Module 49: Health & Medicine

The human body, what goes wrong, and how to make good decisions about your own health

Part A — The Human Body: A Working Map

What this module covers

Medicine is perhaps the most consequential body of knowledge an ordinary person can acquire. It decides how you interpret symptoms, evaluate treatments, navigate healthcare systems, and ultimately understand your own mortality. This module is organised in twelve parts: body architecture (A), disease mechanisms (B), diagnostics (C), treatments (D), the evidence base (E), public health (F), healthcare systems (G), nutrition (H), mental health (I), emergency first aid (J), rational decision-making (K), and the future of medicine (L). You do not need a science background. You need the willingness to think precisely about uncertain information.

Organ systems — interactive map

Click a system to explore it

Cardiovascular
🫁
Respiratory
🦴
Digestive
🧠
Nervous
🧪
Endocrine
🛡
Immune
🫘
Renal
🦴
Musculoskeletal

Cardiovascular System

The cardiovascular system is a closed-loop pressure circuit: the heart (a 300-gram muscular pump) ejects roughly 70 mL of blood per beat, around 5 litres per minute at rest and up to 25 litres during intense exercise. Arteries carry oxygenated blood outward under high pressure; veins return deoxygenated blood under low pressure. The heart's two pumps work in parallel: the right side sends blood to the lungs (pulmonary circuit) and the left side sends it to the rest of the body (systemic circuit). The coronary arteries, which feed the heart muscle itself, are only 2–3 mm in diameter — which is why a small plaque can trigger a massive heart attack. Blood pressure is expressed as systolic/diastolic mmHg; 120/80 is textbook normal. Above 140/90 consistently constitutes hypertension, which silently damages artery walls for decades before causing strokes or kidney failure.

The cell — life's basic unit

The human body contains roughly 37 trillion cells, each a self-contained chemical factory. Every cell contains a nucleus (DNA blueprint), mitochondria (ATP generators, roughly 2,000 per cell), a cell membrane (selectively permeable lipid bilayer), and specialised organelles. A red blood cell lasts 120 days; a gut epithelial cell lasts 3–5 days; a neuron may last your entire lifetime. The key insight: disease almost always starts at the cellular level — a mutation, a pathogen entering, a receptor malfunctioning.

Homeostasis — the body's self-regulation

Core body temperature must stay within 36.1–37.2°C. Blood pH must stay within 7.35–7.45. Blood glucose must remain roughly 4–7 mmol/L. The body achieves this through negative feedback loops: a deviation triggers a corrective response that reverses the deviation. The thermostat analogy is imperfect — biological systems use hormone cascades, neural reflexes, and organ-organ communication simultaneously, which is why a single-point failure (e.g., insulin deficiency in type 1 diabetes) destabilises multiple systems at once.

Part B — How Disease Works

Disease taxonomy — click to explore

Infectious Disease

Infectious diseases are caused by pathogens: bacteria (living cells that replicate independently, killed by antibiotics), viruses (genetic material in a protein coat that hijack host cell machinery, requiring antivirals or vaccines), fungi (eukaryotes, treatable with antifungals), and parasites (complex organisms like Plasmodium falciparum, which kills 600,000 people yearly via malaria). The key concept is transmission: each pathogen has a characteristic R0 (basic reproduction number) — the number of new infections one case generates in a fully susceptible population. Measles has an R0 of 12–18. Influenza is around 2–3. SARS-CoV-2 (original strain) was 2–3; Omicron was 8–15. R0 above 1 means exponential growth; herd immunity requires a fraction (1 − 1/R0) of the population to be immune.

Global causes of death — top conditions (millions/year, WHO 2019 data)

Source: WHO Global Health Estimates 2020. Figures are pre-COVID 2019 baselines. Cardiovascular disease alone accounts for 32% of all global deaths.

