Pushing on the chest

Lesson 2 ended with your hands on the chest. That's where most people's knowledge stops, at "push hard and fast", and it's where a frightened person's pushing tends to go wrong: too shallow, too slow, too many pauses, or hands in the wrong place. This lesson gives every number behind the push and the reason the guidelines give for it, where they give one. It also answers the worry that comes with the first push, which is what you are doing to the person underneath you.

Your hands are the pump

In cardiac arrest the heart has stopped pumping. The British Heart Foundation puts it in one line: "When a person is in cardiac arrest their heart stops pumping blood, which means their brain gets no oxygen."⁷ Chest compressions are the stand-in, and the BHF's summary of what they do is just as short: "CPR keeps the heart pumping blood around the body until emergency help arrives."⁷

The AHA's 2025 guidelines add the part that makes every detail in this lesson matter, and it's about stopping.

The numbers, and the fact that everyone agrees on them

On the adult numbers, the US and UK guidelines say the same thing, and so do the ERC and ILCOR; the American Red Cross's public page gives the same rate and the same 2-inch floor. This is one of the places where the guidance is settled.

AHA 2025 Part 7, sections 7.1 and 7.3;² RCUK 2025 Adult basic life support.¹ COR and LOE are the AHA's class of recommendation and level of evidence. Lesson 4 explains them; for now, COR 1 means "should" and COR 2a "is reasonable".

The AHA grades the depth COR 1 but the rate only 2a. That isn't a disagreement with RCUK, which gives the same rate; it's the AHA saying how strongly it recommends each, on the same level of evidence (B-NR). And it is candid that all of these targets rest on thinner evidence than their precision suggests: "Although there are numerous retrospective observational studies, there is a paucity of prospective studies or randomized trials specifically examining CPR quality targets."² It adds that "evidence suggests interactions between CPR components (eg, rate and depth) confound studying them in isolation."² So a finding about rate can't be cleanly separated from depth, or the reverse. Read each number's reason with that in mind: the numbers work as a set, and mostly on observational data.

Where your hands go

Kneel beside the person. Put the heel of one hand on the lower half of the breastbone (the sternum), in the middle of the chest, and the heel of the other hand on top. RCUK 2025 adds: "Interlock your fingers of the hands to ensure that pressure is not applied over the ribs."¹

The AHA's supporting text says why the hands go low on the breastbone. "Radiographic studies show the left ventricle is typically located inferior to the internipple line, corresponding with the lower half of the sternum."² The left ventricle is the heart's main pumping chamber, and "inferior to the internipple line" means below an imaginary line between the nipples. There's a lower limit too: "hand placement inferior to the internipple line may result in compression over the xiphoid, which may be less effective."² The xiphoid is the small tip at the bottom of the breastbone.

How firm is this? The AHA's evidence on exact hand position is three small studies with 57 patients between them. Two found better physiological readings, such as blood pressure, over the lower third of the sternum than over the middle, one found no difference, and none found a difference in outcomes such as survival.² So the instruction is COR 1 on limited data (C-LD): sound anatomy, not a trial.

Then the rest of you. RCUK 2025: "Keep your arms straight." "Position your shoulders vertically above the person’s chest."¹ The American Red Cross: "Shoulders directly over hands; elbows locked".⁸ None of the guidelines this course read gives a physical reason for the posture in words, so this lesson won't invent one. What the AHA does report is that kneeling beside the person gave deeper compressions than standing, in several manikin trials.²

Clothes. You don't need to undress someone to push on their chest. The BHF: "Unfastening a thick coat will help but you do not need to remove any thin clothing to do CPR. Clothing does however need to be removed to attach a defibrillator."⁷ RCUK 2025 allows more when you need it: "If you are unable to adequately visualise the sternum due to clothing, it is reasonable to displace or remove such garments so you can identify the correct anatomic landmark."¹ Lesson 5 covers clothing and the defibrillator pads, including bras.

