User login
The Accuracy of Community-Based Automated Blood Pressure Machines
METHODS: We randomly selected 25 pharmacies and compared blood pressure readings obtained from their automated machines with from a mercury manometer. We used 3 volunteers with arm circumferences at the low, medium, and high ends of the acceptable range of a normal adult cuff size.
RESULTS: For the subject with the small arm size, store machines reported systolic pressure readings that were, on average, 10 mm Hg higher than those obtained by the clinician (P <.001) and diastolic pressures 9 mm Hg higher (P <.001). The mean systolic pressure readings for the subject with the medium arm size were not significantly different between the store machine and the mercury manometer, and the readings were only modestly different for diastolic pressure. For the subject with the large arm size, store machines reported diastolic pressure readings that were, on average, 8.3 mm Hg lower than those obtained using the mercury manometer (P <.001), but with no significant difference in the systolic pressure.
CONCLUSIONS: We found that automated blood pressure machines from a representative community-based sample of pharmacies did not meet the accepted standards of accuracy and reliability. Accuracy of readings is especially uncertain for patients having arm sizes larger or smaller than average.
Hypertension affects approximately 50 million adults in the United States.1 Frequent blood pressure readings are needed for diagnosing disorders and for monitoring patients’ health status. Self-monitoring may provide a more accurate record of patients’ blood pressure if they have white-coat syndrome and can be used to construct a record of the patient’s response to medications and to improve compliance.
Patients can conveniently monitor their blood pressure by using free automated machines found in pharmacies. Despite their widespread availability, the accuracy of these machines has been questioned.2-6 The Association for the Advancement of Medical Instrumentation (AAMI) published standards for the accuracy of blood pressure measuring devices that state, “for systolic and diastolic pressures, treated separately, the mean difference of the paired measurements of the test system and the comparison system shall be ±5 mm Hg or less, with a standard deviation of 8 mm Hg or less.”7 We sought to determine if the readings obtained from the automated machines available in the community meet the recommended standards and whether these machines are accurate and reliable when compared with the mercury manometer. We also sought to determine if arm size would affect the accuracy of automated machine readings.
Methods
Between June and November 1998, we randomly selected 25 of the 80 pharmacies within 7 miles of the University medical clinics. On the basis of previously reported variances we calculated that 25 stores would be required to provide approximately 80% power to detect a true blood pressure difference of 5 mm Hg between machine and clinician.2 We used 3 volunteers with arm circumferences at the low (26 cm), medium (29 cm), and high (33 cm) ends of the acceptable range of normal adult cuff size. Standard cuff sizes allow for arm circumferences between 26 and 33 cm.8 None of the subjects was known to be hypertensive, and none was taking antihypertensive medications or medications known to increase blood pressure. All readings were taken during a 2-month period.
Mercury manometer readings were obtained using the standardized technique and a standard cuff recommended by the American Heart Association.8 One member of the research team obtained all mercury manometer readings. The subjects obtained store-derived blood pressure readings by following each machine’s printed instructions. The model and manufacturer of each machine were recorded, and the pharmacy staff was asked about the machine’s most recent calibration or maintenance.
Blood pressures were measured after the subjects had a 5-minute rest. Six measurements were recorded for each subject in random order: 3 readings with the store blood pressure machine and 3 with the mercury manometer. Readings were taken from the same arm with a 1-minute rest between them. This is the method recommended by the AAMI and the British Hypertension Society to compare 2 methods of measurement.9
The subjects self-recorded the machine values, and the clinician was blinded to those readings. The 3 systolic and diastolic readings obtained using each method were averaged to determine the final estimate for using each source. We assessed differences between machine and clinician readings using the paired t test. The variance ratio of these readings was compared using the F statistic. We conducted analyses separately by arm size and for systolic and diastolic readings. All reported P values are 2-tailed with statistical significance set at an a of 0.05.
