Under-documentation of chronic kidney disease in the electronic health record in outpatients

The ICD-9 to ICD-10 transition has not improved identification of rapidly progressing stage 3 and stage 4 chronic kidney disease patients: a diagnostic test study

BACKGROUND

The International Classification of Diseases (ICD) coding system is the industry standard tool for billing, disease classification, and epidemiology purposes. Prior research has demonstrated ICD codes to have poor accuracy, particularly in relation to rapidly progressing chronic kidney disease (CKD) patients. In 2016, the ICD system moved to revision 10. This study examines subjects in a large insurer database to determine the accuracy of ICD-10 CKD-staging codes to diagnose patients rapidly progressing towards end-stage kidney disease (ESKD).

PATIENTS AND METHODS

Serial observations of outpatient serum creatinine measurements from 2016 to 2021 of 315,903 patients were transformed to estimated glomerular filtration rate (eGFR) to identify CKD stage-3 and advanced patients diagnosed clinically (eGFR-CKD). CKD-staging codes from the same time period of 59,386 patients and used to identify stage-3 and advanced patients diagnosed by ICD-code (ICD-CKD). eGFR-CKD and ICD-CKD diagnostic accuracy was compared between a total of 334,610 patients.

RESULTS

5,618 patients qualified for the progression analysis; 72 were identified as eGFR rapid progressors; 718 had multiple codes to qualify as ICD rapid progressors. Sensitivity was 5.56%, with positive predictive value (PPV) 5.6%. 34,858 patients were diagnosed as eGFR-CKD stage-3 patients; 17,549 were also diagnosed as ICD-CKD stage-3 patients, for a sensitivity of 50.34%, with PPV of 58.71%. 4,069 patients reached eGFR-CKD stage-4 with 2,750 ICD-CKD stage-4 patients, giving a sensitivity of 67.58%, PPV of 42.43%. 959 patients reached eGFR-CKD stage-5 with 566 ICD-CKD stage-5 patients, giving a sensitivity of 59.02%, PPV of 35.85%.

CONCLUSION

This research shows that recent ICD revisions have not improved identification of rapid progressors in diagnostic accuracy, although marked increases in sensitivity for stage-3 (50.34% vs. 24.68%), and PPV in stage-3 (58.71% vs. 40.08%), stage-4 (42.43% vs. 18.52%), and stage-5 (35.85% vs. 4.51%) were observed. However, sensitivity in stage-5 compares poorly (59.02% vs. 91.05%).

Applying Natural Language Processing to Textual Data From Clinical Data Warehouses: Systematic Review

BACKGROUND

In recent years, health data collected during the clinical care process have been often repurposed for secondary use through clinical data warehouses (CDWs), which interconnect disparate data from different sources. A large amount of information of high clinical value is stored in unstructured text format. Natural language processing (NLP), which implements algorithms that can operate on massive unstructured textual data, has the potential to structure the data and make clinical information more accessible.

OBJECTIVE

The aim of this review was to provide an overview of studies applying NLP to textual data from CDWs. It focuses on identifying the (1) NLP tasks applied to data from CDWs and (2) NLP methods used to tackle these tasks.

METHODS

This review was performed according to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. We searched for relevant articles in 3 bibliographic databases: PubMed, Google Scholar, and ACL Anthology. We reviewed the titles and abstracts and included articles according to the following inclusion criteria: (1) focus on NLP applied to textual data from CDWs, (2) articles published between 1995 and 2021, and (3) written in English.

RESULTS

We identified 1353 articles, of which 194 (14.34%) met the inclusion criteria. Among all identified NLP tasks in the included papers, information extraction from clinical text (112/194, 57.7%) and the identification of patients (51/194, 26.3%) were the most frequent tasks. To address the various tasks, symbolic methods were the most common NLP methods (124/232, 53.4%), showing that some tasks can be partially achieved with classical NLP techniques, such as regular expressions or pattern matching that exploit specialized lexica, such as drug lists and terminologies. Machine learning (70/232, 30.2%) and deep learning (38/232, 16.4%) have been increasingly used in recent years, including the most recent approaches based on transformers. NLP methods were mostly applied to English language data (153/194, 78.9%).

CONCLUSIONS

CDWs are central to the secondary use of clinical texts for research purposes. Although the use of NLP on data from CDWs is growing, there remain challenges in this field, especially with regard to languages other than English. Clinical NLP is an effective strategy for accessing, extracting, and transforming data from CDWs. Information retrieved with NLP can assist in clinical research and have an impact on clinical practice.

Determining the readiness of patients with renal failure to use health information technology

INTRODUCTION

Using information technology (IT) for purposes such as patient education and disease prevention and management is effective when patients are ready to use it. The objective of this study was to determine the readiness of patients with renal failure to use health IT.

METHODS

This study was performed on all dialysis patients in South Khorasan province (n = 263) using a 28-item questionnaire. The questionnaire consisted of (1) demographic information of participants and (2) questions concerning eight main factors including the need for information, desire to receive information, ability to use computers and the Internet, computers and the Internet anxiety, communication with physicians, using mobile phones and concerns about security and confidentiality of information. Descriptive statistics and Mann-Whitney and Kruskal-Wallis statistical tests were used to analyze the data.

