Hypertension Risk
The Hypertension Risk is a key metric designed to estimate an individual's likelihood of consistenlty experiencing blood pressure levels above the normal range. The Hypertension Risk can serve as an early risk indicator for the potential development of hypertension within the next year. It helps individuals recognize possible risk factors, encouraging proactive self-care and increasing awareness.
What is blood pressure?
Blood pressure is the force exerted by circulating blood against the walls of blood vessels as the heart pumps blood throughout the body. It is a vital sign that reflects the efficiency of the cardiovascular system and is essential for ensuring that oxygen and nutrients reach tissues and organs.
What is an elevated blood pressure? An elevated blood pressure refers to blood pressure readings that are higher than normal but not yet in the hypertension range. This condition serves as an early warning that there is a risk of developing high blood pressure (hypertension) in the future.
Specifically, it refers to a blood pressure of 130–139 mmHg systolic and/or 85–89 mmHg diastolic.
What is hypertension?
Hypertension is a medical condition where the force of the blood against the walls of arteries is consistently too high. Over time, this increased pressure can damage arteries and lead to serious health problems. Specifically, following guidelines from the European Society of Cardiology, it refers to a blood pressure of more than 139 mmHg systolic and/or more than 89 mmHg diastolic.
IntelliProve’s processing engine uses machine learning and context-based reasoning to determine the risk of an increased blood pressure. The model uses features extracted from the rPPG signal measured during the Face Scan combined with user-reported data. A risk prediction score, described in detail in Parikh et al., is developed from Bèta-coefficients for variables associated with hypertension incidence in multivariable Weibull regression models.
The factors incorporated in the model include data extracted from the Face Scan, including sytolic and diastolic blood pressure, alongside the following user attributes:
- Age
- Sex
- BMI (Height, weight)
- Lifestyle habits such as smoking
- Parental hypertension
These factors can be reported by the user during the Face Scan, which is included in the journey by default.
Alternatively, you can also provide these via the API. Age and sex can be provided when creating the user: refer to Create new user. Other parameters can be provided using the following endpoint: Update physical health profile.
The link to the rPPG signal
The rPPG signal reflects the dynamics of the blood pressure. During the systole phase, indicated in red in the figure below, the heart contracts and pumps oxygenated blood into the arteries. The systolic phase ends with the heart valve shutting again, inducing the dicrotic notch. During this phase, the blood pressure rises to its maximum value (i.e. systolic blood pressure).
The diastolic phase, indicated in green, is the time during which the heart refills with blood. This phase is longer than the systolic phase. The blood flow in the arteries gradually decreases, and the blood pressure drops to its minimum value (i.e. diastolic blood pressure).
With higher blood pressure, the systolic peak of the signal may become more pronounced due to the increased force of blood being pushed into the arteries. Furthermore, the elasticity of the muscle layer around the arteries – also referred to as arterial compliance – reduces as we age, which reduces the contrast between the systolic and diastolic phases of the rPPG signal. As the arteries can’t expand and contract as efficiently, there’s less differentiation in blood flow between these two phases, making the waveform less dynamic overall.
IntelliProve leverages these types of insights to extract meaningful, predictive features for the Hypertension Risk.
The model
The model was developed based on data of a longitudinal cohort study including 1717 nonhypertensive individuals (age: 20 to 69 years, mean age: 42 years, sex: 54% female), without diabetes and with both parents in the original cohort of the Framingham Heart Study (Drawber et al. 1963). During follow-up (median time over all person-examinations: 3.8 years), 796 persons (sex: 52% female) developed new-onset hypertension. In multivariable analyses, age, sex, systolic and diastolic blood pressure, body mass index, parental hypertension, and cigarette smoking were significant predictors of hypertension. According to the risk score based on these factors, the risk for incident hypertension was classified as low (<5%) in 34% of participants, medium (5% to 10%) in 19%, and high (>10%) in 47%.
When do you get a first reading?
The first result will be available the day after the first Face Scan.
Why do I need to take multiple measurements?
Hypertension is typically identified after multiple readings over time confirm the elevated levels, as blood pressure can temporarily rise due to factors like stress, exercise, or caffeine.
Performance of the prediction algorithm was assessed by using calibration and discrimination measures. The reported predication model has a very good discrimination (c-statistic, 78.8 [95% CI, 73.3 to 80.3]) and was well calibrated (Hosmer–Lemeshow chi-square statistic, 4.35 [values < 20 indicate good calibration]; P value for lack of fit 0.88.