Results
The most common machines in pharmacies were Vita-Stat models 90550 (n=11), 8000 (n=9), and 9000 (n=3). Two stores had Health Clinic Cardio-Analysis devices. Pharmacy staff were unable to state when each machine was last calibrated or had received maintenance, indicating that responsibility for maintaining these machines fell solely on the manufacturer’s representative. Calls to the information numbers on the machines were met with nonspecific answers that the machines are calibrated regularly.
The Table 1 shows each subject’s average systolic and diastolic pressure as determined by the machine and the clinician and the variance between readings using the same method. For the subject with the small arm size, store machines reported systolic values that averaged 10 mm Hg higher than those from the mercury manometer (P <.001) and diastolic pressures that were 9 mm Hg higher (P <.001). The mean systolic pressures for the subject with the medium-sized arm were not significantly different between store and clinician methods and were only modestly different for diastolic readings. For the subject with the large arm size, store machines reported diastolic pressures that averaged 8.3 mm Hg lower than those obtained by the clinician P <.001). The difference in systolic readings between the 2 methods for this subject was not significant. Standard deviations of the store mercury blood pressure differences were greater than the 8.0 mm Hg allowable limit set by the AAMI for systolic readings for the subject with a medium arm and for the diastolic readings for the subject a large arm.7
To determine if there is more variability in readings from store machines than with those from a clinician, we also compared the variance of readings for each method . For the subject with medium arm size, there was much greater variance with store machines, compared with that from the mercury manometer, for both systolic (P=.001) and diastolic readings (P=.002). For the subject with the large arm size, there was significantly more variance in diastolic pressures obtained by store machines than with the clinician’s readings (P=.03).
Discussion
We found that automated blood pressure machines from a representative community-based sample of pharmacies did not meet accepted standards of accuracy and reliability of measurement. Depending on the subject’s arm size, automated blood pressure machines significantly underestimated or overestimated the subject’s blood pressure compared with the reading obtained by a clinician with a standard mercury manometer. For the subject with the medium arm size, the automated machines were generally more accurate but less reliable (higher variance). For the subject with the large arm size, automated machines tended to both underestimate diastolic pressure and show an unacceptably large variance, and for the subject with a small arm size, automated machines overestimated both systolic and diastolic blood pressure.
Although previous limited studies have found automated blood pressure machines less reliable than manometers, our study was unique in several ways.2-6 Unlike previous studies, the same 3 volunteers were used for all readings, rather than random passersby. Each subject also had a different arm size to assess any potential effect on readings. Also, our study used a random selection of the machines that are available and widely used in the community, rather than a few specific machines and models.
The magnitude of difference between the 2 methods of blood pressure measurement was substantial and clinically meaningful. Chronic underestimation of blood pressure will result in inadequate treatment, while overestimation can lead to unnecessary concern by the patient and overtreatment by the physician.
Limitations
Our study has a few limitations that should be considered. The reason for the variation of automated blood pressure machine accuracy with arm size is uncertain and deserves further study. During the study it was discovered that the subject with the large arm size was hypertensive, and the findings attributed to large arm size may be due to a difference in accuracy for subjects with and without hypertension. Previous studies of these machines showed no differences in accuracy between patients with and without hypertension, however.2,6 Although the clinician was blinded to the machine readings, there is a possibility of recall bias when the same clinician measured the same subject’s blood pressure repeatedly. We were not able to assess the degree to which automated machines were maintained or calibrated, so we are unable to comment on the role that machine maintenance may have on the results. Nonetheless, these are the machines used frequently by patients in the real world and not the best-case scenario submitted by the manufacturer or factory. We were unable to compare the relative accuracy of specific brands or models because of the size of our sample, and that may merit further study.
These findings are not meant to imply that all or even most automated machines were inaccurate but that using automated machines as a general method of assessing blood pressure is inaccurate. Unfortunately, it is impossible for physicians or patients to know which of the many automated blood pressure machines available in their community are accurate or reliable.
