Identifying priorities in methodological research using ICD-9-CM and ICD-10 administrative data: report from an international consortium

Barriers and facilitators for using administrative data for surveillance purpose: a narrative overview

Although administrative data are not originally intended for surveillance purposes, they are frequently used for monitoring public health and patient safety. This article provides a narrative overview of the barriers and facilitators for the use of administrative data for surveillance, with a focus on healthcare-associated infection (HAI) in Canada. In this case, only articles on administrative data in general or related to HAI were included. Validation study and meta-analyses on administrative data accuracy were excluded. Medline, Embase and Google Scholar were searched as well as references list of all included articles, for a total of 90 articles included. Our analysis identified 78 barriers at the individual, organizational and systemic levels and outlined 75 facilitators and solutions to improve administrative data utilization and quality. This narrative overview will help to understand barriers, facilitators and offer practical recommendations for optimizing the use of administrative data.

Development and validation of a claims-based algorithm to identify patients with Neuromyelitis Optica Spectrum disorder

INTRODUCTION

No validated algorithm exists to identify patients with neuromyelitis optica spectrum disorder (NMOSD) in healthcare claims data. We developed and tested the performance of a healthcare claims-based algorithm to identify patients with NMOSD.

METHODS

Using medical record data of 101 adults with NMOSD, multiple sclerosis (MS), or myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), we tested the sensitivity and specificity of claims-based algorithms developed through interviews with neurologists. We tested the best-performing algorithm's face validity using 2016-2019 data from IBM MarketScan Commercial and Medicare Supplemental databases. Demographics and clinical characteristics were reported.

RESULTS

Algorithm inclusion criteria were age ≥ 18 years and (≥1 NMO diagnosis [or ≥ 1 transverse myelitis (TM) and ≥ 1 optic neuritis (ON) diagnosis] and ≥ 1 NMOSD drug) or (≥2 NMO diagnoses ≥90 days apart). Exclusion criteria were MS diagnosis or use of MS-specific drug after last NMO diagnosis or NMOSD drug; sarcoidosis diagnosis after last NMO diagnosis; or use of ≥1 immune checkpoint inhibitor. In medical record billing data of 50 patients with NMOSD, 30 with MS, and 21 with MOGAD, the algorithm had 82.0% sensitivity and 70.6% specificity. When applied to healthcare claims data, demographic and clinical features of the identified cohort were similar to known demographics of NMOSD.

CONCLUSIONS

This clinically derived algorithm performed well in medical records. When tested in healthcare claims, demographics and clinical characteristics were consistent with previous clinical findings. This algorithm will enable a more accurate estimation of NMOSD disease burden using insurance claims datasets.

Risk of Ischemic Stroke in Relation to Helicobacter pylori Infection and Eradication Status: A Large-Scale Prospective Observational Cohort Study

Background/Aims

: A few studies have suggested the association between Helicobacter pylori (HP) infection and ischemic stroke. However, the impact of HP eradication on stroke risk has not been well evaluated. This study aimed to assess the influence of HP eradication on the incidence of ischemic stroke, considering the potential effect of sex.

Methods

: This prospective observational cohort study was conducted at Seoul National University Bundang Hospital, from May 2003 to February 2023, and involved gastroscopy-based HP testing. Propensity score (PS) matching was employed to ensure balanced groups by matching patients in the HP eradicated group (n=2,803) in a 3:1 ratio with patients in the HP non-eradicated group (n=960). Cox proportional hazard regression analysis was used to evaluate the risk of ischemic stroke.

Results

: Among 6,664 patients, multivariate analysis after PS matching indicated that HP eradication did not significantly alter the risk of ischemic stroke (hazard ratio, 0.531; 95% confidence interval, 0.221 to 1.270; p=0.157). Sex-specific subgroup analyses, both univariate and multivariate, did not yield statistically significant differences. However, Kaplan-Meier analysis revealed a potential trend: the females in the HP eradicated group exhibited a lower incidence of ischemic stroke than those in the HP non-eradicated group, although this did not reach statistical significance (p=0.057).

Conclusions

: This finding suggests that HP eradication might not impact the risk of ischemic stroke. However, there was a trend showing that females potentially had a lower risk of ischemic stroke following HP eradication, though further investigation is required to establish definitive evidence.

Accuracy of routinely collected hospital administrative discharge data and death certificate ICD-10 diagnostic coding in progressive supranuclear palsy and corticobasal syndrome: a systematic review and validation study

BACKGROUND

We conducted a systematic review to identify existing ICD-10 coding validation studies in progressive supranuclear palsy and corticobasal syndrome [PSP/CBS]) and, in a new study, evaluated the accuracy of ICD-10 diagnostic codes for PSP/CBS in Scottish hospital inpatient and death certificate data.

METHODS

Original studies that assessed the accuracy of specific ICD-10 diagnostic codes in PSP/CBS were sought. Separately, we estimated the positive predictive value (PPV) of specific codes for PSP/CBS in inpatient hospital data (SMR01, SMR04) compared to clinical diagnosis in four regions. Sensitivity was assessed in one region due to a concurrent prevalence study. For PSP, the consistency of the G23.1 code in inpatient and death certificate coding was evaluated across Scotland.

RESULTS

No previous ICD-10 validation studies were identified. 14,767 records (SMR01) and 1497 records (SMR04) were assigned the candidate ICD-10 diagnostic codes between February 2011 and July 2019. The best PPV was achieved with G23.1 (1.00, 95% CI 0.93-1.00) in PSP and G23.9 in CBS (0.20, 95% CI 0.04-0.62). The sensitivity of G23.1 for PSP was 0.52 (95% CI 0.33-0.70) and G31.8 for CBS was 0.17 (95% CI 0.05-0.45). Only 38.1% of deceased G23.1 hospital-coded cases also had this coding on their death certificate: the majority (49.0%) erroneously assigned the G12.2 code.

DISCUSSION

The high G23.1 PPV in inpatient data shows it is a useful tool for PSP case ascertainment, but death certificate coding is inaccurate. The PPV and sensitivity of existing ICD-10 codes for CBS are poor due to a lack of a specific code.

Police and hospital data linkage for traffic injury surveillance: A systematic review

This systematic review examines studies of traffic injury that involved linkage of police crash data and hospital data and were published from 1994 to 2023 worldwide in English. Inclusion and exclusion criteria were the basis for selecting papers from PubMed, Web of Science, and Scopus, and for identifying additional relevant papers using PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and supplementary snowballing (n = 60). The selected papers were reviewed in terms of research objectives, data items and sample size included, temporal and spatial coverage, linkage methods and software tools, as well as linkage rates and most significant findings. Many studies found that the number of clinically significant road injury cases was much higher according to hospital data than crash data. Under-estimation of cases in crash data differs by road user type, pedestrian cases commonly being highly under-counted. A limited number of the papers were from low- and middle-income countries. The papers reviewed lack consistency in what was reported and how, which limited comparability.

Construction and evaluation of neonatal respiratory failure risk prediction model for neonatal respiratory distress syndrome

BACKGROUND

Neonatal respiratory distress syndrome (NRDS) is a common respiratory disease in preterm infants, often accompanied by respiratory failure. The aim of this study was to establish and validate a nomogram model for predicting the probability of respiratory failure in NRDS patients.

METHODS

Patients diagnosed with NRDS were extracted from the MIMIC-iv database. The patients were randomly assigned to a training and a validation cohort. Univariate and stepwise Cox regression analyses were used to determine the prognostic factors of NRDS. A nomogram containing these factors was established to predict the incidence of respiratory failure in NRDS patients. The area under the receiver operating characteristic curve (AUC), receiver operating characteristic curve (ROC), calibration curves and decision curve analysis were used to determine the effectiveness of this model.

RESULTS

The study included 2,705 patients with NRDS. Univariate and multivariate stepwise Cox regression analysis showed that the independent risk factors for respiratory failure in NRDS patients were gestational age, pH, partial pressure of oxygen (PO2), partial pressure of carbon dioxide (PCO2), hemoglobin, blood culture, infection, neonatal intracranial hemorrhage, Pulmonary surfactant (PS), parenteral nutrition and respiratory support. Then, the nomogram was constructed and verified.

CONCLUSIONS

This study identified the independent risk factors of respiratory failure in NRDS patients and used them to construct and evaluate respiratory failure risk prediction model for NRDS. The present findings provide clinicians with the judgment of patients with respiratory failure in NRDS and help clinicians to identify and intervene in the early stage.

Exploring the reliability of inpatient EMR algorithms for diabetes identification

INTRODUCTION

Accurate identification of medical conditions within a real-time inpatient setting is crucial for health systems. Current inpatient comorbidity algorithms rely on integrating various sources of administrative data, but at times, there is a considerable lag in obtaining and linking these data. Our study objective was to develop electronic medical records (EMR) data-based inpatient diabetes phenotyping algorithms.

