The Validity of ICD-9-CM Codes in Identifying Postoperative Deep Vein Thrombosis and Pulmonary Embolism

Exploring the Applicability of Using Natural Language Processing to Support Nationwide Venous Thromboembolism Surveillance: Model Evaluation Study

Background

Venous thromboembolism (VTE) is a preventable, common vascular disease that has been estimated to affect up to 900,000 people per year. It has been associated with risk factors such as recent surgery, cancer, and hospitalization. VTE surveillance for patient management and safety can be improved via natural language processing (NLP). NLP tools have the ability to access electronic medical records, identify patients that meet the VTE case definition, and subsequently enter the relevant information into a database for hospital review.

Objective

We aimed to evaluate the performance of a VTE identification model of IDEAL-X (Information and Data Extraction Using Adaptive Learning; Emory University)-an NLP tool-in automatically classifying cases of VTE by "reading" unstructured text from diagnostic imaging records collected from 2012 to 2014.

Methods

After accessing imaging records from pilot surveillance systems for VTE from Duke University and the University of Oklahoma Health Sciences Center (OUHSC), we used a VTE identification model of IDEAL-X to classify cases of VTE that had previously been manually classified. Experts reviewed the technicians' comments in each record to determine if a VTE event occurred. The performance measures calculated (with 95% CIs) were accuracy, sensitivity, specificity, and positive and negative predictive values. Chi-square tests of homogeneity were conducted to evaluate differences in performance measures by site, using a significance level of .05.

Results

The VTE model of IDEAL-X "read" 1591 records from Duke University and 1487 records from the OUHSC, for a total of 3078 records. The combined performance measures were 93.7% accuracy (95% CI 93.7%-93.8%), 96.3% sensitivity (95% CI 96.2%-96.4%), 92% specificity (95% CI 91.9%-92%), an 89.1% positive predictive value (95% CI 89%-89.2%), and a 97.3% negative predictive value (95% CI 97.3%-97.4%). The sensitivity was higher at Duke University (97.9%, 95% CI 97.8%-98%) than at the OUHSC (93.3%, 95% CI 93.1%-93.4%; P<.001), but the specificity was higher at the OUHSC (95.9%, 95% CI 95.8%-96%) than at Duke University (86.5%, 95% CI 86.4%-86.7%; P<.001).

Conclusions

The VTE model of IDEAL-X accurately classified cases of VTE from the pilot surveillance systems of two separate health systems in Durham, North Carolina, and Oklahoma City, Oklahoma. NLP is a promising tool for the design and implementation of an automated, cost-effective national surveillance system for VTE. Conducting public health surveillance at a national scale is important for measuring disease burden and the impact of prevention measures. We recommend additional studies to identify how integrating IDEAL-X in a medical record system could further automate the surveillance process.

Validation of Postsurgical Venous Thromboembolism Diagnoses of Patients Undergoing Lower Limb Orthopedic Surgery in the Danish National Patient Registry

Background

Healthcare databases can be a valuable source of epidemiological research regarding postoperative venous thromboembolism (VTE), ie, deep vein thrombosis (DVT) and pulmonary embolism (PE), following orthopedic procedures, but only if the diagnoses are valid. We examined the validity of VTE diagnosis codes in the Danish National Patient Registry (DNPR) by calculating their positive predictive value (PPV) and negative predictive value (NPV) versus actual medical records.

Methods

We identified patients who had undergone lower limb surgery during the period 2009–2019 at a hospital in the North Denmark Region. Of these, 420 patients had at least one VTE diagnosis registered in the DNPR within 180 days after lower limb surgery. Each patient with a VTE diagnosis was matched with two patients on age and sex, as well as type, location and period of surgery. The entire medical record and diagnostic imaging were reviewed to confirm VTE diagnosis.

Results

The overall PPVs was 85.2% (95% CI: 81.5–88.5%) for first time VTE diagnosis following lower limb surgery, 82.6% (95% CI: 77.5–82.8%) for DVT, and 90.3% (95% CI: 84.3–94.6%) for PE. We found improvement in PPV during the study period when stratifying for three periods of the whole period. There were no significant differences when stratifying for sex, age, or surgery site. All negative predictive values were higher than 99%. A total of 113 additional VTE diagnoses were registered among 88 VTE patients during follow-up. Only four of the suspected recurrent VTEs were confirmed to be true recurrent VTEs.

Conclusion

The VTE diagnosis codes in the DNPR after lower limb orthopedic surgery were highly valid against the actual medical records, and we observed better PPV over recent years.

High-throughput multimodal automated phenotyping (MAP) with application to PheWAS

OBJECTIVE

Electronic health records linked with biorepositories are a powerful platform for translational studies. A major bottleneck exists in the ability to phenotype patients accurately and efficiently. The objective of this study was to develop an automated high-throughput phenotyping method integrating International Classification of Diseases (ICD) codes and narrative data extracted using natural language processing (NLP).

MATERIALS AND METHODS

We developed a mapping method for automatically identifying relevant ICD and NLP concepts for a specific phenotype leveraging the Unified Medical Language System. Along with health care utilization, aggregated ICD and NLP counts were jointly analyzed by fitting an ensemble of latent mixture models. The multimodal automated phenotyping (MAP) algorithm yields a predicted probability of phenotype for each patient and a threshold for classifying participants with phenotype yes/no. The algorithm was validated using labeled data for 16 phenotypes from a biorepository and further tested in an independent cohort phenome-wide association studies (PheWAS) for 2 single nucleotide polymorphisms with known associations.

RESULTS

The MAP algorithm achieved higher or similar AUC and F-scores compared to the ICD code across all 16 phenotypes. The features assembled via the automated approach had comparable accuracy to those assembled via manual curation (AUCMAP 0.943, AUCmanual 0.941). The PheWAS results suggest that the MAP approach detected previously validated associations with higher power when compared to the standard PheWAS method based on ICD codes.

CONCLUSION

The MAP approach increased the accuracy of phenotype definition while maintaining scalability, thereby facilitating use in studies requiring large-scale phenotyping, such as PheWAS.

Risk Factors for Postoperative Venous Thromboembolic Events in Patients Undergoing Lumbar Spine Surgery

STUDY DESIGN

Retrospective, database study.

OBJECTIVES

The aim of this study was to investigate incidence and risk factors associated with venous thromboembolic events (VTEs) after lumbar spine surgery.

METHODS

Patients who underwent lumbar surgery between 2007 and 2014 were identified using the Humana within PearlDiver database. ICD-9 (International Classification of Diseases Ninth Revision) diagnosis codes were used to search for the incidence of VTEs among surgery types, patient demographics and comorbidities. Complications including DVT and PE were queried each day from the day of surgery to postoperative day 7 and for periods 0 to 1 week, 0 to 1 month, 0 to 2 months, and 0 to 3 months postoperatively.

RESULTS

A total of 64 892 patients within the Humana insurance database received lumbar surgery between 2007 and 2014. Overall VTE rate was 0.9% at 1 week, 1.8% at 1 month, and 2.6% at 3 months postoperatively. Among patients that developed a VTE within 1 week postoperatively, 45.3% had a VTE on the day of surgery. Patients with 1 or more identified risk factors had a VTE incidence of 2.73%, compared with 0.95% for patients without risk factors (P < .001). Risk factors associated with the highest VTE incidence and odds ratios (ORs) were primary coagulation disorder (10.01%, OR 4.33), extremity paralysis (7.49%, OR 2.96), central venous line (6.70%, OR 2.87), and varicose veins (6.51%, OR 2.58).

CONCLUSIONS

This study identified several patient comorbidities that were independent predictors of postoperative VTE occurrence after lumbar surgery. Clinical VTE risk assessment may improve with increased focus toward patient comorbidities rather than surgery type or patient demographics.

Outcomes research on obstetric venous thromboembolism

Outcomes research on obstetric venous thromboembolism (VTE) involves a number of major challenges. While obstetric VTE, including deep vein thrombosis and pulmonary embolism, is relatively common on a population basis, diagnoses during pregnancy are relatively rare in comparison to high-risk scenarios such as orthopedic surgery. This review characterizes outcomes research on obstetric VTE with a focus on strengths, limitations, and appropriate inferences from existing research. It is divided into four sections. First, evidence regarding validity of diagnosis codes for VTE in administrative data is reviewed. Second, limitations of both clinical research and administrative-data study models are analyzed. Third, examples of high-quality obstetric VTE research from the literature and opportunities for improved research in the future are reviewed. Fourth, future directions for research are explored.