Conclusions
We found that automated blood pressure machines from a representative community-based sample did not meet accepted standards of accuracy and reliability of measurement. Accuracy of readings was especially uncertain for patients having larger- or smaller-than-average arm sizes. Even patients with a medium-sized arm can expect significant and unacceptable variability in blood pressure readings, even with the same machine. We agree with the recommendations of the American College of Physicians, which caution against managing patients solely using blood pressure readings obtained with self-monitoring blood pressure devices.10 Although further study in this area may be useful to strengthen this recommendation, we recommend that patients not rely on the results of in-store automated blood pressure devices. The manufacturers of these devices should also monitor the accuracy and reliability of these devices frequently and make maintenance information easily available.
Acknowledgments
Dr Roetzheim was supported through a Generalist Physician Faculty Scholars Award from the Robert Wood Johnson Foundation. Marney Goldstein is a student at the University of South Florida College of Medicine and was supported by a summer research assistantship sponsored by the Florida Academy of Family Physicians and the American Academy of Family Physicians. We also wish to thank Christopher Roever for his assistance in the collection of the data.
1. The sixth report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 1997;157:2413-46.
2. Whitcomb B, Prochazka A, LoVerde M, Byyny R. Failure of the community-based Vita-Stat automated blood pressure device to accurately measure blood pressure. Arch Fam Med 1995;4:419-24.
3. Salaita K, Whelton P, Seidler A. A community-based evaluation of the Vita-Stat automatic blood pressure recorder. Am J Hypertens 1990;3:366-72.
4. Polk B, Rosner B, Feudo R, Vandenburg D. An evaluation of the Vita-Stat automatic blood pressure measuring device. Hypertension 1980;2:221-7.
5. Whelton P, Thompson S, Barnes G, Miall W. Evaluation of the Vita-Stat automatic blood pressure recorder: a comparison with the random-zero sphygmomanometer. Am J Epidemiol 1983;117:46-54.
6. Berkson D, Whipple I, Shireman L, Brown M, Raynor W, Shekelle R. Evaluation of an automated blood pressure measuring device intended for general public use. Am J Public Health 1979;69:473-9.
7. White W, Berson A, Robbins C, et al. National standard for measurement of resting and ambulatory blood pressures with automated sphygmomanometers. Hypertension 1993;21:504-9.
8. Frohlich E, Grim C, Labarthe D, Maxwell M, Perloff D, Weidman W. Recommendations for human blood pressure determination by sphygmomanometers: report of a special task force appointed by the steering committee, American Heart Association. Hypertension 1988;11:210A-21A.
9. O’Brien E, Petrie J, Littler W, et al. The British hypertension society protocol for the evaluation of automated and semi-automated blood pressure measuring devices with special reference to ambulatory systems. J Hypertension 1990;8:607-19.
10. Appel L, Stason W. Ambulatory blood pressure monitoring and blood pressure self-measurement in the diagnosis and management of hypertension. Ann Intern Med 1993;118:867-82.
METHODS: We randomly selected 25 pharmacies and compared blood pressure readings obtained from their automated machines with from a mercury manometer. We used 3 volunteers with arm circumferences at the low, medium, and high ends of the acceptable range of a normal adult cuff size.
RESULTS: For the subject with the small arm size, store machines reported systolic pressure readings that were, on average, 10 mm Hg higher than those obtained by the clinician (P <.001) and diastolic pressures 9 mm Hg higher (P <.001). The mean systolic pressure readings for the subject with the medium arm size were not significantly different between the store machine and the mercury manometer, and the readings were only modestly different for diastolic pressure. For the subject with the large arm size, store machines reported diastolic pressure readings that were, on average, 8.3 mm Hg lower than those obtained using the mercury manometer (P <.001), but with no significant difference in the systolic pressure.
CONCLUSIONS: We found that automated blood pressure machines from a representative community-based sample of pharmacies did not meet the accepted standards of accuracy and reliability. Accuracy of readings is especially uncertain for patients having arm sizes larger or smaller than average.
Hypertension affects approximately 50 million adults in the United States.1 Frequent blood pressure readings are needed for diagnosing disorders and for monitoring patients’ health status. Self-monitoring may provide a more accurate record of patients’ blood pressure if they have white-coat syndrome and can be used to construct a record of the patient’s response to medications and to improve compliance.