The immune response — how your body fights infection

Innate and adaptive immunity — a schematic

Pathogen enters body Innate Immunity Immediate (0–96 hr) Neutrophils, macrophages Inflammation, fever Non-specific Antigen Presentation Dendritic cells process and present fragments to T cells Adaptive Immunity Days 4–14 T cells kill infected cells B cells produce antibodies Highly specific (lock-and-key) Immunological Memory Decades of protection Vaccines skip steps 1–3 and go straight to memory Threat First responders Hand-off Targeted response

The innate immune system responds within hours but cannot distinguish between specific pathogens. The adaptive system takes days but generates exquisitely specific antibodies and cytotoxic T cells, plus memory that makes future responses faster. Vaccines exploit this memory mechanism.

Part C — Diagnostics

Normal vital signs — what each measures

Heart Rate

60–100

beats per minute. Below 60 is bradycardia (normal in athletes); above 100 is tachycardia. Trained endurance athletes often rest at 40–50.

Blood Pressure

120/80

mmHg (systolic/diastolic). Above 140/90 on repeated measures is hypertension. Below 90/60 is hypotension.

Respiratory Rate

12–20

breaths per minute. Often the most ignored vital sign — a rate above 25 in an adult is a serious warning sign of respiratory or metabolic distress.

Temperature

36.1–37.2°C

Oral temperature. Above 38.0°C is fever. Average core temperature has been declining since the 19th century, from 37.0 to around 36.6°C.

O2 Saturation

95–100%

Pulse oximetry. Below 90% requires urgent attention. COPD patients may baseline at 88–92%. SpO2 is not the same as PaO2 (arterial oxygen).

Blood Glucose

4.0–7.8

mmol/L fasting. Above 7.0 fasting or 11.1 random = diabetes threshold. HbA1c over 6.5% confirms 3-month average hyperglycaemia.

Sensitivity, specificity & the base-rate problem — calculator

A test with 99% sensitivity and 99% specificity sounds almost perfect. But when the disease prevalence is low, most positive results are false positives. Enter numbers to see how misleading high-accuracy tests can be.

Press Calculate to see results.

Medical imaging — when each modality is used

X-ray (Radiograph)

X-rays use ionising electromagnetic radiation that passes through soft tissue but is absorbed by dense structures like bone and calcium deposits. The image is a 2D shadow. Dose per chest X-ray is roughly 0.1 mSv — equivalent to about 10 days of natural background radiation. Ideal for: fractures, pneumonia (lung consolidation appears white), pneumothorax (collapsed lung), foreign bodies. Not ideal for: soft tissue, brain, abdomen organs (poor contrast). Speed: 5 minutes from request to image.

Part D — Treatment

Drug development — from discovery to prescription

Average time from molecule identification to approval: 10–15 years. Average cost: $1–2 billion (including failures). Only 1 in ~10,000 candidate molecules reaches clinical use.

Timeline is approximate. Phase III trials alone can enrol 10,000–30,000 participants. The FDA's expedited pathway (Breakthrough Therapy designation) can compress timelines by 30%.

Treatment strength vs. side-effect burden — a spectrum

Approximate relative positioning. Not a clinical guide. Individual drugs within each class vary widely.

Antibiotics — the resistance crisis

Antibiotics are bactericidal or bacteriostatic — they kill bacteria or stop their replication. They work by targeting bacterial-specific structures: the cell wall (penicillins, cephalosporins), protein synthesis (macrolides, tetracyclines), or DNA replication (fluoroquinolones). They have zero effect on viruses. Resistance emerges because bacteria evolve: random mutations that happen to inactivate an antibiotic are selected for when antibiotics are present. The WHO considers antimicrobial resistance one of the greatest threats to global health — 700,000 people die from resistant infections annually today; projections put this at 10 million by 2050.

Vaccines — how immunisation works

Vaccines present the immune system with an antigen (or instructions to make one) without causing disease. The immune system generates antibodies and memory B and T cells. On future exposure, the response is faster and stronger — before you develop symptoms. Types: live-attenuated (MMR, highest efficacy), inactivated (flu, polio Salk), subunit (hepatitis B, pertussis), and mRNA (Pfizer/Moderna COVID-19, the first mRNA vaccines approved for mass use). The mRNA in COVID-19 vaccines degrades within days; it does not enter the nucleus or alter DNA.

