Validation study in four health-care databases: upper gastrointestinal bleeding misclassification affects precision but not magnitude of drug-related upper gastrointestinal bleeding risk

Quality of heart failure registration in primary care: observations from 1 million electronic health records in the Amsterdam Metropolitan Area

OBJECTIVES

Proper coding of heart failure (HF) in electronic health records (EHRs) is an important prerequisite for adequate care and research towards this vulnerable patient population. We set out to evaluate the accuracy of registration of HF diagnoses in primary care EHRs.

METHODS

In a routine primary care database covering the Amsterdam Metropolitan Area, we identified all episodes of care with International Classification of Primary Care (ICPC) codes K77 (decompensatio cordis) or K84.03 (cardiomyopathy) up to 31/12/2021. We also performed two text-based searches to identify HF episodes without an appropriate ICPC-code. An expert panel evaluated all ICPC and text matches for congruence between the assigned codes and notes.

RESULTS

From a database of 968,433 records we identified 19,106 patients (2.0 %) with a total of 24,011 ICPC-coded HF episodes. Removal of 1,324 episodes found to concern other or uncertain diagnoses and inclusion of 4,582 validated HF episodes identified through text search led to exclusion of 909 (overregistration: 4.8 %) and inclusion of 2,266 additional patients (underregistration: 11.1 %). The inclusion of miscoded HF episodes advanced the first known date of HF diagnosis in 3.9 % of records, with a median shift of 3.45 years. Episode-level underregistration decreased significantly over time, from 23.8 % in 2006 to 10.0 % in 2021.

CONCLUSIONS

While there is improvement over time, there are still substantial levels of over- and underregistration of HF, emphasizing the need for cautious interpretation of ICPC-coded data. The findings contribute to the understanding of HF registration issues in primary care and provide insights for improving registration practices.

A methodology to assess the population size and estimate the needed resources for new licensed medications by combining clinical and administrative databases: The example of glycated haemoglobin in type 2 diabetes

PURPOSE

To develop and validate a model to estimate glycated haemoglobin (HbA1c) values in patients with type 2 diabetes mellitus (T2DM) using a clinical data source, with the aim to apply this equation to administrative databases.

METHODS

Using a primary care and administrative Italian databases, namely the Health Search database (HSD) and the ReS (Ricerca e Salute) database, we selected all patients aged 18 years or older on 31 December 2018 being diagnosed with T2DM and without prior prescription of sodium-glucose cotransporter-2 (SGLT-2) inhibitors. We included patients prescribed with and adherent to metformin. HSD was used to develop and test (using 2019 data as well) the algorithm imputing HbA1c values ≥7% according to a series of covariates. The algorithm was gathered by combining beta-coefficients being estimated by logistic regression models using complete case (excluding missing values) and imputed (after multiple imputation) dataset. The final algorithm was applied to ReS database using the same covariates.

RESULTS

The tested algorithms were able to explain 17%-18% variation in assessing HbA1c values. Good discrimination (70%) and calibration were obtained as well. The best algorithm (three) cut-offs, namely those providing correct classifications ranging 66%-70% was therefore calculated and applied to ReS database. By doing so, from 52 999 (27.9, 95% CI: 27.7%-28.1%) to 74 250 (40.1%, 95% CI: 38.9%-39.3%) patients were estimated with HbA1c ≥7%.

CONCLUSION

Through this methodology, healthcare authorities should be able to quantify the population eligible to a new licensed medication, such as SGLT-2 inhibitors, and to simulate scenarios to assess reimbursement criteria according to precise estimates.

Doctors Identify Hemorrhage Better during Chart Review when Assisted by Artificial Intelligence

OBJECTIVES

This study evaluated if medical doctors could identify more hemorrhage events during chart review in a clinical setting when assisted by an artificial intelligence (AI) model and medical doctors' perception of using the AI model.

METHODS

To develop the AI model, sentences from 900 electronic health records were labeled as positive or negative for hemorrhage and categorized into one of 12 anatomical locations. The AI model was evaluated on a test cohort consisting of 566 admissions. Using eye-tracking technology, we investigated medical doctors' reading workflow during manual chart review. Moreover, we performed a clinical use study where medical doctors read two admissions with and without AI assistance to evaluate performance when using and perception of using the AI model.

RESULTS

The AI model had a sensitivity of 93.7% and a specificity of 98.1% on the test cohort. In the use studies, we found that medical doctors missed more than 33% of relevant sentences when doing chart review without AI assistance. Hemorrhage events described in paragraphs were more often overlooked compared with bullet-pointed hemorrhage mentions. With AI-assisted chart review, medical doctors identified 48 and 49 percentage points more hemorrhage events than without assistance in two admissions, and they were generally positive toward using the AI model as a supporting tool.

CONCLUSION

Medical doctors identified more hemorrhage events with AI-assisted chart review and they were generally positive toward using the AI model.

Intermittent Proton Pump Inhibitor Therapy in Low-Risk Non-Variceal Upper Gastrointestinal Bleeding May Be Significantly Cost-Saving

BACKGROUND

High-dose proton pump inhibitor (PPI) therapy, given either intermittently or continuously for non-variceal upper gastrointestinal bleeding (NV-UGIB), is efficacious. Using intermittent PPI for low-risk patients may be cost-saving. Our objective was to estimate the annual cost savings if all low-risk NV-UGIB patients received intermittent PPI therapy.

METHODS

Patients who presented to hospital in Calgary, Alberta, who received a PPI for NV-UGIB from July 2015 to March 2017 were identified using ICD-10 codes. Patients were stratified into no endoscopy, high-risk, and low-risk lesion groups and further subdivided into no PPI, oral PPI, intermittent intravenous (IV), and continuous IV subgroups. Average length of stay (LOS) in each subgroup and costs were calculated.

RESULTS

We identified 4141 patients with NV-UGIBs, (median age 61, 57.4% male). One-thousand two-hundred and thirty-one low-risk patients received continuous IV PPI, with an average LOS of 6.8 days (95% CI 6.2-7.3) versus 4.9 days (95% CI 3.9-5.9) for intermittent IV patients. If continuous IV PPI patients instead received intermittent IV PPI, 3852 patient days and CAD 11,714,390 (2017 CAD)/year could be saved.

CONCLUSIONS

Using real-world administrative data, we demonstrate that a sizable portion of low-risk patients with NV-UGIB who were given continuous IV PPI if switched to intermittent IV therapy could generate significant potential cost savings.

Cluster analysis of angiotensin biomarkers to identify antihypertensive drug treatment in population studies

BACKGROUND

The recent progress in molecular biology generates an increasing interest in investigating molecular biomarkers as markers of response to treatments. The present work is motivated by a study, where the objective was to explore the potential of the molecular biomarkers of renin-angiotensin-aldosterone system (RAAS) to identify the undertaken antihypertensive treatments in the general population. Population-based studies offer an opportunity to assess the effectiveness of treatments in real-world scenarios. However, lack of quality documentation, especially when electronic health record linkage is unavailable, leads to inaccurate reporting and classification bias.

METHOD

We present a machine learning clustering technique to determine the potential of measured RAAS biomarkers for the identification of undertaken treatments in the general population. The biomarkers were simultaneously determined through a novel mass-spectrometry analysis in 800 participants of the Cooperative Health Research In South Tyrol (CHRIS) study with documented antihypertensive treatments. We assessed the agreement, sensitivity and specificity of the resulting clusters against known treatment types. Through the lasso penalized regression, we identified clinical characteristics associated with the biomarkers, accounting for the effects of cluster and treatment classifications.