Patients can conveniently monitor their blood pressure by using free automated machines found in pharmacies. Despite their widespread availability, the accuracy of these machines has been questioned.2-6 The Association for the Advancement of Medical Instrumentation (AAMI) published standards for the accuracy of blood pressure measuring devices that state, “for systolic and diastolic pressures, treated separately, the mean difference of the paired measurements of the test system and the comparison system shall be ±5 mm Hg or less, with a standard deviation of 8 mm Hg or less.”7 We sought to determine if the readings obtained from the automated machines available in the community meet the recommended standards and whether these machines are accurate and reliable when compared with the mercury manometer. We also sought to determine if arm size would affect the accuracy of automated machine readings.
Methods
Between June and November 1998, we randomly selected 25 of the 80 pharmacies within 7 miles of the University medical clinics. On the basis of previously reported variances we calculated that 25 stores would be required to provide approximately 80% power to detect a true blood pressure difference of 5 mm Hg between machine and clinician.2 We used 3 volunteers with arm circumferences at the low (26 cm), medium (29 cm), and high (33 cm) ends of the acceptable range of normal adult cuff size. Standard cuff sizes allow for arm circumferences between 26 and 33 cm.8 None of the subjects was known to be hypertensive, and none was taking antihypertensive medications or medications known to increase blood pressure. All readings were taken during a 2-month period.
Mercury manometer readings were obtained using the standardized technique and a standard cuff recommended by the American Heart Association.8 One member of the research team obtained all mercury manometer readings. The subjects obtained store-derived blood pressure readings by following each machine’s printed instructions. The model and manufacturer of each machine were recorded, and the pharmacy staff was asked about the machine’s most recent calibration or maintenance.
Blood pressures were measured after the subjects had a 5-minute rest. Six measurements were recorded for each subject in random order: 3 readings with the store blood pressure machine and 3 with the mercury manometer. Readings were taken from the same arm with a 1-minute rest between them. This is the method recommended by the AAMI and the British Hypertension Society to compare 2 methods of measurement.9
The subjects self-recorded the machine values, and the clinician was blinded to those readings. The 3 systolic and diastolic readings obtained using each method were averaged to determine the final estimate for using each source. We assessed differences between machine and clinician readings using the paired t test. The variance ratio of these readings was compared using the F statistic. We conducted analyses separately by arm size and for systolic and diastolic readings. All reported P values are 2-tailed with statistical significance set at an a of 0.05.
Results
The most common machines in pharmacies were Vita-Stat models 90550 (n=11), 8000 (n=9), and 9000 (n=3). Two stores had Health Clinic Cardio-Analysis devices. Pharmacy staff were unable to state when each machine was last calibrated or had received maintenance, indicating that responsibility for maintaining these machines fell solely on the manufacturer’s representative. Calls to the information numbers on the machines were met with nonspecific answers that the machines are calibrated regularly.
The Table 1 shows each subject’s average systolic and diastolic pressure as determined by the machine and the clinician and the variance between readings using the same method. For the subject with the small arm size, store machines reported systolic values that averaged 10 mm Hg higher than those from the mercury manometer (P <.001) and diastolic pressures that were 9 mm Hg higher (P <.001). The mean systolic pressures for the subject with the medium-sized arm were not significantly different between store and clinician methods and were only modestly different for diastolic readings. For the subject with the large arm size, store machines reported diastolic pressures that averaged 8.3 mm Hg lower than those obtained by the clinician P <.001). The difference in systolic readings between the 2 methods for this subject was not significant. Standard deviations of the store mercury blood pressure differences were greater than the 8.0 mm Hg allowable limit set by the AAMI for systolic readings for the subject with a medium arm and for the diastolic readings for the subject a large arm.7
To determine if there is more variability in readings from store machines than with those from a clinician, we also compared the variance of readings for each method . For the subject with medium arm size, there was much greater variance with store machines, compared with that from the mercury manometer, for both systolic (P=.001) and diastolic readings (P=.002). For the subject with the large arm size, there was significantly more variance in diastolic pressures obtained by store machines than with the clinician’s readings (P=.03).
