Search

Background A computable phenotype is an algorithm used to identify a group of patients within an electronic medical record system. Developing a computable phenotype that can accurately identify patients with autosomal dominant polycystic kidney disease (ADPKD) will assist researchers in defining patients eligible to participate in clinical trials and other studies. Our objective was to assess the accuracy of a computable phenotype using International Classification of Diseases 9th and 10th revision (ICD-9/10) codes to identify patients with ADPKD. Methods We reviewed four random samples of approximately 250 patients on the basis of ICD-9/10 codes from the EHR from the Kansas University Medical Center database: patients followed in nephrology clinics who had ICD-9/10 codes for ADPKD (Neph+), patients seen in nephrology clinics without ICD codes for ADPKD (Neph−), patients who were not followed in nephrology clinics with ICD codes for ADPKD (No Neph+), and patients not seen in nephrology clinics without ICD codes for ADPKD (No Neph−). We reviewed the charts and determined ADPKD status on the basis of internationally accepted diagnostic criteria for ADPKD. Results The computable phenotype to identify patients with ADPKD who attended nephrology clinics has a sensitivity of 99% (95% confidence interval [95% CI], 96.4 to 99.7) and a specificity of 84% (95% CI, 79.5 to 88.1). For those who did not attend nephrology clinics, the sensitivity was 97% (95% CI, 93.3 to 99.0), and a specificity was 82% (95% CI, 77.4 to 86.1). Conclusion A computable phenotype using the ICD-9/10 codes can correctly identify most patients with ADPKD, and can be utilized by researchers to screen health care records for cohorts of patients with ADPKD with acceptable accuracy.

Background: The National Organization for Civil Registration (NOCR) of Iran has reported causes of death data by ICD-10 chapters for the first time in 2020. Methods: We used this report to review the share of ICD chapters among all deaths in each province and compare them with the Global Burden of Disease (GBD) study. Results: There are major changes in the distribution of causes of death between 2017 and 2018, especially in D50-D89 (Diseases of the blood and blood-forming organs and certain disorders involving the immune mechanism) from 0.76% to 38.94%, and I00-I99 (Diseases of the circulatory system) from 39.27% to 7.09%. Such dramatic changes are probably the results of changes in coding practices or definitions or issues in analysis. Conclusion: Causes of death reports should be timely, clear, and robust on methods. They should contain a minimum level of details, at least 3-digit ICD codes to be useful for public health and medical professionals.

States that have legalized cannabis for medicinal or retail use have noted increases in emergency department (ED) visits associated with cannabis-associated harms. However, most studies are retrospective and identify cannabis-related visits using diagnosis codes or THC-positive urine drug immunoassays. These identification methods are limited. ICD codes are infrequently used in adult ED cases of exposure to drugs of abuse and clinicians may not explicitly document diagnosis of drug use, even when highly suspected or confirmed by patients.

Purpose Population-level studies on the treatment practices and comparative effectiveness of therapies in autoimmune hepatitis (AIH) are lacking due to the absence of validated methods to identify patients with AIH in large databases, such as administrative claims or electronic health records. This study ascertained the performance of International Classification of Diseases (ICD) codes for AIH, and developed and validated a novel algorithm that reliably identifies patients with AIH in health administrative data and claims. Methods This was a cross-sectional study of patients with ≥1 inpatient or ≥2 outpatient ICD codes for AIH between 2008 and 2019 at a single health system. In a random sample of 250 patients, definite or probable AIH was determined using the Simplified AIH score, Revised AIH score or expert adjudication. The positive predictive value (PPV) was obtained. Variations of this base algorithm were evaluated using additional criteria to increase its performance. Results Of the 250 patients, 143 (57.2%) patients had sufficient records available for review. The PPV of the base algorithm was 77.6% (95% CI: 69.9-84.2%). Exclusion of patients with ≥1 ICD code for primary biliary cholangitis or primary sclerosing cholangitis yielded a PPV of 89.7% (95% CI: 82.8-94.6%). Further exclusion of patients with recent immune checkpoint inhibitor therapy increased the PPV to 92.9% (95% CI: 86.5-96.9%). Conclusions The use of ICD codes for AIH alone are insufficient to reliably identify patients with AIH in health administrative data and claims. Our proposed algorithm that includes additional diagnostic and medication-related coding criteria demonstrates excellent performance.

Background Recent publications have increasingly demonstrated a link between superficial-vein thrombosis (SVT) and deep-vein thrombosis (DVT) in the adult population and have led to changes in SVT treatment considerations. A similar relationship between SVT and DVT in pediatric populations, however, is not currently well established. Objectives We sought to evaluate the temporal and anatomic relationship between SVT and DVT among pediatric inpatients in order to determine to what degree SVT is associated with DVT. Methods We first retrospectively reviewed our institution's local prospective hospital-acquired VTE (HA-VTE) registry to identify all children (age 0–21 years, inclusive) admitted to Children's Hospital Colorado for more than 48 h between January 2012 and September 2017 who developed a DVT while hospitalized. We then reviewed each patient's electronic health record for evidence of SVT to identify SVT + DVT cases. Afterwards, we utilized a list of ICD codes to identify all patients during this time frame who developed an SVT and removed patients whom we previously identified as SVT + DVT cases to obtain the number of patients with isolated SVT. Results Of 59,910 patients admitted during the study period, 438 (0.7%) developed a thrombosis while hospitalized - 197 (0.3%) with isolated SVT, 161 (0.3%) with isolated DVT, and 80 (0.1%) with both SVT + DVT. These 80 SVT + DVT patients represent 33% of the 241 total DVT patients and 29% of the 277 total SVT patients. Of the 12 SVT + DVT patients in whom the SVT was diagnosed before the DVT, the subsequent DVT occurred within a mean of 6.4 (range 1–22) days and at the same anatomic site in 6 (50%). The age breakdown for this cohort was: 0 (0%) 0–1 months, 2 (17%) 1 month-2 years, 3 (25%) 2–12 years, 3 (25%) 12–16 years, and 4 (33%) 16–21 years. Conclusions Our results indicate a temporal and anatomic relationship between SVT and DVT in hospitalized children, particularly those with central venous catheters.

Introduction The incidence of and risk factors for exertional heat illness (EHI) and cold weather injury (CWI) in the U.S. Army have been well documented. The “heat season”, when the risk of EHI is highest and application of risk mitigation procedures is mandatory, has been arbitrarily defined as May 1 through September 30, while the “cold season” is understood to occur from October 1 to April 30 each year. The proportions of EHI and CWI that occur outside of the traditional heat and cold seasons are unknown. Additionally, it is unknown if either of the seasonal definitions are appropriate. The primary purpose of this study was to determine the proportion of EHI and of CWI that occur within the commonly accepted seasonal definitions. We also report the location-specific variability, seasonal definitions, and the demographic characteristics of the populations. Methods The U.S. Army installations with the highest frequency of EHI and of CWI from 2008 to 2013 were identified and used for analysis. In total there were 15 installations included in the study, with five installations used for analysis in both the EHI and CWI projects. In- and out-patient EHI and CWI data (ICD-9-CM codes 992.0 to 992.9 and ICD codes 991.0 to 991.9, respectively) were obtained from the Defense Medical Surveillance System. Installation-specific denominator data were obtained from the Defense Manpower Data Center, and incidence rates were calculated by week, for each installation. Segmental (piecewise) regression analysis was used to determine the start and end of the heat and cold seasons. Results Our analysis indicates that the heat season starts around April 22 and ends around September 9. The cold season starts on October 3 and ends on March 24. The majority (n = 6,445, 82.3%) of EHIs were diagnosed during the “heat season” of May 1 to September 30, while 10.3% occurred before the heat season started (January1 to April 30) and 7.3% occurred after the heat season ended (October 1 to December 31). Similar to EHI, 90.5% of all CWIs occurred within the traditionally defined cold season, while 5.7% occurred before and 3.8% occurred after the cold season. The locations with the greatest EHI frequency were Ft Bragg (n = 2,129), Ft Benning (n = 1,560), and Ft Jackson (n = 1,538). The bases with the largest proportion of CWI in this sample were Ft Bragg (17.8%), Ft Wainwright (17.2%), and Ft Jackson (12.7%). There were considerable inter-installation differences for the start and end dates of the respective seasons. Conclusions The present study indicates that the traditional heat season definition should be revised to begin ∼3 weeks earlier than the current date of May 1; our data indicate that the current cold season definition is appropriate. Inter-installation variability in the start of the cold season was much larger than that for the heat season. Exertional heat illnesses are a year-round problem, with ∼17% of all cases occurring during non-summer months, when environmental heat strain and vigilance are lower. This suggests that EHI mitigation policies and procedures require greater year-round emphasis, particularly at certain locations.

