The validity of the diagnosis of inflammatory arthritis in a large population-based primary care database

Electronic Medical Record-Based Case Phenotyping for the Charlson Conditions: Scoping Review

Background

Electronic medical records (EMRs) contain large amounts of rich clinical information. Developing EMR-based case definitions, also known as EMR phenotyping, is an active area of research that has implications for epidemiology, clinical care, and health services research.

Objective

This review aims to describe and assess the present landscape of EMR-based case phenotyping for the Charlson conditions.

Methods

A scoping review of EMR-based algorithms for defining the Charlson comorbidity index conditions was completed. This study covered articles published between January 2000 and April 2020, both inclusive. Embase (Excerpta Medica database) and MEDLINE (Medical Literature Analysis and Retrieval System Online) were searched using keywords developed in the following 3 domains: terms related to EMR, terms related to case finding, and disease-specific terms. The manuscript follows the Preferred Reporting Items for Systematic reviews and Meta-analyses extension for Scoping Reviews (PRISMA) guidelines.

Results

A total of 274 articles representing 299 algorithms were assessed and summarized. Most studies were undertaken in the United States (181/299, 60.5%), followed by the United Kingdom (42/299, 14.0%) and Canada (15/299, 5.0%). These algorithms were mostly developed either in primary care (103/299, 34.4%) or inpatient (168/299, 56.2%) settings. Diabetes, congestive heart failure, myocardial infarction, and rheumatology had the highest number of developed algorithms. Data-driven and clinical rule–based approaches have been identified. EMR-based phenotype and algorithm development reflect the data access allowed by respective health systems, and algorithms vary in their performance.

Conclusions

Recognizing similarities and differences in health systems, data collection strategies, extraction, data release protocols, and existing clinical pathways is critical to algorithm development strategies. Several strategies to assist with phenotype-based case definitions have been proposed.

Increased primary care use for musculoskeletal symptoms, infections and comorbidities in the years before the diagnosis of inflammatory arthritis

Objectives

Little is known about relevant events in the at-risk phase of rheumatoid arthritis before the development of clinically apparent inflammatory arthritis (IA). The present study assessed musculoskeletal symptoms, infections and comorbidity in future IA patients.

Methods

In a nested case–control study using electronic health records of general practitioners, the frequency and timing of 192 symptoms or diseases were evaluated before a diagnosis of IA, using the International Classification of Primary Care coding system. Cases were 2314 adults with a new diagnosis IA between 2012 and 2016; controls were matched 1:2. The frequency of primary care visits was compared using logistic regression.

Results

The frequency of visits for musculoskeletal symptoms (mostly of shoulders, wrists, fingers and knees) and carpal tunnel syndrome was significantly higher in IA patients vs controls within the final 1.5 years before diagnosis, with ORs of 3.2 (95% CI 2.8 to 3.5), 2.8 (95% CI 2.5 to 3.1) and 2.5 (95% CI 2.2 to 2.8) at 6, 12 and 18 months before diagnosis, respectively. Also, infections (notably of the genital and urinary tracts), IA-comorbidities and chronic diseases were more prevalent in cases than controls, but more evenly spread out over the whole 6-year period before IA. A decision tree was created including all symptoms and diseases.

Conclusion

There was an increased frequency of primary care visits for musculoskeletal symptoms, infections and comorbidities prior to the diagnosis of IA. This diverging trend is present for 4–6 years, but becomes statistically significant 1.5 years before the diagnosis. Validation of these results is warranted.

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).

Automatic infection detection based on electronic medical records

BACKGROUND

Making accurate patient care decision, as early as possible, is a constant challenge, especially for physicians in the emergency department. The increasing volumes of electronic medical records (EMRs) open new horizons for automatic diagnosis. In this paper, we propose to use machine learning approaches for automatic infection detection based on EMRs. Five categories of information are utilized for prediction, including personal information, admission note, vital signs, diagnose test results and medical image diagnose.

RESULTS

Experimental results on a newly constructed EMRs dataset from emergency department show that machine learning models can achieve a decent performance for infection detection with area under the receiver operator characteristic curve (AUC) of 0.88. Out of all the five types of information, admission note in text form makes the most contribution with the AUC of 0.87.

CONCLUSIONS

This study provides a state-of-the-art EMRs processing system to automatically make medical decisions. It extracts five types of features associated with infection and achieves a decent performance on automatic infection detection based on machine learning models.

