The Local Pickup and Delivery Problem (LPDP) has drawn much attention, and optimization models and algorithms have been developed to address this problem. However, for real world applications, the large-scale and dynamic nature of the problem causes difficulties in getting good solutions within acceptable time through standard optimization approaches. Meanwhile, actual dispatching solutions made by field experts in transportation companies contain embedded dispatching rules. This paper introduces a Data Mining-based Dispatching System (DMDS) to first learn dispatching rules from historical data and then generate dispatch solutions, which are shown to be as good as those generated by expert dispatchers in the intermodal freight industry. Three additional benefits of DMDS are: (1) it provides a simulation platform for strategic decision making and analysis; (2) the learned dispatching rules are valuable to combine with an optimization algorithm to improve the solution quality for LPDPs; (3) by adding optimized solutions to the training data, DMDS is capable to generate better-than-actuals solutions very quickly.

An 18-month-old Indian girl presented to her pediatrician with a history of shivering spells for more than a month, but worsening during the past week. The episodes occurred 1–2 times daily and lasted 15–30 minutes. The parents had not noted any fevers. The child had a poor appetite for 2 days and developed a rash on her extremities on the day of presentation. She had no respiratory symptoms, vomiting, or diarrhea. In the pediatrician’s office, the patient was febrile to 40°C. A complete blood count (CBC) showed a white blood cell (WBC) count of 42.5 × 109/L, hemoglobin of 10.4 g/dL, and platelet count of 21 × 109/L. She was referred to the emergency department (ED) for further evaluation. Additional history obtained in the ED revealed that the child had emigrated from Mumbai, India 1 month previously. She was a full-term infant, born by vaginal delivery. At 3 months of age, she developed an intracranial hemorrhage, which was complicated by hydrocephalus, necessitating ventriculoperitoneal (VP) shunt placement at 4 months of age. The patient received blood products before this surgery. She also had an isolated seizure at the time and was given a short course of anticonvulsant therapy. According to her parents, a coagulopathy evaluation performed in India revealed a factor II (prothrombin) deficiency. Thereafter, the child had no further seizure activity or unusual bleeding or bruising. The child had no other major medical problems. There were no significant exposures apart from mosquito bites while still in India. The parents denied contact with anyone who might have tuberculosis. There was no family history of bleeding diathesis or consanguinity. Immunizations were up to date, and the patient was not currently taking any medications. Examination in the ED revealed a small, nontoxic child with height of 69 cm, weight of 7 kg, and head circumference of 43 cm—-all less than the 3rd percentile for age. The patient had a temperature of 35.4°C, heart rate of 137 beats/min, respiratory rate of 24 breaths/min, and blood pressure of 93/70 mm Hg. There were petechiae on her face, arms, and legs, but no purpura or other rashes on her body. The cardiac examination was notable for a harsh III/VI systolic murmur, but lungs were clear to auscultation and abdomen had no organomegaly. There was no lymphadenopathy. Initial laboratory evaluation revealed a WBC count of 35.6 × 109/L, with 51% segmented neutrophils, 9% band forms, 31% lymphocytes, 6% monocytes, 1% eosinophils, and 2% reactive lymphocytes; hemoglobin of 9.1 g/dL; and platelet count of 8 × 109/L. C-reactive protein was 9.3 mg/dL. PT was prolonged to 19.5 seconds with an INR of 1.7, PTT was 35.9 seconds, and fibrinogen was 101 mg/dL. Metabolic panel, including creatinine and liver transaminases, was normal, except for bicarbonate of 13 meq/L and albumin of 3 g/dL. Urinalysis showed trace protein, negative nitrite, trace leukocyte esterase, 18 WBC/mm3, and 6 red blood cells/mm3. Urine culture was negative. After 22 hours, the admission blood culture grew coagulase-negative staphylococcus (CoNS), which was suspected to be a contaminant. Thin and thick smears of the peripheral blood were negative for malarial parasites. Parvovirus and HIV antibody testing was negative; IgG for dengue fever was positive but the IgM was negative. Stool culture showed normal flora, and a PPD skin test was negative. Chest radiograph demonstrated low lung volumes with peribronchial thickening but no focal infiltrate; the VP shunt tubing was coursing along the right neck and hemithorax and into the abdomen, where it looped upward with the tip terminating at the right hemidiaphragm. A computed tomography (CT) scan of the head was notable for moderate ventriculomegaly with postsurgical changes in the posterior fossa and VP shunt in situ. The child was admitted to the hospital and was empirically treated with ceftriaxone for possible bacterial sepsis. For several days, she required multiple transfusions of fresh frozen plasma, platelets, and packed red blood cells as a result of a presumed consumptive coagulopathy. In addition to the admission blood culture, 4 subsequent blood cultures drawn on consecutive days were also positive for CoNS, further identified as methicillin-resistant Staphylococcus epidermidis. An echocardiogram revealed a large, mobile vegetation attached to the wall of the right atrium that intermittently prolapsed across the tricuspid valve, extending into the right ventricle during diastole and flopping back into the right atrium during systole. The tricuspid valve leaflets appeared normal but moderate insufficiency was present. The echocardiogram otherwise demonstrated normal mitral, pulmonary, and aortic valves without insufficiency, normal biventricular size and function, and no pleural or pericardial effusions. Given the presence of this intra-atrial vegetation associated with multiple positive blood cultures, the patient was diagnosed with endocarditis, caused by CoNS, and antimicrobial therapy was changed from ceftriaxone to intravenous vancomycin and gentamicin, along with oral rifampin. Further evaluation revealed an unexpected finding which explained the diagnosis. For denouement see p. 91.