Infective endocarditis (IE) secondary to injection drug use (IDU-IE) is a disease with high morbidity, cost, and rapid demographic evolution. Studies frequently utilize combinations of International Statistical Classification of Diseases (ICD) codes to identify IDU-IE cases in electronic medical records. This is a validation of this identification strategy in a US cohort.Records from January 1, 2004 to September 31, 2015 for those aged ≥18yo with any ICD-coded IE encounter (inpatient or outpatient) were retrieved from the electronic medical record and then manually reviewed and classified as IDU-IE by strict and inclusive criteria. This registry was then used to assess the diagnostic accuracy of 10 identification algorithms that combined substance use, hepatitis C, and IE ICD codes.IE was present in 629 of the 2055 manually reviewed records; 109 reported IDU within 3 months of IE diagnosis and an additional 32 during their lifetime (141 cases). In contrast, no algorithm identified more than 46 (33%) of these cases. Algorithms assessing encounters with both an IE and substance use code had specificities99% but sensitivities ≤11% with negative predictive values of 83% to 84% and positive predictive values ranging from 75% to 91%. Use of a hepatitis C OR substance use code with an IE-coded encounter resulted in higher sensitivities of 22% to 32% but more false positives and overall positive predictive value of70%. This algorithm limited to age ≤45yo had the best, but still low, discrimination ability with an area under the receiver operating characteristic curve of 0.62.Substance use and hepatitis C codes have poor ability to accurately classify an IE-coded encounter as IDU-IE or routine IE.

Background Little is known about the impact of the ecosystem disruption and its contribution on the non-tuberculosis mycobacteria (NTM) diseases (cases) rate in Florida (FL), a state with a high prevalence of NTM in the United States. We aimed to evaluate the epidemiological distribution of NTM in FL and identify its association with extreme weather events. Methods We used OneFlorida Clinical Research Consortium dataset and extracted data on NTM cases using ICD codes 9- CM 031.0 and ICD-10 A31 during 2012–2018. The number of hurricanes during the study period which affected FL were extracted data from the National Hurricane Center (NHC) and the National Oceanic and Atmospheric Administration (NOAA). Results Prevalence of NTM gradually increased during the study period. The rate was 2012: 14.3/100,000, 2015; 20.1/100,000 and 2018; 22.6/100,00 except in 2014 where there was an 8% decrease. The incidences were 2012; 6.5/100,00, 2015; 4.9/100,000 and in 2015; 5.4/100,000. Geographical analysis demonstrated a gradual expansion of the NTM cases in Alachua, and Marion Counties throughout the study period. Notably, the 2018 heat map showed higher prevalence of NTM in the northwestern, panhandle region of FL which had been absent in the heat maps for years 2012–2018. High number of the hurricanes was associated with the higher number of the new cases of NTM infection for years 2012, 2016–2018, while the lower number of the hurricanes was associated with the lower number of the new cases of NTM infection for years 2014–2015. Conclusion The current study found the prevalence rates of NTM disease in FL rose from 2012 to 2018. A higher prevalence was seen following the hurricanes.

Objective This study introduces a temporal condition pattern mining methodology to address the sparse nature of coded condition concept utilization in electronic health record data. As a validation study, we applied this method to reveal condition patterns surrounding an initial diagnosis of pediatric asthma. Materials and Methods The SPADE (Sequential PAttern Discovery using Equivalence classes) algorithm was used to identify common temporal condition patterns surrounding the initial diagnosis of pediatric asthma in a study population of 71 824 patients from the Children’s Hospital of Philadelphia. SPADE was applied to a dataset with diagnoses coded using International Classification of Diseases (ICD) concepts and separately to a dataset with the ICD codes mapped to their corresponding expanded diagnostic clusters (EDCs). Common temporal condition patterns surrounding the initial diagnosis of pediatric asthma ascertained by SPADE from both the ICD and EDC datasets were compared. Results SPADE identified 36 unique diagnoses in the mapped EDC dataset, whereas only 19 were recognized in the ICD dataset. Temporal trends in condition diagnoses ascertained from the EDC data were not discoverable in the ICD dataset. Discussion Mining frequent temporal condition patterns from large electronic health record datasets may reveal previously unknown associations between diagnoses that could inform future research into causation or other relationships. Mapping sparsely coded medical concepts into homogenous groups was essential to discovering potentially useful information from our dataset. Conclusions We expect that the presented methodology is applicable to the study of diagnostic trajectories for other clinical conditions and can be extended to study temporal patterns of other coded medical concepts such as medications and procedures.

Literature regarding outcomes of cardiac arrest with associated NSTEMI is limited. We aim to study the predictors and survival outcomes of cardiac arrest patients presenting to the emergency department who were diagnosed with non-ST elevated myocardial infarction (NSTEMI).Data from the nationwide emergency department sample (NEDS) that constitutes 20% sample of hospital-owned emergency departments in the United States was analyzed for the cardiac arrest related visits from 2009-2018. Cardiac arrest was defined by the ICD codes.Out of 3,235,555 cardiac arrests (mean age 64.0 ± 19.5 years, 40.7% females) there were 163,970 (5.1%) patients diagnosed with NSTEMI during the years 2009-2018. Among cardiac arrest patients, the survival for NSTEMI patients was higher than patients without NSTEMI (46.7% vs. 22.7%). These patients were more likely to be males and elderly. Among the predictors for NSTEMI cardiac arrests, hypertension (OR 1.12, p 0.001), peripheral vascular disease (OR 1.16, p 0.001), prior-coronary artery bypass graft (OR 1.20, p 0.001) were the predominant ones. Cardiovascular interventions were more common in NSTEMI cardiac arrests and were associated with lower mortality rates (p 0.001). However, trend for coronary interventions remained steady over study years. We observed an increase in prevalence of NSTEMI cardiac arrests with a worsening trend in survival from 2009-2018.NSTEMI was not uncommon in patients with cardiac arrest. NSTEMI cardiac arrest had a better prognosis than patients without NSTEMI. Cardiovascular interventions might have survival benefits. More research is required to identify NSTEMI in cardiac arrest patients and further evaluate the effect of cardiovascular interventions on survival.

Adverse drug reactions (ADRs) are common yet are often underreported, making them difficult to track and study. Prospective pharmacovigilance programs significantly increase detection and reporting of ADRs. The aim of this pilot study was to apply triggers used by a prospective pharmacovigilance program at a freestanding children's hospital to retrospectively detect ADRs at our institution, therefore determining if these methods could be replicated and provide the basis for implementation of a prospective pharmacovigilance program. In 2019, our institution had 22 000 inpatient admissions and 51 000 emergency room visits and had 21 ADRs voluntarily reported in an electronic medication safety tracking system. Additional ADRs were identified by methods including new or modified entries to a patient's allergy profile in the electronic medical record (EMR) and International Classification of Disease (ICD) codes. We identified 754 unique patients with changes to allergy profile and 5719 ICD codes in 3966 unique patients to evaluate. These triggers prompted screening of the EMR to validate the ADR, and we identified 280 ADRs occurring in 2019. Eight (2.8%) were identified solely by the electronic medication safety tracking system, 64 (23%) were identified by the allergy list, 110 (39%) were identified only by ICD coding, and the remaining 98 (35%) were identified by multiple methods. The use of triggers followed by review of the EMR identified 13-fold more ADRs than were voluntarily reported, illustrating the need for an active pharmacovigilance service and the successful use of multimodal methods to detect and track ADRs.

