Maximizing Interpretability and Cost-Effectiveness of Surgical Site Infection (SSI) Predictive Models Using Feature-Specific Regularized Logistic Regression on Preoperative Temporal Data.

This study describes a novel approach to solve the surgical site infection (SSI) classification problem. Feature engineering has traditionally been one of the most important steps in solving complex classification problems, especially in cases with temporal data. The described novel approach is based on abstraction of temporal data recorded in three temporal windows. Maximum likelihood L1-norm (lasso) regularization was used in penalized logistic regression to predict the onset of surgical site infection occurrence based on available patient blood testing results up to the day of surgery. Prior knowledge of predictors (blood tests) was integrated in the modelling by introduction of penalty factors depending on blood test prices and an early stopping parameter limiting the maximum number of selected features used in predictive modelling. Finally, solutions resulting in higher interpretability and cost-effectiveness were demonstrated. Using repeated holdout cross-validation, the baseline C-reactive protein (CRP) classifier achieved a mean AUC of 0.801, whereas our best full lasso model achieved a mean AUC of 0.956. Best model testing results were achieved for full lasso model with maximum number of features limited at 20 features with an AUC of 0.967. Presented models showed the potential to not only support domain experts in their decision making but could also prove invaluable for improvement in prediction of SSI occurrence, which may even help setting new guidelines in the field of preoperative SSI prevention and surveillance.

Using anchors from free text in electronic health records to diagnose postoperative delirium.

OBJECTIVES: Postoperative delirium is a common complication after major surgery among the elderly. Despite its potentially serious consequences, the complication often goes undetected and undiagnosed. In order to provide diagnosis support one could potentially exploit the information hidden in free text documents from electronic health records using data-driven clinical decision support tools. However, these tools depend on labeled training data and can be both time consuming and expensive to create. METHODS: The recent learning with anchors framework resolves this problem by transforming key observations (anchors) into labels. This is a promising framework, but it is heavily reliant on clinicians knowledge for specifying good anchor choices in order to perform well. In this paper we propose a novel method for specifying anchors from free text documents, following an exploratory data analysis approach based on clustering and data visualization techniques. We investigate the use of the new framework as a way to detect postoperative delirium. RESULTS: By applying the proposed method to medical data gathered from a Norwegian university hospital, we increase the area under the precision-recall curve from 0.51 to 0.96 compared to baselines. CONCLUSIONS: The proposed approach can be used as a framework for clinical decision support for postoperative delirium.

Predicting colorectal surgical complications using heterogeneous clinical data and kernel methods.

OBJECTIVE: In this work, we have developed a learning system capable of exploiting information conveyed by longitudinal Electronic Health Records (EHRs) for the prediction of a common postoperative complication, Anastomosis Leakage (AL), in a data-driven way and by fusing temporal population data from different and heterogeneous sources in the EHRs. MATERIAL AND METHODS: We used linear and non-linear kernel methods individually for each data source, and leveraging the powerful multiple kernels for their effective combination. To validate the system, we used data from the EHR of the gastrointestinal department at a university hospital. RESULTS: We first investigated the early prediction performance from each data source separately, by computing Area Under the Curve values for processed free text (0.83), blood tests (0.74), and vital signs (0.65), respectively. When exploiting the heterogeneous data sources combined using the composite kernel framework, the prediction capabilities increased considerably (0.92). Finally, posterior probabilities were evaluated for risk assessment of patients as an aid for clinicians to raise alertness at an early stage, in order to act promptly for avoiding AL complications. DISCUSSION: Machine-learning statistical model from EHR data can be useful to predict surgical complications. The combination of EHR extracted free text, blood samples values, and patient vital signs, improves the model performance. These results can be used as a framework for preoperative clinical decision support.

Support Vector Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-Words in Electronic Health Records.