Large hospital variation in the utilization of Post-procedural CT to detect pulmonary embolism/Deep Vein Thrombosis in Patients Undergoing Total Knee or Hip Replacement Surgery: Japanese Nationwide Diagnosis Procedure Combination Database Study

OBJECTIVE

The purpose of the study was to investigate variation in the use of in-hospital CT for venous thromboembolism (VTE) detection after total knee or hip replacement (TKR/THR) among surgical patients, using a nationwide Japanese in-hospital administrative database.

METHODS

This retrospective study using a national administrative database (4/2012-3/2013) extracted patients who underwent TKR/THR surgeries at hospitals meeting the annual case-volume threshold of ≥ 30. Hospitals were categorized into three equally sized groups by frequency of postoperative CT use (low, middle, and high CT use group) to compare baseline patient-level and hospital-level characteristics. To further investigate between-hospital variation in CT usage, we fitted a hierarchical logistic regression model including hospital-specific random intercepts and fixed patient- and hospital-level effects. The intra class correlation coefficient was used to measure the amount of variability in CT use attributable to between-hospital variation.

RESULTS

A total of 39,127 patients discharged from 447 hospitals met the inclusion criteria. The median hospital stay was 25 days (interquartile range, 20 - 32) and 7,599 (19.4%) patients underwent CT for VTE. CT utilization varied greatly among the hospitals; the crude frequency ranged from 0 to 100 % (median, 7.3 %; interquartile range, 1.8 - 18.3 %). After adjustment for known hospital- and patient-level factors related to CT use, 47 % of the variation in CT use remained attributable to the behavior of individual hospitals.

CONCLUSION

We observed large inter hospital variability in the utilization of post-procedure CT for VTE detection in this Japanese TKR/THR cohort, suggesting that CT utilization is not optimized across the nation.

ADVANCES IN KNOWLEDGE

We observed significant variability in the utilization of post-procedure CT for VTE detection among inpatients who underwent TKR/THR surgeries in a large sample of Japanese hospitals. The large variation suggests that CT utilization is not optimized across the nation, and that there may be potential overutilization of the technology in the highest CT use hospitals.

Comparing external and internal validation methods in correcting outcome misclassification bias in logistic regression: A simulation study and application to the case of postsurgical venous thromboembolism following total hip and knee arthroplasty

PURPOSE

We assessed the validity of postsurgery venous thromboembolism (VTE) diagnoses identified from administrative databases and compared Bayesian and multiple imputation (MI) approaches in correcting for outcome misclassification in logistic regression models.

METHODS

Sensitivity and specificity of postsurgery VTE among patients undergoing total hip or knee replacement (THR/TKR) were assessed against chart review in six Montreal hospitals in 2009 to 2010. Administrative data on all THR/TKR Quebec patients in 2009 to 2010 were obtained. The performance of Bayesian external, Bayesian internal, and MI approaches to correct the odds ratio (OR) of postsurgery VTE in tertiary versus community hospitals was assessed using simulations. Bayesian external approach used prior information from external sources, while Bayesian internal and MI approaches used chart review.

RESULTS

In total, 17 319 patients were included, 2136 in participating hospitals, among whom 75 had VTE in administrative data versus 81 in chart review. VTE sensitivity was 0.59 (95% confidence interval, 0.48-0.69) and specificity was 0.99 (0.98-0.99), overall. The adjusted OR of VTE in tertiary versus community hospitals was 1.35 (1.12-1.64) using administrative data, 1.45 (0.97-2.19) when MI was used for misclassification correction, and 1.53 (0.83-2.87) and 1.57 (0.39-5.24) when Bayesian internal and external approaches were used, respectively. In simulations, all three approaches reduced the OR bias and had appropriate coverage for both nondifferential and differential misclassification.

CONCLUSION

VTE identified from administrative data had low sensitivity and high specificity. The Bayesian external approach was useful to reduce outcome misclassification bias in logistic regression; however, it required accurate specification of the misclassification properties and should be used with caution.

Impact of Medicare's Nonpayment Program on Venous Thromboembolism Following Hip and Knee Replacements

OBJECTIVE

To determine whether Medicare's Nonpayment Program was associated with decline in venous thromboembolism (VTE) following hip and knee replacements; and whether the decline was greater among hospitals at risk of larger financial losses from the Program.

DATA SOURCES

State Inpatient Database for New York (NY) from 2005 to 2013.

STUDY DESIGN

The primary outcome was an occurrence of VTE. Medicare Utilization Ratio (MUR), which is the proportion of inpatient days in a hospital that is financed by Medicare, represented a hospital's financial sensitivity. We used hierarchical logistic regressions with difference-in-differences estimation to study the Program effects.

PRINCIPAL FINDINGS

A total of 98,729 hip replacement and 111,361 knee replacement stays were identified. For hip replacement, the Program was associated with significant reduction (Range: 44% to 53%) in VTE incidence among hospitals in MUR quartiles 2 to 4. For knee replacement, the Program was associated with significant reduction (47%) in VTE incidence only among quartile 2 hospitals.

CONCLUSION

Implementation of the Program was associated with a reduction in VTE, especially for hip replacements, in higher MUR hospitals. Payment reforms such as Medicare's Nonpayment Program that withhold payments for complications are effective and should be continued.

Chart validation of inpatient International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) administrative diagnosis codes for venous thromboembolism (VTE) among intravenous immune globulin (IGIV) users in the Sentinel Distributed Database

The Sentinel Distributed Database (SDD) is a database of patient administrative healthcare records, derived from insurance claims and electronic health records, sponsored by the US Food and Drug Administration for evaluation of medical product outcomes. There is limited information on the validity of diagnosis codes for acute venous thromboembolism (VTE) in the SDD and administrative healthcare data more generally.In this chart validation study, we report on the positive predictive value (PPV) of inpatient administrative diagnosis codes for acute VTE-pulmonary embolism (PE) or lower-extremity or site-unspecified deep vein thrombosis (DVT)-within the SDD. As part of an assessment of thromboembolic adverse event risk following treatment with intravenous immune globulin (IGIV), charts were obtained for 75 potential VTE cases, abstracted, and physician-adjudicated.VTE status was determined for 62 potential cases. PPVs for lower-extremity DVT and/or PE were 90% (95% CI: 73-98%) for principal-position diagnoses, 80% (95% CI: 28-99%) for secondary diagnoses, and 26% (95% CI: 11-46%) for position-unspecified diagnoses (originating from physician claims associated with an inpatient stay). Average symptom onset was 1.5 days prior to hospital admission (range: 19 days prior to 4 days after admission).PPVs for principal and secondary VTE discharge diagnoses were similar to prior study estimates. Position-unspecified diagnoses were less likely to represent true acute VTE cases.

Surrogate-assisted feature extraction for high-throughput phenotyping

Objective

Phenotyping algorithms are capable of accurately identifying patients with specific phenotypes from within electronic medical records systems. However, developing phenotyping algorithms in a scalable way remains a challenge due to the extensive human resources required. This paper introduces a high-throughput unsupervised feature selection method, which improves the robustness and scalability of electronic medical record phenotyping without compromising its accuracy.

Methods

The proposed Surrogate-Assisted Feature Extraction (SAFE) method selects candidate features from a pool of comprehensive medical concepts found in publicly available knowledge sources. The target phenotype's International Classification of Diseases, Ninth Revision and natural language processing counts, acting as noisy surrogates to the gold-standard labels, are used to create silver-standard labels. Candidate features highly predictive of the silver-standard labels are selected as the final features.

Results

Algorithms were trained to identify patients with coronary artery disease, rheumatoid arthritis, Crohn's disease, and ulcerative colitis using various numbers of labels to compare the performance of features selected by SAFE, a previously published automated feature extraction for phenotyping procedure, and domain experts. The out-of-sample area under the receiver operating characteristic curve and F -score from SAFE algorithms were remarkably higher than those from the other two, especially at small label sizes.