MATERIALS AND METHODS

A chart review on 3040 individuals was completed, and 583 had diabetes. We linked EMR data on these individuals to the International Classification of Disease (ICD) administrative databases. The following EMR-data-based diabetes algorithms were developed: (1) laboratory data, (2) medication data, (3) laboratory and medications data, (4) diabetes concept keywords and (5) diabetes free-text algorithm. Combined algorithms used or statements between the above algorithms. Algorithm performances were measured using chart review as a gold standard. We determined the best-performing algorithm as the one that showed the high performance of sensitivity (SN), and positive predictive value (PPV).

RESULTS

The algorithms tested generally performed well: ICD-coded data, SN 0.84, specificity (SP) 0.98, PPV 0.93 and negative predictive value (NPV) 0.96; medication and laboratory algorithm, SN 0.90, SP 0.95, PPV 0.80 and NPV 0.97; all document types algorithm, SN 0.95, SP 0.98, PPV 0.94 and NPV 0.99.

DISCUSSION

Free-text data-based diabetes algorithm can yield comparable or superior performance to a commonly used ICD-coded algorithm and could supplement existing methods. These types of inpatient EMR-based algorithms for case identification may become a key method for timely resource planning and care delivery.

Emergency department performance assessment using administrative data: A managerial framework

Administrative data play an important role in performance monitoring of healthcare providers. Nonetheless, little attention has been given so far to the emergency department (ED) evaluation. In addition, most of existing research focuses on a single core ED function, such as treatment or triage, thus providing a limited picture of performance. The goal of this study is to harness the value of routinely produced records proposing a framework for multidimensional performance evaluation of EDs able to support internal decision stakeholders in managing operations. Starting with the overview of administrative data, and the definition of the desired framework's characteristics from the perspective of decision stakeholders, a review of the academic literature on ED performance measures and indicators is conducted. A performance measurement framework is designed using 224 ED performance metrics (measures and indicators) satisfying established selection criteria. Real-world feedback on the framework is obtained through expert interviews. Metrics in the proposed ED performance measurement framework are arranged along three dimensions: performance (quality of care, time-efficiency, throughput), analysis unit (physician, disease etc.), and time-period (quarter, year, etc.). The framework has been judged as "clear and intuitive", "useful for planning", able to "reveal inefficiencies in care process" and "transform existing data into decision support information" by the key ED decision stakeholders of a teaching hospital. Administrative data can be a new cornerstone for health care operation management. A framework of ED-specific indicators based on administrative data enables multi-dimensional performance assessment in a timely and cost-effective manner, an essential requirement for nowadays resource-constrained hospitals. Moreover, such a framework can support different stakeholders' decision making as it allows the creation of a customized metrics sets for performance analysis with the desired granularity.

Diagnostic accuracy of algorithms to define incident and second hip fractures: A Taiwan validation study

BACKGROUND

Previous epidemiological researchers have used various algorithms to identify a second hip fracture; however, there has been no validation of these algorithms to date. This study aimed to verify existing algorithms for identifying second hip fracture under the International Classification of Diseases diagnostic coding systems. Furthermore, we examined the validity of two newly proposed algorithms that integrated the concept of periprosthetic fractures and laterality of the ICD-10 coding system.

METHODS

Claims data of patients hospitalized for hip fracture from National Taiwan University Hospitals between 2007 and 2020 were retrieved. Hip fracture was confirmed by 2 orthopaedic surgeons with medical records and imaging data as gold standards. The validity of 9 existing and 2 newly proposed algorithms for identifying second hip fracture was evaluated.

RESULTS

The positive predictive value (PPV) range between 84% and 90% in existing algorithms for identifying second hip fractures. Noteworthy, the longer time interval for discrimination resulted in slightly increased PPV (from 87% to 90%), while decreased sensitivity noticeably (from 87% to 72%). When considering the information about periprosthetic fracture, the PPV increased to 91% without diminished sensitivity. The PPV of the newly proposed ICD-10-specific algorithm was 100%.

CONCLUSION

Algorithms integrated clinical insights of periprosthetic fractures and laterality concept of ICD-10 coding system provided satisfactory validity and help precisely define second hip fracture in future database research.

Evaluation of the reported data linkage process and associated quality issues for linked routinely collected healthcare data in multimorbidity research: a systematic methodology review

OBJECTIVE

The objective of this systematic review was to examine how the record linkage process is reported in multimorbidity research.

METHODS

A systematic search was conducted in Medline, Web of Science and Embase using predefined search terms, and inclusion and exclusion criteria. Published studies from 2010 to 2020 using linked routinely collected data for multimorbidity research were included. Information was extracted on how the linkage process was reported, which conditions were studied together, which data sources were used, as well as challenges encountered during the linkage process or with the linked dataset.

RESULTS

Twenty studies were included. Fourteen studies received the linked dataset from a trusted third party. Eight studies reported variables used for the data linkage, while only two studies reported conducting prelinkage checks. The quality of the linkage was only reported by three studies, where two reported linkage rate and one raw linkage figures. Only one study checked for bias by comparing patient characteristics of linked and non-linked records.

CONCLUSIONS

The linkage process was poorly reported in multimorbidity research, even though this might introduce bias and potentially lead to inaccurate inferences drawn from the results. There is therefore a need for increased awareness of linkage bias and transparency of the linkage processes, which could be achieved through better adherence to reporting guidelines.

PROSPERO REGISTRATION NUMBER

CRD42021243188.

Operational Definition Identifying Osteoporotic Vertebral Fractures in the Claims Database

BACKGROUND

We analyzed the International Classification of Diseases, 10th edition (ICD-10) diagnostic codes, procedure codes, and radiographic image codes for vertebral fracture (VF) used in the database of Health Insurance Review and Assessment Service (HIRA) of Korea to establish a validated operational definition for identifying patients with osteoporotic VF in claims data.

METHODS

We developed three operational definitions for detecting VFs using 9 diagnostic codes, 5 procedure codes and 4 imaging codes. Medical records and radiographs of 2,819 patients, who had primary and subordinated codes of VF between January 2016 and December 2016 at two institutions, were reviewed to detect true vertebral fractures. We evaluated the sensitivity and positive predictive value (PPV) of the operational definition in detecting true osteoporotic VF and obtained the receiver operating characteristic (ROC) curve.

RESULTS

Among the 2,819 patients who had primary or secondary diagnosis codes for VF, 995 patients satisfied at least one of the criteria for the operational definition of osteoporotic VF. Of these patients, 594 were judged as having true fractures based on medical records and radiographic examinations. The sensitivity and PPV were 62.5 (95% confidence interval [CI], 59.4-65.6) and 59.7(95% CI, 56.6-62.8) respectively. In the receiver operating characteristic analysis, area under the curve (AUC) was 0.706 (95% CI, 0.688-0.724).

CONCLUSION

Our findings demonstrate the validity of our operational definitions to identify VFs more accurately using claims data. This algorithm to identify VF is likely to be useful in future studies for diagnosing osteoporotic VF.

Validation of Neurologic Impairment Diagnosis Codes as Signifying Documented Functional Impairment in Hospitalized Children

OBJECTIVE

To assess the performance of previously published high-intensity neurologic impairment (NI) diagnosis codes in identification of hospitalized children with clinical NI.

METHODS

Retrospective study of 500 randomly selected discharges in 2019 from a freestanding children's hospital. All charts were reviewed for 1) NI discharge diagnosis codes and 2) documentation of clinical NI (a neurologic diagnosis and indication of functional impairment like medical technology). Test statistics of clinical NI were calculated for discharges with and without an NI diagnosis code. A sensitivity analysis varied the threshold for "substantial functional impairment." Secondary analyses evaluated misclassified discharges and a more stringent definition for NI.

RESULTS

Diagnosis codes identified clinically documented NI with 88.1% (95% confidence interval [CI]: 84.7, 91) specificity, and 79.4% (95% CI: 67.3, 88.5) sensitivity; negative predictive value (NPV) was 96.7% (95% CI: 94.8, 98.0), and positive predictive value (PPV) was 49% (95% CI: 42, 56.1). Including children with milder functional impairment (lower threshold) resulted in NPV of 95.7% and PPV of 77.5%. Restricting to children with more severe functional impairment (higher threshold) resulted in NPV of 98.2% and PPV of 44.1%. Misclassification was primarily due to inclusion of children without functional impairments. A more stringent NI definition including diagnosis codes for NI and feeding tubes had a specificity of 98.4% (95% CI: 96.7-99.3) and sensitivity of 28.6% (19.4-41.3).

CONCLUSIONS

All scenarios evaluated demonstrated high NPV and low-to-moderate PPV of the diagnostic code list. To maximize clinical utility, NI diagnosis codes should be used with strategies to mitigate the risk of misclassification.

Comparison of crosswalk methods for translating ICD-9 to ICD-10 diagnosis codes for child maltreatment

BACKGROUND

A validated source of International Classification of Disease (1CD) 10th revision diagnostic codes to identify child maltreatment has not been developed. Such a reference would be essential for the practical purposes of administrative data-based research and public health surveillance.

