Effective Identification of Similar Patients Through Sequential Matching over ICD Code Embedding.

Evidence-based medicine often involves the identification of patients with similar conditions, which are often captured in ICD (International Classification of Diseases (World Health Organization 2013)) code sequences. With no satisfying prior solutions for matching ICD-10 code sequences, this paper presents a method which effectively captures the clinical similarity among routine patients who have multiple comorbidities and complex care needs. Our method leverages the recent progress in representation learning of individual ICD-10 codes, and it explicitly uses the sequential order of codes for matching. Empirical evaluation on a state-wide cancer data collection shows that our proposed method achieves significantly higher matching performance compared with state-of-the-art methods ignoring the sequential order. Our method better identifies similar patients in a number of clinical outcomes including readmission and mortality outlook. Although this paper focuses on ICD-10 diagnosis code sequences, our method can be adapted to work with other codified sequence data.

Proportion of patients with cancer among high-cost Medicare beneficiaries: Who they are and what drives their spending.

BACKGROUND: A small proportion of patients account for the majority of health care spending. Of this group, little is known about what proportion have a cancer diagnosis and how their spending pattern compares to those without cancer. METHODS: Using national Medicare data of enrollees 65 or older, we identified patients in the top decile of spending in 2014 and designated them as high-cost. We used ICD-9 codes to identify patients with a cancer diagnosis and examined cancer prevalence among both high-cost and non-high-cost patients. We examined patterns of spending for high-cost patients with and without cancer. RESULTS: While 14.8% of all Medicare beneficiaries have a cancer diagnosis, we found that the prevalence of a cancer diagnosis was much higher among high-cost patients (32.5% versus 12.9% of non-high-cost patients). Thus, having a cancer diagnosis was associated with a 3.1 times greater odds of being high-cost, even after accounting for age (odds ratio 3.09, 95% CI 3.07-3.11; P < 0.001). High-cost patients with cancer had higher total annual spending than high-cost patients without cancer ($66,685 vs. $59,427; p < 0.0001); costs among high-cost cancer patients were driven by greater use of outpatient treatments (19.2% of total spending vs. 13.6% among non-cancer high-cost patients, p < 0.0001) and more prescription drugs (11.9% vs. 9.9%; p < 0.0001). CONCLUSIONS: There is a high prevalence of cancer diagnoses among high-cost Medicare patients. IMPLICATIONS: Programs that target high-cost patients may need to customize interventions based on whether the patient has a cancer diagnosis.

Exploring the relationship between nonalcoholic fatty liver disease and pancreatic cancer by computed tomographic survey.

Pancreatic cancer is difficult to diagnose in an early stage, and has the highest mortality of all types of cancer. Obesity, high body mass index, and increased abdominal girth are established risk factors. Some studies have postulated that there is a correlation between organ steatosis and pancreatic cancer. This study aims to explore whether nonalcoholic fatty liver disease (NAFLD) is a risk factor and a prognostic factor for pancreatic cancer. The study enrolled 557 patients (143 with and 414 without pancreatic cancer) who were diagnosed between January 2009 and December 2013. We reviewed the abdominal computed tomographic scans of the patients to confirm the diagnosis of NAFLD. Clinical parameters, laboratory data, and personal information were analyzed. NAFLD is an independent risk factor for pancreatic cancer according to adjusted multivariate logistic regression analysis (OR 2.63, 95% CI 1.24-5.58, p = 0.011). The Kaplan-Meier survival curve reveals that patients without NAFLD have longer survival than patients with NAFLD (p = 0.005, log-rank test). NAFLD is positively correlated with pancreatic cancer, a result suggesting that NAFLD may increase the incidence and risk of pancreatic cancer. Patients with pancreatic cancer and NAFLD have poorer overall survival than patients without NAFLD, perhaps, because dysregulated cytokine status leads to progression of pancreatic cancer. NAFLD may be a prognostic factor for pancreatic cancer.

The Value of Electronically Extracted Data for Auditing Outpatient Antimicrobial Prescribing.