RESULTS

About 15% of the participants stated that they do not want to receive information from the Internet. Anxiety and concern about Internet security and confidentiality were higher in women, married people, people over 60, villagers, and illiterate people (p < 0.05). Married people and people over 60 years had a higher desire to get information (p < 0.05). The rate of computer anxiety and Internet privacy concern was higher than average (p < 0.001). Most patients (34.2%) could only send text messages using mobile phones.

CONCLUSION

Despite the need of most patients to online health information, they do not use this information due to a lack of skills and experience to use IT. Therefore, the ability of users should be considered when developing IT-based interventions. Due to patients' concerns about Internet privacy, it is required to teach patients how to protect their privacy while using the Internet.

Natural Language Processing in Nephrology

Unstructured data in the electronic health records contain essential patient information. Natural language processing (NLP), teaching a computer to read, allows us to tap into these data without needing the time and effort of manual chart abstraction. The core first step for all NLP algorithms is preprocessing the text to identify the core words that differentiate the text while filtering out the noise. Traditional NLP uses a rule-based approach, applying grammatical rules to infer meaning from the text. Newer NLP approaches use machine learning/deep learning which can infer meaning without explicitly being programmed. NLP use in nephrology research has focused on identifying distinct disease processes, such as CKD, and extraction of patient-oriented outcomes such as symptoms with high sensitivity. NLP can identify patient features from clinical text associated with acute kidney injury and progression of CKD. Lastly, inclusion of features extracted using NLP improved the performance of risk-prediction models compared to models that only use structured data. Implementation of NLP algorithms has been slow, partially hindered by the lack of external validation of NLP algorithms. However, NLP allows for extraction of key patient characteristics from free text, an infrequently used resource in nephrology.

Association of clinic setting with quality indicator performance in systemic lupus erythematosus: a cross-sectional study

BACKGROUND

Healthcare quality for systemic lupus erythematosus (SLE) is a modifiable target for improving patient outcomes. We aimed to assess the quality of care processes in different clinic settings, comparing a subspecialty lupus clinic with hospital-based and private general rheumatology clinics.

METHODS

Patients with SLE (n = 258) were recruited in 2016 from a subspecialty lupus clinic (n = 147), two hospital general rheumatology clinics (n = 56) and two private rheumatology clinics (n = 55). Data were collected from medical records and patient questionnaires. Quality of care was assessed using 31 validated SLE quality indicators (QI) encompassing diagnostic work-up, disease and comorbidity assessments, drug monitoring, preventative care and reproductive health. Per-QI performance was measured as a percentage of patients that met the QI relative to the number of patients eligible. Per-patient QI performance was calculated as a percentage of QIs met relative to the number of eligible QIs for each patient. Per-QI and per-patient QI performance were compared between the three clinic settings, and multiple regression performed to adjust for sociodemographic, disease and healthcare factors.

RESULTS

Per-QI performance was generally high across all clinic settings for diagnostic work-up, comorbidity assessment, lupus nephritis, drug monitoring, prednisolone taper, osteoporosis and pregnancy care. Median [IQR] per-patient performance on eligible QIs was higher in the subspeciality lupus clinic (66.7% [57.1-74.1]) than the hospital general rheumatology (52.7% [47.5-58.1]) and private rheumatology (50.0% [42.9-60.9]) clinics (p <0.001) and the difference remained significant after multivariable adjustment. The subspecialty lupus clinic recorded higher per-QI performance for documentation of disease activity, disease damage, cardiovascular risk factor and drug toxicity assessments, pre-immunosuppression hepatitis and tuberculosis screening, new medication counselling, vaccinations, sun avoidance education and contraception counselling.

CONCLUSIONS

SLE patients managed in a subspecialty lupus clinic recorded higher per-patient QI performance compared to hospital general rheumatology and private rheumatology clinics, in part related to better documentation on certain QIs.

Augmenting the Clinical Data Sources for Enigmatic Diseases: A Cross-Sectional Study of Self-Tracking Data and Clinical Documentation in Endometriosis

BACKGROUND

Self-tracking through mobile health technology can augment the electronic health record (EHR) as an additional data source by providing direct patient input. This can be particularly useful in the context of enigmatic diseases and further promote patient engagement.

OBJECTIVES

This study aimed to investigate the additional information that can be gained through direct patient input on poorly understood diseases, beyond what is already documented in the EHR.

METHODS

This was an observational study including two samples with a clinically confirmed endometriosis diagnosis. We analyzed data from 6,925 women with endometriosis using a research app for tracking endometriosis to assess prevalence of self-reported pain problems, between- and within-person variability in pain over time, endometriosis-affected tasks of daily function, and self-management strategies. We analyzed data from 4,389 patients identified through a large metropolitan hospital EHR to compare pain problems with the self-tracking app and to identify unique data elements that can be contributed via patient self-tracking.