OBJECTIVE

To evaluate the validity of diagnosis code classifications for child maltreatment following conversion from ICD 9th edition, clinical modification (ICD-9-CM) to 10th revision.

PARTICIPANTS AND SETTING

Children receiving inpatient or emergency medical care in the United States with ICD-9-CM/ICD-10-CM diagnosis codes for child maltreatment, identified using two large multicenter hospital-based datasets.

METHODS

We evaluated the performance of general equivalence mappings (GEMs) and network-based mappings for previously-validated ICD-9-CM diagnosis codes for child maltreatment in the 2013-2014 PHIS and 2012 KID and resulting ICD-10-CM codes in the 2018-2019 PHIS and 2016 KID datasets.

RESULTS

Of 56 previously-validated ICD-9-CM diagnoses, GEMs identified 15 with a similar proportion of diagnosed children in the KID ICD-9-CM and ICD-10 eras and 14 diagnoses in PHIS. Network-based mapping identified 18 diagnoses with similar proportions in the KID datasets, and 13 diagnoses in PHIS. For six diagnoses, the proportion of children identified in the ICD-10 era using network-based mapping was more than ten times the proportion identified in the ICD-9-CM era.

CONCLUSION

Neither crosswalk method provided consistently reliable conversions, due to both crosswalk methodology and changes introduced by the ICD 10th revision. These findings highlight the need for independent construction and validation of ICD-10-based definitions of child maltreatment as a precursor to administrative data-based research and public health surveillance.

Techniques to Assess Long-Term Outcomes after Burn Injuries

Burn injuries have a tremendous impact on not only the physical health of the burn survivor, but also mental health and social outcomes of the individual and their support systems. While much effect occurs at the point of injury, post-injury pain, infection, scarring, inflammatory response and metabolic changes all impact the long-term health of the burn survivor. The goal of the following article is to explore how to examine long term outcomes associated with burn injury, including mental disorders, suicide, loss of work and quality of life in the context of risk factors for burn injury, including social determinants of health. We then discuss ways to examine post-burn outcomes, including the important role of administrative data, the advantages of mixed methodology research studies including qualitative research, and the importance of considering sex, gender and vulnerable populations, not only in study design, but in prevention and intervention programs.

Transitions between versions of the International Classification of Diseases and chronic disease prevalence estimates from administrative health data: a population-based study

Background

Diagnosis codes in administrative health data are routinely used to monitor trends in disease prevalence and incidence. The International Classification of Diseases (ICD), which is used to record these diagnoses, have been updated multiple times to reflect advances in health and medical research. Our objective was to examine the impact of transitions between ICD versions on the prevalence of chronic health conditions estimated from administrative health data.

Methods

Study data (i.e., physician billing claims, hospital records) were from the province of Manitoba, Canada, which has a universal healthcare system. ICDA-8 (with adaptations), ICD-9-CM (clinical modification), and ICD-10-CA (Canadian adaptation; hospital records only) codes are captured in the data. Annual study cohorts included all individuals 18 + years of age for 45 years from 1974 to 2018. Negative binomial regression was used to estimate annual age- and sex-adjusted prevalence and model parameters (i.e., slopes and intercepts) for 16 chronic health conditions. Statistical control charts were used to assess the impact of changes in ICD version on model parameter estimates. Hotelling’s T² statistic was used to combine the parameter estimates and provide an out-of-control signal when its value was above a pre-specified control limit.

Results

The annual cohort sizes ranged from 360,341 to 824,816. Hypertension and skin cancer were among the most and least diagnosed health conditions, respectively; their prevalence per 1,000 population increased from 40.5 to 223.6 and from 0.3 to 2.1, respectively, within the study period. The average annual rate of change in prevalence ranged from -1.6% (95% confidence interval [CI]: -1.8, -1.4) for acute myocardial infarction to 14.6% (95% CI: 13.9, 15.2) for hypertension. The control chart indicated out-of-control observations when transitioning from ICDA-8 to ICD-9-CM for 75% of the investigated chronic health conditions but no out-of-control observations when transitioning from ICD-9-CM to ICD-10-CA.

Conclusions

The prevalence of most of the investigated chronic health conditions changed significantly in the transition from ICDA-8 to ICD-9-CM. These results point to the importance of considering changes in ICD coding as a factor that may influence the interpretation of trend estimates for chronic health conditions derived from administrative health data.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12889-022-13118-8.

Hospitalization for ambulatory care sensitive conditions: What conditions make inter-country comparisons possible?

Hospitalizations for ambulatory care sensitive conditions have been extensively used in health services research to assess access, quality and performance of primary health care. Inter-country comparisons can assist policy-makers in pursuing better health outcomes by contrasting policy design, implementation and evaluation. The objective of this study is to identify the conceptual, methodological, contextual and policy dimensions and factors that need to be accounted for when comparing these types of hospitalizations across countries. A conceptual framework for inter-country comparisons was drawn based on a review of 18 studies with inter-country comparison of ambulatory care sensitive conditions hospitalizations. The dimensions include methodological choices; population's demographic, epidemiologic and socio-economic profiles and features of the health services and system. Main factors include access and quality of primary health care, availability of health workforce and health facilities, health interventions and inequalities. The proposed framework can assist in designing studies and interpreting findings of inter-country comparisons of ambulatory care sensitive conditions hospitalizations, accelerating learning and progress towards universal health coverage.

Validation of an Operational Definition to Identify Distal Radius Fractures in a National Health Insurance Database

PURPOSE

To develop and validate identification criteria for distal radius fractures (DRFs) and their treatment using a national health insurance database.

METHODS

Patients who had at least 1 wrist radiograph taken in 2018 were recruited from a single academic referral hospital. After excluding patients who lacked immobilization code for wrist pathology, we collected data on the overall population. Because some patients might have undergone wrist radiography at another institution or had DRFs without an immobilization code, we additionally included patients who had a DRF diagnosis code at our institution. Reviews of medical records and wrist radiographs were considered for the diagnosis of DRF. We evaluated the sensitivity, specificity, and positive predictive value (PPV) of 3 operational definitions of fractures that were based on a single primary or secondary diagnosis code; all diagnosis codes, including primary and secondary codes; and all diagnosis and procedure codes.

RESULTS

Among 768 patients included in the study, true DRFs were confirmed in 305. The sensitivity, specificity, and PPV for definition 1 were 91.5% (95% CI, 88.3%-94.6%), 97.5% (95% CI, 95.9%-99.1%), and 96.9% (95% CI, 94.9%-98.9%), respectively. Although the sensitivity of definition 2 was higher (92.1%; 95% CI, 89.1%-95.2%), its specificity and PPV were lower (96.4% [95% CI, 94.4%-98.3%] and 95.6% [95% CI, 93.2%-97.9%], respectively). The sensitivity of definition 3 was the lowest (88.2%; 95% CI, 84.6%-91.8%), but its specificity and PPV were the highest among the 3 definitions (98.6% [95% CI, 97.4%-98.8%] and 98.2% [95% CI, 96.6%-99.8%], respectively).

CONCLUSIONS

Patients with DRFs can be identified from claims databases with high accuracy using an operational definition based on DRF diagnosis and procedure codes, including codes for surgical and nonoperative methods.

CLINICAL RELEVANCE

Verified operational definitions will increase the consistency of results in future national health insurance database studies related to DRFs.

Maternal and perinatal outcomes of pregnant women with SARS-CoV-2 infection at the time of birth in England: national cohort study

BACKGROUND

Some studies have suggested that women with SARS-CoV-2 infection during pregnancy are at increased risk of adverse pregnancy and neonatal outcomes, but these associations are still not clear.

OBJECTIVE

This study aimed to determine the association between SARS-CoV-2 infection at the time of birth and maternal and perinatal outcomes.

STUDY DESIGN

This is a population-based cohort study in England. The inclusion criteria were women with a recorded singleton birth between May 29, 2020, and January 31, 2021, in a national database of hospital admissions. Maternal and perinatal outcomes were compared between pregnant women with a laboratory-confirmed SARS-CoV-2 infection recorded in the birth episode and those without. Study outcomes were fetal death at or beyond 24 weeks' gestation (stillbirth), preterm birth (<37 weeks' gestation), small for gestational age infant (small for gestational age; birthweight at the <tenth centile), preeclampsia or eclampsia, induction of labor, mode of birth, specialist neonatal care, composite neonatal adverse outcome indicator, maternal and neonatal length of hospital stay after birth (3 days or more), and 28-day neonatal and 42-day maternal hospital readmission. Adjusted odds ratios and their 95% confidence interval for the association between SARS-CoV-2 infection status and outcomes were calculated using logistic regression, adjusting for maternal age, ethnicity, parity, preexisting diabetes mellitus, preexisting hypertension, and socioeconomic deprivation measured using the Index of Multiple Deprivation 2019. Models were fitted with robust standard errors to account for hospital-level clustering. The analysis of the neonatal outcomes was repeated for those born at term (≥37 weeks' gestation) because preterm birth has been reported to be more common in pregnant women with SARS-CoV-2 infection.

