Dataset Information

Comparing Methods for Record Linkage for Public Health Action: Matching Algorithm Validation Study.

ABSTRACT: BACKGROUND:Many public health departments use record linkage between surveillance data and external data sources to inform public health interventions. However, little guidance is available to inform these activities, and many health departments rely on deterministic algorithms that may miss many true matches. In the context of public health action, these missed matches lead to missed opportunities to deliver interventions and may exacerbate existing health inequities. OBJECTIVE:This study aimed to compare the performance of record linkage algorithms commonly used in public health practice. METHODS:We compared five deterministic (exact, Stenger, Ocampo 1, Ocampo 2, and Bosh) and two probabilistic record linkage algorithms (fastLink and beta record linkage [BRL]) using simulations and a real-world scenario. We simulated pairs of datasets with varying numbers of errors per record and the number of matching records between the two datasets (ie, overlap). We matched the datasets using each algorithm and calculated their recall (ie, sensitivity, the proportion of true matches identified by the algorithm) and precision (ie, positive predictive value, the proportion of matches identified by the algorithm that were true matches). We estimated the average computation time by performing a match with each algorithm 20 times while varying the size of the datasets being matched. In a real-world scenario, HIV and sexually transmitted disease surveillance data from King County, Washington, were matched to identify people living with HIV who had a syphilis diagnosis in 2017. We calculated the recall and precision of each algorithm compared with a composite standard based on the agreement in matching decisions across all the algorithms and manual review. RESULTS:In simulations, BRL and fastLink maintained a high recall at nearly all data quality levels, while being comparable with deterministic algorithms in terms of precision. Deterministic algorithms typically failed to identify matches in scenarios with low data quality. All the deterministic algorithms had a shorter average computation time than the probabilistic algorithms. BRL had the slowest overall computation time (14 min when both datasets contained 2000 records). In the real-world scenario, BRL had the lowest trade-off between recall (309/309, 100.0%) and precision (309/312, 99.0%). CONCLUSIONS:Probabilistic record linkage algorithms maximize the number of true matches identified, reducing gaps in the coverage of interventions and maximizing the reach of public health action.

SUBMITTER: Avoundjian T

PROVIDER: S-EPMC7226047 | biostudies-literature | 2020 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Comparing Methods for Record Linkage for Public Health Action: Matching Algorithm Validation Study.

Avoundjian Tigran T Dombrowski Julia C JC Golden Matthew R MR Hughes James P JP Guthrie Brandon L BL Baseman Janet J Sadinle Mauricio M

JMIR public health and surveillance 20200430 2

<h4>Background</h4>Many public health departments use record linkage between surveillance data and external data sources to inform public health interventions. However, little guidance is available to inform these activities, and many health departments rely on deterministic algorithms that may miss many true matches. In the context of public health action, these missed matches lead to missed opportunities to deliver interventions and may exacerbate existing health inequities.<h4>Objective</h4>T ...[more]

PMID: 32352389

Similar Datasets

Project description:BackgroundThe linkage of records which refer to the same entity in separate data collections is a common requirement in public health and biomedical research. Traditionally, record linkage techniques have required that all the identifying data in which links are sought be revealed to at least one party, often a third party. This necessarily invades personal privacy and requires complete trust in the intentions of that party and their ability to maintain security and confidentiality. Dusserre, Quantin, Bouzelat and colleagues have demonstrated that it is possible to use secure one-way hash transformations to carry out follow-up epidemiological studies without any party having to reveal identifying information about any of the subjects - a technique which we refer to as "blindfolded record linkage". A limitation of their method is that only exact comparisons of values are possible, although phonetic encoding of names and other strings can be used to allow for some types of typographical variation and data errors.MethodsA method is described which permits the calculation of a general similarity measure, the n-gram score, without having to reveal the data being compared, albeit at some cost in computation and data communication. This method can be combined with public key cryptography and automatic estimation of linkage model parameters to create an overall system for blindfolded record linkage.ResultsThe system described offers good protection against misdeeds or security failures by any one party, but remains vulnerable to collusion between or simultaneous compromise of two or more parties involved in the linkage operation. In order to reduce the likelihood of this, the use of last-minute allocation of tasks to substitutable servers is proposed. Proof-of-concept computer programmes written in the Python programming language are provided to illustrate the similarity comparison protocol.ConclusionAlthough the protocols described in this paper are not unconditionally secure, they do suggest the feasibility, with the aid of modern cryptographic techniques and high speed communication networks, of a general purpose probabilistic record linkage system which permits record linkage studies to be carried out with negligible risk of invasion of personal privacy.

