Welcome to chapter nine, circulation, breathing and the wider body. This chapter has two purposes that pull in different directions. The first is comprehension: enough physiology to follow a clinical conversation without stopping it every third sentence. The second is restraint. This is where a technologist is most likely to build something dangerous without touching a single sharp edge, because a monitor display is easy to make, and it can look authoritative while showing a stale value in the wrong units.
You need to recognise when a seemingly small design change touches an essential body function. Can you make the strap a bit tighter is a different question if the strap crosses a superficial artery. And you need restraint. The gap between displaying a number and supporting a clinical decision is the most important boundary in this book, and this chapter is where it is easiest to cross without noticing.
The heart is two pumps side by side. The right heart pumps to the lungs; the left heart pumps to the body. Memorise the sequence, because you will use it constantly. Blood returns from the body in the venae cavae to the right atrium, through the tricuspid valve to the right ventricle, through the pulmonary valve into the pulmonary arteries and the lungs. It comes back in the pulmonary veins to the left atrium, through the mitral valve to the left ventricle, and out through the aortic valve into the aorta and the body.
Arteries and veins are defined by the direction of flow, not by how much oxygen they carry. An artery carries blood away from the heart; a vein carries it toward the heart. That is the whole definition. So the pulmonary arteries carry relatively oxygen-poor blood from the right ventricle to the lungs, and the pulmonary veins carry oxygen-rich blood back. The shortcut that arteries are red and veins are blue works for the body circuit and is wrong for the lung circuit, which is exactly why it is the standard exam question.
Arteries have thick elastic and muscular walls, and the aorta's elastic recoil smooths the pulsing flow, a Windkessel or air-chamber effect, which is a genuine engineering analogy. Arterioles are mostly smooth muscle and are where resistance is controlled. Capillaries are a single cell thick, five to ten micrometres across, and they are where exchange happens. Veins are thin and distensible, return blood at low pressure and act as a reservoir holding most of the blood volume, with valves and the squeezing of skeletal muscle helping return from the limbs. Lymphatics return tissue fluid, and their failure causes lymphoedema, relevant to any compressive design.
In a typical resting adult the heart beats sixty to eighty times a minute, ejecting about seventy millilitres each time, so cardiac output is about five litres a minute, and the whole blood volume of about five litres goes round roughly once a minute. Cardiac output equals heart rate times stroke volume is the single most useful equation in cardiovascular physiology. Output can rise by beating faster or ejecting more per beat. Both have limits; at very high rates there is too little time to fill, so stroke volume falls. And the heart sounds: lub is the valves between atria and ventricles closing, dub is the outflow valves closing.
The heart generates its own rhythm. The sinoatrial node fires spontaneously as the pacemaker, the signal spreads across the atria, is delayed briefly at the atrioventricular node so the ventricles can fill, then races down the bundle of His and the Purkinje fibres to the ventricles. On an ECG the P wave is the atria depolarising, the QRS complex is the ventricles depolarising, and the T wave is the ventricles recovering. The ECG records electrical activity, not the pumping, and the two can come apart. Knowing what the letters represent is the right depth here. Interpreting a trace or setting alarm thresholds is a clinical skill, and nothing in this chapter moves you toward it.
Now the engineering bridge. The circulation is a pressurised flow network, and the analogies are genuinely productive, provided you know where they break.
Pressure drop equals flow times resistance, the Ohm's law analogue. For steady laminar flow through a rigid tube, Poiseuille's law says flow goes with the fourth power of radius. Halve a vessel's radius and flow falls to a sixteenth. That is why arterioles, wrapped in smooth muscle, are the body's flow-control valves. It is also directly why compression matters: anything pressing on tissue is a flow-control device whether you meant it to be or not. But the law fails in the body. Blood is a suspension, not a Newtonian fluid. Vessels are compliant and branch. Flow is pulsatile and can be turbulent, which is what makes murmurs. And the network is actively regulated. Use the fourth power to understand sensitivity, never to calculate a flow in a person.
Blood pressure is reported as systolic over diastolic, for example a hundred and twenty over eighty. Mean arterial pressure is roughly diastolic plus a third of the difference, because the heart spends about twice as long relaxing as contracting. Pressure falls through the arterioles, and in the capillaries, where exchange happens, it is only about thirty-five falling to fifteen. The lung circuit runs at about a fifth of the body circuit's pressure with the same flow, so it must have much lower resistance, which is why the right ventricle is thinner-walled. And notice: an external interface pressure of twenty kilopascals is a hundred and fifty millimetres of mercury, several times capillary pressure. The numbers a maker generates casually are large compared with the microcirculation.
