Accelerating Surgical Site Infection Abstraction With a Semi-automated Machine-learning Approach.

OBJECTIVE: To demonstrate that a semi-automated approach to health data abstraction provides significant efficiencies and high accuracy. BACKGROUND: Surgical outcome abstraction remains laborious and a barrier to the sustainment of quality improvement registries like ACS-NSQIP. A supervised machine learning algorithm developed for detecting SSi using structured and unstructured electronic health record data was tested to perform semi-automated SSI abstraction. METHODS: A Lasso-penalized logistic regression model with 2011-3 data was trained (baseline performance measured with 10-fold cross-validation). A cutoff probability score from the training data was established, dividing the subsequent evaluation dataset into "negative" and "possible" SSI groups, with manual data abstraction only performed on the "possible" group. We evaluated performance on data from 2014, 2015, and both years. RESULTS: Overall, 6188 patients were in the 2011-3 training dataset and 5132 patients in the 2014-5 evaluation dataset. With use of the semi-automated approach, applying the cut-off score decreased the amount of manual abstraction by >90%, resulting in < 1% false negatives in the "negative" group and a sensitivity of 82%. A blinded review of 10% of the "possible" group, considering only the features selected by the algorithm, resulted in high agreement with the gold standard based on full chart abstraction, pointing towards additional efficiency in the abstraction process by making it possible for abstractors to review limited, salient portions of the chart. CONCLUSION: Semi-automated machine learning-aided SSI abstraction greatly accelerates the abstraction process and achieves very good performance. This could be translated to other post-operative outcomes and reduce cost barriers for wider ACS-NSQIP adoption.

Colorectal Lymphoma: A Contemporary Case Series.

BACKGROUND: Colon and rectal lymphomas are rare and can occur in the context of posttransplant lymphoproliferative disorder. Evidence-based management guidelines are lacking. OBJECTIVE: The purpose of this study was to characterize the presentation, diagnosis, and management of colorectal lymphoma and to identify differences within the transplant population. DESIGN: This was a retrospective review of patients evaluated for colorectal lymphoma between 2000 and 2017. Patients were identified through clinical note queries. SETTINGS: Four hospitals within a single health system were included. PATIENTS: Fifty-two patients (64% men; mean age = 64 y; range, 26-91 y) were identified. No patient had <3 months of follow-up. Eight patients (15%) had posttransplant lymphoproliferative disorder. MAIN OUTCOME MEASURES: Overall survival, recurrence, and complications in treatment pathway were measured. RESULTS: Most common presentations were rectal bleeding (27%), abdominal pain (23%), and diarrhea (23%). The most common location was the cecum (62%). Most frequent histologies were diffuse large B-cell lymphoma (48%) and mantle cell lymphoma (25%). Posttransplant lymphoproliferative disorder occurred in the cecum (n = 4) and rectum (n = 4). Twenty patients (38%) were managed with chemotherapy; 25 patients (48%) underwent primary resection. Mass lesions had a higher risk of urgent surgical resection (35% vs 8%; p = 0.017). Three patients (15%) treated with chemotherapy presented with perforation requiring emergency surgery. Overall survival was 77 months (range, 25-180 mo). Patients with cecal involvement had longer overall survival (96 vs 26 mo; p = 0.038); immunosuppressed patients had shorter survival (16 vs 96 mo; p = 0.006). Survival in patients treated with surgical management versus chemotherapy was similar (67 vs 105 mo; p = 0.62). LIMITATIONS: This was a retrospective chart review, with data limited by the contents of the medical chart. This was a small sample size. CONCLUSIONS: Colorectal lymphoma is rare, with variable treatment approaches. Patients with noncecal involvement and chronic immunosuppression had worse overall survival. Patients with mass lesions, particularly cecal masses, are at higher risk to require urgent intervention, and primary resection should be considered. See Video Abstract at http://links.lww.com/DCR/A929.

Decreased Use of Sphincter-Preserving Procedures Among African Americans with Rectal Cancer.

