- Email: [email protected]
Routine Blood Pressure Measurements Do Not Predict Adverse Events in Hospitalized Patients
The American Journal of Medicine (2006) 119, 70.e17-70.e22
CLINICAL RESEARCH STUDY
Routine Blood Pressure Measurements do not Predict Adverse Events in Hospitalized Patients David Conen, MD, Bernd M. Leimenstoll, MD, André P. Perruchoud, MD, Benedict Martina, MD Department of Internal Medicine, University Hospital Basel, Switzerland. ABSTRACT PURPOSE: Countless blood pressure measurements are performed every day for almost every hospitalized patient. We analyzed the value of routine blood pressure measurements on patient care in an unselected group of hospitalized patients. METHODS: The study included 639 patients who were admitted to the hospital with a broad range of medical conditions. Two independent investigators reviewed the medical charts of the patients. Routine blood pressure values were abstracted from the patient charts and evaluated with respect to the occurrence of adverse clinical events in the study group. Changes in blood pressure between the last measurement just before adverse clinical events and the mean blood pressure values 72 hours before the adverse events were calculated and compared with mean normal day-to-day variations in blood pressure. RESULTS: In every patient, a mean of 1.6 ⫾ 0.6 routine blood pressure measurements per day were performed. Of the 639 patients in the study, 122 (19%) had clinical complications. The most commonly occurring complications were gastrointestinal bleeding (n ⫽ 15), falls (n ⫽ 13), other bleeding (n ⫽ 12) and pneumonia (n ⫽ 8). In patients who experienced clinical complications, pre-event systolic and diastolic blood pressure changes of at least 10 mm Hg occurred in 41% and 24% of the group, respectively, but this was not different from the normal day-to-day variations observed in patients who had no clinical complications. The results also were similar for patients who died or who had a severe adverse event that required admission to an intensive care unit. CONCLUSION: Routine blood pressure measurements in a general hospital patient population do not predict clinical adverse events. © 2006 Elsevier Inc. All rights reserved. KEYWORDS: Blood pressure determination; Hospitals; Inpatients; Patient monitoring; Complications
Routine blood pressure measurements are performed daily for almost every hospitalized patient. Countless blood pressure values are measured and written down in patient charts, which requires a significant amount of time on the part of hospital personnel. However, there are few published studies on the value of routine blood pressure measurements in the care of hospital patients.1 The usefulness of other tools that are routinely used for the management of hospital patients, such as physical examination or cardiac auscultation, has been demonstrated in observational studies.2,3 For example, a skilled physical examination may have a sub-
Requests for reprints should be addressed to Benedict Martina, MD, University Hospital, Petersgraben 4, 4031 Basel, Switzerland. E-mail address: [email protected].
0002-9343/$ -see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2005.09.060
stantial impact on the care of hospital patients2 and represents an effective filter for the reasonable use of more expensive technology.3-5 In contrast, the mere act of being hospitalized directly influences a patient’s routine blood pressure and impairs the usefulness of these values for detecting elevated blood pressure in hypertensive patients.6-8 Furthermore, patient hospital stays are usually not long enough to correctly adjust long-term blood pressure. Therefore, the only reason for routinely measuring blood pressure in all hospital patients should be to enhance or achieve earlier detection of clinical complications or deterioration in a patient’s condition. The few studies that currently are available show that the routine measurement of vital signs has no effect on detecting deterioration at an earlier
70.e18
The American Journal of Medicine, Vol 119, No 1, January 2006
stage, on the length of patient hospital stays, or on the Patients clinical outcome in several selected subgroups of hospital A total of 749 patients were screened for inclusion. Of patients.9-11 However, these earlier studies did not assess this group, 35 patients (5%) were terminally ill; in 54 the value of routine blood pressure measurement in pre(7%), standardized blood pressure measurements were dicting in-hospital complications. Rather these studies not feasible; and for 21 (3%), the medical charts were examined the usefulness of rouincomplete. Ultimately, 639 patine blood pressure measuretients (85%) were retained for ments in the management of spefinal evaluation. CLINICAL SIGNIFICANCE cific clinical conditions for Baseline characteristics for the which patients were admitted to study population are shown in Table ● In hospitalized patients, adverse events the hospital. For example, a ret1. Mean age in years was 64 ⫾ 17, and substantial day-to-day blood presrospective survey of patients and 322 (50%) patients were male. sure changes are common. with severe thoracoabdominal The patients were hospitalized for a ● Blood pressure changes prior to an adtrauma found that one third of wide range of ailments, most frethese patients had a normal sysverse event are indistinguishable from quently acute coronary syndrome tolic blood pressure despite the normal day-to-day variations in patients (9.4%), fractures (5.3%), malignanoccurrence of major bleeding.9 cies (miscellaneous 5.3%, lung canwithout adverse events. The investigators therefore concer 2.5%, colon cancer 1.4%), pneu● Routine blood pressure measurements in cluded that normal vital signs do monia (5.2%), heart failure (4.7%), a general hospital patient population do not preclude the presence of liferenal transplantation (3.4%), exacnot predict clinical adverse events. threatening hemorrhage. Botti et erbation of chronic obstructive pulal also showed that after coromonary disease (3.1%), and bacte● Considering the large number of routine nary angiography, blood presremia (2.7%). blood pressure measurements taken, sure and heart rate were no difsubstantial resources could be freed for ferent in patients with local Procedures more useful interventions. bleeding complications from The physicians in charge of the those with no complications.11 patients prescribed the frequency Therefore, our hypothesis was of routine blood pressure meathat routine blood pressure measurements. A routine blood pressure measurement was desurement is not useful for the early detection of adverse fined as being a measurement prescribed as part of the clinical events in an unselected group of hospital patients. routine monitoring plan for the patient, independent of spe-
METHODS Study Design The study was conducted at the University Hospital Basel, a 784-bed, academic primary and tertiary care hospital in Switzerland. Patients hospitalized in the wards of the internal medicine, visceral surgery, and orthopedics departments were assessed for eligibility during the period of June 1 through December 16, 2003. For the purpose of the present study, all patients aged 18 years or older in these departments were included according to the following pattern: During one week, patients of one of the wards were included. The next week, patients of another ward were screened for eligibility. The chronological order of rotation among wards was predefined. We excluded terminally ill patients in whom diagnostic and therapeutic procedures were discontinued and also patients in whom standardized blood pressure measurements were not feasible (for example in the case of bilateral arm fracture). The local ethics committee approved the study protocol and the study was performed according to the Declaration of Helsinki. All authors participated in data analysis and interpretation. The manuscript was prepared by the authors.