Discussion
We found that automated blood pressure machines from a representative community-based sample of pharmacies did not meet accepted standards of accuracy and reliability of measurement. Depending on the subject’s arm size, automated blood pressure machines significantly underestimated or overestimated the subject’s blood pressure compared with the reading obtained by a clinician with a standard mercury manometer. For the subject with the medium arm size, the automated machines were generally more accurate but less reliable (higher variance). For the subject with the large arm size, automated machines tended to both underestimate diastolic pressure and show an unacceptably large variance, and for the subject with a small arm size, automated machines overestimated both systolic and diastolic blood pressure.
Although previous limited studies have found automated blood pressure machines less reliable than manometers, our study was unique in several ways.2-6 Unlike previous studies, the same 3 volunteers were used for all readings, rather than random passersby. Each subject also had a different arm size to assess any potential effect on readings. Also, our study used a random selection of the machines that are available and widely used in the community, rather than a few specific machines and models.
The magnitude of difference between the 2 methods of blood pressure measurement was substantial and clinically meaningful. Chronic underestimation of blood pressure will result in inadequate treatment, while overestimation can lead to unnecessary concern by the patient and overtreatment by the physician.
Limitations
Our study has a few limitations that should be considered. The reason for the variation of automated blood pressure machine accuracy with arm size is uncertain and deserves further study. During the study it was discovered that the subject with the large arm size was hypertensive, and the findings attributed to large arm size may be due to a difference in accuracy for subjects with and without hypertension. Previous studies of these machines showed no differences in accuracy between patients with and without hypertension, however.2,6 Although the clinician was blinded to the machine readings, there is a possibility of recall bias when the same clinician measured the same subject’s blood pressure repeatedly. We were not able to assess the degree to which automated machines were maintained or calibrated, so we are unable to comment on the role that machine maintenance may have on the results. Nonetheless, these are the machines used frequently by patients in the real world and not the best-case scenario submitted by the manufacturer or factory. We were unable to compare the relative accuracy of specific brands or models because of the size of our sample, and that may merit further study.
These findings are not meant to imply that all or even most automated machines were inaccurate but that using automated machines as a general method of assessing blood pressure is inaccurate. Unfortunately, it is impossible for physicians or patients to know which of the many automated blood pressure machines available in their community are accurate or reliable.
Conclusions
We found that automated blood pressure machines from a representative community-based sample did not meet accepted standards of accuracy and reliability of measurement. Accuracy of readings was especially uncertain for patients having larger- or smaller-than-average arm sizes. Even patients with a medium-sized arm can expect significant and unacceptable variability in blood pressure readings, even with the same machine. We agree with the recommendations of the American College of Physicians, which caution against managing patients solely using blood pressure readings obtained with self-monitoring blood pressure devices.10 Although further study in this area may be useful to strengthen this recommendation, we recommend that patients not rely on the results of in-store automated blood pressure devices. The manufacturers of these devices should also monitor the accuracy and reliability of these devices frequently and make maintenance information easily available.
Acknowledgments
Dr Roetzheim was supported through a Generalist Physician Faculty Scholars Award from the Robert Wood Johnson Foundation. Marney Goldstein is a student at the University of South Florida College of Medicine and was supported by a summer research assistantship sponsored by the Florida Academy of Family Physicians and the American Academy of Family Physicians. We also wish to thank Christopher Roever for his assistance in the collection of the data.
METHODS: We randomly selected 25 pharmacies and compared blood pressure readings obtained from their automated machines with from a mercury manometer. We used 3 volunteers with arm circumferences at the low, medium, and high ends of the acceptable range of a normal adult cuff size.
RESULTS: For the subject with the small arm size, store machines reported systolic pressure readings that were, on average, 10 mm Hg higher than those obtained by the clinician (P <.001) and diastolic pressures 9 mm Hg higher (P <.001). The mean systolic pressure readings for the subject with the medium arm size were not significantly different between the store machine and the mercury manometer, and the readings were only modestly different for diastolic pressure. For the subject with the large arm size, store machines reported diastolic pressure readings that were, on average, 8.3 mm Hg lower than those obtained using the mercury manometer (P <.001), but with no significant difference in the systolic pressure.
