The 'white-coat' effect - where blood pressure rises during a check by a doctor - is even worse in someone whose level is already high, researchers say.
The effect is due to patients becoming stressed by being in a doctor's surgery or a hospital.
Writing in the British Medical Journal, an Australian team say giving people a cuff to wear for 24 hours is a better way of checking blood pressure.
A UK expert said it showed clearly that external factors affected readings.
High blood pressure affects about 40% of adults in the UK and is a major risk factor for heart attack, heart failure, kidney disease and stroke.
In 2001, 90 million prescriptions for blood pressure lowering drugs were issued by the NHS at a cost of £840 million.
It can either be measured in a clinical setting, or by the patient wearing a cuff as they go about their daily lives - known as ambulatory blood pressure checks.
The researchers monitored over 8,500 patients who were being assessed at 11 blood pressure clinics around Australia.
They compared ambulatory blood pressure measurements with those taken by doctors and nurses and found that there can be a difference of as much as 29 units if a doctor checked it, compared with a rise of 17 units if a nurse took the measurement.
The differences also varied depending on the sex and age of the patient.
However, the study also found that the closer the patient's blood pressure to normal levels, the less of a difference between measurements taken by ambulatory monitoring and those taken by a nurse or doctor.
No 'one size fits all'
Professor Arduino Mangoni, who recently joined the University of Aberdeen from Flinders University in Adelaide, said: "Ambulatory blood pressure monitoring is the tool of choice to correctly diagnose high blood pressure.
"Clearly, if you're going to be treating a person for the rest of their life, you want to get the readings right, and often the reading in the doctor's office is much higher."
Professor Mangoni said the fact there was a difference between doctors' readings and ambulatory measurements was already known - but the surprise finding was how big the gap was for those with high blood pressure.
He said the team's findings should be used to help shape new blood pressure monitoring guidelines.
"Current guidelines for the diagnosis and treatment of hypertension don't pay enough attention to the role of ambulatory monitoring, often adopting a one- size-fits-all approach which doesn't properly address different patient groups."
Writing in the BMJ, Professor Richard McManus, a cardiovascular expert from the University of Birmingham, said a patient's care should be managed using clinic-based and ambulatory measurements, taking into account where the test was done and by whom.
Professor Graham MacGregor, chairman of the Blood Pressure Association, said: "This is interesting research which clearly illustrates how external factors such as environment and who is checking blood pressure can have a significant impact on blood pressure readings.
"Many people feel slightly anxious when going to see a doctor, which is why we have always encouraged blood pressure measuring at home as well as in the clinic, and promotes the use of home blood pressure monitors and ambulatory testing where indicated.
"This research may well be considered as part of the review of the NICE [National Institute of health and Clinical Excellence] guidelines for the treatment of hypertension which is currently being conducted, and will be released next year."