Constructing and evaluating alternative prediction models for identifying clinically relevant cases in an STD clinic

Thesis (Master's)--University of Washington, 2016-06

Development of prediction models for freeway incident durations using data mining techniques

The nation's freeway systems are becoming increasingly congested. A major contribution to traffic congestion on freeways is due to traffic incidents. Traffic incidents are non-recurring events such as accidents or stranded vehicles that cause a temporary roadway capacity reduction, and they can account for as much as 60 percent of all traffic congestion on freeways. One major freeway incident management strategy involves diverting traffic to avoid incident locations by relaying timely information through Intelligent Transportation Systems (ITS) devices such as dynamic message signs or real-time traveler information systems. The decision to divert traffic depends foremost on the expected duration of an incident, which is difficult to predict. In addition, the duration of an incident is affected by many contributing factors. Determining and understanding these factors can help the process of identifying and developing better strategies to reduce incident durations and alleviate traffic congestion. A number of research studies have attempted to develop models to predict incident durations, yet with limited success. ^ This dissertation research attempts to improve on this previous effort by applying data mining techniques to a comprehensive incident database maintained by the District 4 ITS Office of the Florida Department of Transportation (FDOT). Two categories of incident duration prediction models were developed: "offline" models designed for use in the performance evaluation of incident management programs, and "online" models for real-time prediction of incident duration to aid in the decision making of traffic diversion in the event of an ongoing incident. Multiple data mining analysis techniques were applied and evaluated in the research. The multiple linear regression analysis and decision tree based method were applied to develop the offline models, and the rule-based method and a tree algorithm called M5P were used to develop the online models. ^ The results show that the models in general can achieve high prediction accuracy within acceptable time intervals of the actual durations. The research also identifies some new contributing factors that have not been examined in past studies. As part of the research effort, software code was developed to implement the models in the existing software system of District 4 FDOT for actual applications. ^

Software stability assessment using multiple prediction models

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Software stability assessment using multiple prediction models

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Performance of protective factors assessment in risk prediction for adults: systematic review and meta-analysis

Protective factors are neglected in risk assessment in adult psychiatric and criminal justice populations. This review investigated the predictive efficacy of selected tools that assess protective factors. Five databases were searched using comprehensive terms for records up to June 2014, resulting in 17 studies (n = 2,198). Results were combined in a multilevel meta-analysis using the R (R Core Team, R: A Language and Environment for Statistical Computing, Vienna, Austria: R Foundation for Statistical Computing, 2015) metafor package (Viechtbauer, Journal of Statistical Software, 2010, 36, 1). Prediction of outcomes was poor relative to a reference category of violent offending, with the exception of prediction of discharge from secure units. There were no significant differences between the predictive efficacy of risk scales, protective scales, and summary judgments. Protective factor assessment may be clinically useful, but more development is required. Claims that use of these tools is therapeutically beneficial require testing.

Estimating glomerular filtration rate in acute coronary syndromes: Different equations, different mortality risk prediction

AIMS: Renal dysfunction is a powerful predictor of adverse outcomes in patients hospitalized for acute coronary syndrome. Three new glomerular filtration rate (GFR) estimating equations recently emerged, based on serum creatinine (CKD-EPIcreat), serum cystatin C (CKD-EPIcyst) or a combination of both (CKD-EPIcreat/cyst), and they are currently recommended to confirm the presence of renal dysfunction. Our aim was to analyse the predictive value of these new estimated GFR (eGFR) equations regarding mid-term mortality in patients with acute coronary syndrome, and compare them with the traditional Modification of Diet in Renal Disease (MDRD-4) formula. METHODS AND RESULTS: 801 patients admitted for acute coronary syndrome (age 67.3±13.3 years, 68.5% male) and followed for 23.6±9.8 months were included. For each equation, patient risk stratification was performed based on eGFR values: high-risk group (eGFR<60ml/min per 1.73m2) and low-risk group (eGFR⩾60ml/min per 1.73m2). The predictive performances of these equations were compared using area under each receiver operating characteristic curves (AUCs). Overall risk stratification improvement was assessed by the net reclassification improvement index. The incidence of the primary endpoint was 18.1%. The CKD-EPIcyst equation had the highest overall discriminate performance regarding mid-term mortality (AUC 0.782±0.20) and outperformed all other equations (ρ<0.001 in all comparisons). When compared with the MDRD-4 formula, the CKD-EPIcyst equation accurately reclassified a significant percentage of patients into more appropriate risk categories (net reclassification improvement index of 11.9% (p=0.003)). The CKD-EPIcyst equation added prognostic power to the Global Registry of Acute Coronary Events (GRACE) score in the prediction of mid-term mortality. CONCLUSION: The CKD-EPIcyst equation provides a novel and improved method for assessing the mid-term mortality risk in patients admitted for acute coronary syndrome, outperforming the most widely used formula (MDRD-4), and improving the predictive value of the GRACE score. These results reinforce the added value of cystatin C as a risk marker in these patients.

