176 | Chapter 7 Materials and Methods In this study, a retrospective database was used 3. All included patients (n = 242) were laryngectomized between January 2000 and December 2012 and in regular follow-up in the Netherlands Cancer Institute. The last date of follow-up was January 05, 2017. After the initial inclusion of 242 patients, patients with less than two VPs replacements or no follow-up data were excluded. This resulted in 194 included patients (figure 1). Patients were analyzed based on the used type of VP and indication for TL. All VPs analyzed in this article are manufactured by Atos Medical AB (Hörby, Sweden). Analysis were done for the total group and two subgroups. The following data were collected for each VP replacement: date of insertion, replacement or removal, type and size of the VP, and the reason for replacement or removal. VPG1 beholds patients only using regular VPs (Provox2 or Provox Vega). VPG2 are patients alternating between regular and problem-solving VPs (ActiValve Light, Strong or Xtra Strong). There were no patients only using problem-solving VPs. Patient Group 1 (PG1) underwent primary laryngectomy with or without additional treatment. Patient Group 2 (PG2) had either a salvage laryngectomy for recurrence, second primary tumor or a laryngectomy for a functional reason. All VP replacements were performed by qualified and trained speech-language-pathologists, residents or Head-and-Neck surgeons. Statistical analysis Descriptive analysis was used to summarize patient and device characteristics. The VP lifetime was measured in days, calculated from the insertion date to the date of removal or last follow-up date. Kaplan Meier analyses were used to assess device lifetimes and a moment for prophylactic replacement. Device lifetime of VPs ongoing at the end of the observation period are right-censored, as well as device lifetimes of VPs still in situ when the patient was lost to follow-up or died. Standard deviations (SDs) were used to describe variability in device lifetime, and Mann-Whitney-U was used to test for differences in device lifetime and variability. Statistical analyses were performed in R version 4.1.0 and SPSS version 27.0. Prophylactic voice prosthesis replacement prediction model In this study PVPR is considered feasible if at least 70% of VPs are replaced before leaking, allowing a 30% leakage fraction as acceptable. This cut-off point is chosen by consensus of the research team and called DeviceLife70 (DL70, expressed in days). The current standard policy (Wait to Leak, WtL) has a 100% leakage fraction. To train the prediction
RkJQdWJsaXNoZXIy MTk4NDMw