RESULTS

Pelvic pain was the most prevalent problem in the self-tracking sample (57.3%), followed by gastrointestinal-related (55.9%) and lower back (49.2%) pain. Unique problems that were captured by self-tracking included pain in ovaries (43.7%) and uterus (37.2%). Pain experience was highly variable both across and within participants over time. Within-person variation accounted for 58% of the total variance in pain scores, and was large in magnitude, based on the ratio of within- to between-person variability (0.92) and the intraclass correlation (0.42). Work was the most affected daily function task (49%), and there was significant within- and between-person variability in self-management effectiveness. Prevalence rates in the EHR were significantly lower, with abdominal pain being the most prevalent (36.5%).

CONCLUSION

For enigmatic diseases, patient self-tracking as an additional data source complementary to EHR can enable learning from the patient to more accurately and comprehensively evaluate patient health history and status.

Optimized Identification of Advanced Chronic Kidney Disease and Absence of Kidney Disease by Combining Different Electronic Health Data Resources and by Applying Machine Learning Strategies

Automated identification of advanced chronic kidney disease (CKD ≥ III) and of no known kidney disease (NKD) can support both clinicians and researchers. We hypothesized that identification of CKD and NKD can be improved, by combining information from different electronic health record (EHR) resources, comprising laboratory values, discharge summaries and ICD-10 billing codes, compared to using each component alone. We included EHRs from 785 elderly multimorbid patients, hospitalized between 2010 and 2015, that were divided into a training and a test (n = 156) dataset. We used both the area under the receiver operating characteristic (AUROC) and under the precision-recall curve (AUCPR) with a 95% confidence interval for evaluation of different classification models. In the test dataset, the combination of EHR components as a simple classifier identified CKD ≥ III (AUROC 0.96[0.93-0.98]) and NKD (AUROC 0.94[0.91-0.97]) better than laboratory values (AUROC CKD 0.85[0.79-0.90], NKD 0.91[0.87-0.94]), discharge summaries (AUROC CKD 0.87[0.82-0.92], NKD 0.84[0.79-0.89]) or ICD-10 billing codes (AUROC CKD 0.85[0.80-0.91], NKD 0.77[0.72-0.83]) alone. Logistic regression and machine learning models improved recognition of CKD ≥ III compared to the simple classifier if only laboratory values were used (AUROC 0.96[0.92-0.99] vs. 0.86[0.81-0.91], p < 0.05) and improved recognition of NKD if information from previous hospital stays was used (AUROC 0.99[0.98-1.00] vs. 0.95[0.92-0.97]], p < 0.05). Depending on the availability of data, correct automated identification of CKD ≥ III and NKD from EHRs can be improved by generating classification models based on the combination of different EHR components.

Natural language processing of electronic health records is superior to billing codes to identify symptom burden in hemodialysis patients

Symptoms are common in patients on maintenance hemodialysis but identification is challenging. New informatics approaches including natural language processing (NLP) can be utilized to identify symptoms from narrative clinical documentation. Here we utilized NLP to identify seven patient symptoms from notes of maintenance hemodialysis patients of the BioMe Biobank and validated our findings using a separate cohort and the MIMIC-III database. NLP performance was compared for symptom detection with International Classification of Diseases (ICD)-9/10 codes and the performance of both methods were validated against manual chart review. From 1034 and 519 hemodialysis patients within BioMe and MIMIC-III databases, respectively, the most frequently identified symptoms by NLP were fatigue, pain, and nausea/vomiting. In BioMe, sensitivity for NLP (0.85 - 0.99) was higher than for ICD codes (0.09 - 0.59) for all symptoms with similar results in the BioMe validation cohort and MIMIC-III. ICD codes were significantly more specific for nausea/vomiting in BioMe and more specific for fatigue, depression, and pain in the MIMIC-III database. A majority of patients in both cohorts had four or more symptoms. Patients with more symptoms identified by NLP, ICD, and chart review had more clinical encounters. NLP had higher specificity in inpatient notes but higher sensitivity in outpatient notes and performed similarly across pain severity subgroups. Thus, NLP had higher sensitivity compared to ICD codes for identification of seven common hemodialysis-related symptoms, with comparable specificity between the two methods. Hence, NLP may be useful for the high-throughput identification of patient-centered outcomes when using electronic health records.

Can billing codes accurately identify rapidly progressing stage 3 and stage 4 chronic kidney disease patients: a diagnostic test study

Background

The International Classification of Diseases (ICD) coding system is the industry standard tool for billing, disease classification, and epidemiology purposes. However, ICD codes are often not assigned or incorrectly given, particularly among Chronic Kidney disease (CKD) patients. Our study evaluated the diagnostic accuracy of CKD-staging ICD codes among CKD patients from a large insurer database in identifying individuals rapidly progressing towards end-stage renal disease (ESRD).

