Measurement protocol
Take two readings one minute apart, average the pair, and treat that average as your number for that sitting. Then repeat it morning and evening for seven days in a row, throw away everything you recorded on day one, and average days two through seven. That final average is the only number worth acting on. One reading on its own tells you very little, because pressure moves minute to minute and the first reading of any session is usually the highest one your machine will produce all day. Once you have a real average, put it through the mean arterial pressure calculator to see the number your organs actually feel.
Almost everyone who buys a home monitor gets this part wrong, and it is not their fault. The box tells you how to wrap the cuff. It does not tell you that a reading is a sample, not a measurement, or that the thing you are trying to estimate is an average across days rather than a value at a moment. So people take one reading, see 147, panic, take another, see 132, feel relieved, and learn nothing at all. This article is about turning a pile of scattered numbers into a single number that means something, and about how often you should be generating those numbers in the first place.
Scope note before you scroll: this page owns how many readings, how far apart, how to average and how to log. The separate question of which hour of the day to measure has its own guide on the best time to check blood pressure, and the physical technique of sitting, cuffing and holding still lives in how to get a good blood pressure reading. Read those alongside this one and your numbers will improve more than any change of monitor could manage.
The short answer: two readings, seven days, one average
Here is the whole protocol, stated plainly. Every major guideline group agrees on the shape of it even where they argue about the thresholds.
Sit quietly for five minutes before the first reading
Feet flat, back supported, arm resting at heart level, bladder empty, no talking. Those five minutes are not a formality. They are the difference between measuring your circulation and measuring your last flight of stairs. The full setup is covered in the step by step technique guide, and the arm you use should be the same one every time, chosen using the method in which arm is best to check blood pressure.
Take two readings, one minute apart
Leave the cuff on the arm, loosen it slightly if it feels tight, wait a full sixty seconds, then press start again. Do not stand up between them. Do not check your phone. The gap gives the veins in your forearm time to empty after the first inflation, which is a real physiological reason and not superstition.
Average the pair and write both readings down
Add the two systolic numbers and halve them. Do the same with the diastolic numbers. That pair is your session average. Record the raw readings as well, because the raw numbers are what a doctor will want to see if anything looks odd later.
Do it morning and evening, seven days running
Two sessions a day for seven consecutive days gives you twenty-eight readings. Pick a morning slot and an evening slot and keep them fixed, using the guidance in the best time of day guide so that you are not comparing a pre-breakfast number one day with a post-dinner number the next.
Delete day one, average days two through seven
Day one is contaminated by novelty. A new machine, a new routine and a new reason to be worried all push the numbers up. Guideline groups on both sides of the Atlantic tell you to discard it, which leaves twenty-four readings across six days. Average those, and you have a number you can hand to a doctor without apology.
The single most useful sentence in this article: never make a decision about your health based on one reading. Not one high one, not one reassuringly low one. A reading is one sample from a distribution that swings 20 or 30 mmHg over a normal day, and you are trying to estimate the center of that distribution, not catch its extremes.
Why two readings rather than one? Because the difference between them is systematic rather than random. Averaging a pair removes most of the first-reading effect described further down this page, and the remaining error is small enough to live with. Why not five readings? Because after the second or third the returns collapse, the session becomes a chore, and people quietly stop doing it. Two is the number that survives contact with real life, which is why it made it into the guidelines.
Why seven days rather than three? Day-to-day variation is larger than most people expect. Sleep quality, salt on the previous evening, a rough commute, a cold coming on, all of it moves the number. The European Society of Hypertension will accept a minimum of three days, and says twelve readings is the floor for a usable average, but seven days is the standard because it smooths out the weekday-weekend difference and gives you enough readings that one strange morning cannot drag the mean around.
Turn your average into one perfusion number
Systolic and diastolic are two numbers moving semi-independently. Mean arterial pressure collapses them into the single figure that determines whether your organs are being supplied. Feed your seven-day average into the MAP calculator and read the explanation in how to find mean blood pressure.
Quick reference: how many readings and how often
Print this, screenshot it, tape it to the monitor case. The right answer changes with your situation, and most of the confusion online comes from people applying a newly-diagnosed schedule to a stable, well-controlled situation and driving themselves up the wall doing it.
| Your situation | Readings per session | How often | What you report |
|---|---|---|---|
| Brand new monitor, no diagnosis | 2, one minute apart | Morning and evening for 7 days, then stop | Average of days 2 to 7 |
| Being assessed for high blood pressure | 2, sometimes 3 | Morning and evening for 7 consecutive days before the appointment | Full log plus the six-day average, alongside the range your doctor is working to |
| Medication just started or dose just changed | 2, one minute apart | Once or twice daily for the first two weeks, or as your doctor asks | Daily averages plus the date of the change |
| Stable on medication, numbers at target | 2, one minute apart | A 7-day block once a month, or before each review | Block average and any outliers with a note |
| Elevated but not treated, watching it | 2, one minute apart | A 7-day block every 3 to 6 months | Trend across blocks rather than single numbers |
| Normal readings, no risk factors | 2, one minute apart | Once or twice a year, plus at any routine check | Whatever the block produced |
| Pregnancy or a pregnancy complication | 2, one minute apart | Exactly as your maternity team instructs, often daily | Every reading, and any 140/90 or above reported immediately |
| Feeling unwell right now | 1 to confirm, 1 to repeat after 5 minutes | Once, then act on the result | Both readings and the symptoms you had |
| Reading above 180 systolic or 120 diastolic | Repeat once after 5 minutes of rest | Immediately | Call for help if it stays there, see the danger level guide |
Two things in that table surprise people. The first is that most of the schedules end. A seven-day block is a task with a finish line, not a permanent habit, and for someone with no diagnosis and normal numbers, monitoring twice a year is genuinely enough. The second is that the frequency goes down as your control gets better, not up. Once a treated blood pressure is sitting where it should, checking it every day adds noise to your life and nothing to your medical record.
Sessions, not readings
Think in sessions. A session is five minutes of sitting plus two readings plus one average. Everything in the table above counts sessions, so a schedule of two sessions a day means four inflations, not two.
Blocks, not streaks
A seven-day block is a defined project. Start on a Monday, finish on a Sunday, calculate, close the notebook. People who monitor in blocks stick with it. People who try to monitor forever give up in about eleven days.