The free text in electronic health records (EHRs) conveys a huge amount of clinical information about health state and patient history. Despite a rapidly growing literature on the use of machine learning techniques for extracting this information, little effort has been invested toward feature selection and the features' corresponding medical interpretation. In this study, we focus on the task of early detection of anastomosis leakage (AL), a severe complication after elective surgery for colorectal cancer (CRC) surgery, using free text extracted from EHRs. We use a bag-of-words model to investigate the potential for feature selection strategies. The purpose is earlier detection of AL and prediction of AL with data generated in the EHR before the actual complication occur. Due to the high dimensionality of the data, we derive feature selection strategies using the robust support vector machine linear maximum margin classifier, by investigating: 1) a simple statistical criterion (leave-one-out-based test); 2) an intensive-computation statistical criterion (Bootstrap resampling); and 3) an advanced statistical criterion (kernel entropy). Results reveal a discriminatory power for early detection of complications after CRC (sensitivity 100%; specificity 72%). These results can be used to develop prediction models, based on EHR data, that can support surgeons and patients in the preoperative decision making phase.

Data-Driven Personalized Feedback to Patients with Type 1 Diabetes: A Randomized Trial.

BACKGROUND: A mobile phone-based application can be useful for patients with type 1 diabetes in managing their disease. This results in large datasets accumulated on the patient's devices, which can be used for individualized feedback. The effect of such feedback is investigated in this article. MATERIALS AND METHODS: We developed an application that included a data-driven feedback module known as Diastat for patients on self-measured blood glucose regimens. Using a stepped-wedge design, both groups initially received an application without Diastat. Group 1 activated Diastat after 4 weeks, whereas Group 2 activated Diastat 12 weeks after startup (T1). End points were glycated hemoglobin (HbA1c) level and number of out-of-range (OOR) measurements (i.e., outside the range 72-270 mg/dL). RESULTS: Thirty patients were recruited to the study, and 15 were assigned to each group after the initial meeting. There were no significant differences between groups at T1 in HbA1c or OOR events. Overall, all patients had a decrease of 0.6 percentage points in mean HbA1c (P < 0.001) and 14.5 in median OOR events over 2 weeks (P < 0.001). CONCLUSIONS: The study does not provide evidence that data-driven feedback improves glycemic control. The decrease in HbA1c was sizeable and significant, even though the study was not powered to detect this. The overall improvement in glycemic control suggests that, in general, mobile phone-based interventions can be useful in diabetes self-management.

Data-driven Temporal Prediction of Surgical Site Infection.

Analysis of data from Electronic Health Records (EHR) presents unique challenges, in particular regarding nonuniform temporal resolution of longitudinal variables. A considerable amount of patient information is available in the EHR - including blood tests that are performed routinely during inpatient follow-up. These data are useful for the design of advanced machine learning-based methods and prediction models. Using a matched cohort of patients undergoing gastrointestinal surgery (101 cases and 904 controls), we built a prediction model for post-operative surgical site infections (SSIs) using Gaussian process (GP) regression, time warping and imputation methods to manage the sparsity of the data source, and support vector machines for classification. For most blood tests, wider confidence intervals after imputation were obtained in patients with SSI. Predictive performance with individual blood tests was maintained or improved by joint model prediction, and non-linear classifiers performed consistently better than linear models.

Data-driven approach for assessing utility of medical tests using electronic medical records.

OBJECTIVE: To precisely define the utility of tests in a clinical pathway through data-driven analysis of the electronic medical record (EMR). MATERIALS AND METHODS: The information content was defined in terms of the entropy of the expected value of the test related to a given outcome. A kernel density classifier was used to estimate the necessary distributions. To validate the method, we used data from the EMR of the gastrointestinal department at a university hospital. Blood tests from patients undergoing surgery for gastrointestinal surgery were analyzed with respect to second surgery within 30 days of the index surgery. RESULTS: The information content is clearly reflected in the patient pathway for certain combinations of tests and outcomes. C-reactive protein tests coupled to anastomosis leakage, a severe complication show a clear pattern of information gain through the patient trajectory, where the greatest gain from the test is 3-4 days post index surgery. DISCUSSION: We have defined the information content in a data-driven and information theoretic way such that the utility of a test can be precisely defined. The results reflect clinical knowledge. In the case we used the tests carry little negative impact. The general approach can be expanded to cases that carry a substantial negative impact, such as in certain radiological techniques.