Administrative data is useful in population-based studies in hemophilia, but few studies validated the coding accuracy.We validated the accuracy of a case definition using a combination of International Classification of Diseases diagnostic codes and coagulation factor level for identifying hemophilia in administrative data.This is a retrospective population-based study of all residents of Alberta, Canada, who underwent testing for coagulation factor VIII (FVIII) or factor IX (FIX) activity between 2009 and 2017 using linked administrative data. Our predefined algorithm was a combination of the relevant ICD codes and FVIII/FIX activity0.4 IU/ml. Medical charts of 2114 randomly selected patients tested for FVIII and 528 patients tested for FIX were reviewed to identify physician diagnoses of hemophilia. Sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) were calculated.With our algorithm, 72 (3.4%) patients tested for FVIII and 21 (4.0%) tested for FIX met the combined ICD code and laboratory criteria for hemophilia A and hemophilia B, respectively, whereas 63 (3.0%) and 22 (4.2%) had confirmed hemophilia A and hemophilia B upon chart review. Our algorithm had a sensitivity of 93.7% and specificity of 99.4% for identifying hemophilia A, and a sensitivity of 90.9% and specificity of 99.8% for identifying hemophilia B.This study showed that using a case definition of ICD codes and coagulation factor activities can identify hemophilia in administrative data with very high accuracy and can be used for future research.

Psychiatric disorders are common in pediatric HTx recipients. However, the impact of psychiatric comorbidities on patient outcomes is unknown. We aimed to assess the impact of disorders of adjustment, depression, and anxiety on HTx outcomes in children; hypothesizing that the presence of psychiatric disorders during or preceding HTx would negatively impact outcomes. All pediatric HTx recipients ≥8 years of age who survived to hospital discharge were identified from a novel linkage between the PHIS and SRTR databases (2002-2016). Psychiatric disorders were identified using ICD codes during or preceding the HTx admission. Post-transplant graft survival, freedom from readmission, and freedom from rejection were analyzed using the Kaplan-Meier method. Multivariable Cox proportional hazard models were used to adjust for covariates. A total of 1192 patients were included, of which 133 (11.2%) had depression, 197 (16.5%) had anxiety, and 218 (18.3%) had adjustment disorders. The presence of depression was independently associated with higher rates of readmission (60.9% vs 54.1% at 6 months) (AHR 1.63, 95% CI 1.22-2.18, P = .001) and inferior graft survival (70.2% vs 83.4% at 5 years) (AHR 1.62, 95% CI 1.14-2.3, P = .007). Anxiety was independently associated with higher rates of readmission (60.4% vs 53.9% at 6 months) (AHR 1.46, 95% CI 1.09-1.94, P = .01). Anxiety and depression in the pretransplant period are independently associated with outcomes following HTx in children. Evaluation and management of psychiatric comorbidities represents an important component of care in this vulnerable population.

Objective The objective of this study was to compare the performance of an RA algorithm developed and trained in 2010 utilizing natural language processing and machine learning, using updated data containing ICD10, new RA treatments, and a new electronic medical records (EMR) system. Methods We extracted data from subjects with ≥1 RA International Classification of Diseases (ICD) codes from the EMR of two large academic centres to create a data mart. Gold standard RA cases were identified from reviewing a random 200 subjects from the data mart, and a random 100 subjects who only have RA ICD10 codes. We compared the performance of the following algorithms using the original 2010 data with updated data: (i) a published 2010 RA algorithm; (ii) updated algorithm, incorporating ICD10 RA codes and new DMARDs; and (iii) published algorithm using ICD codes only, ICD RA code ≥3. Results The gold standard RA cases had mean age 65.5 years, 78.7% female, 74.1% RF or antibodies to cyclic citrullinated peptide (anti-CCP) positive. The positive predictive value (PPV) for ≥3 RA ICD was 54%, compared with 56% in 2010. At a specificity of 95%, the PPV of the 2010 algorithm and the updated version were both 91%, compared with 94% (95% CI: 91, 96%) in 2010. In subjects with ICD10 data only, the PPV for the updated 2010 RA algorithm was 93%. Conclusion The 2010 RA algorithm validated with the updated data with similar performance characteristics as the 2010 data. While the 2010 algorithm continued to perform better than the rule-based approach, the PPV of the latter also remained stable over time.

Population-based data on the prevalence of and real-life treatment for the autoimmune liver diseases (AILD), primary biliary cholangitis (PBC) and autoimmune hepatitis (AIH), are scarce, and such knowledge may help to improve patient care. Data of 8.1 million individuals having health insurance with the "Techniker Krankenkasse," the largest German health insurer, were analyzed with regard to the prevalence of and prescribed medication for AILD in Germany from 2011 until 2014. Patients with viral hepatitis B infection (HBV) and alcoholic liver cirrhosis (ALC) served as control groups. Case definition was based on ICD codes. The prevalences of PBC and AIH were 36.9/100 000 inhabitants (95 % CI: 35.6-38.2) and 23.0/100 000 inhabitants (95 % CI: 22.0-24.0) in 2014, respectively. The prevalences of AILD increased from 2011 to 2014 (for PBC by 31 % and for AIH by 29 %), with the largest increase for male patients with PBC. In contrast, the prevalence of HBV declined while that of ALC remained stable. The analysis of prescribed real-life treatment revealed considerable deviations from standard treatment recommendations. Specifically, in older patients with PBC or AIH, undertreatment was common. The prevalence of PBC and AIH based on ICD codes is increasing in Germany. The analysis of real-life treatment in this large and population-based cohort points to the unmet need to improve the implementation of treatment guidelines for autoimmune liver diseases in the broader medical community. Es gibt nur wenige populationsbasierte Studien zur Prävalenz und zur „Real-Life“-Therapie der autoimmunen Lebererkrankungen Primär Biliäre Cholangitis (PBC) und Autoimmune Hepatitis (AIH). Es wurden 8,1 Millionen Versichertendaten der Techniker Krankenkasse analysiert bezüglich der Prävalenz und rezeptierten Medikation von Patienten mit autoimmunen Lebererkrankungen in Deutschland von 2011 bis 2014. Patienten mit einer Virushepatitis-B-Infektion (HBV) und mit alkoholischer Leberzirrhose (C2 LZX) dienten als Kontrollgruppen. Die Patienten wurden anhand von ICD-Codes identifiziert. Die PBC und die AIH weisen in Deutschland im Jahr 2014 eine Prävalenz von 36,9 pro 100 000 Einwohner (95 %-KI: 35,6–38,2) beziehungsweise 23,0 pro 100 000 Einwohner (95 %-KI: 22,0–24,0) auf. Die Prävalenz autoimmuner Lebererkrankungen stieg von 2011 bis 2014 an (für die PBC um 31 % und für die AIH um 29 %), wobei männliche PBC-Patienten den stärksten Anstieg aufwiesen. Die Prävalenz von HBV und C2 LZX nahm ab beziehungsweise blieb stabil. Eine Analyse der „Real-Life“-Therapie der autoimmunen Lebererkrankungen ergab erhebliche Abweichungen von Therapieleitlinien; insbesondere bei älteren AIH- und PBC-Patienten lag eine Untertherapie vor. Die Prävalenz der PBC und AIH basierend auf ICD-Codes steigt in Deutschland an. Die populationsbasierte Analyse der „Real-Life“-Therapie autoimmuner Lebererkrankungen in Deutschland weist auf Defizite in der breiten medizinischen Versorgung hinsichtlich der Umsetzung von Therapieleitlinien hin.

Introduction: Whether data on International Classification of Diseases (ICD)-codes from the Swedish National Patient Register (NPR) correctly correspond to subtypes of inflammatory bowel disease (IBD) and phenotypes of the Montreal classification scheme among patients with prevalent disease is unknown. Materials and methods: We obtained information on IBD subtypes and phenotypes from the medical records of 1403 patients with known IBD who underwent biological treatment at ten Swedish hospitals and retrieved information on their IBD-associated diagnostic codes from the NPR. We used previously described algorithms to define IBD subtypes and phenotypes. Finally, we compared these register-generated subtypes and phenotypes with the corresponding information from the medical records and calculated positive predictive values (PPV) with 95% confidence intervals. Results: Among patients with clinically confirmed disease and diagnostic listings of IBD in the NPR (N = 1401), the PPV was 97 (96–99)% for Crohn’s disease, 98 (97–100)% for ulcerative colitis, and 8 (4–11)% for IBD-unclassified. The overall accuracy for age at diagnosis was 95% (when defined as A1, A2, or A3). Examining the validity of codes representing disease phenotype, the PPV was 36 (32–40)% for colonic Crohn’s disease (L2), 61 (56–65)% for non-stricturing/non-penetrating Crohn’s disease behaviour (B1) and 83 (78–87)% for perianal disease. Correspondingly, the PPV was 80 (71–89)% for proctitis (E1)/left-sided colitis (E2) in ulcerative colitis. Conclusions: Among people with known IBD, the NPR is a reliable source of data to classify most subtypes of prevalent IBD, even though misclassification commonly occurred in Crohn’s disease location and behaviour and also among IBD-unclassified patients.