Part E — Evidence-Based Medicine

The hierarchy of evidence — from weakest to strongest

The hierarchy is not absolute: a well-designed observational study can be more informative than a poorly designed RCT. But it captures the general principle: the higher up, the more the study design controls for confounding variables and bias.

Absolute vs relative risk — the numbers that get manipulated

Drug manufacturers often report relative risk reduction (RRR) because it sounds more impressive. Absolute risk reduction (ARR) is what matters for individual decisions. Adjust the baseline risk to see how misleading relative framing can be.

Baseline annual risk of event (e.g., heart attack) in control group:

0.5%20%

Your baseline risk: 5%

Publication bias — what gets published vs what gets done

Estimates suggest only 50–60% of clinical trials ever publish results. Positive trials are 2.4x more likely to be published than negative ones (Turner et al., NEJM 2008). This systematically overstates treatment efficacy in the literature. Pre-registration of trials at ClinicalTrials.gov partially addresses this.

Part F — Public Health

What actually determines your health — proportional contribution (estimates)

These proportions (from US County Health Rankings model) are contested but directionally robust. The counterintuitive insight: clinical healthcare — what we typically mean by "medicine" — accounts for only about 20% of health outcomes. Your postcode predicts your lifespan better than your doctor.

Major pandemic timeline — death toll and origin

Bars show approximate years of peak mortality. Death tolls are historical estimates and vary by source. The Black Death (1347–1351) killed an estimated 30–60% of Europe's population — unmatched in proportional terms by any subsequent pandemic.

Part G — Healthcare Systems

The four basic healthcare financing models

Beveridge Model

Named after William Beveridge, architect of the 1948 NHS. Government collects taxes, owns or contracts hospitals, employs or contracts doctors, and provides care free at the point of use. The logic: healthcare is a public good, like roads or defence. Examples: UK (NHS), Spain, Scandinavia. Advantages: universal coverage, no administrative complexity for patients, lower per-capita cost. Disadvantages: waiting times, political vulnerability (funding decisions tied to electoral cycles), potential for rationing. The NHS costs the UK roughly £3,000 per person per year — less than half the US per-capita spend — with comparable or better outcomes on most population health metrics.

Healthcare spending vs outcomes — selected countries

Annual per-capita healthcare spending in USD (2022, OECD). Life expectancy at birth shown alongside.

The US spends more than $12,000 per person per year — more than twice the OECD average — yet ranks last among high-income countries on life expectancy, infant mortality, and chronic disease burden (Commonwealth Fund, 2023).

Part H — Nutrition & Lifestyle Medicine

Daily calorie needs estimator (Mifflin-St Jeor equation)

The Mifflin-St Jeor equation is currently the most validated formula for estimating resting metabolic rate (RMR). Note: this is an estimate; individual variation can be ±15%.

Press Calculate to see your estimated daily needs.

Ultraprocessed foods — the evidence

The NOVA classification defines ultraprocessed foods (UPF) as industrial formulations containing ingredients not used in domestic cooking: emulsifiers, flavour enhancers, stabilisers, hydrogenated oils. UPFs now account for 57% of UK caloric intake and 60% in the US. The NOVA system was controversial when proposed by Carlos Monteiro in 2010 — critics argued it ignored nutrient content. But prospective cohort studies have since linked high UPF intake to 32% higher cardiovascular mortality, 12% higher cancer risk, and 53% higher dementia risk (JAMA Internal Medicine, 2023). The mechanism is unclear: it may be the additives, the lack of fibre, the speed of digestion, or the food matrix disruption.

Sleep — the most underrated intervention

Chronic sleep restriction to 6 hours per night produces cognitive deficits equivalent to 24 hours of total sleep deprivation, yet subjects report feeling only "slightly sleepy" — they lose the ability to perceive their own impairment. During sleep, the glymphatic system (brain's waste clearance mechanism) pumps cerebrospinal fluid through brain tissue, flushing amyloid beta and tau — the proteins that aggregate in Alzheimer's disease. Adults need 7–9 hours. No amount of caffeine or adaptation compensates for the biological deficit. Matthew Walker's book "Why We Sleep" (2017) overstates some claims, but the core evidence base — from laboratory studies and population epidemiology — is solid.

