Comparison between self‐reported and actigraphy‐derived sleep measures in patients receiving home parenteral nutrition: Secondary analysis of observational data

Abstract

BackgroundPatients receiving home parenteral nutrition (HPN) frequently report disrupted sleep. However, there are often inconsistencies between objectively measured and questionnaire‐derived sleep measures. We compared sleep measures estimated from wrist actigraphy and self‐report in adults receiving HPN.MethodsIn this secondary analysis, we pooled data from two sleep‐related studies enrolling adults receiving habitual HPN. We compared measures from 7‐day averages of wrist actigraphy against comparable responses to a sleep questionnaire. Sleep measures included bedtime, wake time, time in bed, total sleep time, and sleep onset latency (SOL). Spearman correlation coefficients, Bland‐Altman plots, and linear regression models for each set of sleep measures provided estimates of agreement.ResultsParticipants (N = 35) had a mean age of 52 years, body mass index of 21.6 kg/m2, and 77% identified as female. Correlation coefficients ranged from 0.35 to 0.90, were highest for wake time (r = 0.90) and bedtime (r = 0.74), and lowest for total sleep time (r = 0.35). Actigraphy overestimated self‐reported bedtime, wake time, and total sleep time and underestimated self‐reported time in bed and SOL. Regression coefficients indicated the highest calibration for bedtime and wake time and lower calibration for time in bed, total sleep time, and SOL.ConclusionWe observed strong‐to‐moderate agreement between sleep measures derived from wrist actigraphy and self‐report in adults receiving HPN. Weaker correlations for total sleep time and SOL may indicate low wrist actigraphy sensitivity. Low‐quality sleep resulting from sleep disruptions may have also contributed to an underreporting of perceived sleep quantity and lower concordance.

Home parenteral nutrition, sleep patterns, and depressive symptoms: Secondary analysis of cross‐sectional data

“…Interestingly, higher HPN volume showed an association with shorter sleep onset latency but not wake after sleep onset and remained robust in sensitivity analyses and when T A B L E 2 Associations between HPN characteristics and survey-derived and actigraphy-derived measures of sleep. 18 and, therefore, these findings should be interpreted with caution. Lastly, we found that higher total energy was associated with worse sleep quality.…”

“…The study was a secondary analysis of cross-sectional data pooled from two studies examining sleep patterns in those receiving HPN using comparable subjective and objective measures, as previously conducted. 18 The first was a 7-day observational sleep study aimed at characterizing sleep patterns among 20 adults receiving HPN. 10 The second was a quasi-experimental, controlled, 2-week clinical trial (NCT04743960) aimed at comparing nocturnal (week 1) with daytime (week 2) HPN infusions among adult patients living with short bowel syndrome.…”

“…Eligibility criteria for both protocols have been previously described. 10,18,19 Briefly, enrolled participants were adult men or women who were not pregnant (aged 18-79 years) in the United States receiving HPN for at least 3 months without major changes to current HPN regimen. The clinical trial only enrolled patients living with short bowel syndrome.…”

“…The wrist actigraph was equipped with an accelerometer to detect wrist immobility, a proxy for sleep; this method of monitoring sleep has been validated against the criterion standard of polysomnography 23 and further compared with self-report in this population. 18 Objective sleep measures were calculated using a previously described validated algorithm 10,18…”