To investigate the impact of insulin plus other risk factors on cardiovascular disease (CVD) events in patients with type 2 diabetes (T2D), we conducted a retrospective study among patients with T2D from a U.S. managed care plan (n = 342,692). Date of first CVD event was defined as index date. For patients without events, index date was date of last-observed claim. CVD event rates were calculated; odds of CVD event were compared for patients with/without insulin use. Events analyzed included stroke, MI, other cardiac/cerebrovascular events.Among insulin group (n = 14,167), 22 patients/1000 patient-years experienced a CVD event, compared with 19/1000 patient-years in non-insulin group (n = 328,077). Adjusting for risk factors and comorbidities, odds of event for insulin group compared to non-insulin group was 0.66, ranging from 25% lower for patients aged 65+ to 42% lower for patients aged 31-45.Insulin treatment of T2D was associated with reduced risk of CVD events compared to other or no therapy. (C) 2008 Elsevier Ireland Ltd. All rights reserved.

The Local Pickup and Delivery Problem (LPDP) has drawn much attention during recent years. In the literature, optimization models and algorithms have been developed to address this problem. However, for some real world applications, the large-scale and dynamic nature of the problem causes some difficulties in getting good solutions within an acceptable time through standard optimization approaches. On the other hand, actual dispatching solutions made by field experts in transportation companies contain embedded useful dispatching rules. This paper presents a general Data Mining-Based Decision System (DMBDS) framework to mimic current dispatch processes and generate solutions for LPDPs by learning from historical data. An application in the intermodal freight industry is presented, where the DMBDS provides good solutions which are comparable to the real solutions provided by expert dispatchers, with respect to a set of key performance indicators.

PURPOSE:We examined whether men with erectile dysfunction are more likely to have diabetes mellitus than men without erectile dysfunction, and whether erectile dysfunction can be used as an observable early marker of diabetes mellitus. MATERIALS AND METHODS:Using a nationally representative managed care claims database from 51 health plans and 28 million members in the United States, we conducted a retrospective cohort study to compare the prevalence rates of diabetes mellitus between men with erectile dysfunction (285,436) and men without erectile dysfunction (1,584,230) during 1995 to 2001. Logistic regression models were used to isolate the effect of erectile dysfunction on the likelihood of having diabetes mellitus with adjustment for age, region and 7 concurrent diseases. RESULTS:The diabetes mellitus prevalence rates were 20.0% in men with erectile dysfunction and 7.5% in men without erectile dysfunction. With adjustment for age, region and concurrent diseases, the odds ratio of having diabetes mellitus between men with erectile dysfunction and without erectile dysfunction was 1.60 (p <0.0001). With adjustment for regions and concurrent diseases, the age specific odds ratios ranged from 2.94 (p <0.0001, age 26 to 35) to 1.05 (p = 0.1717, age 76 to 85). CONCLUSIONS:Men with erectile dysfunction were more than twice as likely to have diabetes mellitus as men without erectile dysfunction. Erectile dysfunction is an observable marker of diabetes mellitus, strongly so for men 45 years old or younger and likely for men 46 to 65 years old, but it is not a marker for men older than 66 years.

