Where Should Automated Blood Pressures be Done in Primary Care Offices?

• mean value of the last five systolic blood pressures for each location within the offices [ Time Frame: 1 hour ]
  comparison between the mean systolic values of the automated blood pressure measurements using a BpTRU device in private (quiet, no interactions with health care personnel) and non private areas of the offices.
  
  

• comparison of first and second sets of blood pressure readings [ Time Frame: 1 hour ]
  Determination of rank effect on automated blood pressure measurements
  
  
• comparisons of automated BP readings for patients with and without previously documented hypertension [ Time Frame: 1 hour ]
  
• effect of office decibel levels on automated office BP readings [ Time Frame: 1 hour ]
  

Following the design of our pilot study, patients consenting to two sets of blood pressure readings will be randomly allocated using a random number table to either automated blood pressure measurement using a BpTRU device in an exam room (private area without any traffic or staff) with the door closed, or BpTRU in an open (non-private area) of the clinic with routine office traffic and noise levels. Patients randomized to the private area first will be subsequently tested in the non-private area and those randomized to the non-private area first will then be tested in the private area. A single portable BpTRU machine will be used for all recordings. A research assistant will be trained on the use of the equipment with at least five supervised training readings. The BpTRU device will be placed on the patient's arm and readings will be recorded by the research assistant. The first reading will be done with the research assistant present to ensure proper placement and recording. The patient will then be left alone for the subsequent five measurements separated by one minute intervals. This will be immediately followed by a second set of readings in the alternate location. During both sets of readings the patient will be seated comfortably in a chair with arms and will be instructed not to talk or cross their legs. The same arm will be used for both sets of measurements with the blood pressure cuff at heart level. The research assistant will record the average of the last five out of six blood pressure readings for each office location, in accordance with the manufacturer's recommendations and previous studies.

The research assistant will record decibel levels in each location during BP readings using a Reed Sound Level Meter C-322 (Reed Instruments, Ste-Anne-de-Bellevue, Quebec, Canada). The sound level meter includes a data logger, and the mean decibel level per location for each set of readings will be recorded. The assistant will be trained in the use of the sound meter and logger and will use it during the five supervised BP training readings. The research assistant will also measure and record the patient's weight in kg, height in cm, and will collect information on patient's gender and self reported history of hypertension.

Planned sample size and analysis Our sample size is based on equivalence of BpTRU measurement done in non private office areas as compared to measurement done in private office areas, using the results of our pilot study. We consider differences of less than 5 mm Hg in systolic blood pressure to be clinically equivalent. In order to demonstrate equivalence for our primary outcome, and accounting for clustering effect in primary care offices, using a power of 80% and a two sided alpha level of 0.05 to conclude equivalency for a difference of 5 mm Hg or less, we will need to recruit 350 patients. Our planned recruitment is therefore fifty patients per office and seven office locations. 175 patients (25 per office location) will be randomly allocated to BpTRU measurement in a private exam room first and 175 will be allocated to an open area first.

Following the analytic methods used in our pilot study, we will use a random effect regression model to account for repeated measures within each participant for analyzing the data. Regression analysis will be undertaken to compare the measures and results will be adjusted for clustering effect. All tests will be two-sided using an alpha level of 0.05.