Utilising International Statistical Classification of Diseases and Related Health Conditions (ICD)-10 Australian Modification Classifications of "Health Conditions" to Achieve Population Health Surveillance in an Australian Spinal Cord Injury Cohort.

Abstract

Retrospective, non-randomised, registry controlled. To develop a conceptual ICD-10 taxonomic framework for population health surveillance across all-phases of spinal cord injury and disorders (SCI/D). Public Hospital Admitted Patient Care (APC) collection, South Australian Dept. Health, South Australia, Australia. A core ICD-10-Australian Modification (AM) coded dataset was retrieved from the APC hospital patient admission collection (2012-2017). Search filters and key words referenced to the National Library of Medicine thesaurus identified and quantified incident SCI/D cases. Incident SCI/D case data held in the Australian Spinal Cord Injury Registry (ASCIR) of South Australia (2012-2017) tested fidelity. Data linkage to the South Australian Death Registry controlled for cohort attrition. Both unadjusted and case-mix adjusted core data set yields were evaluated. Outcomes were assessed in terms of APC frequency difference (Δ%) versus ASCIR. 3,504 APC cases were extracted, of which 504 (mean, SD age 55 ± 20 yrs; 348 [69%] male, 202 [39%] traumatic; 135 [32%]) cervical; 51 [10.1%] thoracic and (16 [3.2%]) lumbar met criteria. Comparator data were 385 ASCIR new index cases mean, SD age 56 ± 19 yrs, 229 [75%] male, 162 [42%] traumatic. Case-mix adjusted analysis yielded 336 (APC Δ33%) all-cause incident cases (vs. ASCIR -13 Δ%) and 131 incident cases of traumatic aetiologies (vs. ASCIR -19 Δ%). The ICD-10 core "Health Condition" data-set assembled extends our understanding of SCI/D epidemiology and with further development may create a cost-efficient and sustainable framework that will improve health system performance and equity within and between countries. The Lifetime Support Authority of South Australia sponsored the study.