OBJECTIVE The optimal approach to auditing outpatient antimicrobial prescribing has not been established. We assessed how different types of electronic data-including prescriptions, patient-visits, and International Classification of Disease, Tenth Revision (ICD-10) codes-could inform automated antimicrobial audits. DESIGN Outpatient visits during 2016 were retrospectively reviewed, including chart abstraction, if an antimicrobial was prescribed (cohort 1) or if the visit was associated with an infection-related ICD-10 code (cohort 2). Findings from cohorts 1 and 2 were compared. SETTING Primary care clinics and the emergency department (ED) at the Iowa City Veterans Affairs Medical Center. RESULTS In cohort 1, we reviewed 2,353 antimicrobial prescriptions across 52 providers. ICD-10 codes had limited sensitivity and positive predictive value (PPV) for validated cases of cystitis and pneumonia (sensitivity, 65.8%, 56.3%, respectively; PPV, 74.4%, 52.5%, respectively). The volume-adjusted antimicrobial prescribing rate was 13.6 per 100 ED visits and 7.5 per 100 primary care visits. In cohort 2, antimicrobials were not indicated in 474 of 851 visits (55.7%). The antimicrobial overtreatment rate was 48.8% for the ED and 59.7% for primary care. At the level of the individual prescriber, there was a positive correlation between a provider's volume-adjusted antimicrobial prescribing rate and the individualized rates of overtreatment in both the ED (r=0.72; P<.01) and the primary care setting (r=0.82; P=0.03). CONCLUSIONS In this single-center study, ICD-10 codes had limited sensitivity and PPV for 2 infections that typically require antimicrobials. Electronically extracted data on a provider's rate of volume-adjusted antimicrobial prescribing correlated with the frequency at which unnecessary antimicrobials were prescribed, but this may have been driven by outlier prescribers. Infect Control Hosp Epidemiol 2018;39:64-70.

The impact of the 2006 Massachusetts health care reform law on spine surgery patient payer-mix status and age.

OBJECTIVE Several similarities exist between the Massachusetts health care reform law of 2006 and the Affordable Care Act (ACA). The authors' prior neurosurgical research showed a decrease in uninsured surgeries without a significant change in surgical volume after the Massachusetts reform. An analysis of the payer-mix status and the age of spine surgery patients, before and after the policy, should provide insight into the future impact of the ACA on spine surgery in the US. METHODS Using the Massachusetts State Inpatient Database and spine ICD-9-CM procedure codes, the authors obtained demographic information on patients undergoing spine surgery between 2001 and 2012. Payer-mix status was assigned as Medicare, Medicaid, private insurance, uninsured, or other, which included government-funded programs and workers' compensation. A comparison of the payer-mix status and patient age, both before and after the policy, was performed. The New York State data were used as a control. RESULTS The authors analyzed 81,821 spine surgeries performed in Massachusetts and 248,757 in New York. After 2008, there was a decrease in uninsured and private insurance spine surgeries, with a subsequent increase in the Medicare and "other" categories for Massachusetts. Medicaid case numbers did not change. This correlated to an increase in surgeries performed in the age group of patients 65-84 years old, with a decrease in surgeries for those 18-44 years old. New York showed an increase in all insurance categories and all adult age groups. CONCLUSIONS After the Massachusetts reform, spine surgery decreased in private insurance and uninsured categories, with the majority of these surgeries transitioning to Medicare. Moreover, individuals who were younger than 65 years did not show an increase in spine surgeries, despite having greater access to health insurance. In a health care system that requires insurance, the decrease in private insurance is primarily due to an increasing elderly population. The Massachusetts model continues to show that this type of policy is not causing extreme shifts in the payer mix, and suggests that spine surgery will continue to thrive in the current US health care system.

Variability in Standard Outcomes of Posterior Lumbar Fusion Determined by National Databases.