RESULTS

The analysis included 342,080 women, of whom 3527 had laboratory-confirmed SARS-CoV-2 infection. Laboratory-confirmed SARS-CoV-2 infection was more common in women who were younger, of non-White ethnicity, primiparous, or residing in the most deprived areas or had comorbidities. Fetal death (adjusted odds ratio, 2.21; 95% confidence interval, 1.58-3.11; P<.001) and preterm birth (adjusted odds ratio, 2.17; 95% confidence interval, 1.96-2.42; P<.001) occurred more frequently in women with SARS-CoV-2 infection than those without. The risk of preeclampsia or eclampsia (adjusted odds ratio, 1.55; 95% confidence interval, 1.29-1.85; P<.001), birth by emergency cesarean delivery (adjusted odds ratio, 1.63; 95% confidence interval, 1.51-1.76; P<.001), and prolonged admission after birth (adjusted odds ratio, 1.57; 95% confidence interval, 1.44-1.72; P<.001) were significantly higher for women with SARS-CoV-2 infection than those without. There were no significant differences (P>.05) in the rate of other maternal outcomes. The risk of neonatal adverse outcome (adjusted odds ratio, 1.45; 95% confidence interval, 1.27-1.66; P<.001), need for specialist neonatal care (adjusted odds ratio, 1.24; 95% confidence interval, 1.02-1.51; P=.03), and prolonged neonatal admission after birth (adjusted odds ratio, 1.61; 95% confidence interval, 1.49-1.75; P<.001) were all significantly higher for infants with mothers with laboratory-confirmed SARS-CoV-2 infection. When the analysis was restricted to pregnancies delivered at term (≥37 weeks), there were no significant differences in neonatal adverse outcome (P=.78), need for specialist neonatal care after birth (P=.22), or neonatal readmission within 4 weeks of birth (P=.05). Neonates born at term to mothers with laboratory-confirmed SARS-CoV-2 infection were more likely to have prolonged admission after birth (21.1% compared with 14.6%; adjusted odds ratio, 1.61; 95% confidence interval, 1.49-1.75; P<.001).

CONCLUSION

SARS-CoV-2 infection at the time of birth is associated with higher rates of fetal death, preterm birth, preeclampsia, and emergency cesarean delivery. There were no additional adverse neonatal outcomes, other than those related to preterm delivery. Pregnant women should be counseled regarding risks of SARS-CoV-2 infection and should be considered a priority for vaccination.

A study to design minimum data set of COVID-19 registry system

BACKGROUND

From the beginning of the COVID-19 pandemic, the development of infrastructures to record, collect and report COVID-19‏ ‏data has become a fundamental necessity in the world. The disease registry system can help build an infrastructure to collect data systematically. The study aimed to design a minimum data set for the COVID-19 registry system.

METHODS

A qualitative study to design an MDS for the COVID-19 registry system was performed in five phases at Ahvaz University of Medical Sciences in Khuzestan Province in southwestern Iran, 2020-2021. In the first phase, assessing the information requirements was performed for the COVID-19 registry system. Data elements were identified in the second phase. In the third phase, the MDS was selected, and in the four phases, the COVID-19 registry system was implemented as a pilot study to test the MDS. Finally, based on the experiences gained from the COVID-19 registry system implementation, the MDS were evaluated, and corrections were made.

RESULTS

MDS of the COVID-19 registry system contains eight top groups including administrative (34 data elements), disease exposure (61 data elements), medical history and physical examination (138 data elements), findings of clinical diagnostic tests (101 data elements), disease progress and outcome of treatment (55 data elements), medical diagnosis and cause of death (12 data elements), follow-up (14 data elements), and COVID-19 vaccination (19 data elements) data, respectively.

CONCLUSION

Creating a standard and comprehensive MDS can help to design any national data dictionary for COVID-19 and improve the quality of COVID-19 data.

Combining structured and unstructured data in EMRs to create clinically-defined EMR-derived cohorts

Background

There have been few studies describing how production EMR systems can be systematically queried to identify clinically-defined populations and limited studies utilising free-text in this process. The aim of this study is to provide a generalisable methodology for constructing clinically-defined EMR-derived patient cohorts using structured and unstructured data in EMRs.

Methods

Patients with possible acute coronary syndrome (ACS) were used as an exemplar. Cardiologists defined clinical criteria for patients presenting with possible ACS. These were mapped to data tables within the production EMR system creating seven inclusion criteria comprised of structured data fields (orders and investigations, procedures, scanned electrocardiogram (ECG) images, and diagnostic codes) and unstructured clinical documentation. Data were extracted from two local health districts (LHD) in Sydney, Australia. Outcome measures included examination of the relative contribution of individual inclusion criteria to the identification of eligible encounters, comparisons between inclusion criterion and evaluation of consistency of data extracts across years and LHDs.

Results

Among 802,742 encounters in a 5 year dataset (1/1/13–30/12/17), the presence of an ECG image (54.8% of encounters) and symptoms and keywords in clinical documentation (41.4–64.0%) were used most often to identify presentations of possible ACS. Orders and investigations (27.3%) and procedures (1.4%), were less often present for identified presentations. Relevant ICD-10/SNOMED CT codes were present for 3.7% of identified encounters. Similar trends were seen when the two LHDs were examined separately, and across years.

Conclusions

Clinically-defined EMR-derived cohorts combining structured and unstructured data during cohort identification is a necessary prerequisite for critical validation work required for development of real-time clinical decision support and learning health systems.

The accuracy of administrative health data for identifying patients with rheumatoid arthritis: a retrospective validation study using medical records in Western Australia

The use of administrative health datasets is increasingly important for research on disease trends and outcome. The Western Australian (WA) Rheumatic Disease Epidemiological Registry contains longitudinal health data for over 10,000 patients with rheumatoid arthritis (RA). Accurate coding for RA is essential to the validity of this dataset. Investigate the diagnostic accuracy of International Classification of Diseases (ICD)-based discharge codes for RA at WA's largest tertiary hospital. Medical records for a sample of randomly selected patients with ICD-10 codes (M05.00-M06.99) in the hospital discharge database between 2008 and 2020 were retrospectively reviewed. Rheumatologist-reported diagnoses and ACR/EULAR classification criteria were used as reference standards to determine accuracy measures. Medical chart review was completed for 87 patients (mean (± SD) age 64.7 ± 17.2 years), 67.8% female). A total of 80 (91.9%) patients had specialist confirmed RA diagnosis, while seven patients (8%) had alternate clinical diagnoses. Among 87 patients, 69 patients (79.3%) were fulfilled ACR/EULAR classification criteria. The agreement between the reference standards was moderate (Kappa 0.41). Based on rheumatologist-reported diagnoses and ACR/EULAR classification criteria, primary diagnostic codes for RA alone had a sensitivity of (90% vs 89.8%), and PPV (90.9% vs 63.6%), respectively. A combination of a diagnostic RA code with biologic infusion codes in two or more codes increased the PPV to 97.9%. Hospital discharge diagnostic codes in WA identify RA patients with a high degree of accuracy. Combining a primary diagnostic code for RA with biological infusion codes can further increase the PPV.

Validity of Algorithms for Identification of Individuals Suffering from Chronic Noncancer Pain in Administrative Databases: A Systematic Review

BACKGROUND

Secondary analysis of health administrative databases is indispensable to enriching our understanding of health trajectories, health care utilization, and real-world risks and benefits of drugs among large populations.

OBJECTIVES

This systematic review aimed at assessing evidence about the validity of algorithms for the identification of individuals suffering from nonarthritic chronic noncancer pain (CNCP) in administrative databases.

METHODS

Studies reporting measures of diagnostic accuracy of such algorithms and published in English or French were searched in the Medline, Embase, CINAHL, AgeLine, PsycINFO, and Abstracts in Social Gerontology electronic databases without any dates of coverage restrictions up to March 1, 2018. Reference lists of included studies were also screened for additional publications.

RESULTS

Only six studies focused on commonly studied CNCP conditions and were included in the review. Some algorithms showed a ≥60% combination of sensitivity and specificity values (back pain disorders in general, fibromyalgia, low back pain, migraine, neck/back problems studied together). Only algorithms designed to identify fibromyalgia cases reached a ≥80% combination (without replication of findings in other studies/databases).

CONCLUSIONS

In summary, the present investigation informs us about the limited amount of literature available to guide and support the use of administrative databases as valid sources of data for research on CNCP. Considering the added value of such data sources, the important research gaps identified in this innovative review provide important directions for future research. The review protocol was registered with PROSPERO (CRD42018086402).

Validation study: evaluation of the metrological quality of French hospital data for perinatal algorithms

OBJECTIVE

The aim of our validation study was to assess the metrological quality of hospital data for perinatal algorithms on a national level.

DESIGN

Validation study.

