Search

ObjectiveThis article describes the implementation of a privacy-preserving record linkage (PPRL) solution across PCORnet®, the National Patient-Centered Clinical Research Network.Material and MethodsUsing a PPRL solution from Datavant, we quantified the degree of patient overlap across the network and report a de-duplicated analysis of the demographic and clinical characteristics of the PCORnet population.ResultsThere were ∼170M patient records across the responding Network Partners, with ∼138M (81%) of those corresponding to a unique patient. 82.1% of patients were found in a single partner and 14.7% were in 2. The percentage overlap between Partners ranged between 0% and 80% with a median of 0%. Linking patients’ electronic health records with claims increased disease prevalence in every clinical characteristic, ranging between 63% and 173%.DiscussionThe overlap between Partners was variable and depended on timeframe. However, patient data linkage changed the prevalence profile of the PCORnet patient population.ConclusionsThis project was one of the largest linkage efforts of its kind and demonstrates the potential value of record linkage. Linkage between Partners may be most useful in cases where there is geographic proximity between Partners, an expectation that potential linkage Partners will be able to fill gaps in data, or a longer study timeframe.

The Fellegi–Sunter model has been widely used in probabilistic record linkage despite its often invalid conditional independence assumption. Prior research has demonstrated that conditional dependence latent class models yield improved match performance when using the correct conditional dependence structure. With a misspecified conditional dependence structure, these models can yield worse performance. It is, therefore, critically important to correctly identify the conditional dependence structure. Existing methods for identifying the conditional dependence structure include the correlation residual plot, the log-odds ratio check, and the bivariate residual, all of which have been shown to perform inadequately. Bootstrap bivariate residual approach and score test have also been proposed and found to have better performance, with the score test having greater power and lower computational burden. In this paper, we extend the score-test-based approach to account for different conditional dependence structures. Through a simulation study, we develop practical recommendations on the utilisation of the score test and assess the match performance with conditional dependence identified by the proposed method. Performance of the proposed method is further evaluated using a real-world record linkage example. Findings show that the proposed method leads to improved matching accuracy relative to the Fellegi–Sunter model.

Background Data assessing protection conferred from COVID-19 mRNA vaccination and/or prior SARS-CoV-2 infection during Delta and Omicron predominance periods in the United States are limited. Methods This cohort study included persons ≥18 years who had ≥1 health care encounter across 4 health systems and had been tested for SARS-CoV-2 before 26 August 2021. COVID-19 mRNA vaccination and prior SARS-CoV-2 infection defined the exposure. Cox regression estimated hazard ratios (HRs) for the Delta and Omicron periods; protection was calculated as (1−HR)×100%. Results Compared to unvaccinated and previously uninfected persons, during Delta predominance, protection against COVID-19–associated hospitalizations was high for those 2- or 3-dose vaccinated and previously infected, 3-dose vaccinated alone, and prior infection alone (range, 91%–97%, with overlapping 95% confidence intervals [CIs]); during Omicron predominance, estimates were lower (range, 77%–90%). Protection against COVID-19–associated emergency department/urgent care (ED/UC) encounters during Delta predominance was high for those exposure groups (range, 86%–93%); during Omicron predominance, protection remained high for those 3-dose vaccinated with or without a prior infection (76%; 95% CI = 67%–83% and 71%; 95% CI = 67%–73%, respectively). Conclusions COVID-19 mRNA vaccination and/or prior SARS-CoV-2 infection provided protection against COVID-19–associated hospitalizations and ED/UC encounters regardless of variant. Staying up-to-date with COVID-19 vaccination still provides protection against severe COVID-19 disease, regardless of prior infection.

Objectives. To assess the effectiveness of vaccine-induced immunity against new infections, all-cause emergency department (ED) and hospital visits, and mortality in Indiana. Methods. Combining statewide testing and immunization data with patient medical records, we matched individuals who received at least 1 dose of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccines with individuals with previous SARS-CoV-2 infection on index date, age, gender, race/ethnicity, zip code, and clinical diagnoses. We compared the cumulative incidence of infection, all-cause ED visits, hospitalizations, and mortality. Results. We matched 267 847 pairs of individuals. Six months after the index date, the incidence of SARS-CoV-2 infection was significantly higher in vaccine recipients (6.7%) than the previously infected (2.9%). All-cause mortality in the vaccinated, however, was 37% lower than that of the previously infected. The rates of all-cause ED visits and hospitalizations were 24% and 37% lower in the vaccinated than in the previously infected. Conclusions. The significantly lower rates of all-cause ED visits, hospitalizations, and mortality in the vaccinated highlight the real-world benefits of vaccination. The data raise questions about the wisdom of reliance on natural immunity when safe and effective vaccines are available. (Am J Public Health. 2023;113(1):96–104. https://doi.org/10.2105/AJPH.2022.307112 )