CONCLUSIONS: We found that automated blood pressure machines from a representative community-based sample of pharmacies did not meet the accepted standards of accuracy and reliability. Accuracy of readings is especially uncertain for patients having arm sizes larger or smaller than average.
Hypertension affects approximately 50 million adults in the United States.1 Frequent blood pressure readings are needed for diagnosing disorders and for monitoring patients’ health status. Self-monitoring may provide a more accurate record of patients’ blood pressure if they have white-coat syndrome and can be used to construct a record of the patient’s response to medications and to improve compliance.
Patients can conveniently monitor their blood pressure by using free automated machines found in pharmacies. Despite their widespread availability, the accuracy of these machines has been questioned.2-6 The Association for the Advancement of Medical Instrumentation (AAMI) published standards for the accuracy of blood pressure measuring devices that state, “for systolic and diastolic pressures, treated separately, the mean difference of the paired measurements of the test system and the comparison system shall be ±5 mm Hg or less, with a standard deviation of 8 mm Hg or less.”7 We sought to determine if the readings obtained from the automated machines available in the community meet the recommended standards and whether these machines are accurate and reliable when compared with the mercury manometer. We also sought to determine if arm size would affect the accuracy of automated machine readings.
Methods
Between June and November 1998, we randomly selected 25 of the 80 pharmacies within 7 miles of the University medical clinics. On the basis of previously reported variances we calculated that 25 stores would be required to provide approximately 80% power to detect a true blood pressure difference of 5 mm Hg between machine and clinician.2 We used 3 volunteers with arm circumferences at the low (26 cm), medium (29 cm), and high (33 cm) ends of the acceptable range of normal adult cuff size. Standard cuff sizes allow for arm circumferences between 26 and 33 cm.8 None of the subjects was known to be hypertensive, and none was taking antihypertensive medications or medications known to increase blood pressure. All readings were taken during a 2-month period.
Mercury manometer readings were obtained using the standardized technique and a standard cuff recommended by the American Heart Association.8 One member of the research team obtained all mercury manometer readings. The subjects obtained store-derived blood pressure readings by following each machine’s printed instructions. The model and manufacturer of each machine were recorded, and the pharmacy staff was asked about the machine’s most recent calibration or maintenance.
Blood pressures were measured after the subjects had a 5-minute rest. Six measurements were recorded for each subject in random order: 3 readings with the store blood pressure machine and 3 with the mercury manometer. Readings were taken from the same arm with a 1-minute rest between them. This is the method recommended by the AAMI and the British Hypertension Society to compare 2 methods of measurement.9
The subjects self-recorded the machine values, and the clinician was blinded to those readings. The 3 systolic and diastolic readings obtained using each method were averaged to determine the final estimate for using each source. We assessed differences between machine and clinician readings using the paired t test. The variance ratio of these readings was compared using the F statistic. We conducted analyses separately by arm size and for systolic and diastolic readings. All reported P values are 2-tailed with statistical significance set at an a of 0.05.
Results
The most common machines in pharmacies were Vita-Stat models 90550 (n=11), 8000 (n=9), and 9000 (n=3). Two stores had Health Clinic Cardio-Analysis devices. Pharmacy staff were unable to state when each machine was last calibrated or had received maintenance, indicating that responsibility for maintaining these machines fell solely on the manufacturer’s representative. Calls to the information numbers on the machines were met with nonspecific answers that the machines are calibrated regularly.
