{"id":{"repo_id":"mcmaster","oai_identifier":"oai:macsphere.mcmaster.ca:11375/33837"},"canonical_url":"https://search.ndltd.org/etd/mcmaster/oai:macsphere.mcmaster.ca:11375/33837","repository":{"repo_id":"mcmaster","name":"McMaster University","base_url":"https://macsphere.mcmaster.ca/server/oai/request"},"display":{"title":"Predicting the Severity of Cardiac Implantable Electronic Device Infection with a Desirability of Outcome Ranking Scale: a secondary analysis of the PADIT Trial","abstract":"Background. Risk scores for cardiac implantable electronic device (CIED) infection predict whether an infection occurs and often collapse infections of differing severity. The Desirability of Outcome Ranking (DOOR) framework ranks each patient's overall outcome on an ordered scale but has only been used to compare treatments in trials. This thesis asked whether the severity of CIED infection can be predicted from baseline characteristics, and whether an ordinal representation gives information beyond binary prediction. Methods. Secondary analysis of the Prevention of Arrhythmia Device Infection Trial (PADIT), a cluster-randomized crossover trial of 19,603 patients at 28 centres. A full-cohort model ranked every patient at one year across four tiers, from no infection through non-hospitalized and hospitalized infection to death. An events-only model ranked the 384 infected patients from mild through moderate and severe infection to death. Both models used elastic-net penalized ordinal logistic regression under a pre-specified proportional-odds, partial proportional-odds and multinomial fallback sequence, multiple imputation, and optimism-corrected bootstrap internal validation. Discrimination was summarized by the ordinal C-statistic and by per-boundary c-statistics. Results. After optimism correction, the full-cohort model reached an ordinal C-statistic of 0.61, with boundary c-statistics of 0.68, 0.70 and 0.72 at the infection, hospitalization and mortality boundaries, respectively. The events-only model reached 0.52, with 0.59, 0.60 and 0.69, respectively. Baseline characteristics separated the extremes of the severity scale but did not reliably distinguish adjacent tiers. Age reduced the odds of infection while increasing the odds of death, which a binary composite would conceal. Conclusions. Using DOOR, CIED infection severity is predictable at the extremes of the scale, not within it, something a binary outcome would have obscured. In principle, the tier definitions, pre-specified ordinal pipeline and sample-size arithmetic transfer to other conditions whose outcomes span a severity gradient.","abstract_html":"Background. Risk scores for cardiac implantable electronic device (CIED) infection predict whether an infection occurs and often collapse infections of differing severity. The Desirability of Outcome Ranking (DOOR) framework ranks each patient&#x27;s overall outcome on an ordered scale but has only been used to compare treatments in trials. This thesis asked whether the severity of CIED infection can be predicted from baseline characteristics, and whether an ordinal representation gives information beyond binary prediction. Methods. Secondary analysis of the Prevention of Arrhythmia Device Infection Trial (PADIT), a cluster-randomized crossover trial of 19,603 patients at 28 centres. A full-cohort model ranked every patient at one year across four tiers, from no infection through non-hospitalized and hospitalized infection to death. An events-only model ranked the 384 infected patients from mild through moderate and severe infection to death. Both models used elastic-net penalized ordinal logistic regression under a pre-specified proportional-odds, partial proportional-odds and multinomial fallback sequence, multiple imputation, and optimism-corrected bootstrap internal validation. Discrimination was summarized by the ordinal C-statistic and by per-boundary c-statistics. Results. After optimism correction, the full-cohort model reached an ordinal C-statistic of 0.61, with boundary c-statistics of 0.68, 0.70 and 0.72 at the infection, hospitalization and mortality boundaries, respectively. The events-only model reached 0.52, with 0.59, 0.60 and 0.69, respectively. Baseline characteristics separated the extremes of the severity scale but did not reliably distinguish adjacent tiers. Age reduced the odds of infection while increasing the odds of death, which a binary composite would conceal. Conclusions. Using DOOR, CIED infection severity is predictable at the extremes of the scale, not within it, something a binary outcome would have obscured. In principle, the tier definitions, pre-specified ordinal pipeline and sample-size arithmetic transfer to other conditions whose outcomes span a severity gradient.","abstract_has_math":false,"creators":["Elliott-Mohamed, Shakeap"],"institution":null,"degree_name":null,"degree_level":null,"degree_discipline":null,"degree_department":"Health Research Methodology","school":null,"contributors":[],"advisors":["Mertz, Dominik"],"committee_chairs":[],"committee_members":[],"year":2026,"date_issued":"2026","date_published":"2026","updated_at":"2026-11-18T19:36:13Z","subjects":[],"languages":[],"rights":[],"rights_urls":[],"identifier_entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://doi.org/10.71548/25607"],"render_values":[{"text":"https://doi.org/10.71548/25607","href":"https://doi.org/10.71548/25607","code":true}]}],"person_identifier_entries":[]},"links":{"outbound_url":"https://hdl.handle.net/11375/33837","outbound_label":"Handle","outbound_source":"dc:identifier.uri","record_url":"https://hdl.handle.net/11375/33837","record_label":"Repository record","record_kind":"handle","record_source":"dc:identifier.uri","download_url":null,"download_label":null,"download_source":null,"access_status":null,"embargo_until":null,"access_note":null},"source_record":{"url":"https://macsphere.mcmaster.ca/server/oai/request?verb=GetRecord&metadataPrefix=dim&identifier=oai%3Amacsphere.mcmaster.ca%3A11375%2F33837","prefix":"dim"},"metadata_groups":[{"id":"people","label":"People","entries":[{"key":"dc:contributor.advisor","label":"Advisor","values":["Mertz, Dominik"]},{"key":"dc:contributor.department","label":"Department","values":["Health