What dietary patterns actually do — risk reduction estimates

Mediterranean diet data from PREDIMED trial (7,447 participants, Spain, 2013). Effect sizes are approximate. The PREDIMED trial was partially retracted due to randomisation issues but re-analysis confirmed the main findings. Note: dietary epidemiology is notoriously difficult — all estimates are from observational studies except Mediterranean diet, which has RCT evidence.

Part I — Mental Health

Major mental health conditions — click to explore

Depression (Major Depressive Disorder)

Depression affects 280 million people globally — 5% of adults — making it the leading cause of disability worldwide (WHO, 2023). The DSM-5 requires at least 5 of 9 symptoms for 2+ weeks, one of which must be depressed mood or anhedonia (inability to feel pleasure). The "chemical imbalance" explanation (low serotonin = depression) is a significant oversimplification. The serotonin theory was never well-supported, as a major 2022 umbrella review by Moncrieff et al. (Molecular Psychiatry) confirmed. Antidepressants do work — particularly SSRIs for moderate-severe depression — but likely through mechanisms beyond serotonin reuptake. Effect size vs placebo is around 0.3 (NNT roughly 7–8). Cognitive-behavioural therapy (CBT) has comparable efficacy for mild-moderate depression and better durability after treatment ends.

Global mental health burden — disability-adjusted life years (DALYs, millions)

DALYs combine years of life lost to premature death with years lived with disability. Mental disorders collectively account for around 13% of global DALYs, yet receive only 2% of health budgets in most low-income countries (WHO Mental Health Atlas, 2020).

Part J — First Aid & Emergency Medicine

The primary survey — ABCDE in any emergency

Before doing anything else in any medical emergency, assess in order: Airway, Breathing, Circulation, Disability, Exposure. Each step is a "treat as you find" approach — fix a problem before moving on.

A
Airway
Is it open and clear? Look for gurgling, stridor. Head-tilt chin-lift if safe. Jaw thrust if spinal injury suspected.
B
Breathing
Rate, depth, symmetry. SpO2 if available. Give O2 if SpO2 <94%. Look for paradoxical movement (tension pneumothorax).
C
Circulation
Pulse rate, BP, capillary refill (<2 s is normal). Control major bleeding with direct pressure. IV access if trained.
D
Disability
AVPU (Alert/Voice/Pain/Unresponsive) or GCS. Blood glucose. Pupils equal and reactive?
E
Exposure
Expose the body to look for hidden injuries, rashes, burns. Then cover to prevent hypothermia. Preserve dignity.

CPR survival — how compression depth affects outcomes

Survival from out-of-hospital cardiac arrest without bystander CPR is around 5–10%. Optimal CPR (100–120 compressions/min, 5–6 cm depth) can double or triple survival rates. Use this slider to understand the relationship.

No CPRPerfect CPR

Emergency conditions — recognition and response

Stroke — FAST

Stroke occurs when blood supply to part of the brain is cut off (ischaemic, 85% of cases) or when a blood vessel bleeds into brain tissue (haemorrhagic, 15%). Brain cells die at the rate of 1.9 million per minute during a large ischaemic stroke. FAST: Face (drooping on one side?), Arms (one arm drifts down?), Speech (slurred, confused, unable to speak?), Time (call emergency services immediately). Time is brain: thrombolysis (clot-busting drug tPA) must be given within 4.5 hours, and mechanical thrombectomy within 24 hours for large vessel occlusion. Do not give aspirin until haemorrhagic stroke is excluded by CT scan — aspirin worsens a bleed. Do not let the person eat or drink; stroke impairs swallowing and aspiration is fatal.

Part K — Medical Decision-Making

Evaluating health claims — spotting pseudoscience

Red flags: testimonials instead of trials, mechanism too vague to be falsifiable ("boosts immunity"), cherry-picked studies, appeal to nature, appeal to conspiracy (doctors don't want you to know), before/after photos, products sold by the same person making the claims. The single most reliable question: what would change your mind? Pseudoscience producers have no answer. Also check: is the study on humans or petri dishes? Is the effect size clinically meaningful? Was it replicated? Who funded it?

Overdiagnosis — when more medicine harms

Overdiagnosis occurs when a condition is detected that would never have caused symptoms or death in a patient's lifetime. PSA screening for prostate cancer overdiagnoses roughly 20–50% of cases — men undergo surgery, radiation, and hormone therapy for cancers that would never have harmed them. CT colonoscopy finds polyps in 30–40% of adults; most will never progress to cancer. Overdiagnosis is not medical error — it's a structural consequence of improving test sensitivity without improving prognostic discrimination. The Choosing Wisely campaign (launched 2012) has identified over 600 tests and treatments that are overused.

Complementary & alternative medicine — evidence quality spectrum

Approximate positioning based on RCT evidence. "Some evidence" = effect size detectable but modest; "no reliable evidence" = trials fail to show effects beyond placebo or have been retracted.

Acupuncture has the strongest CAM evidence base, particularly for chronic pain and nausea — though effect sizes are small and sham acupuncture often performs as well as "real" acupuncture. Homeopathy is water: no plausible mechanism and no replicated trial evidence above placebo exists. Chiropody for musculoskeletal issues has better evidence than chiropractic for visceral disease claims.

Part L — The Future of Medicine

Genomics and personalised medicine — what it means in practice

Genomics and Personalised Medicine

Whole genome sequencing now costs under $200 (down from $3 billion for the first human genome in 2003). Pharmacogenomics is already clinical reality: variants in the CYP2C19 gene determine whether clopidogrel (an antiplatelet drug) is metabolically active; roughly 30% of people carry loss-of-function variants and receive ineffective therapy. The UK Biobank (500,000 participants with genotype + health records) has identified hundreds of genetic variants associated with disease risk. Polygenic risk scores can predict lifetime risk of type 2 diabetes, coronary artery disease, and breast cancer more accurately than family history alone. The key limitation: most variants explain tiny proportions of variance, and gene-environment interactions are poorly understood. Personalised medicine so far works best for cancer (selecting targeted therapies) and drug dosing — not yet for population-level prevention.

Antimicrobial resistance — projected annual deaths (millions)

Projections from the O'Neill Commission on AMR (2016). Current deaths (~700k/year) are already equivalent to one pandemic per year. Without action, resistant infections could overtake cancer as the leading cause of death by 2050. New antibiotic development has been commercially unattractive because drugs are used briefly and should be reserved — creating a market failure that requires public funding.

Part M · Q&A

If vaccines train the immune system, why does the flu vaccine have such low efficacy some years?

Influenza mutates rapidly via antigenic drift (gradual accumulation of mutations) and antigenic shift (sudden reassortment of genetic segments between strains). The WHO must predict 6–12 months in advance which strains will circulate, using global surveillance data. If the prediction is wrong, the vaccine's antibodies don't match the circulating strain's surface proteins. In years with a good match, flu vaccine efficacy is 40–60%; in mismatched years it can fall below 20%. This is a genuine limitation — not a reason to avoid vaccination, since partial protection, lower hospitalisation rates, and herd effects still confer benefit at the population level.

Why do randomised controlled trials occasionally produce results that turn out to be wrong?

RCTs can fail for several reasons. Publication bias means negative trials often go unreported, so meta-analyses draw on a skewed sample. Industry-funded trials are 2–4x more likely to report favourable results than independently funded ones (a combination of design choices, outcome switching, and selective reporting). Statistical significance at p<0.05 means, by definition, a 5% false positive rate — run enough trials and chance findings appear. Small trials are underpowered and noisy. And surrogate outcomes (e.g., "reduces blood pressure" instead of "reduces strokes") can be misleading if the surrogate doesn't translate to clinical benefit. ALLHAT (2002) showed that cheaper diuretics outperformed newer antihypertensives despite years of contrary trial evidence funded by pharmaceutical companies.