SETTING

This was a multicentre study of the French medicoadministrative database on perinatal indicators.

PARTICIPANTS

In each hospital, we selected 150 discharge abstracts for delivery (after 22 weeks of gestation), in 2014, and their corresponding medical records. Overall, 22 hospitals were included.

INTERVENTIONS

A single investigator performed blind data collection from medical records in order to compare data from discharge abstracts with data from medical records. Finally, 3246 discharge abstracts were studied.

PRIMARY AND SECONDARY OUTCOME MEASURES

Seventy items, including maternal and delivery characteristics and maternal morbidity, were collected for each delivery stay.

RESULTS

The concordance rate of maternal age at delivery was 94.8% (95% CI 93.8 to 95.4). Combining the two forms of pre-existing diabetes, the algorithm presented a PPV of 65.9% and a sensitivity of 75.7%. The concordance rate of gestational age at delivery was 91.8% (90.9 to 92.7). Regarding gestational diabetes, the PPV was 80.8% (79.4 to 82.2) and the sensitivity was 79.5% (78.1 to 80.9). Regardless of the algorithm explored, the PPV for vaginal delivery was over 99%. For the diagnosis codes corresponding to immediate postpartum haemorrhage, the PPV was 77.7% (76.3 to 79.1) and the sensitivity was 75.5% (74.0 to 77.0). The algorithm for stillbirth presented a PPV of 89.4% (88.3 to 90.5) and a sensitivity of 95.4% (94.7 to 96.1).

CONCLUSIONS

This first national validation study of many perinatal algorithms suggests that the French national hospital database is an appropriate data source for epidemiological studies, except for some indicators which presented low PPV and/or sensitivity.

Empowering knowledge generation through international data network: the IMeCCHI-DATANETWORK

Introduction

The International Methodology Consortium for Coded Health Information (IMeCCHI) is a collaboration of health services researchers who promote methodological advances in coded health information. The IMeCCHI-DATANETWORK initiative focuses on developing a multi-purpose distributed data infrastructure and common data model (CDM) to enable cross-border data sharing and international comparisons.

Methods

IMeCCHI consortium partners from six different countries – Canada, Denmark, Italy, New Zealand, South Korea, and Switzerland – used a questionnaire to describe their original databases which differ in size, structure, content and coding systems. To standardize these data, they agreed on a CDM and mapped their population-based databases to meet the CDM specifications. At the end of this process, local data had a more homogenous content and structure, which made them syntactically and semantically interoperable. Data transformation was performed using a common data management software called TheMatrix.

Results

The CDM encompasses four tables of structured data (person characteristics, hospitalizations, outpatient prescription medication and death), linked at the individual level through a person identifier. It can be used to answer research questions across countries using locally converted databases, which facilitates study replication in a distributed fashion. As a proof-of-concept study, an initial research question was addressed using an agreed protocol. Local data were transformed in csv files in the CDM structure and TheMatrix was tested to transform the standardized data from each partner into local analytical datasets. This allowed results to be shared between countries, whilst maintaining local control over each region’s data.

Conclusion

The IMeCCHI-DATANETWORK, a model of a distributed data network, demonstrated that it is feasible to analyze international data using standardized analytical methods that enable independent analyses by regions, without relocating datasets thereby protecting local confidentiality obligations. The distributed data infrastructure can produce results that can be generalized to several countries, while facilitating cross-border data sharing and international comparisons.

Keywords

Common data model, international comparison, cross-border data sharing, interoperability, observational data

Data accuracy in the Ontario birth Registry: a chart re-abstraction study

Background

Ontario’s birth Registry (BORN) was established in 2009 to collect, interpret, and share critical data about pregnancy, birth and the early childhood period to facilitate and improve the provision of healthcare. Since the use of routinely-collected health data has been prioritized internationally by governments and funding agencies to improve patient care, support health system planning, and facilitate epidemiological surveillance and research, high quality data is essential. The purpose of this study was to verify the accuracy of a selection of data elements that are entered in the Registry.

Methods

Data quality was assessed by comparing data re-abstracted from patient records to data entered into the Ontario birth Registry. A purposive sample of 10 hospitals representative of hospitals in Ontario based on level of care, birth volume and geography was selected and a random sample of 100 linked mother and newborn charts were audited for each site. Data for 29 data elements were compared to the corresponding data entered in the Ontario birth Registry using percent agreement, kappa statistics for categorical data elements and intra-class correlation coefficients (ICCs) for continuous data elements.

Results

Agreement ranged from 56.9 to 99.8%, but 76% of the data elements (22 of 29) had greater than 90% agreement. There was almost perfect (kappa 0.81–0.99) or substantial (kappa 0.61–0.80) agreement for 12 of the categorical elements. Six elements showed fair-to-moderate agreement (kappa <0.60). We found moderate-to-excellent agreement for four continuous data elements (ICC >0.50).

Conclusion

Overall, the data elements we evaluated in the birth Registry were found to have good agreement with data from the patients’ charts. Data elements that showed moderate kappa or low ICC require further investigation.

The impact of diabetes on multiple avoidable admissions: a cross-sectional study

BACKGROUND

Multiple admissions for ambulatory care sensitive conditions (ACSC) are responsible for an important proportion of health care expenditures. Diabetes is one of the conditions consensually classified as an ACSC being considered a major public health concern. The aim of this study was to analyse the impact of diabetes on the occurrence of multiple admissions for ACSC.

METHODS

We analysed inpatient data of all public Portuguese NHS hospitals from 2013 to 2015 on multiple admissions for ACSC among adults aged 18 or older. Multiple ACSC users were identified if they had two or more admissions for any ACSC during the period of analysis. Two logistic regression models were computed. A baseline model where a logistic regression was performed to assess the association between multiple admissions and the presence of diabetes, adjusting for age and sex. A full model to test if diabetes had no constant association with multiple admissions by any ACSC across age groups.

RESULTS

Among 301,334 ACSC admissions, 144,209 (47.9%) were classified as multiple admissions and from those, 59,436 had diabetes diagnosis, which corresponded to 23,692 patients. Patients with diabetes were 1.49 times (p < 0,001) more likely to be admitted multiple times for any ACSC than patients without diabetes. Younger adults with diabetes (18-39 years old) were more likely to become multiple users.

CONCLUSION

Diabetes increases the risk of multiple admissions for ACSC, especially in younger adults. Diabetes presence is associated with a higher resource utilization, which highlights the need for the implementation of adequate management of chronic diseases policies.

Comparative research aspects on hospitalizations for ambulatory care sensitive conditions: the case of Brazil and Portugal

Hospitalizations for ambulatory care sensitive conditions have been used to measure access, quality and performance of the primary health care delivery system, as timely and adequate care could potentially avoid the need of hospitalization. Comparative research provides the opportunity for cross-country learning process. Brazil and Portugal have reformed their primary health care services in the last years, with similar organizational characteristics. We used hospitalization data of Brazil and Portugal for the year 2015 to compare hospitalizations for ambulatory care sensitive conditions between the two countries, and discussed conceptual and methodological aspects to be taken into consideration in the comparative approach. Brazil and Portugal presented similarities in causes and standardized rates of hospitalizations for ambulatory care sensitive conditions. There was great sensitivity on rates according to the methodology employed to define conditions. Hospitalizations for ambulatory care sensitive conditions are important sources of pressure for both Brazil and Portugal, and there are conceptual and methodological aspects that are critical to render the country-comparison approach useful.

Challenges of Using ICD-9-CM and ICD-10-CM Codes for Soft-Tissue Sarcoma in Databases for Health Services Research

OBJECTIVES

Soft-tissue sarcoma (STS) is a heterogeneous group of rare solid tumors that arise from various soft tissues in the body, such as muscle, fat, nerves, and blood vessels. Current International Classification of Diseases (ICD) coding systems include a set of nonspecific codes for malignancies of connective and soft tissue (ICD-9-CM code 171 and ICD-10-CM code C49). The goal of this study was to evaluate the use of these codes for health services research involving patients with a diagnosis of this rare malignancy.

METHODS

Two databases were utilized to explore ICD coding for STS: claims data from Truven MarketScan and electronic medical records (EMRs) from Flatiron Health. Eligible patients from claims data were those with at least two ICD-9-CM codes of 171.x on two different days between July 1, 2004, and March 30, 2014. The treatment patterns of these cases were evaluated for consistency with known therapeutic approaches for STS. Eligible patients from the Flatiron EMR system were those who received olaratumab (a drug indicated only for use in patients diagnosed with STS) after its US Food and Drug Administration approval in October 2016 through the end of the data set (November 2017). ICD-10-CM codes were evaluated for this known STS cohort.