RESULTS

We identified three well-separated clusters: cluster 1 (n = 444) preferentially including individuals not receiving RAAS-targeting drugs; cluster 2 (n = 235) identifying angiotensin type 1 receptor blockers (ARB) users (weighted kappa κw = 74%; sensitivity = 73%; specificity = 83%); and cluster 3 (n = 121) well discriminating angiotensin-converting enzyme inhibitors (ACEi) users (κw = 81%; sensitivity = 55%; specificity = 90%). Individuals in clusters 2 and 3 had higher frequency of diabetes as well as higher fasting glucose and BMI levels. Age, sex and kidney function were strong predictors of the RAAS biomarkers independently of the cluster structure.

CONCLUSIONS

Unsupervised clustering of angiotensin-based biomarkers is a viable technique to identify individuals on specific antihypertensive treatments, pointing to a potential application of the biomarkers as useful clinical diagnostic tools even outside of a controlled clinical setting.

Positive Predictive Value of Non-Traumatic Bleeding Diagnoses in the Danish National Patient Register

Purpose

The majority of bleeding diagnoses in the Danish National Patient Registry have not been validated despite extensive use in epidemiological research. Therefore, we examined the positive predictive value (PPV) of non-traumatic bleeding diagnoses in the Danish National Patient Registry.

Study Design

Population-based validation study.

Patients and Methods

Based on a manual review of electronic medical records, we estimated the PPV of diagnostic coding (International Classification of Diseases, Tenth Revision (ICD-10)) for non-traumatic bleeding for all patients ≥65 years of age with any hospital contact in the North Denmark Region during March-December 2019 as registered in the Danish National Patient Registry. We calculated PPVs and associated 95% confidence intervals (CI) for non-traumatic bleeding diagnoses overall and stratified according to primary or secondary diagnosis, and according to major anatomical sites.

Results

A total of 907 electronic medical records were available for review. The population mean age was 79.33 years (standard deviation (SD)=7.73) and 57.6% were males. Primary bleeding diagnoses accounted for 766 of the records and 141 were secondary bleeding diagnoses. The overall PPV for bleeding diagnoses was 94.0% (95% CI: 92.3-95.4). The PPV was 98.7% (95% CI: 97.6-99.3) for the primary diagnoses and 68.8% (95% CI: 60.7-75.9) for the secondary diagnoses. When stratified according to subgroups of major anatomical sites, the PPVs ranged between 94.1% and 100% for the primary diagnoses, and between 53.8% and 100% for secondary diagnoses.

Conclusion

The overall validity of non-traumatic bleeding diagnoses in the Danish National Patient Registry is high and considered acceptable for epidemiological research. However, PPVs were substantially higher for primary than for secondary diagnosis.

Incident peptic ulcers and concomitant treatment of direct oral anticoagulants and oral bisphosphonates-a real-world cohort study

Oral bisphosphonates and direct oral anticoagulants are related to upper gastrointestinal ulcers. The present study investigated whether concomitant use of these drugs increase the risk of upper gastrointestinal ulcers and report no increased risk of upper gastrointestinal ulcers compared to the use of either drug alone, when individuals with previous upper gastrointestinal ulcers are excluded.

INTRODUCTION

This study examines whether concomitant use of oral bisphosphonates (oBP) and direct oral anticoagulants (DOAC) increases the risk of peptic ulcers more than any drug alone.

METHODS

A population-based cohort study was performed. We sampled a cohort of oBP and DOAC users from a sample of 2,622,742 individuals, consisting of diabetes patients and age- and gender-matched controls, obtained from the Danish National Patient Register. The exposures were concomitant use of oBP and DOAC and single use of DOAC and single use of oBP. The primary endpoint was the first incident peptic ulcer. Information on exposure and outcome were collected from national registries. The period of observation was from 01.01.2008 until 31.12.2018. Unadjusted and adjusted Cox regressions were performed.

RESULTS

8077 individuals received concomitant treatment with DOAC and oBP; 96,451 individuals used DOAC and no oBP; and 118,675 used oBP and no DOAC. The mean duration of follow-up was 1.9 years for concomitant users, 2.5 years for DOAC users, and 4.5 years for oBP users. A total of 4742 individuals with incident peptic ulcers were collected. We observed an increased risk of incident ulcer in users of DOAC and oBP compared to single DOAC treatment in the adjusted analysis (HR = 1.23, 95% CI: 1.03; 1.48). However, the effects were abolished when excluding individuals with a previous ulcer. We observed an increased risk of incident ulcer in users of DOAC and oBP compared to users of oBP in the adjusted model (HR = 1.34, 95% CI: 1.11; 1.63).

CONCLUSION

Based on our results, concomitant use of oBP and DOAC is associated with a slight increase in the risk of peptic ulcers compared to either drug alone. The prescribing physician should weigh the slight increased risk of ulcer in concomitant users of oBP and DOAC with beneficial reductions in stroke and fractures.

Modern Learning from Big Data in Critical Care: Primum Non Nocere

Large and complex data sets are increasingly available for research in critical care. To analyze these data, researchers use techniques commonly referred to as statistical learning or machine learning (ML). The latter is known for large successes in the field of diagnostics, for example, by identification of radiological anomalies. In other research areas, such as clustering and prediction studies, there is more discussion regarding the benefit and efficiency of ML techniques compared with statistical learning. In this viewpoint, we aim to explain commonly used statistical learning and ML techniques and provide guidance for responsible use in the case of clustering and prediction questions in critical care. Clustering studies have been increasingly popular in critical care research, aiming to inform how patients can be characterized, classified, or treated differently. An important challenge for clustering studies is to ensure and assess generalizability. This limits the application of findings in these studies toward individual patients. In the case of predictive questions, there is much discussion as to what algorithm should be used to most accurately predict outcome. Aspects that determine usefulness of ML, compared with statistical techniques, include the volume of the data, the dimensionality of the preferred model, and the extent of missing data. There are areas in which modern ML methods may be preferred. However, efforts should be made to implement statistical frameworks (e.g., for dealing with missing data or measurement error, both omnipresent in clinical data) in ML methods. To conclude, there are important opportunities but also pitfalls to consider when performing clustering or predictive studies with ML techniques. We advocate careful valuation of new data-driven findings. More interaction is needed between the engineer mindset of experts in ML methods, the insight in bias of epidemiologists, and the probabilistic thinking of statisticians to extract as much information and knowledge from data as possible, while avoiding harm.

Risk of drug-related upper gastrointestinal bleeding in the total population of the Netherlands: a time-trend analysis

Objective

Many prescribed and over-the-counter medications, for example, non-steroidal anti-inflammatory drugs (NSAIDs) are associated with upper gastrointestinal bleeding (UGIB). Recently, a decrease in prescribing of NSAIDs was observed in the Netherlands, but whether a similar decreasing trend could be observed in the incidence of severe UGIB (either fatal or requiring hospitalisation), contingent on medication prescription, is unknown.

Design

We conducted a cohort study using Dutch national statistics on pharmacy claims, hospitalisation and mortality between 2013 and 2018. We explored the incidence of sex-specific and age-specific severe UGIB in four (sub)populations: (A) total population, (B) without a filled prescrption for NSAIDs, (C) without filled prescriptions for NSAIDs and antithrombotic agents, (D) without any risk factors for UGIB.

Results

The cumulative incidence of severe UGIB did not decrease throughout the study period, regardless of the subgroup analysis. In the total population, it was 199 per 100 000 inhabitants (95% CI 197 to 201) in 2013–2014 and 260 (95% CI 258 to 263) in 2017–2018. The absolute risk of severe UGIB was 50% lower in the subgroup B than in the full cohort. It decreased further by 50% in the subgroup D when compared with subgroup B. The risk of severe UGIB was 1.5–1.9 fold higher in young women than in young men; an indication of over-the-counter NSAIDs use being more prevalent in women than men in this age group.