Why the first reading is almost always the highest
Sit down, wrap the cuff, press start, and your body responds to the cuff itself. The squeeze on your arm is mildly unpleasant, your attention narrows onto the machine, and the small burst of alertness that follows nudges heart rate and vessel tone upward. Researchers call this the alerting response, and it is the same mechanism behind the more famous white coat effect, just smaller and happening in your own kitchen.
The size of the drop between the first and second reading varies a lot between people. Most published home-monitoring work puts the typical fall at somewhere around 5 mmHg systolic, and in anxious first-time users it can be considerably more. In people who are entirely relaxed about the whole business it can be zero, or the second reading can even come out slightly higher. What almost never happens is a consistent, large rise across the session. If your numbers climb steadily reading after reading, suspect a technique problem rather than physiology, most commonly an arm held below heart level or a cuff that is slowly slipping.
There is a second, purely mechanical reason for the first reading being high. The cuff inflates well above your systolic pressure and traps blood in the forearm veins. If you restart immediately, that congestion is still there and the device has to squeeze harder against a stiffer limb. Sixty seconds of loosening lets the arm drain. This is why the guidelines specify an interval rather than saying take two readings back to back, and it is why the auto-repeat feature on cheaper monitors sometimes produces a nonsense second number.
Discard the first reading, or average it in?
Both conventions are in use and the difference between them is smaller than the internet suggests. Averaging the two readings is the standard home-monitoring instruction in the United States, and it is what your monitor manual almost certainly tells you to do. The European approach leans more toward taking the average of the second and third readings when three are taken, on the grounds that the first is contaminated. Some clinic protocols discard the first reading entirely and average the next two.
What matters is consistency. If you average two readings this month and discard the first of three next month, your trend line is measuring your method rather than your circulation. Pick one convention, write it at the top of your log, and keep it for as long as you keep the log. My own preference is the two-reading average, because it is the simplest thing that works and because the first-reading effect is largely handled by discarding day one of the week anyway.
A useful sanity check: if your first reading of a session is regularly more than 10 mmHg above your second, you are probably not resting properly before you start. Try eight minutes of sitting instead of five and see whether the gap shrinks. It usually does.
This same alerting effect is why a reading taken the instant you walk through a doctor door is close to meaningless, and why a single number from a pharmacy machine should never be treated as a diagnosis. If a supermarket kiosk once told you 158 over 96 and you have been quietly worried ever since, the honest response is to run a proper seven-day block at home and see what the average says. It is very common for that average to land somewhere completely different, and worth reading what a 140/90 reading actually means before you assume the worst.
Average blood pressure means two different things
Search for average blood pressure and you get two completely different questions tangled together. One is arithmetic about your own readings. The other is a population statistic about everyone else. They need different answers and confusing them causes real harm, because people start treating the population figure as a target to hit.
Your personal average
The mean of your own readings across a defined block of time, taken with a consistent method on a consistent arm. This is the number that gets you diagnosed, treated or reassured. It is the subject of the rest of this article.
The population average
The mean reading across a group of people, usually reported by age and sex from national surveys. Useful for public health planning. Almost useless for deciding what your own reading means, because the population it averages includes millions of people with untreated disease.
Here is the trap. Adult blood pressure in the United States averages somewhere in the low 120s systolic over the low-to-mid 70s diastolic, and that sounds comfortingly close to the classic 120/80. But close to half of American adults meet the current definition of high blood pressure, which is roughly 120 million people, so the population average is being dragged upward by an enormous number of people who are not healthy. Matching the national average is not a health goal. Being below it is not an achievement. The category boundaries are what count, and those are set out in the normal range guide and what actually counts as a good number.
The same problem shows up with age and sex breakdowns. Average systolic pressure rises steadily across adult life in almost every population that has been studied, because arteries stiffen. That does not make a rising number normal in the sense of harmless. Men tend to run slightly higher than women through early and middle adulthood, and women catch up and often overtake in the decades after menopause, a pattern explored in the average blood pressure for a woman. Both sexes are compared against the same treatment thresholds, so the difference in averages changes the odds of being diagnosed rather than the number that triggers a diagnosis.
Heart rate gets pulled into this too, usually in the same search. A resting pulse for a healthy adult sits roughly between 60 and 100 beats per minute, with fit people often in the 50s. Your monitor displays it alongside the pressure because the same cuff oscillations reveal it, not because the two numbers are versions of each other. They are not, as the comparison of bpm and blood pressure lays out, and a slow pulse carries no promise of a low pressure.
One more version of the question that deserves a straight answer: newborns. A healthy term newborn runs a systolic in roughly the 60 to 80 range with a diastolic somewhere around 30 to 50, and mean arterial pressure in the first days of life is often compared against the baby gestational age in weeks as a rough floor. None of that can be measured with an adult home monitor. Infant pressures need a neonatal cuff and a device calibrated for them, so if you have a question about a baby, it belongs with a pediatric team rather than with the machine in your bathroom cabinet.
The averaging arithmetic, worked out properly
This is simpler than people fear and easier to get subtly wrong than people expect. The rule in one line: average the systolic column and the diastolic column separately, then write the two results as a pair.
Average systolic = (sum of all systolic readings) / (number of readings)
Average diastolic = (sum of all diastolic readings) / (number of readings)
You do not average the two numbers of a single reading together. You never turn 138/86 into 112. The systolic values form one series and the diastolic values form another, and they get averaged independently even though they came from the same inflations.
Worked example one: a single session
You sit for five minutes and take two readings a minute apart. The machine shows 138/86, then 130/82.
Systolic: (138 + 130) / 2 = 268 / 2 = 134
Diastolic: (86 + 82) / 2 = 168 / 2 = 84
Session average = 134/84 mmHg
Neither raw reading was 134/84. That is the point. The average is an estimate of something neither individual reading measured cleanly, and it is closer to the truth than either of them. Note also how differently the two readings would land if you treated them as answers in their own right: 138/86 sits solidly in stage 1 territory under the current American thresholds, while 130/82 sits right at the boundary. The averaged 134/84 is the honest description, and what a number in that region means is set out in the guide to 130/80 readings.
Worked example two: three readings, awkward numbers
Averages rarely come out whole. Say your three readings are 141/88, 134/85 and 132/84.
