Positive predictive value of computerized medical records for uncomplicated and complicated upper gastrointestinal ulcer

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).

Exploring novel diabetes surveillance methods: a comparison of administrative, laboratory and pharmacy data case definitions using THIN

Background

The objective of this study was to identify patients with diabetes in a comprehensive primary care electronic medical records database using a number of different case definitions (clinical, pharmacy, laboratory definitions and a combination thereof) and understand the differences in patient populations being captured by each definition.

Methods

Data for this population-based retrospective cohort study was obtained from The Health Information Network (THIN). THIN is a longitudinal, primary care medical records database of over 9 million patients in UK. Primary outcome was a diagnosis of diabetes, defined by the presence of a diabetes read code, or an abnormal laboratory result, or a prescription for an Oral Anti-diabetic drug or insulin. A 2-year washout period was applied prior to the index of diabetes to avoid inclusion of prevalent cases for each case definition.

Results

This study demonstrated that different case definitions of diabetes identify different sub-populations of patients. When the cohorts were observed based on any measure of central tendency, each of the cohorts were reasonably comparable to each other. However, the distribution of each of the cohorts when grouped by age categories and sex, reveal differences. For example, using pharmacy case definition results in a bimodal distribution among women, one between 1-19 year and 35-39 age categories, and then again between 60-64 and 85 years-however, the histogram becomes more normally distributed when metformin was removed from the case definition.

Conclusion

Our results suggest that clinical, pharmacy, laboratory case definitions identify different sub-populations and using multiple case definitions is likely required to optimally identify the entire diabetes population within THIN. Our study also suggests that age and sex of patients may affect the indexing of diabetes in THIN and is critical to better understand these variations.

Risk of bleeding events among patients with systemic sclerosis and the general population in the UK: a large population-based cohort study

INTRODUCTION/OBJECTIVES

To compare the risk of different bleeding outcomes between patients with systemic sclerosis (SSc) and the general population free of SSc.

METHODS

Using UK electronic primary care data (2000-2012), 1314 patients with SSc and a matched SSc-free comparison cohort (n = 19,992) were followed until December 2013 to identify bleeding, confirmed following manual review of patient records including free text comments. Incidence rates were calculated and Cox regression used to estimate adjusted hazard ratios (HRs; SSc cohort vs. matched general population cohort) adjusted for confounders.

RESULTS

One hundred and twenty-seven bleeding events occurred in the SSc cohort and 1762 in the general population cohort; incidence rates per 1000 person-years for the SSc cohort and general population cohort were 0.5 versus 0.3 for hemorrhagic stroke, 4.1 versus 3.3 for gastrointestinal bleeding, 2.5 versus 1.7 for pulmonary hemorrhage, 8.4 versus 7.5 for urogenital bleeding, and 15.5 versus 12.9 for any of the aforementioned bleedings. Adjusted HRs (95% confidence intervals) were 1.21 (1.00-1.46) for any bleeding, 1.51 (0.54-4.21) for hemorrhagic stroke, 1.50 (0.96-2.35) for pulmonary hemorrhage, 1.08 (0.75-1.54) for gastrointestinal bleeds, and 1.28 (1.00-1.64) for urogenital bleeds. HRs were more often higher in SSc patients with organ involvement than without organ involvement and in those with diffuse cutaneous SSc.

CONCLUSION

Our results are consistent with a moderately increased risk of bleeding in SSc patients. Further evidence from large SSc patient cohorts is needed to confirm this finding.Key Points• The risk of experiencing a major bleed may be higher among patients with SSc than the general population.• Further large and well-designed studies are needed to corroborate our findings.

Incidence of Upper and Lower Gastrointestinal Bleeding in New Users of Low-Dose Aspirin

BACKGROUND & AIMS

There are few data on the incidence of upper and lower gastrointestinal bleeding (UGIB and LGIB) from observational studies of low-dose aspirin users. We aimed to estimate incidence rates of UGIB and LGIB in a large cohort of new users of low-dose aspirin in the United Kingdom, with subanalyses of hospitalization status and fatalities.

METHODS

We performed a population-based study of 199,079 new users of low-dose aspirin (median age, 64.0 years) identified from the Health Improvement Network primary care database (2000-2012). Individuals were followed for a median 5.4 years (maximum, 14 years) to identify new cases of UGIB and LGIB. Following multistep validation, we calculated overall and age- and sex-specific incidence rates; we performed subanalyses for health care use and death within 30 days of GIB. We also estimated rates within a matched (1:1) cohort of nonusers of low-dose aspirin at the start of the follow-up period.

RESULTS

The low-dose aspirin users had 1115 UGIB events and 1936 LGIB events; most subjects with UGIB events (58.9%) were hospitalized, whereas most subjects with LGIB events were referred to secondary care (72.8%). Crude incidence rates of GIB per 1000 person-years were 0.97 for subjects with UGIB (95% CI, 0.91-1.02) and 1.68 for subjects with LGIB (95% CI, 1.60-1.75). Incidence rates per 1000 person-years for patients hospitalized for GIB were 0.57 for UGIB (95% CI, 0.53-0.61) and 0.45 for LGIB (95% CI, 0.42-0.49); for referred (but not hospitalized) cases, these values were 0.39 for UGIB (95% CI, 0.36-0.43) and 1.22 for LGIB (1.16-1.29). Incidence rates per 1000 person-years were 0.06 for fatal UGIB (95% CI, 0.04-0.07), 0.01 for fatal LGIB (95% CI, 0.01-0.02), 0.91 for nonfatal UGIB (95% CI, 0.86-0.97), and 1.66 for nonfatal LGIB (95% CI, 1.59-1.74). Among nonusers of low-dose aspirin, incidence rates per 1000 person-years were 0.67 (95% CI, 0.63-0.75) for UGIB and 0.76 (95% CI, 0.72-0.82) for LGIB.

