Validation of Algorithm to Identify Persons with Non-traumatic Spinal Cord Dysfunction in Canada Using Administrative Health Data

Abstract

Administrative health data, such as the hospital Discharge Abstract Database (DAD), can potentially be used to identify patients with non-traumatic spinal cord dysfunction (NTSCD). Algorithms utilizing administrative health data for this purpose should be validated before clinical use. To validate an algorithm designed to identify patients with NTSCD through DAD. DAD between 2006 and 2016 for Southern Alberta in Canada were obtained through Alberta Health Services. Cases of NTSCD were identified using the algorithm designed by the research team. These were then validated by chart review using electronic medical records where possible and paper records where electronic records were unavailable. Measures of diagnostic accuracy including sensitivity, specificity, and positive and negative predictive values and 95% confidence intervals (CI) were computed. Two hundred and eighty cases were identified to have both the administrative codes for neurological impairments and NTSCD etiology. Twenty-eight cases were excluded from analysis as 5 had inadequate medical record information, 17 had traumatic spinal cord injury, and 6 were considered "other" non-spinal cord conditions. Measures of diagnostic accuracy that were computed were sensitivity 97% (95% CI, 94%-98%), specificity 60% (95% CI, 47%-73%), positive predictive value (PPV) 92% (95% CI, 88%-95%), and negative predictive value (NPV) 80% (95% CI, 65%-90%). The most prevalent etiologies were degenerative (36.9%), infection (19.0%), oncology malignant (15.1%), and vascular (10.3%). Our algorithm has high sensitivity and PPV and satisfactory specificity and NPV for the identification of persons with NTSCD using DAD, though the limitations for using this method should be recognized.

A Narrative Review of Pediatric Nontraumatic Spinal Cord Dysfunction

“…Researchers have used the ICD codes to identify patients with SCDys from health databases for research purposes. 12,62,63 Given the rarity of SCDys, population-based studies using health databases that code etiology using the ICD codes are an important source of research data for this group. However, a challenge for researchers is that there is no single ICD-10/11 code specifically for SCDys, and there is no accepted gold standard list of codes for the various conditions causing SCDys.…”

Routinely collected patient data in neurology research: a systematic mapping review

“…A small number of papers (nine articles and six abstracts) were dedicated to developing administrative data algorithms. Typical of these papers is Ho C et al who used a set of rules applied to data stored in a discharge database to identify patients with nontraumatic spinal cord dysfunction [20]. There were relatively few papers using computationally intensive methods such as Natural Language Processing (NLP) (five articles and five abstracts) and machine learning (ML) (five articles and two abstracts).…”

Methods for development of structure, process and outcome indicators for prioritized spinal cord injury rehabilitation Domains: SCI-High Project

“…A priori , the Project Leaders determined that data for all patients with SCI/D was to be collected from the time of rehabilitation admission to 18 months thereafter. The rationale for this decision was that: (1) the current case-finding strategies for patients with non-traumatic SCI 49 are not well established and an international classification of diseases codes for case finding was proposed for the first time in 2018; 50 , 51 (2) the Project Team wanted to allow for data collection beyond one year post-injury due to the current lack of community datasets 52 describing patient outcomes following rehab discharge; (3) recognizing that the median 78 day length-of-stay in Canada, and current length-of-stay targets 53 , 54 may limit the time for outcome indicator data collection; and (4) recognizing the high rates of divorce rates in the first 3 years after injury, 55 and the high rates of depression onset 6 months post discharge 56 and the role for poor coping skills to impede self-management 57 in the first few months following discharge from the inpatient rehabilitation setting. Figure 4 is a conceptual diagram illustrating how elements from the Canadian Institute for Health Information National Rehabilitation Reporting System ( www.cihi.ca/en/national-rehabilitation-reporting-system-metadata ), RHSCIR 3.0 ( www.rickhanseninstitute.org ) and local site health records are planned to be combined with indicator data to enable reporting in the fiscal year 2019–2020.…”