Systolic: (141 + 134 + 132) / 3 = 407 / 3 = 135.67 → 136
Diastolic: (88 + 85 + 84) / 3 = 257 / 3 = 85.67 → 86
Session average = 136/86 mmHg
Round to the nearest whole millimeter of mercury and round a half upward. Do not round to the nearest five or the nearest ten. Terminal digit preference, the human habit of writing numbers ending in 0, is a documented problem in manual measurement and it is one of the reasons automated devices took over, as discussed in the manual technique guide. Rounding 135.67 to 140 because it feels tidier throws away exactly the precision you spent a week collecting.
Averaging mean arterial pressure and pulse pressure
Two derived numbers are worth calculating once you have an average, and both behave nicely under averaging because both are linear.
MAP = diastolic + (systolic − diastolic) / 3
Pulse pressure = systolic − diastolic
Take the 134/84 session average from the first example. MAP works out at 84 + (134 − 84) / 3, which is 84 + 16.7, giving about 100.7 mmHg. Pulse pressure is 134 − 84, which is 50. Normal MAP sits roughly between 70 and 100 mmHg, with about 60 as the rough floor below which organ perfusion starts to suffer, and a typical pulse pressure is around 40. Because both formulas are linear, the mean arterial pressure of your averaged reading is identical to the average of the mean arterial pressures of every individual reading, so you can safely calculate it once at the end rather than for all twenty-four readings. The MAP calculator does the arithmetic for you, and how to find mean blood pressure explains why the diastolic gets the heavier weighting.
That linearity property is genuinely useful in practice. It means you can average first and derive second, which cuts the work by a factor of twenty-four, and it means nobody can argue that your MAP is wrong because you calculated it from an average. The two routes give the same answer.
Keep morning and evening separate as well as combined
Calculate three numbers from a seven-day block, not one. The combined average is what gets compared against diagnostic thresholds. The morning-only average and the evening-only average are what reveal the shape of your day. A morning average sitting well above the evening one points at the early surge that follows waking, and if you take medication at night, a high morning number can mean the dose is wearing off before the next one, which is exactly the observation behind the debate about when to take blood pressure medication. None of that is visible if you collapse everything into a single figure.
A full seven-day log, worked from start to finish
Numbers on a page beat any amount of explanation. Below is a complete, realistic week from someone who bought a monitor after a slightly high reading at a checkup. Two readings each session, two sessions a day, twenty-eight readings in total.
| Day | Morning readings | Morning average | Evening readings | Evening average |
|---|---|---|---|---|
| Day 1 (discarded) | 142/90, 136/86 | 139/88 | 138/88, 134/84 | 136/86 |
| Day 2 | 134/86, 130/84 | 132/85 | 132/84, 128/82 | 130/83 |
| Day 3 | 136/88, 130/84 | 133/86 | 130/82, 128/82 | 129/82 |
| Day 4 | 132/84, 128/82 | 130/83 | 134/86, 130/84 | 132/85 |
| Day 5 | 130/84, 126/80 | 128/82 | 128/82, 126/80 | 127/81 |
| Day 6 | 134/86, 132/84 | 133/85 | 130/84, 126/80 | 128/82 |
| Day 7 | 128/82, 126/80 | 127/81 | 132/84, 128/82 | 130/83 |
Now do the arithmetic on the twenty-four readings from days two through seven. Add the twenty-four systolic values and divide by twenty-four. Do the same for the diastolic values.
Systolic sum (days 2–7) = 3118 → 3118 / 24 = 129.9
Diastolic sum (days 2–7) = 1996 → 1996 / 24 = 83.2
Seven-day home average = 130/83 mmHg
MAP = 83.2 + (129.9 − 83.2) / 3 = 98.8 mmHg
Pulse pressure = 129.9 − 83.2 = 46.7 mmHg
What that week actually tells you
Four separate lessons fall out of this one table, and they are the reason the protocol exists.
The worst reading was 142/90. The truth was 130/83.
If this person had taken one reading on day one and acted on it, they would have walked into an appointment convinced they had stage 2 hypertension. The averaged number sits 12 mmHg lower on systolic. A single high number is a sample from the top of a distribution, nothing more, and the usual causes of a short-lived spike explain most of them.
Dropping day one moved the average by about one point.
Including day one gives 131.0/83.7 instead of 129.9/83.2. On this log the correction is small and honest reporting says so. On a log from someone genuinely rattled by a new machine, day one can sit 8 or 10 mmHg above the rest of the week, and then the correction changes the conclusion. You cannot know in advance which kind of week you are having, so you discard day one either way.
Mornings ran higher than evenings, by a little.
The morning average for days two through seven is 130.5/83.7 and the evening average is 129.3/82.7. That gap is modest and probably means nothing on its own, but tracking it across several blocks is how you spot a genuine morning pattern developing. Sleep quality feeds directly into that gap, which is why the relationship between sleep and blood pressure is worth understanding before you blame your medication.
The answer depends on which country your doctor trained in.
An average of 130/83 measured at home meets the American definition of hypertension, because the 2017 ACC and AHA guidelines put the home threshold at 130/80. The European and UK home threshold is 135/85, and by that standard the same person is not hypertensive at all. Nobody is lying. The line is drawn in a different place, which is why the label matters less than the trend and the total risk picture, and why what level counts as high blood pressure needs more than one paragraph to answer.
Home thresholds sit lower than office thresholds for a straightforward reason. Being measured by a stranger in a clinic raises most people numbers by a few mmHg, so a clinic cutoff of 140/90 corresponds to a home cutoff nearer 135/85, and the American 130/80 office cutoff maps to the same figure at home. When you present a home average, say clearly that it is a home average. Handing over 130/83 without that context invites it to be compared against the wrong line.
When your two readings disagree by more than 5 mmHg
You take two readings and get 148/92 followed by 136/86. Twelve points apart on systolic. Averaging them gives 142/89, which feels like a number invented by a committee rather than measured by a machine. So what do you do?
The standard instruction is to take a third reading after another minute of rest. Both the American and European approaches say the same thing in slightly different words: when the first two disagree substantially, keep going until the readings settle, then average. A gap of more than about 5 mmHg systolic is the usual trigger, and a gap of more than 10 mmHg is a strong hint that something in your setup is wrong rather than something in your circulation being interesting.
The arithmetic branch point
Suppose the third reading comes in at 134/84. You now have a decision, and it changes the answer by several mmHg.