How deep, and why there's a ceiling

At least 5 cm (2 in), and not more than 6 cm (2.4 in), in both countries.¹² The BHF puts the everyday version plainly: "Usually, it’s best to press harder rather than softer," because "If you press too softly then your compressions won’t be effective."⁷

The floor comes from a 2020 ILCOR review of 12 studies and more than 10,700 patients, which the AHA summarises: "Several studies found improved survival to hospital discharge when compression depth was at least 5 cm, compared to less than 4 cm."²

The ceiling runs against instinct. If deep is good, why not deeper? The AHA's reason is survival: "Observational research has suggested reduced survival with chest compressions of excessive depth (greater than 6cm)."² The AHA doesn't give injury as a reason for the ceiling, but it's the worry most people have, and one study bears on it. Hellevuo and colleagues, in 2013, studied 170 adults who had a cardiac arrest in one Finnish hospital, with compression depth measured by the defibrillator and injuries found at post-mortem, on CT scans or on chest X-rays.¹³

Bars drawn to scale from zero by this course, from the three percentages in the study's abstract.¹³

Look at the first two bars. Pushing to the recommended 5 to 6 cm carried no more injury than pushing too shallow, 27% against 28%. The rise comes above 6 cm. But read the grey line under the title too. Across all 170 patients, the gap between the groups was not quite big enough to rule out chance by the usual test: the p value was 0.06, and by convention a result counts as statistically significant only below 0.05. The clear link was in men; in women the study found none.¹³ The authors' own conclusion runs both ways: "there is an increased risk of complications with deeper compressions", but "the injuries were by and large not fatal."¹³

So the fair reading is that the 6 cm ceiling is a sensible caution, backed by a survival signal the AHA cites and an injury signal in one hospital study, and neither is proof. It is not a reason to hold back. Too shallow is the mistake the BHF's "harder rather than softer" is aimed at.

How fast

100 to 120 compressions a minute, in both countries (AHA 2025, COR 2a, LOE B-NR; RCUK 2025).¹² The BHF says "2 per second" and offers a tune: "Try pushing to the beat of Stayin’ Alive by the Bee Gees."⁷

The AHA's supporting text reports three observational studies, of more than 13,700 patients, that "suggested improved survival to hospital discharge with compression rates of 100 to 119/min, compared with lower or higher rates."² So in the data there is a penalty on both sides. The guideline doesn't explain why faster than 120 would do worse, beyond its warning that rate and depth are tangled together in the studies, so this lesson does not guess.

The AHA is also open about a finding that points the other way. In one study the heart restarted more often at 121 to 140 a minute, but "the writing group placed higher value on survival compared with ROSC when making this recommendation."² A heart restarting and a person leaving hospital alive are different outcomes, and the writers chose the second. And "One randomized trial found no difference in survival between chest compression rates of 100 and 120/min."² Anywhere in the range is fine. Outside it, on either side, is the mistake.

Letting the chest come back up, and not pausing

After each push, let the chest rise all the way before the next. RCUK 2025 says so in the table above; the American Red Cross puts it as "Allow chest to return to normal position after each compression".⁸ The mistake this guards against is leaning: resting some of your weight on your hands between pushes, so the chest never quite comes back up. The AHA's evidence comes from animal work: "Porcine study data suggest decreased coronary perfusion with rescuers leaning on the chest."² Porcine means pigs; in pig studies, leaning lowered the same pressure that pauses lower. The guideline text does not go further into the physiology, so neither does this lesson. Your hands stay in place, and each push ends with your weight off.

As for pauses, RCUK wants compressions "with as few interruptions as possible";¹ the American Red Cross, for trained rescuers who stop to give breaths, says "Minimize interruptions to chest compressions to less than 10 seconds."⁸ Lesson 4 covers breaths and lesson 5 the defibrillator, the two planned pauses.

Where they're lying

The floor, and knee level

New in the AHA's 2025 guidelines, and one of its top ten messages: "In adult cardiac arrest, rescuers should perform chest compressions with the patient’s torso at approximately the level of the rescuer’s knees whenever possible" (COR 1, LOE B-NR).² The evidence is from manikin studies: kneeling beside the manikin gave deeper compressions than standing, and two studies found better depth when the manikin was no more than 10 cm below the rescuer's knee.² For a person on the floor, that means kneel beside them. The AHA also says, at COR 1, that "resuscitation should generally be conducted where the person is found, as long as high-quality CPR can be administered safely and effectively."²

The bed: the one real difference

This is where the two countries' guidance, and one public page, part ways.

Everyone agrees a firm surface is better: "A firm surface improves the likelihood of adequate chest compression depth," as the AHA says.² ILCOR, whose evidence reviews both draw on, says "We suggest performing chest compressions on a firm surface when possible", and, for patients in hospital, "we suggest against moving a patient from a bed to floor" (ILCOR 2024, weak recommendation, very low certainty).⁵ The difference is what a bystander should do about the bed.