Implementation of an electronic surgical referral service. Collaboration, consensus and cost of the surgeon - general practitioner Delphi approach.

BACKGROUND: Poor coordination between levels of care plays a central role in determining the quality and cost of health care. To improve patient coordination, systematic structures, guidelines, and processes for creating, transferring, and recognizing information are needed to facilitate referral routines. METHODS: Prospective observational survey of implementation of electronic medical record (EMR)-supported guidelines for surgical treatment. RESULTS: One university clinic, two local hospitals, 31 municipalities, and three EMR vendors participated in the implementation project. Surgical referral guidelines were developed using the Delphi method; 22 surgeons and seven general practitioners (GPs) needed 109 hours to reach consensus. Based on consensus guidelines, an electronic referral service supported by a clinical decision support system, fully integrated into the GPs' EMR, was developed. Fifty-five information technology personnel and 563 hours were needed (total cost 67,000 £) to implement a guideline supported system in the EMR for 139 GPs. Economical analyses from a hospital and societal perspective, showed that 504 (range 401-670) and 37 (range 29-49) referred patients, respectively, were needed to provide a cost-effective service. CONCLUSION: A considerable amount of resources were needed to reach consensus on the surgical referral guidelines. A structured approach by the Delphi method and close collaboration between IT personnel, surgeons and primary care physicians were needed to reach consensus.

Power law approximations of movement network data for modeling infectious disease spread.

Globalization and increased mobility of individuals enable person-to-person transmitted infectious diseases to spread faster to distant places around the world, making good models for the spread increasingly important. We study the spatiotemporal pattern of spread in the remotely located and sparsely populated region of North Norway in various models with fixed, seasonal, and random effects. The models are applied to influenza A counts using data from positive microbiology laboratory tests as proxy for the underlying disease incidence. Human travel patterns with local air, road, and sea traffic data are incorporated as well as power law approximations thereof, both with quasi-Poisson regression and based on the adjacency structure of the relevant municipalities. We investigate model extensions using information about the proportion of positive laboratory tests, data on immigration from outside North Norway and by connecting population to the movement network. Furthermore, we perform two separate analyses for nonadults and adults as children are an important driver for influenza A. Comparisons of one-step-ahead predictions generally yield better or comparable results using power law approximations.

Performance of a dermoscopy-based computer vision system for the diagnosis of pigmented skin lesions compared with visual evaluation by experienced dermatologists.

BACKGROUND: It is often difficult to differentiate early melanomas from benign melanocytic nevi even by expert dermatologists, and the task is even more challenging for primary care physicians untrained in dermatology and dermoscopy. A computer system can provide an objective and quantitative evaluation of skin lesions, reducing subjectivity in the diagnosis. OBJECTIVE: Our objective is to make a low-cost computer aided diagnostic tool applicable in primary care based on a consumer grade camera with attached dermatoscope, and compare its performance to that of experienced dermatologists. METHODS AND MATERIALS: We propose several new image-derived features computed from automatically segmented dermoscopic pictures. These are related to the asymmetry, color, border, geometry, and texture of skin lesions. The diagnostic accuracy of the system is compared with that of three dermatologists. RESULTS: With a data set of 206 skin lesions, 169 benign and 37 melanomas, the classifier was able to provide competitive sensitivity (86%) and specificity (52%) scores compared with the sensitivity (85%) and specificity (48%) of the most accurate dermatologist using only dermoscopic images. CONCLUSION: We show that simple statistical classifiers can be trained to provide a recommendation on whether a pigmented skin lesion requires biopsy to exclude skin cancer with a performance that is comparable to and exceeds that of experienced dermatologists.

A semiautomatic tool for prostate segmentation in radiotherapy treatment planning.