Average all three: (148 + 136 + 134) / 3 = 139.3 → 139 systolic; (92 + 86 + 84) / 3 = 87.3 → 87 diastolic → 139/87
Average the last two: (136 + 134) / 2 = 135; (86 + 84) / 2 = 85 → 135/85
Four mmHg of systolic separates those results, and 139/87 versus 135/85 can be the difference between two different conversations with a doctor. There is no universal ruling that settles it. Averaging the last two readings is defensible when the first is obviously the alerting response and the second and third agree closely. Averaging all three is defensible when the readings scatter without an obvious pattern. What is not defensible is choosing whichever method gives the lower number on that particular day. Decide your rule in advance, write it in your log, and apply it even when you dislike the result.
Work through the causes before you blame your arteries
| What you see | Most likely cause | What to do about it |
|---|---|---|
| First reading far higher, then two that agree | Alerting response, not enough rest | Rest longer before starting, use the second and third |
| Readings scattered with no pattern | Irregular heartbeat, movement, talking | Check for an irregular heartbeat symbol, sit completely still, consider a manual check |
| Every reading high, all close together | Cuff too small or wrapped over clothing | Measure your arm and follow the cuff sizing guide |
| Numbers climbing steadily through the session | Arm dropping below heart level, or rising discomfort | Support the arm on a table, review positioning basics |
| Wildly different from your usual | Different arm, different cuff, or a different device | Standardize on one arm, one cuff and one device, then repeat the session |
| Wrist device disagreeing with an upper-arm device | Wrist position, or an unvalidated sensor | Trust the validated upper-arm device, see the reliable monitor comparison |
An irregular heartbeat deserves its own note. Oscillometric monitors estimate pressure from the pattern of pulsations under the cuff, and an irregular rhythm scrambles that pattern. Devices with an irregular heartbeat indicator will flag it, and when the flag appears repeatedly across a week, the readings are less trustworthy and the flag itself is worth showing a doctor. That is one of the few situations where a listening technique with a stethoscope still beats the machine on your kitchen table.
One last cause that catches people out: readings taken too soon after something obvious. Coffee, a cigarette, a heavy meal, a brisk walk in from the car or a full bladder will each move the numbers, sometimes by a lot. Caffeine effects can persist for a couple of hours in people who do not drink it regularly, exercise leaves pressure disturbed for a while as covered in blood pressure after exercise, and digestion shifts things for an hour or two after a large meal. Those readings are not wrong, they are just answering a different question. Note the context and move on.
The seven-day protocol in detail
The seven-day home block is the closest thing home monitoring has to a gold standard, and it is worth understanding why each rule is there rather than following it blindly.
Seven consecutive days, not seven days when you feel like it
Consecutive matters because the thing you are averaging out is day-to-day variation, and days are not interchangeable. Monitoring on the three days you felt calm and skipping the two stressful ones produces a beautiful average that describes a person who does not exist. If you miss a day, do not patch the gap by doubling up the next day. Extend the block by one day at the end, or start over if you missed several. Seven clean consecutive days beats ten patchy ones.
Two sessions a day, at fixed times
Morning and evening sessions capture different parts of the daily rhythm, and both go into the average. The morning session should happen before medication and before breakfast, which is the point at which any overnight dose has had the longest to wear off. The evening session goes in the hours before bed. Fixing the times matters more than picking clever ones, and the reasoning behind the specific windows belongs to the best time to check guide rather than here.
Discard day one
The instruction to bin the first day is one of the few places where the guidelines are unusually direct. Day one carries the novelty of the device, the awkwardness of the routine and whatever prompted you to start monitoring, all of which inflate the numbers. It is not that day one is wrong. It is that day one is measuring a different thing, which is your reaction to measuring.
Twelve readings is the floor, twenty-four is the target
If seven days is genuinely impossible, three days of morning and evening sessions gets you twelve readings once day one is discarded, and twelve is the widely quoted minimum for a usable home average. Below that the average is too easily distorted by one strange session. The full week gives twenty-four, and the added precision from those extra twelve readings is the difference between a number a doctor will act on and a number they will ask you to repeat.
When to run a block
A block is a project with a trigger. Run one in the week before a scheduled appointment so the data is fresh. Run one starting a week or two after any change to medication, so the new dose has had time to settle. Run one if you have started a serious change to diet, exercise or drinking, since those levers take weeks rather than days to show up in the numbers. Run one if you feel something has changed. Do not run one continuously, forever, out of habit.
Blocks answer questions, habits answer nothing
A seven-day block with a clear question behind it produces a number worth having. Random checking produces a folder of screenshots. If you are not sure what question you are asking, start with what a good number even is and work backward from there.
How many readings a diagnosis actually requires
Nobody should be diagnosed with high blood pressure on one reading. Everybody knows this and it still happens, usually because a single number got recorded in a chart and nothing followed it up.
In the office
American guidance asks for at least two readings taken on each of at least two separate occasions, averaged. That is a minimum of four readings across a minimum of two visits, and the readings on any one visit should be taken after five minutes of quiet sitting with the patient properly positioned. Many clinics now use automated office measurement, where a device takes three or more readings on its own with nobody in the room, precisely because the presence of staff pushes the number up.
At home
The seven-day block described above, giving twenty-four readings across six days after day one is dropped, is what home monitoring contributes. Its diagnostic threshold is 130/80 under American guidance and 135/85 under European and UK guidance. Home data is not a second-class substitute for clinic readings. For predicting future cardiovascular events, home and ambulatory averages generally outperform office measurement, which is why guideline groups have moved steadily toward requiring out-of-office confirmation before starting treatment.
Ambulatory monitoring
A 24-hour ambulatory monitor takes a reading every 15 to 30 minutes through the day and every 30 to 60 minutes overnight, producing somewhere in the region of 50 to 70 usable readings. It is the reference standard, and the only common method that captures what happens to your pressure while you are asleep. Its daytime threshold matches the home figure of 135/85 in European practice, with a lower nighttime threshold, because pressure normally dips during sleep. People whose pressure fails to dip overnight carry extra risk, and that is information no home routine can give you.
The two conditions that only averaging can reveal
White coat hypertension
High in the clinic, normal at home. Common, and it means a large fraction of people started on treatment from office readings alone did not need it. The seven-day home block is how it gets caught.