Conclusion

We found no evidence to support a relationship between reduced prescribing of NSAIDs and the incidence of severe UGIB in the Netherlands since 2013. The relationship was also not observed when we removed the effect of risk factors.

Risk and Outcome of Venous and Arterial Thrombosis in Patients With Cirrhosis: A Danish Nation-wide Cohort Study

BACKGROUND AND AIMS

Cirrhosis affects hemostasis, but its effects across the spectrum of thromboses remain poorly understood. We examined risks and outcomes of venous and arterial thrombosis.

APPROACH AND RESULTS

We used nation-wide Danish health care registries to identify outpatients with cirrhosis and a sex- and age-matched comparison cohort without cirrhosis from the general population. Patients with cirrhosis and comparators were followed until they had a venous thromboembolism (VTE), acute myocardial infarction (AMI), or ischemic stroke (IS) or died. We computed absolute risks and HRs of thrombosis and compared outcomes after thrombosis. We included 5,854 patients with cirrhosis (median Model for End-Stage Liver Disease score, 9; interquartile range, 7-13), and their risk of any of the thrombotic events was 0.8% after 1 year and 6.3% after 10 years. They were more likely than the 23,870 matched comparators to have a VTE (adjusted hazard ratio [aHR], 2.0; 95% CI, 1.5-2.6) or IS (aHR, 1.7; 95% CI, 1.3-2.3), but not AMI (aHR, 0.7; 95% CI, 0.5-0.9). Among patients with cirrhosis, decompensation increased the risk of AMI, but not the other thromboses. Following thrombosis, patients with cirrhosis had higher 90-day mortality than comparators (after VTE: 17% vs. 7%; after AMI: 27% vs. 5%; after IS: 10% vs. 7%) and were less likely to receive antithrombotic treatment.

CONCLUSIONS

Patients with cirrhosis had an increased risk of VTE and IS, but not AMI. Among patients with cirrhosis, decompensation increased the risk of AMI, exclusively. Mortality after thrombosis was higher in patients with cirrhosis than in other patients. These findings are relevant for decisions about antithrombotic prophylaxis in patients with cirrhosis.

Deep learning detects and visualizes bleeding events in electronic health records

BACKGROUND

Bleeding is associated with a significantly increased morbidity and mortality. Bleeding events are often described in the unstructured text of electronic health records, which makes them difficult to identify by manual inspection.

OBJECTIVES

To develop a deep learning model that detects and visualizes bleeding events in electronic health records.

PATIENTS/METHODS

Three hundred electronic health records with International Classification of Diseases, Tenth Revision diagnosis codes for bleeding or leukemia were extracted. Each sentence in the electronic health record was annotated as positive or negative for bleeding. The annotated sentences were used to develop a deep learning model that detects bleeding at sentence and note level.

RESULTS

On a balanced test set of 1178   sentences, the best-performing deep learning model achieved a sensitivity of 0.90, specificity of 0.90, and negative predictive value of 0.90. On a test set consisting of 700 notes, of which 49 were positive for bleeding, the model achieved a note-level sensitivity of 1.00, specificity of 0.52, and negative predictive value of 1.00. By using a sentence-level model on a note level, the model can explain its predictions by visualizing the exact sentence in a note that contains information regarding bleeding. Moreover, we found that the model performed consistently well across different types of bleedings.

CONCLUSIONS

A deep learning model can be used to detect and visualize bleeding events in the free text of electronic health records. The deep learning model can thus facilitate systematic assessment of bleeding risk, and thereby optimize patient care and safety.

Characterizing Bias Due to Differential Exposure Ascertainment in Electronic Health Record Data

Data derived from electronic health records (EHR) are heterogeneous with availability of specific measures dependent on the type and timing of patients' healthcare interactions. This creates a challenge for research using EHR-derived exposures because gold-standard exposure data, determined by a definitive assessment, may only be available for a subset of the population. Alternative approaches to exposure ascertainment in this case include restricting the analytic sample to only those patients with gold-standard exposure data available (exclusion); using gold-standard data, when available, and using a proxy exposure measure when the gold standard is unavailable (best available); or using a proxy exposure measure for everyone (common data). Exclusion may induce selection bias in outcome/exposure association estimates, while incorporating information from a proxy exposure via either the best available or common data approaches may result in information bias due to measurement error. The objective of this paper was to explore the bias and efficiency of these three analytic approaches across a broad range of scenarios motivated by a study of the association between chronic hyperglycemia and five-year mortality in an EHR-derived cohort of colon cancer survivors. We found that the best available approach tended to mitigate inefficiency and selection bias resulting from exclusion while suffering from less information bias than the common data approach. However, bias in all three approaches can be severe, particularly when both selection bias and information bias are present. When risk of either of these biases is judged to be more than moderate, EHR-based analyses may lead to erroneous conclusions.

Gastrointestinal Disorders and Risk of First-Ever Ischemic Stroke

BACKGROUND AND PURPOSE

Recent studies suggest that alteration of the normal gut microbiome contributes to atherosclerotic burden and cardiovascular disease. While many gastrointestinal diseases are known to cause disruption of the normal gut microbiome in humans, the clinical impact of gastrointestinal diseases on subsequent cerebrovascular disease remains unknown. We conducted an exploratory analysis evaluating the relationship between gastrointestinal diseases and ischemic stroke.

METHODS

We performed a retrospective cohort study using claims between 2008 and 2015 from a nationally representative 5% sample of Medicare beneficiaries. We included only beneficiaries ≥66 years of age. We used previously validated diagnosis codes to ascertain our primary outcome of ischemic stroke. In an exploratory manner, we categorized gastrointestinal disorders by anatomic location, disease chronicity, and disease mechanism. We used Cox proportional hazards models to examine associations of gastrointestinal disorder categories and ischemic stroke with adjustment for demographics and established vascular risk factors.

RESULTS

Among a mean of 1 725 246 beneficiaries in each analysis, several categories of gastrointestinal disorders were associated with an increased risk of ischemic stroke after adjustment for established stroke risk factors. The most notable positive associations included disorders of the stomach (hazard ratio, 1.17 [95% CI, 1.15-1.19]) and functional (1.16 [95% CI, 1.15-1.17]), inflammatory (1.13 [95% CI, 1.12-1.15]), and infectious gastrointestinal disorders (1.13 [95% CI, 1.12-1.15]). In contrast, we found no associations with stroke for diseases of the anus and rectum (0.97 [95% CI, 0.94-1.00]) or neoplastic gastrointestinal disorders (0.97 [95% CI, 0.94-1.00]).

CONCLUSIONS

In exploratory analyses, several categories of gastrointestinal disorders were associated with an increased risk of future ischemic stroke after adjustment for demographics and established stroke risk factors.

Risk of Upper Gastrointestinal Bleeding and Gastroduodenal Ulcers in Persons With Schizophrenia: A Danish Cohort Study

There is little evidence about gastrointestinal (GI) disorders in patients with schizophrenia. We examined association of schizophrenia with upper GI bleeding (UGIB) and nonbleeding ulcers and associated risk factors and mortality.

Identification of validated case definitions for medical conditions used in primary care electronic medical record databases: a systematic review

Objectives

Data derived from primary care electronic medical records (EMRs) are being used for research and surveillance. Case definitions are required to identify patients with specific conditions in EMR data with a degree of accuracy. The purpose of this study is to identify and provide a summary of case definitions that have been validated in primary care EMR data.

Materials and Methods

We searched MEDLINE and Embase (from inception to June 2016) to identify studies that describe case definitions for clinical conditions in EMR data and report on the performance metrics of these definitions.