CONCLUSION

In a population-based study of low-dose aspirin users, the incidence of LGIB was higher than the incidence of UGIB. However, incidence rates of hospitalized GI bleeds and 30-day mortality rates were lower for LGIB than for UGIB. These estimates are valuable for benefit-risk assessments of low-dose aspirin for cardiovascular and colorectal cancer prevention.

Low-dose aspirin and risk of upper/lower gastrointestinal bleeding by bleed severity: a cohort study with nested case-control analysis using primary care electronic health records from the United Kingdom

Introduction: Risks of low-dose aspirin-associated upper and lower gastrointestinal bleeds (UGIB/LGIB) may vary by severity and presence of cardiovascular disease (CVD). No study has quantified these risks for UGIB and LGIB in the same real-world study population. Patients and methods: Using UK primary care data, 199,049 new users of low-dose aspirin (75-300 mg/day) and 1:1 matched non-users were followed to identify incident UGIB (N = 1843)/LGIB (N = 2763) cases. Nested case-control analyses compared current low-dose aspirin vs. non-use on UGIB/LGIB risk. Results: Adjusted incidence rate ratios (ORs; 95% CIs) were 1.62 (1.42-1.86) for non-fatal UGIB, 1.63 (1.47-1.81) for non-fatal LGIB, 0.77 (0.51-1.16) for fatal UGIB, 1.29 (0.50-3.36) for fatal LGIB. For hospitalizations, adjusted ORs (95% CIs) were 1.55 (1.32-1.81) for UGIB and 1.89 (1.58-2.27) for LGIB; for referred only cases, they were 1.52 (1.26-1.84) for UGIB and 1.54 (1.37-1.73) for LGIB. In primary CVD prevention, adjusted ORs (95% CI) were 1.62 (1.38-1.90) for UGIB and 1.60 (1.42-1.81) for LGIB; in secondary CVD prevention, they were 1.16 (0.89-1.50) for UGIB and 1.67 (1.34-2.09) for LGIB. Conclusion: Low-dose aspirin was associated with increased risks of non-fatal but not fatal UGIB/LGIB. Key message Low-dose aspirin is associated with an increased risks of non-fatal UGIB/LGIB but not fatal UGIB/LGIB.

Relationships of H. pylori infection and its related gastroduodenal morbidity with metabolic syndrome: a large cross-sectional study

The few published studies on the relationship between Helicobacter pylori infection and metabolic homeostasis were relatively small and yielded inconsistent results. We examined the prevalence of metabolic syndrome in relation to H. pylori infection and its symptoms in a large and unselected population. Coded data from the computerised database of a large health maintenance organisation in Israel were accessed for 147,936 individuals 25–95 years of age who performed the urea breath test during 2002–2012. The classification of metabolic syndrome followed a modified definition of the international diabetes federation. Prevalences of H. pylori infection and metabolic syndrome were 52.0% and 11.4% respectively. H. pylori infected patients had increased likelihood of metabolic syndrome: adjusted odds ratio (aOR) 1.15 (95% confidence intervals (CI) 1.10–1.19), as did patients with gastric ulcer: aOR 1.15 (95% CI 1.03–1.28) vs patients without these conditions. Duodenal ulcer was associated with metabolic syndrome only in persons aged 25–34 years: aOR 1.59 (95% CI 1.19-2.13), but not in older persons (P = 0.001 for heterogeneity). In conclusion, the likelihood of metabolic syndrome appeared significantly increased in relation to H. pylori infection and gastric and duodenal ulcers. These findings suggest that H. pylori long-term gastric inflammation might play a role in metabolic homeostasis.

Informatics and machine learning to define the phenotype

INTRODUCTION

For the past decade, the focus of complex disease research has been the genotype. From technological advancements to the development of analysis methods, great progress has been made. However, advances in our definition of the phenotype have remained stagnant. Phenotype characterization has recently emerged as an exciting area of informatics and machine learning. The copious amounts of diverse biomedical data that have been collected may be leveraged with data-driven approaches to elucidate trait-related features and patterns. Areas covered: In this review, the authors discuss the phenotype in traditional genetic associations and the challenges this has imposed.Approaches for phenotype refinement that can aid in more accurate characterization of traits are also discussed. Further, the authors highlight promising machine learning approaches for establishing a phenotype and the challenges of electronic health record (EHR)-derived data. Expert commentary: The authors hypothesize that through unsupervised machine learning, data-driven approaches can be used to define phenotypes rather than relying on expert clinician knowledge. Through the use of machine learning and an unbiased set of features extracted from clinical repositories, researchers will have the potential to further understand complex traits and identify patient subgroups. This knowledge may lead to more preventative and precise clinical care.

Positive predictive value of peptic ulcer diagnosis codes in the Danish National Patient Registry

Background

Diagnoses of peptic ulcer are registered in the Danish National Patient Registry (DNPR) for administrative as well as research purposes, but it is unknown whether the coding validity depends on the location of the ulcer.

Objective

To validate the International Classification of Diseases, 10^th revision diagnosis codes of peptic ulcer in the DNPR by estimating positive predictive values (PPVs) for gastric and duodenal ulcer diagnoses.

Methods

We identified all patients registered with a hospital discharge diagnosis of peptic ulcer from Aarhus University Hospital, Denmark, in 1995–2006. Among them, we randomly selected 200 who had an outpatient gastroscopy at the time of ulcer diagnosis. We reviewed the findings from these gastroscopies to confirm the presence of peptic ulcer and its location. We calculated PPVs and corresponding 95% confidence intervals (CIs) of gastric and duodenal ulcer diagnoses, using descriptions from the gastroscopic examinations as standard reference.

