Question 6 of 6: Pulse Oximetry and the ICU Bedside Monitor

Question 6: Pulse Oximetry and the ICU Bedside Monitor (25 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

(i) Biophysical principles of pulse oximetry

Light at two wavelengths — red (~660 nm) and infrared (~940 nm) — is transmitted through a vascular tissue bed (finger, earlobe). By the Beer-Lambert law, transmitted intensity is $I=I_0e^{-\varepsilon cL}$, where the absorption coefficient $\varepsilon$ depends on wavelength and on which haemoglobin species (oxy- or deoxyhaemoglobin) is present. Deoxyhaemoglobin (Hb) absorbs red light substantially more strongly than oxyhaemoglobin (HbO₂) does — this is exactly why deoxygenated blood looks darker/bluer — while at the infrared wavelength the two species' absorption coefficients are much closer (HbO₂ absorbing slightly more). Because arterial blood pulses with each heartbeat, the light path length through blood (and hence transmitted intensity) has a small pulsatile (AC) component riding on a much larger, steady (DC) component from non-pulsatile tissue, venous blood, bone and skin. Taking the ratio of AC to DC at each wavelength isolates the arterial-blood-only contribution and cancels out the (unknown, wavelength-independent) fixed path length and non-arterial absorbers — the principle of photoplethysmography. The normalised ratio $R=\dfrac{AC_{red}/DC_{red}}{AC_{IR}/DC_{IR}}$ is then compared against an empirically-derived calibration curve (built from co-oximeter reference measurements on volunteers) relating $R$ to arterial oxygen saturation, SpO₂.

(ii) Instrumentation for the ICU bedside monitor

ECG electrodes (a standard 3- or 5-lead set) feed a differential biopotential amplifier as in Question 2/5; the SpO2 channel uses the dual-wavelength LED/photodiode pulse-oximetry front end described in (i) (AC/DC separation per wavelength, ratio-of-ratios computation); respiration is obtained by impedance pneumography, re-using the same ECG electrodes to inject a small, safe, high-frequency (tens of kHz), low-amplitude (well below the microshock threshold) AC current and measuring the resulting transthoracic impedance, which rises and falls as the lungs fill with air during breathing; and blood pressure is obtained from a programmed oscillometric sphygmomanometer — a cuff and pump/valve module that, on a timer (e.g. every 15 min), inflates above systolic pressure and then deflates slowly while a pressure transducer detects the envelope of small cuff-pressure oscillations produced by the underlying arterial pulse; mean arterial pressure is read at the point of maximum oscillation amplitude, and systolic/diastolic pressures are derived from fixed fractions of that maximum (the standard oscillometric algorithm). Each channel's front-end amplifier and filter is matched to its own bandwidth (as in Question 2), and all channels are multiplexed onto a shared ADC and read by a central microcontroller that extracts heart rate (from ECG or the SpO2 pulse), SpO2, respiration rate and NIBP values, drives alarm/limit logic for each, and presents waveforms and numerics together on one bedside display.

(iii) Reducing cable/hose tangling

The block diagram of (ii) can be redesigned by converting each sensing modality into a small, low-power, battery-operated wireless module worn at or near the sensing site, reporting over a short-range body-area-network link (e.g. Bluetooth Low Energy) to a single bedside receiver/hub, rather than each sensor running its own wire back to the central monitor. Concretely: the ECG electrodes and impedance-pneumography measurement are combined into a single adhesive chest patch with its own wireless transmitter (since both already share the same electrodes, this removes one whole cable run at no loss of function); the SpO2 probe becomes a self-contained wireless finger/ear clip; and the NIBP module's pump and pressure-transducer electronics are moved into a small wireless unit strapped near the cuff, so only the short, unavoidable pneumatic cuff tubing remains local to the arm, rather than a long electrical/pneumatic cable running the full distance to the bedside monitor. All modules report to one wireless hub at the bedside, which aggregates the streams into the same microcontroller processing chain as (ii) and drives the same display. This does not, and cannot, eliminate the IV line itself (a fluid-carrying physical tube has no wireless equivalent), but it does substantially reduce the number of electrical leads and their length draped across the bed and crossing the IV line's own path, which is where the entanglement risk the question describes actually arises; smart infusion pumps that report their own status/rate over the same wireless hub (rather than a separate wired data cable to a nurse-call or monitor system) further reduce the remaining electrical cable count without touching the fluid tubing.

(iv) Two noise sources in arterial O2 saturation measurement

(1) Motion artifact — patient movement (common in an agitated or repositioned ICU patient) changes the optical coupling and effective path length through the tissue, generating spurious signal components in the same low-frequency band as the genuine pulsatile signal, which can swamp or mimic it and produce false low-saturation alarms. (2) Ambient light interference — external light sources (particularly other pulsed or flickering sources such as fluorescent or certain surgical/procedure lighting common in an ICU bay) leak into the photodetector and can beat against the LED multiplexing frequency, producing spurious apparent pulsation; this is mitigated by shielding the sensor from ambient light and by synchronous detection referenced precisely to the LED drive timing.