Abstract

Introduction. Sickle cell disease (SCD) is a hematologic disorder with a multitude of complications occurring throughout a patient's life. Pain is the most common complication and causes significant morbidity. Due to pain being inherently subjective, patients and medical providers have difficultly treating and managing pain. Therefore, recent efforts focus on better quantification and predictors of pain. These include the use of mobile apps to better record symptoms and wearable devices to record objective measures such as heart rate (HR) and galvanic skin response (GSR). We have previously shown the usefulness and validity of our mHealth app for patients with SCD and recently improved our user interface based on feedback to foster patient engagement.

Wearable technology devices now allow easier capability to record very granular changes in objective measures such as HR and GSR, which have been correlated with pain. In our study, we use our newly revised mHealth app (Technology Recordings to Understand Pain - TRU-Pain) to record symptoms and a wearable device to record objective patient data during treatment. We hypothesize that with an improvement in clinical pain, we will also find an improvement in objective measures during pain treatment.

Methods. Following Institutional Review Board approval, patients were approached within the day hospital when presenting for pain treatment. Once consented, patients were given an iPad with TRU-Pain, instructed on its use, and fitted with a Microsoft Band 2 wearable device. Patients were asked to record all symptoms including pain into TRU-Pain. Patients returned devices prior to discharge. Patients admitted to the hospital continued to use the app and wearable, however, this analysis only includes data collected prior to time of the admission order.

Results. We enrolled 14 adult patients who presented to the Duke University Adult Day Hospital (50% males, ages 21 to 66 years-old). Of the 14 patients, eight had SS disease (57%), five had SC disease (36%). Their average stay for pain treatment within the day hospital was 4 hours, 1 minute. Data from the TRU-Pain app was available for 11/14 patients, with three patients making entries without data saving to the cloud-based server. Nine patients used the TRU-Pain app multiple times during their day hospital visit (median 4, range 2 to 5). All fourteen patients provided wearable data (average data points recorded 6,099) and two were admitted.

Visual analog pain scores (scale 0-10) via TRU-Pain decreased by 2.36 over their stay (0.95 decrease for admitted vs. 2.54 for non-admitted). In contrast, nursing records revealed an average decrease of 3.56 (3 for admitted vs. 3.63 for non-admitted). When examining initial and discharge vitals, HR decreased on average 9.64 bpm. Based on wearable data, HR changed by -0.82 bpm/hr or -3.31 bpm over their stay (95% confidence interval (CI) for slope [-2.26, 0.06]; Figure 1). GSR increased by 26,000 kOhms over the average session time, or 6,450 kOhm/hr (95% CI for slope [-126, 28,200]). RR interval standard deviation increased for 11/14 patients (79%).

Conclusions. We found interesting insight from our TRU-Pain app and wearable data. TRU-Pain was useful in recording additional symptoms and had differences from pain scores recorded by nurses. As hoped in the treatment of pain, a decrease in pain scores was observed with an expected less dramatic effect when patients are admitted. Although data was not saved for three patients in our pilot study, improvements have been made to now confirm data flow.

We also found changes in objective measures. As expected, a decrease in heart rate was observed over the time with medication administration. The RR interval variability should be investigated further as it suggests a potential correlation with successful discharge, i.e. increased RR variability has been previously seen with improved pain. GSR showed an increase over each patient's stay, potentially due to patient discomfort or distress while in the day hospital.

Further analysis of this data will focus on establishing correlations between HR and pain and connecting HR data with medication administration. We aim to expand our enrollment further to better understand the wide data variability seen in pilot enrollment. We are optimistic that more in-depth analysis is likely to continue to reveal interesting insight into objective data to potentially correlate with subjective pain data.

Disclosures

Jonassaint: Sicklesoft: Other: Owner. Shah: Novartis: Other: Speaker; Alexion: Other: Speaker.

Author notes

*

Asterisk with author names denotes non-ASH members.