We recruited 125 participants from the general population of Ciudad Real using snowball sampling, from May to August 2016. A total of 100 participants completed the study; 25 participants did not return for the second measurement. According to the ESH International Protocol [35], a minimum of one-third of the total number of participants must be men and a minimum of one-third must be women. The participants satisfied the eligibility criteria in the latest revision of the ESH International Protocol [33-35], which are as follows:

Inclusion criteria:

Exclusion criteria:

All participants gave their written informed consent to participate in this study.

Before recruitment, we calculated with the GRANMO sample size calculation program (Institut Hospital del Mar d’Investigacions Mèdiques, Barcelona, Spain) that a sample size of 125 would give 80% statistical power (α=.05) and two-tailed test to detect a difference equal to or greater than 0.9 mm Hg in BP. This calculation was based on the assumptions of a standard deviation of 3.2 mm Hg [37] and a loss to follow-up rate of 20%.

Participants visited the study site for two measurement sessions separated by 5 to 7 days. The first author (VMP) collected all the data. Both BP and HR were measured in the left arm. The central portion of the cuff was placed at heart level, with the bottom edge 1 to 2 cm above the elbow.

During the first session, each participant reported his/her sex and date of birth. Measurements were taken of the participant’s weight (with an automatic digital scale), height (with a wall-mounted measuring rod), and arm circumference (with a tape measure at the midpoint between the acromion and olecranon). The participant then relaxed for 10 minutes before baseline BP and HR were measured with the criterion BP monitor Omron M3 (HEM-7200-E, Omrom Healthcare Co Kyoto, Japan) (Figure 2). Next, BP and HR were measured with the test device QardioArm only to confirm that the test device was working correctly on the participant (these were not included in analysis).

The rest of the first session involved seven BP and HR measurements, alternating between the criterion and test devices, with 2 to 3 minute intervals between measurements. The first, third, fifth, and seventh measurements were made with the criterion device; the second, fourth, and sixth measurements were made with the test device. The second measurement session involved the same series of seven measurements, with identical procedures carried out by the same researcher, which occurred on the same day of the week and at the same time after lunch, in the same room with identical conditions, as far as it was possible.

During measurements, participants were calm and quiet while sitting with their feet parallel and flat on the floor, their legs uncrossed, and their left hands resting palm side up on a flat surface. The room was maintained at an appropriate temperature and kept free from noises and distractions during the sessions [33,34].

Of the 125 participants who completed the first session, 100 also completed the second session. Our analysis focused on data from just these 100 participants.

The Omron M3 Intellisense was the criterion device used as a benchmark in our study. It has CE 0197 certification and has been validated by the ESH between the Omron M3 and a different criterion device with mean 1.7 (SD 3.2) mm Hg for systolic and mean –0.9 (SD 2.6) mm Hg for diastolic [36].

The Omron M3 is a fully automatic BP monitor operating on the oscillometric principle. It has a range of 0 to 299 mm Hg for BP and 40 to 180 beats per minute (bpm) for HR. The cuff is inflated using an electric pump and deflated by means of a pressure release valve.

After each measurement, the SBP, DBP, and HR are shown on the Omron M3’s LCD screen. The device can also display a symbol on the screen indicating an irregular heartbeat, which is detected during measurement of SBP and DBP.

The QardioArm app for mobile phones and tablets was the test device in our study. It has CE 0734 certification and is a fully automatic, noninvasive wireless BP monitor for measuring SBP, DBP, and HR in adults. The QardioArm also operates on the oscillometric principle and has an inflatable cuff, which is placed around the upper arm. The cuff is inflated automatically and deflated using a controlled pressure release valve. The cuff is suitable for an arm circumference of between 22 and 37 cm. The measuring range is 40 to 250 mm Hg for BP and 40 to 200 bpm for HR.

The QardioArm has an automatic screen with graphics and visuals to facilitate interpretation of the data. The monitor connects to the free Qardio app on any device with Bluetooth 4.0 that runs iOS 7.0 or later or Android 4.4 or later. It can be used with iPhones, iPods, iPads, Apple Watches, and Android mobile phones and tablets. The data are stored on the mobile phone or tablet and therefore the pattern of values over time can be viewed. The app can be configured to issue reminders and warnings, and the measurements and progress can be shared in real time with other users .

The QardioArm weighs approximately 310 g without batteries and measures 6.8 cm in width, 3.8 cm in height, and 14 cm in length, with the cuff closed. It requires four AAA batteries.

All data were analyzed using SPSS version 19 (IBM Corp, Armonk, NY, USA). Statistical significance was set at P<.05.

We computed univariate summary statistics on participants’ characteristics. Body mass index (BMI) was defined as weight in kilograms divided by height (in meters) squared. To assess whether the BP and HR variables were normally distributed, we conducted one-sample Kolmogorov-Smirnov tests, applying the Lilliefors significance correction. Significant probability values in this test indicate nonnormality.

To evaluate the reliability of the QardioArm, we computed the intraclass correlation coefficient (ICC) for comparing measurements taken within the same session (intrasession) and in different sessions (intersession) [38]. The participants were the groups in each ICC analysis. Each intrasession evaluation involved the three measurements obtained using the test device for a particular primary variable (SBP, DBP, and HR). We used the session means of a variable in calculating the intersession ICC.

We interpreted ICC values with the guide proposed by Landis and Koch [39]: ≤0.20=slight agreement, 0.21-0.40=fair agreement, 0.41-0.60=moderate agreement, 0.61-0.80=substantial agreement, and ≥0.81=almost perfect agreement. We also followed the recommendations of Portney and Watkins [40], who suggested that an ICC of 0.90 or higher for clinical measurements indicates they are reliable.

We also calculated the standard error of measurement (SEM) for each intrasession and intersession primary variable. SEM values indicate the degree of error, with low values indicating low error. We used the SEM formula of the standard deviation multiplied by the square root of (1–ICC). To identify systematic error over time, we also compared first and second session means on each primary variable with paired-sample Student t tests.

To determine the validity of the QardioArm, we compared the QardioArm measurements with those made by the Omron M3, using both the ESH’s International Protocol for the validation of BP measuring devices and the Student t test. Following the ESH protocol [33,34], we computed the absolute values of the differences between successive pairs of the seven measurements in a session (second–first, second–third, fourth–third, fourth–fifth, sixth–fifth, and sixth–seventh). This gives three paired differences for each variable (SBP, DBP, and HR), participant, and session. We classified whether the paired differences were ≤5, ≤10, ≤15, or >15 mm Hg for BP and ≤3, ≤5, ≤8, or >8 bpm in the case of HR. The four levels of difference correspond to very accurate, slightly inaccurate, moderately inaccurate, and very inaccurate, respectively [33,34]. If the device passed during both sessions, it could be recommended for clinical use. If it did not pass during the second session, it could not be recommended for clinical use.

We compared the observed number of differences (out of 300: three paired differences for a variable × 100 participants in a session) falling into these categories with the standards specified in the ESH International Protocol. There were two phases of this comparison. If a test device did not pass during the first session, the validation process was terminated. If it passed during the first session, the second session was started.

In addition to following the ESH International Protocol, we computed Student t test for independent samples to check whether there were differences between devices in BP and HR measurements. For these tests, we used the mean values of the three study variables obtained from both sessions for the different devices. We also compared the mean intersession differences for the two devices. Furthermore, we calculated the Pearson correlations between the devices for the mean participant values on the three primary variables [41].

Finally, we produced Bland-Altman plots [42] to display the agreement between the two devices. These plots show the difference between each pair of measurements on the y-axis against the mean of each pair of measurements on the x-axis.