RESULTS

In claims data, 4,159 patients were eligible for inclusion. Although national treatment guidelines include only a limited number of drugs used to treat STS, 98 unique anticancer drugs were identified as being used to treat patients in a claims data cohort. Only 7.7 percent of patients had claims for doxorubicin-based therapy and 3.8 percent had claims for ifosfamide-based therapy as initial treatment for STS, despite these being a standard of care. In the EMR data, 350 patients were eligible; only 170 patients (48.6 percent) had any evidence in the database of a connective or soft-tissue ICD-10-CM malignancy code within 60 days before or after initiation of olaratumab.

CONCLUSIONS

ICD coding for STS using the "Malignant neoplasm of connective and soft tissue" code is not reliable as a method to identify patients diagnosed with STS. Although codes reflecting the primary site of disease may have clinical relevance, lack of consistency in ICD coding for the diagnosis and treatment of this disease is a limiting factor in the ability to conduct real-world observational research of this rare disease. In the absence of consistent use of this code, an algorithm needs to be developed and validated to accurately identify patients with STS in these databases.

Key considerations when using health insurance claims data in advanced data analyses: an experience report

Health claims have become a popular source of data for healthcare analytics, with numerous applications ranging from disease burden estimation and policy evaluation to drug event detection and advanced predictive analytics. Independent of the application, a researcher utilising claims information will likely encounter challenges in using the data, which include dealing with several coding systems and coding irregularities. We highlight some of these challenges and approaches for successful analysis that may reduce implementation time and help in avoiding common pitfalls. We describe the experiences of a group of academic researchers in using an extensive seven-year repository of US medical and pharmaceutical claims data in a research study, and provide an overview of the challenges encountered with handling claims records for data analysis while sharing suggestions on how to address these challenges. To illustrate our experiences, we use the example of defining episodes of care for a bundled payment reimbursement system in the US context.

Validation of algorithms to identify knee osteoarthritis patients in the claims database

BACKGROUND

To identify knee osteoarthritis (OA) patients among OA patients in the claims database.

METHODS

All patients with OA diagnostic codes for any sites (M15 to M19) in 2014 were recruited from a single academic referral hospital. After excluding patients who had inflammatory arthritis or were less than 50 years of age, we identified data for the overall OA population. Radiographic knee OA of Kellgren and Lawrence grades ≥2 is considered the gold standard for knee OA, and we evaluated the sensitivity, specificity, and positive predictive value (PPV) of three operational definitions using the diagnostic codes in the claims database. The operational definitions were: (1) gonarthrosis (M17); (2) any site of OA (M15 to M19) with knee X-ray; and (3) (1) or (2).

RESULTS

A total of 7959 OA patients were included in this study of whom 74.5% were women. The PPV of gonarthrosis (M17) was 0.67 (95% CI 0.65-0.69), and sensitivity was 0.44 (95% CI 0.42-0.46). The PPV and sensitivity of any OA site (M15 to M19) with knee X-ray were 0.65 (95% CI 0.62-0.67), and 0.37 (95% CI 0.35-0.39), respectively. When knee OA was defined as satisfying either of the two above definitions, PPV was 0.63 (95% CI 0.62-0.65) and sensitivity 0.55 (95% CI 0.53-0.57).

CONCLUSIONS

Knee OA patients can be identified in a claims database using the algorithms of gonarthrosis (M17) or any site of OA (M15 to M19) with a performed knee X-ray.

Accuracy of Algorithms to Identify Pulmonary Arterial Hypertension in Administrative Data: A Systematic Review

BACKGROUND

The diagnosis of pulmonary arterial hypertension (PAH) is challenging, and there is significant overlap with the more heterogenous diagnosis of pulmonary hypertension (PH). Clinical and research efforts that rely on administrative data are limited by current coding systems that do not adequately reflect the clinical classification scheme. The aim of this systematic review is to investigate current algorithms to detect PAH using administrative data and to appraise the diagnostic accuracy of these algorithms against a reference standard.

METHODS

We conducted comprehensive searches of Medline, Embase, and Web of Science from their inception. We included English-language articles that applied an algorithm to an administrative or electronic health record database to identify PAH in adults.

RESULTS

Of 2,669 unique citations identified, 32 studies met all inclusion criteria. Only four of these studies validated their algorithm against a reference standard. Algorithms varied widely, ranging from single International Classification of Diseases (ICD) codes to combinations of visit, procedure, and pharmacy codes. ICD codes alone performed poorly, with positive predictive values ranging from 3.3% to 66.7%. The addition of PAH-specific therapy and diagnostic procedures to the algorithm improved the diagnostic accuracy.

CONCLUSIONS

Algorithms to identify PAH in administrative databases vary widely, and few are validated. The sole use of ICD codes performs poorly, potentially leading to biased results. ICD codes should be revised to better discriminate between PH groups, and universally accepted algorithms need to be developed and validated to capture PAH in administrative data, better informing research and clinical efforts.

A Meta-Data Manifesto: The Need for Global Health Meta-Data

Administrative health data recorded for individual health episodes (such as births, deaths, physician visits, and hospital stays) are being widely used to study policy-relevant scientific questions about population health, health services, and quality of care. An increasing number of international health comparisons are undertaken with these data. An essential pre-requisite to such international comparative work is a detailed characterization of existing international health data resources, so that they can be more readily used for comparisons across counties. A major challenge to such international comparative work is the variability across countries in the extent, content, and validity of existing administrative data holdings. Recognizing this, we have undertaken an international proof of concept pilot compiling detailed data about data – i.e., a “meta-data catalogue” – for existing international administrative health data holdings. We describe the methodological process for collecting these meta-data, along with some general descriptive results for selected countries included in the pilot.

Sensitivity and specificity of breast cancer ICD-9-CM codes in three Italian administrative healthcare databases: a diagnostic accuracy study

OBJECTIVES

To assess the accuracy of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes in identifying patients diagnosed with incident carcinoma in situ and invasive breast cancer in three Italian administrative databases.

DESIGN

A diagnostic accuracy study comparing ICD-9-CM codes for carcinoma in situ (233.0) and for invasive breast cancer (174.x) with medical chart (as a reference standard). Case definition: (1) presence of a primary nodular lesion in the breast and (2) cytological or histological documentation of cancer from a primary or metastatic site.

SETTING

Administrative databases from Umbria Region, Azienda Sanitaria Locale (ASL) Napoli 3 Sud (NA) and Friuli VeneziaGiulia (FVG) Region.

PARTICIPANTS

Women with breast carcinoma in situ (n=246) or invasive breast cancer (n=384) diagnosed (in primary position) between 2012 and 2014.

OUTCOME MEASURES

Sensitivity and specificity for codes 233.0 and 174.x.

RESULTS

For invasive breast cancer the sensitivities were 98% (95% CI 93% to 99%) for Umbria, 96% (95% CI 91% to 99%) for NA and 100% (95% CI 97% to 100%) for FVG. Specificities were 90% (95% CI 82% to 95%) for Umbria, 91% (95% CI 83% to 96%) for NA and 91% (95% CI 84% to 96%) for FVG.For carcinoma in situ the sensitivities were 100% (95% CI 93% to 100%) for Umbria, 100% (95% CI 95% to 100%) for NA and 100% (95% CI 96% to 100%) for FVG. Specificities were 98% (95% CI 93% to 100%) for Umbria, 86% (95% CI 78% to 92%) for NA and 90% (95% CI 82% to 95%) for FVG.

CONCLUSIONS

Administrative healthcare databases from Umbria, NA and FVG are accurate in identifying hospitalised news cases of carcinoma of the breast. The proposed case definition is a powerful tool to perform research on large populations of newly diagnosed patients with breast cancer.

Accuracy of colorectal cancer ICD-9-CM codes in Italian administrative healthcare databases: a cross-sectional diagnostic study

Objectives To assess the accuracy of International Classification of Diseases, Ninth Revision - Clinical Modification (ICD-9-CM) codes in identifying subjects with colorectal cancer.

DESIGN

A diagnostic accuracy study comparing ICD-9-CM codes (index test) for colorectal cancers with medical chart (as a reference standard). Case ascertainment based on neoplastic lesion(s) within the colon/rectum and histological documentation from a primary or metastatic site positive for colorectal cancer.

SETTING

Administrative databases from the Umbria region, Azienda Sanitaria Locale (ASL) Napoli 3 Sud (NA) region and Friuli Venezia Giulia (FVG) region.

PARTICIPANTS

We randomly selected 130 incident patients from each hospital discharge database, admitted between 2012 and 2014, having colorectal cancer ICD-9 codes located in primary position, and 94 non-cases, that is, patients having a diagnosis of cancer (ICD-9 140-239) other than colorectal cancer in primary position.

OUTCOME MEASURES

Sensitivity, specificity and predictive values for 153.x code (colon cancer) and for 154.x code (rectal cancer).

RESULTS

The positive predictive value (PPV) for colon cancer diagnoses was 80% for Umbria (95% CI 73% to 87%), 81% for NA (95% CI 73% to 88%) and 80% for FVG (95% CI 72% to 87%).The sensitivity ranged from 98% to 99%, while the specificity ranged from 78% to 80% in the three units.For rectal cancer, the PPV was 84% for Umbria (95% CI 77% to 90%), 80% for NA (95% CI 72% to 87%) and 81% for FVG (95% CI 73% to 87%). The sensitivities ranged from 98% to 100%, while the specificity estimates from 79% to 82%.