The Table 1 shows each subject’s average systolic and diastolic pressure as determined by the machine and the clinician and the variance between readings using the same method. For the subject with the small arm size, store machines reported systolic values that averaged 10 mm Hg higher than those from the mercury manometer (P <.001) and diastolic pressures that were 9 mm Hg higher (P <.001). The mean systolic pressures for the subject with the medium-sized arm were not significantly different between store and clinician methods and were only modestly different for diastolic readings. For the subject with the large arm size, store machines reported diastolic pressures that averaged 8.3 mm Hg lower than those obtained by the clinician P <.001). The difference in systolic readings between the 2 methods for this subject was not significant. Standard deviations of the store mercury blood pressure differences were greater than the 8.0 mm Hg allowable limit set by the AAMI for systolic readings for the subject with a medium arm and for the diastolic readings for the subject a large arm.7
To determine if there is more variability in readings from store machines than with those from a clinician, we also compared the variance of readings for each method . For the subject with medium arm size, there was much greater variance with store machines, compared with that from the mercury manometer, for both systolic (P=.001) and diastolic readings (P=.002). For the subject with the large arm size, there was significantly more variance in diastolic pressures obtained by store machines than with the clinician’s readings (P=.03).
Discussion
We found that automated blood pressure machines from a representative community-based sample of pharmacies did not meet accepted standards of accuracy and reliability of measurement. Depending on the subject’s arm size, automated blood pressure machines significantly underestimated or overestimated the subject’s blood pressure compared with the reading obtained by a clinician with a standard mercury manometer. For the subject with the medium arm size, the automated machines were generally more accurate but less reliable (higher variance). For the subject with the large arm size, automated machines tended to both underestimate diastolic pressure and show an unacceptably large variance, and for the subject with a small arm size, automated machines overestimated both systolic and diastolic blood pressure.
Although previous limited studies have found automated blood pressure machines less reliable than manometers, our study was unique in several ways.2-6 Unlike previous studies, the same 3 volunteers were used for all readings, rather than random passersby. Each subject also had a different arm size to assess any potential effect on readings. Also, our study used a random selection of the machines that are available and widely used in the community, rather than a few specific machines and models.
The magnitude of difference between the 2 methods of blood pressure measurement was substantial and clinically meaningful. Chronic underestimation of blood pressure will result in inadequate treatment, while overestimation can lead to unnecessary concern by the patient and overtreatment by the physician.
Limitations
Our study has a few limitations that should be considered. The reason for the variation of automated blood pressure machine accuracy with arm size is uncertain and deserves further study. During the study it was discovered that the subject with the large arm size was hypertensive, and the findings attributed to large arm size may be due to a difference in accuracy for subjects with and without hypertension. Previous studies of these machines showed no differences in accuracy between patients with and without hypertension, however.2,6 Although the clinician was blinded to the machine readings, there is a possibility of recall bias when the same clinician measured the same subject’s blood pressure repeatedly. We were not able to assess the degree to which automated machines were maintained or calibrated, so we are unable to comment on the role that machine maintenance may have on the results. Nonetheless, these are the machines used frequently by patients in the real world and not the best-case scenario submitted by the manufacturer or factory. We were unable to compare the relative accuracy of specific brands or models because of the size of our sample, and that may merit further study.
These findings are not meant to imply that all or even most automated machines were inaccurate but that using automated machines as a general method of assessing blood pressure is inaccurate. Unfortunately, it is impossible for physicians or patients to know which of the many automated blood pressure machines available in their community are accurate or reliable.
Conclusions
We found that automated blood pressure machines from a representative community-based sample did not meet accepted standards of accuracy and reliability of measurement. Accuracy of readings was especially uncertain for patients having larger- or smaller-than-average arm sizes. Even patients with a medium-sized arm can expect significant and unacceptable variability in blood pressure readings, even with the same machine. We agree with the recommendations of the American College of Physicians, which caution against managing patients solely using blood pressure readings obtained with self-monitoring blood pressure devices.10 Although further study in this area may be useful to strengthen this recommendation, we recommend that patients not rely on the results of in-store automated blood pressure devices. The manufacturers of these devices should also monitor the accuracy and reliability of these devices frequently and make maintenance information easily available.