Impact of a Restriction in Reimbursement on Proton Pump Inhibitors in Patients with an Increased Risk of Gastric Complications

Governments have several options to reduce the increasing costs of health care, including restrictions for the reimbursement of medicines. Next to the intended effect of reduced costs for medicines, reimbursement restriction can have unintended effects such as patients refraining from their treatment which may lead to health problems and increased use of health care. An example of a reimbursement restriction is the one for proton pump inhibitors (PPIs) that became effective in the Netherlands in January 2012. A major unintended effect of this measure could be that high-risk patients who start with non-steroidal anti-inflammatory drugs (NSAIDs) or low-dose aspirin (aspirin) and who have an increased risk of gastric complications for which they are prescribed PPIs refrain from this PPI treatment. The aim of this study was to evaluate the effect of the reimbursement restriction among high-risk users of NSAIDs or aspirin. Do these patients refrain from their PPI treatment and if so do they have an increased risk of gastric complications? Part of the patients starting with NSAIDs or aspirin have an increased risk of gastric complications due to their age, comorbidities, or co-medication. The incidence of PPI use during the 2 years before the reimbursement restriction (2010 and 2011) and 2 years after the introduction of the reimbursement restriction was compared for patients on NSAIDs or aspirin with an increased risk of developing gastric complications. Impact of age, sex, and social economic status (SES) was taken into account. Hospital admissions due to gastric complications were studied over the same period (2010–2013). Data were obtained from a large population-based primary care database and a hospital database. The use of PPIs in patients with an increased risk of gastric complications who started NSAID/aspirin increased from 40% in 2010 to 55% in 2013. No impact was found of age, sex, or SES. There was no increase in hospital admissions due to gastric complications after the reimbursement restriction. The reimbursement restriction on PPIs was not associated with any detectable unintended effects for patients with an increased risk of gastric complications.

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.

Do GPs know their patients with cancer? Assessing the quality of cancer registration in Dutch primary care: a cross-sectional validation study

OBJECTIVES

To assess the quality of cancer registry in primary care.

DESIGN AND SETTING

A cross-sectional validation study using linked data from primary care electronic health records (EHRs) and the Netherlands Cancer Registry (NCR).

POPULATION

290 000 patients, registered with 120 general practitioners (GPs), from 50 practice centres in the Utrecht area, the Netherlands, in January 2013.

INTERVENTION

Linking the EHRs of all patients in the Julius General Practitioners' Network database at an individual patient level to the full NCR (∼1.7 million tumours between 1989 and 2011), to determine the proportion of matching cancer diagnoses. Full-text EHR extraction and manual analysis for non-matching diagnoses.

MAIN OUTCOME MEASURES

Proportions of matching and non-matching breast, lung, colorectal and prostate cancer diagnoses between 2007 and 2011, stratified by age category, cancer type and EHR system. Differences in year of diagnosis between the EHR and the NCR. Reasons for non-matching diagnoses.

RESULTS

In the Primary Care EHR, 60.6% of cancer cases were registered and coded in accordance with the NCR. Of the EHR diagnoses, 48.9% were potentially false positive (not registered in the NCR). Results differed between EHR systems but not between age categories or cancer types. The year of diagnosis corresponded in 80.6% of matching coded diagnoses. Adding full-text EHR analysis improved results substantially. A national disease registry (the NCR) proved incomplete.

CONCLUSIONS

Even though GPs do know their patients with cancer, only 60.6% are coded in concordance with the NCR. Reusers of coded EHR data should be aware that 40% of cases can be missed, and almost half can be false positive. The type of EHR system influences registration quality. If full-text manual EHR analysis is used, only 10% of cases will be missed and 20% of cases found will be wrong. EHR data should only be reused with care.

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).

Calcium channel blockers and cancer: a risk analysis using the UK Clinical Practice Research Datalink (CPRD)

OBJECTIVE

The evidence of an association between calcium channel blockers (CCBs) and cancer is conflicting. The objective of the present study was to evaluate the risk of cancer (all, breast, prostate and colon cancers) in association with exposure to CCB.

METHODS

This is a population-based cohort study in patients exposed to CCBs from across the UK, using two comparison cohorts: (1) patients with no exposure to CCB (non-CCB) matched on age and gender and (2) unmatched patients unexposed to CCB and at least one other antihypertensive (AHT) prescription. Cancer incidence rates computed in the exposed and the two unexposed groups were compared using HRs and 95% CIs obtained from multivariate Cox regression analyses.

RESULTS

Overall, 150,750, 557,931 and 156,966 patients were included, respectively, in the CCB, non-CCB and AHT cohorts. Crude cancer incidence rates per 1000 person-years were 16.51, 15.75 and 10.62 for the three cohorts, respectively. Adjusted HRs (CI) for all cancers comparing CCB, non-CCB and AHT cohorts were 0.88 (0.86 to 0.89) and 1.01 (0.98 to 1.04), respectively. Compared to the AHT cohort, adjusted HRs (CI) for breast, prostate and colon cancer for the CCB cohort were 0.95 (0.87 to 1.04), 1.07 (0.98 to 1.16) and 0.89 (0.81 to 0.98), respectively. Analyses by duration of exposure to CCB did not show excess risk.