Redefining dermatologists’ role in skin cancer early detection and follow-up care

Early diagnosis of melanoma leads to the best prognosis for patients and may be more likely achieved when those who are at high risk for melanoma undergo regular and systematic monitoring. However, many people rarely or never see a dermatologist. Risk prediction models (recently reviewed by Usher-Smith et al ) could assist to triage people into preventive care appropriate for their risk profile. Most risk prediction models contain measures of phenotype including skin, eye and hair colour as well as genetic mutations. Almost all also contain the number and size of naevi, as well as the presence of naevi with atypical features which are independently associated with melanoma risk. In the absence of formal population-based screening programs for melanoma in most countries worldwide, people with high risk phenotypes may need to consider regular monitoring or self-monitoring of their naevi , especially since the vast majority of melanomas are found by people themselves or their friend and relatives. Another group of patients that will require regular monitoring are patients who have been successfully treated for their first melanoma, whose risk to develop a second melanoma is greatly increased . In a US study of 89,515 melanoma survivors those with a previous diagnosis of melanoma had a 9-fold increased risk of developing subsequent melanoma compared with the general population, equating to a rate of 3.76 per 1000 person-years, while in an Australian study, risk of subsequent melanoma was 6 per 1000 person-years. Regular follow-up is therefore essential for melanoma survivors, especially during the first few years after initial melanoma diagnosis.

Surveillance for ocular hypertension: An evidence synthesis and economic evaluation

OBJECTIVES: To determine effective and efficient monitoring criteria for ocular hypertension [raised intraocular pressure (IOP)] through (i) identification and validation of glaucoma risk prediction models; and (ii) development of models to determine optimal surveillance pathways.

DESIGN: A discrete event simulation economic modelling evaluation. Data from systematic reviews of risk prediction models and agreement between tonometers, secondary analyses of existing datasets (to validate identified risk models and determine optimal monitoring criteria) and public preferences were used to structure and populate the economic model.

SETTING: Primary and secondary care.

PARTICIPANTS: Adults with ocular hypertension (IOP > 21 mmHg) and the public (surveillance preferences).

INTERVENTIONS: We compared five pathways: two based on National Institute for Health and Clinical Excellence (NICE) guidelines with monitoring interval and treatment depending on initial risk stratification, 'NICE intensive' (4-monthly to annual monitoring) and 'NICE conservative' (6-monthly to biennial monitoring); two pathways, differing in location (hospital and community), with monitoring biennially and treatment initiated for a ≥ 6% 5-year glaucoma risk; and a 'treat all' pathway involving treatment with a prostaglandin analogue if IOP > 21 mmHg and IOP measured annually in the community.

MAIN OUTCOME MEASURES: Glaucoma cases detected; tonometer agreement; public preferences; costs; willingness to pay and quality-adjusted life-years (QALYs).

RESULTS: The best available glaucoma risk prediction model estimated the 5-year risk based on age and ocular predictors (IOP, central corneal thickness, optic nerve damage and index of visual field status). Taking the average of two IOP readings, by tonometry, true change was detected at two years. Sizeable measurement variability was noted between tonometers. There was a general public preference for monitoring; good communication and understanding of the process predicted service value. 'Treat all' was the least costly and 'NICE intensive' the most costly pathway. Biennial monitoring reduced the number of cases of glaucoma conversion compared with a 'treat all' pathway and provided more QALYs, but the incremental cost-effectiveness ratio (ICER) was considerably more than £30,000. The 'NICE intensive' pathway also avoided glaucoma conversion, but NICE-based pathways were either dominated (more costly and less effective) by biennial hospital monitoring or had a ICERs > £30,000. Results were not sensitive to the risk threshold for initiating surveillance but were sensitive to the risk threshold for initiating treatment, NHS costs and treatment adherence.

LIMITATIONS: Optimal monitoring intervals were based on IOP data. There were insufficient data to determine the optimal frequency of measurement of the visual field or optic nerve head for identification of glaucoma. The economic modelling took a 20-year time horizon which may be insufficient to capture long-term benefits. Sensitivity analyses may not fully capture the uncertainty surrounding parameter estimates.