Research Methodology"]},{"key":"dc:creator","label":"Author","values":["Elliott-Mohamed, Shakeap"]}]},{"id":"academic_context","label":"Academic Context","entries":[{"key":"dc:date.accessioned","label":"Dc Date Accessioned","values":["2026-10-14T22:27:45Z"]},{"key":"dc:date.issued","label":"Date","values":["2026"]},{"key":"dc:type","label":"Dc Type","values":["Thesis"]}]},{"id":"identifiers","label":"Identifiers","entries":[{"key":"dc:identifier.uri","label":"Identifier URI","values":["https://hdl.handle.net/11375/33837","https://doi.org/10.71548/25607"]}]},{"id":"additional","label":"Additional Metadata","entries":[{"key":"dc:description.abstract","label":"Abstract","values":["Background. Risk scores for cardiac implantable electronic device (CIED) infection predict whether an infection occurs and often collapse infections of differing severity. The Desirability of Outcome Ranking (DOOR) framework ranks each patient's overall outcome on an ordered scale but has only been used to compare treatments in trials. This thesis asked whether the severity of CIED infection can be predicted from baseline characteristics, and whether an ordinal representation gives information beyond binary prediction. Methods. Secondary analysis of the Prevention of Arrhythmia Device Infection Trial (PADIT), a cluster-randomized crossover trial of 19,603 patients at 28 centres. A full-cohort model ranked every patient at one year across four tiers, from no infection through non-hospitalized and hospitalized infection to death. An events-only model ranked the 384 infected patients from mild through moderate and severe infection to death. Both models used elastic-net penalized ordinal logistic regression under a pre-specified proportional-odds, partial proportional-odds and multinomial fallback sequence, multiple imputation, and optimism-corrected bootstrap internal validation. Discrimination was summarized by the ordinal C-statistic and by per-boundary c-statistics. Results. After optimism correction, the full-cohort model reached an ordinal C-statistic of 0.61, with boundary c-statistics of 0.68, 0.70 and 0.72 at the infection, hospitalization and mortality boundaries, respectively. The events-only model reached 0.52, with 0.59, 0.60 and 0.69, respectively. Baseline characteristics separated the extremes of the severity scale but did not reliably distinguish adjacent tiers. Age reduced the odds of infection while increasing the odds of death, which a binary composite would conceal. Conclusions. Using DOOR, CIED infection severity is predictable at the extremes of the scale, not within it, something a binary outcome would have obscured. In principle, the tier definitions, pre-specified ordinal pipeline and sample-size arithmetic transfer to other conditions whose outcomes span a severity gradient."]},{"key":"dc:description.degree","label":"Dc Description Degree","values":["Master of Health Sciences (MSc)"]},{"key":"dc:title","label":"Title","values":["Predicting the Severity of Cardiac Implantable Electronic Device Infection with a Desirability of Outcome Ranking Scale: a secondary analysis of the PADIT Trial"]},{"key":"others:access_status","label":"Others Access Status","values":["open.access"]}]}],"canonical_facts":{"dc:contributor.advisor":["Mertz, Dominik"],"dc:contributor.department":["Health Research Methodology"],"dc:creator":["Elliott-Mohamed, Shakeap"],"dc:date.accessioned":["2026-10-14T22:27:45Z"],"dc:date.issued":["2026"],"dc:description.abstract":["Background. Risk scores for cardiac implantable electronic device (CIED) infection predict whether an infection occurs and often collapse infections of differing severity. The Desirability of Outcome Ranking (DOOR) framework ranks each patient's overall outcome on an ordered scale but has only been used to compare treatments in trials. This thesis asked whether the severity of CIED infection can be predicted from baseline characteristics, and whether an ordinal representation gives information beyond binary prediction. Methods. Secondary analysis of the Prevention of Arrhythmia Device Infection Trial (PADIT), a cluster-randomized crossover trial of 19,603 patients at 28 centres. A full-cohort model ranked every patient at one year across four tiers, from no infection through non-hospitalized and hospitalized infection to death. An events-only model ranked the 384 infected patients from mild through moderate and severe infection to death. Both models used elastic-net penalized ordinal logistic regression under a pre-specified proportional-odds, partial proportional-odds and multinomial fallback sequence, multiple imputation, and optimism-corrected bootstrap internal validation. Discrimination was summarized by the ordinal C-statistic and by per-boundary c-statistics. Results. After optimism correction, the full-cohort model reached an ordinal C-statistic of 0.61, with boundary c-statistics of 0.68, 0.70 and 0.72 at the infection, hospitalization and mortality boundaries, respectively. The events-only model reached 0.52, with 0.59, 0.60 and 0.69, respectively. Baseline characteristics separated the extremes of the severity scale but did not reliably distinguish adjacent tiers. Age reduced the odds of infection while increasing the odds of death, which a binary composite would conceal. Conclusions. Using DOOR, CIED infection severity is predictable at the extremes of the scale, not within it, something a binary outcome would have obscured. In principle, the tier definitions, pre-specified ordinal pipeline and sample-size arithmetic transfer to other conditions whose outcomes span a severity gradient."],"dc:description.degree":["Master of Health Sciences (MSc)"],"dc:identifier.uri":["https://hdl.handle.net/11375/33837","https://doi.org/10.71548/25607"],"dc:title":["Predicting the Severity of Cardiac Implantable Electronic Device Infection with a Desirability of Outcome Ranking Scale: a secondary analysis of the PADIT Trial"],"dc:type":["Thesis"],"others:access_status":["open.access"]},"updated_at":"2026-11-18T19:36:13Z"}