Acknowledgments
Dr Roetzheim was supported through a Generalist Physician Faculty Scholars Award from the Robert Wood Johnson Foundation. Marney Goldstein is a student at the University of South Florida College of Medicine and was supported by a summer research assistantship sponsored by the Florida Academy of Family Physicians and the American Academy of Family Physicians. We also wish to thank Christopher Roever for his assistance in the collection of the data.
1. The sixth report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 1997;157:2413-46.
2. Whitcomb B, Prochazka A, LoVerde M, Byyny R. Failure of the community-based Vita-Stat automated blood pressure device to accurately measure blood pressure. Arch Fam Med 1995;4:419-24.
3. Salaita K, Whelton P, Seidler A. A community-based evaluation of the Vita-Stat automatic blood pressure recorder. Am J Hypertens 1990;3:366-72.
4. Polk B, Rosner B, Feudo R, Vandenburg D. An evaluation of the Vita-Stat automatic blood pressure measuring device. Hypertension 1980;2:221-7.
5. Whelton P, Thompson S, Barnes G, Miall W. Evaluation of the Vita-Stat automatic blood pressure recorder: a comparison with the random-zero sphygmomanometer. Am J Epidemiol 1983;117:46-54.
6. Berkson D, Whipple I, Shireman L, Brown M, Raynor W, Shekelle R. Evaluation of an automated blood pressure measuring device intended for general public use. Am J Public Health 1979;69:473-9.
7. White W, Berson A, Robbins C, et al. National standard for measurement of resting and ambulatory blood pressures with automated sphygmomanometers. Hypertension 1993;21:504-9.
8. Frohlich E, Grim C, Labarthe D, Maxwell M, Perloff D, Weidman W. Recommendations for human blood pressure determination by sphygmomanometers: report of a special task force appointed by the steering committee, American Heart Association. Hypertension 1988;11:210A-21A.
9. O’Brien E, Petrie J, Littler W, et al. The British hypertension society protocol for the evaluation of automated and semi-automated blood pressure measuring devices with special reference to ambulatory systems. J Hypertension 1990;8:607-19.
10. Appel L, Stason W. Ambulatory blood pressure monitoring and blood pressure self-measurement in the diagnosis and management of hypertension. Ann Intern Med 1993;118:867-82.
1. The sixth report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure. Arch Intern Med 1997;157:2413-46.
2. Whitcomb B, Prochazka A, LoVerde M, Byyny R. Failure of the community-based Vita-Stat automated blood pressure device to accurately measure blood pressure. Arch Fam Med 1995;4:419-24.
3. Salaita K, Whelton P, Seidler A. A community-based evaluation of the Vita-Stat automatic blood pressure recorder. Am J Hypertens 1990;3:366-72.
4. Polk B, Rosner B, Feudo R, Vandenburg D. An evaluation of the Vita-Stat automatic blood pressure measuring device. Hypertension 1980;2:221-7.
5. Whelton P, Thompson S, Barnes G, Miall W. Evaluation of the Vita-Stat automatic blood pressure recorder: a comparison with the random-zero sphygmomanometer. Am J Epidemiol 1983;117:46-54.
6. Berkson D, Whipple I, Shireman L, Brown M, Raynor W, Shekelle R. Evaluation of an automated blood pressure measuring device intended for general public use. Am J Public Health 1979;69:473-9.
7. White W, Berson A, Robbins C, et al. National standard for measurement of resting and ambulatory blood pressures with automated sphygmomanometers. Hypertension 1993;21:504-9.
8. Frohlich E, Grim C, Labarthe D, Maxwell M, Perloff D, Weidman W. Recommendations for human blood pressure determination by sphygmomanometers: report of a special task force appointed by the steering committee, American Heart Association. Hypertension 1988;11:210A-21A.
9. O’Brien E, Petrie J, Littler W, et al. The British hypertension society protocol for the evaluation of automated and semi-automated blood pressure measuring devices with special reference to ambulatory systems. J Hypertension 1990;8:607-19.
10. Appel L, Stason W. Ambulatory blood pressure monitoring and blood pressure self-measurement in the diagnosis and management of hypertension. Ann Intern Med 1993;118:867-82.