Conclusion

SAFE advances high-throughput phenotyping methods by automatically selecting a succinct set of informative features for algorithm training, which in turn reduces overfitting and the needed number of gold-standard labels. SAFE also potentially identifies important features missed by automated feature extraction for phenotyping or experts.

Validity of Using Inpatient and Outpatient Administrative Codes to Identify Acute Venous Thromboembolism: The CVRN VTE Study

BACKGROUND

Administrative data are frequently used to identify venous thromboembolism (VTE) for research and quality reporting. However, the validity of these codes, particularly in outpatients, has not been well-established.

OBJECTIVE

To determine how well International Classification of Diseases, Ninth Revision (ICD-9) codes for VTE predict chart-confirmed acute VTE in inpatient and outpatients.

PATIENTS AND METHODS

We selected 4642 adults with an incident ICD-9 diagnosis of VTE between years 2004 and 2010 from the Cardiovascular Research Network Venous Thromboembolism cohort study. Medical charts were reviewed to determine validity of events. Positive predictive values (PPVs) of ICD-9 codes were calculated as the number of chart-validated VTE events divided by the number with specific VTE codes. Analyses were stratified by VTE type [pulmonary embolism (PE), deep venous thrombosis (DVT)], code position (primary, secondary), and setting [hospital/emergency department (ED), outpatient].

RESULTS

The PPV for any diagnosis of VTE was 64.6% for hospital/ED patients and 30.9% for outpatients. Primary diagnosis codes from hospital/ED patients were more likely to represent acute VTE than secondary diagnosis codes (78.9% vs. 44.4%, P<0.001). Primary hospital/ED codes for PE and lower extremity DVT had higher PPV than for upper extremity DVT (89.1%, 74.9%, and 58.1%, respectively). Outpatient codes were poorly predictive of acute VTE: 28.0% for PE and 53.6% for lower extremity DVT.

CONCLUSIONS

ICD-9 codes for VTE obtained from outpatient encounters or from secondary diagnosis codes do not reliably reflect acute VTE. More accurate ways of identifying VTE in outpatients are needed before these codes can be adopted for research or policy purposes.

Statin use and venous thromboembolism in cancer: A large, active comparator, propensity score matched cohort study

BACKGROUND

Statins have been shown to have a protective effect for venous thromboembolism (VTE) in the general population. This study sought to assess the association between statins and the risk for cancer-associated deep vein thrombosis (DVT) and pulmonary embolism (PE).

METHODS

Patients with newly diagnosed cancer were followed for up to one year in a healthcare claims database (2010-2013). Three treatment groups included statin users, non-statin cholesterol lowering medication users, and an untreated group with pre-existing indications for statin therapy (hyperlipidemia, diabetes, or heart disease). Propensity score matched groups were compared using competing risks survival models for DVT and PE outcomes reporting the hazard ratios (HR) between the treatment groups. Sensitivity analyses assessed the influence of age and individual medications.

RESULTS

The total cohort included 170,459 patients, which, after matching, were similar on baseline characteristics. The overall model showed a statistically significant protective effect for statins compared to no treatment attributed only to leukemia for DVT (HR=0.77, 95% CI 0.61-0.99) and colorectal cancers for PE (HR=0.80, 95% CI 0.64-0.99) in stratified analyses. There were generally no differences in outcomes between statins and non-statins and no individual statin use showed results different from the class effect.

CONCLUSIONS

In this propensity score matched sample of patients with cancer, statins were shown to have a small protective effect in some cancers for DVT or PE compared to no treatment and little difference compared to an active control group. The lack of effect was consistent across statins and was also not found for any of the sensitivity analyses included.

Impact of Time-Varying Treatment Exposures on the Risk of Venous Thromboembolism in Multiple Myeloma

Multiple myeloma (MM) has one of the highest risks of venous thromboembolism (VTE) of all cancers due to pathologic changes and treatment-related exposures. This study assessed the one-year incidence of VTE in newly diagnosed MM and to determine the baseline and time-varying treatment-related factors associated with VTE risk in a U.S.-based cohort. MM patients were identified and age, gender, and baseline comorbidities were determined. Treatment-related exposures included thalidomide derivatives (IMIDs), proteasome inhibitors, cytotoxic chemotherapy, steroids, erythropoietin-stimulating agents (ESAs), stem cell transplants (SCT), hospitalizations, infection, and central venous catheters (CVC). Multiple statistical models were used including a baseline competing risks model, a time-varying exposure Cox proportional hazard (CPH) model, and a case-time-control analysis. The overall incidence of VTE was 107.2 per 1000 person-years with one-half of the VTEs occurring in the first 90 days. The baseline model showed that increasing age, heart failure, and hypertension were associated with one-year incidence of VTE. MM-specific IMID treatment had lower than expected associations with VTE based on prior literature. Instead, exposure to ESAs, SCT, CVC, and infection had higher associations. Based on these results, VTE risk in MM may be less straightforward than considering only chemotherapy exposures, and other treatment-related exposures should be considered to determine patient risk.

Direct Medical Costs Attributable to Cancer-Associated Venous Thromboembolism: A Population-Based Longitudinal Study

PURPOSE

The purpose of this study is to estimate medical costs attributable to venous thromboembolism among patients with active cancer.

METHODS

In a population-based cohort study, we used Rochester Epidemiology Project (REP) resources to identify all Olmsted County, Minn. residents with incident venous thromboembolism and active cancer over the 18-year period, 1988-2005 (n = 374). One Olmsted County resident with active cancer without venous thromboembolism was matched to each case on age, sex, cancer diagnosis date, and duration of prior medical history. Subjects were followed forward in REP provider-linked billing data for standardized, inflation-adjusted direct medical costs from 1 year prior to index (venous thromboembolism event date or control-matched date) to the earliest of death, emigration from Olmsted County, or December 31, 2011, with censoring on the shortest follow-up to ensure a similar follow-up duration for each case-control pair. We used generalized linear modeling to predict costs for cases and controls and bootstrapping methods to assess uncertainty and significance of mean adjusted cost differences. Outpatient drug costs were not included in our estimates.

RESULTS

Adjusted mean predicted costs were 1.9-fold higher for cases ($49,351) than for controls ($26,529) (P < .001) from index to up to 5 years post index. Cost differences between cases and controls were greatest within the first 3 months (mean difference = $13,504) and remained significantly higher from 3 months to 5 years post index (mean difference = $12,939).

CONCLUSIONS

Venous thromboembolism-attributable costs among patients with active cancer contribute a substantial economic burden and are highest from index to 3 months, but may persist for up to 5 years.

Limited Hospital Variation in the Use and Yield of CT for Pulmonary Embolism in Patients Undergoing Total Hip or Total Knee Replacement Surgery

Purpose To evaluate the variation among U.S. hospitals in overall use and yield of in-hospital computed tomographic (CT) pulmonary angiography (PA) in patients undergoing total hip replacement (THR) or total knee replacement (TKR) surgery. Materials and Methods Patients in the Premier Research Database who underwent elective TKR or THR between 2007 and 2011 were enrolled in this HIPAA-compliant, institutional review board-approved retrospective observational study. The informed consent requirement was waived. Hospitals were categorized into low, medium, and high tertiles of CT PA use to compare baseline patient- and hospital-level characteristics and pulmonary embolism (PE) positivity rates. To further investigate between-hospital variation in CT PA use, a hierarchical logistic regression model that included hospital-specific random effects and fixed patient- and hospital-level effects was used. The intraclass correlation coefficient (ICC) was used to measure the amount of variability in CT PA use attributable to between-hospital variation. Results The cohort included 205 198 patients discharged from 178 hospitals (median of 734.5 patients discharged per hospital; interquartile range, 316-1461 patients) with 3647 CT PA studies (1.8%). The crude frequency of CT PA scans among the hospitals ranged from 0% to 6.2% (median, 1.6%); more than 90% of the hospitals performed CT PA in less than 3% of their patients. The mean hospital-level PE positivity rate was 12.3% (median, 9.1%); there was no significant difference in PE positivity rate across low through high CT PA use tertiles (11.3%, 11.9%, 12.9%, P = .37). After adjustment for hospital- and patient-level factors, the remaining amount of interhospital variation was relatively low (ICC, 9.0%). Conclusion Limited interhospital variation in use and yield of in-hospital CT PA was observed among patients undergoing TKR or THR in the United States. ^© RSNA, 2016 Online supplemental material is available for this article.