OBJECTIVE: National databases are used with increasing frequency in spine surgery literature to evaluate patient outcomes. The differences between individual databases in relationship to outcomes of lumbar fusion are not known. We evaluated the variability in standard outcomes of posterior lumbar fusion between the University HealthSystem Consortium (UHC) database and the Healthcare Cost and Utilization Project National Inpatient Sample (NIS). METHODS: NIS and UHC databases were queried for all posterior lumbar fusions (International Classification of Diseases, Ninth Revision code 81.07) performed in 2012. Patient demographics, comorbidities (including obesity), length of stay (LOS), in-hospital mortality, and complications such as urinary tract infection, deep venous thrombosis, pulmonary embolism, myocardial infarction, durotomy, and surgical site infection were collected using specific International Classification of Diseases, Ninth Revision codes. RESULTS: Analysis included 21,470 patients from the NIS database and 14,898 patients from the UHC database. Demographic data were not significantly different between databases. Obesity was more prevalent in UHC (P = 0.001). Mean LOS was 3.8 days in NIS and 4.55 in UHC (P < 0.0001). Complications were significantly higher in UHC, including urinary tract infection, deep venous thrombosis, pulmonary embolism, myocardial infarction, surgical site infection, and durotomy. In-hospital mortality was similar between databases. CONCLUSIONS: NIS and UHC databases had similar demographic patient populations undergoing posterior lumbar fusion. However, the UHC database reported significantly higher complication rate and longer LOS. This difference may reflect academic institutions treating higher-risk patients; however, a definitive reason for the variability between databases is unknown. The inability to precisely determine the basis of the variability between databases highlights the limitations of using administrative databases for spinal outcome analysis.

The reporting and diagnosis of uterine fibroids in the UK: an observational study.

BACKGROUND: Uterine fibroids (UFs) are the most common benign tumour in women, and many undergo hysterectomy or uterus-preserving procedures (UPPs) to manage their symptoms. We aimed to validate the recording of UFs in a primary care database, The Health Improvement Network (THIN), and to determine the incidence of UFs in the UK. METHODS: In this observational study, women in THIN aged 15-54 years between January 2000 and December 2009 with no previous record of UFs, hysterectomy or UPPs were identified. Individuals were followed up until there was a Read code indicating UFs, they reached 55 years of age or died, or the study ended. Among those without a UF code, women were identified with a code for hysterectomy, UPPs or heavy menstrual bleeding (HMB). Anonymized patient profiles from each category were randomly selected and reviewed. Subsequently, primary care physicians were asked to complete questionnaires to verify the diagnosis for a randomly selected subgroup. RESULTS: In total, 737,638 women were identified who met the initial inclusion criteria. The numbers of women with a code for UFs, hysterectomy, UPPs and HMB were 9380, 11,002, 3220 and 60,915, respectively; the proportions of confirmed cases of UFs were 88.8, 29.7, 57.7 and 15.9 %. The estimated number of women with UFs was 23,140 (64.0 % without a recorded UF diagnosis). The overall incidence of UFs was 5.8 per 1000 woman-years. CONCLUSIONS: UFs were confirmed in a high proportion of women with UF Read codes. However, almost two-thirds of cases were identified among women with a code for hysterectomy, UPPs or HMB. These results show that UFs are under-recorded in UK primary care, and suggest that primary care physicians tend to code the symptoms of UFs more often than the diagnosis.

[Sigmoid colon diverticulitis : Treatment modalities 2011-2013]. / Sigmadivertikulitis : Behandlungsmodalitäten 2011 bis 2013.

INTRODUCTION: Diverticulosis is a relevant disease in Germany with a prevalence of over 60 % in patients aged ≥70 years. The S2k guidelines for the treatment of diverticulosis were recently published. Systematic epidemiological data on treatment modalities do not exist. METHODS: Analysis of in-hospital treatment modalities for diverticulosis based on data from the Federal Office of Statistics. RESULTS: Approximately 130,000 inpatient cases of diverticulosis are treated in Germany per year. Approximately 25 % undergo surgery and of these slightly under 50 % (12,000 procedures) are carried out by laparoscopy. The complication rates are 18 % in a best case scenario and up to 85 % in a worst case scenario. A stage-adjusted classification of treatment modalities based on data from the Federal Office of Statistics is currently practically impossible. CONCLUSION: To enable stage-adjusted epidemiological analysis of diverticulosis, a standardized and transparent documentation system enabling systematic analysis is necessary, which does not currently exist (e. g. ICD 10 coding); moreover, information on conservative and interventional treatment options are not included in the operations and procedures key (OPS) coding system.

