, Germany). The Drager Oxylog is a completely mechanical ventilator, so is not affected by EMI. The Oxylog was placed 3 m away from the source of the EMI, where the field strength would be negligible and unable to interfere with ventilator function. No change in performance of the VT-1 or measurements was observed.
All the radio devices being tested were programmed to transmit and were placed in contact with each ventilator with special attention being paid to areas thought likely to be at risk such as ports, cable exits and switches. The protocol used during the experiment broadly followed the American National Standard Recommended Practice for On-Site, Ad Hoc Testing ANSI C63 . The radio device being tested was moved in towards the device. [Ventilator performance was measured with the device 100, 80, 60, 40, 30, 20 and 10 cm and directly in contact with the ventilator.] This was repeated in the six axes (Fig. 2). The ventilator display was observed for changes in performance values or display errors.
At each point, the radio device being tested was made to operate in a high-power-output mode. The mobile phones were called using a landline located in the same bay and allowed to ring for 5 s. The radio handset transmitted for 5 s on an unused band. The Bluetooth® devices were made to transmit a large file to the partner device, located 5 m away in the same bay.
At each point, the wireless device was rotated through 360° during peak power output and the devices were allowed to ‘rock’ within the holding cradle. If a positive result occurred, the test was repeated to ensure the results were replicable (Fig. 3).
All data were analysed using Microsoft Excel. Graphs were produced detailing normal ventilator performance, as well as ventilator performance when exposed to EMI. P-values were calculated using the t-test if any variation was observed.
Four of the five ventilators tested showed an observed error when subjected to EMI (Tables 2 and 3).
The most marked effect was that of the Puritan Bennett 840. At a distance of 1 m, the Simoco 8020 caused the ventilator to stop ventilating completely, and required a full reboot with extended self test to return to normal function. This whole procedure took around 45 min. An attempt was made to replicate this effect with a different Puritan Bennett 840. This ventilator did not demonstrate the same effect. However, at a distance of 40 cm the graphical display unit on the front of the ventilator did appear to show a prolonged inspiratory hold but no alteration in ventilator performance was measured by the ventilator tester.
Effects on other ventilators included display reset, with the ventilator restoring normal display function within 2 s, and low-power/low-pressure alarms. Both these alarms were cleared by pressing alarm reset, and normal function was restored. All these effects were temporary and function did not alter after the transmitting device was removed.
The most important findings of this study are that:
- a hand held walkie-talkie within 1 m of a Puritan Bennett 840 ventilator can render it inoperable;
- mobile phone handsets may cause alarms and errors with ICU ventilators;
- generally, alarm and error codes caused by EMI have no effect on ventilator performance;
- Bluetooth® appears to have no effect on the ventilators tested.
The effect of EMI on a ventilator is a function of the distance and terrain between the devices, the shielding of the device, and the type of EMI being produced by the transmitting device. The Medical Devices Agency made a number of recommendations regarding EMI in their 1997 publication: walkie-talkie handsets use should be restricted, and serious consideration should be given to alternatives; mobile phones should be used only at a distance of 2 m or greater from treatment areas where sensitive devices are used, or at patients' bedsides where electromedical devices are present .
We found security and porters' radios pose a more serious risk than mobile phones, and as such these staff need to be aware of the risk and radio handsets should be turned off before entering clinical areas where susceptible equipment is in operation.
The results in our study suggest that mobile phones only present a risk to ventilators when held in close proximity. Although this risk may appear slight, it should be remembered that visitors to ICU patients may be unaware of the risk of EMI and are likely to be in close proximity to the ventilator when at the bedside. A mobile phone in a pocket or handbag will be held at the optimum height for causing interference. Even with prominent signs reminding people to turn off their phone many people may instead turn their phone to silent mode, which removes the audible ring tone but does nothing to reduce the risk of EMI. In intensive care environments, relatives may be distressed and forget to turn off their phones. Therefore adequate shielding of ICU ventilators is essential to reduce the risk of interference and potential harm to patients.
Although mobile phone use should be regulated in certain environments, evidence and common sense suggest that blanket bans are unnecessary. Patients often feel bored and isolated during their time in hospital. They are unable to easily contact, friends, family and colleagues. Relatives may need to urgently contact other people. This may be particularly true in intensive care environments, where the majority of ventilators are deployed. The Medicines and Healthcare products Regulatory Agency (MHRA) has recently issued a supplement to the 1997 guideline suggesting the development of designated areas for mobile phone use in non-clinical areas including waiting and visitors rooms.