cific interventions or symptoms. All routine blood pressure values for each patient during the patient’s hospital stay were entered into a database and double-checked by a second investigator. Two experienced physicians independently reviewed all patient charts for adverse clinical events according to a standardized protocol, namely, review of discharge letters, charge notes, monitoring sheets, and laboratory values. The findings of the 2 physicians were compared and any disagreement was resolved by consensus finding. The kappa value as a measure for inter-observer concordance in the detection of clinical complications was 0.79. To evaluate the usefulness of routine blood pressure measurements in predicting clinical complications, the last routine blood pressure measurement taken before an adverse event was compared with the patient’s mean blood pressure during the 72 hours preceding the event. For patients with no clinical complications, the day-to-day variations in routine blood pressure measurement values were determined. We did this by randomly choosing a day between day 1 and day 10 (day 4) and a day between day 11 and day 20 (day 11) of the hospital stay. We then compared the mean values obtained on those days with the mean values on the preceding day (ie, day 4 compared with day 3 and day 11 compared with day 10).
Conen et al Table 1
Routine Blood Pressure in Hospitalized Patients
Reasons for admitting patients N ⫽ 639
Infections Pneumonia (%) Exacerbation of COPD (%) Bacteremia (%) Miscellaneous infections (%) Osteomyelitis (%) Endocarditis (%) Abscess (%) Urinary tract infection (%) Cardiovascular disease Acute coronary syndrome (%) Heart failure (%) Valvular heart disease (%) Peripheral artery disease (%) Pulmonary embolism (%) Malignancies Miscellaneous malignancies (%) Lung cancer (%) Hematological malignancies (%) Colon cancer (%) Musculoskeletal system disorder Fractures (%) Prosthetic joint implantation (%) Osteoarthritis (%) Discal herniation (%) Gastrointestinal disorder Gastrointestinal bleeding (%) Diverticulitis (%) Colitis (%) Cholecystitis (%) Liver cirrhosis (%) Diarrhea (%) Miscellaneous conditions Renal transplantation (%) Falls (%) Alcoholism (%) Syncope (%) Diabetes (%) Other diagnoses (%)
33 20 17 15 15 10 10 7
(5.2) (3.1) (2.7) (2.3) (2.3) (1.6) (1.6) (1.1)
60 30 9 7 6
(9.4) (4.7) (1.4) (1.1) (0.9)
33 16 11 9
(5.2) (2.5) (1.7) (1.4)
34 11 9 8
(5.3) (1.7) (1.4) (1.3)
10 8 8 8 7 7
(1.6) (1.3) (1.3) (1.3) (1.1) (1.1)
22 6 6 5 5 177
(3.4) (0.9) (0.9) (0.8) (0.8) (27.7)
Data are number of patients (%). Because of rounding, the percentages might not equal 100. COPD ⫽ chronic obstructive pulmonary disease.
70.e19 treatment were included in the analysis. Also included in the analysis were: the death of a patient, any documented thrombotic or ischemic event, any exacerbation of heart failure that resulted in additional treatment, any symptomatic and documented cardiac arrhythmias, any acute renal failure, any witnessed syncope or seizure, any documented fall, and any complication that arose from a procedure in which routine postprocedural blood pressure monitoring usually is not performed (eg, lumbar puncture, central venous catheters).
Statistical Analysis Assuming a relative frequency of 40% (10%) for blood pressure variations of at least 10 (20) mm Hg, the present sample sizes allowed us to detect a true 14% (9%) difference in the respective frequencies between the patient populations with and without complications with a probability of 80% and a significance level of 5%. Continuous variables were expressed as mean ⫾ standard deviation. Kappa statistics were used to assess the interobserver concordance for the detection of clinical complications through independent chart review. To assess day-to-day variability in patients who did not experience clinical complications, absolute blood pressure differences were calculated for mean blood pressure on day 4 versus day 3 and mean blood pressure on day 11 versus day 10. In patients who experienced a clinical complication, blood pressure differences were calculated for the last blood pressure value taken just before the event versus mean blood pressure 72 hours before the event. The proportion of patients with an absolute blood pressure difference of at least 10 mm Hg and 20 mm Hg between these days was calculated. Differences in proportions were assessed using Fisher’s exact test. A 2-tailed P value of less than .05 was considered to indicate statistical significance. Pearson correlation coefficients, mean differences, standard deviation of the differences (SDD), and coefficients of variation (SDD divided by mean blood pressure) were calculated. All analyses were performed using SPSS software version 11.5 (SPSS Inc, Chicago, Ill).
RESULTS Adverse Clinical Events An adverse clinical event was defined as a complication that arose during the patient’s hospital stay, including those complications with a causal relationship to the admission diagnosis. Adverse events that arose during a stay in the emergency room or in the intensive care unit or within the first 24 hours after surgery were excluded. Any event that led to an unplanned admission to the intensive care unit was included in the analysis. Significant bleeding episodes were defined as bleeding that required the transfusion of at least two units of blood, any intracranial hemorrhage, any gastrointestinal bleeding other than hemorrhoidal, any bleeding that necessitated surgical intervention or admission to the intensive care unit, and disabling bleeding from any site. All documented infections that required medical or surgical
We detected adverse events in 122 patients (19%) (Table 2). The most frequent clinical adverse events were gastrointestinal bleeding (12.3%), falls (10.7%), other bleeding (9.8%), pneumonia (6.6%), and bacteremia (6.6%). In 27 cases (22%), the patient was admitted to an intensive care unit. Among those admitted to an intensive care unit, heart failure (22%), gastrointestinal bleeding (19%), arrhythmias (19%), and other bleedings (15%) were the most frequent complications (Table 2). Each day, the mean number of routine blood pressure readings per patient was 1.6 ⫾ 0.6. As shown in Table 3, the proportion of patients with a blood pressure change of at least 10 mm Hg just before a clinical event was 40.7% for systolic blood pressure and 23.7% for diastolic blood pressure. The proportion of patients with a blood pressure dif-
70.e20
The American Journal of Medicine, Vol 119, No 1, January 2006
Table 2 Type and number of clinical complications in the study group Number of Number of patients admitted to patients (n ⫽ 122) ICU (n ⫽ 27) GI bleeding (%) Falls (%) Bleeding (other than GI) (%) Pneumonia (%) Bacteremia (%) Heart failure (%) Acute renal failure (%) Arrhythmias (%) Death (%) Febrile urinary tract infection (%) DVT/pulmonary embolism (%) Empyema (%) Syncope (%) Acute urinary retention (%) Acute coronary syndrome (%) Abscess (%) Other infections (%) Other complications (%)*
15 13 12 8 8 6 5 5 4 4 4 4 4 3 3 3 7 14
(12.3) (10.7) (9.8) (6.6) (6.6) (4.9) (4.1) (4.1) (3.3) (3.3) (3.3) (3.3) (3.3) (2.5) (2.5) (2.5) (5.7) (11.5)
5 4 2 6 1 5 1
2 1
Data are number of patients (%). Because of rounding, % might not equal 100. GI ⫽ gastrointestinal; ICU ⫽ intensive care unit; DVT ⫽ deep venous thrombosis. *Other complications included transient ischemic attack (2), pneumothorax (2), ileus (2), seizure (2), hypoglycemia with neurological manifestations, allergic drug reaction, thrombopenia, dyspnea, decrease in oxygen saturation (with admission to ICU), and radiculary syndrome after lumbar puncture.