The Prevalence and Incidence of Diabetes in Germany.

BACKGROUND: The database of the German Institute of Medical Documentation and Information makes it possible for the first time to compute statistics on diabetes for all insurees of the statutory health insurance scheme in Germany. Data from this comprehensive source are less likely to be biased by differences in the membership structures of individual insurance carriers or by the underrepresentation of persons over age 80 that is seen in most population-based studies. METHODS: International Classification of Diseases (ICD)-coded diagnosis data from the inpatient and outpatient sectors were used to define persons as having diabetes. Incidences were estimated from differences in prevalence from one year to the next and the expected mortality of persons with and without diabetes. RESULTS: A diabetes diagnosis was present in 6.4 million out of a total of 65.6 million insurees in 2009 and in 6.7 million out of 64.9 million insures in 2010. The corresponding age and sex standardized prevalences of diabetes were 9.7% in 2009 and 9.9% in 2010, respectively. The number of persons with type 2 diabetes was 4.6 million in 2009 and 4.7 million in 2010. The prevalence and incidence of type 2 diabetes rose steeply from age 50 to age 80. Peak incidence was at age 85, with 24 newly diagnosed cases of diabetes per 1000 person-years. CONCLUSION: On the basis of these data, we estimate that 5.8 million persons with type 2 diabetes are living in Germany today. The database used in this study is a valuable complement to population-based studies for monitoring the prevalence of diabetes, particularly in persons over age 80.

Validation of Case Finding Algorithms for Hepatocellular Cancer From Administrative Data and Electronic Health Records Using Natural Language Processing.

BACKGROUND: Accurate identification of hepatocellular cancer (HCC) cases from automated data is needed for efficient and valid quality improvement initiatives and research. We validated HCC International Classification of Diseases, 9th Revision (ICD-9) codes, and evaluated whether natural language processing by the Automated Retrieval Console (ARC) for document classification improves HCC identification. METHODS: We identified a cohort of patients with ICD-9 codes for HCC during 2005-2010 from Veterans Affairs administrative data. Pathology and radiology reports were reviewed to confirm HCC. The positive predictive value (PPV), sensitivity, and specificity of ICD-9 codes were calculated. A split validation study of pathology and radiology reports was performed to develop and validate ARC algorithms. Reports were manually classified as diagnostic of HCC or not. ARC generated document classification algorithms using the Clinical Text Analysis and Knowledge Extraction System. ARC performance was compared with manual classification. PPV, sensitivity, and specificity of ARC were calculated. RESULTS: A total of 1138 patients with HCC were identified by ICD-9 codes. On the basis of manual review, 773 had HCC. The HCC ICD-9 code algorithm had a PPV of 0.67, sensitivity of 0.95, and specificity of 0.93. For a random subset of 619 patients, we identified 471 pathology reports for 323 patients and 943 radiology reports for 557 patients. The pathology ARC algorithm had PPV of 0.96, sensitivity of 0.96, and specificity of 0.97. The radiology ARC algorithm had PPV of 0.75, sensitivity of 0.94, and specificity of 0.68. CONCLUSIONS: A combined approach of ICD-9 codes and natural language processing of pathology and radiology reports improves HCC case identification in automated data.

[Incidence and Costs of 1:1 Care in Psychiatric Hospitals in Germany - A Descriptive Analysis Based on the VIPP Project Data Set]. / Häufigkeit und Kosten der 1:1-Betreuung in der Erwachsenenpsychiatrie in Deutschland - Eine deskriptive Analyse anhand des VIPP-Datensatzes.