Predictive Validity of the Beers and Screening Tool of Older Persons' Potentially Inappropriate Prescriptions (STOPP) Criteria to Detect Adverse Drug Events, Hospitalizations, and Emergency Department Visits in the United States

OBJECTIVES

To compare the predictive validity of the 2003 Beers, 2012 American Geriatrics Society (AGS) Beers, and Screening Tool of Older Persons' potentially inappropriate Prescriptions (STOPP) criteria.

DESIGN

Retrospective cohort.

SETTING

Managed care administrative claims data from 2006 to 2009.

PARTICIPANTS

Commercially insured persons aged 65 and older in the United States (N=174,275).

MEASUREMENTS

Association between adverse drug events (ADEs), emergency department (ED) visits, and hospitalization outcomes and inappropriate medication use using time-varying Cox proportional hazard models. Measures of model discrimination (c-index) and hazard ratios (HRs) were calculated to compare unadjusted and adjusted models for associations.

RESULTS

The prevalence of inappropriate prescribing was 34.1% for the 2012 AGS Beers criteria, 32.2% for the 2003 Beers criteria, and 27.6% for the STOPP criteria. Each set of criteria modestly discriminated ADEs in unadjusted analyses (STOPP criteria: hazard ratio (HR)=2.89, 95% confidence interval (CI)=2.68-3.12, C-index=0.607; 2012 AGS Beers criteria: HR=2.51, 95% CI=2.33-2.70, C-index=0.603; 2003 Beers criteria: HR=2.65, 95% CI=2.46-2.85, C-index=0.605). Similar results were observed for ED visits and hospitalizations. The c-indices increased to between 0.65 and 0.70 in adjusted analyses. The kappa for agreement between criteria was 0.80 for the 2003 and 2012 AGS Beers criteria, 0.58 for the 2012 AGS Beers and STOPP criteria, and 0.59 for the 2003 Beers and STOPP criteria. For the three outcomes, the 2012 AGS Beers criteria had the highest sensitivity (61.2-71.2%) and the lowest specificity (41.2-70.7%), and the STOPP criteria had the lowest sensitivity (53.8-64.7%) but the highest specificity (47.8-78.1%).

CONCLUSION

All three criteria were modestly prognostic for ADEs, EDs, and hospitalizations, with the STOPP criteria slightly outperforming both Beers criteria. With low sensitivity, low specificity, and low agreement between the criteria, they can be used in a complementary fashion to enhance sensitivity in detecting ADEs.

Incidence and Risk Factors of Thromboembolism with Multiple Myeloma in the Presence of Death as a Competing Risk: An Empirical Comparison of Statistical Methodologies

Multiple myeloma (MM) has an inherent high risk of thromboembolic events associated with patient as well as disease- and treatment-related factors. Previous studies have assessed the association of MM-related thromboembolism using "traditional" Kaplan-Meier (KM) and/or Cox proportional hazard (PH) regression. In the presence of high incidence of death, as would be the case in cancer patients with advanced age, these statistical models will produce bias estimates. Instead, a competing risk framework should be used. This study assessed the baseline patient demographic and clinical characteristics associated with MM-related thromboembolism and compared the cumulative incidence and the measures of association obtained using each statistical approach. The cumulative incidence of thromboembolism was 9.2% using the competing risk framework and nearly 12% using the KM approach. Bias in the measures of covariate risk associations was highest for factors related to risk of death such as increased age (75% bias) and severe liver disease (50%) for the Cox PH model compared to the competing risk model. These results show that correct specification of statistical techniques can have a large impact on the results obtained.

The effect of chronic liver disease on venous thromboembolism among medically managed patients in Singapore General Hospital

BACKGROUND

Chronic liver disease (CLD) has been suggested to be associated with venous thromboembolism (VTE) in western populations. However, little is known about the risk effects of CLD on VTE among Asians.

OBJECTIVE

To compute the prevalence of VTE among hospitalised Asian patients, and to evaluate the pattern and scale of risk effects of CLD on VTE occurrence.

METHOD

Retrospective study of hospital discharge database from 2004 to 2011 to identify patients with VTE and CLD using International Statistical Classification of Diseases and Related Health Problems, 9th Revision, Australian Modification (ICD-9-AM) codes.

RESULTS

Of 199904 medically managed inpatients during the 8years, 1744 (0.9%) patients had VTE. Patients with CLD had significant higher prevalence of VTE (non-cirrhosis CLD 1.5%, cirrhosis 2.0%) than patients without CLD (0.8%, p<0.001). In the logistic regression analyses, non-cirrhosis CLD (odds ratio, OR 1.4, 95% CI 1.2-1.7, p<0.001) and cirrhosis (OR 1.5, 95% CI 1.2-2.0, p=0.002) were significant predictors of VTE after adjustment for age, gender, ethnicity, hospital long stayer, cancer, infectious disease, and other comorbid conditions such as diabetic mellitus, anaemia, and cardiovascular, cerebrovascular, renal and pulmonary diseases.

CONCLUSION

CLD, particular liver cirrhosis, increases the risks of VTE in hospitalised Asian patients. As CLD patients are perceived to be at risks of bleeding due to the prolonged clotting times and thrombocytopenia, the results of this study brings attention to opposite end of the haemostatic pendulum in patients with chronic liver disease.

Toward high-throughput phenotyping: unbiased automated feature extraction and selection from knowledge sources

OBJECTIVE

Analysis of narrative (text) data from electronic health records (EHRs) can improve population-scale phenotyping for clinical and genetic research. Currently, selection of text features for phenotyping algorithms is slow and laborious, requiring extensive and iterative involvement by domain experts. This paper introduces a method to develop phenotyping algorithms in an unbiased manner by automatically extracting and selecting informative features, which can be comparable to expert-curated ones in classification accuracy.

MATERIALS AND METHODS

Comprehensive medical concepts were collected from publicly available knowledge sources in an automated, unbiased fashion. Natural language processing (NLP) revealed the occurrence patterns of these concepts in EHR narrative notes, which enabled selection of informative features for phenotype classification. When combined with additional codified features, a penalized logistic regression model was trained to classify the target phenotype.

RESULTS

The authors applied our method to develop algorithms to identify patients with rheumatoid arthritis and coronary artery disease cases among those with rheumatoid arthritis from a large multi-institutional EHR. The area under the receiver operating characteristic curves (AUC) for classifying RA and CAD using models trained with automated features were 0.951 and 0.929, respectively, compared to the AUCs of 0.938 and 0.929 by models trained with expert-curated features.

DISCUSSION

Models trained with NLP text features selected through an unbiased, automated procedure achieved comparable or slightly higher accuracy than those trained with expert-curated features. The majority of the selected model features were interpretable.

CONCLUSION

The proposed automated feature extraction method, generating highly accurate phenotyping algorithms with improved efficiency, is a significant step toward high-throughput phenotyping.

The validity of ICD codes coupled with imaging procedure codes for identifying acute venous thromboembolism using administrative data

The purpose of this study was to evaluate the accuracy of using a combination of International Classification of Diseases (ICD) diagnostic codes and imaging procedure codes for identifying deep vein thrombosis (DVT) and pulmonary embolism (PE) within administrative databases. Information from the Alberta Health (AH) inpatients and ambulatory care administrative databases in Alberta, Canada was obtained for subjects with a documented imaging study result performed at a large teaching hospital in Alberta to exclude venous thromboembolism (VTE) between 2000 and 2010. In 1361 randomly-selected patients, the proportion of patients correctly classified by AH administrative data, using both ICD diagnostic codes and procedure codes, was determined for DVT and PE using diagnoses documented in patient charts as the gold standard. Of the 1361 patients, 712 had suspected PE and 649 had suspected DVT. The sensitivities for identifying patients with PE or DVT using administrative data were 74.83% (95% confidence interval [CI]: 67.01-81.62) and 75.24% (95% CI: 65.86-83.14), respectively. The specificities for PE or DVT were 91.86% (95% CI: 89.29-93.98) and 95.77% (95% CI: 93.72-97.30), respectively. In conclusion, when coupled with relevant imaging codes, VTE diagnostic codes obtained from administrative data provide a relatively sensitive and very specific method to ascertain acute VTE.