ference of at least 20 mm Hg just before an event was 10.2% for systolic blood pressure and 5.1% for diastolic blood pressure. These changes were very similar to the normal day-to-day variability in blood pressure in the study population, as illustrated by the proportion of patients who had no clinical complications and who showed the same abso-
Table 3
lute mean blood pressure changes between day 3 and day 4 and between day 10 and day 11 of their hospital stay (Table 3). None of these comparisons between the two patient groups reached statistical significance (all P values ⬎.27). When only patients who experienced clinical complications are considered, the proportion of patients with a difference in systolic blood pressure of at least 10 mm Hg was 51.7% between day 3 and day 4 and 42.9% between day 10 and day 11. The proportion with a difference in systolic blood pressure of at least 20 mm Hg was 17.2% between day 3 and day 4, and 16.3% between day 10 and day 11. These percentages were not significantly different from those related to blood pressure changes just before a clinical event (all P values ⬎.25) (Table 3). When only death or severe complications (n ⫽ 31) that led to admission to an intensive care unit are considered, the proportion of patients with a blood pressure change of at least 10 mm Hg before the event was 30.0% for systolic blood pressure and 33.3% for diastolic blood pressure. Only 1 patient who was admitted to the intensive care unit and 1 patient who subsequently died had an event-related systolic blood pressure change of at least 20 mm Hg. Both patients showed clinical symptoms before the last routine blood pressure measurement before the event. In 4 patients, no routine blood pressure measurements were performed before, during, and after the occurrence of the adverse event (2 patients had central venous catheterrelated blood stream infection, 1 patient had heart failure, and 1 patient had pneumothorax). All 4 patients survived the remainder of their stay at the hospital without further complications. Calculation of correlation coefficients, SDD, and coefficients of variation are shown in Table 4. The SDD for day-to-day differences in mean blood pressure values was 16.6 for systolic and 11.6 for diastolic blood pressure between day 3 and day 4, and 16.2 for systolic and 11.8 for diastolic blood pressure between day 10 and day 11. The corresponding SDD for blood pressure variations before a
Proportion of patients with pre-specified blood pressure (BP) changes Patients with no adverse clinical events
Difference in mean BP ⬎10 mm Hg Systolic % Diastolic % ⬎20 mm Hg Systolic % Diastolic %
Patients with adverse clinical events (n ⫽ 118)*
3rd vs 4th day of hospital stay (n ⫽ 517)†
10th vs 11th day of hospital stay (n ⫽ 344)‡
40.7 23.7
41.3 24.4
35.4 22.4
10.2 5.1
14.2 4.9
17.7 4.1
*The difference between the last routine blood pressure value before the event and the mean blood pressure value 72 hours before the event was calculated. These numbers do not include 4 patients for whom no routine blood pressure measurements were made before the event. †All patients with no clinical adverse events during their hospital stay. The difference between the mean routine blood pressure values of the 3rd and 4th day of hospitalization was calculated. ‡All patients with no clinical adverse events during their hospital stay who were hospitalized for at least 11 days. The difference between the mean routine blood pressure values of the 10th and 11th day of hospitalization was calculated.
Conen et al Table 4
Routine Blood Pressure in Hospitalized Patients
70.e21
Assessment of blood pressure (BP) variability
Patients with adverse events* Systolic Diastolic Patients with no adverse events Systolic 3rd vs 4th day† Diastolic 3rd vs 4th day† Systolic 11th vs 10th day‡ Diastolic 11th vs 10th day‡
Mean difference (mm Hg)
R
SDD
CV (%)
⫺1.03 ⫺0.14
0.818 0.543
13.9 11.1
10.8 15.6
⫺0.20 0.04 ⫺1.53 0.84
0.762 0.651 0.763 0.622
16.6 11.6 16.2 11.8
12.4 15.0 12.6 14.6
R ⫽ correlation coefficient; SDD ⫽ standard deviation of the differences; CV ⫽ coefficient of variation. *The last routine blood pressure value before the event and the mean blood pressure value 72 hours before the event were compared. These numbers do not include 4 patients for whom no routine blood pressure measurements were made before the event. †Mean routine blood pressure on the 4th day of hospitalization was compared with the mean blood pressure on the 3rd day of hospitalization. Only patients with no adverse events during their hospital stay are included in this group (n ⫽ 517). ‡Mean routine blood pressure on the 11th day of hospitalization was compared with the mean blood pressure on the 10th day of hospitalization. Only patients with no adverse events during their hospital stay are included in this group (n ⫽ 344).
clinical event were 13.9 for systolic and 11.1 for diastolic blood pressure. Coefficients of variation also were similar for comparisons of days 3 and 4, days 10 and 11, and for changes noted in patients who experienced clinical complications.