BACKGROUND: Improved multimodality rectal cancer treatment has increased the use of sphincter-preserving surgery. This study sought to determine whether African American (AA) patients with rectal cancer receive sphincter-preserving surgery at the same rate as non-AA patients. METHODS: The study used the Nationwide Inpatient Sample for years 1998-2012 to compare AA and non-AA patients with rectal cancer undergoing low anterior resection or abdominoperineal resection. The logistic regression model was used to adjust for age, gender, admission type, Elixhauser comorbidity index, and hospital factors such as size, location (urban vs.rural), teaching status, and procedure volume. RESULTS: The search identified 22,697 patients, 1600 of whom were identified as AA. After adjustment for age and gender, the analysis showed that AA patients were less likely to undergo sphincter-preserving surgery than non-AA patients [odds ratio (OR) 0.70; 95% confidence interval (CI) 0.63-0.78; p < 0.0001). After further adjustment for the Elixhauser comorbidity index, admission type, hospital-specific factors, and insurance status, the analysis showed that AA patients still were less likely to undergo sphincter-preserving surgery (OR 0.78; 95% CI 0.70-0.87; p < 0.0001). Although the proportion of non-AA patients undergoing sphincter-preserving surgery increased during the study period (p = 0.0003), this trend was not significant for the AA patients (p = 0.13). CONCLUSION: In this data analysis, the AA patients with rectal cancer had lower rates of sphincter-preserving surgery than the non-AA patients, even after adjustment for patient- and hospital-specific factors. Further work is required to elucidate why. Eliminating racial disparities in rectal cancer treatment should continue to be a priority for the surgical community.

Survival Rates for Patients with Resected Gastric Adenocarcinoma Finally have Increased in the United States.

BACKGROUND: In the United States, the overall survival rates for gastric adenocarcinoma have remained low, with surgical resection as the only therapy for many patients. Given the advances in multimodality treatment and the development of guidelines recommending adequate lymph node evaluation, the authors determined whether overall survival rates for patients with gastric adenocarcinoma have increased in the United States. METHODS: The study used the Surveillance Epidemiology and End Results (SEER) database to examine overall survival for patients with the diagnosis of gastric adenocarcinoma between 1988 and 2013. The study cohort was divided into five periods: 1988-1992, 1993-1997, 1998-2002, 2003-2007, and 2008-2013. Kaplan-Meier methods and Cox proportional hazards modeling were used to determine the effect that year of diagnosis had on overall survival. RESULTS: The diagnosis was determined for 13,470 patients between 1988 and 2013. The use of radiation therapy and the proportion of patients who had at least 15 lymph nodes evaluated significantly increased during the study period. Unadjusted Kaplan-Meier estimates demonstrated significantly better survival rates for the patients with a diagnosis of gastric cancer in the later periods (2003-2007 and 2008-2013) than for those in the three earlier periods. In our Cox proportional hazards model, recent period was associated with a significantly lower hazard of 5-year mortality. CONCLUSION: This analysis demonstrated for the first time that gastric cancer survival rates have significantly improved in the United States during the past 2 decades. This observation likely reflects improved adherence to cancer treatment guidelines, including adequate lymph node evaluation and delivery of adjuvant treatment more consistently.

Strategies for handling missing clinical data for automated surgical site infection detection from the electronic health record.

Proper handling of missing data is important for many secondary uses of electronic health record (EHR) data. Data imputation methods can be used to handle missing data, but their use for analyzing EHR data is limited and specific efficacy for postoperative complication detection is unclear. Several data imputation methods were used to develop data models for automated detection of three types (i.e., superficial, deep, and organ space) of surgical site infection (SSI) and overall SSI using American College of Surgeons National Surgical Quality Improvement Project (NSQIP) Registry 30-day SSI occurrence data as a reference standard. Overall, models with missing data imputation almost always outperformed reference models without imputation that included only cases with complete data for detection of SSI overall achieving very good average area under the curve values. Missing data imputation appears to be an effective means for improving postoperative SSI detection using EHR clinical data.

Detecting clinically relevant new information in clinical notes across specialties and settings.

BACKGROUND: Automated methods for identifying clinically relevant new versus redundant information in electronic health record (EHR) clinical notes is useful for clinicians and researchers involved in patient care and clinical research, respectively. We evaluated methods to automatically identify clinically relevant new information in clinical notes, and compared the quantity of redundant information across specialties and clinical settings. METHODS: Statistical language models augmented with semantic similarity measures were evaluated as a means to detect and quantify clinically relevant new and redundant information over longitudinal clinical notes for a given patient. A corpus of 591 progress notes over 40 inpatient admissions was annotated for new information longitudinally by physicians to generate a reference standard. Note redundancy between various specialties was evaluated on 71,021 outpatient notes and 64,695 inpatient notes from 500 solid organ transplant patients (April 2015 through August 2015). RESULTS: Our best method achieved at best performance of 0.87 recall, 0.62 precision, and 0.72 F-measure. Addition of semantic similarity metrics compared to baseline improved recall but otherwise resulted in similar performance. While outpatient and inpatient notes had relatively similar levels of high redundancy (61% and 68%, respectively), redundancy differed by author specialty with mean redundancy of 75%, 66%, 57%, and 55% observed in pediatric, internal medicine, psychiatry and surgical notes, respectively. CONCLUSIONS: Automated techniques with statistical language models for detecting redundant versus clinically relevant new information in clinical notes do not improve with the addition of semantic similarity measures. While levels of redundancy seem relatively similar in the inpatient and ambulatory settings in the Fairview Health Services, clinical note redundancy appears to vary significantly with different medical specialties.