Results

In total, 182 records (91%) were available for review. The overall PPV of peptic ulcer diagnoses in DNPR was 95.6% (95% CI 91.5–98.1), with PPVs of 90.3% (95% CI 82.4–95.5) for gastric ulcer diagnoses, and 94.4% (95% CI 87.4–98.2) for duodenal ulcer diagnoses. PPVs were constant over time.

Conclusion

The PPV of uncomplicated peptic ulcer diagnoses in the DNPR is high, and the location of the ulcers is registered correctly in most cases, indicating that the diagnoses are useful for research purposes.

The Role of Hemoglobin Laboratory Test Results for the Detection of Upper Gastrointestinal Bleeding Outcomes Resulting from the Use of Medications in Observational Studies

INTRODUCTION

The identification of upper gastrointestinal (UGI) bleeding and perforated ulcers in claims data typically relies on inpatient diagnoses. The use of hemoglobin laboratory results might increase the detection of UGI events that do not lead to hospitalization.

OBJECTIVES

Our objective was to evaluate whether hemoglobin results increase UGI outcome identification in electronic databases, using non-steroidal anti-inflammatory drugs (NSAIDs) as a test case.

METHODS

From three data partner sites within the Mini-Sentinel Distributed Database, we identified NSAID initiators aged ≥18 years between 2008 and 2013. Numbers of events and risks within 30 days after NSAID initiation were calculated for four mutually exclusive outcomes: (1) inpatient UGI diagnosis of bleeding or gastric ulcer (standard claims-based definition without laboratory results); (2) non-inpatient UGI diagnosis AND ≥3 g/dl hemoglobin decrease; (3) ≥3 g/dl hemoglobin decrease without UGI diagnosis in any clinical setting; (4) non-inpatient UGI diagnosis, without ≥3 g/dl hemoglobin decrease.

RESULTS

We identified 2,289,772 NSAID initiators across three sites. Overall, 45.3% had one or more hemoglobin result available within 365 days before or 30 days after NSAID initiation; only 6.8% had results before and after. Of 7637 potential outcomes identified, outcome 1 accounted for 21.7%, outcome 2 for 0.8%, outcome 3 for 34.3%, and outcome 4 for 43.3%. Potential cases identified by outcome 3 were largely not suggestive of UGI events. Outcomes 1, 2, and 4 had similar distributions of specific UGI diagnoses.

CONCLUSIONS

Using available hemoglobin result values combined with non-inpatient UGI diagnoses identified few additional UGI cases. Non-inpatient UGI diagnostic codes may increase outcome detection but would require validation.

Validity and completeness of colorectal cancer diagnoses in a primary care database in the United Kingdom

Purpose

To validate the recorded diagnoses of colorectal cancer (CRC) and identify false negatives in The Health Improvement Network (THIN) primary care database.

Methods

We conducted a validation study of incident CRC cases in THIN among patients aged 40–89 years from 2000–2011. CRC Read code entries (N = 3805) were verified by manual review of patients' electronic medical records (EMRs) including free‐text comments. Incident CRC cases in THIN ascertained following manual review were validated against two data sources deemed gold standards: (i) questionnaires sent to primary care practitioners (PCPs; for a random sample of 100 potential CRC cases), and (ii) Hospital Episode Statistics (HES) among linked practices. False negatives in THIN were identified by searching for International Classification of Diseases‐10 codes related to CRC in HES.

Results

Of 3805 CRC cases identified in THIN via Read codes, 3033 patients (80.0%) were considered definite cases after manual review of EMRs. The positive predictive value (PPV) of CRC Read codes was 86.0% after removing patients identified from THIN via a Read code for ‘fast track referral for suspected CRC’. The response rate from PCPs was 87.0% (n = 87), and the PPV of CRC in THIN was 100% based on PCP questionnaires. Using HES, the PPV for CRC in THIN was 97.9% (556/568), and false negative rate was 6.1% (36/592).

Conclusions

CRC diagnostic Read codes in THIN have a high PPV, which is increased further following manual review of free‐text comments. The false negative rate of CRC diagnoses in THIN is low. © 2015 The Authors. Pharmacoepidemiology and Drug Safety published by John Wiley & Sons Ltd.

Patterns in the use of low-dose acetylsalicylic acid and other therapies following upper gastrointestinal bleeding

BACKGROUND

Anticoagulants and/or antiplatelet agents such as acetylsalicylic acid (ASA) are important in prevention of cardiovascular (CV) events, but may be associated with upper gastrointestinal bleeding (UGIB). However, discontinuing these agents may leave patients at risk of CV events.

OBJECTIVES

This study aimed to assess patterns of therapy after UGIB in routine clinical practice.

METHODS

The Health Improvement Network UK primary care database was used to identify a cohort of patients aged 40-84 years with a UGIB event between 2000 and 2007 (n = 2,036). Patients were followed up for 1 year from the recorded UGIB. Re-prescription rates for antithrombotics and drugs that can modify the risk of UGIB were estimated at 30, 90, 180, and 365 days.

RESULTS

At 365 days, the re-prescription rate was 43 % for ASA, 66 % for warfarin, 69 % for clopidogrel, and 49 % for dipyridamole. The re-prescription rate of gastroprotective agents at 365 days for current users of histamine H2-receptor antagonists was 36 % and that of proton pump inhibitors (PPIs) was 97 %. In patients who were prescribed ASA before UGIB (n = 572), only 24 % were prescribed a PPI in the previous year. In patients who were prescribed ASA in the year after UGIB (n = 337), 92 % were prescribed a PPI.

CONCLUSIONS

Antiplatelet use fell after UGIB events. In patients who were prescribed a PPI after a UGIB event, there was increased re-prescription of antiplatelet agents and antithrombotics.