CONCLUSIONS

Administrative databases in Italy can be a valuable tool for cancer surveillance as well as monitoring geographical and temporal variation of cancer practice.

Systematic review of atopic dermatitis disease definition in studies using routinely collected health data

BACKGROUND

Routinely collected electronic health data obtained for administrative and clinical purposes are increasingly used to study atopic dermatitis (AD). Methods for identifying AD patients in routinely collected electronic health data differ, and it is unknown how this might affect study results.

OBJECTIVES

To evaluate how patients with AD have been identified in studies using routinely collected electronic health data, to determine whether these methods were validated and to estimate how the method for identifying patients with AD affected variability in prevalence estimates.

METHODS

We systematically searched PubMed, Embase and Web of Science for studies using routinely collected electronic health data that reported on AD as a primary outcome. Studies of localized AD and other types of dermatitis were excluded. The protocol for this review was registered in PROSPERO (CRD42016037968).

RESULTS

In total, 59 studies met eligibility criteria. Medical diagnosis codes for inclusion and exclusion, number of occasions of a code, type of provider associated with a code and prescription data were used to identify patients with AD. Only two studies described validation of their methods and no study reported on disease severity. Prevalence estimates ranged from 0·18% to 38·33% (median 4·91%) and up to threefold variation in prevalence was introduced by differences in the method for identifying patients with AD.

CONCLUSIONS

This systematic review highlights the need for clear reporting of methods for identifying patients with AD in routinely collected electronic health data to allow for meaningful interpretation and comparison of results.

ICD-10 procedure codes produce transition challenges

The transition of procedure coding from ICD-9-CM-Vol-3 to ICD-10-PCS has generated problems for the medical community at large resulting from the lack of clarity required to integrate two non-congruent coding systems. We hypothesized that quantifying these issues with network topology analyses offers a better understanding of the issues, and therefore we developed solutions (online tools) to empower hospital administrators and researchers to address these challenges. Five topologies were identified: "identity"(I), "class-to-subclass"(C2S), "subclass-toclass"(S2C), "convoluted(C)", and "no mapping"(NM). The procedure codes in the 2010 Illinois Medicaid dataset (3,290 patients, 116 institutions) were categorized as C=55%, C2S=40%, I=3%, NM=2%, and S2C=1%. Majority of the problematic and ambiguous mappings (convoluted) pertained to operations in ophthalmology cardiology, urology, gyneco-obstetrics, and dermatology. Finally, the algorithms were expanded into a user-friendly tool to identify problematic topologies and specify lists of procedural codes utilized by medical professionals and researchers for mitigating error-prone translations, simplifying research, and improving quality.http://www.lussiergroup.org/transition-to-ICD10PCS.

Accuracy of lung cancer ICD-9-CM codes in Umbria, Napoli 3 Sud and Friuli Venezia Giulia administrative healthcare databases: a diagnostic accuracy study

OBJECTIVES

To assess the accuracy of International Classification of Diseases 9th Revision-Clinical Modification (ICD-9-CM) codes in identifying subjects with lung cancer.

DESIGN

A cross-sectional diagnostic accuracy study comparing ICD-9-CM 162.x code (index test) in primary position with medical chart (reference standard). Case ascertainment was based on the presence of a primary nodular lesion in the lung and cytological or histological documentation of cancer from a primary or metastatic site.

SETTING

Three operative units: administrative databases from Umbria Region (890 000 residents), ASL Napoli 3 Sud (NA) (1 170 000 residents) and Friuli Venezia Giulia (FVG) Region (1 227 000 residents).

PARTICIPANTS

Incident subjects with lung cancer (n=386) diagnosed in primary position between 2012 and 2014 and a population of non-cases (n=280).

OUTCOME MEASURES

Sensitivity, specificity and positive predictive value (PPV) for 162.x code.

RESULTS

130 cases and 94 non-cases were randomly selected from each database and the corresponding medical charts were reviewed. Most of the diagnoses for lung cancer were performed in medical departments.True positive rates were high for all the three units. Sensitivity was 99% (95% CI 95% to 100%) for Umbria, 97% (95% CI 91% to 100%) for NA, and 99% (95% CI 95% to 100%) for FVG. The false positive rates were 24%, 37% and 23% for Umbria, NA and FVG, respectively. PPVs were 79% (73% to 83%)%) for Umbria, 58% (53% to 63%)%) for NA and 79% (73% to 84%)%) for FVG.

CONCLUSIONS

Case ascertainment for lung cancer based on imaging or endoscopy associated with histological examination yielded an excellent sensitivity in all the three administrative databases. PPV was moderate for Umbria and FVG but lower for NA.

Validating malignant melanoma ICD-9-CM codes in Umbria, ASL Napoli 3 Sud and Friuli Venezia Giulia administrative healthcare databases: a diagnostic accuracy study

OBJECTIVES

To assess the accuracy of International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes in identifying subjects with melanoma.

DESIGN

A diagnostic accuracy study comparing melanoma ICD-9-CM codes (index test) with medical chart (reference standard). Case ascertainment was based on neoplastic lesion of the skin and a histological diagnosis from a primary or metastatic site positive for melanoma.

SETTING

Administrative databases from Umbria Region, Azienda Sanitaria Locale (ASL) Napoli 3 Sud (NA) and Friuli Venezia Giulia (FVG) Region.

PARTICIPANTS

112, 130 and 130 cases (subjects with melanoma) were randomly selected from Umbria, NA and FVG, respectively; 94 non-cases (subjects without melanoma) were randomly selected from each unit.

OUTCOME MEASURES

Sensitivity and specificity for ICD-9-CM code 172.x located in primary position.

RESULTS

The most common melanoma subtype was malignant melanoma of skin of trunk, except scrotum (ICD-9-CM code: 172.5), followed by malignant melanoma of skin of lower limb, including hip (ICD-9-CM code: 172.7). The mean age of the patients ranged from 60 to 61 years. Most of the diagnoses were performed in surgical departments.The sensitivities were 100% (95% CI 96% to 100%) for Umbria, 99% (95% CI 94% to 100%) for NA and 98% (95% CI 93% to 100%) for FVG. The specificities were 88% (95% CI 80% to 93%) for Umbria, 77% (95% CI 69% to 85%) for NA and 79% (95% CI 71% to 86%) for FVG.

CONCLUSIONS

The case definition for melanoma based on clinical or instrumental diagnosis, confirmed by histological examination, showed excellent sensitivities and good specificities in the three operative units. Administrative databases from the three operative units can be used for epidemiological and outcome research of melanoma.

Exploration of association rule mining for coding consistency and completeness assessment in inpatient administrative health data

OBJECTIVE

Data quality assessment is a challenging facet for research using coded administrative health data. Current assessment approaches are time and resource intensive. We explored whether association rule mining (ARM) can be used to develop rules for assessing data quality.

MATERIALS AND METHODS

We extracted 2013 and 2014 records from the hospital discharge abstract database (DAD) for patients between the ages of 55 and 65 from five acute care hospitals in Alberta, Canada. The ARM was conducted using the 2013 DAD to extract rules with support ≥0.0019 and confidence ≥0.5 using the bootstrap technique, and tested in the 2014 DAD. The rules were compared against the method of coding frequency and assessed for their ability to detect error introduced by two kinds of data manipulation: random permutation and random deletion.

RESULTS

The association rules generally had clear clinical meanings. Comparing 2014 data to 2013 data (both original), there were 3 rules with a confidence difference >0.1, while coding frequency difference of codes in the right hand of rules was less than 0.004. After random permutation of 50% of codes in the 2014 data, average rule confidence dropped from 0.72 to 0.27 while coding frequency remained unchanged. Rule confidence decreased with the increase of coding deletion, as expected. Rule confidence was more sensitive to code deletion compared to coding frequency, with slope of change ranging from 1.7 to 184.9 with a median of 9.1.

CONCLUSION

The ARM is a promising technique to assess data quality. It offers a systematic way to derive coding association rules hidden in data, and potentially provides a sensitive and efficient method of assessing data quality compared to standard methods.

Appropriateness of ICD-coded Diagnostic Inpatient Hospital Discharge Data for Medical Practice Assessment

Objectives: We performed a systematic review to investigate the quality of diagnostic hospital discharge data (DHDD) in order to gain insight in the usefulness of these data for medical practice assessment. We investigated the methods used to evaluate data quality, factors that determine data quality and its consequences for medical practice assessment.

Methods: We selected studies in which both completeness (or sensitivity: SENS) and correctness (or positive predictive value: PPV) were measured. We used the random-effects model to calculate mean SENS and PPV and to explore the effect of a number of covariates.