Hospital-wide policies banning mobile phone use may inhibit effective delivery of care. For example, increased use of portable communication and data transfer technology by health care teams is being advocated by the ‘Hospital at Night’ project in the UK. The guidelines of the Intensive Care Society for the transport of critically ill adult patients state that during transport of the critically ill, there should be ‘[a] mobile telephone to enable communication with referring/receiving hospital (Compatibility with medical equipment to be ensured)’ . Our study specifically tested a new generation of transport ventilators and demonstrated that these ventilators can be affected by mobile phone EMI although these effects are transient and do not alter the performance of ventilation.
Bluetooth® appeared to have no effect on ventilator function or output. This was expected, as Bluetooth® communicates at both low power and high frequency, both of which make it an unlikely candidate for EMI. Very little investigation of EMI by Bluetooth® has been reported, and that which has been done did not look at ventilator performance, nor did it look at the ventilators we tested in this study. Further research is now needed to examine the possibility of Bluetooth® communication being used to replace cables in an intensive care environment.
This is one of the few reports of EMI to have attempted to quantify ventilator function during interference. This is important as although a number of false alarms and errors were produced, there was often no change in ventilator performance. However, it is important to realize that an alarm or error code may cause interruption or alteration to patient management by giving misinformation to the health care professional caring for the patient. Reassuringly, alteration in ventilator performance which was not detected by the ventilator itself was not detected by the ventilator tester.
A number of these ventilators had not been tested before, and those that had been tested had not been subjected to EMI from the devices we used. Shaw and colleagues reported mechanical failure of the Puritan Bennett 840 following exposure to EMI in laboratory conditions in a shielded room and computer control of the power output from the phone . Our experiment aimed to provide in situ measurements, which would relate to the risk during clinical practice and revealed similar results. We believe that these studies complement each other, as previous studies have suggested that EMI with medical devices which occurs in clinical environments may not be detected in controlled experiments as additional sources of radiation may be present in the clinical environment, justifying the rationale of in situ testing .
This series of experiments had limitations. Although five commonly used ventilators were tested, other machines were not. A restricted number of communications devices were tested and we did not test other portable telephones connected to landlines via digital enhanced cordless telecommunication (DECT) which are used on some ICUs.
An attempt was made to replicate all positive results which were produced in this study. In every instance, an effect was observed. However, the effect was not always the same as the one originally recorded. For example, a different display alarm was observed. This could be for a number of reasons. Although every effort was made to ensure that exactly the same ventilator was used throughout the experiments, it was not always possible to do this. We tested two Puritan Bennett 840C ventilators and found different responses. Different conduiting of cables within different ventilators may affect their susceptibility to EMI. Some models may have had more shielding inserted during manufacture. Secondly, the amount of EMI produced by a device varies. It is possible that different amounts of EMI were generated during different experiments, and this would explain why ventilators were affected in seemingly different ways. We did not measure the power output of the communication device during the experiment. The Nokia 6230 has a specific absorption rate of 0.59 W kg−1 and the Nokia 7210 has a specific absorption rate 0.63 W kg−1. Both phones used the Orange Network which transmits at a frequency of 1750-2000 MHz. The Simoco radio handset transmits anywhere between 68 and 470 MHz with a power output of anything between 1 and 5 W.
In conclusion, radio handsets produce significant amounts of EMI that can seriously affect ventilator performance. Mobile telephones can alter ventilator function, including false alarm messages and error codes. However, these problems do not appear to affect ventilator performance. Bluetooth communication appears to have no effect on ventilator function or performance. All electrical devices used in the clinical environment should be adequately shielded to minimize the risk from mobile communication devices which may have an important role to play in the rapid transfer of clinical data to and from the bedside.
1. Klein AA, Djaiani GN. Mobile phones in the hospital - past, present and future. Anaesthesia
2. IEEE. Radiofrequency interference with medical devices. IEEE Engineering in Medicine and Biology Magazine
3. Myerson SG, Mitchell ARJ. Mobile phones in hospitals. BMJ
4. American National Standards Institute Recommended Practice for an On-Site, Ad Hoc Test Method for Estimating Radiated Electromagnetic
Immunity of Medical Devices to Specific Radio-Frequency Transmitters
. ANSI C63.18-1997.
5. Medical Devices Agency. Device Bulletin; Electromagnetic
Compatibility of Medical Devices with Mobile Communications. MDA DB 9702. March 1997.
6. The Intensive Care Society. Guidelines for the Transport of the Critically Ill Patient
. London, 2002.
7. Shaw CI, Kacmarek RM et al.
Cellular phone interference with the operation of mechanical ventilators. Crit Care Med
8. Turcotte J, Witters D. A practical technique for assessing electromagnetic
interference in the clinical setting: ad hoc testing. Biomed Instrum Technol