Results

We identified 40 studies reporting on case definitions for 47 unique clinical conditions. The studies used combinations of International Classification of Disease version 9 (ICD-9) codes, Read codes, laboratory values, and medications in their algorithms. The most common validation metric reported was positive predictive value, with inconsistent reporting of sensitivity and specificity.

Discussion

This review describes validated case definitions derived in primary care EMR data, which can be used to understand disease patterns and prevalence among primary care populations. Limitations include incomplete reporting of performance metrics and uncertainty regarding performance of case definitions across different EMR databases and countries.

Conclusion

Our review found a significant number of validated case definitions with good performance for use in primary care EMR data. These could be applied to other EMR databases in similar contexts and may enable better disease surveillance when using clinical EMR data. Consistent reporting across validation studies using EMR data would facilitate comparison across studies.

Systematic review registration

PROSPERO CRD42016040020 (submitted June 8, 2016, and last revised June 14, 2016).

Bleeding in cardiac patients prescribed antithrombotic drugs: electronic health record phenotyping algorithms, incidence, trends and prognosis

BACKGROUND

Clinical guidelines and public health authorities lack recommendations on scalable approaches to defining and monitoring the occurrence and severity of bleeding in populations prescribed antithrombotic therapy.

METHODS

We examined linked primary care, hospital admission and death registry electronic health records (CALIBER 1998-2010, England) of patients with newly diagnosed atrial fibrillation, acute myocardial infarction, unstable angina or stable angina with the aim to develop algorithms for bleeding events. Using the developed bleeding phenotypes, Kaplan-Meier plots were used to estimate the incidence of bleeding events and we used Cox regression models to assess the prognosis for all-cause mortality, atherothrombotic events and further bleeding.

RESULTS

We present electronic health record phenotyping algorithms for bleeding based on bleeding diagnosis in primary or hospital care, symptoms, transfusion, surgical procedures and haemoglobin values. In validation of the phenotype, we estimated a positive predictive value of 0.88 (95% CI 0.64, 0.99) for hospitalised bleeding. Amongst 128,815 patients, 27,259 (21.2%) had at least 1 bleeding event, with 5-year risks of bleeding of 29.1%, 21.9%, 25.3% and 23.4% following diagnoses of atrial fibrillation, acute myocardial infarction, unstable angina and stable angina, respectively. Rates of hospitalised bleeding per 1000 patients more than doubled from 1.02 (95% CI 0.83, 1.22) in January 1998 to 2.68 (95% CI 2.49, 2.88) in December 2009 coinciding with the increased rates of antiplatelet and vitamin K antagonist prescribing. Patients with hospitalised bleeding and primary care bleeding, with or without markers of severity, were at increased risk of all-cause mortality and atherothrombotic events compared to those with no bleeding. For example, the hazard ratio for all-cause mortality was 1.98 (95% CI 1.86, 2.11) for primary care bleeding with markers of severity and 1.99 (95% CI 1.92, 2.05) for hospitalised bleeding without markers of severity, compared to patients with no bleeding.

CONCLUSIONS

Electronic health record bleeding phenotyping algorithms offer a scalable approach to monitoring bleeding in the population. Incidence of bleeding has doubled in incidence since 1998, affects one in four cardiovascular disease patients, and is associated with poor prognosis. Efforts are required to tackle this iatrogenic epidemic.

Quantifying outcome misclassification in multi-database studies: The case study of pertussis in the ADVANCE project

BACKGROUND

The Accelerated Development of VAccine beNefit-risk Collaboration in Europe (ADVANCE) is a public-private collaboration aiming to develop and test a system for rapid benefit-risk (B/R) monitoring of vaccines using European healthcare databases. Event misclassification can result in biased estimates. Using different algorithms for identifying cases of Bordetella pertussis (BorPer) infection as a test case, we aimed to describe a strategy to quantify event misclassification, when manual chart review is not feasible.

METHODS

Four participating databases retrieved data from primary care (PC) setting: BIFAP: (Spain), THIN and RCGP RSC (UK) and PEDIANET (Italy); SIDIAP (Spain) retrieved data from both PC and hospital settings. BorPer algorithms were defined by healthcare setting, data domain (diagnoses, drugs, or laboratory tests) and concept sets (specific or unspecified pertussis). Algorithm- and database-specific BorPer incidence rates (IRs) were estimated in children aged 0-14 years enrolled in 2012 and 2014 and followed up until the end of each calendar year and compared with IRs of confirmed pertussis from the ECDC surveillance system (TESSy). Novel formulas were used to approximate validity indices, based on a small set of assumptions. They were applied to approximately estimate positive predictive value (PPV) and sensitivity in SIDIAP.

RESULTS

The number of cases and the estimated BorPer IRs per 100,000 person-years in PC, using data representing 3,173,268 person-years, were 0 (IR = 0.0), 21 (IR = 4.3), 21 (IR = 5.1), 79 (IR = 5.7), and 2 (IR = 2.3) in BIFAP, SIDIAP, THIN, RCGP RSC and PEDIANET respectively. The IRs for combined specific/unspecified pertussis were higher than TESSy, suggesting that some false positives had been included. In SIDIAP the estimated IR was 45.0 when discharge diagnoses were included. The sensitivity and PPV of combined PC specific and unspecific diagnoses for BorPer cases in SIDIAP were approximately 85% and 72%, respectively.

CONCLUSION

Retrieving BorPer cases using only specific concepts has low sensitivity in PC databases, while including cases retrieved by unspecified concepts introduces false positives, which were approximately estimated to be 28% in one database. The share of cases that cannot be retrieved from a PC database because they are only seen in hospital was approximately estimated to be 15% in one database. This study demonstrated that quantifying the impact of different event-finding algorithms across databases and benchmarking with disease surveillance data can provide approximate estimates of algorithm validity.

A Review of Automatic Phenotyping Approaches using Electronic Health Records

Electronic Health Records (EHR) are a rich repository of valuable clinical information that exist in primary and secondary care databases. In order to utilize EHRs for medical observational research a range of algorithms for automatically identifying individuals with a specific phenotype have been developed. This review summarizes and offers a critical evaluation of the literature relating to studies conducted into the development of EHR phenotyping systems. This review describes phenotyping systems and techniques based on structured and unstructured EHR data. Articles published on PubMed and Google scholar between 2013 and 2017 have been reviewed, using search terms derived from Medical Subject Headings (MeSH). The popularity of using Natural Language Processing (NLP) techniques in extracting features from narrative text has increased. This increased attention is due to the availability of open source NLP algorithms, combined with accuracy improvement. In this review, Concept extraction is the most popular NLP technique since it has been used by more than 50% of the reviewed papers to extract features from EHR. High-throughput phenotyping systems using unsupervised machine learning techniques have gained more popularity due to their ability to efficiently and automatically extract a phenotype with minimal human effort.

Sensitivity and specificity of an algorithm based on medico-administrative data to identify hospitalized patients with major bleeding presenting to an emergency department

Background

Validation studies on an ICD-10-based algorithm to identify major bleeding events are scarce, and mostly focused on positive predictive values.

Objective

To evaluate the sensitivity and specificity of an ICD-10-based algorithm in adult patients referred to hospital.

Methods

This was a cross-sectional, retrospective analysis. Among all hospital stays of adult patients referred to Rennes University Hospital, France, through the emergency ward in 2014, we identified major bleeding events according to an index test based on a list of ICD-10 diagnoses. As a reference, a two-step process was applied: firstly, a computerized request for electronic health records from the emergency ward, using several hemorrhage-related diagnostic codes and specific emergency therapies so as to discard stays with a very low probability of bleeding; secondly, a chart review of selected records was conducted by a medical expert blinded to the index test results and each hospital stay was classified into one of two exclusive categories: major bleeding or no major bleeding, according to pre-specified criteria.