Risk factors associated with uncomplicated peptic ulcer and changes in medication use after diagnosis

Background

Few epidemiologic studies have investigated predictors of uncomplicated peptic ulcer disease (PUD) separately from predictors of complicated PUD.

Objective

To analyze risk factors associated with uncomplicated PUD and medication use after diagnosis.

Methods

Patients diagnosed with uncomplicated PUD (n = 3,914) were identified from The Health Improvement Network database among individuals aged 40–84 years during 1997–2005, with no previous history of PUD. Prescription records for the year after the date of diagnosis were reviewed and a nested case–control analysis was performed to calculate the odds ratios for the association of potential risk factors with PUD.

Results

Medications associated with developing uncomplicated PUD included current use of acetylsalicylic acid (ASA), nonsteroidal anti-inflammatory drugs (NSAIDs), paracetamol, selective serotonin reuptake inhibitors, antidepressants, antihypertensives or acid suppressants. Uncomplicated PUD was significantly associated with being a current or former smoker and having had a score of at least 3 on the Townsend deprivation index. Approximately 50% of patients who were users of ASA (19% of patients) or chronic users of NSAIDs (7% of patients) at diagnosis did not receive another prescription of the medication in the 60 days after diagnosis, and 30% were not represcribed therapy within a year. Among patients who were current users of ASA or chronic NSAIDs at the time of the PUD diagnosis and received a subsequent prescription for their ASA or NSAID during the following year, the vast majority (80–90%) also received a proton pump inhibitor coprescription.

Conclusions

Our results indicate that several risk factors for upper gastrointestinal bleeding are also predictors of uncomplicated PUD, and that some patients do not restart therapy with ASA or NSAIDs after a diagnosis of uncomplicated PUD. Further investigation is needed regarding the consequences for these patients in terms of increased cardiovascular burden due to discontinuation of antiplatelet therapy.

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

OBJECTIVE

To evaluate the accuracy of disease codes and free text in identifying upper gastrointestinal bleeding (UGIB) from electronic health-care records (EHRs).

STUDY DESIGN AND SETTING

We conducted a validation study in four European electronic health-care record (EHR) databases such as Integrated Primary Care Information (IPCI), Health Search/CSD Patient Database (HSD), ARS, and Aarhus, in which we identified UGIB cases using free text or disease codes: (1) International Classification of Disease (ICD)-9 (HSD, ARS); (2) ICD-10 (Aarhus); and (3) International Classification of Primary Care (ICPC) (IPCI). From each database, we randomly selected and manually reviewed 200 cases to calculate positive predictive values (PPVs). We employed different case definitions to assess the effect of outcome misclassification on estimation of risk of drug-related UGIB.

RESULTS

PPV was 22% [95% confidence interval (CI): 16, 28] and 21% (95% CI: 16, 28) in IPCI for free text and ICPC codes, respectively. PPV was 91% (95% CI: 86, 95) for ICD-9 codes and 47% (95% CI: 35, 59) for free text in HSD. PPV for ICD-9 codes in ARS was 72% (95% CI: 65, 78) and 77% (95% CI: 69, 83) for ICD-10 codes (Aarhus). More specific definitions did not have significant impact on risk estimation of drug-related UGIB, except for wider CIs.

CONCLUSIONS

ICD-9-CM and ICD-10 disease codes have good PPV in identifying UGIB from EHR; less granular terminology (ICPC) may require additional strategies. Use of more specific UGIB definitions affects precision, but not magnitude, of risk estimates.

Low-dose aspirin and upper gastrointestinal bleeding in primary versus secondary cardiovascular prevention: a population-based, nested case-control study

BACKGROUND

The benefit-risk profile of low-dose aspirin in primary prevention of cardiovascular disease is unclear. We sought to quantify upper gastrointestinal bleeding (UGIB) risk associated with low-dose aspirin in secondary versus primary prevention patients.

METHODS AND RESULTS

We performed a population-based nested case-control study using The Health Improvement Network (THIN) Database between 2000 and 2007. We identified 2049 cases of UGIB and 20,000 controls, frequency-matched to the cases on age, sex, and calendar year, who were subdivided into primary (without previous cardiovascular disease) and secondary (with previous cardiovascular disease) prevention populations. We estimated the relative risk of UGIB associated with the use of low-dose aspirin by multivariate logistic regression. The UGIB risk in patients taking low-dose aspirin relative to nonusers was significantly higher in the primary (adjusted relative risk, 1.90; 95% confidence interval, 1.59-2.26) than in the secondary (relative risk, 1.40; 95% confidence interval, 1.14-1.72; P value for the difference=0.0014) prevention cohort. However, as the baseline risk of UGIB was lower in the primary than in the secondary prevention cohort, numbers needed to harm per 1 year of low-dose aspirin use were 601 and 391 for primary and secondary prevention, respectively.

CONCLUSIONS

The relative risk of UGIB in patients taking low-dose aspirin is higher when used for primary than for secondary cardiovascular disease prevention, but this difference is more than compensated by the lower baseline risk in the primary prevention population. Such estimates are important for an assessment of the net clinical benefit in primary prevention.

Determining the predictive value of Read codes to identify congenital cardiac malformations in the UK Clinical Practice Research Datalink

PURPOSES

The purposes of this study were to determine (i) the positive predictive value (PPV) of multiple Read codes used to identify congenital cardiac malformation (CCM) cases in the UK Clinical Practice Research Datalink (CPRD); (ii) the accuracy of the diagnosis date; and (iii) the source of information that the general practitioners (GPs) use for validating the diagnosis suggested by the code.

METHODS

Eight hundred eighty-eight records with Read diagnostic and procedures codes for CCM, between January 1996 and November 2010, were identified from CPRD. Questionnaires were sent to GPs to verify the diagnoses and date of the code-identified events.