Patients and methods

Serial observations including outpatient serum creatinine measurements collected from 2007 through 2014 of 216,529 patients were examined. The progression of CKD using a serum creatinine based longitudinal mixed-model was contrasted with that documented by CKD-staging ICD codes. Rapid progressors, defined as those with yearly estimated glomerular filtration rate (eGFR) loss greater than 4 ml/min/1.73m²) were identified. The diagnosis of CKD using eGFR was also compared to diagnosis using a set of CKD related ICD codes.

Results

Of 10,927 clinically identified CKD patients qualifying for inclusion in the progression analysis, 323 were clinically identified as rapid progressors. CKD-staging ICD codes identified 83 of these, for a sensitivity of 25.7% with positive predictive value (PPV) of 13.74%, and specificity 95.09% with negative predictive value (NPV) of 97.68%. Of 28,762 laboratory-confirmed CKD patients, 9249 had a qualifying ICD code, for a sensitivity of 16% with PPV of 63.10%; Of 187,767 patients with laboratory-confirmed absence of CKD, 182,359 also did not have a qualifying ICD code, for a specificity of 97.12% with NPV of 90.33%.

Conclusion

This study depicts the novel finding that ICD-codes display poor capacity to identify rapidly progressing CKD patients when compared to gold standard eGFR measures, and further demonstrates the limitations of coding in CKD diagnosis. This analysis further defines the limitations of ICD codes in addressing diagnosis of disease severity or disease progression for clinical or epidemiological purposes.

Electronic supplementary material

The online version of this article (10.1186/s12882-019-1429-4) contains supplementary material, which is available to authorized users.

Departmental collaborative approach for improving in-patient clinical documentation (five years experience)

Introduction

Health care institutes are cooperative areas where multiple health care services come together and work closely; physician, nurses and paramedics etc,. These multidisciplinary teams usually communicate with each other by documentation. Therefore, accurate documentation in health care organization is considered one of the vital processes. To make the documentation useful, it needs to be accurate, relevant, complete and confidential.

Objectives

The aim of this paper is to demonstrate the effect of the collaborative work in the Department of Pediatrics on improving the quality of inpatient clinical documentation over 5 years.

Methods

Improving clinical documentations went through several collaborative approaches, these include: Departmental Administration involvement, establishment of quality management team, regular departmental collaborative meeting as a monitoring and motivating tool, establishment of the residents quality team, Integration of quality projects into the new residents annual orientation, considering it as a part of the trainee personal evaluation, sending reminders to the consultants and residents on the adherence for admission note initiating and 24 h's verification, utilization of standardized template of admission note and progress note and emphasizing on the adherence to the approved medical abbreviation list only for any abbreviation to be used.

Results

During the period between the first quarter of 2012 to the fourth quarter of 2017; a significant improvement was noticed in the overall in-patient clinical documentation compliance rate, as it was ranging from lower 50% in 2012 and 2013, and increased gradually to reach upper 80% in the last quarters of 2016 and 2017. These figures are based on an independent audit that being done by the hospital quality management department and received by the department in a quarterly basis.

Conclusion

Despite multiple challenges for improving the compliance for clinical documentations, major improvement can be achieved when the collaboration and efforts among all stakeholders being shared and set as a common goal.

Clinical Informatics Researcher's Desiderata for the Data Content of the Next Generation Electronic Health Record

OBJECTIVE

Clinical informatics researchers depend on the availability of high-quality data from the electronic health record (EHR) to design and implement new methods and systems for clinical practice and research. However, these data are frequently unavailable or present in a format that requires substantial revision. This article reports the results of a review of informatics literature published from 2010 to 2016 that addresses these issues by identifying categories of data content that might be included or revised in the EHR.

MATERIALS AND METHODS

We used an iterative review process on 1,215 biomedical informatics research articles. We placed them into generic categories, reviewed and refined the categories, and then assigned additional articles, for a total of three iterations.

RESULTS

Our process identified eight categories of data content issues: Adverse Events, Clinician Cognitive Processes, Data Standards Creation and Data Communication, Genomics, Medication List Data Capture, Patient Preferences, Patient-reported Data, and Phenotyping.

DISCUSSION

These categories summarize discussions in biomedical informatics literature that concern data content issues restricting clinical informatics research. These barriers to research result from data that are either absent from the EHR or are inadequate (e.g., in narrative text form) for the downstream applications of the data. In light of these categories, we discuss changes to EHR data storage that should be considered in the redesign of EHRs, to promote continued innovation in clinical informatics.

CONCLUSION

Based on published literature of clinical informaticians' reuse of EHR data, we characterize eight types of data content that, if included in the next generation of EHRs, would find immediate application in advanced informatics tools and techniques.

Ensuring Patient Safety During the Transition to ESRD

Patients with chronic kidney disease (CKD) are at risk for complications both inherent to the disease and as a consequence of its treatment. The dangers that CKD patients face change across the spectrum of the disease. Providers who are well-versed in these safety threats are best poised to safeguard patients as their CKD progresses.

Early recognition of multiple sclerosis using natural language processing of the electronic health record

BACKGROUND

Diagnostic accuracy might be improved by algorithms that searched patients' clinical notes in the electronic health record (EHR) for signs and symptoms of diseases such as multiple sclerosis (MS). The focus this study was to determine if patients with MS could be identified from their clinical notes prior to the initial recognition by their healthcare providers.