Implementation and Analysis of an Institution-Wide EHR-Integrated Handoff Note.

EHR-Integrated Handoff Notes are becoming increasingly prevalent, especially among inpatient clinical service teams composed of physicians in training (resident physicians). We describe the implementation of such a tool at our institution, and then describe the usage of the tool, as well as changes made to its content, structure, and format, years after original implementation. We also describe frequency and temporal distribution of updates made to free text elements of the Handoff Note. At the initial implementation there were three versions available; medical/surgical, pediatric, and ICU. Years after implementation, 57% of clinical service teams continue to utilize the note, including all medical, pediatric, behavioral health, obstetrics/gynecology, and neurology services, as well as most surgical services, that serve as "primary" teams. Several interesting changes were noted to the content and structure of the Handoff Note, namely that more complicated versions were abandoned in favor of simpler versions. The Patient Summary and To Do free text boxes are updated an average of 1.0 and 1.6 times per day. Around 60% of updates to both free text boxes occur between 12 pm - 5:59 pm, likely in preparation for the daytime-nightime team handoff.

Is there a role for specialized geriatric centers in treating geriatric cancer patients?

As the population with colorectal cancer ages, the tailored approach required to manage older patients becomes all the more important for all providers and institutions treating colorectal cancer to adopt and improve the outcomes and well-being of this important and increasingly prevalent population. Joint guidelines from the American College of Surgeons and American Geriatric Association should be followed. Older cancer patients undergoing colorectal cancer surgery should be referred to centers with expertise in minimally invasive surgery. Likewise, older rectal cancer patients should be referred to centers with expertise in treating rectal cancer.

A NSQIP analysis of trends in surgical outcomes for rectal cancer: What can we improve upon?

BACKGROUND: There is significant variation in rectal cancer outcomes in the USA, and reported outcomes have been inferior to those in other countries. In recognition of this fact, the American College of Surgeons (ACS) recently launched the Commission on Cancer (CoC) National Accreditation Program for Rectal Cancer (NAPRC) in an effort to further optimize rectal cancer care. Large surgical databases will play an important role in tracking surgical and oncologic outcomes. Our study sought to explore the trends in surgical outcomes over the decade prior to the NAPRC using a large national database. METHODS: The ACS National Surgical Quality Improvement Program (NSQIP) database from 2005 to 2017 was used to select colorectal cancer cases which were divided into abdominal-colonic (AC) and pelvic-rectal (PR) cohorts based upon the operation performed. Outcomes of interest were occurrence of any major surgical complication, mortality within 30 days of procedure, and postoperative length of stay (LOS). Chi-square and two sample t-tests were used to evaluate association between various risk factors and outcomes. Modified Poisson regression was used to compare and estimate the unadjusted and adjusted effect of procedure type on the outcomes. STATA 15.1 was used for analysis and statistical significance was set at 0.05. RESULTS: A total of 34,159 patients were analyzed. AC cases constituted 50.7% of the overall cohort. The two groups were relatively similar in demographic distribution, but the PR patients had higher rates of hypoalbuminemia and were sicker (ASA class 3 or greater). Rates of non-sphincter preserving operations ranged from 30 to 34%. Higher complication rates in the PR cohort were mainly infectious and surgical site complications, while rates of deep vein thrombosis and pulmonary embolism were similar between the two cohorts. On bivariate analysis, rates of mortality were similar between the two groups (AC: 1.02% vs PR: 0.91%, p = 0.395), while PR patients were found to be 1.36 times (95% CI: 1.32-1.41) more likely to have major complications and 1.40 times (95% CI: 1.35-1.44) more likely to have an extended LOS as compared to the AC patients. After multivariable analysis, PR patients continued to have a higher likelihood of major complications (IRR: 1.31, 95% CI 1.25-1.36) and extended LOS (IRR: 1.38, 95% CI: 1.33-1.43). 10-year trends showed a significant reduction in the percentage of patients with prolonged lengths of hospitalization as well as a reduction of nearly 20% in the mean LOS, but without improvement in morbidity or mortality. CONCLUSIONS: Patients undergoing PR operations were more likely to have had major complications than were patients who underwent AC procedures; unfortunately no improvement in the rate of these complications or in mortality occurred. Perhaps the significant reduction in LOS is due in part to an increased prevalence of minimally invasive surgery and/or enhanced recovery protocols. Data were found to be lacking within NSQIP for several important variables including key oncologic data, stratification by surgical volume, and patient geographic location. We anticipate that the NAPRC should help improve PR surgical and oncologic outcomes including decreasing morbidity and mortality rates during the next decade.