RESULTS

A total of 719 questionnaires were returned (81% response rate). The PPV of the CCM codes was 93% (670/719). Thirty-one percent of cases had a different event date than the one recorded in the electronic medical record (EMR); 10% of these differing dates were within 30 days of the code-identified CCM date. GPs used a variety of data sources to confirm CCM diagnoses. Although the EMR was the most frequently used data source (70%), 66% reported using consultation letters, 9% reported using clinical notes or paper charts, and 35% of GPs reported using the hospital record to confirm the CCM diagnosis.

CONCLUSIONS

Clinical Practice Research Datalink Read codes for CCMs have 93% PPV and most likely point to true cases. However, the accuracy of diagnosis dates and the age at diagnosis may not be as reliable. The findings of this study indicate that GPs use information beyond what is available for researchers in the EMR to confirm clinical diagnoses when responding to validation questionnaires. Published 2013. This article is a U.S. Government work and is in the public domain in the USA.

Clopidogrel use and short-term mortality after peptic ulcer bleeding: a population-based cohort study

Clopidogrel therapy increases bleeding risk, but whether it influences short-term mortality after peptic ulcer bleeding (PUB) is unknown. The objective was to examine whether clopidogrel use at the time of PUB increases 30-day mortality. We conducted this cohort study in northern Denmark (population 1.7 million). We used the Danish National Patient Registry, covering all hospitals, to identify all patients with a first-ever inpatient diagnosis of endoscopically or surgically confirmed PUB between 1998 and 2008 and their comorbidities. From the prescription database in the region, we ascertained the use of clopidogrel at the time of admission (current use) or before admission (former use) and use of concurrent medications. We obtained mortality data from the Danish Civil Registration System. We used regression modeling to compute mortality rate ratios (MRRs) with 95% confidence intervals (CIs), controlling for potential confounders. We identified 6951 patients with bleeding peptic ulcers. At admission, 122 (1.8%) were current users of clopidogrel, 143 (2.1%) were former users, and 6686 (96.2%) were nonusers. Thirty-day mortality was 5.7% for current users, 7.0% for former users, and 8.0% for nonusers. The adjusted 30-day MRR was reduced in both current and former users, compared with nonusers (MRR = 0.72, 95% CI 0.34, 1.52 and MRR = 0.71, 95% CI 0.38, 1.32, respectively). There was no notable modification of the association within gender or age strata. Although the use of clopidogrel increases the risk of PUB, former use and current use of clopidogrel were not associated with increased short-term mortality after admission for this condition.

Comorbidities affect risk of nonvariceal upper gastrointestinal bleeding

BACKGROUND & AIMS

The incidence of upper gastrointestinal bleeding (GIB) has not been reduced despite the decreasing incidence of peptic ulcers, strategies to eradicate Helicobacter pylori infection, and prophylaxis against ulceration from nonsteroidal anti-inflammatory drugs. Other factors might therefore be involved in the pathogenesis of GIB. Patients with GIB have increasing nongastrointestinal comorbidity, so we investigated whether comorbidity itself increased the risk of GIB.

METHODS

We conducted a matched case-control study using linked primary and secondary care data collected in England from April 1, 1997 through August 31, 2010. Patients older than 15 years with nonvariceal GIB (n = 16,355) were matched to 5 controls by age, sex, year, and practice (n = 81,636). All available risk factors for GIB were extracted and modeled using conditional logistic regression. Adjusted associations with nongastrointestinal comorbidity, defined using the Charlson Index, were then tested and sequential population attributable fractions calculated.

RESULTS

Comorbidity had a strong graded association with GIB; the adjusted odds ratio for a single comorbidity was 1.43 (95% confidence interval [CI]: 1.35-1.52) and for multiple or severe comorbidity was 2.26 (95% CI: 2.14%-2.38%). The additional population attributable fraction for comorbidity (19.8%; 95% CI: 18.4%-21.2%) was considerably larger than that for any other measured risk factor, including aspirin or nonsteroidal anti-inflammatory drug use (3.0% and 3.1%, respectively).

CONCLUSIONS

Nongastrointestinal comorbidity is an independent risk factor for GIB, and contributes to a greater proportion of patients with bleeding in the population than other recognized risk factors. These findings could help in the assessment of potential causes of GIB, and also explain why the incidence of GIB remains high in an aging population.

Risk of complications after a peptic ulcer diagnosis: effectiveness of proton pump inhibitors

BACKGROUND

Few studies have evaluated the prevention of upper gastrointestinal complications (UGIC; bleeding or perforation) in patients with uncomplicated peptic ulcer (PU). We assessed the effect of proton pump inhibitors (PPI) in a non-randomized setting. To maximize exchangeability of exposed and unexposed groups we restricted the study to patients with a new diagnosis of PU, i.e., a clear indication. To minimize selection bias we mimicked an intention to treat approach by assessing the effect of PPI prescription after PU diagnosis.

METHODS

Within a population of subjects aged 40-84 years from The Health Improvement Network database, 1997-2006, we identified 3,850 patients with incident PU. Among them, we confirmed 74 first UGIC episodes during a mean follow-up of 4 years. Exposure was prescription coverage during the month following PU diagnosis. We performed a nested case-control analysis and compared UGIC cases with 400 controls matched for age, sex, year and duration of follow-up. Relative risks (RR) and 95 % confidence intervals (CI) were estimated.

RESULTS

The overall incidence of UGIC was 4.6 cases/1,000 person-years; it was highest during the months after PU diagnosis, increased with age, and it was higher in men and subjects with Helicobacter pylori infection, anemia, and alcohol use at PU diagnosis. The RR for UGIC associated with PPI prescriptions during the month after PU diagnosis was 0.56 (95 % CI 0.31-1.0). The RR for NSAIDs with and without a PPI was 1.72 (0.68-4.45) and 3.27 (0.85-12.67), respectively.