Masked hypertension
Normal in the clinic, high at home. More dangerous, because it is invisible to the health system. It is found only by someone who measures properly at home, averages honestly and reports the result even though the clinic said they were fine.
Both of these are arguments for doing the home block properly rather than for skipping the doctor. What a specific averaged number means once you have it is a separate topic, and the batch of readings-focused guides covers the common landing points: whether 120/80 is as good as its reputation, what 122/70 signifies, the reading of 112/75, and at the lower end whether 100/70 counts as low and what to make of 90/60.
A brief reminder of the American category boundaries, since they are what a home average gets compared against. Normal is below 120 and below 80. Elevated is 120 to 129 with a diastolic still below 80. Stage 1 is 130 to 139 systolic or 80 to 89 diastolic. Stage 2 is 140 or above or 90 or above. Anything above 180 systolic or above 120 diastolic is a crisis reading that needs immediate attention. Europe, the UK and the World Health Organization still open hypertension at 140/90, which is why the same average can be called two different things depending on where you are standing.
How often should you actually check
This is where advice online falls apart, because the honest answer is that it depends on why you are measuring, and most sources pick one schedule and apply it to everyone. Here is the version that matches how blood pressure is actually managed.
If you take medication and something has just changed
A new drug, a dose increase, a switch between classes: this is the period where home data earns its keep. Morning and evening sessions daily for the first one to two weeks after the change give your prescriber a picture of how the new regimen behaves, including whether it is working too well. Dizziness on standing during this window is worth reporting rather than tolerating, since it often means the pressure has been pulled down further or faster than intended. Never change a dose yourself on the strength of your own numbers, and never stop a drug because your readings look good, which is the whole argument of can you stop taking blood pressure tablets.
If you take medication and everything is stable
Once your average is where your doctor wants it and the dose has been unchanged for months, daily monitoring stops adding information. A seven-day block once a month is generous. A block before each review appointment is the minimum. Plenty of well-controlled people run a block quarterly and do nothing in between, and their care is not worse for it.
If you have raised numbers but no treatment yet
Run a block every three to six months and compare block averages rather than individual readings. The comparison you care about is 132 last quarter against 129 this quarter, not Tuesday against Wednesday. This is also the group for whom lifestyle changes matter most, so time your blocks to sit a couple of months after a genuine change rather than a week after it.
If your numbers are normal and you have no risk factors
Once or twice a year is enough, plus whatever gets taken at routine appointments. Many people in this group own a monitor because a relative had a scare, and end up checking weekly out of vague anxiety. That is not monitoring, it is worrying with equipment.
If you are pregnant
Follow the maternity team schedule exactly, because the thresholds and the response to a high reading are different in pregnancy and the risks move fast. Home monitoring in pregnancy is often daily and sometimes twice daily, and a reading of 140/90 or above is reported the same day rather than averaged into a weekly block. Pregnancy and blood pressure covers why the rules change.
Nothing in this section is a reason to check less often than your own doctor has asked. If you have been told to measure daily, measure daily. The schedules above describe the general case, and a specific instruction from someone who knows your history beats a general case every time.
How to record and chart your readings
A log is not a diary of numbers. It is a small dataset, and a dataset with missing columns is worth much less than one with all of them. Here is what belongs in it and why each column earns its place.
| Column | Example | Why it matters |
|---|---|---|
| Date | 14 Mar | Lets you check the days were consecutive and identify which block a reading belongs to |
| Time | 07:10 | Separates morning from evening averages and exposes drift in your routine |
| Arm | Left | Arms differ, sometimes by 10 mmHg, per the arm comparison |
| Reading 1 | 134/86 | The raw number, always recorded even though it is usually the high one |
| Reading 2 | 130/84 | The second of the pair, taken a full minute later |
| Reading 3 | — | Only filled in when the first two disagreed by more than 5 mmHg |
| Session average | 132/85 | The number that goes into the weekly calculation |
| Pulse | 68 | Free information from the same inflation, and useful context |
| Medication taken | Yes, 22:00 | Tells you whether a morning reading is a trough or a peak |
| Notes | Poor sleep, coffee at 06:30 | Turns an unexplained outlier into an explained one |
The notation, since people ask
Blood pressure is written systolic over diastolic followed by the unit: 132/85 mmHg, said out loud as one thirty-two over eighty-five. Systolic is the higher number and represents the pressure while the heart is contracting, diastolic is the lower one and represents the pressure between beats. If you want the mechanics behind those two figures, how blood pressure works explains the plumbing behind both of them. Pulse is written separately as a number of beats per minute, never as part of the fraction.
Three rules that make a log trustworthy
Write every reading, including the ugly ones
The temptation to leave out a 156/98 because you had just run up the stairs is enormous and you must resist it. Record it and write the reason next to it. A log with obvious outliers and honest explanations is far more useful to a doctor than a suspiciously smooth one, and the outliers themselves sometimes turn out to be the interesting part.
Write it down immediately, on paper or in an app
Not in ten minutes. Not from device memory at the weekend. Transcription errors from memory are common and always run in the direction of what you expected to see. If you use an app, check that it stores the raw readings and not just an average it computed for you.
One person per log, one device per log
Household monitors get shared, and multi-user modes get forgotten. If two people use the same machine, keep separate logs and check which user profile is selected before each session. Mixing two people readings into one average produces a number that describes nobody.
Paper, spreadsheet or app
All three work and each fails differently. Paper is the most reliable because nothing syncs, nothing updates and nothing loses your history in a migration, but it makes calculating a twenty-four reading average tedious. A spreadsheet does the arithmetic for you and charts it in two clicks, which is why it is my default recommendation for anyone comfortable with one. Apps are convenient and often connect straight to the monitor by Bluetooth, but they vary wildly in whether they let you export, whether they average the way you want and whether they will still exist in three years.
Whatever you choose, the export test is the one that matters. Before you commit to a system, try to get your data out of it as a plain table. If you cannot, your log lives at the mercy of a company you have never met, and the doctor who eventually wants to see it will get a screenshot of a chart rather than the numbers underneath.
Charting it so the trend is visible
Plot session averages, not individual readings. Two lines, systolic on top and diastolic below, with the date along the horizontal axis. Twenty-eight raw readings plotted as points looks like static and hides everything. Fourteen session averages, or better still seven daily averages, show you a shape. Mark medication changes on the chart with a vertical line and a date, because the whole reason for charting is to see whether something you did moved the number.