METHODS

An MS-enriched cohort of patients with well-established MS (n = 165) and controls (n = 545), was generated from the adult outpatient clinic. A random sample cohort was generated from randomly selected patients (n = 2289) from the same adult outpatient clinic, some of whom had MS (n = 16). Patients' notes were extracted from the data warehouse and signs and symptoms mapped to UMLS terms using MedLEE. Approximately 1000 MS-related terms occurred significantly more frequently in MS patients' notes than controls'. Synonymous terms were manually clustered into 50 buckets and used as classification features. Patients were classified as MS or not using Naïve Bayes classification.

RESULTS

Classification of patients known to have MS using notes of the MS-enriched cohort entered after the initial ICD9[MS] code yielded an ROC AUC, sensitivity, and specificity of 0.90 [0.87-0.93], 0.75[0.66-0.82], and 0.91 [0.87-0.93], respectively. Similar classification accuracy was achieved using the notes from the random sample cohort. Classification of patients not yet known to have MS using notes of the MS-enriched cohort entered before the initial ICD9[MS] documentation identified 40% [23-59%] as having MS. Manual review of the EHR of 45 patients of the random sample cohort classified as having MS but lacking an ICD9[MS] code identified four who might have unrecognized MS.

CONCLUSIONS

Diagnostic accuracy might be improved by mining patients' clinical notes for signs and symptoms of specific diseases using NLP. Using this approach, we identified patients with MS early in the course of their disease which could potentially shorten the time to diagnosis. This approach could also be applied to other diseases often missed by primary care providers such as cancer. Whether implementing computerized diagnostic support ultimately shortens the time from earliest symptoms to formal recognition of the disease remains to be seen.

Chances and challenges of using routine data collections for renal health care research

BACKGROUND

Collections of electronic medical records (EMRs) can provide a rich source of information for renal health care research. However, their use in statistical analyses requires many preparatory steps, including coding of freetext entries and clear definitions of time windows for harvesting prognostic factors and outcomes. We analyse a large collection of EMRs to identify prognostic factors of adequate health care in diabetic patients at risk for chronic kidney disease, and discuss benefits and risks of such re-use of routine data.

METHODS

In a representative sample of 695 068 patient records collected in 58 Austrian general practitioners' offices, we could identify 31 374 patients with diabetes mellitus. As outcomes, we investigated whether a patient received a serum creatinine measurement, and the time elapsing between two consecutive serum creatinine measurements. Prognostic factors were defined by extracting previous diagnoses, laboratory measurements, drug prescriptions and demographic characteristics from the records.

RESULTS

Serum creatinine was measured annually in 44.4% of diabetic patients with previous signs of reduced kidney function and in 20.5% of the patients without such signs. Within 1 year after the first measurement, a follow-up measurement was made in 79.4 and 68.4% of the patients, respectively. Previous diagnoses, laboratory measurements, drug prescriptions and demographic characteristics explained 41% of the observed variance of kidney function monitoring. With 24% explained variance, previous referrals to laboratories were identified as the most important prognostic factor group.

CONCLUSIONS

The analysis of large routine data collections poses various challenges, among which the need for coding free text into variables and various sources of biases are most demanding. However, routine data collections represent the daily practice of health care and offer many chances for renal health services and outcomes research.

Is routine hospital episode data sufficient for identifying individuals with chronic kidney disease? A comparison study with laboratory data

Internationally, investment in the availability of routine health care data for improving health, health surveillance and health care is increasing. We assessed the validity of hospital episode data for identifying individuals with chronic kidney disease compared to biochemistry data in a large population-based cohort, the Grampian Laboratory Outcomes, Morbidity and Mortality Study-II (n = 70,435). Grampian Laboratory Outcomes, Morbidity and Mortality Study-II links hospital episode data to biochemistry data for all adults in a health region with impaired kidney function and random samples of individuals with normal and unmeasured kidney function in 2003. We compared identification of individuals with chronic kidney disease by hospital episode data (based on International Classification of Diseases-10 codes) to the reference standard of biochemistry data (at least two estimated glomerular filtration rates <60 mL/min/1.73 m(2) at least 90 days apart). Hospital episode data, compared to biochemistry data, identified a lower prevalence of chronic kidney disease and had low sensitivity (<10%) but high specificity (>97%). Using routine health care data from multiple sources offers the best opportunity to identify individuals with chronic kidney disease.

Pharmacist-driven renal medication dosing intervention in a primary care patient-centered medical home

PURPOSE

The purposes of this population management intervention were to identify patients with stage 3, 4, or 5 chronic kidney disease (CKD) and to improve care in a patient-centered medical home (PCMH). Objectives of the intervention were to (i) increase the identification of CKD, (ii) increase the use of aspirin and angiotensin-converting enzyme inhibitors (ACE-Is) or angiotensin II receptor blockers (ARBs) in patients with CKD, and (iii) ensure that all medications prescribed to patients with CKD were dosed appropriately based on renal function.