Now breathing.
The airway branches about twenty-three times from the trachea to the alveoli, of which there are three to five hundred million, with an exchange area of about seventy square metres across a barrier less than a micrometre thick. The conducting zone, from nose to the smallest bronchioles, moves and conditions air but does no exchange. Its volume, about a hundred and fifty millilitres, is dead space. That explains why shallow rapid breathing is inefficient. With a five hundred millilitre breath, three hundred and fifty is useful. Halve the breath and double the rate, and minute volume is unchanged, but useful ventilation per breath collapses from three hundred and fifty to a hundred.
Quiet inspiration is active. The diaphragm contracts and flattens, the ribs swing up and out, chest volume increases, pressure in the lungs falls below atmospheric, and air flows in. That is Boyle's law: the body changes the volume and the pressure follows. Quiet expiration is mostly passive elastic recoil. The lungs are not muscles and do not inflate themselves; they follow the chest wall because the thin pleural space couples them to it. Let air into that space, a pneumothorax, and the lung collapses even though nothing is wrong with it. That is also why a balloon in a jar is a good demonstration of ventilation and says nothing at all about gas exchange.
Three mechanical properties govern breathing. Compliance is volume change per unit pressure change: stiff lungs from fibrosis need more pressure for the same volume, while lungs that have lost their recoil, as in emphysema, make breathing out hard. Airway resistance is dominated by the radius of the airways, so narrowing in asthma raises it steeply: the fourth power again. And surfactant. By Laplace's law a smaller bubble has a higher internal pressure, so small alveoli should empty into big ones and collapse. Surfactant lowers surface tension more in smaller alveoli, equalising the pressures. Its absence in premature infants causes respiratory distress. An elegant fix for a real instability.
A quiet breath is about five hundred millilitres, twelve to twenty times a minute, giving about six litres a minute. Vital capacity, the most you can breathe out after the deepest breath in, is around four and a half litres in a typical young man, and about a litre and a bit always stays behind as residual volume. These vary substantially with sex, height, age and fitness, and spirometry reports results against predicted values built from those variables, which is itself an interesting measurement-design problem.
Ventilation is air movement. Diffusion across the barrier is gas exchange. Perfusion is blood flow. All three must be adequate and matched: blood flowing past unventilated alveoli is a shunt; ventilated alveoli with no blood flow are dead space. Oxygen diffuses from alveolar air at about a hundred and four millimetres of mercury into blood arriving at forty. Most oxygen is carried on haemoglobin, and the saturation curve is S-shaped. Its flat top keeps saturation high in the lungs even if alveolar oxygen dips; its steep middle unloads a lot of oxygen in the tissues for a small fall. It shifts right, unloading more readily, with heat, carbon dioxide and acidity, exactly the conditions in working tissue.
Each vital sign measures something specific, and each is routinely over-interpreted. Heart rate is not cardiac output. Blood pressure is not flow; pressure can be maintained while flow is inadequate. Respiratory rate tells you nothing about how much air each breath moves. Oxygen saturation is an estimate of how much haemoglobin is carrying oxygen; it is not ventilation, not carbon dioxide, and not oxygen delivery. Temperature depends on where it was measured, so the site must be stated. And a pain score is not a physical measurement at all; it is a report, and it belongs to the person.
The pulse oximeter is the single most instructive measurement device for a maker: a simple principle, enormous clinical reliance, well-documented limits. An emitter shines red light at about six hundred and sixty nanometres and infrared at about nine hundred and forty through a finger. Oxygenated and deoxygenated haemoglobin absorb the two differently. Only the pulsing component is used, to separate arterial blood from the steady background of tissue, bone and veins. The ratio of the two signals is mapped to a saturation by an empirical calibration curve built from healthy volunteer studies. So SpO2 is an estimate from a calibration curve, and motion, cold fingers, ambient light, nail polish and carbon monoxide all degrade it.
Pulse oximeters have been shown to overestimate oxygen saturation more often in people with darker skin pigmentation, so true low oxygen can be missed. Regulators have issued safety communications, and this is an active area of research. Three lessons generalise far beyond oximetry. A device calibrated on a non-representative population performs worse on the people it was not calibrated on; it is the same failure as a machine learning model trained on skewed data. An estimate presented as a measurement conceals its own uncertainty. And if you build an interface, you inherit responsibility for how that uncertainty is communicated. Displaying a bare number is a design decision with clinical consequences.