OBJECTIVE: 1:1 care is applied for patients requiring close psychiatric monitoring and care like patients with acute suicidality. The article describes the frequency of 1:1 care across different diagnoses and age groups in German psychiatric hospitals. METHODS: The analysis was based on the VIPP Project from the years 2011 and 2012. A total of 47 hospitals with more than 120,000 cases were included. Object of the analysis was the OPS code 9-640.0 1:1 care. The evaluation was performed on case level. RESULTS: Data of 47 hospitals were included. Of the 121,454 cases evaluated in 2011 3.8 % documented a 1:1 care within the meaning of OPS 9-640.0 additional code. Of the 66 245 male cases a 1:1 care was documented in 3.5 % and the 55 207 female cases was 4.1 %. Compared to 2011, the proportion of 1:1 care in 2012 rose to 4.8 %. CONCLUSION: The results show that 1:1 care is frequently applied in German psychiatric hospitals. The Data of the VIPP project have proven to be a useful tool to gain information on the frequency of cost-intensive interventions in German psychiatric hospitals. Further analyses should create the possibility of evaluation at the level of the individual codes.

Charts versus Discharge ICD-10 Coding for Sternal Wound Infection Following Coronary Artery Bypass Grafting.

BACKGROUND: Sternal wound infection (SWI) in patients undergoing coronary artery bypass grafting (CABG) can carry a significant risk of morbidity and mortality. The objective of this work is to describe the methods used to identify cases of SWI in an administrative database and to demonstrate the effectiveness of using an International Classification of Diseases, Tenth Revision (ICD-10) coding algorithm for this purpose. METHODS: ICD-10 codes were used to identify cases of SWI within one year of CABG between April 2002 and November 2009. We randomly chose 200 charts for detailed chart review (100 from each of the groups coded as having SWI and not having SWI) to determine the utility of the ICD-10 coding algorithm. RESULTS: There were 2,820 patients undergoing CABG. Of these, 264 (9.4 percent) were coded as having SWI. Thirty-eight cases of SWI were identified by chart review. The ICD-10 coding algorithm of T81.3 or T81.4 was able to identify incident SWI with a positive predictive value of 35 percent and a negative predictive value of 97 percent. The agreement between the ICD-10 coding algorithm and presence of SWI remained fair, with an overall kappa coefficient of 0.32 (95 percent confidence interval, 0.22-0.43). The effectiveness of identifying deep SWI cases is also presented. CONCLUSIONS: This article describes an effective algorithm for identifying a cohort of patients with SWI following open sternotomy in large databases using ICD-10 coding. In addition, alternative search strategies are presented to suit researchers' needs.

Evaluation of an ICD-10 algorithm to detect osteonecrosis of the jaw among cancer patients in the Danish National Registry of Patients.

PURPOSE: This study aimed to validate a predefined algorithm for osteonecrosis of the jaw (ONJ) among cancer patients in the Danish National Registry of Patients and to assess the nature of clinical information recorded in medical charts of ONJ patients. METHODS: We identified potential ONJ cases recorded in 2005-2010 among cancer patients at the hospital Departments of Oral and Maxillofacial Surgery (DOMS) in three Danish regions, using a set of codes from the International Classification of Diseases, 10th revision (ICD-10). We abstracted DOMS charts of the potential cases, had the ONJ status adjudicated by an expert ONJ adjudication committee (ONJAC), and computed positive predictive values. For patients with ONJAC-confirmed ONJ, we abstracted the charts for information on ONJ clinical course. Sensitivity of the algorithm was computed using a separate sample of 101 known ONJ cases accrued in 2005-2011. RESULTS: We identified 212 potential ONJ cases, of which 197 (93%) had charts available for abstraction. Eighty-three potential cases were confirmed by ONJAC, with a positive predictive value of 42% (95% confidence interval [CI] 35%-49%). DOMS charts of these 83 cases contained complete information on ONJ clinical course. Information about antiresorptive treatment was recorded for 84% of the patients. Among the 101 known ONJ cases, 74 had at least one prespecified ICD-10 code recorded in the Danish National Registry of Patients within ±90 days of the ONJ diagnosis (sensitivity 73%; 95%CI [64%-81%]). CONCLUSIONS: The predefined algorithm is not adequate for monitoring ONJ in pharmacovigilance studies. Additional case-finding approaches, coupled with adjudication, are necessary to estimate ONJ incidence accurately.

