Article Text

Download PDFPDF
64 Linking statewide hospital admission and mortality datasets to investigate events after opioid-related hospitalizations in pennsylvania
  1. Thomas Songer1,
  2. Zhen-qiang Ma2,
  3. Evelyn Talbott1,
  4. Michelle Mellers1,
  5. Stephen Liu1,
  6. Judith Rager1
  1. 1US University of Pittsburgh
  2. 2US Pennsylvania Dept. of Health


The burden of opioid misuse relies upon information from one surveillance system. Linked data systems to describe the natural history of opioid misuse are needed to clarify the patterns of risk for opioid users.

Methods The project linked hospital admission and mortality statewide data systems in Pennsylvania from 2000–2014 to identify a cohort of subjects with an index opioid-related hospital admission. Index cases were identified from the first presence of ISW7 ICD-9-CM opioid code in a primary or secondary diagnosis field. The risk for subsequent admission (2000–2014) and mortality (2000–2010) and patterns in readmission and underlying cause of death was assessed. 1 73 849 index cases were identified; of whom 1 03 312 were re-hospitalised, and 17 975 died.

Results 53% of the cohort was identified as having a subsequent hospital admission (for any cause) in the follow-up period. The average time to the next admission was 1.2 years, and the probability of being re-admitted in the first year was 33%. This pattern did not change over time of enrollment in the cohort. About 10% of the cohort from 2000–2010 had died by the end of 2010, with a mean time to death of 11 months following the baseline hospital admission. 800 subjects died at the baseline hospital admission. Drug-related deaths (defined by CDC ICD-10 codes) represented 20% of all causes of death in the cohort. The probability of dying the first year following the index admission was 2% and did not change over time. Death rates varied markedly by age group, with high PMRs for drug-related deaths in the younger age groups (15–44 years).

Conclusions Health care use and mortality is significant and elevated in persons with an initial opioid hospital admission.

Significance Linked surveillance data provide depth to describe the natural history of opioid misuse and the risk for mortality.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.