Resist the urge to set the vertical axis from zero. A chart running from 60 to 160 makes a genuine 10 mmHg improvement look like nothing. Set the axis to a range that spans your actual readings plus a little headroom, and the changes become legible. Equally, resist reading meaning into every wiggle. Blood pressure moves several mmHg between sessions for no reason at all, and the only movements worth interpreting are ones that persist across a whole block.
Build the log around the calculation
If your log has a session average column, the weekly average is one formula away and the mean arterial pressure is one more. Keep the MAP calculator bookmarked next to the spreadsheet, and browse the rest of the health calculators when you need pulse pressure or body measurements alongside it.
What to take to the appointment
Turning up with a phone full of screenshots wastes the four useful minutes of a consultation. Turn up with a page. Here is what that page should say and what should be in your bag alongside it.
The headline average, with its denominator
Say it as a sentence: my home average over six days was 130/83, from twenty-four readings, taken morning and evening on my left arm, day one discarded. That sentence contains everything a doctor needs to decide how much weight to give it. An average with no denominator is just a number you found somewhere.
The full log, not a summary
One page, all readings, dates and times visible. If your doctor wants to check whether your mornings run high or whether one bad day is doing the work, they need the raw rows. Print it if it lives in an app.
The monitor itself
Take the actual device and its cuff. A clinic can compare it against their own equipment in a minute, and a monitor reading 8 mmHg low has been quietly reassuring you for a year. Bring the cuff you use, since a cuff that does not fit your arm invalidates everything, as the sizing guide explains. It also helps to know whether your model appears on a validation list, which is the subject of the reliable monitor comparison.
Your medication list with times
Names, doses and the time of day you take each one, plus any recent changes with dates. Blood pressure numbers are uninterpretable without knowing when the last dose went in, and over-the-counter drugs count too, since anti-inflammatory painkillers and some decongestants push pressure up.
Anything odd, written down before you forget
Dizziness, headaches, palpitations, a week of terrible sleep, an irregular heartbeat symbol that appeared four times. Symptoms are hard to recall under pressure in a consultation room, and what high blood pressure does and does not feel like is worth reading beforehand so you know what is worth mentioning.
One more thing to bring: a question. The most useful one is usually what number are we aiming for in my case, because targets are individualized and the answer for a 45-year-old with no other problems differs from the answer for someone with diabetes or kidney disease. Getting that number written down turns your future logs from anxiety generators into progress reports.
The over-monitoring trap
There is a specific pattern that home monitors create, and if you recognize yourself in it you are not unusual. It goes like this. You take a reading and it is higher than you hoped. That bothers you, so you take another one to check. The second is higher still, because you are now anxious about the first. So you take a third. By the fourth you are measuring your own alarm, and the machine is faithfully reporting it back to you.
The physiology behind this is not disputed. Acute stress raises blood pressure through the sympathetic nervous system within seconds, which is the same mechanism that produces the white coat effect. Attaching a cuff to a worried person and asking them to sit still while it squeezes reliably produces a higher number than attaching it to a calm one. So the feedback loop is real, and the more anxious you are about the result, the more the act of measuring changes what you are measuring. What is less well established is how much harm the habit does over the long run, since most of the evidence is clinical observation rather than trials. Nobody has run a randomized study of obsessive checking, and nobody is going to.
Signs you have crossed the line
You retake until you get a number you like
If a session ends when the reading is acceptable rather than after a fixed number of inflations, you are not measuring, you are sampling until you find the answer you want. The resulting log understates your pressure and misleads the person treating you.
You check more than twice a day without being asked to
Outside a medication change or a specific instruction, more than two sessions a day adds nothing to the average and a great deal to the anxiety. Twenty-four readings from six days beats sixty readings from two frantic ones.
You check because you feel something
Headache, warm face, a thumping in the ears. These sensations correlate poorly with blood pressure, and the checking usually confirms a number that the checking itself produced. The exception is a genuine symptom pattern, which belongs in the emergency section further down rather than in a habit.
The monitor lives on the table rather than in a cupboard
Visibility drives use. People who put the device away between blocks check less and worry less, and their averages are no worse for it.
What to do instead
Set the schedule in advance and hold to it regardless of the results. If a reading alarms you, the correct response is to note it, finish the session as planned and look at it in the context of the block, not to start an unscheduled investigation. Put the monitor away when a block ends. If the anxiety itself is the problem rather than the numbers, that is a legitimate thing to raise with a doctor, and the techniques in how to calm blood pressure down are more useful than another inflation. Slow breathing before a session, done consistently every time, is fine. Slow breathing until the number drops and then recording that one is not.
A reasonable rule for anyone who tends toward anxious checking: two sessions a day, seven days, then the monitor goes in a drawer for a month. Write the next block date on the calendar. The structure does more for accuracy than any amount of willpower.
There is a mirror image of this problem worth naming: people who bought a monitor, took three readings in 2022 and have not touched it since. Under-monitoring is quieter but at least as common, and it is how masked hypertension goes undetected for years. The fix is the same one, which is a scheduled block rather than a vague intention. General background on the whole subject sits in the blood pressure section of the blog if you want the wider picture before committing to a routine.
Letting the machine do the averaging
Many monitors will average for you, and the feature is worth using once you understand which of the several different things your model does.
Triple measurement mode
The most useful feature. Press start once and the device takes three readings, usually 15 to 60 seconds apart, then displays the average of the three. It removes the temptation to fiddle between readings and it produces a session average automatically. If your monitor has it, use it, and record both the average and the individual readings if the display will show them.
Memory averages
Almost every device stores a history and offers an average of it. The question is which readings that average includes. Some average the last three. Some average everything from the past seven days, which is close to what you want. Some average every reading ever stored, which is useless, because it blends a reading from before your medication change with one from last night. Read the manual, find out which behavior yours has, and if it is the lifetime version, ignore the number entirely and do the arithmetic yourself.
App and cloud averages
Connected monitors push readings to an app that computes averages and draws charts. Convenient, and the charts are usually better than anything you would build. Two cautions. First, apps sometimes include readings you would have excluded, such as the one you took after running for the bus, so look for a way to flag or annotate rather than delete. Second, check what the app calls an average before you trust it, because a rolling 30-day average and a seven-day block average answer different questions.