DISCUSSION To our knowledge, this is the first study to focus on the value of routine blood pressure measurements in an unselected population of hospitalized patients. The principal finding of the study is that when hospitalized patients experience complications during their hospital stay, significant blood pressure changes do not occur or do not occur early enough to detect these complications at an earlier stage. Of the patients in the study, 19% had clinical complications; the most common were gastrointestinal hemorrhage, falls, other bleeding, pneumonia, and bacteremia. Although changes in systolic blood pressure of at least 10 mm Hg did occur just before a clinical event in 41% of the patients, these changes were indistinguishable from normal day-today variations in patients who did not experience complications. Even when cases of death or severe complications that required admission to an intensive care unit are taken into account, systolic blood pressure changes of more than 10 mm Hg occurred in only 27% of this group of patients. Even when blood pressure differences of at least 20 mm Hg were considered, the findings were identical. The results of the present study agree with previous studies in selected patients.9-11 However, these studies assessed the value of blood pressure measurement for the management of specific admission diagnoses. It is interesting to note that one third of patients with major hemorrhage after trauma have normal blood pressure upon admission to the hospital.9 We found that patients with severe complications that led to admission to an intensive care unit do not have higher variations in blood pressure than do patients who do not experience complications.
In 4 patients with significant complications, no blood pressure readings were obtained before, during, and after the event. The outcomes for these patients were excellent. These examples show that the early detection and accurate treatment of adverse events in hospitalized patients is possible without regularly monitoring blood pressure. The usefulness of routine blood pressure measurements in the care of hospitalized patients may be impaired as a result of inaccurate and unstandardized blood pressure measurement procedures. This reduces the sensitivity and specificity of blood pressure changes. Only about 3% of health care workers measure blood pressure accurately,12 and nurses only rarely measure blood pressure according to recommended techniques.13 But even if education programs could improve the standardization and accuracy of blood pressure measurement,14 it is unlikely that these improvements would significantly increase the predictive power of routine in-hospital blood pressure measurement. In addition, hypotension is a late and rather abrupt event, at least in hemorrhage, the most frequent complication we observed.15 Therefore, routine blood pressure monitoring would not be an appropriate monitoring tool to predict this condition. But then how should hospitalized patients be assessed for clinical complications? It is the authors’ conviction that clinical observation and assessment is of primary importance in monitoring a patient’s condition, although stringent studies are lacking to support this conclusion. However, some indirect evidence is provided by Botti et al.11 They found that most bleeding complications after coronary angiography are detected during regular inspections by the staff or by the patient. Mean blood pressure and heart rate did not differ between patients with local bleeding complications and those without. Eliminating or reducing the number of routine blood pressure measurements in most inpatients would make more staff time available for more useful interventions. During ward rounds, more time could be used for evaluating the patient’s
70.e22 clinical condition. In this way, nurses could substantially improve the use of their time and also possibly the outcome and satisfaction of their patients. However, if blood pressure measurements are reduced or stopped, in no way should the number of ward rounds or patient inspections be reduced. Our study has several limitations. A randomized controlled trial would be necessary to provide definite evidence that routine blood pressure measurements have no benefit in the care of hospitalized patients. Because the measurement of blood pressure is a fundamental part of routine hospital patient care at the moment, we deemed it unlikely that such a trial would be feasible at this time. In this study, some events could have been missed because patient charts were reviewed retrospectively. However, it is unlikely that major events were missed by the reviewers or that an event was not noted at all in the patient chart. The more severe an event, the more documentation is available. In addition, patients who are admitted to an intensive care unit are easily identified because of the different type of flow charts associated with that type of care. In conclusion, this study suggests that routine blood pressure measurements in hospitalized patients are of little value in predicting clinical complications. Considering the large number of routine blood pressure measurements taken every day, and the total amount of staff time required for even this simple procedure, substantial resources could be freed for more useful interventions.
ACKNOWLEDGMENTS This study received financial support from the Scientific fund, University Hospital Basel, and the Astra Zeneca Foundation, Department of Internal Medicine, University Hospital Basel. We are indebted to our study nurse, C. Schlienger for her assistance with the present project. We also would like to thank the other nurses and the physicians for their excellent collaboration with the study team.
The American Journal of Medicine, Vol 119, No 1, January 2006
References 1. Evans D, Hodgkinson B, Berry J. Vital signs in hospital patients: a systematic review. Int J Nurs Stud. 2001;38:643-650. 2. Reilly BM. Physical examination in the care of medical inpatients: an observational study. Lancet. 2003;362:1100-1105. 3. Reichlin S, Dieterle T, Camli C, Leimenstoll B, Schoenenberger RA, Martina B. Initial clinical evaluation of cardiac systolic murmurs in the ED by noncardiologists. Am J Emerg Med. 2004;22:71-75. 4. Drazner MH, Rame JE, Stevenson LW, Dries DL. Prognostic importance of elevated jugular venous pressure and a third heart sound in patients with heart failure. N Engl J Med. 2001;345:574-581. 5. Martina B, Bucheli B, Stotz M, Battegay E, Gyr N. First clinical judgment by primary care physicians distinguishes well between nonorganic and organic causes of abdominal or chest pain. J Gen Intern Med. 1997;12:459-465. 6. Young MA, Rowlands DB, Stallard TJ, Watson RD, Littler WA. Effect of environment on blood pressure: home versus hospital. Br Med J (Clin Res Ed). 1983;286:1235-1236. 7. Fotherby MD, Critchley D, Potter JF. Effect of hospitalization on conventional and 24-hour blood pressure. Age Ageing. 1995;24:2529. 8. Hossmann V, FitzGerald GA, Dollery CT. Influence of hospitalization and placebo therapy on blood pressure and sympathetic function in essential hypertension. Hypertension. 1981;3:113-118. 9. Luna GK, Eddy AC, Copass M. The sensitivity of vital signs in identifying major thoracoabdominal hemorrhage. Am J Surg. 1989; 157:512-515. 10. Potti A, Panwalkar A, Hebert B, Sholes K, Lewis MJ, Hanley J. Ineffectiveness of measuring routine vital signs in adult inpatients with deep venous thrombosis. Clin Appl Thromb Hemost. 2003;9: 163-166. 11. Botti M, Williamson B, Steen K. Coronary angiography observations: evidence-based or ritualistic practice? Heart Lung. 2001;30:138-145. 12. Villegas I, Arias IC, Botero A, Escobar A. Evaluation of the technique used by health-care workers for taking blood pressure. Hypertension. 1995;26:1204-1206. 13. Bogan B, Kritzer S, Deane D. Nursing student compliance to standards for blood pressure measurement. J Nurs Educ. 1993;32:90-92. 14. Scherwitz LW, Evans LA, Hennrikus DJ, Vallbona C. Procedures and discrepancies of blood pressure measurements in two community health centers. Med Care. 1982;20:727-738. 15. Lohmeier TE. Neurohumoral regulation of arterial pressure in hemorrhage and heart failure. Am J Physiol Regul Integr Comp Physiol. 2002;283:R810-R814.
