ImportanceClinicians use blood pressure (BP) readings obtained during clinical encounters to detect hypertension and determine the adequacy of treatment. Variations in office-based BP measurements may obscure a hypertension diagnosis or overwhelm a signal of treatment response.ObjectivesTo quantify visit-to-visit variability (VVV) in BP values and its association with patient factors in real-world practice.Design, Setting and, ParticipantsRetrospective cohort analysis of adult patients (age ≥18 years) with at least two outpatient visits in the Yale-New Haven Health System between January 1, 2014 to October 31, 2018.Main Outcome and MeasuresPatient-level measures of VVV included standard deviation (SD) and coefficient of variation (CV) of a given patient’s BP across visits. We introduced a metric to determine the VVV between any two visits (dyad) to characterize the BP information that clinicians have as they formulate their recommendations. Dyad-level measures of VVV included difference, absolute difference, standardized difference, and absolute standardized difference between the two visits of a dyad.ResultsThe study population included 537,245 adults, with a total of 7,721,864 BP measurements. The mean age was 53.4 years (SD of 19.0), 60.4% were women, 69.4% were non-Hispanic White, and 18.1% with hypertension treatment. At the patient level, the mean intra-individual SD and CV were 10.6 mmHg and 0.08 mmHg. At the dyad level, the mean difference, absolute difference, standardized difference, and absolute standardized difference were -0.7 mmHg, 11.6 mmHg, 0 mmHg, and 0.09 mmHg, respectively. Given the observed VVV, if an antihypertensive medication truly reduced a patient’s SBP by 10 mmHg (the average BP-lowering effect reported in previous review), clinicians would expect to observe a reduction of SBP < 5mm Hg at the next visit 36.9% of the time. In the multivariable linear regression model, only 2% of the variance in absolute standardized difference was attributable to patient characteristics.Conclusions and RelevanceThe large VVV poses challenges for diagnosis, treatment, and monitoring of patients with hypertension based on BP readings in outpatient settings, supporting recent guidelines recommending home BP monitoring and ambulatory BP monitoring as out-of-office alternatives to establish diagnosis of hypertension and BP control.KEY POINTSQuestionWhat is the visit-to-visit variability (VVV) in blood pressure (BP) values and its association with patient factors in real-world practice?FindingsIn this retrospective cohort analysis that included 537,245 adults and 7,721,864 office-based BP measurements from a large health system, marked VVV was observed in BP values and the median absolute change between two consecutive visits was about 12 mmHg. The VVV was not significantly associated with patient demographic and clinical characteristics.MeaningThe large VVV poses challenges for diagnosis, treatment, and monitoring of patients with hypertension based on BP readings in outpatient settings, supporting recent guidelines recommending home BP monitoring and ambulatory BP monitoring as out-of-office alternatives to establish a new diagnosis of hypertension and BP control.