BACKGROUND: Delineation of the target volume is a time-consuming task in radiotherapy treatment planning, yet essential for a successful treatment of cancers such as prostate cancer. To facilitate the delineation procedure, the paper proposes an intuitive approach for 3D modeling of the prostate by slice-wise best fitting ellipses. METHODS: The proposed estimate is initialized by the definition of a few control points in a new patient. The method is not restricted to particular image modalities but assumes a smooth shape with elliptic cross sections of the object. A training data set of 23 patients was used to calculate a prior shape model. The mean shape model was evaluated based on the manual contour of 10 test patients. The patient records of training and test data are based on axial T1-weighted 3D fast-field echo (FFE) sequences. The manual contours were considered as the reference model. Volume overlap (Vo), accuracy (Ac) (both ratio, range 0-1, optimal value 1) and Hausdorff distance (HD) (mm, optimal value 0) were calculated as evaluation parameters. RESULTS: The median and median absolute deviation (MAD) between manual delineation and deformed mean best fitting ellipses (MBFE) was Vo (0.9 ± 0.02), Ac (0.81 ± 0.03) and HD (4.05 ± 1.3)mm and between manual delineation and best fitting ellipses (BFE) was Vo (0.96 ± 0.01), Ac (0.92 ± 0.01) and HD (1.6 ± 0.27)mm. Additional results show a moderate improvement of the MBFE results after Monte Carlo Markov Chain (MCMC) method. CONCLUSIONS: The results emphasize the potential of the proposed method of modeling the prostate by best fitting ellipses. It shows the robustness and reproducibility of the model. A small sample test on 8 patients suggest possible time saving using the model.

Mobile patient applications within diabetes - from few and easy to advanced functionalities.

Patient diaries as apps on mobile phones are becoming increasingly common, and can be a good support tool for patients who need to organize information relevant for their disease. Self-management is important to achieving diabetes treatment goals and can be a tool for lifestyle changes for patients with Type 2 diabetes. The autoimmune disease Type 1 diabetes requires a more intensive management than Type 2 - thus more advanced functionalities is desirable for users. Both simple and easy-to-use and more advanced diaries have their respective benefits, depending on the target user group and intervention. In this poster we summarize main findings and experience from more than a decade of research and development in the diabetes area. Several versions of the mobile health research platform-the Few Touch Application (FTA) are presented to illustrate the different approaches and results.

Model-driven diabetes care: study protocol for a randomized controlled trial.

BACKGROUND: People with type 1 diabetes who use electronic self-help tools register a large amount of information about their disease on their participating devices; however, this information is rarely utilized beyond the immediate investigation. We have developed a diabetes diary for mobile phones and a statistics-based feedback module, which we have named Diastat, to give data-driven feedback to the patient based on their own data. METHOD: In this study, up to 40 participants will be given a smartphone on which is loaded a diabetes self-help application (app), the Few Touch Application (FTA). Participants will be randomized into two groups to be given access to Diastat 4 or 12 weeks, respectively after receiving the smartphone, and will use the FTA with Diastat for 8 weeks after this point. The primary endpoint is the frequency of high and low blood-glucose measurements. DISCUSSION: The study will investigate the effect of data-driven feedback to patients. Our hypothesis is that this will improve glycemic control and reduce variability. The endpoints are robust indicators that can be assembled with minimal effort by the patient beyond normal routine. TRIAL REGISTRATION: Clinicaltrials.gov: NCT01774149.

Cost-effectiveness and quality of life in surgeon versus general practitioner-organised colon cancer surveillance: a randomised controlled trial.