SETTING

This intervention was completed at a National Committee for Quality Assurance tier 3 PCMH associated with a major, academic health system.

PATIENTS

A list of 328 patients with an estimated glomerular filtration rate of < 60 mL/min per 1.73 m(2) was generated using the electronic medical record; 146 patients underwent the intervention.

MEASUREMENTS AND OBSERVATIONS

A pharmacist review of the electronic medical record was completed to confirm stage 3, 4, or 5 CKD based on estimated glomerular filtration rate, to ensure that ACE-Is/ARBs and aspirin were prescribed, and to ensure that all medications were dosed appropriately based on renal dosing adjustment recommendations. Recommendations were made to improve medication use and safety in patients with CKD. Before intervention, 73% of patients were prescribed an ACE-I/ARB and 72% of patients were prescribed aspirin. After the intervention, use of these medications increased to 77% and 82% of patients, respectively. Pharmacist recommendations to adjust medication dosing based on Cockcroft-Gault calculated creatinine clearance were made for 138 medications (0.95 medication per patient); 90 (65.2%) recommendations were accepted by the patients' physicians.

CONCLUSION

These results demonstrate the impact of a pharmacist-driven renal medication dosing intervention within a PCMH on medication use and safety for patients with CKD.

Survival Analysis with Electronic Health Record Data: Experiments with Chronic Kidney Disease

This paper presents a detailed survival analysis for chronic kidney disease (CKD). The analysis is based on the EHR data comprising almost two decades of clinical observations collected at New York-Presbyterian, a large hospital in New York City with one of the oldest electronic health records in the United States. Our survival analysis approach centers around Bayesian multiresolution hazard modeling, with an objective to capture the changing hazard of CKD over time, adjusted for patient clinical covariates and kidney-related laboratory tests. Special attention is paid to statistical issues common to all EHR data, such as cohort definition, missing data and censoring, variable selection, and potential for joint survival and longitudinal modeling, all of which are discussed alone and within the EHR CKD context.

Electronic problem list documentation of chronic kidney disease and quality of care

Background

Chronic kidney disease (CKD) is increasingly common and under-recognized in primary care clinics, leading to low rates of stage-appropriate monitoring and treatment. Our objective was to determine whether electronic problem list documentation of CKD is associated with monitoring and treatment.

Methods

This is a cross-sectional observational study of patients with stage 3 or 4 CKD, defined as two past estimated glomerular filtration rates (eGFR) 15-60 mL/min/1.73 m² separated by 90 days and collected between 2007-2008. We examined the association of problem list documentation with: 1) serum eGFR monitoring test, 2) urine protein or albumin monitoring test, 3) an angiotensin converting enzyme inhibitor or angiotensin receptor blocker (ACE/ARB) prescription, 4) mean systolic blood pressure (BP), and 5) BP control.

Results

Out of 3,149 patients with stage 3 or 4 CKD, only 16% of patients had CKD documented on the problem list. After adjustment for eGFR, gender, and race/ethnicity and after clustering by physician, problem list documentation of CKD was associated with serum eGFR testing (97% with problem list documentation vs. 94% without problem list documentation, p = 0.02) and urine protein testing (47% with problem list documentation vs. 40% without problem list documentation, p = 0.04). After adjustment, problem list documentation was not associated with ACE/ARB prescription, mean systolic BP, or BP control.

Conclusions

Documentation of CKD on the electronic problem list is rare. Patients with CKD documentation have better stage-appropriate monitoring of the disease, but do not have higher rates of blood pressure treatment or better blood pressure control. Interventions aimed at increasing documentation of CKD on the problem list may improve stage-appropriate monitoring, but may not improve clinical outcomes.

Chronic kidney disease in an electronic health record problem list: quality of care, ESRD, and mortality

BACKGROUND

Whether chronic kidney disease (CKD) recognition in an electronic health record (EHR) problem list improves processes of care or clinical outcomes of end-stage renal disease (ESRD) and death is unclear.

METHODS

We identified patients who had at least 1 year of follow-up (2005-2009) in our EHR-based CKD registry (n = 25,742). CKD recognition was defined by having ICD-9 codes for CKD, diabetic kidney disease, or hypertensive kidney disease in the problem list. We calculated proportions of patients with and without CKD recognition and examined differences by demographics, clinical factors, and development of ESRD or mortality. We evaluated differences in the proportion of patients with CKD-specific laboratory results checked before and after recognition among cases and propensity-matched controls.

RESULTS

Only 11% (n = 2,735) had CKD recognition in the problem list and they were younger (68 vs. 71 years), a higher proportion were male (61 vs. 37%) and African-American (21 vs. 10%) compared to those unrecognized. CKD-specific laboratory results for patients with estimated glomerular filtration rate (eGFR) 30-59 including intact parathyroid hormone (23 vs. 6%), vitamin D (22 vs. 18%), phosphorus (29 vs. 7%), and a urine check for proteinuria (55 vs. 36%) were significantly more likely to be done among those with CKD recognition (all p < 0.05). Similar results were found for eGFR <30 except for proteinuria and in our propensity score-matched control analysis. There was no independent association of CKD recognition with ESRD or mortality.