Project description:BackgroundThe Enhanced Matching System (EMS) is a probabilistic record linkage program developed by the tuberculosis section at Public Health England to match data for individuals across two datasets. This paper outlines how EMS works and investigates its accuracy for linkage across public health datasets.MethodsEMS is a configurable Microsoft SQL Server database program. To examine the accuracy of EMS, two public health databases were matched using National Health Service (NHS) numbers as a gold standard unique identifier. Probabilistic linkage was then performed on the same two datasets without inclusion of NHS number. Sensitivity analyses were carried out to examine the effect of varying matching process parameters.ResultsExact matching using NHS number between two datasets (containing 5931 and 1759 records) identified 1071 matched pairs. EMS probabilistic linkage identified 1068 record pairs. The sensitivity of probabilistic linkage was calculated as 99.5% (95%CI: 98.9, 99.8), specificity 100.0% (95%CI: 99.9, 100.0), positive predictive value 99.8% (95%CI: 99.3, 100.0), and negative predictive value 99.9% (95%CI: 99.8, 100.0). Probabilistic matching was most accurate when including address variables and using the automatically generated threshold for determining links with manual review.ConclusionWith the establishment of national electronic datasets across health and social care, EMS enables previously unanswerable research questions to be tackled with confidence in the accuracy of the linkage process. In scenarios where a small sample is being matched into a very large database (such as national records of hospital attendance) then, compared to results presented in this analysis, the positive predictive value or sensitivity may drop according to the prevalence of matches between databases. Despite this possible limitation, probabilistic linkage has great potential to be used where exact matching using a common identifier is not possible, including in low-income settings, and for vulnerable groups such as homeless populations, where the absence of unique identifiers and lower data quality has historically hindered the ability to identify individuals across datasets.

Project description:Linkage of medical databases, including insurer claims and electronic health records (EHRs), is increasingly common. However, few studies have investigated the behavior and output of linkage software. To determine how linkage quality is affected by different algorithms, blocking variables, methods for string matching and weight determination, and decision rules, we compared the performance of 4 nonproprietary linkage software packages linking patient identifiers from noninteroperable inpatient and outpatient EHRs. We linked datasets using first and last name, gender, and date of birth (DOB). We evaluated DOB and year of birth (YOB) as blocking variables and used exact and inexact matching methods. We compared the weights assigned to record pairs and evaluated how matching weights corresponded to a gold standard, medical record number. Deduplicated datasets contained 69,523 inpatient and 176,154 outpatient records, respectively. Linkage runs blocking on DOB produced weights ranging in number from 8 for exact matching to 64,273 for inexact matching. Linkage runs blocking on YOB produced 8 to 916,806 weights. Exact matching matched record pairs with identical test characteristics (sensitivity 90.48%, specificity 99.78%) for the highest ranked group, but algorithms differentially prioritized certain variables. Inexact matching behaved more variably, leading to dramatic differences in sensitivity (range 0.04-93.36%) and positive predictive value (PPV) (range 86.67-97.35%), even for the most highly ranked record pairs. Blocking on DOB led to higher PPV of highly ranked record pairs. An ensemble approach based on averaging scaled matching weights led to modestly improved accuracy. In summary, we found few differences in the rankings of record pairs with the highest matching weights across 4 linkage packages. Performance was more consistent for exact string matching than for inexact string matching. Most methods and software packages performed similarly when comparing matching accuracy with the gold standard. In some settings, an ensemble matching approach may outperform individual linkage algorithms.

Project description:BackgroundFinancial codes are often used to extract diagnoses from electronic health records. This approach is prone to false positives. Alternatively, queries are constructed, but these are highly center and language specific. A tantalizing alternative is the automatic identification of patients by employing machine learning on format-free text entries.ObjectiveThe aim of this study was to develop an easily implementable workflow that builds a machine learning algorithm capable of accurately identifying patients with rheumatoid arthritis from format-free text fields in electronic health records.MethodsTwo electronic health record data sets were employed: Leiden (n=3000) and Erlangen (n=4771). Using a portion of the Leiden data (n=2000), we compared 6 different machine learning methods and a naïve word-matching algorithm using 10-fold cross-validation. Performances were compared using the area under the receiver operating characteristic curve (AUROC) and the area under the precision recall curve (AUPRC), and F1 score was used as the primary criterion for selecting the best method to build a classifying algorithm. We selected the optimal threshold of positive predictive value for case identification based on the output of the best method in the training data. This validation workflow was subsequently applied to a portion of the Erlangen data (n=4293). For testing, the best performing methods were applied to remaining data (Leiden n=1000; Erlangen n=478) for an unbiased evaluation.ResultsFor the Leiden data set, the word-matching algorithm demonstrated mixed performance (AUROC 0.90; AUPRC 0.33; F1 score 0.55), and 4 methods significantly outperformed word-matching, with support vector machines performing best (AUROC 0.98; AUPRC 0.88; F1 score 0.83). Applying this support vector machine classifier to the test data resulted in a similarly high performance (F1 score 0.81; positive predictive value [PPV] 0.94), and with this method, we could identify 2873 patients with rheumatoid arthritis in less than 7 seconds out of the complete collection of 23,300 patients in the Leiden electronic health record system. For the Erlangen data set, gradient boosting performed best (AUROC 0.94; AUPRC 0.85; F1 score 0.82) in the training set, and applied to the test data, resulted once again in good results (F1 score 0.67; PPV 0.97).ConclusionsWe demonstrate that machine learning methods can extract the records of patients with rheumatoid arthritis from electronic health record data with high precision, allowing research on very large populations for limited costs. Our approach is language and center independent and could be applied to any type of diagnosis. We have developed our pipeline into a universally applicable and easy-to-implement workflow to equip centers with their own high-performing algorithm. This allows the creation of observational studies of unprecedented size covering different countries for low cost from already available data in electronic health record systems.

Dataset Information

Comparing Methods for Record Linkage for Public Health Action: Matching Algorithm Validation Study.

Publications

Comparing Methods for Record Linkage for Public Health Action: Matching Algorithm Validation Study.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets