Preventox - a Digital Intervention for Patients With Intoxications in Intensive Care

Background Internet-based healthcare interventions offer an alternative to conventional care and may reach patients who would otherwise not seek help, for instance owing to stigma. Intensive care provides a natural point of contact with, and an opportunity to influence patients with substance use disorders, however, research on psychiatric interventions in the intensive care context is lacking. At present, referral to psychiatric or addiction services is offered only where the responsible physician deems it relevant, and constitutes the sole aftercare option following intoxication requiring intensive care. Preventox is an internet-based self-help tool developed by The Stockholm Centre for Dependency Disorders together with the medical intensive care unit Södersjukhuset. In-depth interviews with patients who have received intensive care for intoxications have been conducted prior to study commencement, in order to provide a basis for modifying the application so that it better corresponds to the intensive care context and to the patient group concerned. Aim To determine whether a larger randomised controlled trial (RCT) would be feasible in practice, and to evaluate whether the self-help tool Preventox could contribute to reduced mortality and to a reduction in readmissions and emergency department visits within somatic acute care related to alcohol and illicit drug use. Methods A two-arm randomised controlled pragmatic pilot/feasibility study in patients treated in an intensive care unit or intermediate care unit in Stockholm for intoxication with alcohol or illicit drugs, withdrawal seizures, or delirium tremens. A power calculation for the intervention study conducted as a full RCT indicated that approximately 300 patients per arm would be required to demonstrate a 10% reduction in the readmission rate. The present study is instead planned as an exploratory feasibility study, with patients recruited over a limited period of approximately one year. Recruitment and randomisation take place once the patient is somatically ready for discharge from intensive or intermediate care, immediately prior to discharge. All patients receive information about the study, and consent is obtained. Participants are thereafter enrolled and randomised, using the RedCap software, either to the intervention arm with access to Preventox or to standard discharge procedures. As some patients are discharged in a disoriented state, randomisation is in such cases performed at a later stage on the general ward by outreach intensive care staff. Patient recruitment takes place during a predefined period. Inclusion criteria: patient aged \>18 years; treated at the IMA/MIVA units at Södersjukhuset, during a defined period; assigned a discharge code according to ICD-10 relating to unintentional or intentional poisoning by intoxicants (F10.0-F17.0, F19.0), or a discharge code according to ICD-10 relating to alcohol-induced withdrawal with delirium (F10.4), or treated for seizures related to alcohol or benzodiazepine withdrawal; Swedish- or English-speaking. Exclusion criteria: patients without a Swedish personal identity number. All patients are followed up after 12 months via the National Patient Register, the Swedish Intensive Care Registry and the Cause of Death Register with respect to readmission, emergency department visits within somatic acute care, and death at 1, 3 and 12 months after the index admission. Outcome measures: the primary outcomes are readmissions and emergency department visits within somatic acute care at 1, 3 and 12 months after the index admission, related to intoxication with alcohol and/or illicit drugs, withdrawal seizures, or delirium tremens. The secondary outcome is mortality at 1, 3 and 12 months after the admission. The feasibility of the study will be analysed descriptively on the basis of the feasibility measures of acceptability, recruitment and resources, together with preliminary outcomes. For readmission and death, the investigators will principally employ the Kaplan-Meier method. Binary outcomes will be analysed as proportions, with p-values estimated using the chi-square test or Fisher's exact test, depending on the number of events. Continuous outcomes will be examined using parametric or non-parametric methods, such as the t-test or the Wilcoxon test, depending on the variable and its distribution. The purpose is to test the entire study procedure ahead of a full-scale RCT.