Results

Out of 16,012 hospital stays, the reference identified 736 major bleeding events and left 15,276 stays considered as without the target condition. The index test identified 637 bleeding events: 293 intracranial hemorrhages, 197 gastrointestinal hemorrhages and 147 other bleeding events. Overall, sensitivity was 65% (95%CI, 62 to 69), and specificity was 99.0%. We observed differential sensitivity and specificity across bleeding types, with the highest values for intracranial hemorrhage. Positive predictive values ranged from 59% for “other” bleeding events, to 71% (95%CI, 65 to 78) for gastrointestinal hemorrhage, and 96% for intracranial hemorrhage.

Conclusions

Low sensitivity and differential measures of accuracy across bleeding types support the need for specific data collection and medical validation rather than using an ICD-10-based algorithm for assessing the incidence of major bleeding.

TASKA: A modular task management system to support health research studies

Background

Many healthcare databases have been routinely collected over the past decades, to support clinical practice and administrative services. However, their secondary use for research is often hindered by restricted governance rules. Furthermore, health research studies typically involve many participants with complementary roles and responsibilities which require proper process management.

Results

From a wide set of requirements collected from European clinical studies, we developed TASKA, a task/workflow management system that helps to cope with the socio-technical issues arising when dealing with multidisciplinary and multi-setting clinical studies. The system is based on a two-layered architecture: 1) the backend engine, which follows a micro-kernel pattern, for extensibility, and RESTful web services, for decoupling from the web clients; 2) and the client, entirely developed in ReactJS, allowing the construction and management of studies through a graphical interface. TASKA is a GNU GPL open source project, accessible at https://github.com/bioinformatics-ua/taska. A demo version is also available at https://bioinformatics.ua.pt/taska.

Conclusions

The system is currently used to support feasibility studies across several institutions and countries, in the context of the European Medical Information Framework (EMIF) project. The tool was shown to simplify the set-up of health studies, the management of participants and their roles, as well as the overall governance process.

Manual review of electronic medical records as a reference standard for case definition development: a validation study

BACKGROUND

The Canadian Primary Care Sentinel Surveillance Network (CPCSSN) previously carried out a validation study of case definitions for 8 chronic diseases (diabetes mellitus, hypertension, osteoarthritis, depression, dementia, chronic obstructive pulmonary disease, parkinsonism and epilepsy) using direct review of "raw" electronic medical record data. Although effective, this method is time-consuming and can present methodological and organizational challenges. We aimed to determine whether the processed and standardized data contained with the CPCSSN database might function as a reference standard for case definition validation.

METHODS

Using a traditional validation study design, we compared the case identification results of the chart reviews for the 8 chronic diseases with the results of a manual review of the CPCSSN processed data for the same conditions in the same patient sample. Patients were randomly sampled from the June 30, 2012 CPCSSN database, with oversampling of patients with rare conditions.

RESULTS

We analyzed data for 1906 patients. Manual review of the CPCSSN records for case ascertainment yielded sensitivity ranging from 77.5% (95% confidence interval [CI] 73.3%-81.6%) for depression to 97.2% (95% CI 95.4%-99.0%) for diabetes. Specificity was high for all definitions (range 93.1% [95% CI 91.4%-94.7%] to 99.4% [95% CI 99.0%-99.8%]). Positive predictive values and negative predictive values also showed high accuracy of the manual CPCSSN record review relative to review of the raw chart data.

INTERPRETATION

The use of CPCSSN records as the reference standard to validate case definitions substantially reduces the burden on sentinel physicians and clinic managers as well as on researchers while offering a reference standard that is a reasonable substitution for chart review.

Assessment of the coding accuracy of warfarin-related bleeding events

INTRODUCTION

Using International Classification of Diseases, 9th edition (ICD-9) diagnosis codes to identify potential warfarin-related bleeding events from administrative datasets is highly efficient but may be prone to identifying non-events. The objective of this study was to evaluate the ability of bleeding-related ICD-9 codes to identify true bleeding events in patients who were receiving warfarin therapy at the time of hospitalization.

METHODS

This was a cross-sectional study conducted in an integrated healthcare delivery system. Anticoagulated patients aged ≥18years and hospitalized between January 1, 2014 and March 31, 2014 were identified using administrative data queries. All hospitalizations were manually chart reviewed by a trained abstractor blinded to hospitalization diagnoses to assess for true bleeding events. Identification of the presence or lack of bleeding-related ICD-9 diagnosis code(s) for each hospitalization was then performed. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for each ICD-9 code present.

RESULTS

There were 486 hospitalizations in 468 anticoagulated patients with 57 true bleeding events identified. Patients had a mean age of 73.4years and 50% were female. For codes in the principal position, sensitivity, specificity, PPV, and NPV were 7.0%, 99.8%, 80.0%, and 89.0%, respectively. For codes in any position, sensitivity, specificity, PPV, and NPV were 94.7%, 90.9%, 58.1%, and 99.2%, respectively. For major bleeding, sensitivity, specificity, PPV, and NPV were 100%, 83.1%, 14.0%, and 100%, respectively.

CONCLUSIONS

While the absence of a bleeding ICD-9 code reliably ruled-out hospitalization for warfarin-related bleeding, bleeding ICD-9 codes in the principal position were rarely used and undesirable false positive rates were identified when ICD-9 codes when recorded in any position and for major bleeding. Manual chart review is recommended to validate bleeding events from administrative data.

Factors influencing the development of primary care data collection projects from electronic health records: a systematic review of the literature

BACKGROUND

Primary care data gathered from Electronic Health Records are of the utmost interest considering the essential role of general practitioners (GPs) as coordinators of patient care. These data represent the synthesis of the patient history and also give a comprehensive picture of the population health status. Nevertheless, discrepancies between countries exist concerning routine data collection projects. Therefore, we wanted to identify elements that influence the development and durability of such projects.

METHODS

A systematic review was conducted using the PubMed database to identify worldwide current primary care data collection projects. The gray literature was also searched via official project websites and their contact person was emailed to obtain information on the project managers. Data were retrieved from the included studies using a standardized form, screening four aspects: projects features, technological infrastructure, GPs' roles, data collection network organization.

RESULTS

The literature search allowed identifying 36 routine data collection networks, mostly in English-speaking countries: CPRD and THIN in the United Kingdom, the Veterans Health Administration project in the United States, EMRALD and CPCSSN in Canada. These projects had in common the use of technical facilities that range from extraction tools to comprehensive computing platforms. Moreover, GPs initiated the extraction process and benefited from incentives for their participation. Finally, analysis of the literature data highlighted that governmental services, academic institutions, including departments of general practice, and software companies, are pivotal for the promotion and durability of primary care data collection projects.

CONCLUSION

Solid technical facilities and strong academic and governmental support are required for promoting and supporting long-term and wide-range primary care data collection projects.

A rule-based electronic phenotyping algorithm for detecting clinically relevant cardiovascular disease cases

Background

The implementation of electronic medical records (EMR) is becoming increasingly common. Error and data loss reduction, patient-care efficiency increase, decision-making assistance and facilitation of event surveillance, are some of the many processes that EMRs help improve. In addition, they show a lot of promise in terms of data collection to facilitate observational epidemiological studies and their use for this purpose has increased significantly over the recent years. Even though the quantity and availability of the data are clearly improved thanks to EMRs, still, the problem of the quality of the data remains. This is especially important when attempting to determine if an event has actually occurred or not. We sought to assess the sensitivity, specificity, and agreement level of a codes-based algorithm for the detection of clinically relevant cardiovascular (CaVD) and cerebrovascular (CeVD) disease cases, using data from EMRs.