CLINICAL RESEARCH STUDY
Routine Blood Pressure Measurements do not Predict Adverse Events in Hospitalized Patients David Conen, MD, Bernd M. Leimenstoll, MD, André P. Perruchoud, MD, Benedict Martina, MD Department of Internal Medicine, University Hospital Basel, Switzerland. ABSTRACT PURPOSE: Countless blood pressure measurements are performed every day for almost every hospitalized patient. We analyzed the value of routine blood pressure measurements on patient care in an unselected group of hospitalized patients. METHODS: The study included 639 patients who were admitted to the hospital with a broad range of medical conditions. Two independent investigators reviewed the medical charts of the patients. Routine blood pressure values were abstracted from the patient charts and evaluated with respect to the occurrence of adverse clinical events in the study group. Changes in blood pressure between the last measurement just before adverse clinical events and the mean blood pressure values 72 hours before the adverse events were calculated and compared with mean normal day-to-day variations in blood pressure. RESULTS: In every patient, a mean of 1.6 ⫾ 0.6 routine blood pressure measurements per day were performed. Of the 639 patients in the study, 122 (19%) had clinical complications. The most commonly occurring complications were gastrointestinal bleeding (n ⫽ 15), falls (n ⫽ 13), other bleeding (n ⫽ 12) and pneumonia (n ⫽ 8). In patients who experienced clinical complications, pre-event systolic and diastolic blood pressure changes of at least 10 mm Hg occurred in 41% and 24% of the group, respectively, but this was not different from the normal day-to-day variations observed in patients who had no clinical complications. The results also were similar for patients who died or who had a severe adverse event that required admission to an intensive care unit. CONCLUSION: Routine blood pressure measurements in a general hospital patient population do not predict clinical adverse events. © 2006 Elsevier Inc. All rights reserved. KEYWORDS: Blood pressure determination; Hospitals; Inpatients; Patient monitoring; Complications
Routine blood pressure measurements are performed daily for almost every hospitalized patient. Countless blood pressure values are measured and written down in patient charts, which requires a significant amount of time on the part of hospital personnel. However, there are few published studies on the value of routine blood pressure measurements in the care of hospital patients.1 The usefulness of other tools that are routinely used for the management of hospital patients, such as physical examination or cardiac auscultation, has been demonstrated in observational studies.2,3 For example, a skilled physical examination may have a sub-
Requests for reprints should be addressed to Benedict Martina, MD, University Hospital, Petersgraben 4, 4031 Basel, Switzerland. E-mail address: [email protected].
0002-9343/$ -see front matter © 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.amjmed.2005.09.060
stantial impact on the care of hospital patients2 and represents an effective filter for the reasonable use of more expensive technology.3-5 In contrast, the mere act of being hospitalized directly influences a patient’s routine blood pressure and impairs the usefulness of these values for detecting elevated blood pressure in hypertensive patients.6-8 Furthermore, patient hospital stays are usually not long enough to correctly adjust long-term blood pressure. Therefore, the only reason for routinely measuring blood pressure in all hospital patients should be to enhance or achieve earlier detection of clinical complications or deterioration in a patient’s condition. The few studies that currently are available show that the routine measurement of vital signs has no effect on detecting deterioration at an earlier
70.e18
The American Journal of Medicine, Vol 119, No 1, January 2006
stage, on the length of patient hospital stays, or on the Patients clinical outcome in several selected subgroups of hospital A total of 749 patients were screened for inclusion. Of patients.9-11 However, these earlier studies did not assess this group, 35 patients (5%) were terminally ill; in 54 the value of routine blood pressure measurement in pre(7%), standardized blood pressure measurements were dicting in-hospital complications. Rather these studies not feasible; and for 21 (3%), the medical charts were examined the usefulness of rouincomplete. Ultimately, 639 patine blood pressure measuretients (85%) were retained for ments in the management of spefinal evaluation. CLINICAL SIGNIFICANCE cific clinical conditions for Baseline characteristics for the which patients were admitted to study population are shown in Table ● In hospitalized patients, adverse events the hospital. For example, a ret1. Mean age in years was 64 ⫾ 17, and substantial day-to-day blood presrospective survey of patients and 322 (50%) patients were male. sure changes are common. with severe thoracoabdominal The patients were hospitalized for a ● Blood pressure changes prior to an adtrauma found that one third of wide range of ailments, most frethese patients had a normal sysverse event are indistinguishable from quently acute coronary syndrome tolic blood pressure despite the normal day-to-day variations in patients (9.4%), fractures (5.3%), malignanoccurrence of major bleeding.9 cies (miscellaneous 5.3%, lung canwithout adverse events. The investigators therefore concer 2.5%, colon cancer 1.4%), pneu● Routine blood pressure measurements in cluded that normal vital signs do monia (5.2%), heart failure (4.7%), a general hospital patient population do not preclude the presence of liferenal transplantation (3.4%), exacnot predict clinical adverse events. threatening hemorrhage. Botti et erbation of chronic obstructive pulal also showed that after coromonary disease (3.1%), and bacte● Considering the large number of routine nary angiography, blood presremia (2.7%). blood pressure measurements taken, sure and heart rate were no difsubstantial resources could be freed for ferent in patients with local Procedures more useful interventions. bleeding complications from The physicians in charge of the those with no complications.11 patients prescribed the frequency Therefore, our hypothesis was of routine blood pressure meathat routine blood pressure measurements. A routine blood pressure measurement was desurement is not useful for the early detection of adverse fined as being a measurement prescribed as part of the clinical events in an unselected group of hospital patients. routine monitoring plan for the patient, independent of spe-
METHODS Study Design The study was conducted at the University Hospital Basel, a 784-bed, academic primary and tertiary care hospital in Switzerland. Patients hospitalized in the wards of the internal medicine, visceral surgery, and orthopedics departments were assessed for eligibility during the period of June 1 through December 16, 2003. For the purpose of the present study, all patients aged 18 years or older in these departments were included according to the following pattern: During one week, patients of one of the wards were included. The next week, patients of another ward were screened for eligibility. The chronological order of rotation among wards was predefined. We excluded terminally ill patients in whom diagnostic and therapeutic procedures were discontinued and also patients in whom standardized blood pressure measurements were not feasible (for example in the case of bilateral arm fracture). The local ethics committee approved the study protocol and the study was performed according to the Declaration of Helsinki. All authors participated in data analysis and interpretation. The manuscript was prepared by the authors.