Wearables and cuffless devices
A wrist band that claims to track blood pressure continuously would in theory solve the averaging problem completely, giving hundreds of readings a day with no effort. In practice the technology is not there yet for most consumer devices. Cuffless optical estimates drift, usually need regular calibration against a real cuff, and very few models have passed independent validation. Products in this category come and go quickly, and a band marketed on blood pressure tracking is not the same as a band that has been shown to measure it accurately. The current state of play is covered honestly in can smart watches measure blood pressure. Until that changes, an average built from a validated upper-arm cuff beats an average built from thousands of estimates of unknown accuracy.
Whatever averages your device or app calculates, keep your own record of the raw readings. Firmware updates change behavior, apps get discontinued, and device memory fills up and overwrites. A single page of handwritten numbers has outlasted a lot of expensive ecosystems.
Common averaging and logging mistakes
Every one of these is something people do in good faith, and every one of them corrupts the number they end up reporting.
| The mistake | What it does to your average | The fix |
|---|---|---|
| Recording only the lowest reading of the session | Understates pressure, often by 5 to 10 mmHg | Record both, average them, live with the result |
| Recording only the second reading | Slight understatement, and an inconsistent method if you sometimes average instead | Pick one convention and write it in the log header |
| Rounding to the nearest 5 or 10 | Adds several mmHg of noise and can flip a category boundary | Round to the nearest whole mmHg |
| Averaging systolic and diastolic together | Produces a meaningless single figure | Average each column separately, and use mean arterial pressure if you want one combined number |
| Switching arms mid-block | Introduces a systematic offset that looks like a real change | One arm for the whole block, per the arm guide |
| Averaging across a medication change | Blends two different treatment states into one meaningless mean | Start a new block on the day the change takes effect |
| Comparing a home average against an office threshold | Makes normal home readings look better than they are | Use 130/80 for home under US guidance, 135/85 under European |
| Averaging a week with three missing days | Gives a small sample the authority of a full block | Extend the block, or report it honestly as a partial one |
| Using the device lifetime average | Mixes years of unrelated readings | Calculate the block average yourself |
| Deleting readings you cannot explain | Removes exactly the data a doctor would find interesting | Keep them and annotate them |
The one that does the most damage is the first: quietly keeping the number you prefer. It feels harmless because you are only fooling yourself, but the log is going to be used to make a decision about medication, and a log that runs 8 mmHg low can leave a genuine problem untreated for years. High blood pressure does almost nothing you can feel until it has already done damage. The log is your only instrument. Do not put a thumb on the scale.
When one reading does matter
Everything above says do not act on a single reading. There are exceptions, and they are the situations where waiting for an average could hurt you.
A reading above 180 over 120 accompanied by symptoms is a medical emergency and the averaging rules do not apply. That combination can mean organ damage is happening while you are looking at the display, and the correct response is emergency services rather than a second reading. Without symptoms, the standard advice is to rest for five minutes and repeat once, because a single very high number is sometimes a measurement artifact, and to contact a doctor the same day if it stays there. What counts as really high blood pressure and the relationship between pressure and stroke risk both go further into the thresholds.
At the other end, a low reading with symptoms deserves the same urgency even though the numbers look reassuring. Fainting, confusion or clammy skin alongside a low pressure can point to something acute, and the reading is a clue rather than the problem. The signs of low blood pressure covers what to watch for, and a symptom-free low reading in someone who always runs low is usually nothing at all.
There is a third case that no threshold captures: a sudden, sustained departure from your own baseline. If your average has sat around 118/74 for two years and this week every session reads in the 150s, the fact that 150 is not an emergency number is beside the point. Something changed. New medication, an illness, a new supplement or a genuine development in your circulation are all candidates, and it is worth a conversation regardless of which side of a guideline line the number falls on. Personal baselines beat population thresholds for spotting change, which is the strongest argument for keeping a log in the first place.
Frequently asked questions
How many times should I take my blood pressure in one sitting?
Two, one minute apart, and average them. Take a third if those two differ by more than about 5 mmHg on the systolic number. Beyond three you are usually measuring your own reaction to being measured rather than anything useful, and the extra readings drag the session out to the point where people stop bothering. If your device has a triple measurement mode that does the three inflations and the average automatically, use it.
How do you calculate average blood pressure?
Add up all the systolic values and divide by the number of readings. Do the same separately for the diastolic values. Write the result as a pair, rounded to the nearest whole mmHg. So readings of 138/86 and 130/82 give (138+130)/2 = 134 and (86+82)/2 = 84, which is 134/84 mmHg. You never average the systolic and diastolic numbers together. If you want a single combined figure, that is mean arterial pressure, worked out as the diastolic plus one third of the gap between the two numbers.
How do you work out average blood pressure over a week?
Take two readings each morning and each evening for seven consecutive days, which gives twenty-eight readings. Discard all four from day one. Add the remaining twenty-four systolic values and divide by twenty-four, then do the same for the diastolic values. That pair is your seven-day home average, and it is the number a doctor will treat as meaningful. Calculate the morning-only and evening-only averages as well, since the gap between them tells you something the combined figure hides.
What is an average blood pressure reading?
Two different questions hide in that one. Your own average blood pressure reading is the mean of your readings across a defined block, and for most healthy adults it lands somewhere below 120/80. The population average for American adults sits around the low 120s systolic over the low-to-mid 70s diastolic, but that figure includes tens of millions of people with untreated hypertension, so it is not a target. Compare your number against the category boundaries in the normal range guide, not against the national mean.
What is the average blood pressure for a man?
Population surveys generally put average systolic pressure in men slightly above that of women through early and middle adulthood, with both figures climbing steadily with age as arteries stiffen. A man in his twenties averages considerably lower than a man in his sixties. The clinically useful point is that men and women are assessed against identical thresholds, so a higher male average means more men cross into a diagnosis, not that a higher number is acceptable in men. Age-adjusted averages describe a population, and the number that matters for you personally is your own six-day home average.
What is the average blood pressure for a woman?
Women tend to average slightly lower than men in early and middle adulthood, then rise more steeply in the decades after menopause, frequently overtaking men of the same age in later life. Pregnancy changes the picture again, with pressure typically falling in the middle of pregnancy before returning toward baseline. The full breakdown lives in the guide to average blood pressure for a woman, and pregnancy-specific behavior is covered in does pregnancy raise blood pressure.