BACKGROUND The COVID-19 pandemic has highlighted the inability of health systems to leverage existing system infrastructure in order to rapidly develop and apply broad analytical tools that could inform state- and national-level policymaking, as well as patient care delivery in hospital settings. The COVID-19 pandemic has also led to highlighted systemic disparities in health outcomes and access to care based on race or ethnicity, gender, income-level, and urban-rural divide. Although the United States seems to be recovering from the COVID-19 pandemic owing to widespread vaccination efforts and increased public awareness, there is an urgent need to address the aforementioned challenges. OBJECTIVE This study aims to inform the feasibility of leveraging broad, statewide datasets for population health–driven decision-making by developing robust analytical models that predict COVID-19–related health care resource utilization across patients served by Indiana’s statewide Health Information Exchange. METHODS We leveraged comprehensive datasets obtained from the Indiana Network for Patient Care to train decision forest-based models that can predict patient-level need of health care resource utilization. To assess these models for potential biases, we tested model performance against subpopulations stratified by age, race or ethnicity, gender, and residence (urban vs rural). RESULTS For model development, we identified a cohort of 96,026 patients from across 957 zip codes in Indiana, United States. We trained the decision models that predicted health care resource utilization by using approximately 100 of the most impactful features from a total of 1172 features created. Each model and stratified subpopulation under test reported precision scores >70%, accuracy and area under the receiver operating curve scores >80%, and sensitivity scores approximately >90%. We noted statistically significant variations in model performance across stratified subpopulations identified by age, race or ethnicity, gender, and residence (urban vs rural). CONCLUSIONS This study presents the possibility of developing decision models capable of predicting patient-level health care resource utilization across a broad, statewide region with considerable predictive performance. However, our models present statistically significant variations in performance across stratified subpopulations of interest. Further efforts are necessary to identify root causes of these biases and to rectify them.

BACKGROUND Community Health Information Exchange (HIEs) organizations were originally designed to support individual patient care, but their ability to aggregate health and non-health information about populations offers significant opportunities. OBJECTIVE The aim of this paper is to discuss and illustrate current opportunities for HIEs using the status and trajectory of the Indiana Network for Patient Care (INPC), the oldest and one of the largest HIEs in the United States. METHODS We reviewed the research and clinical applications the INPC has developed and introduced numerous innovations and initiatives. RESULTS HIEs have the positioning and opportunity to provide highly relevant services in today’s healthcare, and public and population health contexts, such as research-ready data sets, state registry collaborations, data commons, population health applications, clinical applications, nationwide interoperability, and support of accountable care organizations. CONCLUSIONS Community HIEs can help advance the practice of medicine and population health and help shift sick-focused to health-focused care. To fully take advantage of their potential, community HIEs must continue to innovate and evolve.

BACKGROUND The United States, unlike other high-income countries, currently has no national unique patient identifier to facilitate health information exchange. Because of security and privacy concerns, Congress, in 1998, prevented the government from promulgating a unique patient identifier. The Health and Human Services funding bill that was enacted in 2019 requires that Health and Human Services report their recommendations on patient identification to Congress. While there are anecdotes of incomplete health care data due to patient misidentification, to date there have been insufficient large-scale analyses measuring improvements to patient care that a unique patient identifier might provide. This lack of measurement has made it difficult for policymakers to balance security and privacy concerns against the value of potential improvements. OBJECTIVE We sought to determine the frequency of serious drug-drug interaction alerts discovered because a pharmacy benefits manager uses a universal patient identifier and estimate undiscovered serious drug-drug interactions because pharmacy benefit managers do not yet fully share patient records. METHODS We conducted a retrospective study of serious drug-drug interaction alerts provided from September 1, 2016 to August 31, 2019 to retail pharmacies by a national pharmacy benefit manager that uses a unique patient identifier. We compared each alert to the contributing prescription and determined whether the unique patient identifier was necessary in order to identify the crossover alert. We classified each alert’s disposition as override, abandonment, or replacement. Using the crossover alert rate and sample population size, we inferred a rate of missing serious drug-drug interaction alerts for the United States. We performed logistic regression in order to identify factors correlated with crossover and alert outcomes. RESULTS Among a population of 49.7 million patients, 242,646 serious drug-drug interaction alerts occurred in 3 years. Of these, 2388 (1.0%) crossed insurance and were discovered because the pharmacy benefit manager used a unique patient identifier. We estimate that up to 10% of serious drug-drug alerts in the United States go undetected by pharmacy benefit managers because of unexchanged information or pharmacy benefit managers that do not use a unique patient identifier. These information gaps may contribute, annually, to up to 6000 patients in the United States receiving a contraindicated medication. CONCLUSIONS Comprehensive patient identification across disparate data sources can help protect patients from serious drug-drug interactions. To better safeguard patients, providers should (1) adopt a comprehensive patient identification strategy and (2) share patient prescription history to improve clinical decision support.