Cervical Fracture Stabilization within 72 Hours of Injury is Associated with Decreased Hospitalization Costs with Comparable Perioperative Outcomes in a Propensity Score-Matched Cohort

INTRODUCTION

Prior studies have indicated that early decompression of traumatic cervical fractures can be performed safely and is associated with improved outcomes, though the economic impact of the timing of surgery in the American population has not been studied. After adjusting for patient, hospital, and injury confounders, we performed propensity score modeling (PSM) on a large clinical administrative database to determine associated costs depending upon timing of surgery for acute cervical fracture.

METHODS

A total of 3,348 patients with surgically treated, traumatic, cervical fractures were identified. Patients were sorted into early (within 72 hours of admission) and late (beyond 72 hours) surgery groups. PSM was able to match 2,132 early and late surgery patients on age, comorbidity, expected payer, trauma severity, hospital type, urgent admission, and surgical approach. Perioperative complications, mortality, and resource utilization were assessed.

RESULTS

Late surgery was more frequently associated with increased age, more comorbidities, higher ICISS score, and non-private insurance. Following PSM matching, there were no significant, preoperative differences between early and late surgery groups. Surgery performed after 72 hours was associated with an increase in in-hospital complications (OR=1.3). The early surgery group was associated with decreased length of stay (11 days vs. 16 days, p <0.0001) and hospital charges ($237,786 v. $282,727, p <0.0001).

CONCLUSIONS

After controlling for potential confounding differences through PSM matching and multivariate analyses, we found late surgery independently associated with increased in-hospital complications, length of stay, and hospital resource utilization. These data suggest surgery within 72 hours may decrease resource utilization without a corresponding increase in postoperative morbidity.

Risk of venous thromboembolism after shoulder arthroplasty in the Medicare population

BACKGROUND

Chemoprophylaxis for venous thromboembolism (VTE) is considered standard of care after lower but not after upper extremity arthroplasty. Medicare claims data were analyzed to determine the national incidence of symptomatic VTE after shoulder arthroplasty during the index surgical admission and after discharge.

METHODS

Incidence of VTE was compared between shoulder arthroplasty and lower extremity arthroplasty, and patient-specific risk factors for pulmonary embolism were determined for shoulder arthroplasty. Finally, the incidence of surgical site bleeding was determined for each type of procedure.

RESULTS

VTE complications occurred in 1.2% of lower extremity and 0.53% of shoulder arthroplasties. Patient factors that were associated with an increased risk of VTE included a principal diagnosis of fracture, history of VTE, cardiac arrhythmia, presence of a metastatic tumor, coagulopathy, congestive heart failure, alcohol abuse, and obesity. Adjusted rates of VTE were generally higher for lower extremity compared with shoulder arthroplasties. Prevalence of readmission for VTE was higher for shoulder hemiarthroplasty than for total shoulder arthroplasty. Wound hematoma rates were similar between shoulder and lower extremity procedures.

CONCLUSIONS

VTE rates after shoulder arthroplasty were generally lower than those after lower extremity arthroplasty. We believe that the risk of bleeding combined with the lower rates of VTE with existing lower rates of chemoprophylaxis does not warrant the routine use of anticoagulation. Use of mechanical prophylaxis combined with aspirin may be sufficient for shoulder arthroplasty patients who are not at increased risk of VTE. Chemoprophylaxis with agents other than aspirin may be warranted in patients with a demonstrated risk of VTE.

Classification of CT pulmonary angiography reports by presence, chronicity, and location of pulmonary embolism with natural language processing

In this paper we describe an efficient tool based on natural language processing for classifying the detail state of pulmonary embolism (PE) recorded in CT pulmonary angiography reports. The classification tasks include: PE present vs. absent, acute PE vs. others, central PE vs. others, and subsegmental PE vs. others. Statistical learning algorithms were trained with features extracted using the NLP tool and gold standard labels obtained via chart review from two radiologists. The areas under the receiver operating characteristic curves (AUC) for the four tasks were 0.998, 0.945, 0.987, and 0.986, respectively. We compared our classifiers with bag-of-words Naive Bayes classifiers, a standard text mining technology, which gave AUC 0.942, 0.765, 0.766, and 0.712, respectively.

Decreased incidence of venous thromboembolism after spine surgery with early multimodal prophylaxis

Object

Venous thromboembolism (VTE) represents a significant complication after spine surgery, with reported rates as high as 2%–4%. Published institutional practices for VTE prophylaxis are highly variable. In 2008, the authors implemented a departmental protocol for early VTE prophylaxis consisting of combined compressive devices and subcutaneous heparin initiated either preoperatively or on the same day of surgery. In this study, the authors compared the incidence of VTE in spine surgery patients before and after implementing this protocol.

Methods

An institutional review board–approved retrospective review of outcomes in patients undergoing spine surgery 2 years before protocol implementation (representing the preprotocol group) and of outcomes in patients treated 2 years thereafter (the postprotocol group) was conducted. Inclusion criteria were that patients were 18 years or older and had been admitted for 1 or more days. Before 2008 (preprotocol), VTE prophylaxis was variable and provider dependent without any uniform protocol. Since 2008 (postprotocol), a new VTE-prophylaxis protocol was administered, starting either preoperatively or on the same day of surgery and continuing throughout hospitalization. The new protocol consisted of 5000 U heparin administered subcutaneously 3 times daily, except in patients older than 75 years or weighing less than 50 kg, who received this dose twice daily. All patients also received sequential compression devices (SCDs). The incidence of VTE in the 2 protocol phases was identified by codes of the International Classification of Diseases, Ninth Revision (ICD-9) codes for deep vein thrombosis (DVT) and pulmonary embolus (PE). Bleeding complications arising from anticoagulation treatments were evaluated by the Current Procedural Terminology (CPT) code for postoperative epidural hematoma (EDH) requiring evacuation.

Results

In total, 941 patients in the preprotocol group met the inclusion criteria: 25 had DVT (2.7%), 6 had PE (0.6%), and 6 had postoperative EDH (0.6%). In the postprotocol group, 992 patients met the criteria: 10 had DVT (1.0%), 5 had PE (0.5%), and 4 had postoperative EDH (0.4%). This reduction in DVT after the protocol's implementation was statistically significant (p = 0.009). Despite early aggressive prophylaxis, the incidence of postoperative EDH did not increase and compared favorably to the published literature.

Conclusions

At a high-volume tertiary center, an aggressive protocol for early VTE prophylaxis after spine surgery decreases VTE incidence without increasing morbidity.

Validity of AHRQ patient safety indicators derived from ICD-10 hospital discharge abstract data (chart review study)

OBJECTIVE

To assess if the Agency for Healthcare Research and Quality  patient safety indictors (PSIs) could be used for case findings in the International Classification of Disease 10th revision (ICD-10) hospital discharge abstract data.

DESIGN

We identified and randomly selected 490 patients with a foreign body left during a procedure (PSI 5-foreign body), selected infections (IV site) due to medical care (PSI 7-infection), postoperative pulmonary embolism (PE) or deep vein thrombosis (DVT; PSI 12-PE/DVT), postoperative sepsis (PSI 13-sepsis)and accidental puncture or laceration (PSI 15-laceration) among patients discharged from three adult acute care hospitals in Calgary, Canada in 2007 and 2008. Their charts were reviewed for determining the presence of PSIs and used as the reference standard, positive predictive value (PPV) statistics were calculated to determine the proportion of positives in the administrative data representing 'true positives'.

RESULTS

The PPV for PSI 5-foreign body was 62.5% (95% CI 35.4% to 84.8%), PSI 7-infection was 79.1% (67.4% to 88.1%), PSI 12-PE/DVT was 89.5% (66.9% to 98.7%), PSI 13-sepsis was 12.5% (1.6% to 38.4%) and PSI 15-laceration was 86.4% (75.0% to 94.0%) after excluding those who presented to the hospital with the condition.

CONCLUSIONS

Several PSIs had high PPV in the ICD administrative data and are thus powerful tools for true positive case finding. The tools could be used to identify potential cases from the large volume of admissions for verification through chart reviews. In contrast, their sensitivity has not been well characterised and users of PSIs should be cautious if using them for 'quality of care reporting' presenting the rate of PSIs because under-coded data would generate falsely low PSI rates.