OBJECTIVE: To assess whether colon cancer follow-up can be organised by general practitioners (GPs) without a decline in the patient's quality of life (QoL) and increase in cost or time to cancer diagnoses, compared to hospital follow-up. DESIGN: Randomised controlled trial. SETTING: Northern Norway Health Authority Trust, 4 trusts, 11 hospitals and 88 local communities. PARTICIPANTS: Patients surgically treated for colon cancer, hospital surgeons and community GPs. INTERVENTION: 24-month follow-up according to national guidelines at the community GP office. To ensure a high follow-up guideline adherence, a decision support tool for patients and GPs were used. MAIN OUTCOME MEASURES: Primary outcomes were QoL, measured by the global health scales of the European Organisation for Research and Treatment of Cancer QoL Questionnaire (EORTC QLQ C-30) and EuroQol-5D (EQ-5D). Secondary outcomes were cost-effectiveness and time to cancer diagnoses. RESULTS: 110 patients were randomised to intervention (n=55) or control (n=55), and followed by 78 GPs (942 follow-up months) and 70 surgeons (942 follow-up months), respectively. Compared to baseline, there was a significant improvement in postoperative QoL (p=0.003), but no differences between groups were revealed (mean difference at 1, 3, 6, 9, 12, 15, 18, 21 and 24-month follow-up appointments): Global Health; Δ-2.23, p=0.20; EQ-5D index; Δ-0.10, p=0.48, EQ-5D VAS; Δ-1.1, p=0.44. There were no differences in time to recurrent cancer diagnosis (GP 35 days vs surgeon 45 days, p=0.46); 14 recurrences were detected (GP 6 vs surgeon 8) and 7 metastases surgeries performed (GP 3 vs surgeon 4). The follow-up programme initiated 1186 healthcare contacts (GP 678 vs surgeon 508), 1105 diagnostic tests (GP 592 vs surgeon 513) and 778 hospital travels (GP 250 vs surgeon 528). GP organised follow-up was associated with societal cost savings (£8233 vs £9889, p<0.001). CONCLUSIONS: GP-organised follow-up was associated with no decline in QoL, no increase in time to recurrent cancer diagnosis and cost savings. TRIAL REGISTRATION: ClinicalTrials.gov identifier NCT00572143.

Long-term engagement with a mobile self-management system for people with type 2 diabetes.

BACKGROUND: In a growing number of intervention studies, mobile phones are used to support self-management of people with Type 2 diabetes mellitus (T2DM). However, it is difficult to establish knowledge about factors associated with intervention effects, due to considerable differences in research designs and outcome measures as well as a lack of detailed information about participants' engagement with the intervention tool. OBJECTIVE: To contribute toward accumulating knowledge about factors associated with usage and usability of a mobile self-management application over time through a thorough analysis of multiple types of investigation on each participant's engagement. METHODS: The Few Touch application is a mobile-phone-based self-management tool for patients with T2DM. Twelve patients with T2DM who have been actively involved in the system design used the Few Touch application in a real-life setting from September 2008 until October 2009. During this period, questionnaires and semistructured interviews were conducted. Recorded data were analyzed to investigate usage trends and patterns. Transcripts from interviews were thematically analyzed, and the results were further analyzed in relation to the questionnaire answers and the usage trends and patterns. RESULTS: The Few Touch application served as a flexible learning tool for the participants, responsive to their spontaneous needs, as well as supporting regular self-monitoring. A significantly decreasing (P<.05) usage trend was observed among 10 out of the 12 participants, though the magnitude of the decrease varied widely. Having achieved a sense of mastery over diabetes and experiences of problems were identified as reasons for declining motivation to continue using the application. Some of the problems stemmed from difficulties in integrating the use of the application into each participant's everyday life and needs, although the design concepts were developed in the process where the participants were involved. The following factors were identified as associated with usability and/or usage over time: Integration with everyday life; automation; balance between accuracy and meaningfulness of data with manual entry; intuitive and informative feedback; and rich learning materials, especially about foods. CONCLUSION: Many grounded design implications were identified through a thorough analysis of results from multiple types of investigations obtained through a year-long field trial of the Few Touch application. The study showed the importance and value of involving patient-users in a long-term trial of a tool to identify factors influencing usage and usability over time. In addition, the study confirmed the importance of detailed analyses of each participant's usage of the provided tool for better understanding of participants' engagement over time.

Using noninferiority tests to evaluate telemedicine and e-health services: systematic review.