Resident-Physician Preferences for Electronic Handoff Note Content: Implications for Implementation of a System-Wide Electronic Health Record-Integrated Handoff Tool.

Increasing attention is paid to the Handoff Process and EHR- integrated tools to enhance the Handoff Process and aid in creating Handoff Notes are becoming more prevalent. In this study, we attempted to determine the ideal content of the Handoff Note based on the preferences of the resident physicians for whom the tool was being constructed. This commenced with an initial semi-structured interview and culminated in a large-scale survey. Overall, 315 resident physicians completed the survey. Plan of Care, Illness Severity, and Patient Summary were the most important content elements to resident physicians. The importance and trustworthiness of other content elements, as well as their preferred display and level of granularity within the Handoff Note, varied considerably. Subjective assessment by a colleague of a patient's hospital course and plan of care, rather than any single piece of objective data, are preferred as ideal content for Handoff Note composition by resident physicians.

Supplementing Resident Research Funding Through a Partnership With Local Industry.

OBJECTIVE: To develop a model for the supplementation of resident research funding through a resident-hosted clinical immersion with local industry. DESIGN: Designated research residents hosted multiple groups of engineers and business professionals from local industry in general surgery-focused clinical immersion weeks. The participants in these week-long programs are educated about general surgery and brought to the operating room to observe a variety of surgeries. SETTING: This study was performed at the University of Minnesota, in Minneapolis, Minnesota, at a tertiary medical center. PARTICIPANTS: Ten designated research residents hosted general surgery immersion programs. Fifty-seven engineers and business professionals from 5 different local biomedical firms have participated in this program. RESULTS: General surgery research residents (in collaboration with the University of Minnesota's Institute for Engineering in Medicine) have hosted 9 clinical immersion programs since starting the collaborative in 2015. Immersion participant response to the experiences was very positive. Two full-time resident research positions can be funded annually through participation in this program. CONCLUSIONS: With decreasing funding available for surgical research, particularly resident research, innovative ways to fund resident research are needed. The general surgery clinical immersion program at the University of Minnesota has proven its value as a supplement for resident research funding and may be a sustainable model for the future.

Characterizing Functional Health Status of Surgical Patients in Clinical Notes.

Functional health status is an important factor not only for determining overall health, but also for measuring risks of adverse events. Our hypothesis is that important functional status data is contained in clinical notes. We found that several categories of phrases related to functional status including diagnoses, activity and care assessments, physical exam, functional scores, assistive equipment, symptoms, and surgical history were important factors. Use of functional health status level terms from our chart review compared to National Surgical Quality Improvement Program determination had varying sensitivities for correct functional status category identification, with 96% for independent patients, 60% for partially dependent patients, and 44% for totally dependent patients. Inter-rater agreement assessing term relevance to functional health status was high at 91% (Kappa=0.74). Functional status-related terms in clinical notes show potential for use in future methodologies for automated detection of functional health status for quality improvement registries and other clinical assessments.

Assessing Provider-Generated Free-Text Quality in EHR-Integrated Handoff Notes.

Handoff notes are increasingly integrated within electronic health record (EHR) systems and often contain data automatically generated from the EHR and free-text narratives. We examined the quality of data entered by providers in the free-text portion of our institutional EHR handoff tool. Overall, 65% of handoff notes contained at least one error (average 1.7 errors per note). Most errors were omissions in information around patient plan/management or assessment/diagnosis rather than entry of false data. Factors associated with increased error rate were increasing hospital day number; weekend note; medical (vs. surgical) service team; and authorship by a medical student, first or fourth year resident physician, or attending physician. Our findings suggest that errors are common in handoff notes, and while these errors are not completely false data, they may provide individuals caring for patients an inaccurate understanding of patient status.

African Americans and Short-Term Outcomes after Surgery for Crohn's Disease: An ACS-NSQIP Analysis.