BACKGROUND Emerging interest in precision health and the increasing availability of patient- and population-level data sets present considerable potential to enable analytical approaches to identify and mitigate the negative effects of social factors on health. These issues are not satisfactorily addressed in typical medical care encounters, and thus, opportunities to improve health outcomes, reduce costs, and improve coordination of care are not realized. Furthermore, methodological expertise on the use of varied patient- and population-level data sets and machine learning to predict need for supplemental services is limited. OBJECTIVE The objective of this study was to leverage a comprehensive range of clinical, behavioral, social risk, and social determinants of health factors in order to develop decision models capable of identifying patients in need of various wraparound social services. METHODS We used comprehensive patient- and population-level data sets to build decision models capable of predicting need for behavioral health, dietitian, social work, or other social service referrals within a safety-net health system using area under the receiver operating characteristic curve (AUROC), sensitivity, precision, F1 score, and specificity. We also evaluated the value of population-level social determinants of health data sets in improving machine learning performance of the models. RESULTS Decision models for each wraparound service demonstrated performance measures ranging between 59.2%% and 99.3%. These results were statistically superior to the performance measures demonstrated by our previous models which used a limited data set and whose performance measures ranged from 38.2% to 88.3% (behavioural health: F1 score PP=.01; social work: F1 score PP=.03; dietitian: F1 score P=.001, AUROC P=.001; other: F1 score P=.01, AUROC P=.02); however, inclusion of additional population-level social determinants of health did not contribute to any performance improvements (behavioural health: F1 score P=.08, AUROC P=.09; social work: F1 score P=.16, AUROC P=.09; dietitian: F1 score P=.08, AUROC P=.14; other: F1 score P=.33, AUROC P=.21) in predicting the need for referral in our population of vulnerable patients seeking care at a safety-net provider. CONCLUSIONS Precision health–enabled decision models that leverage a wide range of patient- and population-level data sets and advanced machine learning methods are capable of predicting need for various wraparound social services with good performance.

The conditional independence assumption of the Felligi and Sunter (FS) model in probabilistic record linkage is often violated when matching real-world data. Ignoring conditional dependence has been shown to seriously bias parameter estimates. However, in record linkage, the ultimate goal is to inform the match status of record pairs and therefore, record linkage algorithms should be evaluated in terms of matching accuracy. In the literature, more flexible models have been proposed to relax the conditional independence assumption, but few studies have assessed whether such accommodations improve matching accuracy. In this paper, we show that incorporating the conditional dependence appropriately yields comparable or improved matching accuracy than the FS model using three real-world data linkage examples. Through a simulation study, we further investigate when conditional dependence models provide improved matching accuracy. Our study shows that the FS model is generally robust to the conditional independence assumption and provides comparable matching accuracy as the more complex conditional dependence models. However, when the match prevalence approaches 0% or 100% and conditional dependence exists in the dominating class, it is necessary to address conditional dependence as the FS model produces suboptimal matching accuracy. The need to address conditional dependence becomes less important when highly discriminating fields are used. Our simulation study also shows that conditional dependence models with misspecified dependence structure could produce less accurate record matching than the FS model and therefore we caution against the blind use of conditional dependence models.

CMS-2728 form (Medical Evidence Report) assesses 23 comorbidities chosen to reflect poor outcomes and increased mortality risk. Previous studies questioned the validity of physician reporting on forms CMS-2728. We hypothesize that reporting of comorbidities by computer algorithms identifies more comorbidities than physician completion, and, therefore, is more reflective of underlying disease burden.We collected data from CMS-2728 forms for all 296 patients who had incident ESRD diagnosis and received chronic dialysis from 2005 through 2014 at Indiana University outpatient dialysis centers. We analyzed patients' data from electronic medical records systems that collated information from multiple health care sources. Previously utilized algorithms or natural language processing was used to extract data on 10 comorbidities for a period of up to 10 years prior to ESRD incidence. These algorithms incorporate billing codes, prescriptions, and other relevant elements. We compared the presence or unchecked status of these comorbidities on the forms to the presence or absence according to the algorithms.Computer algorithms had higher reporting of comorbidities compared to forms completion by physicians. This remained true when decreasing data span to one year and using only a single health center source. The algorithms determination was well accepted by a physician panel. Importantly, algorithms use significantly increased the expected deaths and lowered the standardized mortality ratios.Using computer algorithms showed superior identification of comorbidities for form CMS-2728 and altered standardized mortality ratios. Adapting similar algorithms in available EMR systems may offer more thorough evaluation of comorbidities and improve quality reporting.