If the US spends far more per capita on healthcare than any other country, why does it have worse outcomes?

The US healthcare paradox has several components. First, spending is wildly unequal: ~50% of healthcare costs are incurred by 5% of patients, while 30 million Americans have no insurance and forgo preventive care. Second, high spending reflects high prices (administrative overhead, provider negotiating power, drug pricing), not high volume of care. A hip replacement costs $30,000–$50,000 in the US and $10,000–$15,000 in Germany. Third, the social determinants of health — gun violence, opioid addiction, poverty, poor diet — are worse in the US than in peer nations and aren't addressed by healthcare spending. Fourth, lack of universal primary care means people arrive at emergency rooms with late-stage preventable disease.

Is the "serotonin hypothesis" of depression wrong, and does that mean antidepressants don't work?

The 2022 umbrella review by Moncrieff et al. found no consistent evidence that people with depression have lower serotonin levels, synthesis capacity, or transporter density than people without depression. This undermines the "chemical imbalance" narrative — but does not mean SSRIs are ineffective. A drug can be effective without the original mechanism hypothesis being correct: aspirin was used for decades before COX inhibition was understood. Large meta-analyses (Cipriani et al., Lancet 2018 — covering 522 trials and 116,000 patients) confirm antidepressants are superior to placebo for moderate-severe depression. The effect size is modest (NNT ~7), which means for every 7 patients treated, one benefits who wouldn't have benefited from placebo. This is comparable to many accepted cardiovascular drugs.

Antibiotics have no effect on viruses — so why do doctors still prescribe them for viral infections?

This is a real problem. Estimates suggest 30–50% of antibiotic prescriptions in primary care settings are unnecessary or inappropriate (CDC, 2020). The drivers are multifactorial: patient expectation (patients who expect antibiotics are more likely to receive them); diagnostic uncertainty (distinguishing bacterial from viral infections is hard without laboratory tests); fear of missing bacterial superinfection; and time pressure. Some primary care systems now use point-of-care CRP tests (a blood inflammation marker) that can rapidly distinguish bacterial from viral illness, halving antibiotic prescriptions without increasing complications. Every unnecessary prescription accelerates resistance, shortens the era of effective antibiotics, and disrupts the gut microbiome.

What is a QALY and why is it controversial for rationing healthcare?

A Quality-Adjusted Life Year (QALY) is one year of perfect health. A year with a serious disability might count as 0.5 QALYs; a year in severe pain, 0.3 QALYs. The UK's NICE uses a threshold of approximately £20,000–30,000 per QALY: treatments more expensive than this are generally not funded on the NHS. The QALY is controversial because: it implicitly values the life-years of disabled people less than those of healthy people; it disadvantages treatments for the elderly (fewer life-years remaining); it embeds subjective quality-of-life assessments derived from population surveys rather than patients' own preferences; and it ignores the value of hope or the severity of conditions for small patient groups. Despite these flaws, QALYs remain the least-bad tool for systematic healthcare rationing in resource-constrained systems. The alternative — informal clinical discretion — introduces bias and inconsistency.

Can you prevent cancer through lifestyle changes, and by how much?

About 40–45% of cancers are attributable to modifiable risk factors (Cancer Research UK, 2022). The five biggest modifiable factors are: tobacco (15% of all cancers), excess body weight (7%), alcohol (4%), UV radiation (4%), and physical inactivity (1%). Not smoking is the single most impactful lifestyle change: smoking causes lung, bladder, kidney, pancreatic, throat, and other cancers. The caveat: "lifestyle causes cancer" messaging can produce harmful blame narratives. Cancer involves both environmental exposures and random DNA replication errors (Tomasetti and Vogelstein estimate roughly 60–70% of cancer-driver mutations are due to replication errors, not external carcinogens). Screening and early detection (smear tests, colonoscopy, mammography) further shift the balance of survivable outcomes but don't prevent cancer at the cellular level.