CONCLUSIONS: For confirmed ocular hypertension, findings suggest that there is no clear benefit from intensive monitoring. Consideration of the patient experience is important. A cohort study is recommended to provide data to refine the glaucoma risk prediction model, determine the optimum type and frequency of serial glaucoma tests and estimate costs and patient preferences for monitoring and treatment.

FUNDING: The National Institute for Health Research Health Technology Assessment Programme.

Genetics of Diabetic Nephropathy: a Long Road of Discovery

The global prevalence of diabetic nephropathy is rising in parallel with the increasing incidence of diabetes in most countries. Unfortunately, up to 40 % of persons diagnosed with diabetes may develop kidney complications. Diabetic nephropathy is associated with substantially increased risks of cardiovascular disease and premature mortality. An inherited susceptibility to diabetic nephropathy exists, and progress is being made unravelling the genetic basis for nephropathy thanks to international research collaborations, shared biological resources and new analytical approaches. Multiple epidemiological studies have highlighted the clinical heterogeneity of nephropathy and the need for better phenotyping to help define important subgroups for analysis and increase the power of genetic studies. Collaborative genome-wide association studies for nephropathy have reported unique genes, highlighted novel biological pathways and suggested new disease mechanisms, but progress towards clinically relevant risk prediction models for diabetic nephropathy has been slow. This review summarises the current status, recent developments and ongoing challenges elucidating the genetics of diabetic nephropathy.

Monitoring ocular hypertension, how much and how often? A cost-effectiveness perspective

OBJECTIVE: To assess the efficiency of alternative monitoring services for people with ocular hypertension (OHT), a glaucoma risk factor.

DESIGN: Discrete event simulation model comparing five alternative care pathways: treatment at OHT diagnosis with minimal monitoring; biennial monitoring (primary and secondary care) with treatment if baseline predicted 5-year glaucoma risk is ≥6%; monitoring and treatment aligned to National Institute for Health and Care Excellence (NICE) glaucoma guidance (conservative and intensive).

SETTING: UK health services perspective.

PARTICIPANTS: Simulated cohort of 10 000 adults with OHT (mean intraocular pressure (IOP) 24.9 mm Hg (SD 2.4).

MAIN OUTCOME MEASURES: Costs, glaucoma detected, quality-adjusted life years (QALYs).

RESULTS: Treating at diagnosis was the least costly and least effective in avoiding glaucoma and progression. Intensive monitoring following NICE guidance was the most costly and effective. However, considering a wider cost-utility perspective, biennial monitoring was less costly and provided more QALYs than NICE pathways, but was unlikely to be cost-effective compared with treating at diagnosis (£86 717 per additional QALY gained). The findings were robust to risk thresholds for initiating monitoring but were sensitive to treatment threshold, National Health Service costs and treatment adherence.

CONCLUSIONS: For confirmed OHT, glaucoma monitoring more frequently than every 2 years is unlikely to be efficient. Primary treatment and minimal monitoring (assessing treatment responsiveness (IOP)) could be considered; however, further data to refine glaucoma risk prediction models and value patient preferences for treatment are needed. Consideration to innovative and affordable service redesign focused on treatment responsiveness rather than more glaucoma testing is recommended.

Perioperative cardiovascular evaluation: heads or tails?

When dealing with surgical patients, a perioperative evaluation is essential to anticipate complications and institute measures to reduce the risks. Several algorithms and exams have been used to identify postoperative cardiovascular events, which account for more than 50% of perioperative mortality. However, they are far from ideal. Some of these algorithms and exams were proposed before important advances in cardiology, at a time when pharmacological risk reduction strategies for surgical patients were not available. New biomarkers and exams, such as C-reactive protein, brain natriuretic peptide, and multislice computed tomography have been used in cardiology and have provided important prognostic information. The ankle-brachial index is another significant marker of atherosclerosis. However, specific information regarding the perioperative context of all these methods is still needed. The objective of this article is to evaluate cardiovascular risk prediction models after noncardiac surgery.

Collaborative Research in Childhood Cancer Survivorship: The Current Landscape.

Survivors of childhood cancer carry a substantial burden of morbidity and are at increased risk for premature death. Furthermore, clear associations exist between specific therapeutic exposures and the risk for a variety of long-term complications. The entire landscape of health issues encountered for decades after successful completion of treatment is currently being explored in various collaborative research settings. These settings include large population-based or multi-institutional cohorts and single-institution studies. The ascertainment of outcomes has depended on self-reporting, linkage to registries, or clinical assessments. Survivorship research in the cooperative group setting, such as the Children's Oncology Group, has leveraged the clinical trials infrastructure to explore the molecular underpinnings of treatment-related adverse events, and to understand specific complications in the setting of randomized risk-reduction strategies. This review highlights the salient findings from these large collaborative initiatives, emphasizing the need for life-long follow-up of survivors of childhood cancer, and describing the development of several guidelines and efforts toward harmonization. Finally, the review reinforces the need to identify populations at highest risk, facilitating the development of risk prediction models that would allow for targeted interventions across the entire trajectory of survivorship.

