Childhood Adenotonsillectomy Trial
3.1 Data Acquisition
Data collection for PSG was standardized among various manufactured instruments by use of standardized collection and display montages (see tables in this section). Each unit used for collection was certified before use to maximize signal quality integrity throughout the EDF conversion and import into Compumedics Profusion software used for scoring. Technicians performing CHAT studies also underwent a formal certification procedure.
Due to the new sites that joined CHAT in 2009 (SLU, Boston Childrens, Montifiore), several modifications were made to the standardized collection due to problems with EDF conversion and or equipment limitations at the new sites and or original sites. Although initial CHAT sites were required to use a Nonin Oximeter board and Novametric Capnograph, later CHAT sites were unable to comply with this and thus may have used a Masimo oximeter or BCI capnograph. In addition, with expansion of beds in the original CHAT sites, BCI capnographs were allowed as an alternative to the Novametrix. Data has been generated from 19 studies where both capnographs were used. Analysis results not yet available.
Xltek Oximetry. Xltek units used an internal Masimo oximeter. Nonin external sensor did not convert through EDF. Studies done with Xltek were scored using Masimo oximetry. SLU collected several studies for analysis using an external Nonin so the values could be compared. Preliminary data indicate no significant difference; however, additional reports are being run for further analysis.
Biologic Oximetry. Biologic units used an internal Nonin. However, when units were certified it was noted that the oximetry values after EDF export/import were approximately 1.5% lower in the scoring software than the screen displays in the Biologic software.
Pleth Wave Form. The Pleth wave form was not available with the Biologic, Xltek or Sandman equipment.
Body Position. There was no universal position sensor that worked successfully with all equipment and converted properly through EDF. EDF imported entire recording as non-supine (right). Most sites were able to use a sensor that output numerical values that could be used to manually score Supine. Supine/Non-Supine was manually scored for all other PSGs based on technician notes, when available, if numerical values were not output.
Capnography/EtCO2. Novametrix capnograph was standard to be used for all research data collection. BCI capnograph was approved for use in baseline studies done in clinical lab. With addition of new beds at existing sites and addition of three new sites it was agreed that BCI could be used for both baseline and follow-up studies. Collection of EtCO2 data at Boston and SLU was problematic, particularly with Biologic® equipment. As a result, there is limited useable EtCO2 data from SLU collection and no EtCO2 useable data from Boston or Louisville. The signal wave forms are present in many of these studies, but no numerical data were acquired so report data could not be generated on CO2 levels. Comment: The EtCO2 signal waveform was present on many of these studies but it was not collected in a calibrated channel. Therefore, the automatic analysis of the EtCO2 waveform in the Profusion scoring software could not do the EtCO2 peak analysis of this signal. There were two channels of data from the Capnograph: the waveform [CAP (V)] which had to be in a calibrated channel which is used for the EtCO2 analysis and the numerical EtCO2 peak values [EtCO2 (mmHg)] from the Capnograph shown as trend in another channel [this channel of numbers can not be used for analysis, only used to compare numbers to the peak values from the analysis of the waveform and sleep tech recorded the peak number for EtCO2 on the PSG Signal Integrity form every 30 minutes]. Numerical data for analysis of EtCO2 must come from the waveform channel.
Time/Epoch differences. Edf exports from Biologic® units had start times that differ from the original files and the epoch assignments differ because the instruments pause and restart data collection automatically, which affects timing of the epoch assignments. Adjustments were made in tech notes so scorer could identify correct start time of study.