Skip to main content
. 2022 Aug 9;8(3):e36879. doi: 10.2196/36879

Table 2.

Summary of key technical and procedural deployment challenges and resulting iterative changes.

Deployment number Total days of active data collection Technical and procedural deployment challenges How system and deployment procedures were changed or improved and lessons learned
1. Install: 190 min; teardown: 35 min 12

  • Lengthy installation time (due in part to smart watches not properly paired with base station; teaching took 45 min).

  • Unable to remotely monitor smart watches due to bug in code logic; this required 2 members of engineering team to make additional trip to dyad home to fix.

  • Inconsistent delivery of EMAsa on caregiver smart watch.

  • Patient stopped wearing smart watch in final days of deployment due to a fall.

  • Created standardized predeployment protocol checklists for both engineers and nurses to streamline deployment installation (eg, asking better dyad screening questions about size of home; developed environmental sensor placement protocol).

  • Cross-trained nurse team members to help engineers place environmental sensors to expedite installation process.

  • Established time goal of 1 h for installation; 30 min for teardown.

  • Revised structure of daily EMAs; decreased smart watch touchscreen sensitivity; added a “do not disturb/sleep” option on smart watch app.

  • Created a “ground truth” daily log for patients and caregivers to record important events that may occur during deployment (such as a fall or injury).
2. Install: 75 min; teardown: 38 min 9

  • Patient reported they had stable internet, but this was not the case when we arrived in home. Mobile hot spot was set up.

  • Smart watch battery life lasting 6-7 h (vs desired 10-12 h); patient smart watch had to be factory reset due to running out of power, which resulted in loss of data.

  • EMAs not generating or coming at wrong time; smart watches not displaying correct date or time; base station went offline and did not connect properly to hot spot.

  • Smart watches “locking” after deployment resulting in difficulty offloading collected data.

  • Patient consented to study alone in clinic; caregiver unaware of pending deployment until study team arrived at dyad home.

  • Ask more detailed questions about internet and cellular service before in-home visit; be prepared to set up mobile hot spot if needed.

  • Allow more time during installation for participants to practice using app and answering EMAs.

  • Investigation regarding battery life undertaken.

  • Avoid plugging in base station to switch-controlled electrical outlet.

  • Implemented automatic data download script to download smart watch data when they are charging to prevent any data loss.

  • Enhanced predeployment testing.

  • Changed recruitment and consenting processes to ensure caregiver aware of scheduled deployment.
3. Install: 95 min; teardown: 45 min 12

  • Environmental sensors would not stick to wood paneling with standard 3M strips.

  • Participant confusion regarding EMAs; did not feel like they could answer some questions properly.

  • Issues with button press activation of EMAs due to patient neuropathy (numbness in fingers).

  • Battery life of smart watches still problematic, lasting 4-5 h.

  • Smart watches displaying correct data/time, but daily EMAs behaving inconsistently, not coming at all or generating at wrong time.

  • Ensure other measures are available to adhere environmental sensors to walls, such as sticky putty.

  • Added “unsure” option to relevant EMAs.

  • Changed all EMAs to "touch to wake" or screen tap.

  • Refined sampling times for heart rate and accelerometer and operating system settings to optimize battery life.

  • Changed daily EMA to be manually available between 5 PM to midnight with a reminder sent at 8:30 PM.
4. Install: 75 min; teardown: 47 min 14

  • Continued issues with daily EMAs not generating at correct times and smart watch battery life.

  • Caregiver did not understand she should continue to wear the smart watch even if she is not physically with patient.

  • Smart watch time going out of sync after battery dies.

  • Safety concerns for study team related to unsecured firearms in dyad home.

  • System lock turned on to help with time sync issues with smart watches; code changed to help with processing power and accelerometer efficiency.

  • Smart watch wearing instructions revised.

  • Began deploying an Android smart phone to help sync the time and date on the smart watch when the smart watch battery dies.

  • Created home-safety protocol for team.
5. Install: 100 min; teardown: 100 minb 15

  • Blue light on environmental sensor in bedroom kept patient awake at night.

  • One Bluetooth beacon that was placed on top of refrigerator fell into the freezer.

  • One environmental sensor lost connectivity to the system and was not able to be put back online.

  • Patient smart watch not seen with remote monitoring; possibly due to system lock out turned off (to help with time sync issue and prevent smart watch from powering down) or from bug in code; follow-up EMAs not consistently being generated; random buzzes; long lag time with “touch-to-wake” feature of smart watch.

  • New operating system update of the smart watches came with battery consumption reduction mode called “doze mode”; this interfered with EMAs being generated.

  • Ensure tape is placed over environmental sensors to prevent sleep disturbance.

  • Caution with placement of Bluetooth beacons.

  • Adding redundant environmental sensors in monitored rooms to ensure adequate data capture.

  • Code changed to ensure smart watches do not go into “doze mode” and to address other inconsistencies with EMA delivery.

aEMA: ecological momentary assessment.

bIncreased teardown time primarily due to particularly social or talkative dyad; also, iPads for survey data collection took longer to use with this deployment.