CONCLUSIONS

Findings suggest that prescription of PPIs after a PU diagnosis is associated with a reduced risk of UGIC.

Statins and risk of diabetes: an analysis of electronic medical records to evaluate possible bias due to differential survival

OBJECTIVE

Two meta-analyses of randomized trials of statins found increased risk of type 2 diabetes. One possible explanation is bias due to differential survival when patients who are at higher risk of diabetes survive longer under statin treatment.

RESEARCH DESIGN AND METHODS

We used electronic medical records from 500 general practices in the U.K. and included data from 285,864 men and women aged 50-84 years from January 2000 to December 2010. We emulated the design and analysis of a hypothetical randomized trial of statins, estimated the observational analog of the intention-to-treat effect, and adjusted for differential survival bias using inverse-probability weighting.

RESULTS

During 1.2 million person-years of follow-up, there were 13,455 cases of type 2 diabetes and 8,932 deaths. Statin initiation was associated with increased risk of type 2 diabetes. The hazard ratio (95% CI) of diabetes was 1.45 (1.39-1.50) before adjusting for potential confounders and 1.14 (1.10-1.19) after adjustment. Adjusting for differential survival did not change the estimates. Initiating atorvastatin and simvastatin was associated with increased risk of type 2 diabetes.

CONCLUSIONS

In this sample of the general population, statin therapy was associated with 14% increased risk of type 2 diabetes. Differential survival did not explain this increased risk.

Methods and dimensions of electronic health record data quality assessment: enabling reuse for clinical research

OBJECTIVE

To review the methods and dimensions of data quality assessment in the context of electronic health record (EHR) data reuse for research.

MATERIALS AND METHODS

A review of the clinical research literature discussing data quality assessment methodology for EHR data was performed. Using an iterative process, the aspects of data quality being measured were abstracted and categorized, as well as the methods of assessment used.

RESULTS

Five dimensions of data quality were identified, which are completeness, correctness, concordance, plausibility, and currency, and seven broad categories of data quality assessment methods: comparison with gold standards, data element agreement, data source agreement, distribution comparison, validity checks, log review, and element presence.

DISCUSSION

Examination of the methods by which clinical researchers have investigated the quality and suitability of EHR data for research shows that there are fundamental features of data quality, which may be difficult to measure, as well as proxy dimensions. Researchers interested in the reuse of EHR data for clinical research are recommended to consider the adoption of a consistent taxonomy of EHR data quality, to remain aware of the task-dependence of data quality, to integrate work on data quality assessment from other fields, and to adopt systematic, empirically driven, statistically based methods of data quality assessment.

CONCLUSION

There is currently little consistency or potential generalizability in the methods used to assess EHR data quality. If the reuse of EHR data for clinical research is to become accepted, researchers should adopt validated, systematic methods of EHR data quality assessment.

Erosive reflux disease increases risk for esophageal adenocarcinoma, compared with nonerosive reflux

BACKGROUND & AIMS

Gastroesophageal reflux disease is a strong risk factor for esophageal adenocarcinoma, but it is not clear whether the mucosal inflammation that develops in patients with reflux disease promotes this cancer. We determined the development of adenocarcinoma among patients who underwent esophagogastroduodenoscopy and were found to have erosive (with esophagitis) or nonerosive (without esophagitis) reflux.

METHODS

We performed a nationwide cohort study using data from 33,849 patients with reflux disease (52% men; median age, 59.3 y) from population-based Danish medical registries, from 1996 through 2008. The observed incidences of adenocarcinoma were compared with the expected incidence for the general population, standardized by age, sex, and calendar time. Absolute risks were estimated using Kaplan-Meier methods.

RESULTS

In the study cohort, 26,194 of the patients (77%) had erosive reflux disease and 37 subsequently developed esophageal adenocarcinoma after a mean follow-up time of 7.4 years. Their absolute risk after 10 years was 0.24% (95% confidence interval [CI], 0.15%-0.32%). The incidence of cancer among patients with erosive reflux disease was significantly greater than that expected for the general population (standardized incidence ratio, 2.2; 95% CI, 1.6-3.0). In contrast, of the 7655 patients with nonerosive reflux disease, only 1 was diagnosed with esophageal adenocarcinoma after 4.5 years of follow-up evaluation (standardized incidence ratio, 0.3; 95% CI, 0.01-1.5).

CONCLUSIONS

Erosive reflux disease, but not nonerosive disease, increased the risk of esophageal adenocarcinoma, based on analysis of population-based Danish medical registries. Inflammation therefore might be an important factor in the progression from reflux to esophageal adenocarcinoma.

Validation of diabetic retinopathy and maculopathy diagnoses recorded in a U.K. primary care database

OBJECTIVE

To describe the validity of recorded diabetic retinopathy (DR) and diabetic maculopathy (DMP) diagnoses, including edema (DMO) in The Health Improvement Network (THIN) database.

RESEARCH DESIGN AND METHODS

In two independent computer searches, we detected 20,838 patients with diabetes aged 1-84 years with a first DR computer Read entry in 2000-2008 and 4,064 with a first DMP entry. A two-step strategy was used to validate both outcomes as follows: 1) review of patient profiles including free-text comments from primary care practitioners (PCPs) (containing referral information and test results) of a random sample of 500 DR and all DMP computer-detected patients. We classified them in probable, possible, and noncase according to the diagnosis plausibility based on the manual review of the computerized information; and 2) review of questionnaires sent by PCPs and medical records in a random sample (N = 200 for each outcome including 36 diabetic macular edema [DMO]). Gold standard was PCPs' confirmation.