BACKGROUND: Previous reports on racial disparities in the treatment of Crohn's disease [CD] in African American [AA] patients have shown differences in both medical and surgical treatments in this population. No study thus far has examined the effect of AA race on outcomes after surgery for CD. METHODS: Utilizing the National Surgical Quality Improvement Program [NSQIP] Participant User File [PUF] for the years 2005-2013, we examined the effect of AA race on postoperative complications in patients with CD undergoing intestinal surgery. RESULTS: AA patients had a significantly higher rate of complications overall compared to non-AA patients [23.5% vs 18.9%, p = 0.002]. Postoperative sepsis [10.9% vs 6.6%, p < 0.001] and surgical site infection [17.6% vs 14.8%, p = 0.037] were most significant. After adjustment for age, sex, preoperative disease severity and lifestyle factors [smoking], race remained a statistically significant factor in postoperative complication rate. Only after additional adjustment was made for comorbidities and American Society of Anesthesiologists class did race lose significance within our model. CONCLUSION: African Americans experience a greater amount of postoperative complications following surgery for Crohn's disease. Preoperative disease management, addressing smoking status and control of comorbid disease are important factors in addressing the racial disparities in the surgical treatment of Crohn's disease.

Characterizing Surgical Site Infection Signals in Clinical Notes.

Surgical site infections (SSIs) are the most common and costly of hospital acquired infections. An important step in reducing SSIs is accurate SSI detection, which enables measurement quality improvement, but currently remains expensive through manual chart review. Building off of previous work for automated and semi-automated SSI detection using expert-derived "strong features" from clinical notes, we hypothesized that additional SSI phrases may be contained in clinical notes. We systematically characterized phrases and expressions associated with SSIs. While 83% of expert-derived original terms overlapped with new terms and modifiers, an additional 362 modifiers associated with both positive and negative SSI signals were identified and 62 new base observations and actions were identified. Clinical note queries with the most common base terms revealed another 49 modifiers. Clinical notes contain a wide variety of expressions describing infections occurring among surgical specialties which may provide value in improving the performance of SSI detection algorithms.

Accelerating Chart Review Using Automated Methods on Electronic Health Record Data for Postoperative Complications.

Manual Chart Review (MCR) is an important but labor-intensive task for clinical research and quality improvement. In this study, aiming to accelerate the process of extracting postoperative outcomes from medical charts, we developed an automated postoperative complications detection application by using structured electronic health record (EHR) data. We applied several machine learning methods to the detection of commonly occurring complications, including three subtypes of surgical site infection, pneumonia, urinary tract infection, sepsis, and septic shock. Particularly, we applied one single-task and five multi-task learning methods and compared their detection performance. The models demonstrated high detection performance, which ensures the feasibility of accelerating MCR. Specifically, one of the multi-task learning methods, propensity weighted observations (PWO) demonstrated the highest detection performance, with single-task learning being a close second.

Automated Detection of Postoperative Surgical Site Infections Using Supervised Methods with Electronic Health Record Data.

The National Surgical Quality Improvement Project (NSQIP) is widely recognized as "the best in the nation" surgical quality improvement resource in the United States. In particular, it rigorously defines postoperative morbidity outcomes, including surgical adverse events occurring within 30 days of surgery. Due to its manual yet expensive construction process, the NSQIP registry is of exceptionally high quality, but its high cost remains a significant bottleneck to NSQIP's wider dissemination. In this work, we propose an automated surgical adverse events detection tool, aimed at accelerating the process of extracting postoperative outcomes from medical charts. As a prototype system, we combined local EHR data with the NSQIP gold standard outcomes and developed machine learned models to retrospectively detect Surgical Site Infections (SSI), a particular family of adverse events that NSQIP extracts. The built models have high specificity (from 0.788 to 0.988) as well as very high negative predictive values (>0.98), reliably eliminating the vast majority of patients without SSI, thereby significantly reducing the NSQIP extractors' burden.

Characterizing Patient-Generated Clinical Data and Associated Implications for Electronic Health Records.

Patient-facing technologies are increasingly utilized for direct patient data entry for potential incorporation into the electronic health record. We analyzed patient-entered data during implementation of a patient-facing data entry technology using an online patient portal and clinic-based tablet computers at a University-based tertiary medical center clinic, including entries for past medical history, past surgical history, and social history. Entries were assessed for granularity, clinical accuracy, and the addition of novel information into the record. We found that over half of patient-generated diagnoses were duplicates of lesser or equal granularity compared to previous provider-entered diagnoses. Approximately one fifth of patient-generated diagnoses were found to meet the criteria for new, meaningful additions to the medical record. Our findings demonstrate that while patient-generated data provides important additional information, it may also present challenges including generating inaccurate or less granular information.