What makes moving costly is in the AHA's own supporting text: "Importantly, delay in initiation of chest compressions is associated with worse outcomes. In a study of telephone-assisted CPR during OHCA, delays to CPR due to patient repositioning occurred in 41% of cases, most commonly due to physical limitations of the rescuer."² OHCA is out-of-hospital cardiac arrest.

The St John page was reviewed in April 2025, before RCUK's October guidelines. In the UK, RCUK 2025 is what your call handler and your next class will work from. So: in the UK, start on the bed and push deeper (RCUK 2025). In the US, get the person onto the floor if you can do it quickly and safely, but not at the cost of a delay (AHA 2025, COR 2a). On a bed, the AHA's manikin evidence still applies: kneeling in the bed beside the person gave deeper compressions than standing beside it.² The AHA adds that kneeling "needs to be weighed against rescuer safety and stability".²

Not on their side. The AHA says effective CPR cannot be done with someone in the recovery position; compressions go with the person on their back.³ If someone has been put on their side, roll them onto their back first.

Larger bodies

The technique doesn't change. AHA 2025: "CPR for adults with obesity who are in cardiac arrest should be provided by using the same techniques as for the nonobese patient" (COR 1, LOE C-LD).² RCUK's 2025 special circumstances guidelines agree that obese patients get standard resuscitation with no change to basic life support.⁴

What changes is the effort. The AHA's supporting text explains that compressions can be harder "due to increased chest wall thickness, requiring greater force to achieve the recommended depth," and that "The physical demands of performing CPR on an adult patient with obesity can lead to faster rescuer fatigue, highlighting the need for close monitoring to switch rescuers."² So push harder, and swap sooner if someone can take over. On a bed the two problems stack: "During OHCA in which an adult patient with obesity is found in a bed, the time required to move the patient must be carefully considered to avoid delays in chest compression initiation."²

Pregnancy

A pregnant woman who collapses gets CPR as anyone else would, and RCUK 2025 tells rescuers to "Consider pregnancy in any collapsed woman of childbearing age."⁴ Later in pregnancy the guidelines add one step, and the reason is mechanical. The AHA: "The growing uterus can cause position-dependent aortocaval compression when the pregnant person is supine."³ In plain words, when she lies on her back the womb presses on the large blood vessels behind it, including the vein that returns blood to the heart from the lower body, and the AHA says that means less blood reaching the heart and less pumped out, which "must be overcome to enable adequate blood flow."³

AHA 2025 Part 10, pregnancy section;³ RCUK 2025 Special circumstances.⁴

Manual left uterine displacement means pushing the womb over towards her left side by hand while she stays flat on her back. Both prefer it to tilting her whole body, and the AHA gives the evidence: in simulation trials, compressions were better on a manikin lying flat than on one tilted sideways.³

Here's the honest limit. Both texts are written for resuscitation teams, the AHA shows the hand position in a figure this course didn't see, and neither describes a version for a bystander. So this course's reading, for a lay rescuer, is: start compressions at once, exactly as for anyone, and let nothing delay them; tell the call handler she's pregnant; and if a second person is free, let the call handler say whether and how that person should do the displacement.

Swapping, and when to stop

CPR is hard physical work, and the AHA's evidence says it degrades in a way that's easy to miss. With two or more rescuers, "it is reasonable to switch chest compressors approximately every 2 min" (AHA 2025, COR 2a, LOE B-R).² The supporting text gives the reason: "Chest compression depth begins to decrease after 90 to 120 seconds of CPR, although compression rates do not decrease significantly over that time window."²

Read the second half of that again. The part of your technique you can hear, the rhythm, holds up. The part that's hard to judge from above, the depth, is what goes. So "I'm fine, my rhythm's good" can be true about the rhythm and still miss what matters. That is this course's reading of the AHA's two findings, not a line from it, but it is why the swap goes by the clock. ILCOR puts it more simply: "We recommend monitoring for fatigue in all rescuers performing CPR (good practice statement)."⁵

St John Ambulance says to swap "every one to two minutes, but try to minimise interruptions to chest compressions".⁶ A manikin trial the AHA cites found no difference in quality between swapping every minute and every two.² Either works, as long as the changeover is quick: new hands go on as the old ones come off, and the count carries on.

When to stop. St John and the BHF give nearly the same list.⁶⁷ Keep going until emergency help arrives and takes over; the person shows signs of life, such as breathing normally or coming round; you're too exhausted to carry on; or, on St John's list, a defibrillator is ready to use, which is where lesson 5 picks up. Being too exhausted is on both lists; stopping then isn't failing. With a second person there, swapping keeps it from coming to that.