Variáveis predictivas de readmissão hospitalar nas enfermarias de medicina

Trabalho Final do Curso de Mestrado Integrado em Medicina, Faculdade de Medicina, Universidade de Lisboa, 2014

Modeling Hospital Readmissions in Dalarna County, Sweden

Unplanned hospital readmissions increase health and medical care costs and indicate lower the lower quality of the healthcare services. Hence, predicting patients at risk to be readmitted is of interest. Using administrative data of patients being treated in the medical centers and hospitals in the Dalarna County, Sweden, during 2008 – 2016 two risk prediction models of hospital readmission are built. The first model relies on the logistic regression (LR) approach, predicts correctly 2,648 out of 3,392 observed readmission in the test dataset, reaching a c-statistics of 0.69. The second model is built using random forests (RF) algorithm; correctly predicts 2,183 readmission (out of 3,366) and 13,198 non-readmission events (out of 18,982). The discriminating ability of the best performing RF model (c-statistic 0.60) is comparable to that of the logistic model. Although the discriminating ability of both LR and RF risk prediction models is relatively modest, still these models are capable to identify patients running high risk of hospital readmission. These patients can then be targeted with specific interventions, in order to prevent the readmission, improve patients’ quality of life and reduce health and medical care costs.

Factores asociados a la mortalidad en el síndrome de Sjögren

Antecedentes: El síndrome de Sjögren (SS) es una patología crónica, autoinmune, de características multifactoriales en su etiología. También es conocida como una epitelitis autoinmune, caracterizada por síntomas secos como xeroftalmia y xerostomía, pero que también puede tener compromiso sistémico, dado por manifestaciones extra-glandulares. En la actualidad es poco reconocida como tal, y por lo tanto, la tasa de sobrevida en estos pacientes se encuentra disminuida pero poco tenida en cuenta a la hora de la valoración de ellos. Este trabajo describe la evidencia encontrada acerca de las causas de mortalidad y sus factores asociados luego de realizar una revisión sistemática de la literatura. Objetivos: El objetivo de este estudio fue reunir de forma exhaustiva y sistemática toda la evidencia empírica, publicada o no, que cumpla los criterios de búsqueda y elegibilidad sobre factores asociados al incremento de la mortalidad o disminución en la sobrevida de los pacientes con diagnóstico de SS. Métodos: Se realizó una revisión sistemática de la literatura mediante una búsqueda exhaustiva de todos los estudios publicados en las bases de datos electrónicas preestablecidas, hasta abril de 2015, con el fin de determinar las causas más frecuentes de mortalidad en pacientes con SS y los factores asociados a ella. Resultados: Se encontraron 4,654 resultados que coincidían con los criterios de búsqueda establecidos; de estos, 33 cumplieron con los criterios de inclusión y se distribuyeron de la siguiente forma: el 66.6% (22/33) correspondieron a estudios de corte cohorte, 30.3% (10/33) a estudios de corte transversal y el 3.03% (1/33) a estudios casos y controles. Se obtuvieron resultados en cuanto a frecuencias de mortalidad, razón estandarizada de mortalidad, tasas de supervivencia, causas más frecuentes de mortalidad y sus factores asociados. Conclusiones: La mortalidad reportada en los diferentes estudios fue entre el 1.2% hasta el 30%. Aquellos estudios que reportaron una tasa de mortalidad inferior al 5%, tuvieron un tiempo de seguimiento menor 8 años [1,7,33,60,64,86]. La mayoría de los casos sigue un curso relativamente estable, pero hay un porcentaje importante que presenta otras manifestaciones sistémicas con mayor frecuencia de complicaciones durante la evolución del SS. Por tanto, son los que requieren un seguimiento más estrecho, debido a una mayor necesidad de tratamiento sistémico y al mayor riesgo de ingreso hospitalario y de mortalidad, especialmente por el desarrollo de procesos linfoproliferativos B. La presencia de factores pronósticos en el paciente con SS obligará a realizar un seguimiento clínico e inmunológico mucho más estrecho, lo cual permitirá identificar lo antes posible las complicaciones que puedan aparecer e instaurar las correspondientes medidas terapéuticas, para aumentar las tasas de supervivencia.