Release of the Netflix series 13 Reasons Why in March 2017 raised concern over associated suicide attempts. This study aimed to identify trends in self-harm admissions to a tertiary children's hospital with special attention paid to the time after series release.Records for admitted patients ages 4-18 years from January 2012 to October 2017 were identified based on ICD codes indicating self-harm. Admissions were grouped by month, and the ARMA (Auto Regression and Moving Average) model was used in analysis. Log transformation was used to obtain a constant variance, and seasonal terms were added for adjustment. A "postintervention" level shift, temporary shift, and linear growth term were incorporated as predictors in ARMA models to test for differences using the series premier as the intervention. Terms from the best fitting model (without intervention effects) were fit to preintervention data and forecast predictions were compared to the observed data from the postintervention period.Seven hundred seventy-five records were included in analysis. There was an increase of .024 in the log of suicide admission counts per month (p.001). The model that best explained the data was an ARMA (2,2) model with cubic growth curve terms, a post-intervention level shift, and a postintervention linear growth term, indicating an increase in observed over expected admissions following the premiere.Suicide admission counts increased over the time series. Actual suicide admissions following March 2017 were higher than predicted using the optimal model, suggesting an effect that temporally coincides with the release of 13 Reasons Why.

Many competitive health insurance markets adjust payments to participating health plans according to their enrollees’ risk − including based on diagnostic information. We investigate responses of German health plans to the introduction of morbidity-based risk adjustment in the Statutory Health Insurance in 2009, which triggers payments based on “validated” diagnoses by providers. Using the regulator’s data from office-based physicians, we estimate a difference-in-difference analysis of the change in the share and number of validated diagnoses for ICD codes that are inside or outside the risk adjustment but are otherwise similar. We find a differential increase in the share of validated diagnoses of 2.6 and 3.6 percentage points (3–4%) between 2008 and 2013. This increase appears to originate from both a shift from not-validated toward validated diagnoses and an increase in the number of such diagnoses. Overall, our results indicate that plans were successful in influencing physicians’ coding practices in a way that could lead to higher payments.

BACKGROUND: Telephone consultation and triage are used to limit the workload on emergency departments. Lack of visual cues and clinical tests put telephone consultations to a disadvantage compared to face-to-face consultations increasing the risk of under-triage. Under-triage occurs in telephone triage; however why under-triage happens is not explored yet. The aim of the study was to describe situations of under-triage in context, to assess the quality of under-triaged calls, and to identify communication patterns contributing to under-triage in a regional OOH service in the capital region of Denmark.METHODS: Explanatory simultaneous mixed method with thematic analysis and descriptive statistics was chosen. The study was carried out in an Out-Of-Hours service (OOH) in the Capital Region of Denmark, Copenhagen. Under-triage was defined as Potentially Under-Triaged Calls (PUTC) by specific criteria to an OOH Hotline, and identification by integration of three databases: Medical Hotline database, Emergency number database, including the Ambulance database, and electronic patient records. Distribution of PUTC were carried out using ICD-10 codes to identify diagnosis and main themes identified by qualitative analysis of audio recorded under-triaged calls. Study period was October 15(th) to November 30(th) 2014.RESULTS: Three hundred twenty seven PUTC were identified, representing 0.04% of all calls (n = 937.056) to the OOH. Distribution of PUTC according to diagnoses was: digestive (24%), circulatory (19%), respiratory (15%) and all others (42%). Thematic analysis of the voice logs suggested that inadequate communication and non-normative symptom description contributed to under-triage.DISCUSSION: The incidence of potentially under-triage is low (0.04%). However, the over-representation of digestive, circulatory, and respiratory diagnoses might suggest that under-triage is related to inadequate symptom description. We recommend that caller and call-handler collaborate systematically on problem identification and negotiate non-normative symptom description.CONCLUSION: The incidence of under-triage is low (0.04%). However, the over-representation of digestive, circulatory, and respiratory diagnoses might suggest that under-triage is related to inadequate symptom description. We recommend that caller and call-handler collaborate systematically on problem identification and negotiate non-normative symptom description.

Leveraging the Electronic Health Records (EHR) longitudinal data to produce actionable clinical insights has always been a critical issue for recent studies. Non-forecasted extended hospitalizations account for a disproportionate amount of resource use, the mediocre quality of inpatient care, and avoidable fatalities. The capability to predict the Length of Stay (LoS) and mortality in the early stages of the admission provides opportunities to improve care and prevent many preventable losses. Forecasting the in-hospital mortality is important in providing clinicians with enough insights to make decisions and hospitals to allocate resources, hence predicting the LoS and mortality within the first day of admission is a difficult but a paramount endeavor. The biggest challenge is that few data are available by this time, thus the prediction has to bring in the previous admissions history and free text diagnosis that are recorded immediately on admission. We propose a model that uses the multi-modal EHR structured medical codes and key demographic information to classify the LoS in 3 classes; Short Los (LoS ⩽ 10 days), Medium LoS (10 LoS ⩽ 30 days) and Long LoS (LoS > 30 days) as well as mortality as a binary classification of a patient’s death during current admission. The prediction has to use data available only within 24 h of admission. The key predictors include previous ICD9 diagnosis codes, ICD9 procedures, key demographic data, and free text diagnosis of the current admission recorded right on admission. We propose a Hierarchical Attention Network (HAN-LoS and HAN-Mor) model and train it to a dataset of over 45321 admissions recorded in the de-identified MIMIC-III dataset. For improved prediction, our attention mechanisms can focus on the most influential past admissions and most influential codes in these admissions. For fair performance evaluation, we implemented and compared the HAN model with previous approaches. With dataset balancing techniques HAN-LoS achieved an AUROC of over 0.82 and a Micro-F1 score of 0.24 and HAN-Mor achieved AUC-ROC of 0.87 hence outperforming the existing baselines that use structured medical codes as well as clinical time series for LoS and Mortality forecasting. By predicting mortality and LoS using the same model, we show that with little tuning the proposed model can be used for other clinical predictive tasks like phenotyping, decompensation,re-admission prediction, and survival analysis.

Identification of patients for epidemiologic research through administrative coding has important limitations. We investigated the feasibility of a search based on natural language processing (NLP) on the text sections of electronic health records for identification of patients with septic shock.Results of an explicit search strategy (using explicit concept retrieval) and a combined search strategy (using both explicit and implicit concept retrieval) were compared to hospital ICD-9 based administrative coding and to our department's own prospectively compiled infection database.Of 8911 patients admitted to the medical or surgical ICU, 1023 (11.5%) suffered from septic shock according to the combined search strategy. This was significantly more than those identified by the explicit strategy (518, 5.8%), by hospital administrative coding (549, 5.8%) or by our own prospectively compiled database (609, 6.8%) (p .001). Sensitivity and specificity of the automated combined search strategy were 72.7% (95%CI 69.0%-76.2%) and 93.0% (95%CI 92.4%-93.6%), compared to 56.0% (95%CI 52.0%-60.0%) and 97.5% (95%CI 97.1%-97.8%) for hospital administrative coding.An automated search strategy based on a combination of explicit and implicit concept retrieval is feasible to screen electronic health records for septic shock and outperforms an administrative coding based explicit approach.

Background Cardiovascular severe maternal morbidity (CSMM) is rising and has become the leading cause of maternal mortality. Research using administrative data sets may allow for better understanding of this critical group of diseases. Objective To validate a composite variable of CSMM for use in epidemiologic studies. Methods We analysed delivery hospitalisations at an obstetric teaching hospital from 2007 to 2017. We utilised a subset of indicators developed by the Centers for Disease Control and Prevention based on ICD codes to form the composite variable for CSMM. Two expert clinicians manually reviewed all qualifying events using a standardised tool to determine whether these represented true CSMM events. Additionally, we estimated the number of CSMM cases among delivery hospitalisations without qualifying ICD codes by manually reviewing all hospitalisations with severe preeclampsia, a population at high risk of CSMM, and a random sample of 1000 hospitalisations without severe preeclampsia. We estimated validity of the composite variable. Results Among 91 355 admissions for delivery, we captured 113 potential CSMM cases using qualifying ICD codes. Of these, 65 (57.5%) were true CSMM cases. Indicators for acute myocardial infarction, cardiac arrest, and cardioversion had the highest true-positive rates (100% for all). We found an additional 70 CSMM cases in the 2102 admissions with severe preeclampsia and a single CSMM case in the random sample. Assuming a rate of 1 CSMM case per 1000 deliveries in the remaining cohort, the composite variable had a positive predictive value of 57.5% (95% CI 47,9, 66.8), a negative predictive value of 99.8% (95% CI 99.8, 99.9), a sensitivity of 29.0% (95% CI 23.2, 35.4), and a specificity of 100% (95% CI 99.9, 100.0). Conclusion A novel composite variable for CSMM had reasonable PPV but limited sensitivity. This composite variable may enable epidemiologic studies geared towards reducing maternal morbidity and mortality.