Two of you, on the hallway floor

Put it together in one case, at home, where most arrests outside hospital happen.²

You're visiting your sister in Sacramento. Her husband, who is in his fifties, collapses in the hallway. You check he does not respond and is not breathing normally, and your sister calls 911 on speaker while you start, as lesson 2 set out. He is on his back on a wooden floor.

1. Position. You kneel beside his chest. On the floor his torso is below your knees, the position the AHA recommends (COR 1), and you do not move him: this is where he was found, and good CPR is possible here.²
2. Clothes. He's in a T-shirt, and it stays on (BHF). If it hid the breastbone, you'd push it up (RCUK 2025).⁷¹
3. Hands. Heel of one hand on the lower half of his breastbone, the other on top (AHA 2025; RCUK 2025), fingers interlocked (RCUK 2025).¹²
4. Posture. Arms straight, shoulders over your hands (RCUK 2025).¹
5. Depth and rate. At least 2 inches (5 cm) and no more than 2.4 inches (6 cm), at 100 to 120 a minute (AHA 2025; RCUK 2025), with the call-taker counting.¹² You aren't measuring. You're pushing hard, and a class manikin is where you'll learn what 5 cm feels like.
6. Recoil. Hands stay on the chest; weight comes off after every push.¹²
7. Two minutes. Your sister kneels on his other side. At about two minutes you swap (AHA 2025, COR 2a).² You say "swap on thirty", count thirty pushes, and she takes over as your hands come off; the thirty-count is this course's suggestion for a clean changeover, not a guideline's. You swap at two minutes even if you feel fine, because depth drops after 90 to 120 seconds while the rhythm holds;² if either of you is flagging sooner, swap sooner.
8. Four minutes. You swap back. Between turns you answer the call-taker and unlock the front door.

Two things are missing from that list on purpose. Nobody stopped to check whether it was working, and nobody stopped to move him.

The bed at night

This case is harder, because the answer depends on the country, on a public page written before RCUK's 2025 guideline, and on who the rescuer is.

It's two in the morning in Norwich. A woman in her late seventies makes a strange noise in her sleep, then goes still. Her husband is eighty, has arthritis in both knees, and did a first aid course through his bowls club in 2019. He remembers being taught to move them onto the floor first. She is in a soft double bed. He is alone, with his phone on the bedside table.

Now check it against the other two. The AHA prefers a firm surface only "when feasible and does not delay chest compressions."² For an eighty-year-old with bad knees, alone, lifting an adult off a double bed is neither quick nor feasible. St John says to move her only "if it is safe for you to do so",⁶ and for him it is not clearly safe, or quick.

So all three roads end in the same place for this rescuer. The difference between the guidelines is real, but it matters mostly for a strong rescuer with help. Beyond kneeling on the bed beside her (AHA's manikin evidence, which it says has to be weighed against the rescuer's stability, and his knees are bad) and pushing deeper (RCUK), neither text describes technique on a bed, so this course doesn't. The call handler is counting with him.

What CPR does to the person under your hands

Lesson 1 promised numbers for this fear. It comes in two forms.

If they're in arrest: ribs break, and that's expected

The largest review this course read, by Van Wijck and colleagues in 2024, pooled 74 studies of 16,629 people who had CPR for a cardiac arrest not caused by injury. It found some CPR injury in 60%, and rib fractures, the commonest injury, in 55%.¹² That's about one in two.

So a crack under your hands is common, not a sign of failure. The BHF: "Sometimes minor injuries can occur, such as a cracked rib, but you should continue to give CPR to help save their life," and "A broken rib or other injury can be treated."⁷

Three limits on the 55%. The review took in CPR by machine as well as by hand, and found machines carried about 1.36 times the risk of injury,¹² so it isn't a figure for hands alone. The abstract, which is all this course read, doesn't say how long CPR lasted in the included studies or who gave the CPR by hand, so it isn't a measure of what a minute of bystander CPR does. And it doesn't show that depth was wrong: in the chart above, injuries were as common at the correct depth as at too shallow a one.

If they weren't in arrest: injury is uncommon

The sharper fear is pushing on someone whose heart was beating all along: the person who had fainted, or was very drunk, or was having a seizure. Lesson 2 told you not to wait until you're sure. This is the evidence that makes that advice safe.

