SUDDEN employs a population-based data collection process to establish the epidemiology of out-of-hospital sudden unexpected death.
Methodology – Case Ascertainment
The SUDDEN registry is established by electronically filtering death certificates to determine potential OHSUD cases, procuring records on these cases, and manually reviewing the potential cases to determine final inclusion or exclusion.
Record filtering is done by calendar year at the county level and includes victims that were both residents and died in the county. The death certificate data can be amended for a period of time after death. To allow for these changes to cycle through the system, the database is accessed and filtered 9 months after the close of the calendar year, i.e., 2015 death data will be filtered in September of 2016.
Death certificate filters include age of decedent (18-64), location of death (dead on arrival at hospital, decedent’s home, or other), and manner of death (natural only). A final filter is applied that excludes natural deaths attributable to cancer (ICD10 codes C00–C97 as identified anywhere on the death certificate). These filtered cases are deemed “Potential Sudden Unexpected Death” victims and medical records within five years of date of death (by report of medical examiner and EMS clinical data) is procured.
Clinical EMS data is provided electronically by the county EMS data manager. This data is linked to the “Potential Sudden Unexpected Death” cases using Registry Plus™ Link Plus Software. The EMS report includes 29 data elements related to time of death, time since last seen alive, resuscitation status, and chief narrative among others.
The initial medical record request is sent to the signatory of death as identified on the death certificate. Each medical record is reviewed by a post-graduate intern to identify other providers where records should be requested, i.e., the signatory of death or a primary care provider references a referral to a specialist, such as a cardiologist.
Medical examiner cases are so identified on the death certificate. For these subjects the report of medical examiner, which is considered a public record and does not require consent from next of kin, is requested from the state medical examiner’s office. At minimum, the report consists of a report of medical examiner (external examination) and toxicology screen results. Autopsy data is included on subjects where a full post-mortem examination is completed.
Medical record requests are submitted to the local hospital(s), as identified in the Medicare hospital service area database, based on the subject’s zip code.
If no response to a record request is received within 90 days, a follow-up phone call is made to the provider. Record request, receipt and/or follow up are electronically tracked.
Two post-graduate interns, using established protocols, review each case to determine final inclusion or exclusion of the victim. Review criteria are free-living (excluding those who die in a hospital, prison, rehabilitation facility, or receiving hospice care) and manner (natural). Expected deaths, broadly defined as those that are directly attributable to a recent event, such as a drug overdose and evidenced by a lethal toxicology report, or disease, such as end-stage cancer, are excluded. Subjects with sub-lethal levels of recreational drugs are included. If the post-graduate intern reviewers cannot agree on inclusion/exclusion of a case, it is discussed between them with the addition of other post-graduate interns, before coming to a consensus. All cases that pass the death certificate filtering and the manual screening process are deemed “confirmed OHSUD”.
Demographic data is captured from the death certificate and medical records are abstracted in a systematic manner for relevant clinical data. Core medical record abstraction focuses on identification of chronic diseases and other relevant clinical data. A process is in place to create specialized data abstraction tools and abstract data, for specific projects.
Victim residence and location of death are geotagged to allow mapping of each occurrence. Data is entered in REDCap, a clinical data collection platform. Quality assurance checks are performed routinely and the processes are modified as indicated. Method manuals and analysis files are freely disseminated to researchers and the public.
More detailed information on the project methodology and protocol is available in the SUDDEN manual below.