The quality of diagnostic coding in cerebrovascular disease

A Time-Tested Information System in Neurosurgical Oncology

The Brain and Spine Center at The University of Texas MD Anderson Cancer Center is a leading multidisciplinary referral center for patients with nervous system (NS) tumors. It has a wealth of clinical experience and an internationally recognized leadership role in the management of NS cancers. In that context, an informatics infrastructure that allows the archiving of both the prospective and retrospective characterization of patients, diseases, treatments, and outcomes is invaluable. We describe our experience with the Neurosurgical Oncology Database, a database that has provided valuable, extensive, and readily searchable data on multifaceted patient, tumor, and treatment characteristics for many years, successfully serving as an administrative and operational resource and as a resource for retrospective and prospective research endeavors.

Measures of prevalence: which healthcare registers are applicable?

Aims: This study analyses the applicability of some of the registers used within the healthcare system for estimations of disease prevalence .The study focuses on the diagnoses of asthma, diabetes mellitus, chronic bronchitis/emphysema, hypertensive disease, and cerebrovascular disease. Methods: The study population comprised all inhabitants ( n=20,037) in the municipality of Tierp on 31 December 1996. Diagnostic information was collected from primary healthcare and occupational healthcare in the municipality of Tierp and from inpatient and outpatient units at the hospitals in Uppsala County. The proportion of registered patients in the different registers was calculated in relation to the total number of patients who had been registered during 1996 with the selected diagnoses . Results: In the primary healthcare register, between 67% ( cerebrovascular disease) and 85% ( asthma) of all patients with selected diagnoses could be identified. A search on the inpatient care register ( Hospital Discharge Register) led to the identification of between 8% ( hypertensive disease) and 53% ( cerebrovascular disease) of the patients. Conclusions: For all of the examined diagnoses, most patients could be identified in the primary healthcare register. Register data from both primary healthcare and inpatient and outpatient care at hospital are needed to make reasonable estimates of prevalence.

Biases in detection of apparent "weekend effect" on outcome with administrative coding data: population based study of stroke

OBJECTIVES

 To determine the accuracy of coding of admissions for stroke on weekdays versus weekends and any impact on apparent outcome.

DESIGN

 Prospective population based stroke incidence study and a scoping review of previous studies of weekend effects in stroke.

SETTING

 Primary and secondary care of all individuals registered with nine general practices in Oxfordshire, United Kingdom (OXVASC, the Oxford Vascular Study).

PARTICIPANTS

 All patients with clinically confirmed acute stroke in OXVASC identified with multiple overlapping methods of ascertainment in 2002-14 versus all acute stroke admissions identified by hospital diagnostic and mortality coding alone during the same period.

MAIN OUTCOMES MEASURES

 Accuracy of administrative coding data for all patients with confirmed stroke admitted to hospital in OXVASC. Difference between rates of "false positive" or "false negative" coding for weekday and weekend admissions. Impact of inaccurate coding on apparent case fatality at 30 days in weekday versus weekend admissions. Weekend effects on outcomes in patients with confirmed stroke admitted to hospital in OXVASC and impacts of other potential biases compared with those in the scoping review.

RESULTS

 Among 92 728 study population, 2373 episodes of acute stroke were ascertained in OXVASC, of which 826 (34.8%) mainly minor events were managed without hospital admission, 60 (2.5%) occurred out of the area or abroad, and 195 (8.2%) occurred in hospital during an admission for a different reason. Of 1292 local hospital admissions for acute stroke, 973 (75.3%) were correctly identified by administrative coding. There was no bias in distribution of weekend versus weekday admission of the 319 strokes missed by coding. Of 1693 admissions for stroke identified by coding, 1055 (62.3%) were confirmed to be acute strokes after case adjudication. Among the 638 false positive coded cases, patients were more likely to be admitted on weekdays than at weekends (536 (41.0%) v 102 (26.5%); P<0.001), partly because of weekday elective admissions after previous stroke being miscoded as new stroke episodes (267 (49.8%) v 26 (25.5%); P<0.001). The 30 day case fatality after these elective admissions was lower than after confirmed acute stroke admissions (11 (3.8%) v 233 (22.1%); P<0.001). Consequently, relative 30 day case fatality for weekend versus weekday admissions differed (P<0.001) between correctly coded acute stroke admissions and false positive coding cases. Results were consistent when only the 1327 emergency cases identified by "admission method" from coding were included, with more false positive cases with low case fatality (35 (14.7%)) being included for weekday versus weekend admissions (190 (19.5%) v 48 (13.7%), P<0.02). Among all acute stroke admissions in OXVASC, there was no imbalance in baseline stroke severity for weekends versus weekdays and no difference in case fatality at 30 days (adjusted odds ratio 0.85, 95% confidence interval 0.63 to 1.15; P=0.30) or any adverse "weekend effect" on modified Rankin score at 30 days (0.78, 0.61 to 0.99; P=0.04) or one year (0.76, 0.59 to 0.98; P=0.03) among incident strokes.