• White and colleagues, 2010, followed people in King County, Washington, given compressions on a dispatcher's instructions who turned out not to be in arrest. Of 247 with full follow-up, 12% had discomfort, 2% (6 people) had an injury likely or possibly caused by CPR, 2% (5) had a fracture, and none had an injury to an internal organ.⁹
• Haley and colleagues, 2011, reviewed six years of one US county's ambulance records. Of 72 people given bystander CPR whom the ambulance crew found not to be in arrest (the authors call them "likely not in cardiac arrest"), and who could be checked, one (1.4%) had an injury recorded as possibly caused by CPR, and it was rhabdomyolysis, a breakdown of muscle, not a fracture.¹⁰ More than half of the 72 were admitted to intensive care, so these were not healthy people pushed on by mistake.
• Williamson and colleagues, 2024, reviewed five observational studies, 1031 people in all. Under 1% (9 people) had injuries, including rib fractures and internal bleeding; 2% had symptoms such as chest pain; no deaths were attributed to compressions. The authors rated the evidence very low certainty.¹¹

That's where lesson 1's "about 2% or fewer" comes from. Haley's team called what they studied "Short-duration bystander CPR",¹⁰ and this course's reading is that a short spell of compressions, stopped when the person wakes, is a different event from a full resuscitation, which fits the gap between 2% and one in two.

The guidelines act on it. RCUK 2025, in its key points: "The risk of harm from CPR is low. Rescuers should not be concerned that they will cause serious injury if the person is not in cardiac arrest."¹ The AHA's supporting text says that giving compressions to someone unconscious but not in arrest "has not been associated with serious harm, as demonstrated in several observational studies."² Very low certainty, which lesson 4 explains, means the exact figure could move as studies are added; every study read here found injury uncommon.

And what it does to you

RCUK 2025: "The risk of infection to rescuers performing CPR is low," and "The risk of physical injury to the rescuer from performing CPR is low."¹ The AHA says gloves or a mask are reasonable "When immediately available" (COR 2a, LOE C-LD), and warns that "the process of donning PPE may delay the initiation of CPR."² Use them if they're in your hand; do not go looking first.

What people get wrong

"Breaking ribs means you did it wrong." Rib fractures are found in about 55% of people who have CPR (Van Wijck and colleagues, 2024), and the BHF says to keep going.¹²⁷

"You'll hurt someone who's only fainted." In the studies of that case, injuries were about 2% or fewer, and RCUK 2025 tells lay people not to be concerned about it.⁹¹⁰¹¹¹

"Gentle is safer." Depths under 4 cm were linked to worse survival; the BHF's advice is "harder rather than softer".²⁷

"You need to take their clothes off." Not for compressions (BHF); move clothing only if you can't find the breastbone (RCUK 2025). The pads are another matter, in lesson 5.⁷¹

"Always move them to the floor." Not in the UK (RCUK 2025), and in the US only when it doesn't delay compressions (AHA 2025). A UK public page says to move them if it's safe.¹²⁶

"You can do CPR with them on their side." Not effectively (AHA 2025). Roll them onto their back.³

"I'll know when I'm getting tired." Depth falls after 90 to 120 seconds while the rate holds, which is why the AHA puts the swap on a two-minute clock.²

Practice

Connections

Lesson 1 raised the fear of hurting someone; this lesson gave the studies behind "about 2% or fewer". Lesson 2 gave you "push hard and fast in the centre of the chest"; this lesson made each word exact. Lesson 4 covers the pauses you have been told to keep short, rescue breaths, and what the grades in the parentheses mean. Lesson 5 is the defibrillator that interrupts your two-minute cycles, and lesson 6 gives the depths for children and babies, set as a fraction of the chest.

From earlier on the Core, Sleep taught you to compare how strongly guidelines back a recommendation: here the AHA grades depth COR 1 and rate 2a for the same set of numbers. And Memory taught you to ask who was studied, which is what separates the 55% (people in arrest) from the 2% (people who weren't).

Go deeper

• RCUK, 2025 Adult basic life support guidelines, the section "High-quality chest compressions". Free, one screen long, read here in full: the whole of this lesson's UK instruction in thirteen sentences.
• AHA, Part 7: Adult Basic Life Support, 2025, section 7. Free; read here in full from an Internet Archive capture. The supporting text under each recommendation is where the AHA says how thin its evidence is and why it recommends what it does anyway.
• Van Wijck and colleagues, "Rib fractures and other injuries after cardiopulmonary resuscitation for non-traumatic cardiac arrest", 2024, open access. This course read the abstract only. It is the source of the 55%, and the place to start if you want to know how the included studies found the injuries.