Background Policy evaluations and health system interventions often utilize International Classification of Diseases (ICD) codes of opioid use, dependence, and abuse to identify individuals with opioid use disorder (OUD) and assess receipt of evidence-based treatments. However, ICD codes may not map directly onto the Diagnostic and Statistical Manual of Mental Disorder (DSM-5) OUD criteria. This study investigates the positive predictive value of ICD codes in identifying patients with OUD. Methods We conducted a clinical chart review on a national sample of 520 Veterans assigned ICD-9 or ICD-10 codes for opioid use, dependence, or abuse from 2012 to 2017. We extracted evidence of DSM-5 OUD criteria and opioid misuse from clinical documentation in the month preceding and three months following initial ICD code listing, and categorized patients into: 1) high likelihood of OUD, 2) limited aberrant opioid use, 3) prescribed opioid use without evidence of aberrant use, and 4) insufficient information. Positive predictive value was calculated as the percentage of individuals with these ICD codes meeting high likelihood of OUD criteria upon chart review. Results Only 57.7 % of patients were categorized as high likelihood of OUD; 16.5 % were categorized as limited aberrant opioid use, 18.9 % prescribed opioid use without evidence of aberrant use, and 6.9 % insufficient information. Conclusions Patients assigned ICD codes for opioid use, dependence, or abuse often lack documentation of meeting OUD criteria. Many receive long-term opioid therapy for chronic pain without evidence of misuse. Robust methods of identifying individuals with OUD are crucial to improving access to clinically appropriate treatment.

This article describes how maps were developed from the clinical modifications of the 9th and 10th revisions of the International Classification of Diseases (ICD) to the Abbreviated Injury Scale 2005 Update 2008 (AIS08). The development of the mapping methodology is described, with discussion of the major assumptions used in the process to map ICD codes to AIS severities. There were many intricacies to developing the maps, because the 2 coding systems, ICD and AIS, were developed for different purposes and contain unique classification structures to meet these purposes.Experts in ICD and AIS analyzed the rules and coding guidelines of both injury coding schemes to develop rules for mapping ICD injury codes to the AIS08. This involved subject-matter expertise, detailed knowledge of anatomy, and an in-depth understanding of injury terms and definitions as applied in both taxonomies. The official ICD-9-CM and ICD-10-CM versions (injury sections) were mapped to the AIS08 codes and severities, following the rules outlined in each coding manual. The panel of experts was composed of coders certified in ICD and/or AIS from around the world. In the process of developing the map from ICD to AIS, the experts created rules to address issues with the differences in coding guidelines between the 2 schemas and assure a consistent approach to all codes.Over 19,000 ICD codes were analyzed and maps were generated for each code to AIS08 chapters, AIS08 severities, and Injury Severity Score (ISS) body regions. After completion of the maps, 14,101 (74%) of the eligible 19,012 injury-related ICD-9-CM and ICD-10-CM codes were assigned valid AIS08 severity scores between 1 and 6. The remaining 4,911 codes were assigned an AIS08 of 9 (unknown) or were determined to be nonmappable because the ICD description lacked sufficient qualifying information for determining severity according to AIS rules. There were also 15,214 (80%) ICD codes mapped to AIS08 chapter and ISS body region, which allow for ISS calculations for patient data sets.This mapping between ICD and AIS provides a comprehensive, expert-designed solution for analysts to bridge the data gap between the injury descriptions provided in hospital codes (ICD-9-CM, ICD-10-CM) and injury severity codes (AIS08). By applying consistent rules from both the ICD and AIS taxonomies, the expert panel created these definitive maps, which are the only ones endorsed by the Association for the Advancement of Automotive Medicine (AAAM). Initial validation upheld the quality of these maps for the estimation of AIS severity, but future work should include verification of these maps for MAIS and ISS estimations with large data sets. These ICD-AIS maps will support data analysis from databases with injury information classified in these 2 different systems and open new doors for the investigation of injury from traumatic events using large injury data sets.

Background: Adult studies demonstrated that extremes of nutritional status adversely impact clinical outcomes in acute pancreatitis (AP). With rising prevalence of undernutrition/obesity in children, we analyzed the effect of nutritional status on the clinical outcomes in children and adolescents with acute pancreatitis. Methodology: We analyzed the Kids&rsquo, Inpatient Database (KID) between 2003 and 2016 to include all patients with a primary diagnosis of AP using specific International Classification of Diseases (ICD) codes. We classified into (1) undernutrition, (2) obesity and (3) control groups, based on ICD codes, and we compared severe acute pancreatitis and healthcare utilization (length of stay and hospitalization costs). Results: Total number of AP admissions was 39,805. The prevalence of severe AP was higher in the undernutrition and obesity groups than the control group (15.7% vs. 5.8% vs. 3.5% respectively, p &lt, 0.001). Multivariate analyses demonstrated that undernutrition and obesity were associated with 2.5 and 1.6 times increased risk of severe AP, p &lt, 0.001. Undernutrition was associated with an additional six days of hospitalization and almost $16,000 in hospitalization costs. Obesity was associated with an additional 0.5 day and almost $2000 in hospitalization costs, p &lt, 0.001. Conclusion: Undernutrition and obesity were associated with greater severity of AP, as well as prolonged hospitalization stay and costs. It is imperative for treating clinicians to be aware of these high-risk groups to tailor management and strive for improved outcomes.

We read the work of Rudasill et al. [(1)][1] with great interest. The authors did an admirable job in studying the outcomes of infective endocarditis (IE) among injection drug users (IDU). Epidemiological studies indeed are valuable in representing the disease burden in society, and therefore

BACKGROUND: Biliary atresia (BA) is a rare obstructive cholangiopathy that presents in early infancy. The Kasai portoenterostomy (PE) improves survival with native liver. Epidural analgesia is an appealing option to control pain in this fragile patient population, yet its safety, efficacy, and potential benefits remain unproven. METHODS: Patients undergoing PE for BA between 2001 and 2016 at a single institution were identified by ICD codes. Preoperative labs, procedure details, and recovery outcomes were reviewed retrospectively. Outcomes of interest were need for postoperative mechanical ventilation, pain scores, normalized opioid administration, return of bowel function, and length of hospital stay after PE. RESULTS: Of 47 infants undergoing PE for BA, 25 received epidural analgesia, and 22 did not. Infants with epidurals received less systemic opioids over the first 96 hours post-operatively than those without (p