Methods

Three family physicians from the research group selected clinically relevant CaVD and CeVD terms from the international classification of primary care, Second Edition (ICPC-2), the ICD 10 version 2015 and SNOMED-CT 2015 Edition. These terms included both signs, symptoms, diagnoses and procedures associated with CaVD and CeVD. Terms not related to symptoms, signs, diagnoses or procedures of CaVD or CeVD and also those describing incidental findings without clinical relevance were excluded. The algorithm yielded a positive result if the patient had at least one of the selected terms in their medical records, as long as it was not recorded as an error. Else, if no terms were found, the patient was classified as negative. This algorithm was applied to a randomly selected sample of the active patients within the hospital’s HMO by 1/1/2005 that were 40–79 years old, had at least one year of seniority in the HMO and at least one clinical encounter. Thus, patients were classified into four groups: (1) Negative patients (2) Patients with CaVD but without CeVD; (3) Patients with CeVD but without disease CaVD; (4) Patients with both diseases. To facilitate the validation process, a stratified sample was taken so that each of the groups represented approximately 25% of the sample. Manual chart review was used as the gold standard for assessing the algorithm’s performance. One-third of the patients were assigned randomly to each reviewer (Cohen’s kappa 0.91). Both coded and un-coded (free text) sections of the EMR were reviewed. This was done from the first present clinical note in the patients chart to the last one registered prior to 1/1/2005.

Results

The performance of the algorithm was compared against manual chart review. It yielded high sensitivity (0.99, 95% CI 0.938–0.9971) and acceptable specificity (0.86, 95% CI 0.818–0.895) for detecting cases of CaVD and CeVD combined. A qualitative analysis of the false positives and false negatives was performed.

Conclusions

We developed a simple algorithm, using only standardized and non-standardized coded terms within an EMR that can properly detect clinically relevant events and symptoms of CaVD and CeVD. We believe that combining it with an analysis of the free text using an NLP approach would yield even better results.

Electronic supplementary material

The online version of this article (doi:10.1186/s13104-017-2600-2) contains supplementary material, which is available to authorized users.

Readmission after Gastrointestinal Bleeding in Children: A Retrospective Cohort Study

INTRODUCTION

To compare the demographic, clinical, and therapeutic characteristics in a cohort of patients discharged following acute gastrointestinal bleeding, representing to the emergency department (ED) and readmitted within 30 days of discharge with the characteristics of non-readmitted patients.

STUDY DESIGN

Hospitalization data was obtained from the Pediatric Hospital Information System including 49 tertiary children's hospitals in the US. Children 1-21 years of age diagnosed with acute gastrointestinal bleeding, admitted between January 2007 and September 2015 were included. The primary outcomes in this study were 30-day inpatient readmission through the ED and 30-day return to the ED only. Unadjusted, univariate followed by multivariable analysis of the associations between patient characteristics and treatment course at the index encounter using the R statistical package, v. 3.2.3.

RESULTS

During the study period, 9902 patients were admitted with acute gastrointestinal bleeding; in the following month, 1460 (16.1%) represented to the ED and 932 (9%) were readmitted; 68.7% within 14 days from discharge. Readmission was most frequently associated with portal hypertension or esophageal variceal hemorrhage. There was a decreased likelihood of readmission with endoscopy (OR 0.77, 95% CI, 0.661, 0.906) and with Meckel scan (OR 0.513, 95% CI 0.362, 0.727) during the initial admission. Multiple comorbidities, longer initial stay and the early proton pump inhibitor therapy were associated with higher likelihood of readmission.

DISCUSSION

Readmission following acute gastrointestinal bleeding is common and is more likely following variceal hemorrhage, long initial admission, and chronic comorbidities.

Mortality associated with gastrointestinal bleeding in children: A retrospective cohort study

AIM

To determine the clinical characteristics of children with gastrointestinal bleeding (GIB) who died during the course of their admission.

METHODS

We interrogated the Pediatric Hospital Information System database, including International Classification of Diseases, Current Procedural Terminology and Clinical Transaction Classification coding from 47 pediatric tertiary centers extracting the population of patients (1-21 years of age) admitted (inpatient or observation) with acute, upper or indeterminate GIB (1/2007-9/2015). Descriptive statistics, unadjusted univariate and adjusted multivariate analysis of the associations between patient characteristics and treatment course with mortality was performed with mortality as primary and endoscopy a secondary outcome of interest. All analyses were performed using the R statistical package, v.3.2.3.

RESULTS

The population with GIB was 19528; 54.6% were male, overall mortality was 2.07%; (0.37% in patients with the principal diagnosis of GIB). When considering only the mortalities in which GIB was the principal diagnosis, 48% (12 of 25 principal diagnosis GIB mortalities) died within the first 3 d of admission, whereas 19.8% of secondary diagnosis GIB patients died with 3 d of admission. Patients who died were more likely to have received octreotide (19.8% c.f. 4.04%) but tended to have not received proton pump inhibitor therapy in the first 48 h, and far less likely to have undergone endoscopy during their admission (OR = 0.489, P < 0.0001). Chronic liver disease associated with a greater likelihood of endoscopy. Mortalities were significantly more likely to have multiple complex chronic conditions.

CONCLUSION

GIB associated mortality in children is highest within 7 d of admission. Multiple comorbidities are a risk factor whereas early endoscopy during the admission is protective.

Automatic identification of type 2 diabetes, hypertension, ischaemic heart disease, heart failure and their levels of severity from Italian General Practitioners' electronic medical records: a validation study

OBJECTIVES

The Italian project MATRICE aimed to assess how well cases of type 2 diabetes (T2DM), hypertension, ischaemic heart disease (IHD) and heart failure (HF) and their levels of severity can be automatically extracted from the Health Search/CSD Longitudinal Patient Database (HSD). From the medical records of the general practitioners (GP) who volunteered to participate, cases were extracted by algorithms based on diagnosis codes, keywords, drug prescriptions and results of diagnostic tests. A random sample of identified cases was validated by interviewing their GPs.

SETTING

HSD is a database of primary care medical records. A panel of 12 GPs participated in this validation study.

PARTICIPANTS

300 patients were sampled for each disease, except for HF, where 243 patients were assessed.

OUTCOME MEASURES

The positive predictive value (PPV) was assessed for the presence/absence of each condition against the GP's response to the questionnaire, and Cohen's κ was calculated for agreement on the severity level.

RESULTS

The PPV was 100% (99% to 100%) for T2DM and hypertension, 98% (96% to 100%) for IHD and 55% (49% to 61%) for HF. Cohen's kappa for agreement on the severity level was 0.70 for T2DM and 0.69 for hypertension and IHD.

CONCLUSIONS

This study shows that individuals with T2DM, hypertension or IHD can be validly identified in HSD by automated identification algorithms. Automatic queries for levels of severity of the same diseases compare well with the corresponding clinical definitions, but some misclassification occurs. For HF, further research is needed to refine the current algorithm.

The Effect of Safety Warnings on Antipsychotic Drug Prescribing in Elderly Persons with Dementia in the United Kingdom and Italy: A Population-Based Study

BACKGROUND

Antipsychotic (AP) drugs are commonly used to manage the behavioural symptoms of dementia. Nevertheless, international (i.e. the European Medicines Agency in Europe) and national (i.e. the Medicines and Healthcare products Regulatory Agency in the UK and the Italian Drug Agency) regulatory agencies issued safety warnings against AP use in dementia in 2004 and 2009.