The Hypertension Risk metric is a simple, effective tool to detect early signs of increased blood pressure and helps facilitate the proactive management of high-risk individuals with a risk of hypertension. Small lifestyle changes, such as reducing caffeine intake, can be recommended to prevent progression when a low risk is detected.
The metric can also serve as an early warning for hypertension and could prompt an individual to keep closer tabs on their blood pressure levels or measure it more regularly. Since elevated blood pressure often has no directly noticeable symptoms, it can remain under the radar, and this metric can help raise awareness if it is present.
Hypertension Risk is returned as a percentage – float between 0 and 100 – and can be requested as a Widget or via the Rest API.
Name | Unit | Range | programmatic name | Health Profile |
|---|---|---|---|---|
Hypertension Risk | Percentage [%] | 0 - 100 | hypertension_risk | Physical Health |
Value | Meaning | Explanation | Example user text |
|---|---|---|---|
<5% | Low | Minimal risk, associated with a probability of less than 5% of developing hypertension within the next year. | Maintain healthy habits to keep your risk low. |
5% – 10% | Elevated | Increased risk, up to twice that of the low-risk group, with a 5-10% probability of developing hypertension within a year. | Consider learning more about your blood pressure levels, e.g. by measuring it on a regular basis with a blood pressure monitor for some time. Consider lifestyle changes if your blood pressure is elevated. If your BP is elevated, lifestyle changes (e.g., improved sleep, stress reduction, exercise) are recommended. |
>10% | High | Significantly increased risk, associated with a probability greater than 10% of developing hypertension within a year. | If you haven't already, make sure to gain a better understanding of your blood pressure levels by doing regular measurements with a blood pressure monitor. If you notice your levels are in the hypertensive range, consult your GP or another healthcare professional for guidance. |
The table above shows
- Value: The possible values for the health insight
- Meaning: What does the value mean
- Explanation: how to interpret this value
- Example user text: An example of what could be communicated to the user in case this value is measured.
Below are some possible lifestyle changes that can help reduce the risk of an increased blood pressure. You can incorporate these actionables into your platform based on the Hypertension Risk metric to guide users effectively.
- Impact of diet: Reducing sodium intake can lower blood pressure risk by up to 11%.
- Impact of physical activity: Regular aerobic exercise can reduce BP risk by 15% on average.
- Stress management: Engaging in relaxation techniques can reduce BP risk by 8–12%.
- Impact of weight loss: Losing excess weight can lower blood pressure, with studies indicating reductions of approximately 1 mm Hg for every kilogram.
- Getting adequate sleep (7–9 hours per night) is associated with improved blood pressure regulation and a lower risk of hypertension.
- Limiting alcohol intake can reduce blood pressure risk, with moderate drinking leading to lower BP compared to excessive consumption
- Lenfant, C., Chobanian, A. V., Jones, D. W., & Roccella, E. J. (2003). Seventh report of the Joint National Committee on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC 7) resetting the hypertension sails. Circulation, 107(24), 2993-2994.
- Allender, P. S., Cutler, J. A., Follmann, D., Cappuccio, F. P., Pryer, J., & Elliott, P. (1996). Dietary calcium and blood pressure: a meta-analysis of randomized clinical trials. Annals of internal medicine, 124(9), 825-831.
- Pescatello, L. S., Franklin, B. A., Fagard, R., Farquhar, W. B., Kelley, G. A., & Ray, C. A. (2004). Exercise and hypertension. Medicine & science in sports & exercise, 36(3), 533-553.
- Brook, R. D., Appel, L. J., Rubenfire, M., Ogedegbe, G., Bisognano, J. D., Elliott, W. J., ... & Rajagopalan, S. (2013). Beyond medications and diet: alternative approaches to lowering blood pressure: a scientific statement from the American Heart Association. Hypertension, 61(6), 1360-1383.
- Bock, J. M., Vungarala, S., Covassin, N., & Somers, V. K. (2022). Sleep duration and hypertension: epidemiological evidence and underlying mechanisms. American journal of hypertension, 35(1), 3-11.Chicago
- Dawber TR, Kannel WB, Lyell LP. An approach to longitudinal studies in a community: the Framingham Study. Ann N Y Acad Sci. 1963;107:539-56.
- Parikh NI, Pencina MJ, Wang TJ, Benjamin EJ, Lanier KJ, Levy D, D'Agostino RB Sr, Kannel WB, Vasan RS. A risk score for predicting near-term incidence of hypertension: the Framingham Heart Study. Ann Intern Med. 2008 Jan 15;148(2):102-10.