In the Policy Forum “To help aging populations, classify organismal senescence” (1 November 2019, p. [576][1]), S. R. G. Calimport et al. call for classification of senescence as a disease and propose a classification methodology. Unfortunately, their call to action comes a year after the publication of the 11th International Classification of Diseases (ICD) ([ 1 ][2]). The next revision of the ICD is expected in 2028 or later, and much progress needs to be made in basic research and clinical trials before including senescence would be possible. Calimport et al. 's assertion that aging-related diseases can be classified using current information overestimates our understanding of senescence. Multiple studies suggest that aging starts in utero ([ 2 ][3]) and staging of most senescence processes will require massive longitudinal studies in humans and animals. Furthermore, the first clinical trials targeting aging, such as Targeting Aging with Metformin (TAME), are under discussion, and no aging biomarker consensus or hierarchical structuring has yet emerged ([ 3 ][4], [ 4 ][5]). Staging senescence at the tissue-specific level for the purposes of preventative intervention may be a model for private or national health care plans and systems; however, these plans do not require these processes to be classified in the ICD with the individual codes. In addition to providing a general disease framework, ICD codes are used internationally to standardize burden of disease reports and assess cause of death. The concept of staging senescence does not fit the current ICD framework. The addition would likely require a new international initiative to develop a comprehensive set of biomarkers and interventions for senescence and longevity, as well as a set of recommendations for reducing the burden of senescence and aging-associated diseases. Pharmaceutical companies are already working on the interventions targeting the common biological pathways implicated in aging to deliver therapeutics against age-related diseases within the traditional health care paradigm. To accelerate drug development efforts focused on aging as a disease in the pharmaceutical industry, we need to demonstrate that basic research into whether aging is a biological process disease can yield valuable targets or interventions that demonstrate efficacy in treating or preventing age-related diseases. 1. [↵][6]1. A. Zhavoronkov, 2. B. Bhullar , Front. Genet. 6, 326 (2015). [OpenUrl][7] 2. [↵][8]1. B. J. Allison et al ., FASEB 30, 5 (2016). [OpenUrl][9] 3. [↵][10]1. J. N. Justice et al ., Cardiovasc. Endocrinol. Metab. 7, 80 (2018). [OpenUrl][11] 4. [↵][12]American Federation for Aging Research, “TAME trial” (2020); [www.afar.org/research/TAME/][13]. A.Z. is the founder and CEO of Insilico Medicine, a Hong Kong-based artificial intelligence company focused on drug discovery and aging research. He is also an adviser to Haut. AI Ltd., an Estonia-based artificial intelligence company focused on artificial intelligence for skin care and aging research. He is a director and co-founder of Deep Longevity, Inc., an artificial intelligence company focused on aging biomarkers. He is an adviser to the All Party Parliamentary Group (APPG) for Longevity in the UK. He is the seed investor in Retrotope, Inc., a California-based drug discovery company focused on age-related diseases. He may be associated with other startup companies in aging research as an adviser or through investments in Longenesis. [1]: http://www.sciencemag.org/content/366/6465/576 [2]: #ref-1 [3]: #ref-2 [4]: #ref-3 [5]: #ref-4 [6]: #xref-ref-1-1 "View reference 1 in text" [7]: {openurl}?query=rft.jtitle%253DFront.%2BGenet.%26rft.volume%253D6%26rft.spage%253D326%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [8]: #xref-ref-2-1 "View reference 2 in text" [9]: {openurl}?query=rft.jtitle%253DFASEB%26rft.volume%253D30%26rft.spage%253D5%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [10]: #xref-ref-3-1 "View reference 3 in text" [11]: {openurl}?query=rft.jtitle%253DCardiovasc.%2BEndocrinol.%2BMetab.%26rft.volume%253D7%26rft.spage%253D80%26rft.genre%253Darticle%26rft_val_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Ajournal%26ctx_ver%253DZ39.88-2004%26url_ver%253DZ39.88-2004%26url_ctx_fmt%253Dinfo%253Aofi%252Ffmt%253Akev%253Amtx%253Actx [12]: #xref-ref-4-1 "View reference 4 in text" [13]: http://www.afar.org/research/TAME/

This study investigates the use of systemic treatment for clinically amyopathic dermatomyositis using natural language queries and ICD codes for amyopathic and hypomyopathic dermatomyositis.

Opioid-related adverse drug events (ORADEs) include a range of complications, from respiratory arrest to ileus and urinary retention. ORADEs correlate to morbidity, mortality, and increased costs. The Centers for Medicare & Medicaid Services database, which represents approximately 35% of hospital discharges. The authors searched for previously published ICD-9 codes that defined ORADEs. A group of surgical diagnosis-related groups (DRGs) were selected. Recurring queries were programmed using these ICD codes and DRGs and used to update an online dashboard. The dashboard presents an estimate of the burden of ORADEs to frontline clinicians and hospital leadership and allows users to compare local data on ORADEs rates to other hospitals. Users are able to refine their search by surgery type or ORADE type. An interface was created, using national administrative claims data, to allow hospitals to access their ORADEs and benchmark local data against national trends.

Background Heart failure is the predominant cause of hospitalization and amongst the leading causes of death in Germany. However, accurate estimates of prevalence and incidence are lacking. Reported figures originating from different information sources are compromised by factors like economic reasons or documentation quality. Methods We implemented a clinical data warehouse that integrates various information sources (structured parameters, plain text, data extracted by natural language processing) and enables reliable approximations to the real number of heart failure patients. Performance of ICD-based diagnosis in detecting heart failure was compared across the years 2000–2015 with (a) advanced definitions based on algorithms that integrate various sources of the hospital information system, and (b) a physician-based reference standard. Results Applying these methods for detecting heart failure in inpatients revealed that relying on ICD codes resulted in a marked underestimation of the true prevalence of heart failure, ranging from 44% in the validation dataset to 55% (single year) and 31% (all years) in the overall analysis. Percentages changed over the years, indicating secular changes in coding practice and efficiency. Performance was markedly improved using search and permutation algorithms from the initial expert-specified query (F1 score of 81%) to the computer-optimized query (F1 score of 86%) or, alternatively, optimizing precision or sensitivity depending on the search objective. Conclusions Estimating prevalence of heart failure using ICD codes as the sole data source yielded unreliable results. Diagnostic accuracy was markedly improved using dedicated search algorithms. Our approach may be transferred to other hospital information systems.

International audience; AbstractAnaphylaxis is defined as a severe life-threatening generalized or systemic hypersensitivity reaction. The difficulty of coding anaphylaxis fatalities under the World Health Organization (WHO) International Classification of Diseases (ICD) system is recognized as an important reason for under-notification of anaphylaxis deaths. On current death certificates, a limited number of ICD codes are valid as underlying causes of death, and death certificates do not include the word anaphylaxis per se. In this review, we provide evidences supporting the need for changes in WHO mortality coding rules and call for addition of anaphylaxis as an underlying cause of death on international death certificates. This publication will be included in support of a formal request to the WHO as a formal request for this move taking the 11th ICD revision.

To efficiently identify persons with dementia to serve as a comparison group for a dementia care management program, we created and validated an automated electronic health record (EHR) dementia identification method. We used 3 data elements in the UCLA EHR that would indicate the presence of dementia: 1) ICD-9 code for dementia, 2) documentation of dementia medication use (cholinesterase inhibitors, memantine), and 3) natural language processing (NLP) of clinical notes for evidence of dementia. We examined the positive predictive value (PPV) of combinations of these 3 elements to identify dementia compared to physician medical record review. Physicians reviewed 60 records stratified by age (40–64, 65–84, 85+). The 3-element model had a PPV of 87%, but PPV for the 2-element models that included medications were lower (27% and 47%). Thus, we dropped medications from the method and reviewed an additional 64 cases identified only by NLP+ICD9. The NLP+ICD9 approach was less accurate among 40–64 year-olds (PPV 48%); however, when this younger group was excluded, the PPV was high (93%). By applying the NLP+ICD9 approach to 989 patients with verified dementia in a dementia management program, we estimated its sensitivity to be 63%. In summary, using a combination of ICD codes and NLP, we created a method to identify persons with dementia with a high PPV – especially among older patients – and reasonably high sensitivity. Advantages of an automated dementia identification system include readily identifying patients for care management programs, recruitment for research, or creation of observational cohorts and appropriate comparison groups.

Background and objective The volume of healthcare data is growing rapidly with the adoption of health information technology. We focus on automated ICD9 code assignment from discharge summary content and methods for evaluating such assignments. Methods We study ICD9 diagnosis codes and discharge summaries from the publicly available Multiparameter Intelligent Monitoring in Intensive Care II (MIMIC II) repository. We experiment with two coding approaches: one that treats each ICD9 code independently of each other (flat classifier), and one that leverages the hierarchical nature of ICD9 codes into its modeling (hierarchy-based classifier). We propose novel evaluation metrics, which reflect the distances among gold-standard and predicted codes and their locations in the ICD9 tree. Experimental setup, code for modeling, and evaluation scripts are made available to the research community. Results The hierarchy-based classifier outperforms the flat classifier with F-measures of 39.5% and 27.6%, respectively, when trained on 20 533 documents and tested on 2282 documents. While recall is improved at the expense of precision, our novel evaluation metrics show a more refined assessment: for instance, the hierarchy-based classifier identifies the correct sub-tree of gold-standard codes more often than the flat classifier. Error analysis reveals that gold-standard codes are not perfect, and as such the recall and precision are likely underestimated. Conclusions Hierarchy-based classification yields better ICD9 coding than flat classification for MIMIC patients. Automated ICD9 coding is an example of a task for which data and tools can be shared and for which the research community can work together to build on shared models and advance the state of the art.