The incidence and mortality of thromboembolic events in cervical spine surgery

STUDY DESIGN

Retrospective database analysis.

OBJECTIVE

A nationwide population-based database was analyzed to identify the incidence, risk factors, and mortalities associated with venous thromboembolism (VTE) after cervical spine surgery.

SUMMARY OF BACKGROUND DATA

Pulmonary embolism (PE) and deep vein thrombosis (DVT) are potential complications that may occur after orthopedic procedures. Incidences of these complications are not well characterized after cervical spine surgery.

METHODS

Data from the Nationwide Inpatient Sample database were obtained from 2002 to 2009. Patients undergoing anterior cervical fusion, posterior cervical fusion, and posterior cervical decompression (i.e., laminoforaminotomy, laminectomy, laminoplasty) for the diagnosis of cervical myelopathy and/or radiculopathy were identified. Incidences of PE and DVT were calculated. Comorbidities were calculated using the modified Charlson Comorbidity Index. Mortality associated with these complications was assessed in the 3 surgical subgroups. Statistical analysis was performed to assess significant differences between groups. Logistic regression was used to identify independent predictors of VTE. A P value of <0.0005 was used to denote significance.

RESULTS

There were 273,396 cervical procedures recorded in the Nationwide Inpatient Sample database from 2002 to 2009. Posterior cervical fusion-treated patients had statistically the highest incidences of DVT and PE, whereas the lowest PE and DVT rates were found in anterior cervical fusion-treated patients (P < 0.0005). All patients with thromboembolic events had significantly increased rates of mortality, hospitalization, and costs compared with patients without VTE across all procedural groups. Logistic regression analysis demonstrated statistically significant predictors of VTE to be male sex, pulmonary circulation disorders, fluid/electrolyte disorders, and teaching-hospital status.

CONCLUSION

Thromboembolic events are potential complications of cervical spine surgery. The highest rates of VTE were identified in those patients undergoing posterior cervical fusion. Regardless of approach, DVT and PEs resulted in increased mortality rates and hospitalization. We recommend a thorough preoperative assessment to identify patients at risk for VTE and treat accordingly to decrease the incidence of these thromboembolic events.

Vascular disease in women: comparison of diagnoses in hospital episode statistics and general practice records in England

Background

Electronic linkage to routine administrative datasets, such as the Hospital Episode Statistics (HES) in England, is increasingly used in medical research. Relatively little is known about the reliability of HES diagnostic information for epidemiological studies. In the United Kingdom (UK), general practitioners hold comprehensive records for individuals relating to their primary, secondary and tertiary care. For a random sample of participants in a large UK cohort, we compared vascular disease diagnoses in HES and general practice records to assess agreement between the two sources.

Methods

Million Women Study participants with a HES record of hospital admission with vascular disease (ischaemic heart disease [ICD-10 codes I20-I25], cerebrovascular disease [G45, I60-I69] or venous thromboembolism [I26, I80-I82]) between April 1^st 1997 and March 31^st 2005 were identified. In each broad diagnostic group and in women with no such HES diagnoses, a random sample of about a thousand women was selected for study. We asked each woman’s general practitioner to provide information on her history of vascular disease and this information was compared with the HES diagnosis record.

Results

Over 90% of study forms sent to general practitioners were returned and 88% of these contained analysable data. For the vast majority of study participants for whom information was available, diagnostic information from general practice and HES records was consistent. Overall, for 93% of women with a HES diagnosis of vascular disease, general practice records agreed with the HES diagnosis; and for 97% of women with no HES diagnosis of vascular disease, the general practitioner had no record of a diagnosis of vascular disease. For severe vascular disease, including myocardial infarction (I21-22), stroke, both overall (I60-64) and by subtype, and pulmonary embolism (I26), HES records appeared to be both reliable and complete.

Conclusion

Hospital admission data in England provide diagnostic information for vascular disease of sufficient reliability for epidemiological analyses.

Should policy-makers and managers trust PSI? An empirical validation study of five patient safety indicators in a national health service

BACKGROUND

Patient Safety Indicators (PSI) are being modestly used in Spain, somewhat due to concerns on their empirical properties. This paper provides evidence by answering three questions: a) Are PSI differences across hospitals systematic -rather than random?; b) Do PSI measure differences among hospital-providers -as opposed to differences among patients?; and, c) Are measurements able to detect hospitals with a higher than "expected" number of cases?

METHODS

An empirical validation study on administrative data was carried out. All 2005 and 2006 publicly-funded hospital discharges were used to retrieve eligible cases of five PSI: Death in low-mortality DRGs (MLM); decubitus ulcer (DU); postoperative pulmonary embolism or deep-vein thrombosis (PE-DVT); catheter-related infections (CRI), and postoperative sepsis (PS). Empirical Bayes statistic (EB) was used to estimate whether the variation was systematic; logistic-multilevel modelling determined what proportion of the variation was explained by the hospital; and, shrunken residuals, as provided by multilevel modelling, were plotted to flag hospitals performing worse than expected.

RESULTS

Variation across hospitals was observed to be systematic in all indicators, with EB values ranging from 0.19 (CI95%:0.12 to 0.28) in PE-DVT to 0.34 (CI95%:0.25 to 0.45) in DU. A significant proportion of the variance was explained by the hospital, once patient case-mix was adjusted: from a 6% in MLM (CI95%:3% to 11%) to a 24% (CI95%:20% to 30%) in CRI. All PSI were able to flag hospitals with rates over the expected, although this capacity decreased when the largest hospitals were analysed.

CONCLUSION

Five PSI showed reasonable empirical properties to screen healthcare performance in Spanish hospitals, particularly in the largest ones.

Outpatient use of low molecular weight heparin monotherapy for first-line treatment of venous thromboembolism in advanced cancer

BACKGROUND

Evidence-based treatment guidelines recommend low molecular weight heparin (LMWH) monotherapy for cancer-associated venous thromboembolism (VTE). This analysis assessed the first-line treatment strategies for VTE in patients with advanced solid tumors.

METHODS

Using administrative data from advanced lung, prostate, colon, or breast cancer patients diagnosed between January 2000 and December 2007 at four HMOs with integrated delivery systems, patients with an inpatient or outpatient VTE diagnosed within 2 years after cancer diagnosis and an outpatient purchase of warfarin, LMWH, and/or fondaparinux anticoagulant within 7 days of the VTE diagnosis were identified. First-line outpatient VTE pharmacological treatment and factors independently associated with receipt/non-receipt of LMWH monotherapy were assessed.

RESULTS

Overall, 25% of the 1,089 eligible patients received LMWH monotherapy as primary VTE treatment. The percentage increased steadily over time from 18% among patients diagnosed in 2000 to 31% among those diagnosed in 2007. Factors associated with LMWH monotherapy included VTE diagnosis year, chemotherapy within 60 days prior to VTE diagnosis, history of VTE prior to cancer diagnosis, and invasive surgery in the 90 days following VTE diagnosis. Colorectal and prostate cancer patients versus lung cancer patients and stage III versus stage IV patients were less likely to be treated with LMWH monotherapy.

CONCLUSIONS

Adoption of LMWH monotherapy as initial treatment for cancer-associated VTE was low but increased steadily over the study period. Future studies should explore reasons underlying the underutilization of this preferred evidence-based treatment as well as the comparative effectiveness of LMWH versus warfarin-based anticoagulation in real-world cancer patients with VTE.

Pulmonary embolism prophylaxis in more than 30,000 total knee arthroplasty patients: is there a best choice?

Prophylaxis for pulmonary embolism (PE) prevention in total knee arthroplasty remains controversial. A joint registry evaluated venous thromboembolism prophylaxis and anesthesia impact on the incidence of PE, fatal PE, and death. Patients received mechanical prophylaxis alone or chemical with or without mechanical prophylaxis. The overall PE incidence was 0.45%; fatal PE, 0.01%; and death, 0.31%. The only significant difference in any outcome was the incidence of PE between Coumadin and mechanical prophylaxis alone. Variables associated with a higher incidence of PE were age, an American Society of Anesthesiologists score of 3 or higher, and the use of general anesthesia. Based on the findings, general anesthesia can be discouraged, and only Coumadin fared better than mechanical prophylaxis alone, whereas other forms of chemical prophylaxis revealed no significant differences.