CONCLUSION

 Retrospective studies of UK administrative hospital coding data to determine "weekend effects" on outcome in acute medical conditions, such as stroke, can be undermined by inaccurate coding, which can introduce biases that cannot be reliably dealt with by adjustment for case mix.

A Karnaugh map based approach towards systemic reviews and meta-analysis

Studying meta-analysis and systemic reviews since long had helped us conclude numerous parallel or conflicting studies. Existing studies are presented in tabulated forms which contain appropriate information for specific cases yet it is difficult to visualize. On meta-analysis of data, this can lead to absorption and subsumption errors henceforth having undesirable potential of consecutive misunderstandings in social and operational methodologies. The purpose of this study is to investigate an alternate forum for meta-data presentation that relies on humans’ strong pictorial perception capability. Analysis of big-data is assumed to be a complex and daunting task often reserved on the computational powers of machines yet there exist mapping tools which can analyze such data in a hand-handled manner. Data analysis on such scale can benefit from the use of statistical tools like Karnaugh maps where all studies can be put together on a graph based mapping. Such a formulation can lead to more control in observing patterns of research community and analyzing further for uncertainty and reliability metrics. We present a methodological process of converting a well-established study in Health care to its equaling binary representation followed by furnishing values on to a Karnaugh Map. The data used for the studies presented herein is from Burns et al (J Publ Health 34(1):138–148, 2011) consisting of retrospectively collected data sets from various studies on clinical coding data accuracy. Using a customized filtration process, a total of 25 studies were selected for review with no, partial, or complete knowledge of six independent variables thus forming 64 independent cells on a Karnaugh map. The study concluded that this pictorial graphing as expected had helped in simplifying the overview of meta-analysis and systemic reviews.

Accuracy of Electronic Health Record Data for Identifying Stroke Cases in Large-Scale Epidemiological Studies: A Systematic Review from the UK Biobank Stroke Outcomes Group

Objective

Long-term follow-up of population-based prospective studies is often achieved through linkages to coded regional or national health care data. Our knowledge of the accuracy of such data is incomplete. To inform methods for identifying stroke cases in UK Biobank (a prospective study of 503,000 UK adults recruited in middle-age), we systematically evaluated the accuracy of these data for stroke and its main pathological types (ischaemic stroke, intracerebral haemorrhage, subarachnoid haemorrhage), determining the optimum codes for case identification.

Methods

We sought studies published from 1990-November 2013, which compared coded data from death certificates, hospital admissions or primary care with a reference standard for stroke or its pathological types. We extracted information on a range of study characteristics and assessed study quality with the Quality Assessment of Diagnostic Studies tool (QUADAS-2). To assess accuracy, we extracted data on positive predictive values (PPV) and—where available—on sensitivity, specificity, and negative predictive values (NPV).

Results

37 of 39 eligible studies assessed accuracy of International Classification of Diseases (ICD)-coded hospital or death certificate data. They varied widely in their settings, methods, reporting, quality, and in the choice and accuracy of codes. Although PPVs for stroke and its pathological types ranged from 6–97%, appropriately selected, stroke-specific codes (rather than broad cerebrovascular codes) consistently produced PPVs >70%, and in several studies >90%. The few studies with data on sensitivity, specificity and NPV showed higher sensitivity of hospital versus death certificate data for stroke, with specificity and NPV consistently >96%. Few studies assessed either primary care data or combinations of data sources.

Conclusions

Particular stroke-specific codes can yield high PPVs (>90%) for stroke/stroke types. Inclusion of primary care data and combining data sources should improve accuracy in large epidemiological studies, but there is limited published information about these strategies.

Validity of Diagnostic Codes for Acute Stroke in Administrative Databases: A Systematic Review

OBJECTIVE

To conduct a systematic review of studies reporting on the validity of International Classification of Diseases (ICD) codes for identifying stroke in administrative data.

METHODS

MEDLINE and EMBASE were searched (inception to February 2015) for studies: (a) Using administrative data to identify stroke; or (b) Evaluating the validity of stroke codes in administrative data; and (c) Reporting validation statistics (sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), or Kappa scores) for stroke, or data sufficient for their calculation. Additional articles were located by hand search (up to February 2015) of original papers. Studies solely evaluating codes for transient ischaemic attack were excluded. Data were extracted by two independent reviewers; article quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies tool.