CONCLUSIONS

CKD recognition in the EHR problem list was low, but translated into more CKD-specific processes of care; however ESRD or mortality were not affected.

Lessons learned in replicating data-driven experiments in multiple medical systems and patient populations

Electronic health records are an increasingly important source of data for research, allowing for large-scale longitudinal studies on the same population that is being treated. Unlike in controlled studies, though, these data vary widely in quality, quantity, and structure. In order to know whether algorithms can accurately uncover new knowledge from these records, or whether findings can be extrapolated to new populations, they must be validated. One approach is to conduct the same study in multiple sites and compare results, but it is a challenge to determine whether differences are due to artifacts of the medical process, population differences, or failures of the methods used. In this paper we describe the results of replicating a data-driven experiment to infer possible causes of congestive heart failure and their timing using data from two medical systems and two patient populations. We focus on the difficulties faced in this type of work, lessons learned, and recommendations for future research.

A pilot trial of a computerized renal template note to improve resident knowledge and documentation of kidney disease

BACKGROUND

Kidney disease is under-documented in physician notes. The use of template-guided notes may improve physician recognition of kidney disease early in training.

OBJECTIVE

The objective of this study was to determine whether a computerized inpatient renal template note with clinical decision support improves resident knowledge and documentation of kidney disease.

METHODS

In this prospective study, first year medical residents were encouraged to use the renal template note for documentation over a one-month period. The renal template note included an option for classification of acute kidney injury (AKI) and chronic kidney disease (CKD) categories with a link to standard classifications. Pre- and post-knowledge of AKI and CKD categories was tested with a quiz and surveys of resident experience with the intervention were conducted. Appropriate AKI and/or CKD classification was determined in 100 renal template notes and 112 comparable historical internal medicine resident progress notes from approximately one year prior.

RESULTS

2,435 inpatient encounters amongst 15 residents who participated were documented using the renal template note. A significantly higher percent of residents correctly staged earlier stage CKD (CKD3) using the renal template note compared to historical notes (9/46 vs. 0/33, p<0.01). Documentation of AKI and more advanced CKD stages (CKD4 and 5) did not improve. Knowledge based on quiz scores increased modestly but was not significant. The renal template note was well received by residents and was perceived as helping improve knowledge and documentation of kidney disease.

CONCLUSION

The renal template note significantly improved staging of earlier stage CKD (CKD3) with a modest but non-significant improvement in resident knowledge. Given the importance of early recognition and treatment of CKD, future studies should focus on teaching early recognition using template notes with supplemental educational interventions.

Electronic health records: a new tool to combat chronic kidney disease?

Electronic health records (EHRs) were first developed in the 1960s as clinical information systems for document storage and retrieval. Adoption of EHRs has increased in the developed world and is increasing in developing countries. Studies have shown that quality of patient care is improved among health centers with EHRs. In this article, we review the structure and function of EHRs along with an examination of its potential application in CKD care and research. Well-designed patient registries using EHRs data allow for improved aggregation of patient data for quality improvement and to facilitate clinical research. Preliminary data from the United States and other countries have demonstrated that CKD care might improve with use of EHRs-based programs. We recently developed a CKD registry derived from EHRs data at our institution and complimented the registry with other patient details from the United States Renal Data System and the Social Security Death Index. This registry allows us to conduct a EHRs-based clinical trial that examines whether empowering patients with a personal health record or patient navigators improves CKD care, along with identifying participants for other clinical trials and conducting health services research. EHRs use have shown promising results in some settings, but not in others, perhaps attributed to the differences in EHRs adoption rates and varying functionality. Thus, future studies should explore the optimal methods of using EHRs to improve CKD care and research at the individual patient level, health system and population levels.

Documentation and management of CKD in rural primary care

BACKGROUND AND OBJECTIVES

Recognition of CKD by primary care practitioners is essential in rural communities where nephrology access is limited. This study determined the prevalence of undocumented CKD in patients cared for in rural primary care practices and evaluated characteristics associated with undocumented CKD as well as CKD management.

DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS

A retrospective cohort study, conducted within the Oregon Rural Practice Based Research Network, consisted of 865 CKD patients with serum creatinine≥1.5 mg/dl in males and ≥1.3 mg/dl in females and an estimated GFR<60 ml/min per 1.73 m(2). Documentation of a CKD diagnosis and laboratory values were abstracted by chart review.

RESULTS

Of CKD patients, 51.9% had no documentation of CKD. Undocumented CKD occurred more frequently in female patients (adjusted odds ratio=2.93, 95% confidence interval=2.04, 4.21). The association of serum creatinine reporting versus automating reporting of estimated GFR on CKD documentation was dependent on patient sex, years of practitioner experience, and practitioner clinical training. Hypertensive patients with documented CKD were more likely to have a BP medication change than patients with undocumented CKD (odds ratio=2.07, 95% confidence interval=1.15, 3.73). Only 2 of 449 patients with undocumented CKD were comanaged with a nephrologist compared with 20% of patients with documented CKD (odds ratio=53.20, 95% confidence interval=14.90, 189.90).