What is the average blood pressure and heart rate together?
For a healthy adult at rest, a pressure below 120/80 mmHg with a pulse somewhere between 60 and 100 beats per minute is the usual description, and fit people often sit in the 50s for pulse. The two numbers come from the same cuff inflation but they measure different things, and they do not move together in any simple way. A racing pulse does not imply high pressure and a slow pulse does not imply low pressure, which is the point of is bpm the same as blood pressure. Record both in your log, because a pulse that changes alongside a pressure change is useful information.
What is the average blood pressure of a newborn?
A healthy term newborn runs far lower than an adult, with systolic pressure typically somewhere in the 60 to 80 range and diastolic roughly 30 to 50, rising over the first weeks of life. In the first days, mean arterial pressure is often compared against the baby gestational age in weeks as a rough lower bound. None of this can be measured with an adult home monitor, which needs an appropriately sized neonatal cuff and a device validated for infants. Any concern about a baby blood pressure belongs with a pediatric team.
What is the average blood pressure range?
For adults under the current American classification, normal is below 120 systolic and below 80 diastolic. Elevated runs 120 to 129 systolic with a diastolic still under 80. Stage 1 hypertension is 130 to 139 systolic or 80 to 89 diastolic, and stage 2 is 140 or above or 90 or above. Above 180 systolic or above 120 diastolic is a crisis. Europe, the UK and the World Health Organization still begin hypertension at 140/90, so the same reading gets labeled differently depending on where you are. Home averages are compared against slightly lower cutoffs than office readings, as explained in what level is high blood pressure.
How many people in America have high blood pressure?
Close to half of American adults meet the current definition of hypertension, which works out at roughly 120 million people. That figure jumped when the 2017 American guidelines lowered the threshold from 140/90 to 130/80, moving many millions of people into a diagnosis overnight without their readings changing at all. A large share of that group either does not know they have it or does not have it under control, which is the entire argument for measuring properly at home rather than relying on one number a year at a checkup.
How do you record blood pressure?
Write it as systolic over diastolic with the unit, for example 132/85 mmHg, and put the pulse next to it as a separate number rather than part of the fraction. Alongside the reading, record the date, the time, which arm you used, whether you had taken medication and anything unusual about the circumstances. Log both raw readings from the session as well as their average. A log with those columns can answer questions later. A list of averages with no context cannot.
How do I chart blood pressure over time?
Plot session averages or daily averages as two lines, systolic above and diastolic below, with dates along the bottom. Do not plot every individual reading, because the scatter buries the trend. Set the vertical axis to span your actual readings with a little headroom rather than starting at zero, otherwise real changes look flat. Mark medication changes and any major lifestyle change with a dated vertical line, since the whole reason for charting is to see whether an intervention moved the number. A spreadsheet does this in about two minutes.
Does the Hume band track blood pressure?
Wrist bands marketed with blood pressure tracking, including the Hume band and a long list of similar products, generally produce estimates from optical sensors rather than measurements from a cuff. Very few cuffless consumer wearables have passed independent validation against the standards that apply to medical devices, and those that come closest usually require periodic calibration against a real upper-arm cuff. Treat any number from a band as a rough trend indicator at best, and never build a diagnostic average from one. Can smart watches measure blood pressure goes through the evidence in detail.
Should I take my blood pressure more than twice a day?
Only if your doctor has asked you to. Two sessions a day for seven days produces everything a diagnostic average needs, and additional sessions add anxiety rather than accuracy. The exception is the first week or two after a medication change, when more frequent readings genuinely help your prescriber, and pregnancy, where the monitoring schedule is set by your maternity team. If you find yourself checking six or eight times a day because a number worried you, that is a pattern worth breaking rather than a routine worth keeping.
Which single number do I give my doctor if they only want one?
Your six-day home average, stated with its denominator: for example 130 over 83 from twenty-four readings across six days, left arm, morning and evening, day one discarded. That one sentence is worth more than a folder of screenshots. Bring the full log anyway in case they want to look at the spread, and bring the monitor so it can be checked against clinic equipment. Comparing your average against a clinic threshold rather than a home threshold is the most common way that conversation goes wrong, so say clearly that the number came from home.
The whole thing in six lines
Two readings, one minute apart, averaged. Morning and evening. Seven consecutive days. Bin day one. Average the twenty-four readings that remain. Never act on a single number unless it is above 180 over 120 or you feel genuinely unwell.
Everything else is refinement. Keep the same arm and the same cuff so your numbers stay comparable. Write down the readings you dislike as carefully as the ones you like. Calculate morning and evening averages separately as well as combined, because the shape of your day carries information the single figure loses. Put the monitor away between blocks. And when you finally have an average, feed it into the mean arterial pressure calculator to see the perfusion number underneath the pair, then look up what that pair actually means rather than guessing.
The reason all of this matters is that the decision hanging off your number is usually whether to start a medication you will take for the rest of your life. A method that runs 8 mmHg high pushes people onto treatment they do not need. A method that runs 8 mmHg low leaves real hypertension quietly damaging arteries for a decade. Getting the count, the spacing and the arithmetic right costs you twenty minutes a day for one week, a few times a year. That is a good trade. More tools and guides are collected at waldev, and the natural next read is whether 120/80 is really the target everyone assumes, because a surprising number of people are chasing a number that was never meant to be a goal.
Related reading
Medical disclaimer and sources
This article is general information, not medical advice. It cannot diagnose you, and it is not a substitute for assessment by a clinician who knows your history. Nothing here should be used to start, stop or adjust any medication. If a reading is above 180 systolic or above 120 diastolic and you have chest pain, breathlessness, weakness or numbness on one side, difficulty speaking, vision changes or a severe headache, treat it as an emergency and call 911 or your local emergency number. If the reading is that high without symptoms, rest for five minutes, repeat it once and contact a doctor the same day. Anyone who is pregnant, has kidney disease, diabetes or an existing heart condition should follow the monitoring schedule their own clinician has set rather than the general guidance on this page.
American Heart Association
Guidance on home blood pressure monitoring, including resting before measurement, taking two readings a minute apart and averaging them. heart.org
Centers for Disease Control and Prevention
US hypertension prevalence figures and public guidance on measuring blood pressure at home. cdc.gov