RESULTS

After profiles review, we categorized 418 as probable/possible DR. In addition, 3,676 DMP were categorized as probable/possible (including 711 DMO). After review of information sent by PCPs, confirmation rates were 87.3 and 87.2%, respectively (90.3% for DMO). When we applied them to the whole sample of computer-detected patients, the weighted confirmation rate was 78.0% for DR and 78.8% for DMP (86.2% for DMO).

CONCLUSIONS

Read codes for DR, DM, and DMO are moderately accurate in identifying incident case subjects of these ophthalmologic complications. The validity improved when incorporating PCPs' text comments to the patient's profile. THIN database proved to be a valuable resource to study ophthalmological diabetes complications.

Can primary care data be used to monitor regional smoking prevalence? An analysis of The Health Improvement Network primary care data

Background

Accurate and timely regional data on smoking trends allow tobacco control interventions to be targeted at the areas most in need and facilitate the evaluation of such interventions. Electronic primary care databases have the potential to provide a valuable source of such data due to their size, continuity and the availability of socio-demographic data. UK electronic primary care data on smoking prevalence from The Health Improvement Network (THIN) have previously been validated at the national level, but may be less representative at the regional level due to reduced sample sizes. We investigated whether this database provides valid regional data and whether it can be used to compare smoking prevalence in different UK regions.

Methods

Annual estimates of smoking prevalence by government office region (GOR) from THIN were compared with estimates of smoking prevalence from the General Lifestyle Survey (GLF) from 2000 to 2008.

Results

For all regions, THIN prevalence data were generally found to be highly comparable with GLF data from 2006 onwards.

Conclusions

THIN primary care data could be used to monitor regional smoking prevalence and highlight regional differences in smoking in the UK.

Temporal and within practice variability in the health improvement network

BACKGROUND

The Health Improvement Network (THIN) database is a primary care electronic medical record database in the UK designed for pharmacoepidemiologic research. Matching on practice and calendar year often is used to account for secular trends in time and differences across practices. However, little is known about the consistency within practices across observation years and among practices within a given year, in THIN or other large medical record databases.

METHODS

We analyzed mortality rates, cancer incidence rates, prescribing rates, and encounter rates across 415 practices from 2000 to 2007 using a practice-year as the unit of observation in separate random and fixed effects longitudinal Poisson regression models. Adjusted models accounted for aggregate practice-level characteristics (smoking, obesity, age, and Vision software experience).

RESULTS

In adjusted models, subsequent calendar years were associated with lower reported mortality rates, increasing cancer reporting rates, increasing prescriptions per patient, and decreasing encounters per patient, with a corresponding linear trend (p < 0.001 for all analyses). For calendar year 2007, the ratio of the 75th percentile to the 25th percentile for crude rate of cancer, mortality, prescriptions, and encounters was 1.63, 1.63, 1.45, and 1.42, respectively. Adjusting for practice characteristics reduced the among-practice variation by approximately 40%.

CONCLUSIONS

THIN data are characterized by secular trends and among-practice variation, both of which should be considered in the design of pharmacoepidemiology studies. Whether these are trends in data quality or true secular trends could not be definitively differentiated.

Acid suppressants reduce risk of gastrointestinal bleeding in patients on antithrombotic or anti-inflammatory therapy

BACKGROUND & AIMS

We investigated the effect of different prevention strategies against upper gastrointestinal bleeding (UGIB) in the general population and in patients on antithrombotic or anti-inflammatory treatments.

METHODS

We performed a population-based, nested, case-control study using The Health Improvement Network UK primary care database. From 2000 to 2007, we identified 2049 cases of UGIB and 20,000 controls. The relative risk (RR) of UGIB associated with various gastroprotective agents was estimated by comparing current use (defined as use within 30 days of the index date) with nonuse in the previous year, using multivariate logistic regression.

RESULTS

The adjusted RR of UGIB associated with current use of proton pump inhibitors (PPIs) for more than 1 month was 0.58 (95% confidence interval [CI], 0.42-0.79) among patients who received low-dose acetylsalicylic acid (ASA), 0.18 (95% CI, 0.04-0.79) for clopidogrel, 0.17 (95% CI, 0.04-0.76) for dual antiplatelet therapy, 0.48 (95% CI, 0.22-1.04) for warfarin, and 0.51 (95% CI, 0.34-0.78) for nonsteroidal anti-inflammatory drugs. The corresponding estimates for therapy with histamine-2-receptor antagonists (H2RAs) were more unstable, but tended to be of a smaller magnitude. In the general population, PPI use was associated with a reduced risk of UGIB compared with nonuse (RR, 0.80; 95% CI, 0.68-0.94); no such reduction was observed for H2RAs or nitrates.

CONCLUSIONS

PPI use was associated with a lower risk of UGIB in the general population and in patients on antithrombotic or anti-inflammatory therapy compared with nonuse of PPIs. The reduction in risks of UGIB was smaller in H2RA than in PPI users.

Systematic review of peptic ulcer disease incidence rates: do studies without validation provide reliable estimates?

PURPOSE

Incidence rate (IR) estimates for peptic ulcer disease (PUD) vary widely among studies. We conducted a systematic review to quantify and examine the discrepancies.

METHODS

Of 4780 articles identified from PubMed and EMBASE databases, 31 published in the last three decades that had reported IRs of PUD in the general population were included. Random effects meta-analysis and meta-regression were performed to calculate pooled estimates and to identify sources of heterogeneity.

RESULTS

The pooled IR estimate per 1000 person-years was 0.90 (95% confidence interval: 0.78-1.04) for uncomplicated PUD, 0.57 (0.49-0.65) for peptic ulcer bleeding, 0.10 (0.08-0.13) for gastrointestinal perforations, and 3.18 (2.05-4.92) for nonspecific PUD. Within specific outcomes definitions, IR estimates were significantly lower in studies with restriction to hospitalized cases, case validation, and case ascertainment directly from hospital or clinical sources versus computerized health care databases. Younger age, female sex, and later calendar time were also associated with lower PUD incidence.