RESULTS

Seventy-seven studies published from 1976-2015 were included. The sensitivity of ICD-9 430-438/ICD-10 I60-I69 for any cerebrovascular disease was ≥ 82% in most [≥ 50%] studies, and specificity and NPV were both ≥ 95%. The PPV of these codes for any cerebrovascular disease was ≥ 81% in most studies, while the PPV specifically for acute stroke was ≤ 68%. In at least 50% of studies, PPVs were ≥ 93% for subarachnoid haemorrhage (ICD-9 430/ICD-10 I60), 89% for intracerebral haemorrhage (ICD-9 431/ICD-10 I61), and 82% for ischaemic stroke (ICD-9 434/ICD-10 I63 or ICD-9 434&436). For in-hospital deaths, sensitivity was 55%. For cerebrovascular disease or acute stroke as a cause-of-death on death certificates, sensitivity was ≤ 71% in most studies while PPV was ≥ 87%.

CONCLUSIONS

While most cases of prevalent cerebrovascular disease can be detected using 430-438/I60-I69 collectively, acute stroke must be defined using more specific codes. Most in-hospital deaths and death certificates with stroke as a cause-of-death correspond to true stroke deaths. Linking vital statistics and hospitalization data may improve the ascertainment of fatal stroke.

Impact of Documentation Errors on Accuracy of Cause of Death Coding in an Educational Hospital in Southern Iran

Accurate cause of death coding leads to organised and usable death information but there are some factors that influence documentation on death certificates and therefore affect the coding. We reviewed the role of documentation errors on the accuracy of death coding at Shahid Mohammadi Hospital (SMH), Bandar Abbas, Iran. We studied the death certificates of all deceased patients in SMH from October 2010 to March 2011. Researchers determined and coded the underlying cause of death on the death certificates according to the guidelines issued by the World Health Organization in Volume 2 of the International Statistical Classification of Diseases and Health Related Problems-10th revision (ICD-10). Necessary ICD coding rules (such as the General Principle, Rules 1–3, the modification rules and other instructions about death coding) were applied to select the underlying cause of death on each certificate. Demographic details and documentation errors were then extracted. Data were analysed with descriptive statistics and chi square tests. The accuracy rate of causes of death coding was 51.7%, demonstrating a statistically significant relationship (p=.001) with major errors but not such a relationship with minor errors. Factors that result in poor quality of Cause of Death coding in SMH are lack of coder training, documentation errors and the undesirable structure of death certificates.

Registration of acute stroke: validity in the Danish Stroke Registry and the Danish National Registry of Patients

BACKGROUND

The validity of the registration of patients in stroke-specific registries has seldom been investigated, nor compared with administrative hospital discharge registries. The objective of this study was to examine the validity of the registration of patients in a stroke-specific registry (The Danish Stroke Registry [DSR]) and a hospital discharge registry (The Danish National Patient Registry [DNRP]).

METHODS

Assuming that all patients with stroke were registered in either the DSR, DNRP or both, we first identified a sample of 75 patients registered with stroke in 2009; 25 patients in the DSR, 25 patients in the DNRP, and 25 patients registered in both data sources. Using the medical record as a gold standard, we then estimated the sensitivity and positive predictive value of a stroke diagnosis in the DSR and the DNRP. Secondly, we reviewed 160 medical records for all potential stroke patients discharged from four major neurologic wards within a 7-day period in 2010, and estimated the sensitivity, specificity, positive predictive value, and negative predictive value of the DSR and the DNRP.

RESULTS

Using the first approach, we found a sensitivity of 97% (worst/best case scenario 92%-99%) in the DSR and 79% (worst/best case scenario 73%-84%) in the DNRP. The positive predictive value was 90% (worst/best case scenario 72%-98%) in the DSR and 79% (worst/best case scenario 62%-88%) in the DNRP. Using the second approach, we found a sensitivity of 91% (95% confidence interval [CI] 81%-96%) and 58% (95% CI 46%-69%) in the DSR and DNRP, respectively. The negative predictive value was 91% (95% CI 83%-96%) in the DSR and 72% (95% CI 62%-80%) in the DNRP. The specificity and positive predictive value did not differ among the registries.

CONCLUSION

Our data suggest a higher sensitivity in the DSR than the DNRP for acute stroke diagnoses, whereas the positive predictive value was comparable in the two data sources.

Systematic review of discharge coding accuracy

INTRODUCTION

Routinely collected data sets are increasingly used for research, financial reimbursement and health service planning. High quality data are necessary for reliable analysis. This study aims to assess the published accuracy of routinely collected data sets in Great Britain.

METHODS

Systematic searches of the EMBASE, PUBMED, OVID and Cochrane databases were performed from 1989 to present using defined search terms. Included studies were those that compared routinely collected data sets with case or operative note review and those that compared routinely collected data with clinical registries.