cific interventions or symptoms. All routine blood pressure values for each patient during the patient’s hospital stay were entered into a database and double-checked by a second investigator. Two experienced physicians independently reviewed all patient charts for adverse clinical events according to a standardized protocol, namely, review of discharge letters, charge notes, monitoring sheets, and laboratory values. The findings of the 2 physicians were compared and any disagreement was resolved by consensus finding. The kappa value as a measure for inter-observer concordance in the detection of clinical complications was 0.79. To evaluate the usefulness of routine blood pressure measurements in predicting clinical complications, the last routine blood pressure measurement taken before an adverse event was compared with the patient’s mean blood pressure during the 72 hours preceding the event. For patients with no clinical complications, the day-to-day variations in routine blood pressure measurement values were determined. We did this by randomly choosing a day between day 1 and day 10 (day 4) and a day between day 11 and day 20 (day 11) of the hospital stay. We then compared the mean values obtained on those days with the mean values on the preceding day (ie, day 4 compared with day 3 and day 11 compared with day 10).
Conen et al Table 1
Routine Blood Pressure in Hospitalized Patients
Reasons for admitting patients N ⫽ 639
Infections Pneumonia (%) Exacerbation of COPD (%) Bacteremia (%) Miscellaneous infections (%) Osteomyelitis (%) Endocarditis (%) Abscess (%) Urinary tract infection (%) Cardiovascular disease Acute coronary syndrome (%) Heart failure (%) Valvular heart disease (%) Peripheral artery disease (%) Pulmonary embolism (%) Malignancies Miscellaneous malignancies (%) Lung cancer (%) Hematological malignancies (%) Colon cancer (%) Musculoskeletal system disorder Fractures (%) Prosthetic joint implantation (%) Osteoarthritis (%) Discal herniation (%) Gastrointestinal disorder Gastrointestinal bleeding (%) Diverticulitis (%) Colitis (%) Cholecystitis (%) Liver cirrhosis (%) Diarrhea (%) Miscellaneous conditions Renal transplantation (%) Falls (%) Alcoholism (%) Syncope (%) Diabetes (%) Other diagnoses (%)
33 20 17 15 15 10 10 7
(5.2) (3.1) (2.7) (2.3) (2.3) (1.6) (1.6) (1.1)
60 30 9 7 6
(9.4) (4.7) (1.4) (1.1) (0.9)
33 16 11 9
(5.2) (2.5) (1.7) (1.4)
34 11 9 8
(5.3) (1.7) (1.4) (1.3)
10 8 8 8 7 7
(1.6) (1.3) (1.3) (1.3) (1.1) (1.1)
22 6 6 5 5 177
(3.4) (0.9) (0.9) (0.8) (0.8) (27.7)
Data are number of patients (%). Because of rounding, the percentages might not equal 100. COPD ⫽ chronic obstructive pulmonary disease.
70.e19 treatment were included in the analysis. Also included in the analysis were: the death of a patient, any documented thrombotic or ischemic event, any exacerbation of heart failure that resulted in additional treatment, any symptomatic and documented cardiac arrhythmias, any acute renal failure, any witnessed syncope or seizure, any documented fall, and any complication that arose from a procedure in which routine postprocedural blood pressure monitoring usually is not performed (eg, lumbar puncture, central venous catheters).
Statistical Analysis Assuming a relative frequency of 40% (10%) for blood pressure variations of at least 10 (20) mm Hg, the present sample sizes allowed us to detect a true 14% (9%) difference in the respective frequencies between the patient populations with and without complications with a probability of 80% and a significance level of 5%. Continuous variables were expressed as mean ⫾ standard deviation. Kappa statistics were used to assess the interobserver concordance for the detection of clinical complications through independent chart review. To assess day-to-day variability in patients who did not experience clinical complications, absolute blood pressure differences were calculated for mean blood pressure on day 4 versus day 3 and mean blood pressure on day 11 versus day 10. In patients who experienced a clinical complication, blood pressure differences were calculated for the last blood pressure value taken just before the event versus mean blood pressure 72 hours before the event. The proportion of patients with an absolute blood pressure difference of at least 10 mm Hg and 20 mm Hg between these days was calculated. Differences in proportions were assessed using Fisher’s exact test. A 2-tailed P value of less than .05 was considered to indicate statistical significance. Pearson correlation coefficients, mean differences, standard deviation of the differences (SDD), and coefficients of variation (SDD divided by mean blood pressure) were calculated. All analyses were performed using SPSS software version 11.5 (SPSS Inc, Chicago, Ill).
RESULTS Adverse Clinical Events An adverse clinical event was defined as a complication that arose during the patient’s hospital stay, including those complications with a causal relationship to the admission diagnosis. Adverse events that arose during a stay in the emergency room or in the intensive care unit or within the first 24 hours after surgery were excluded. Any event that led to an unplanned admission to the intensive care unit was included in the analysis. Significant bleeding episodes were defined as bleeding that required the transfusion of at least two units of blood, any intracranial hemorrhage, any gastrointestinal bleeding other than hemorrhoidal, any bleeding that necessitated surgical intervention or admission to the intensive care unit, and disabling bleeding from any site. All documented infections that required medical or surgical
We detected adverse events in 122 patients (19%) (Table 2). The most frequent clinical adverse events were gastrointestinal bleeding (12.3%), falls (10.7%), other bleeding (9.8%), pneumonia (6.6%), and bacteremia (6.6%). In 27 cases (22%), the patient was admitted to an intensive care unit. Among those admitted to an intensive care unit, heart failure (22%), gastrointestinal bleeding (19%), arrhythmias (19%), and other bleedings (15%) were the most frequent complications (Table 2). Each day, the mean number of routine blood pressure readings per patient was 1.6 ⫾ 0.6. As shown in Table 3, the proportion of patients with a blood pressure change of at least 10 mm Hg just before a clinical event was 40.7% for systolic blood pressure and 23.7% for diastolic blood pressure. The proportion of patients with a blood pressure dif-
70.e20
The American Journal of Medicine, Vol 119, No 1, January 2006
Table 2 Type and number of clinical complications in the study group Number of Number of patients admitted to patients (n ⫽ 122) ICU (n ⫽ 27) GI bleeding (%) Falls (%) Bleeding (other than GI) (%) Pneumonia (%) Bacteremia (%) Heart failure (%) Acute renal failure (%) Arrhythmias (%) Death (%) Febrile urinary tract infection (%) DVT/pulmonary embolism (%) Empyema (%) Syncope (%) Acute urinary retention (%) Acute coronary syndrome (%) Abscess (%) Other infections (%) Other complications (%)*
15 13 12 8 8 6 5 5 4 4 4 4 4 3 3 3 7 14
(12.3) (10.7) (9.8) (6.6) (6.6) (4.9) (4.1) (4.1) (3.3) (3.3) (3.3) (3.3) (3.3) (2.5) (2.5) (2.5) (5.7) (11.5)
5 4 2 6 1 5 1
2 1
Data are number of patients (%). Because of rounding, % might not equal 100. GI ⫽ gastrointestinal; ICU ⫽ intensive care unit; DVT ⫽ deep venous thrombosis. *Other complications included transient ischemic attack (2), pneumothorax (2), ileus (2), seizure (2), hypoglycemia with neurological manifestations, allergic drug reaction, thrombopenia, dyspnea, decrease in oxygen saturation (with admission to ICU), and radiculary syndrome after lumbar puncture.