CONCLUSIONS

Undocumented CKD in a rural primary care setting is frequent, particularly in female patients. Depending on practitioner characteristics, automatic reporting of estimated GFR might improve documentation of CKD in this population.

Clinical use of an enterprise data warehouse

The enormous amount of data being collected by electronic medical records (EMR) has found additional value when integrated and stored in data warehouses. The enterprise data warehouse (EDW) allows all data from an organization with numerous inpatient and outpatient facilities to be integrated and analyzed. We have found the EDW at Intermountain Healthcare to not only be an essential tool for management and strategic decision making, but also for patient specific clinical decision support. This paper presents the structure and two case studies of a framework that has provided us the ability to create a number of decision support applications that are dependent on the integration of previous enterprise-wide data in addition to a patient's current information in the EMR.

Improving completeness of electronic problem lists through clinical decision support: a randomized, controlled trial

Background

Accurate clinical problem lists are critical for patient care, clinical decision support, population reporting, quality improvement, and research. However, problem lists are often incomplete or out of date.

Objective

To determine whether a clinical alerting system, which uses inference rules to notify providers of undocumented problems, improves problem list documentation.

Study Design and Methods

Inference rules for 17 conditions were constructed and an electronic health record-based intervention was evaluated to improve problem documentation. A cluster randomized trial was conducted of 11 participating clinics affiliated with a large academic medical center, totaling 28 primary care clinical areas, with 14 receiving the intervention and 14 as controls. The intervention was a clinical alert directed to the provider that suggested adding a problem to the electronic problem list based on inference rules. The primary outcome measure was acceptance of the alert. The number of study problems added in each arm as a pre-specified secondary outcome was also assessed. Data were collected during 6-month pre-intervention (11/2009–5/2010) and intervention (5/2010–11/2010) periods.

Results

17 043 alerts were presented, of which 41.1% were accepted. In the intervention arm, providers documented significantly more study problems (adjusted OR=3.4, p<0.001), with an absolute difference of 6277 additional problems. In the intervention group, 70.4% of all study problems were added via the problem list alerts. Significant increases in problem notation were observed for 13 of 17 conditions.

Conclusion

Problem inference alerts significantly increase notation of important patient problems in primary care, which in turn has the potential to facilitate quality improvement.

Trial Registration

ClinicalTrials.gov: NCT01105923.

A method and knowledge base for automated inference of patient problems from structured data in an electronic medical record

BACKGROUND

Accurate knowledge of a patient's medical problems is critical for clinical decision making, quality measurement, research, billing and clinical decision support. Common structured sources of problem information include the patient problem list and billing data; however, these sources are often inaccurate or incomplete.

OBJECTIVE

To develop and validate methods of automatically inferring patient problems from clinical and billing data, and to provide a knowledge base for inferring problems.

STUDY DESIGN AND METHODS

We identified 17 target conditions and designed and validated a set of rules for identifying patient problems based on medications, laboratory results, billing codes, and vital signs. A panel of physicians provided input on a preliminary set of rules. Based on this input, we tested candidate rules on a sample of 100,000 patient records to assess their performance compared to gold standard manual chart review. The physician panel selected a final rule for each condition, which was validated on an independent sample of 100,000 records to assess its accuracy.

RESULTS

Seventeen rules were developed for inferring patient problems. Analysis using a validation set of 100,000 randomly selected patients showed high sensitivity (range: 62.8-100.0%) and positive predictive value (range: 79.8-99.6%) for most rules. Overall, the inference rules performed better than using either the problem list or billing data alone.

CONCLUSION

We developed and validated a set of rules for inferring patient problems. These rules have a variety of applications, including clinical decision support, care improvement, augmentation of the problem list, and identification of patients for research cohorts.

Discovering disease associations by integrating electronic clinical data and medical literature

Electronic health record (EHR) systems offer an exceptional opportunity for studying many diseases and their associated medical conditions within a population. The increasing number of clinical record entries that have become available electronically provides access to rich, large sets of patients' longitudinal medical information. By integrating and comparing relations found in the EHRs with those already reported in the literature, we are able to verify existing and to identify rare or novel associations. Of particular interest is the identification of rare disease co-morbidities, where the small numbers of diagnosed patients make robust statistical analysis difficult. Here, we introduce ADAMS, an Application for Discovering Disease Associations using Multiple Sources, which contains various statistical and language processing operations. We apply ADAMS to the New York-Presbyterian Hospital's EHR to combine the information from the relational diagnosis tables and textual discharge summaries with those from PubMed and Wikipedia in order to investigate the co-morbidities of the rare diseases Kaposi sarcoma, toxoplasmosis, and Kawasaki disease. In addition to finding well-known characteristics of diseases, ADAMS can identify rare or previously unreported associations. In particular, we report a statistically significant association between Kawasaki disease and diagnosis of autistic disorder.