Automated electronic laboratory reporting (ELR) for public health has many potential advantages, but requires mapping local laboratory test codes to a standard vocabulary such as LOINC. Mapping only the most frequently reported tests provides one way to prioritize the effort and mitigate the resource burden. We evaluated the implications of selective mapping on ELR for public health by comparing reportable conditions from an operational ELR system with the codes in the LOINC Top 2000. Laboratory result codes in the LOINC Top 2000 accounted for 65.3% of the reportable condition volume. However, by also including the 129 most frequent LOINC codes that identified reportable conditions in our system but were not present in the LOINC Top 2000, this set would cover 98% of the reportable condition volume. Our study highlights the ways that our approach to implementing vocabulary standards impacts secondary data uses such as public health reporting.

In this paper, we consider the design for comparing the performance of two binary classification rules, for example, two record linkage algorithms or two screening tests. Statistical methods are well developed for comparing these accuracy measures when the gold standard is available for every unit in the sample, or in a two-phase study when the gold standard is ascertained only in the second phase in a subsample using a fixed sampling scheme. However, these methods do not attempt to optimize the sampling scheme to minimize the variance of the estimators of interest. In comparing the performance of two classification rules, the parameters of primary interest are the difference in sensitivities, specificities, and positive predictive values. We derived the analytic variance formulas for these parameter estimates and used them to obtain the optimal sampling design. The efficiency of the optimal sampling design is evaluated through an empirical investigation that compares the optimal sampling with simple random sampling and with proportional allocation. Results of the empirical study show that the optimal sampling design is similar for estimating the difference in sensitivities and in specificities, and both achieve a substantial amount of variance reduction with an over-sample of subjects with discordant results and under-sample of subjects with concordant results. A heuristic rule is recommended when there is no prior knowledge of individual sensitivities and specificities, or the prevalence of the true positive findings in the study population. The optimal sampling is applied to a real-world example in record linkage to evaluate the difference in classification accuracy of two matching algorithms.

Maintaining value sets is a necessary and distinct activity apart from maintaining recognized controlled vocabularies. As an illustration of one such value set, we evaluate the CDC and Regenstrief versions of the notifiable condition mapping tables (NCMT) and illustrate they are not synchronized. We used practical informatics approaches including heuristic queries and similarity measures to accurately identify more than 800 new candidate reportable LOINC codes. To successfully maintain value sets we must establish a clear strategy for coordinating the value sets and process for disseminating among stakeholders. These stakeholders will likely be distinct from, but interface with, the existing standards development organizations (SDOs).

Infections with Methicillin-Resistant Staphylococcus aureus (MRSA) account for almost 20,000 deaths per year. Early identification of patients with MRSA infection or colonization aids in stopping spread. We compared automated identification of MRSA using HL7 lab result messages to current manual infection control practices at a local hospital during July–September 2008. We used data from infection control providers (ICPs), the microbiology lab, and a Regional Healthcare Information Exchange to assess the accuracy of manual and automated methods. Three hundred seventy MRSA cases were identified from July–September 2008. Manual identification recognized 314 (sensitivity 84.9%, positive predictive value 99.4%) MRSA cases and automated detection from HL7 messages identified 341 (sensitivity 92.2%, positive predictive value 98.8%). Automated processing of HL7 lab report messages is a more sensitive method of capturing MRSA cases than current standard infection control practice, with minimal loss of specificity.

The International Society for Disease Surveillance (ISDS) celebrates its 10th annual meeting with the arrival of the 2011 ISDS Annual Conference, ‘Building the Future of Public Health Surveillance’...

Beginning in 2004, the Indiana State Department of Health (ISDH) partnered with the Regenstrief Institute on a 4-year project to implement a statewide biosurveillance system incorporating more than 110 hospitals. This paper describes our evolving experience with the system including ongoing implementation challenges, how the system has helped to uncover events of public health significance, and future directions. The system currently receives emergency department visit data from 50 hospitals totaling nearly 5,000 visits per day, and is projected to have 65 hospitals connected by August 2006.