Venous thromboembolism after thoracic/thoracolumbar spinal fusion

OBJECTIVE

Venous thromboembolism (VTE), which includes deep venous thrombosis and pulmonary embolism, is a serious and potentially fatal surgical complication. The goal of our study was to examine preoperative characteristics, incidence, and outcomes of patients with VTE after elective thoracic/thoracolumbar level spine fusion.

METHODS

We identified 430,081 patients from the Nationwide Inpatient Sample database who underwent spinal fusion between 2002 and 2008. Patients undergoing thoracic/thoracolumbar level fusion (n = 8617) were found to have the greatest concurrent rate of VTE. We then performed multivariate analyses on this cohort to identify predictors of and outcomes after VTE in patients undergoing thoracic/thoracolumbar level fusion.

RESULTS

The overall VTE rate in spinal fusion surgery was 0.40% (cervical = 0.22%, thoracic/thoracolumbar = 1.90%, lumbar/lumbosacral = 0.49%, re-fusions = 0.64%, and fusions not otherwise specified = 0.84%). On multivariate logistic regression analysis of patients undergoing spinal fusion at the thoracic/thoracolumbar level, increasing age, Medicare insurance coverage (vs. private insurance), urban teaching hospital (vs. urban nonteaching hospital), combined anterior/posterior surgical approach (vs. posterior-only approach), and the presence of congestive heart failure or weight loss (Elixhauser comorbidity groups) were each independently associated with an increased odds ratio of VTE complication. VTE after thoracic/thoracolumbar surgery was significantly associated with longer hospital stays (16.6 vs. 6.74 days), increased total hospital costs ($260,208 vs. $115,474), and increased mortality (4.33% vs. 0.33%).

CONCLUSIONS

Multivariate logistic regression analysis reveals age, insurance status, hospital type, combined anterior/posterior surgical approach, and the presence of congestive heart failure or weight loss to be independently associated with an increased odds ratio of VTE complication. This complication is associated with increased hospital costs, length of stay, and overall mortality.

Measuring the cost of hospital adverse patient safety events

This paper estimates the excess cost of hospital inpatient care due to adverse safety events in the U.S. Department of Veterans Affairs (VA) hospitals during fiscal year 2007. We measured adverse events according to the Patient Safety Indicator (PSI) algorithms of the Agency for Healthcare Research and Quality. Patient level cost regression analyses were performed using generalized linear modeling techniques. Accounting for the heavily skewed distribution of costs among patients having adverse safety events, results suggested that the excess cost of nine different PSIs for VA patients are much higher than previously estimated. We tested sensitivity of results to whether costs were measured by VA's Decision Support System (DSS) that uses local costs of specific inputs, or by the average costing system developed by VA's Health Economics Resource Center. DSS costing appeared to better characterize the high cost patients.

Developing public health surveillance for deep vein thrombosis and pulmonary embolism

Deep vein thrombosis (DVT) and pulmonary embolism (PE), collectively known as venous thromboembolism (VTE), are an important and growing public health issue, associated with considerable morbidity and mortality. Presently, there is no national surveillance for DVT and PE. This article provides a summary of an expert workgroup meeting convened January 12, 2010, by the CDC. The purpose of the meeting was to inform CDC on the development of U.S. population-based public health surveillance activities for DVT/PE. Topics discussed included: (1) stakeholders, needs, gaps, and target populations; (2) requirements of surveillance systems; (3) challenges, limitations, and potential barriers to implementation of surveillance activities; and (4) integration of research and education with surveillance activities.

Venous thromboembolism: deep venous thrombosis and pulmonary embolism in a neurosurgical population

Object

Venous thromboembolism (VTE), a combination of deep venous thrombosis (DVT) and pulmonary embolism (PE), is a major cause of morbidity and death in neurosurgical patients. This study evaluates 1) the risk of developing lower-extremity DVT following a neurosurgical procedure; 2) the timing of initiation of pharmacological DVT prophylaxis upon the occurrence of VTE; and 3) the relationship between DVT and PE as related to VTE prophylaxis in neurosurgical patients.

Methods

The records of all neurosurgical patients between January 2006 and December 2008 (2638 total) were reviewed for clinical documentation of VTE. As part of a quality improvement initiative, a subgroup of 1638 patients was studied during the implementation of pharmacological prophylaxis. A high-risk group of 555 neurosurgical patients in the intensive care unit underwent surveillance venous lower-extremity duplex ultrasonography studies twice weekly. All patients throughout the review received mechanical DVT prophylaxis. Pharmacological DVT prophylaxis, consisting of 5000 U of subcutaneous heparin twice daily (initially started within 48 hours of a neurosurgical procedure and subsequently within 24 hours of a procedure) was implemented in combination with mechanical prophylaxis. The DVT and PE rates were calculated for each group.

Results

In the surveillance group (555 patients), 84% of the DVTs occurred within 1 week and 92% within 2 weeks of a neurosurgical procedure. There was a linear correlation between the duration of surgery and DVT development. The use of subcutaneous heparin reduced the rate of DVT from 16% to 9% when medication was given at either 24 or 48 hours postoperatively, without any increase in hemorrhagic complications. In the overall group (2638 patients), there were 94 patients who exhibited clinical signs of a possible PE and therefore underwent spiral CT; 22 of these patients (0.8%) had radiological confirmation of PE. There was no correlation between the use of pharmacological prophylaxis at either time point and the occurrence of PE, despite a 43% reduction in the lower-extremity DVT rate with pharmacological intervention.

Conclusions

The majority of DVTs occurred within the first week after a neurosurgical procedure. There was a linear correlation between the duration of surgery and DVT occurrence. Use of early subcutaneous heparin (at either 24 or 48 hours) was associated with a 43% reduction of developing a lower-extremity DVT, without an increase in surgical site hemorrhage. There was no association of pharmacological prophylaxis with overall PE occurrence.

Validity of selected Patient Safety Indicators: opportunities and concerns

BACKGROUND

The Agency for Healthcare Research and Quality (AHRQ) recently designed the Patient Safety Indicators (PSIs) to detect potential safety-related adverse events. The National Quality Forum has endorsed several of these ICD-9-CM-based indicators as quality-of-care measures. We examined the positive predictive value (PPV) of 3 surgical PSIs: postoperative pulmonary embolus and deep vein thrombosis (pPE/DVT), iatrogenic pneumothorax (iPTX), and accidental puncture and laceration (APL).

STUDY DESIGN

We applied the AHRQ PSI software (v.3.1a) to fiscal year 2003 to 2007 Veterans Health Administration (VA) administrative data to identify (flag) patients suspected of having a pPE/DVT, iPTX, or APL. Two trained nurse abstractors reviewed a sample of 336 flagged medical records (112 records per PSI) using a standardized instrument. Inter-rater reliability was assessed.

RESULTS

Of 2,343,088 admissions, 6,080 were flagged for pPE/DVT (0.26%), 1,402 for iPTX (0.06%), and 7,203 for APL (0.31%). For pPE/DVT, the PPV was 43% (95% CI, 34% to 53%); 21% of cases had inaccurate coding (eg, arterial not venous thrombosis); and 36% featured thromboembolism present on admission or preoperatively. For iPTX, the PPV was 73% (95% CI, 64% to 81%); 18% had inaccurate coding (eg, spontaneous pneumothorax), and 9% were pneumothoraces present on admission. For APL, the PPV was 85% (95% CI, 77% to 91%); 10% of cases had coding inaccuracies and 5% indicated injuries present on admission. However, 27% of true APLs were minor injuries requiring no surgical repair (eg, small serosal bowel tear). Inter-rater reliability was >90% for all 3 PSIs.

CONCLUSIONS

Until coding revisions are implemented, these PSIs, especially pPE/DVT, should be used primarily for screening and case-finding. Their utility for public reporting and pay-for-performance needs to be reassessed.

Medical complications in patients with stroke: data validity in a stroke registry and a hospital discharge registry

Background:

Stroke patients frequently experience medical complications; yet, data on incidence, causes, and consequences are sparse.

