The pre-defined criterion to claim feasibility was clearly missed. Only six instead of ≥ 15 among 20 patients fulfilled the patient specific goal of the primary outcome measure. This was mainly due to missing data because the WD was not worn, i.e., low compliance. We see four relevant and potentially remediable factors leading to low compliance: First, there was no direct benefit of study participation for parents and patients; this may be different during interventional trials. Second, there was no feedback on quality of recording and communication with the application; such feedback can be implemented in an updated application. Third, demotivating technical problems mainly regarding communication, can as well be solved using updated hardware or software. Finally, the only side effects reported were local skin irritations; regularly changing the wearing site can prevent them.
This negative main result contrasts with two encouraging findings: The vast majority of participants considered the WD to be suitable, easy and comfortable to wear. And the age range of patients fulfilling the patient specific goal was wide, from 3 to 16 years. This implies that potential future clinical applications of WDs for continuous and at home recording of vital signs can be envisaged from preschool children to adolescents. Infants were not studied here.
In healthy children, continuous monitoring has been reported for heart rate  and physical activity  with wrist-worn WDs. Good compliance for wrist worn WDs has been reported in a large study (n = 886), but the minimal wear time criterion of ≥ 10 hours per day had been much lower than here . In children undergoing chemotherapy for cancer, studies have reported on physical activity assessment by WDs worn on the ankle , waist [33, 34], or clothes , and two small pilot studies have reported on continuous monitoring of temperature with patches in hospitalized patients  and at home . Continuous recording of vital signs beyond temperature and activity using WDs has not yet been reported in this patient group.
The Everion®, the device used here, has been reported to be comfortable during day- and nighttime without displacement in a study of 115 adult patients with epilepsy . Uninterrupted recording is essential for prediction of imminent fever or infections and requires that the WD is comfortable, remains in place and does not hamper the child’s daily activities. In patients with good compliance, these requirements seemed to be fulfilled in our study.
In contrast to most other WDs, assignment of quality scores to recorded vital signs is routinely implemented in the Everion®. These scores were useful to exclude many non-plausible results from analysis. Non-plausible poor quality was systematically assigned to part of the vital signs in very low or very high ranges, which are physiological for children, however. This issue, which may distort the function of pattern search algorithms, may be due to the use of aggregates. Should it remain unchanged when non-aggregated date are used, it might be solved together with the implementation of off-body detection by updating the firmware without changing the hardware.
The study has several limitations. (i) The WDs were studied for only 14 days and compliance might be different during long-term use. (ii) The small number of patients precluded an analysis of patient factors influencing the primary outcome, limits the validity of the skin type assessment, and allowed only rudimentary exploration for specific patterns preceding episodes of fever and infection. (iii) The planned comparison of discrete measurements versus continuously was not feasible because the exact second-wise time point of discrete measurements was not available.
The major strength of this study is that the patients included represent an unbiased sample of children undergoing chemotherapy for cancer from 2 to 16 years of age. The broad assessment of feasibility including parents’ and patients’ opinion and effort for the investigator, allows a comprehensive judgement on feasibility in pediatric patients undergoing chemotherapy for cancer. The problems identified and described above can be addressed in future studies
In conclusion, we found that continuous recording of vital signs by the Everion® WD is feasible across a very wide age range in pediatric patients undergoing chemotherapy for cancer. In the configuration studied, however, the predefined feasibility criterion was not fulfilled due to important compliance problems. These are essentially independent of the WD itself and correspondingly may be resolved in future studies. The results of this study will influence the design of future WD-studies including those aiming to identify patterns predicting fever or infection.