Where No Universal Health Care Identifier Exists: Comparison and Determination of the Utility of Score-Based Persons Matching Algorithms Using Demographic Data
Bibliographic record
Abstract
BACKGROUND: A universal health care identifier (UHID) facilitates the development of longitudinal medical records in health care settings where follow up and tracking of persons across health care sectors are needed. HIV case-based surveillance (CBS) entails longitudinal follow up of HIV cases from diagnosis, linkage to care and treatment, and is recommended for second generation HIV surveillance. In the absence of a UHID, records matching, linking, and deduplication may be done using score-based persons matching algorithms. We present a stepwise process of score-based persons matching algorithms based on demographic data to improve HIV CBS and other longitudinal data systems. OBJECTIVE: The aim of this study is to compare deterministic and score-based persons matching algorithms in records linkage and matching using demographic data in settings without a UHID. METHODS: We used HIV CBS pilot data from 124 facilities in 2 high HIV-burden counties (Siaya and Kisumu) in western Kenya. For efficient processing, data were grouped into 3 scenarios within (1) HIV testing services (HTS), (2) HTS-care, and (3) within care. In deterministic matching, we directly compared identifiers and pseudo-identifiers from medical records to determine matches. We used R stringdist package for Jaro, Jaro-Winkler score-based matching and Levenshtein, and Damerau-Levenshtein string edit distance calculation methods. For the Jaro-Winkler method, we used a penalty (р)=0.1 and applied 4 weights (ω) to Levenshtein and Damerau-Levenshtein: deletion ω=0.8, insertion ω=0.8, substitutions ω=1, and transposition ω=0.5. RESULTS: We abstracted 12,157 cases of which 4073/12,157 (33.5%) were from HTS, 1091/12,157 (9.0%) from HTS-care, and 6993/12,157 (57.5%) within care. Using the deterministic process 435/12,157 (3.6%) duplicate records were identified, yielding 96.4% (11,722/12,157) unique cases. Overall, of the score-based methods, Jaro-Winkler yielded the most duplicate records (686/12,157, 5.6%) while Jaro yielded the least duplicates (546/12,157, 4.5%), and Levenshtein and Damerau-Levenshtein yielded 4.6% (563/12,157) duplicates. Specifically, duplicate records yielded by method were: (1) Jaro 5.7% (234/4073) within HTS, 0.4% (4/1091) in HTS-care, and 4.4% (308/6993) within care, (2) Jaro-Winkler 7.4% (302/4073) within HTS, 0.5% (6/1091) in HTS-care, and 5.4% (378/6993) within care, (3) Levenshtein 6.4% (262/4073) within HTS, 0.4% (4/1091) in HTS-care, and 4.2% (297/6993) within care, and (4) Damerau-Levenshtein 6.4% (262/4073) within HTS, 0.4% (4/1091) in HTS-care, and 4.2% (297/6993) within care. CONCLUSIONS: Without deduplication, over reporting occurs across the care and treatment cascade. Jaro-Winkler score-based matching performed the best in identifying matches. A pragmatic estimate of duplicates in health care settings can provide a corrective factor for modeled estimates, for targeting and program planning. We propose that even without a UHID, standard national deduplication and persons-matching algorithm that utilizes demographic data would improve accuracy in monitoring HIV care clinical cascades.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.054 | 0.173 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.005 | 0.004 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.003 |
| Open science | 0.002 | 0.003 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".