RESULTS

Thirty-two studies were included. Twenty-five studies compared routinely collected data with case or operation notes. Seven studies compared routinely collected data with clinical registries. The overall median accuracy (routinely collected data sets versus case notes) was 83.2% (IQR: 67.3-92.1%). The median diagnostic accuracy was 80.3% (IQR: 63.3-94.1%) with a median procedure accuracy of 84.2% (IQR: 68.7-88.7%). There was considerable variation in accuracy rates between studies (50.5-97.8%). Since the 2002 introduction of Payment by Results, accuracy has improved in some respects, for example primary diagnoses accuracy has improved from 73.8% (IQR: 59.3-92.1%) to 96.0% (IQR: 89.3-96.3), P= 0.020.

CONCLUSION

Accuracy rates are improving. Current levels of reported accuracy suggest that routinely collected data are sufficiently robust to support their use for research and managerial decision-making.

Organisational factors affecting the quality of hospital clinical coding

The influence of organisational factors on the quality of hospital coding using the International Statistical Classification of Diseases and Health Related Problems, 10th Revision, Australian Modification (ICD-10-AM) was investigated using a mixed quantitative-qualitative approach. The organisational variables studied were: hospital specialty; geographical locality; structural characteristics of the coding unit; education, training and resource supports for Clinical Coders; and quality control mechanisms. Baseline data on the hospitals' coding quality, measured by the Performance Indicators for Coding Quality tool, were used as an independent index measure. No differences were found in error rates between rural and metropolitan hospitals, or general and specialist hospitals. Clinical Coder allocation to "general" rather than "specialist" unit coding resulted in fewer errors. Coding Managers reported that coding quality can be improved by: Coders engaging in a variety of role behaviours; improved Coder career opportunities; higher staffing levels; reduced throughput; fewer time constraints on coding outputs and associated work; and increased Coder interactions with medical staff.

Miscoding as a cause of elevated simple pneumonia mortality

BACKGROUND

An independent health care evaluating organization reported on the Internet that the 1996-1998 mortality rate for simple pneumonia (one of the diagnoses in diagnosis-related group [DRG] 89) was 11.06% (expected rate, 7.69%)--a rate much higher than suggested by the medical center's prospective quality surveillance. Chart review was undertaken to explain this reported higher mortality rate.

METHODS

Two-hundred forty-six charts of patients (123 expired, and 123 alive at discharge) were reviewed; each chart concluded with a principal diagnosis. The differences between the originally coded principal diagnosis and the recoded principal diagnoses were examined.

RESULTS

Application of Coding Clinic guidelines revealed that a principal diagnosis of simple pneumonia should have been coded in only 85 (34.6%) of the charts. The remaining charts should have been coded as respiratory failure (13.8%), congestive heart failure (11.4%), respiratory infections and inflammations (7.7%), and other diagnoses (32.5%). Coding occurred prior to discharge summary dictation in 48.4% of the cases. On the basis of the findings, the actual calculated mortality rate of simple pneumonia was 6.6%.

DISCUSSION

Coding in advance of discharge summary completion and nonexplicit documentation of the principal diagnoses occurred frequently. Reasons for miscoding included failure to distinguish between principal and final diagnoses, delay in discharge summary dictation, and inadequate documentation.

Multimedia computer database for neurosurgery

OBJECTIVE

There is a need for an efficient mechanism of storing and analyzing neurosurgical clinical, imaging, and operative data to facilitate clinical audit, research, education, and preparation of scientific presentations.

METHODS

A computer database was developed to meet this need. The recorded data include diagnoses, digitized neuroimaging studies, operative details (with intraoperative video clips), transcranial Doppler studies, outcomes, complications, admissions, and clinic visits. The anatomy, pathology, and clinical presentation are recorded for each diagnosis.

RESULTS

The database provides an audit of neurosurgery cases, which includes admission Glasgow Coma Scale score, length of intensive care and hospital stays, Glasgow Outcome Scale score, and complications. Clinical research is facilitated by flexible search strategies based on the anatomy, pathology, or clinical presentation of diseases, or any of the recorded intraoperative or outcome factors. The system can be used to assess the influence on outcome of factors, such as transcranial Doppler velocity, intraoperative blood pressure, and the use of ventricular drainage, intraoperative angiography, or temporary clipping. The database can be used to track patients with untreated or partially treated conditions, such as incidental or incompletely coiled aneurysms. The recorded images and video clips are used for teaching and producing multimedia presentations and reports. The database is designed to enable secure Internet connections among institutions so that outcomes and complications can be compared among surgeons and institutions.

CONCLUSION

This multimedia computer database facilitates clinical audit, research, teaching, and presentation activities.