ference of at least 20 mm Hg just before an event was 10.2% for systolic blood pressure and 5.1% for diastolic blood pressure. These changes were very similar to the normal day-to-day variability in blood pressure in the study population, as illustrated by the proportion of patients who had no clinical complications and who showed the same abso-
Table 3
lute mean blood pressure changes between day 3 and day 4 and between day 10 and day 11 of their hospital stay (Table 3). None of these comparisons between the two patient groups reached statistical significance (all P values ⬎.27). When only patients who experienced clinical complications are considered, the proportion of patients with a difference in systolic blood pressure of at least 10 mm Hg was 51.7% between day 3 and day 4 and 42.9% between day 10 and day 11. The proportion with a difference in systolic blood pressure of at least 20 mm Hg was 17.2% between day 3 and day 4, and 16.3% between day 10 and day 11. These percentages were not significantly different from those related to blood pressure changes just before a clinical event (all P values ⬎.25) (Table 3). When only death or severe complications (n ⫽ 31) that led to admission to an intensive care unit are considered, the proportion of patients with a blood pressure change of at least 10 mm Hg before the event was 30.0% for systolic blood pressure and 33.3% for diastolic blood pressure. Only 1 patient who was admitted to the intensive care unit and 1 patient who subsequently died had an event-related systolic blood pressure change of at least 20 mm Hg. Both patients showed clinical symptoms before the last routine blood pressure measurement before the event. In 4 patients, no routine blood pressure measurements were performed before, during, and after the occurrence of the adverse event (2 patients had central venous catheterrelated blood stream infection, 1 patient had heart failure, and 1 patient had pneumothorax). All 4 patients survived the remainder of their stay at the hospital without further complications. Calculation of correlation coefficients, SDD, and coefficients of variation are shown in Table 4. The SDD for day-to-day differences in mean blood pressure values was 16.6 for systolic and 11.6 for diastolic blood pressure between day 3 and day 4, and 16.2 for systolic and 11.8 for diastolic blood pressure between day 10 and day 11. The corresponding SDD for blood pressure variations before a
Proportion of patients with pre-specified blood pressure (BP) changes Patients with no adverse clinical events
Difference in mean BP ⬎10 mm Hg Systolic % Diastolic % ⬎20 mm Hg Systolic % Diastolic %
Patients with adverse clinical events (n ⫽ 118)*
3rd vs 4th day of hospital stay (n ⫽ 517)†
10th vs 11th day of hospital stay (n ⫽ 344)‡
40.7 23.7
41.3 24.4
35.4 22.4
10.2 5.1
14.2 4.9
17.7 4.1
*The difference between the last routine blood pressure value before the event and the mean blood pressure value 72 hours before the event was calculated. These numbers do not include 4 patients for whom no routine blood pressure measurements were made before the event. †All patients with no clinical adverse events during their hospital stay. The difference between the mean routine blood pressure values of the 3rd and 4th day of hospitalization was calculated. ‡All patients with no clinical adverse events during their hospital stay who were hospitalized for at least 11 days. The difference between the mean routine blood pressure values of the 10th and 11th day of hospitalization was calculated.
Conen et al Table 4
Routine Blood Pressure in Hospitalized Patients
70.e21
Assessment of blood pressure (BP) variability
Patients with adverse events* Systolic Diastolic Patients with no adverse events Systolic 3rd vs 4th day† Diastolic 3rd vs 4th day† Systolic 11th vs 10th day‡ Diastolic 11th vs 10th day‡
Mean difference (mm Hg)
R
SDD
CV (%)
⫺1.03 ⫺0.14
0.818 0.543
13.9 11.1
10.8 15.6
⫺0.20 0.04 ⫺1.53 0.84
0.762 0.651 0.763 0.622
16.6 11.6 16.2 11.8
12.4 15.0 12.6 14.6
R ⫽ correlation coefficient; SDD ⫽ standard deviation of the differences; CV ⫽ coefficient of variation. *The last routine blood pressure value before the event and the mean blood pressure value 72 hours before the event were compared. These numbers do not include 4 patients for whom no routine blood pressure measurements were made before the event. †Mean routine blood pressure on the 4th day of hospitalization was compared with the mean blood pressure on the 3rd day of hospitalization. Only patients with no adverse events during their hospital stay are included in this group (n ⫽ 517). ‡Mean routine blood pressure on the 11th day of hospitalization was compared with the mean blood pressure on the 10th day of hospitalization. Only patients with no adverse events during their hospital stay are included in this group (n ⫽ 344).
clinical event were 13.9 for systolic and 11.1 for diastolic blood pressure. Coefficients of variation also were similar for comparisons of days 3 and 4, days 10 and 11, and for changes noted in patients who experienced clinical complications.
DISCUSSION To our knowledge, this is the first study to focus on the value of routine blood pressure measurements in an unselected population of hospitalized patients. The principal finding of the study is that when hospitalized patients experience complications during their hospital stay, significant blood pressure changes do not occur or do not occur early enough to detect these complications at an earlier stage. Of the patients in the study, 19% had clinical complications; the most common were gastrointestinal hemorrhage, falls, other bleeding, pneumonia, and bacteremia. Although changes in systolic blood pressure of at least 10 mm Hg did occur just before a clinical event in 41% of the patients, these changes were indistinguishable from normal day-today variations in patients who did not experience complications. Even when cases of death or severe complications that required admission to an intensive care unit are taken into account, systolic blood pressure changes of more than 10 mm Hg occurred in only 27% of this group of patients. Even when blood pressure differences of at least 20 mm Hg were considered, the findings were identical. The results of the present study agree with previous studies in selected patients.9-11 However, these studies assessed the value of blood pressure measurement for the management of specific admission diagnoses. It is interesting to note that one third of patients with major hemorrhage after trauma have normal blood pressure upon admission to the hospital.9 We found that patients with severe complications that led to admission to an intensive care unit do not have higher variations in blood pressure than do patients who do not experience complications.