The aim of this study was to estimate the prevalence of physician-diagnosed and registered chronic hepatitis C (CHC), and to estimate the reported frequencies of Charlson comorbidities compared with matched comparators from the general population.Patients were identified according to ICD codes for CHC in the Swedish National Patient Register (1997-2013). Prevalence was estimated according to different patient identification algorithms and for different subgroups. Charlson comorbidities were ascertained from the same register and compared with age/sex/county of residence matched general population comparators.A total of 34,633 individuals with physician-diagnosed CHC were alive in Sweden in 2013 (mean age, 49 years; 64% men), corresponding to a physician-diagnosed prevalence of 0.36%. The prevalence varied by case definition (0.22%-0.36%). The estimate dropped to 0.14% for monitored CHC disease (defined as ≥1 CHC-related visit in 2013). Overall, 41.3% of the CHC patients had ≥1 physician-registered Charlson comorbidity; the most common was liver diseases (22.1%). Compared with matched comparators from the general population (n = 171,338), patients with CHC had more physician-diagnosed and registered diseases such as chronic pulmonary disease (10.2% vs. 4.0%), diabetes (10.6% vs. 5.5%) and liver-related cancer (1.3% vs. 0.2%; all p .01). No information on behavioural factors, such as smoking, alcohol consumption or on-going illicit drug use, was available.The physician-diagnosed prevalence of CHC was slightly lower than previously reported estimates, and varied by case definition. The additional comorbidities observed in the CHC group should be taken into consideration, as these comorbidities add to the disease burden.

Objective: To identify early diagnostic profiles such as diagnostic codes and consultation patterns of cancer patients in primary care one year prior to cancer diagnosis. Design: Total population-based case–control study. Setting and subjects: 4562 cancer patients and 17,979 controls matched by age, sex, and primary care unit. Data were collected from the Swedish Cancer Register and the Regional Healthcare Database. Method: We identified cancer patients in the Västra Götaland Region of Sweden diagnosed in 2011 with prostate, breast, colorectal, lung, gynaecological, and skin cancers including malignant melanoma. We studied the symptoms and diagnoses identified by diagnostic codes during a diagnostic interval of 12 months before the cancer diagnosis. Main outcome measures: Consultation frequency, symptom density by cancer type, prevalence and odds ratios (OR) for the diagnostic codes in the cancer population as a whole. Results: The diagnostic codes with the highest OR were unspecified lump in breast, neoplasm of uncertain behaviour, and abnormal serum enzyme levels. The codes with the highest prevalence were hyperplasia of prostate, other skin changes and abdominal and pelvic pain. The frequency of diagnostic codes and consultations in primary care rose in tandem 50 days before diagnosis for breast and gynaecological cancer, 60 days for malignant melanoma and skin cancer, 80 days for prostate cancer and 100 days for colorectal and lung cancer. Conclusion: Eighty-seven percent of patients with the most common cancers consulted a general practitioner (GP) a year before their diagnosis. An increase in consultation frequency and presentation of any symptom should raise the GP’s suspicion of cancer. Key Points Knowledge about the prevalence of early symptoms and other clinical signs in cancer patients in primary care remains insufficient. Eighty-seven percent of the patients with the seven most common cancers consulted a general practitioner 12 months prior to cancer diagnosis. Both the frequency of consultation and the number of symptoms and diseases expressed in diagnostic codes rose in tandem 50–100 days before the cancer diagnosis. Unless it is caused by a previously known disease, an increased consultation rate for any symptom should result in a swift investigation or referral from primary care to confirm or exclude cancer.

Background Kawasaki disease (KD) is a systemic vasculitis with an acute and self-limited course. The incidence of KD differs widely among ethnic groups and is higher in the Asian population. In Italy, no recent data are available. Our purpose is to define the epidemiology of Kawasaki disease in the years 2008–2013 in children aged

Background The international classification of diseases version 10 (ICD-10) uses alphanumeric expanded codes and external cause of injury codes (E-codes). Objective To examine the reliability and validity of emergency department (ED) coders in applying E-codes in ICD-9 and -10. Methods Bicycle and pedestrian injuries were identified from the ED information system from one period before and two periods after transition from ICD-9 to -10 coding. Overall, 180 randomly selected bicycle and pedestrian injury charts were reviewed as the reference standard (RS). Original E-codes assigned by ED coders (ICD-9 in 2001 and ICD-10 in 2004 and 2007) were compared with charts (validity) and also to ICD-9 and -10 codes assigned from RS chart review, to each case by an independent (IND) coder (reliability). Sensitivity, specificity, simple, and chance-corrected agreements (κ statistics) were calculated. Results Sensitivity of E-coding bicycle injuries by the IND coder in comparison with the RS ranged from 95.1% (95% CI 86.3 to 99.0) to 100% (95% CI 94.0 to 100.0) for both ICD-9 and -10. Sensitivity of ED coders in E-coding bicycle injuries ranged from 90.2% (95% CI 79.8 to 96.3) to 96.7% (95% CI 88.5 to 99.6). The sensitivity estimates for the IND coder ranged from 25.0% (95% CI 14.7 to 37.9) to 45.0% (95% CI 32.1 to 58.4) for pedestrian injuries for both ICD-9 and -10. Conclusion Bicycle injuries are coded in a reliable and valid manner; however, pedestrian injuries are often miscoded as falls. These results have important implications for injury surveillance research.

Objective. National estimates of hospi- talizations for rotavirus, the leading cause of acute gas- troenteritis (AGE) in children, have been used to estab- lish the need for rotavirus vaccines. A previous method directly estimated discharges by using the rotavirus-spe- cific International Classification of Diseases (ICD) code, but this method has not been validated. Our study eval- uated the sensitivity of the rotavirus ICD code among children hospitalized for AGE by using active surveil- lance for rotavirus at a tertiary children's hospital. Design. We compared data for rotavirus-coded hospi- tal discharges in 2000-2001 at Cincinnati Children's Hos- pital Medical Center with data on laboratory-confirmed cases of rotavirus obtained from active surveillance. We estimated additional rotavirus hospitalizations by ex- trapolating the proportion of rotavirus-positive results from active-surveillance cases to those with an unknown rotavirus status. Results. Of 767 cases of AGE-related discharge codes, 103 (13%) were coded as rotavirus, 91% (94 of 103) of which were laboratory-confirmed diagnoses. Among all children discharged with an AGE-related illness, 260 (34%) were enrolled in active surveillance, of whom 155 (60%) tested positive for rotavirus. An additional 47 lab- oratory-confirmed rotavirus-case patients not enrolled in active surveillance yielded a total of 202 rotavirus cases and a maximum sensitivity of the rotavirus code of 47%. Extrapolation indicated that an additional 170 untested children might be rotavirus-positive, yielding a total of 372 rotavirus hospitalizations and a minimum sensitivity of the rotavirus code of 25%. Conclusions. Measurement of rotavirus-coded hospi- tal discharges alone seems to greatly underestimate the true burden of rotavirus-associated hospitalizations. The numbers of national rotavirus hospitalization discharges may be substantially greater than previously estimated. Pediatrics 2005;115:78-82; acute gastroenteritis, United States, child, surveillance, hospitalization, ICD codes, sensitivity, specificity.

This study uses 61 years of death certificates for Bexar County, Texas, uniformly coded under ICD 9, to describe the transition in heart disease mortality from 1935-1995. We find that life expectancy for persons dying with heart diseases increased throughout this period, with clear differences in rates of increase for males and females, associated with acute ischemic heart disease. Our data point to an epidemic of AIHD in the 1950s and 1960s, which is now abating. Findings are less clear for chronic ischemic heart disease, while other major heart diseases cannot be traced with any confidence owing to changes over time in the emphasis accorded particular causes. Our findings suggest caution with respect to the socioeconomic analysis of heart disease mortality data, particularly where the instability of the coding conventions has been most acute.

Summary The World Health Organization (WHO) International Classification of Diseases (ICD) has been used to classify causes of morbidity and mortality such as epilepsy for more than 50 years. The aims of this critical commentary are to do the following: (1) Introduce the ICD classification, summarize the ICD-9 and ICD-10 codes for epilepsy and seizures, and discuss the challenges of mapping epilepsy codes between these two versions; (2) discuss how the ICD-9 and ICD-10 relate to the revised International League Against Epilepsy (ILAE) terminology and concepts for classification of seizures and epilepsies; (3) discuss how ICD-coded data have been used for epilepsy care and research and briefly examine the potential impact of the international ICD-10 clinical modifications on research; (4) discuss the upcoming ICD-11 codes and the role of the epilepsy community in their development; and (5) discuss how the ICD-11 will conform more closely to the current ILAE terminology and classification of the epilepsies and seizures and its potential impact on clinical care, surveillance, and public health and research.

