From "wading through treacle" to "making haste, slowly" in patient data sharing: A comprehensive yet parsimonious model of drivers and challenges to data sharing based on an EPaCCS evaluation and four pre-existing literature reviews

Aims: This study aimed to identify comprehensively the challenges and drivers encountered by Electronic Palliative Care Coordination System (EPaCCS) projects in the context of challenges and drivers in other projects on data sharing for individual care (also referred to as Health Information Exchange, HIE). It aimed to organise them in a parsimonious framework that underpins specific and non-trivial recommendations for steps forward.Data and methods: Primary data comprised 40 in-depth interviews with healthcare professionals from general practice, out-of-hours, specialist palliative care and hospital services; patients and carers; project team members and decision makers in Cambridgeshire, UK. Transcripts amounted to approximately 300,000 words. Secondary data were extracted from four pre-existing literature reviews on Health Information Exchange and Health Information Technology implementation covering 135 studies. A seven-stage analysis process was employed.Results: We reduced an initial set of >1,800 parameters into >500 challenges and >300 drivers to implementing EPaCCS and other data sharing projects. Less than a quarter of the 800+ parameters were associated primarily with the IT solution. These challenges and drivers were further condensed into an action-guiding, strategy-informing framework of nine types of “pure challenges”, drawing parallels between patient data sharing and other broad and complex domains of sociotechnical or social practice; four types of “pure drivers”, defined in terms of whether they were internal or external to the IT solution and project team; and nine types of “oppositional or ambivalent forces”, representing factors perceived simultaneously as a challenge and a driver. Conclusions: Teams working on data sharing projects may need to focus less on refining their IT tools and more on shaping the social interactions and structural and contextual parameters in the midst of which they are configured. The high number of “ambivalent forces” speaks of the vital importance for data sharing projects of skills in eliciting stakeholders’ assumptions; managing conflict; and navigating multiple needs, interests and “worldviews”, amongst others.

Resource Profile: The Regenstrief Institute COVID-19 Research Data Commons (CoRDaCo)

The primary objective of the COVID-19 Research Data Commons (CoRDaCo) is to provide broad and efficient access to a large corpus of clinical data related to COVID-19 in Indiana, facilitating research and discovery. This curated collection of data elements provides information on a significant portion of COVID-19 positive patients in the State from the beginning of the pandemic, as well as two years of health information prior its onset. CoRDaCo combines data from multiple sources, including clinical data from a large, regional health information exchange, clinical data repositories of two health systems, and state laboratory reporting and vital records, as well as geographic-based social variables. Clinical data cover information such as healthcare encounters, vital measurements, laboratory orders and results, medications, diagnoses, the Charlson Comorbidity Index and Pediatric Early Warning Score, COVID-19 vaccinations, mechanical ventilation, restraint use, intensive care unit and ICU and hospital lengths of stay, and mortality. Interested researchers can visit ridata.org or email [email protected] to discuss access to CoRDaCo.

Identification of patients at risk of new onset heart failure: Utilizing a large statewide health information exchange to train and validate a risk prediction model

Background New-onset heart failure (HF) is associated with poor prognosis and high healthcare utilization. Early identification of patients at increased risk incident-HF may allow for focused allocation of preventative care resources. Health information exchange (HIE) data span the entire spectrum of clinical care, but there are no HIE-based clinical decision support tools for diagnosis of incident-HF. We applied machine-learning methods to model the one-year risk of incident-HF from the Maine statewide-HIE. Methods and results We included subjects aged ≥ 40 years without prior HF ICD9/10 codes during a three-year period from 2015 to 2018, and incident-HF defined as assignment of two outpatient or one inpatient code in a year. A tree-boosting algorithm was used to model the probability of incident-HF in year two from data collected in year one, and then validated in year three. 5,668 of 521,347 patients (1.09%) developed incident-HF in the validation cohort. In the validation cohort, the model c-statistic was 0.824 and at a clinically predetermined risk threshold, 10% of patients identified by the model developed incident-HF and 29% of all incident-HF cases in the state of Maine were identified. Conclusions Utilizing machine learning modeling techniques on passively collected clinical HIE data, we developed and validated an incident-HF prediction tool that performs on par with other models that require proactively collected clinical data. Our algorithm could be integrated into other HIEs to leverage the EMR resources to provide individuals, systems, and payors with a risk stratification tool to allow for targeted resource allocation to reduce incident-HF disease burden on individuals and health care systems.

A Data-Driven Iterative Approach for Semi-automatically Assessing the Correctness of Medication Value Sets: A Proof of Concept Based on Opioids

Background Value sets are lists of terms (e.g., opioid medication names) and their corresponding codes from standard clinical vocabularies (e.g., RxNorm) created with the intent of supporting health information exchange and research. Value sets are manually-created and often exhibit errors. Objectives The aim of the study is to develop a semi-automatic, data-centric natural language processing (NLP) method to assess medication-related value set correctness and evaluate it on a set of opioid medication value sets. Methods We developed an NLP algorithm that utilizes value sets containing mostly true positives and true negatives to learn lexical patterns associated with the true positives, and then employs these patterns to identify potential errors in unseen value sets. We evaluated the algorithm on a set of opioid medication value sets, using the recall, precision and F1-score metrics. We applied the trained model to assess the correctness of unseen opioid value sets based on recall. To replicate the application of the algorithm in real-world settings, a domain expert manually conducted error analysis to identify potential system and value set errors. Results Thirty-eight value sets were retrieved from the Value Set Authority Center, and six (two opioid, four non-opioid) were used to develop and evaluate the system. Average precision, recall, and F1-score were 0.932, 0.904, and 0.909, respectively on uncorrected value sets; and 0.958, 0.953, and 0.953, respectively after manual correction of the same value sets. On 20 unseen opioid value sets, the algorithm obtained average recall of 0.89. Error analyses revealed that the main sources of system misclassifications were differences in how opioids were coded in the value sets—while the training value sets had generic names mostly, some of the unseen value sets had new trade names and ingredients. Conclusion The proposed approach is data-centric, reusable, customizable, and not resource intensive. It may help domain experts to easily validate value sets.

Semantics Management for a Regional Health Information System in Italy by CTS2 and FHIR

An infrastructure for the management of semantics is being developed to support the regional health information exchange in Veneto – an Italian region which has about 5 million inhabitants. Terminology plays a key role in the management of the information fluxes of the Veneto region, in which the management of electronic health record is given great attention. An architecture for the management of the semantics of laboratory reports has been set up, adopting standards by HL7. The system has been initially developed according to the common terminology service release 2 (CTS2) standard and, in order to overcome complexities of CTS2 is being revised according to the Fast Healthcare Interoperability Resources (FHIR) standard, which has been subsequently introduced. Aspects of CST2 and of FHIR have been considered in order to retain most suitable aspects of both. This integration can be regarded as most worthwhile.