In 4 patients with significant complications, no blood pressure readings were obtained before, during, and after the event. The outcomes for these patients were excellent. These examples show that the early detection and accurate treatment of adverse events in hospitalized patients is possible without regularly monitoring blood pressure. The usefulness of routine blood pressure measurements in the care of hospitalized patients may be impaired as a result of inaccurate and unstandardized blood pressure measurement procedures. This reduces the sensitivity and specificity of blood pressure changes. Only about 3% of health care workers measure blood pressure accurately,12 and nurses only rarely measure blood pressure according to recommended techniques.13 But even if education programs could improve the standardization and accuracy of blood pressure measurement,14 it is unlikely that these improvements would significantly increase the predictive power of routine in-hospital blood pressure measurement. In addition, hypotension is a late and rather abrupt event, at least in hemorrhage, the most frequent complication we observed.15 Therefore, routine blood pressure monitoring would not be an appropriate monitoring tool to predict this condition. But then how should hospitalized patients be assessed for clinical complications? It is the authors’ conviction that clinical observation and assessment is of primary importance in monitoring a patient’s condition, although stringent studies are lacking to support this conclusion. However, some indirect evidence is provided by Botti et al.11 They found that most bleeding complications after coronary angiography are detected during regular inspections by the staff or by the patient. Mean blood pressure and heart rate did not differ between patients with local bleeding complications and those without. Eliminating or reducing the number of routine blood pressure measurements in most inpatients would make more staff time available for more useful interventions. During ward rounds, more time could be used for evaluating the patient’s
70.e22 clinical condition. In this way, nurses could substantially improve the use of their time and also possibly the outcome and satisfaction of their patients. However, if blood pressure measurements are reduced or stopped, in no way should the number of ward rounds or patient inspections be reduced. Our study has several limitations. A randomized controlled trial would be necessary to provide definite evidence that routine blood pressure measurements have no benefit in the care of hospitalized patients. Because the measurement of blood pressure is a fundamental part of routine hospital patient care at the moment, we deemed it unlikely that such a trial would be feasible at this time. In this study, some events could have been missed because patient charts were reviewed retrospectively. However, it is unlikely that major events were missed by the reviewers or that an event was not noted at all in the patient chart. The more severe an event, the more documentation is available. In addition, patients who are admitted to an intensive care unit are easily identified because of the different type of flow charts associated with that type of care. In conclusion, this study suggests that routine blood pressure measurements in hospitalized patients are of little value in predicting clinical complications. Considering the large number of routine blood pressure measurements taken every day, and the total amount of staff time required for even this simple procedure, substantial resources could be freed for more useful interventions.
ACKNOWLEDGMENTS This study received financial support from the Scientific fund, University Hospital Basel, and the Astra Zeneca Foundation, Department of Internal Medicine, University Hospital Basel. We are indebted to our study nurse, C. Schlienger for her assistance with the present project. We also would like to thank the other nurses and the physicians for their excellent collaboration with the study team.
The American Journal of Medicine, Vol 119, No 1, January 2006
References 1. Evans D, Hodgkinson B, Berry J. Vital signs in hospital patients: a systematic review. Int J Nurs Stud. 2001;38:643-650. 2. Reilly BM. Physical examination in the care of medical inpatients: an observational study. Lancet. 2003;362:1100-1105. 3. Reichlin S, Dieterle T, Camli C, Leimenstoll B, Schoenenberger RA, Martina B. Initial clinical evaluation of cardiac systolic murmurs in the ED by noncardiologists. Am J Emerg Med. 2004;22:71-75. 4. Drazner MH, Rame JE, Stevenson LW, Dries DL. Prognostic importance of elevated jugular venous pressure and a third heart sound in patients with heart failure. N Engl J Med. 2001;345:574-581. 5. Martina B, Bucheli B, Stotz M, Battegay E, Gyr N. First clinical judgment by primary care physicians distinguishes well between nonorganic and organic causes of abdominal or chest pain. J Gen Intern Med. 1997;12:459-465. 6. Young MA, Rowlands DB, Stallard TJ, Watson RD, Littler WA. Effect of environment on blood pressure: home versus hospital. Br Med J (Clin Res Ed). 1983;286:1235-1236. 7. Fotherby MD, Critchley D, Potter JF. Effect of hospitalization on conventional and 24-hour blood pressure. Age Ageing. 1995;24:2529. 8. Hossmann V, FitzGerald GA, Dollery CT. Influence of hospitalization and placebo therapy on blood pressure and sympathetic function in essential hypertension. Hypertension. 1981;3:113-118. 9. Luna GK, Eddy AC, Copass M. The sensitivity of vital signs in identifying major thoracoabdominal hemorrhage. Am J Surg. 1989; 157:512-515. 10. Potti A, Panwalkar A, Hebert B, Sholes K, Lewis MJ, Hanley J. Ineffectiveness of measuring routine vital signs in adult inpatients with deep venous thrombosis. Clin Appl Thromb Hemost. 2003;9: 163-166. 11. Botti M, Williamson B, Steen K. Coronary angiography observations: evidence-based or ritualistic practice? Heart Lung. 2001;30:138-145. 12. Villegas I, Arias IC, Botero A, Escobar A. Evaluation of the technique used by health-care workers for taking blood pressure. Hypertension. 1995;26:1204-1206. 13. Bogan B, Kritzer S, Deane D. Nursing student compliance to standards for blood pressure measurement. J Nurs Educ. 1993;32:90-92. 14. Scherwitz LW, Evans LA, Hennrikus DJ, Vallbona C. Procedures and discrepancies of blood pressure measurements in two community health centers. Med Care. 1982;20:727-738. 15. Lohmeier TE. Neurohumoral regulation of arterial pressure in hemorrhage and heart failure. Am J Physiol Regul Integr Comp Physiol. 2002;283:R810-R814.