AIMS: Alcohol can induce diverse serious pathologies, yet this complexity may be obscured when alcohol-related deaths are classified according to a single underlying cause. We sought to quantify this issue and its implications for analysing mortality data. DESIGN, SETTING AND PARTICIPANTS: Cross-sectional study included 554 men aged 25-54 in Estonia undergoing forensic autopsy in 2008-09. MEASUREMENTS: Potentially alcohol-related pathologies were identified following macroscopic and histological examination. Alcohol biomarkers levels were determined. For a subset (26%), drinking behaviour was provided by next-of-kin. The Estonian Statistics Office provided underlying cause of death. FINDINGS: Most deaths (75%) showed evidence of potentially alcohol-related pathologies, and 32% had pathologies in two or more organs. The liver was most commonly affected [60.5%, 95% confidence interval (CI) = 56.3-64.6] followed by the lungs (18.6%, 95% CI = 15.4-22.1), stomach (17.5%, 95% CI = 14.4-20.9), pancreas (14.1%, 95% CI = 11.3-17.3), heart (4.9%, 95% CI = 3.2-7.0) and oesophagus (1.4%, 95% CI = 0.6-2.8). Only a minority with liver pathology had a second pathology. The number of pathologies correlated with alcohol biomarkers (phosphatidylethanol, gamma-glytamyl transpeptidase in blood, ethylglucuronide, ethylsulphate in urine). Despite the high prevalence of liver pathology, few deaths had alcoholic liver disease specified as the underlying cause. CONCLUSION: The majority of 554 men aged 25-54 undergoing forensic autopsy in Estonia in 2008-09 showed evidence of alcohol-related pathology. However, the recording of deaths by underlying cause failed to capture the scale and nature of alcohol-induced pathologies found.

Background Administrative data are increasingly used to conduct research on depression and inform health services and health policy. Depression surveillance using administrative data is an alternative to surveys, which can be more resource-intensive. The objectives of this study were to: (1) systematically review the literature on validated case definitions to identify depression using International Classification of Disease and Related Health Problems (ICD) codes in administrative data and (2) identify individuals with and without depression in administrative data and develop an enhanced case definition to identify persons with depression in ICD-coded hospital data. Methods (1) Systematic review: We identified validation studies using ICD codes to indicate depression in administrative data up to January 2013. (2) Validation: All depression case definitions from the literature and an additional three ICD-9-CM and three ICD-10 enhanced definitions were tested in an inpatient database. The diagnostic accuracy of all case definitions was calculated [sensitivity (Se), specificity (Sp), positive predictive value (PPV) and negative predictive value (NPV)]. Results (1) Systematic review: Of 2,014 abstracts identified, 36 underwent full-text review and three met eligibility criteria. These depression studies used ICD-9 and ICD-10 case definitions. (2) Validation: 4,008 randomly selected medical charts were reviewed to assess the performance of new and previously published depression-related ICD case definitions. All newly tested case definitions resulted in Sp >99%, PPV >89% and NPV >91%. Sensitivities were low (28-35%), but higher than for case definitions identified in the literature (1.1-29.6%). Conclusions Validating ICD-coded data for depression is important due to variation in coding practices across jurisdictions. The most suitable case definitions for detecting depression in administrative data vary depending on the context. For surveillance purposes, the most inclusive ICD-9 & ICD-10 case definitions resulted in PPVs of 89.7% and 89.5%, respectively. In cases where diagnostic certainty is required, the least inclusive ICD-9 and −10 case definitions are recommended, resulting in PPVs of 92.0% and 91.1%. All proposed case definitions resulted in suboptimal levels of sensitivity (ranging from 28.9%-35.6%). The addition of outpatient data (such as pharmacy records) for depression surveillance is recommended and should result in improved measures of validity. Electronic supplementary material The online version of this article (doi:10.1186/s12888-014-0289-5) contains supplementary material, which is available to authorized users.

New Zealand mortality records for the years 1980 to 1993 were analysed to estimate the aggregate burden of infectious disease using a recoding of ICD-9 codes to identify deaths with infectious aetiology. The recoding scheme was modified from one developed by US CDC, which used expert panels to assign ICD codes to categories dependent on the proportion of the code attributable to infection. ICD-9 Chapter One ('Infectious and parasitic diseases') accounted for only 0.7% of total deaths. Following recoding, this proportion increased tenfold, with 6.9% of deaths attributable to infectious disease. This proportion was stable or declined only slowly between 1980 and 1993. While rates varied by age, gender and ethnicity, the results indicate that infectious disease still accounts for a substantial proportion of the burden of disease in New Zealand.

Background: The aim of the present study was to evaluate the association between heart failure (HF) and chronic kidney disease (CKD) in tertiary care centers using the clinical records of patients enrolled in internal medicine departments.Patients and methods: We used the clinical records of 1380 elderly patients to identify patients with a history of HF and CKD using admission ICD codes and glomerular filtration rate (GFR) formulas. Magnitude and strength of such associations were investigated by univariable and multivariable analysis.Results: Of the 1380 patients enrolled, 27.9% had HF (age 80 ± 7, BMI 27 ± 6 kg/m2) and 17.4% CKD (age 81 ± 7, BMI 26.8 ± 6 kg/m2). Both groups were significantly older (P

Background: Over the last two decades, mortality from coronary heart disease (CHD) and cerebrovascular disease (CVD) declined by about 30% in the European Union (EU). Design: We analyzed trends in CHD (X ICD codes: 120-125) and CVD (X ICD codes: 160-169) mortality in young adults (age 35-44 years) in the EU as a whole and in 12 selected European countries, over the period 1980-2007. Methods: Data were derived from the World Health Organization mortality database. With joinpoint regression analysis, we identified significant changes in trends and estimated average annual percent changes (AAPC). Results: CHD mortality rates at ages 35-44 years have decreased in both sexes since the 1980s for most countries, except for Russia (130/100,000 men and 24/100,000 women, in 2005-7). The lowest rates (around 9/100,000 men, 2/100,000 women) were in France, Italy and Sweden. In men, the steepest declines in mortality were in the Czech Republic (AAPC = -6.1%), the Netherlands (-5.2%), Poland (-4.5%), and England and Wales (-4.5%). Patterns were similar in women, though with appreciably lower rates. The AAPC in the EU was -3.3% for men (rate = 16.6/100,000 in 2005-7) and -2.1% for women (rate = 3.5/100,000). For CVD, Russian rates in 2005-7 were 40/100,000 men and 16/100,000 women, 5 to 10-fold higher than in most western European countries. The steepest declines were in the Czech Republic and Italy for men, in Sweden and the Czech Republic for women. The AAPC in the EU was -2.5% in both sexes, with steeper declines after the mid-late 1990s (rates = 6.4/100,000 men and 4.3/100,000 women in 2005-7). Conclusions: CHD and CVD mortality steadily declined in Europe, except in Russia, whose rates were 10 to 15-fold higher than those of France, Italy or Sweden. Hungary and Poland, and also Scotland, where CHD trends were less favourable than in other western European countries, also emerge as priorities for preventive interventions.

Comparison of mortality and morbidity is a commonly used method in health related studies. The International Classification of Disease (ICD) consists of thousands of codes for classifying cause of death and disease categories. A grouping scheme is needed to cluster related categories into a meaningful and manageable number for comparative purposes. Different kinds of grouping schemes have been used; nevertheless, little is known about the comparability among different grouping schemes. In this study, we compared seven grouping schemes; five for mortality and two for morbidity. We found poor comparability between different grouping schemes. Different schemes covered different ranges of codes. Some schemes used the same title, but included different ranges of codes. Features of newly developed grouping schemes were to group disease categories of similar characteristics across traditional ICD chapters and to group disease categories based on health care needs, instead of those based merely on etiology or organ system. Different grouping schemes were developed to reveal the unique mortality and morbidity pattern of different geographical areas. Different grouping logic was used by different grouping schemes. Therefore, it is difficult to make a good comparison between different schemes. An investigator tabulating the mortality or morbidity figures based on a given grouping scheme should explicitly define the exact ICD codes included. Any user of data derived from different grouping schemes, especially for comparisons between countries, should be cautious about the comparability problems.