CONCLUSIONS

We found that the IR of uncomplicated PUD was in the order of one case per 1000 person-years in the general population, and that the IR of peptic ulcer complications was around 0.7 cases per 1000 person-years. Comparisons of IR estimates among studies need to take into account disease definition and other study characteristics, particularly whether outcome validation was performed in computerized claims. The use of claims to identify PUD cases might overestimate the IR by around 45%.

Validity of The Health Improvement Network (THIN) for the study of psoriasis

BACKGROUND

Psoriasis is a common disease frequently studied in large databases. To date the validity of psoriasis information has not been established in The Health Improvement Network (THIN).

OBJECTIVES

To investigate the validity of THIN for identifying patients with psoriasis and to determine if the database can be used to determine the natural history of the disease.

METHODS

First, we conducted a cross-sectional study to determine if psoriasis prevalence in THIN is similar to expected. Second, we created a cohort of 4900 patients, aged 45-64 years, with a psoriasis diagnostic Read Code and surveyed their general practitioners (GPs) to confirm the diagnosis clinically. Third, we created models to determine if psoriasis descriptors (extent, severity, duration and dermatologist confirmation) could be accurately captured from database records.

RESULTS

Psoriasis prevalence was 1·9%, and showed the characteristic age distribution expected. GP questionnaires were received for 4634 of 4900 cohort patients (95% response rate), and psoriasis diagnoses were confirmed in 90% of patients. Duration of disease in the database showed substantial agreement with physician query (κ = 0·69). GPs confirmed that the psoriasis diagnosis was corroborated by a dermatologist in 91% of patients whose database records contained a dermatology referral code associated with a psoriasis code. We achieved good discrimination between patients with and without extensive disease based on the number of psoriasis codes received per year (area under curve = 0·8).

CONCLUSIONS

THIN is a valid data resource for studying psoriasis and can be used to identify characteristics of the disease such as duration and confirmation by a dermatologist.

Risk of upper gastrointestinal bleeding with low-dose acetylsalicylic acid alone and in combination with clopidogrel and other medications

BACKGROUND

This study evaluated the risk of upper gastrointestinal bleeding (UGIB) associated with use of low-dose acetylsalicylic acid (ASA) alone and in combination with other gastrotoxic medications.

METHODS AND RESULTS

The Health Improvement Network UK primary care database was used to identify individuals 40 to 84 years of age with a UGIB diagnosis in 2000 to 2007 (n = 2049). An age-, sex-, and calendar year-matched control group (n = 20,000) was identified from the same source population. The relative risk (RR) of UGIB associated with use of low-dose ASA (75 to 300 mg/d), clopidogrel, and other commonly coadministered medications was estimated by multivariate logistic regression. The risk of UGIB was increased in current users of low-dose ASA (RR, 1.80; 95% confidence interval [CI], 1.59 to 2.03) or clopidogrel (RR, 1.67; 95% CI, 1.24 to 2.24) compared with nonusers. Compared with low-dose ASA monotherapy, the risk of UGIB was significantly increased when low-dose ASA was coadministered with clopidogrel (RR, 2.08; 95% CI, 1.34 to 3.21), oral anticoagulants (RR, 2.00; 95% CI, 1.15 to 3.45), low-/medium-dose nonsteroidal antiinflammatory drugs (RR, 2.63; 95% CI, 1.93 to 3.60), high-dose nonsteroidal antiinflammatory drugs (RR, 2.66; 95% CI, 1.88 to 3.76), or high-dose oral corticosteroids (RR, 4.43; 95% CI, 2.10 to 9.34); this was not apparent with coadministration of statins (RR, 0.99; 95% CI, 0.81 to 1.21) or low-dose oral corticosteroids (RR, 1.01; 95% CI, 0.58 to 1.77).

CONCLUSIONS

Use of low-dose ASA is associated with an almost 2-fold increase in the risk of UGIB compared with nonuse. This risk is increased further in individuals taking low-dose ASA along with clopidogrel, oral anticoagulants, nonsteroidal antiinflammatory drugs, or high-dose oral corticosteroids.

Uncomplicated peptic ulcer in the UK: trends from 1997 to 2005

BACKGROUND

Few studies have examined the incidence of uncomplicated peptic ulcer or the trends in factors affecting its incidence.

AIM

To estimate the incidence rate of uncomplicated peptic ulcer in the UK from 1997 to 2005 and report temporal changes in the main known preventive and risk factors.

METHODS

Population-based cohort study of 1 049 689 patients enrolled in The Health Improvement Network in the UK. We estimated the incidence rate of uncomplicated peptic ulcer and evaluated temporal trends in demographic characteristics and prescription patterns for various anti-inflammatory and gastroprotective agents.

RESULTS

Overall uncomplicated peptic ulcer incidence was 0.75 cases per 1000 persons-years, declining from 1.1 to 0.52 cases per 1000 person-years between 1997 and 2005. Distributions of age, gender and alcohol habits were similar in 1997 and 2005. The proportion of documented Helicobacter pylori-negative cases increased from 5% to 12%. Monthly prevalence of subjects with prescriptions for traditional non-aspirin NSAIDs changed from 7.7% to 6.8%, Coxibs from 0% to 0.7%, and proton pump inhibitors (PPIs) from 2.4% to 7.4%. The proportion of subjects on prescription NSAIDs on PPIs increased continuously over time.

CONCLUSION

A reduction in H. pylori-related peptic ulcers, changing patterns in NSAID use and increasing PPI use may have contributed to a decline in uncomplicated peptic ulcer incidence in the UK.