OBJECTIVE

The aim of this study is to investigate the short- and long-term impact of safety warnings on the use of APs in UK and Italian persons with dementia using two nationwide databases: The Health Improvement Network (THIN) from the UK and the Health Search Database-Cegedim-Strategic Data-Longitudinal Patient Database (HSD-CSD-LPD) from Italy.

METHODS

We calculated the overall quarterly prevalence of AP use by class and by individual drug in persons with dementia aged ≥65 years and used generalized linear models to explore the effect of the safety warnings.

RESULTS

We identified 58,497 and 10,857 individuals aged ≥65 years with dementia from the THIN and HSD-CSD-LPD databases, respectively, over the period 2000-2012. After the 2004 warnings, the use of atypical APs decreased, whereas the use of conventional APs increased, in Italy and the UK until 2009. However, the trend for APs individually showed that the use of risperidone/olanzapine decreased, whereas the use of quetiapine increased in both countries. After the 2009 warnings (until 2012), the use of atypical and conventional APs decreased in the UK (from 11 to 9 and 5 to 3 %, respectively), but such use increased in Italy (from 11 to 18 and 9 to 14 %, respectively).

CONCLUSION

The 2004 warnings led to a reduction in the use of olanzapine and risperidone and increased the use of quetiapine/conventional APs in both countries. From 2009, the use of APs decreased in persons with dementia in the UK but not in Italy. Possible reasons for the difference in AP use between the two countries include a more proactive approach towards reducing the use of APs in the UK than in Italy.

Safety of Direct Oral Anticoagulants: Insights from Postmarketing Studies

Direct oral anticoagulants (DOACs) have been marketed in the United States since 2010. While numerous large-scale prospective phase 3 outcomes studies have documented the effectiveness of DOACs for the prevention of stroke and systemic embolism in patients with nonvalvular atrial fibrillation, the primary safety concern with all of these drugs-as it is with the more established oral anticoagulant warfarin-is the risk of major bleeding. Postmarketing surveillance studies (PMSS) provide the opportunity to evaluate the safety of these recently approved drugs across a spectrum of patients that may be broader than those included in randomized controlled trials. This review will summarize the safety findings of numerous recently performed, large-scale PMSS evaluations, and consider the currently available evidence regarding the risks for bleeding in patients treated with DOACs, in order to give providers and patients additional evidence regarding the safety of DOACs.

Safety of direct oral anticoagulants: insights from postmarketing studies

Direct oral anticoagulants (DOACs) have been marketed in the United States since 2010. While numerous large-scale prospective phase 3 outcomes studies have documented the effectiveness of DOACs for the prevention of stroke and systemic embolism in patients with nonvalvular atrial fibrillation, the primary safety concern with all of these drugs-as it is with the more established oral anticoagulant warfarin-is the risk of major bleeding. Postmarketing surveillance studies (PMSS) provide the opportunity to evaluate the safety of these recently approved drugs across a spectrum of patients that may be broader than those included in randomized controlled trials. This review will summarize the safety findings of numerous recently performed, large-scale PMSS evaluations, and consider the currently available evidence regarding the risks for bleeding in patients treated with DOACs, in order to give providers and patients additional evidence regarding the safety of DOACs.

Useful Interplay Between Spontaneous ADR Reports and Electronic Healthcare Records in Signal Detection

Background and Objective

Spontaneous reporting systems (SRSs) remain the cornerstone of post-marketing drug safety surveillance despite their well-known limitations. Judicious use of other available data sources is essential to enable better detection, strengthening and validation of signals. In this study, we investigated the potential of electronic healthcare records (EHRs) to be used alongside an SRS as an independent system, with the aim of improving signal detection.

Methods

A signal detection strategy, focused on a limited set of adverse events deemed important in pharmacovigilance, was performed retrospectively in two data sources—(1) the Exploring and Understanding Adverse Drug Reactions (EU-ADR) database network and (2) the EudraVigilance database—using data between 2000 and 2010. Five events were considered for analysis: (1) acute myocardial infarction (AMI); (2) bullous eruption; (3) hip fracture; (4) acute pancreatitis; and (5) upper gastrointestinal bleeding (UGIB). Potential signals identified in each system were verified using the current published literature. The complementarity of the two systems to detect signals was expressed as the percentage of the unilaterally identified signals out of the total number of confirmed signals. As a proxy for the associated costs, the number of signals that needed to be reviewed to detect one true signal (number needed to detect [NND]) was calculated. The relationship between the background frequency of the events and the capability of each system to detect signals was also investigated.

Results

The contribution of each system to signal detection appeared to be correlated with the background incidence of the events, being directly proportional to the incidence in EU-ADR and inversely proportional in EudraVigilance. EudraVigilance was particularly valuable in identifying bullous eruption and acute pancreatitis (71 and 42 % of signals were correctly identified from the total pool of known associations, respectively), while EU-ADR was most useful in identifying hip fractures (60 %). Both systems contributed reasonably well to identification of signals related to UGIB (45 % in EudraVigilance, 40 % in EU-ADR) but only fairly for signals related to AMI (25 % in EU-ADR, 20 % in EudraVigilance). The costs associated with detection of signals were variable across events; however, it was often more costly to detect safety signals in EU-ADR than in EudraVigilance (median NNDs: 7 versus 5).

Conclusion

An EHR-based system may have additional value for signal detection, alongside already established systems, especially in the presence of adverse events with a high background incidence. While the SRS appeared to be more cost effective overall, for some events the costs associated with signal detection in the EHR might be justifiable.

Electronic supplementary material

The online version of this article (doi:10.1007/s40264-015-0341-5) contains supplementary material, which is available to authorized users.

Extracting information from the text of electronic medical records to improve case detection: a systematic review

BACKGROUND

Electronic medical records (EMRs) are revolutionizing health-related research. One key issue for study quality is the accurate identification of patients with the condition of interest. Information in EMRs can be entered as structured codes or unstructured free text. The majority of research studies have used only coded parts of EMRs for case-detection, which may bias findings, miss cases, and reduce study quality. This review examines whether incorporating information from text into case-detection algorithms can improve research quality.

METHODS

A systematic search returned 9659 papers, 67 of which reported on the extraction of information from free text of EMRs with the stated purpose of detecting cases of a named clinical condition. Methods for extracting information from text and the technical accuracy of case-detection algorithms were reviewed.

RESULTS

Studies mainly used US hospital-based EMRs, and extracted information from text for 41 conditions using keyword searches, rule-based algorithms, and machine learning methods. There was no clear difference in case-detection algorithm accuracy between rule-based and machine learning methods of extraction. Inclusion of information from text resulted in a significant improvement in algorithm sensitivity and area under the receiver operating characteristic in comparison to codes alone (median sensitivity 78% (codes + text) vs 62% (codes), P = .03; median area under the receiver operating characteristic 95% (codes + text) vs 88% (codes), P = .025).

CONCLUSIONS

Text in EMRs is accessible, especially with open source information extraction algorithms, and significantly improves case detection when combined with codes. More harmonization of reporting within EMR studies is needed, particularly standardized reporting of algorithm accuracy metrics like positive predictive value (precision) and sensitivity (recall).