Objective:

To examine the data validity of medical complications among patients with stroke in a population-based clinical registry and a hospital discharge registry.

Methods:

We examined the predictive values, sensitivity and specificity of medical complications among patients admitted to specialized stroke units and registered in the Danish National Indicator Project (DNIP) and the Danish National Registry of Patients (NRP) between January 2003 and December 2006 (n = 8,024). We retrieved and reviewed medical records from a random sample of patients (n = 589, 7.3%).

Results:

We found substantial variation in the data quality of stroke-related medical complication diagnoses both within the specific complications and between the registries. The positive predictive values ranged from 39.0%–87.1% in the DNIP, and from 0.0%–92.9% in the NRP. The negative predictive values ranged from 71.6%–98.9% in the DNIP and from 63.3% to 97.4% in the NRP. In both registries the specificity of the diagnoses was high. The sensitivity ranged from 23.5% (95% confidence interval [CI]: 14.9–35.4) for falls to 62.9% (95% CI: 54.9–70.4) for urinary infection in the DNIP, and from 0.0 (95% CI: 0.0–4.99) for falls to 18.1% (95% CI: 2.3–51.8) for pressure ulcer in the NRP.

Conclusion:

The DNIP may be useful for studying medical complications among patients with stroke.

Surveillance for deep vein thrombosis and pulmonary embolism: recommendations from a national workshop

Deep vein thrombosis (DVT) and pulmonary embolism (PE), known collectively as venous thromboembolism (VTE), affect an estimated 900,000 people in the U.S. each year, resulting in several hundred thousand hospitalizations and about 300,000 deaths. Despite this substantial public health burden, no systematic collection of VTE-related morbidity and mortality data exists in the U.S. The available information about disease prevalence and incidence consists of estimates based mainly on population-based epidemiologic studies and analysis of hospital discharge or health insurance claims databases. The limited scope of the available data has raised the question of whether a systematic surveillance system for VTE should be established. To help answer this question and to make recommendations for the next steps toward better surveillance of VTE, the CDC asked the American Society of Hematology (ASH) to convene a national workshop of stakeholders representing relevant federal agencies, experts in VTE epidemiology and treatment, public health experts in VTE, and patient representatives. These groups were assembled by ASH for a 1-day meeting in Washington DC. The subspecialty experts included representatives from internal medicine, cardiovascular diseases, adult and pediatric hematology, surgery, obstetrics and gynecology, radiology, emergency medicine and trauma care, hospital practice and critical care, and geriatrics. Experts in epidemiology, healthcare quality, and health policy also participated. During the workshop, experts discussed their perspectives on the burden of disease from VTE and its diagnosis, treatment, and prevention. The workshop also focused on the advisability and feasibility of establishing systematic surveillance for VTE and included preliminary discussion of the advantages and disadvantages of various approaches. The workshop concluded that (1) improved utilization in clinical practice of existing, proven-effective preventive measures is critical to reducing the disease burden from VTE; (2) systematic surveillance of DVT and PE is needed to provide nationally representative data on the prevalence and annual incidence of DVT and PE in the U.S.; (3) tracking and documenting changes in the incidence of DVT and PE through systematic surveillance will be important to enhance prevention efforts; and (4) the CDC should convene a second group of experts to advise the agency in detail on the strengths, weaknesses, and feasibility of possible approaches to systematic surveillance for DVT and PE.

ICD-10 hospital discharge diagnosis codes were sensitive for identifying pulmonary embolism but not deep vein thrombosis

OBJECTIVE

To estimate the sensitivity of International Classification of Diseases, Tenth revision (ICD-10) hospital discharge diagnosis codes for identifying deep vein thrombosis (DVT) and pulmonary embolism (PE).

STUDY DESIGN AND SETTING

We compared predefined ICD-10 discharge diagnosis codes with the diagnoses that were prospectively recorded for 1,375 patients with suspected DVT or PE who were enrolled at 25 hospitals in France. Sensitivity was calculated as the percentage of patients identified by predefined ICD-10 codes among positive cases of acute symptomatic DVT or PE confirmed by objective testing.

RESULTS

The sensitivity of ICD-10 codes was 58.0% (159 of 274; 95% CI: 51.9, 64.1) for isolated DVT and 88.9% (297 of 334; 95% CI: 85.6, 92.2) for PE. Depending on the hospital, the median values for sensitivity were 57.7% for DVT (interquartile range, IQR, 48.6-66.7; intracluster correlation coefficient, 0.02; P=0.31) and 88.9% for PE (IQR, 83.3-96.3; intracluster correlation coefficient, 0.11; P=0.03). The sensitivity of ICD-10 codes was lower for surgical patients and for patients who developed PE or DVT while they were hospitalized.

CONCLUSION

ICD-10 discharge diagnosis codes yield satisfactory sensitivity for identifying objectively confirmed PE. A substantial proportion of DVT cases are missed when using hospital discharge data for complication screening or research purposes.

Effects of resident duty hour reform on surgical and procedural patient safety indicators among hospitalized Veterans Health Administration and Medicare patients

OBJECTIVE

Improving patient safety was a strong motivation behind duty hour regulations implemented by Accreditation Council for Graduate Medical Education on July 1, 2003. We investigated whether rates of patient safety indicators (PSIs) changed after these reforms.

RESEARCH DESIGN

Observational study of patients admitted to Veterans Health Administration (VA) (N = 826,047) and Medicare (N = 13,367,273) acute-care hospitals from July 1, 2000 to June 30, 2005. We examined changes in patient safety events in more versus less teaching-intensive hospitals before (2000-2003) and after (2003-2005) duty hour reform, using conditional logistic regression, adjusting for patient age, gender, comorbidities, secular trends, baseline severity, and hospital site.

MEASURES

Ten PSIs were aggregated into 3 composite measures based on factor analyses: "Continuity of Care," "Technical Care," and "Other" composites.

RESULTS

Continuity of Care composite rates showed no significant changes postreform in hospitals of different teaching intensity in either VA or Medicare. In the VA, there were no significant changes postreform for the technical care composite. In Medicare, the odds of a Technical Care PSI event in more versus less teaching-intensive hospitals in postreform year 1 were 1.12 (95% CI; 1.01-1.25); there were no significant relative changes in postreform year 2. Other composite rates increased in VA in postreform year 2 in more versus less teaching-intensive hospitals (odds ratio, 1.63; 95% CI; 1.10-2.41), but not in Medicare in either postreform year.

CONCLUSIONS

Duty hour reform had no systematic impact on PSI rates. In the few cases where there were statistically significant increases in the relative odds of developing a PSI, the magnitude of the absolute increases were too small to be clinically meaningful.

Validity of selected AHRQ patient safety indicators based on VA National Surgical Quality Improvement Program data

OBJECTIVES

To examine the criterion validity of the Agency for Health Care Research and Quality (AHRQ) Patient Safety Indicators (PSIs) using clinical data from the Veterans Health Administration (VA) National Surgical Quality Improvement Program (NSQIP).

DATA SOURCES

Fifty five thousand seven hundred and fifty two matched hospitalizations from 2001 VA inpatient surgical discharge data and NSQIP chart-abstracted data.

STUDY DESIGN

We examined the sensitivities, specificities, positive predictive values (PPVs), and positive likelihood ratios of five surgical PSIs that corresponded to NSQIP adverse events. We created and tested alternative definitions of each PSI.

DATA COLLECTION

FY01 inpatient discharge data were merged with 2001 NSQIP data abstracted from medical records for major noncardiac surgeries.

PRINCIPAL FINDINGS

Sensitivities were 19-56 percent for original PSI definitions; and 37-63 percent using alternative PSI definitions. PPVs were 22-74 percent and did not improve with modifications. Positive likelihood ratios were 65-524 using original definitions, and 64-744 using alternative definitions. "Postoperative respiratory failure" and "postoperative wound dehiscence" exhibited significant increases in sensitivity after modifications.

CONCLUSIONS

PSI sensitivities and PPVs were moderate. For three of the five PSIs, AHRQ has incorporated our alternative, higher sensitivity definitions into current PSI algorithms. Further validation should be considered before most of the PSIs evaluated herein are used to publicly compare or reward hospital performance.