Results: The 101 included studies were very heterogeneous. We distinguished six typical study designs. We found a mean SENS of 0.67 (95%CI: 0.62– 0.73) and PPV of 0.76 (95%CI: 0.73– 0.79); SENS was significantly lower for comorbidity and complication studies than for some single disease studies. PPV was significantly higher for Scandinavian countries than for other countries. Recoding compared to re-abstracting of the medical record as a gold standard gave a significantly lower PPV. Diagnostic data were considered appropriate by the authors of the studies for quality of care purposes when both SENS and PPV were at least 0.85. Only 13% of the studies fulfilled this criterion.

Conclusions: Variability in quality of care between settings can easily be overshadowed by variability in data quality. However, the use of DHDD by physicians to evaluate their own medical practice may be useful. But only if physicians are willing to critically interpret the meaning of the information for their medical practice assessment.

Navigating a ship with a broken compass: evaluating standard algorithms to measure patient safety

Objective

Agency for Healthcare Research and Quality (AHRQ) software applies standardized algorithms to hospital administrative data to identify patient safety indicators (PSIs). The objective of this study was to assess the validity of PSI flags and report reasons for invalid flagging.

Material and Methods

At a 6-hospital academic medical center, a retrospective analysis was conducted of all PSIs flagged in fiscal year 2014. A multidisciplinary PSI Quality Team reviewed each flagged PSI based on quarterly reports. The positive predictive value (PPV, the percent of clinically validated cases) was calculated for 12 PSI categories. The documentation for each reversed case was reviewed to determine the reasons for PSI reversal.

Results

Of 657 PSI flags, 185 were reversed. Seven PSI categories had a PPV below 75%. Four broad categories of reasons for reversal were AHRQ algorithm limitations (38%), coding misinterpretations (45%), present upon admission (10%), and documentation insufficiency (7%). AHRQ algorithm limitations included 2 subcategories: an "incident" was inherent to the procedure, or highly likely (eg, vascular tumor bleed), or an "incident" was nonsignificant, easily controlled, and/or no intervention was needed.

Discussion

These findings support previous research highlighting administrative data problems. Additionally, AHRQ algorithm limitations was an emergent category not considered in previous research. Herein we present potential solutions to address these issues.

Conclusions

If, despite poor validity, US policy continues to rely on PSIs for incentive and penalty programs, improvements are needed in the quality of administrative data and the standardized PSI algorithms. These solutions require national motivation, research attention, and dissemination support.

Record linkage study of the pathogen-specific burden of respiratory viruses in children

Background

Reliance on hospital discharge diagnosis codes alone will likely underestimate the burden of respiratory viruses.

Objectives

To describe the epidemiology of respiratory viruses more accurately, we used record linkage to examine data relating to all children hospitalized in Western Australia between 2000 and 2012.

Patients/Methods

We extracted hospital, infectious disease notification and laboratory data of a cohort of children born in Western Australia between 1996 and 2012. Laboratory records of respiratory specimens collected within 48 hours of admission were linked to hospitalization records. We calculated the frequency and rates of virus detection. To identify groups where under‐ascertainment for respiratory viruses was greatest, we used logistic regression to determine factors associated with failure to test.

Results and conclusions

Nine percentage of 484 992 admissions linked to a laboratory record for respiratory virus testing. While 62% (n = 26 893) of laboratory‐confirmed admissions received respiratory infection diagnosis codes, 38% (n = 16 734) had other diagnoses, notably viral infection of unspecified sites. Of those tested, incidence rates were highest for respiratory syncytial virus (247 per 100 000 child‐years) followed by parainfluenza (63 per 100 000 child‐years). Admissions among older children and those without a respiratory diagnosis were associated with failure to test for respiratory viruses. Linked data can significantly enhance diagnostic codes when estimating the true burden of disease. In contrast to current emphasis on influenza, respiratory syncytial virus and parainfluenza were the most common viral pathogens among hospitalized children. By characterizing those failing to be tested, we can begin to quantify the under‐ascertainment of respiratory viruses.

Development and Validation of an Algorithm to Accurately Identify Atopic Eczema Patients in Primary Care Electronic Health Records from the UK

Electronic health records hold great promise for clinical and epidemiologic research. Undertaking atopic eczema (AE) research using such data is challenging because of its episodic and heterogeneous nature. We sought to develop and validate a diagnostic algorithm that identifies AE cases based on codes used for electronic records used in the UK Health Improvement Network. We found that at least one of five diagnosis codes plus two treatment codes for any skin-directed therapy were likely to accurately identify patients with AE. To validate this algorithm, a questionnaire was sent to the physicians of 200 randomly selected children and adults. The primary outcome, positive predictive value for a physician-confirmed diagnosis of AE, was 86% (95% confidence interval = 80-91). Additional criteria increased the PPV up to 95% but would miss up to 89% of individuals with physician-confirmed AE. The first and last entered diagnosis codes for individuals showed good agreement with the physician-confirmed age at onset and last disease activity; the mean difference was 0.8 years (95% confidence interval = -0.3 to 1.9) and -1.3 years (95% confidence interval = -2.5 to -0.1), respectively. A combination of diagnostic and prescription codes can be used to reliably estimate the diagnosis and duration of AE from The Health Improvement Network primary care electronic health records in the UK.

Protocol for validating cardiovascular and cerebrovascular ICD-9-CM codes in healthcare administrative databases: the Umbria Data Value Project

INTRODUCTION

Administrative healthcare databases can provide a comprehensive assessment of the burden of diseases in terms of major outcomes, such as mortality, hospital readmissions and use of healthcare resources, thus providing answers to a wide spectrum of research questions. However, a crucial issue is the reliability of information gathered. Aim of this protocol is to validate International Classification of Diseases, 9th Revision-Clinical Modification (ICD-9-CM) codes for major cardiovascular diseases, including acute myocardial infarction (AMI), heart failure (HF), atrial fibrillation (AF) and stroke.

METHODS AND ANALYSIS

Data from the centralised administrative database of the entire Umbria Region (910 000 residents, located in Central Italy) will be considered. Patients with a first hospital discharge for AMI, HF, AF or stroke, between 2012 and 2014, will be identified in the administrative database using the following groups of ICD-9-CM codes located in primary position: (1) 410.x for AMI; (2) 427.31 for AF; (3) 428 for HF; (4) 433.x1, 434 (excluding 434.x0), 436 for ischaemic stroke, 430 and 431 for haemorrhagic stroke (subarachnoid haemorrhage and intracerebral haemorrhage). A random sample of cases, and of non-cases, will be selected, and the corresponding medical charts retrieved and reviewed for validation by pairs of trained, independent reviewers. For each condition considered, case adjudication of disease will be based on symptoms, laboratory and diagnostic tests, as available in medical charts. Divergences will be resolved by consensus. Sensitivity and specificity with 95% CIs will be calculated.

ETHICS AND DISSEMINATION

Research protocol has been granted approval by the Regional Ethics Committee. Study results will be disseminated widely through peer-reviewed publications and presentations at national and international conferences.

How Reliable Are Administrative Data for Capturing Stroke Patients and Their Care

Background/Aims

The reliability of diagnostic coding of acute stroke and transient ischemic attack (TIA) in administrative data is uncertain. The purpose of this study is to determine the agreement between administrative data sources and chart audit for the identification of stroke type, stroke risk factors, and the use of hospital-based diagnostic procedures in patients with stroke or TIA.

Methods

Medical charts for a population-based sample of patients (n = 14,508) with ischemic stroke, intracerebral hemorrhage (ICH), or TIA discharged from inpatient and emergency departments (ED) in Ontario, Canada, between April 1, 2012 and March 31, 2013, were audited by trained abstractors. Audited data were linked and compared with hospital administrative data and physician billing data. The positive predictive value (PPV) of hospital administrative data and kappa agreement for the reporting of stroke type were calculated. Kappa agreement was also determined for stroke risk factors and for select stroke-related procedures.

Results

The PPV for stroke type in inpatient administrative data ranged from 89.5% (95% CI 88.0-91.0) for TIA, 91.9% (95% CI 90.2-93.5) for ICH, and 97.3% (95% CI 96.9-97.7) for ischemic stroke. For ED administrative data, PPV varied from 78.8% (95% CI 76.3-81.2) for ischemic, 86.3% (95% CI 76.8-95.7) for ICH, and 95.3% (95% CI 94.6-96.0) for TIA. The chance-corrected agreement between the audited and administrative data was good for atrial fibrillation (k = 0.60) and very good for diabetes (k = 0.86). Hospital administrative data combined with physician billing data more than doubled the observed agreement for carotid imaging (k = 0.65) and echocardiography (k = 0.66) compared to hospital administrative data alone.

Conclusions

Inpatient and ED administrative data were found to be reliable in the reporting of the International Classification of Diagnosis, 10th revision, Canada (ICD-10-CA)-coded ischemic stroke, ICH and TIA, and for the recording of atrial fibrillation and diabetes. The combination of physician billing data with hospital administrative data greatly improved the capture of some diagnostic services provided to inpatients.