BACKGROUND: An increasing number of studies within the field of telemedicine and e-health are designed as noninferiority studies, aiming to show that the telemedicine/e-health solution is not inferior to the traditional way of treating patients. OBJECTIVE: The objective is to review and sum up the status of noninferiority studies within this field, describing advantages and pitfalls of this approach. METHODS: PubMed was searched according to defined criteria, and 16 relevant articles were identified from the period 2008-June 2011. RESULTS: Most of the studies were related to the fields of psychiatry and emergency medicine, and most were published in journals relating to these fields or in general scientific or general medicine journals. All the studies claimed to be noninferiority studies, but 7 out of 16 tested for statistical differences as a proxy of noninferiority. CONCLUSIONS: The methodological quality of the studies varied. We discuss optimal procedures for future noninferiority studies within the field of telemedicine and e-health and situations in which this approach is most appropriate.

Mobile phone-based pattern recognition and data analysis for patients with type 1 diabetes.

BACKGROUND: Persons with type 1 diabetes who use electronic self-help tools, most commonly blood glucose meters, record a large amount of data about their personal condition. Mobile phones are powerful and ubiquitous computers that have a potential for data analysis, and the purpose of this study is to explore how self-gathered data can help users improve their blood glucose management. SUBJECTS AND METHODS: Thirty patients with insulin-regulated type 1 diabetes were equipped with a mobile phone application for 3-6 months, recording blood glucose, insulin, dietary information, physical activity, and disease symptoms. The data were analyzed in terms of usage of the different modules and which data processing and visualization tools could be constructed to support the use of these data. RESULTS: Eighteen patients (denoted "adopters") recorded complete data for over 80 consecutive days, up to 247 days. Among those who withdrew or did not use the application extensively, the most common reasons given were outdated or difficult-to-use phone. Data analysis using period finding and scale-space trends was found to yield significant patterns for most adopters. Pattern recognition methods to predict low or high blood glucose were found to be performing poorly. CONCLUSIONS: Minimally intrusive mobile applications enable users with type 1 diabetes to record data that can provide data-driven feedback to the user, potentially providing relevant insight into their disease.

Mobile health applications to assist patients with diabetes: lessons learned and design implications.

Self-management is critical to achieving diabetes treatment goals. Mobile phones and Bluetooth® can supportself-management and lifestyle changes for chronic diseases such as diabetes. A mobile health (mHealth) research platform--the Few Touch Application (FTA)--is a tool designed to support the self-management of diabetes. The FTA consists of a mobile phone-based diabetes diary, which can be updated both manually from user input and automatically by wireless data transfer, and which provides personalized decision support for the achievement of personal health goals. Studies and applications (apps) based on FTAs have included: (1) automatic transfer of blood glucose (BG) data; (2) short message service (SMS)-based education for type 1diabetes (T1DM); (3) a diabetes diary for type 2 diabetes (T2DM); (4) integrating a patient diabetes diary with health care (HC) providers; (5) a diabetes diary for T1DM; (6) a food picture diary for T1DM; (7) physical activity monitoring for T2DM; (8) nutrition information for T2DM; (9) context sensitivity in mobile self-help tools; and (10) modeling of BG using mobile phones. We have analyzed the performance of these 10 FTA-based apps to identify lessons for designing the most effective mHealth apps. From each of the 10 apps of FTA, respectively, we conclude: (1) automatic BG data transfer is easy to use and provides reassurance; (2) SMS-based education facilitates parent-child communication in T1DM; (3) the T2DM mobile phone diary encourages reflection; (4) the mobile phone diary enhances discussion between patients and HC professionals; (5) the T1DM mobile phone diary is useful and motivational; (6) the T1DM mobile phone picture diary is useful in identifying treatment obstacles; (7) the step counter with automatic data transfer promotes motivation and increases physical activity in T2DM; (8) food information on a phone for T2DM should not be at a detailed level; (9) context sensitivity has good prospects and is possible to implement on today's phones; and (10) BG modeling on mobile phones is promising for motivated T1DM users. We expect that the following elements will be important in future FTA designs: (A) automatic data transfer when possible; (B) motivational and visual user interfaces; (C) apps with considerable health benefits in relation to the effort required; (D) dynamic usage, e.g., both personal and together with HC personnel, long-/short-term perspective; and (E) inclusion of context sensitivity in apps. We conclude that mHealth apps will empower patients to take a more active role in managing their own health.