This is the mental model to take away. First, the phenomenon, what is physically happening in the body. Second, the sensor, a transduction with noise, drift, artefact, calibration, latency and failure modes. Third, the displayed value: selected, filtered, rounded, averaged, unit-converted, and possibly stale. Fourth, interpretation: what it means for this person, in this context, now. And fifth, the decision. Every arrow is a place where information is lost or distorted. A maker can build layers two and three. Layers four and five are clinical acts, and they are not yours.
Software can look authoritative while its inputs are wrong. A plausible display can conceal a value whose input stopped minutes ago, the wrong units, rounding that hides a trend, or a sensor that has come off. So for any physiological display you ever write, even a classroom mock-up: make simulation status unmissable; show units beside every number; show the age of each value and change its appearance when it goes stale; distinguish no data from zero, because conflating them has killed people; make invalid input explicit rather than clamping it; make paused look different from live; and handle the connection dropping or the device sleeping. And never label a value normal because it fell inside a range you chose.
The chapter closes Part one's tour with a map of the other systems, so you recognise when a project has quietly become about one of them. The kidneys regulate fluid, electrolytes and acidity; urine output is a vital measurement, and urinary catheters are a leading source of hospital infection. The gut breaks down and absorbs food, and enteral feeding connectors are deliberately incompatible with intravenous ones because misconnections killed people. Hormones regulate over minutes to years, and diabetes technology lives here. The immune and lymphatic systems defend and return fluid. Pregnancy changes almost every system. Chapter ten covers kidneys, gut and hormones properly. Two cross-cutting ideas: body water and its compartments, and thermoregulation. Anything that covers skin is a thermal decision.
This is the densest device chapter in medicine. Cardiology, respiratory medicine, anaesthesia and critical care account for a large share of all medical equipment, and almost all of it is life supporting, patient connected, or both. Read it as a map, not an invitation. Alarm design and alarm fatigue are the dominant human factors problem in critical care. The training column is where you can contribute: CPR manikins with feedback, AED trainers, full patient simulators, auscultation trainers, rhythm simulators, test lungs, and moulage, which is casting, fabrics and dressings. And if you write signal processing software, open datasets like PhysioNet are exactly where to learn.
The mechanical test lung is an unusually good project. It is a bellows with a settable spring for compliance and an interchangeable orifice for resistance. It connects a ventilator to something that behaves plausibly, it teaches this chapter's two mechanical properties honestly, it touches no patient, and it needs no electronics. It is a genuinely mixed-workshop build: a sewn or moulded bellows, turned or drilled orifice plates of known diameter, sheet-metal or machined end plates, and springs chosen by measured rate. Add a pressure sensor and a microcontroller later if a display is wanted, with the electronics outside the gas path. And it makes the physics tangible: narrow the orifice and resistance climbs steeply, as the fourth-power law predicts.
An educator wants a fake patient monitor for a teaching scenario. You can build it in an afternoon; here is how to build it responsibly. First establish the learning objective, which the educator defines, and whether the numbers must behave in a clinically meaningful way; if so, the educator specifies the behaviour. Build it with a permanent, undismissable simulation-only banner, obviously fictional patient data, units beside every value, a visible difference between live, paused and stale, and a version number. Do not connect it to any real device, use real data, invent alarms or thresholds, or deploy it anywhere a passer-by could read it as real. Note how little of that is programming.
For the workshop: draw both circuits from memory, and check you have not connected the aorta to the right ventricle. Draw the air path and compare useful ventilation at five hundred and at two hundred and fifty millilitres. Calculate mean arterial pressure for a hundred and eighteen over seventy-six, cardiac output for seventy-two beats of sixty-eight millilitres, and the flow drop when a vessel's radius falls by a fifth, which is about sixty percent. If software is your thing, build the mock monitor, deliberately make a value persist after its input stops, then fix it. And trace a worn object across a superficial artery or nerve, then write the question you would ask a clinician.
The limits. You know what the numbers represent; what they mean for a person, and what to do about it, are clinical acts. Alarm limits and thresholds are clinical and regulatory decisions with a large human factors literature behind them. Breathing circuits, infusion, airway devices and monitoring are specialised regulated projects, not scaled-up classroom mock-ups. A model that shows inflation shows ventilation; exchange needs separate representation and labelling. And the physiological numbers here let you check an order of magnitude, not assess anybody. Next, kidneys, gut and hormones: the internal regulators.