Characterization of inflammatory bowel disease in elderly hospitalized patients in a large central Canadian Health region.

OBJECTIVE: To determine differences in phenotype and treatment among hospitalized elderly and young patients with inflammatory bowel disease (IBD), and the utility of International Classification of Diseases, 10th Revision (ICD)-10 codes in hospital discharge abstracts in diagnosing IBD in elderly patients. METHODS: A large Canadian health region hospitalization discharge database was used to identify elderly (>65 years of age) and young (19 to 50 years of age) patients with IBD admitted between April 1, 2007 and March 31, 2012, and a random sample of elderly patients with other colonic conditions. Medical records were reviewed to confirm IBD diagnosis and extract clinical information. The characteristics of elderly versus young hospitalized IBD patients and accuracy of ICD-10 IBD discharge codes in the elderly were assessed. RESULTS: One hundred forty-three elderly and 82 young patients with an IBD discharge diagnosis, and 135 elderly patients with other gastrointestinal discharge diagnoses were included. Elderly IBD patients were less likely to have ileocolonic Crohn disease (21.4% versus 50.9%; P=0.001), more likely to be prescribed 5-aminosalicylates (61% versus 43%; P=0.04), and less likely to be prescribed biologics (6% versus 21%; P=0.016) or immunomodulators (21% versus 42%; P=0.01). The sensitivity, specificity and positive predictive value of a single ICD code for CD were 98%, 96% and 94%, respectively, and for ulcerative colitis (UC) were 98%, 92% and 70%, respectively. CONCLUSIONS: Treatment approaches in elderly patients were different than in younger IBD patients despite having disease sufficiently severe to require hospitalization. While less accurate in UC, a single ICD-10 IBD code was sufficient to identify elderly CD and UC hospitalized patients.

[The issue of quality of Russian cause-specific mortality statistics at old ages].

The paper presents a brief analysis on the quality of cause-specific mortality statistics at old ages in Russia. Two aspects are in focus: the regional differences in approaches for choosing the underlying cause of death and the comparison of Russian coding practice with the practices of some other European countries. The results obtained indicate that there are certain problems with the quality of cause-of-death coding at old ages in Russia. No unified approach to coding deaths as caused by «Senility¼ is used at a sub-national level. This leads to the distortion of the regional cause-specific mortality structures. Furthermore for many groups of causes Russian death rates are much lower than those in European countries. This peculiarity is caused by the specificity of Russian cause-of-death coding practice, which results in underestimating the mortality from certain causes at old ages.

Accuracy of malignant hyperthermia diagnoses in hospital discharge records.

BACKGROUND: In 1997, the International Classification of Diseases (ICD), 9th Revision Clinical Modification (ICD-9) coding system introduced the code for malignant hyperthermia (MH) (995.86). The aim of this study was to estimate the accuracy of coding for MH in hospital discharge records. METHODS: An expert panel of anesthesiologists reviewed medical records for patients with a discharge diagnosis of MH based on ICD-9 or ICD-10 codes from January 1, 2006 to December 31, 2008 at six tertiary care medical centers in North America. All cases were categorized as possible, probable, or fulminant MH, history of MH (family or personal) or other. RESULTS: A total of 47 medical records with MH diagnoses were reviewed; 68.1% had a documented surgical procedure and general anesthesia, and 23.4% (95% CI, 12.3-38.0%) had a possible, probable, or fulminant MH event. Dantrolene was given in 81% of the MH events. All patients judged to have an incident MH event survived to discharge. Family and personal history of MH accounted for 46.8% of cases. High fever without evidence of MH during admission accounted for 23.4%, and the reason for MH coding was unclear in 6.4% of cases. CONCLUSIONS: Approximately one quarter of ICD-9 or ICD-10 coded MH diagnoses in hospital discharge records refer to incident MH episodes and an additional 47% to MH susceptibility (including personal history or family history). Information such as surgical procedure, anesthesia billing data, and dantrolene administration may aid in identifying incident MH cases among those with an ICD-9 or ICD-10 coded MH diagnosis in their hospital discharge records.