CONCLUSIONS

This large population-based study provides strong evidence that CCB use is not associated with an increased risk of cancer. The analyses yielded robust results across all types of cancer and different durations of exposure to CCBs.

Cardiovascular disease-related hospital admissions of patients with inflammatory arthritis

OBJECTIVE

Patients with inflammatory arthritis (IA) have an increased risk of cardiovascular diseases (CVD), suggesting a high rate of CVD-related hospitalizations, but data on this topic are limited. Our study addressed hospital admissions for CVD in a primary care-based population of patients with IA and controls.

METHODS

All newly diagnosed patients with IA between 2001 and 2010 were selected from electronic medical records of the Netherlands Institute for Health Services Research Primary Care database, representing a national network of general practices. Two control patients matched for age, sex, and practice were selected for each patient with IA. Hospital admission data for all patients was retrieved from the Dutch Hospital Data.

RESULTS

There were 2615 patients with IA and 5555 controls included in our study. CVD-related hospital admissions were observed more frequently among patients with IA as compared with control patients: 48% versus 36% (p < 0.001) in a followup period of 4 years. Patients with IA were more often hospitalized because of ischemic heart disease (OR 1.7, 95% CI 1.2-2.2) and for day-care admission because of cerebrovascular disease (OR 2.2, 95% CI 1.0-4.9).

CONCLUSION

Increased hospital admission rates confirm the higher CVD burden among patients with IA compared with controls, and underscore the need for proper CVD risk management in patients with IA.

Increased risk for chronic comorbid disorders in patients with inflammatory arthritis: a population based study

Background

Studies determining the development of a wide variety of different comorbid disorders in inflammatory arthritis (IA) patients are scarce, however, this knowledge could be helpful in optimising preventive care in IA patients. The aim of this study is to establish the risk that new chronic comorbid disorders in newly diagnosed patients with IA in a primary care setting are developed.

Methods

This is a nested-case–control study from 2001–2010 using data from electronic medical patient records in general practice. In total, 3,354 patients with newly diagnosed IA were selected. Each patient was matched with two control patients of the same age and sex in the same general practice. The development of 121 chronic comorbid disorders of index and control patients was compared using Cox regression.

Results

After a median follow-up period of 2.8 years, 56% of the IA-patients had developed at least one chronic comorbid disorder after the onset of IA, compared to 46% of the control patients (p < 0.05). The most frequent developed comorbid disorders after the onset of IA were of cardiovascular (23%), and musculoskeletal (17%) origin. The highest hazard ratios (HRs) were found for anaemia (HR 2.0 [95% CI: 1.4-2.7]) osteoporosis (HR 1.9 [1.4-2.4]), and COPD (HR 1.8 [1.4-2.3]).

Conclusion

Patients with IA developed more chronic comorbid disorders after the onset of IA than one might expect based on age and sex. Since comorbidity has a large impact on the disease course, quality of life, and possibly on treatment itself, prevention of comorbidity should be one of the main targets in the treatment of IA patients.

Prevalence of chronic diseases at the onset of inflammatory arthritis: a population-based study

OBJECTIVE

Little is known about the presence of chronic morbidity in inflammatory arthritis (IA) patients at disease onset. Previous studies have been mainly performed in established IA patients or they focus on isolated co-morbid diseases. Our aim was to determine the prevalence of chronic diseases at the onset of IA and to determine whether this is different from the number that one might expect based on age and sex. Patients and methods. A nested case-control study from 2001 to 2010 using data from patient electronic medical records in general practice. Totally, 3354 patients with newly diagnosed IA were included. Each patient was matched on age, sex and general practice with two control patients. In total, 121 different chronic diseases were studied.

RESULTS

In total, 70% of the IA patients had at least one chronic disease at the onset of IA, compared with 59% of the control patients (P < 0.001). The highest prevalence in IA patients was found for cardiovascular diseases (35%), musculoskeletal diseases (27%) and neurological diseases (22%). Compared with the control patients, patients with IA had the highest increased risk for musculoskeletal diseases [odds ratio, OR = 1.7 (95% confidence interval: 1.6-19)] and for neurological diseases [OR = 1.6 (1.4-1.7)] at the onset of IA.

CONCLUSION

At the onset of IA, nearly three-quarters of patients with IA had at least one other chronic disease. Since multi-morbidity affects treatment and outcome of the IA patient, these diseases should be taken into account when treating IA patients.