A mild pain was caused with an algometer in the thumbnail in two attempts and measured on paper (4.04 [SD 2.10] and 4.33 [SD 2.23]) and electronic (4.20 [SD 2.09] and 4.52 [SD 2.19]) VASs. Good intermethod (ICC[1,3]=.86) and intramethod (ICCa[1,3]=.86) reliability was supported. Bland-Altman analysis showed a difference of 0.18 (SD 0.49), and LoA ranged from –0.79 to 1.14.

The introduction of mobile devices and tablets in everyday health application is becoming increasingly common [15-17]. New smart health technologies are now available for clinicians and researchers, which may positively impact patient compliance to prescribed treatment and overall health care [18]. The use of mobile apps in pain management has been demonstrated to have a number of benefits, especially in clinical settings: pain apps are easy to use and usually welcomed by patients and clinicians [19,20]. Some concerns may arise in introducing mobile health (mHealth) in an elderly population (aged 65 years and older); however, there is growing evidence of accessibility and successful use of mobile pain apps in this population [21]. It is well recognized that pain assessment is the initial step in the early identification of many pathologies, and it is frequently adopted in effective clinical management plans [22].

However, the quality of some apps is still questionable, especially for pain management [23]. In our study, interrater agreement and an exploratory Bland-Altman plot analysis were presented in order to reach agreement between methods. Regarding ICC analysis, mean of squares from intrasubject and subject-method interaction were very small (0.65) compared with mean of squares from subject (16.46). No systematic effect in methods was found, even when inducing high values of ICC.

Bland-Altman analysis reported no interaction to subject by method or correlation between differences. The analyses of data replicates was accounted for, instead of the mean values, enabling the comparison of repeatability of methods and obtaining more realistic LoA when considering both within and between subject difference variation. Averaging the subject replications would remove variation within the subject. The calculated LoA would be narrower, especially if both within and between subject variations were similar.

Compared with the traditional paper version of the VAS as a gold standard, the results of this study provide very strong evidence of the validity and reliability of the electronic version of the pain level module on the Interactive Clinics app when assessing acute pain in adults. The mean of pain registered by the subjects only differed by 0.18 units between paper and eVAS, a very small difference compared with 1.30 units considered clinically significant [13]. From the obtained LoA, results measured with eVAS may be 0.79 units below or 1.14 units above paper results (Figure 5). It was also possible to estimate the precision of the LoA as 95% confidence intervals. Considering a 95% confidence interval on lower limit of lower LoA and a 95% confidence interval on upper limit of upper LoA, the clinically significant difference of ±1.30 is still not covered.

One of the most widely adopted instruments to measure pain level is the VAS, which has previously proven its validity and reliability as a pain categories tool [24-26]. Pain is a subjective experience, and therefore it may be difficult to measure in terms of physiologic response unless using complex and expensive materials [27]. Hence, patient’s self-reported measures are valuable and frequently used in clinical and research settings.

In order to compare the assessment of pain between the paper version and the electronic device, acute pain was caused to each subject by means of an algometer. This method has been previously considered easy to operate and reliable [28,29]. Furthermore, it has been validated to determine pain threshold [30,31], and it has been found repeatable and stable [32]. As expected, despite the exact same stimulus of pressure being applied to each subject, individual perception was recorded to be different.

The paper VAS format presents with some limitations, especially when measuring the evolution of pain in noninstitutionalized patients. There are some alternatives. The numeric rating scale (NRS) and the verbal rating scale (VRS) can be performed by phone and have demonstrated different levels of consistency and validity. The VAS showed the highest scores [33,34]. Bijur et al [35] concluded that NRS was strongly correlated with VAS in emergency patients, making NRS suitable for these patients. However, VRS and VAS are not interchangeable when measuring pain, whether chronic pain [36] or chronic/idiopathic, nociceptive, and neuropathic pain [37]. As a consequence, the measurement instrument used before, during, and after a surgical procedure should be the same.

Compared with the VRS and NRS, the eVAS is self-reported and self-administered, which allows an unlimited number of measures regarding research costs from an economic and time perspective. This is a significant advantage, especially considering the increasing number of noninstitutionalized postoperative and chronic pain patients. The electronic devices facilitate documentation management and may encourage active patient participation [6]. Furthermore, the eVAS, in an adequate app framework, automatically enables a precise record of the day and time the assessments have been performed, reducing potential human error and time for data collection. Consequently, pain level can be assessed at different time intervals during the day and as frequently as desired. With increased awareness of a patient’s progression of pain intensity, clinicians may be capable of providing more accurate analgesic strategies and improved clinical management. For example, medication administration can be tailored to prevent symptoms during specific times of the day by increasing its power (dosage or active principle) accordingly.

A recent systematic literature review reported that no comparisons had been made between the VAS in paper-and-pen versus electronic versions for pain assessment [38]. However, previous comparisons have been made regarding appetite. The Apple Newton electronic appetite rating system was determined to be as sensitive and reliable as the paper method [39]. Other studies support the use of electronic versions of the VAS for appetite assessment; however, although no superiority was found in terms of validity, it was highlighted that data are not interchangeable between electronic and paper versions [40-42]. Another study compared eVAS, eNRS, and the electronic version of the Roland-Morris Disability Questionnaire in patients with low back pain [43] and concluded they were comparable with their paper versions.

One main difference reported between the studied app and other previously used devices is the actions that the subjects must perform to confer their results. While in most of the electronic linear scales subjects must place their finger on one end of the line (usually on the left side, corresponding to zero) and slide it until the desired point on the line, in this app subjects simply cast their mark directly on the line, replicating more closely the motion used with the traditional paper VAS. This new feature may provide a higher reliability between devices.

Some limitations should be outlined as part of this study. Although no sequence effect (paper or electronic first) has been demonstrated through an ad hoc previous analysis, future papers may take into account its randomization.

A practical question for future research is whether a single patient using the same device will simply trace the fingerprint left on the screen, especially during successive and repeated recordings. A feasible solution to prevent the patient tracing the previous fingerprint left on the screen is to simply ensure that the subject or data collector cleans the screen after each recording. Regarding our study, it must be noted that all of the electronic measures were made using one single device, a tablet with a 7-inch screen; in order to increase validity, future studies should adopt other tablet screen sizes and include smartphones. Another limitation of the study is that acute pain was initiated to record the desired outcome measured. In order to fully investigate digital symptom progression, future studies may include other categories of pain.

The eVAS on the Interactive Clinics app has been demonstrated to be highly reliable and consistent with paper version results. Fully understanding the impact that pain progression has on individual patients has long been a challenge for clinicians. The introduction of this reliable, safe, and noninvasive mHealth solution may have the potential to achieve enduring changes in improving patient awareness of their progression of pain.

Future research is needed to further explore the feasibility of the app using other tablet screen sizes and smartphones accessible by the wider population. Finally, the introduction of this novel translated research approach may significantly increase the quality of reliable data accessible to clinicians to address pain-related issues.