Sources

1. Resuscitation Council UK, 2025 Resuscitation Guidelines: Adult basic life support, 27 October 2025 . Read level: full, direct, re-read 2026-09-24. Every RCUK 2025 compression instruction here, the key point on harm, and rescuer safety. It adapts the ERC's 2025 Adult BLS chapter, cited here through it.
2. M. E. Kleinman, J. E. Buick, N. Huber and colleagues, "Part 7: Adult Basic Life Support: 2025 American Heart Association Guidelines for CPR and Emergency Cardiovascular Care", Circulation 152 (suppl 2), 2025 . Read level: full, from the Internet Archive capture 20251026225858; sections 6.2 and 7.1 to 7.3, the obesity section and the epidemiology preamble re-read 2026-09-24. The AHA's recommendations with COR and LOE, and the supporting text quoted.
3. D. Cao, A. M. Arens, S. L. Chow and colleagues, "Part 10: Adult and Pediatric Special Circumstances of Resuscitation: 2025 AHA Guidelines", Circulation 152 (suppl 2), 2025. **Read level: partial, from the Internet Archive capture 20251027153841**; pregnancy section re-read in full 2026-09-24 (Figure 6 not seen); the sentence on CPR in the recovery position is from the gas embolism section. Uterine displacement, aortocaval compression, and CPR in the recovery position.
4. Resuscitation Council UK, 2025 Resuscitation Guidelines: Special circumstances, 27 October 2025. **Read level: partial, direct** (obese patients; pregnancy).
5. J. E. Bray, M. A. Smyth, G. D. Perkins and colleagues for the ILCOR BLS Task Force, "Basic Life Support: 2025 International Consensus on CPR and ECC Science With Treatment Recommendations", Circulation, 2025, doi 10.1161/CIR.0000000000001364 . Read level: full, direct, from the accepted manuscript on ilcor.org; the quotations used here were then checked word for word against the typeset text. Firm surface, bed to floor (2024 recommendations kept in the 2025 summary), and fatigue monitoring (2023).
6. St John Ambulance, How to do CPR, clinically reviewed 28 April 2025 . Read level: full, direct, re-read 2026-09-24.
7. British Heart Foundation, How to do CPR, undated . Read level: full, direct, re-read 2026-09-24.
8. American Red Cross, CPR Steps, undated . Read level: full, direct, re-read 2026-09-24.
9. L. White and colleagues, "Dispatcher-assisted cardiopulmonary resuscitation: risks for patients not in cardiac arrest", Circulation 121, 2010, pp. 91 to 97 . Read level: abstract only.
10. K. B. Haley and colleagues, "The frequency and consequences of cardiopulmonary resuscitation performed by bystanders on patients who are not in cardiac arrest", Prehospital Emergency Care 15, 2011, pp. 282 to 287 . Read level: abstract only.
11. F. Williamson and colleagues, "Does delivering chest compressions to patients who are not in cardiac arrest cause unintentional injury? A systematic review", Resuscitation Plus 20, 2024, 100828 . Read level: abstract only.
12. S. F. M. Van Wijck and colleagues, "Rib fractures and other injuries after cardiopulmonary resuscitation for non-traumatic cardiac arrest: a systematic review and meta-analysis", European Journal of Trauma and Emergency Surgery 50, 2024, pp. 1331 to 1346. **Read level: abstract only.**
13. H. Hellevuo and colleagues, "Deeper chest compression: more complications for cardiac arrest patients?", Resuscitation 84, 2013, pp. 760 to 765 . Read level: abstract only. The chart uses the abstract's three percentages.
14. R. Greif and colleagues, "Education, Implementation, and Teams: 2025 ILCOR CoSTR", Circulation 152 (suppl 1), 2025 . Read level: Internet Archive capture 20260430180745, recommendation sections.
15. British Heart Foundation, RevivR, and Resuscitation Council UK, Lifesaver. **Read level: both pages full, direct**; neither course was taken.
16. M. Riggs, R. Franklin and L. Saylany, "Associations between CPR knowledge, self-efficacy, training history and willingness to perform CPR and CPR psychomotor skills: a systematic review", Resuscitation 138, 2019, pp. 259 to 272 . Read level: abstract only.
17. US Army Corps of Engineers, Illinois Waterway Visitor Center, CPR drawing, via Wikimedia Commons, public domain . Licence and posture checked 2026-09-24.