Identifying Cases of Type 2 Diabetes in Heterogeneous Data Sources: Strategy from the EMIF Project

Due to the heterogeneity of existing European sources of observational healthcare data, data source-tailored choices are needed to execute multi-data source, multi-national epidemiological studies. This makes transparent documentation paramount. In this proof-of-concept study, a novel standard data derivation procedure was tested in a set of heterogeneous data sources. Identification of subjects with type 2 diabetes (T2DM) was the test case. We included three primary care data sources (PCDs), three record linkage of administrative and/or registry data sources (RLDs), one hospital and one biobank. Overall, data from 12 million subjects from six European countries were extracted. Based on a shared event definition, sixteeen standard algorithms (components) useful to identify T2DM cases were generated through a top-down/bottom-up iterative approach. Each component was based on one single data domain among diagnoses, drugs, diagnostic test utilization and laboratory results. Diagnoses-based components were subclassified considering the healthcare setting (primary, secondary, inpatient care). The Unified Medical Language System was used for semantic harmonization within data domains. Individual components were extracted and proportion of population identified was compared across data sources. Drug-based components performed similarly in RLDs and PCDs, unlike diagnoses-based components. Using components as building blocks, logical combinations with AND, OR, AND NOT were tested and local experts recommended their preferred data source-tailored combination. The population identified per data sources by resulting algorithms varied from 3.5% to 15.7%, however, age-specific results were fairly comparable. The impact of individual components was assessed: diagnoses-based components identified the majority of cases in PCDs (93–100%), while drug-based components were the main contributors in RLDs (81–100%). The proposed data derivation procedure allowed the generation of data source-tailored case-finding algorithms in a standardized fashion, facilitated transparent documentation of the process and benchmarking of data sources, and provided bases for interpretation of possible inter-data source inconsistency of findings in future studies.

Data Extraction and Management in Networks of Observational Health Care Databases for Scientific Research: A Comparison of EU-ADR, OMOP, Mini-Sentinel and MATRICE Strategies

Introduction:

We see increased use of existing observational data in order to achieve fast and transparent production of empirical evidence in health care research. Multiple databases are often used to increase power, to assess rare exposures or outcomes, or to study diverse populations. For privacy and sociological reasons, original data on individual subjects can’t be shared, requiring a distributed network approach where data processing is performed prior to data sharing.

Case Descriptions and Variation Among Sites:

We created a conceptual framework distinguishing three steps in local data processing: (1) data reorganization into a data structure common across the network; (2) derivation of study variables not present in original data; and (3) application of study design to transform longitudinal data into aggregated data sets for statistical analysis. We applied this framework to four case studies to identify similarities and differences in the United States and Europe: Exploring and Understanding Adverse Drug Reactions by Integrative Mining of Clinical Records and Biomedical Knowledge (EU-ADR), Observational Medical Outcomes Partnership (OMOP), the Food and Drug Administration’s (FDA’s) Mini-Sentinel, and the Italian network—the Integration of Content Management Information on the Territory of Patients with Complex Diseases or with Chronic Conditions (MATRICE).

Findings:

National networks (OMOP, Mini-Sentinel, MATRICE) all adopted shared procedures for local data reorganization. The multinational EU-ADR network needed locally defined procedures to reorganize its heterogeneous data into a common structure. Derivation of new data elements was centrally defined in all networks but the procedure was not shared in EU-ADR. Application of study design was a common and shared procedure in all the case studies. Computer procedures were embodied in different programming languages, including SAS, R, SQL, Java, and C++.

Conclusion:

Using our conceptual framework we found several areas that would benefit from research to identify optimal standards for production of empirical knowledge from existing databases.an opportunity to advance evidence-based care management. In addition, formalized CM outcomes assessment methodologies will enable us to compare CM effectiveness across health delivery settings.

Increased risk of microscopic colitis with use of proton pump inhibitors and non-steroidal anti-inflammatory drugs

OBJECTIVES

Microscopic colitis (MC) is characterized by chronic watery diarrhea. Recently, several drugs were reported to increase the risk of MC. However, studies lacked a clear exposure definition, did not address duration relationships, and did not take important biases into account. We estimated the risk of MC during drug use.

METHODS

This is a population-based nested case-control study using a Dutch primary care database (1999-2013). Incident MC cases (aged ≥18 years) were matched to community-based and colonoscopy-negative controls on age, sex, and primary care practice. Drug use was assessed within 1 and 2 years before the index date. Adjusted odds ratios (OR) were calculated by conditional logistic regression.

RESULTS

From the source population of 1,458,410 subjects, 218 cases were matched to 15,045 community controls and 475 colonoscopy-negative controls. Current use (≤3 months) of proton pump inhibitors (PPIs), nonsteroidal anti-inflammatory drugs (NSAIDs), selective serotonin reuptake inhibitors, low-dose aspirin, angiotensin-converting enzyme (ACE) inhibitors and beta-blockers significantly increased the risk of MC compared with never use in community controls. Adjusted ORs ranged from 2.5 (95% confidence interval (CI): 1.5-4.2) for ACE inhibitors to 7.3 (95% CI: 4.5-12.1) for PPIs in the year prior to the index date. After accounting for diagnostic delay, only use of NSAIDs, PPIs, low-dose aspirin, and ACE inhibitors increased the risk of MC. Compared with colonoscopy controls, only use of PPIs (OR-adjusted 10.6; 1.8-64.2) and NSAIDs (OR-adjusted 5.6; 1.2-27.0) increased the risk of MC.

CONCLUSIONS

NSAIDs and PPIs are associated with an increased risk of MC. The association of MC with use of the other drugs is probably explained by worsening of diarrhea/symptoms rather than increasing the risk of MC itself.

Risk of upper gastrointestinal bleeding from different drug combinations

BACKGROUND & AIMS

Concomitant use of nonsteroidal anti-inflammatory drugs (NSAIDs) and low-dose aspirin increases the risk of upper gastrointestinal bleeding (UGIB). Guidelines suggest avoiding certain drug combinations, yet little is known about the magnitude of their interactions. We estimated the risk of UGIB during concomitant use of nonselective (ns)NSAIDs, cyclooxygenase -2 selective inhibitors (COX-2 inhibitors), and low-dose aspirin with other drugs.

METHODS

We performed a case series analysis of data from 114,835 patients with UGIB (930,888 person-years of follow-up) identified from 7 population-based health care databases (approximately 20 million subjects). Each patient served as his or her own control. Drug exposure was determined based on prescriptions of nsNSAIDs, COX-2 inhibitors, or low-dose aspirin, alone and in combination with other drugs that affect the risk of UGIB. We measured relative risk (incidence rate ratio [IRR] during drug exposure vs nonexposure) and excess risk due to concomitant drug exposure (relative excess risk due to interaction [RERI]).

RESULTS

Monotherapy with nsNSAIDs increased the risk of diagnosis of UGIB (IRR, 4.3) to a greater extent than monotherapy with COX-2 inhibitors (IRR, 2.9) or low-dose aspirin (IRR, 3.1). Combination therapy generally increased the risk of UGIB; concomitant nsNSAID and corticosteroid therapies increased the IRR to the greatest extent (12.8) and also produced the greatest excess risk (RERI, 5.5). Concomitant use of nsNSAIDs and aldosterone antagonists produced an IRR for UGIB of 11.0 (RERI, 4.5). Excess risk from concomitant use of nsNSAIDs with selective serotonin reuptake inhibitors (SSRIs) was 1.6, whereas that from use of COX-2 inhibitors with SSRIs was 1.9 and that for use of low-dose aspirin with SSRIs was 0.5. Excess risk of concomitant use of nsNSAIDs with anticoagulants was 2.4, of COX-2 inhibitors with anticoagulants was 0.1, and of low-dose aspirin with anticoagulants was 1.9.

CONCLUSIONS

Based on a case series analysis, concomitant use of nsNSAIDs, COX-2 inhibitors, or low-dose aspirin with SSRIs significantly increases the risk of UGIB. Concomitant use of nsNSAIDs or low-dose aspirin, but not COX-2 inhibitors, with corticosteroids, aldosterone antagonists, or anticoagulants produces significant excess risk of UGIB.