PURPOSE:: We examined whether men with erectile dysfunction (ED) are more likely to have hypertension than men without ED in a managed care setting.MATERIALS AND METHODS:: We used a naturalistic cohort design to compare hypertension prevalence rates in 285,436 men with ED to that in 1,584,230 men without ED from 1995 through 2001. We also used a logistic regression model to isolate the effect of ED on the likelihood of hypertension after controlling for subject age, census regions and 9 concurrent diseases. The ED and the nonED cohort came from a nationally representative, managed care claims database that covers 51 health plans and 28 million members in the United States. Finally, the prevalence rate difference between members with and without ED, and the OR of having hypertension were calculated.RESULTS:: The hypertension prevalence rate was 41.2% in men with ED and 19.2% in men without ED. After controlling for subject age, census region and 9 concurrent diseases the OR was 1.383 (p <0.0001), which implies that the odds for men with ED to have hypertension were 38.3% higher than the odds for men without ED.CONCLUSIONS:: Men with ED were more likely to have hypertension than men without ED. This evidence supports the hypothesis that ED shares common risk factors with hypertension. It also suggests that men with ED and clinicians could use ED as an alerting signal to detect and treat undiagnosed hypertension earlier.

PURPOSE:We examined the direct costs of erectile dysfunction (ED) empirically. MATERIALS AND METHODS:A naturalistic cohort study was done in 285,436 patients with ED and 51 health plans that covered 28 million lives nationwide from 1999 through 2001. Based on claims that had an ED related diagnosis code, procedure code or medication code we categorized the cost structure of ED care and calculated the annual costs of ED care per patient with ED, per user and per member monthly for individual and for all categories of ED care. RESULTS:A patient with ED in a health plan spent about an average of 83.91 dollars in 1999, 95.41 dollars in 2000 and 119.26 dollars in 2001 for ED care. In 2001, 37.08% of ED care costs per patient with ED were spent on phosphodiesterase type 5 (PDE-5) inhibitor therapy, 14.36% were spent on physician office visits, 10.19% were spent on diagnosis procedures, 8.45% were spent on testosterone hormone therapy, 3.85% were spent on penile prosthesis implantation, 4.41% were spent on intracavernous injection, 2.68% were spent on alprostadil pellet insertion and 0.81% was spent on vacuum erection devices. Of the 7 commonly used ED treatments PDE-5 inhibitor therapy has the lowest annual cost per user. CONCLUSIONS:In 2001 ED imposed a 122,669 dollars annual burden to a health plan with 100,000 members, that is or 0.108 dollars per member monthly. Each patient with ED spent 119.26 dollars annually for all ED related services or treatments. Of the 7 commonly used treatments PDE-5 inhibitor therapy had the lowest annual cost per user.

Objectives: The primary objectives of this research were to: (i) identify and present methodologies for estimating three types of ‘cost-of-illness’ measures using healthcare and disability claims data — specifically ‘cost of treatment’, ‘incremental cost of patient’, and ‘incremental cost of illness’; and (ii) perform a case-study analysis of these cost measures for women treated for stress urinary incontinence (SUI). Study Design and Methods: In this paper, we discuss aspects of cost-of-illness methodologies in the context of SUI. We first distinguish between ‘cost of treatment’ (i.e. the costs of treating a specific condition), ‘incremental cost of patient’ (i.e. the additional costs associated with patients with a particular condition, irrespective of any comorbid conditions they may also have), and ‘incremental cost of illness’ (i.e. the additional costs resulting from a particular illness, as distinct from the costs of other conditions that the patient might have, including conditions which might have caused the illness in question). The latter case is in many ways the most complex to model, requiring controls for related causal conditions. We then applied these three methodologies by analysing the costs associated with SUI. Using data from a large employer claims database (n >100 000), we estimated a series of regression models that reflected cost of treatment, incremental cost of patient, and incremental cost of illness for SUI. Results: The three approaches yielded substantially different results. For many purposes the incremental cost-of-illness model provides the most appropriate results, as it controls for comorbid conditions, as well as patient demographics. On a per capita basis using the incremental cost-of-illness model, patients with SUI had direct costs that were 134% more than those for their controls and indirect costs that were 163% more than those for controls. Estimating costs for the average (i.e. mean) person results in dollar-termed estimates of the costs of SUI. In particular, we found that in 1998, the average direct medical cost of SUI was $US5642 and the indirect workplace cost of SUI was $US4208. Conclusions: Since the various methods yield substantially different results, it is important that the end user of cost-of-illness analyses of claims data have a clear purpose in mind when reporting the cost of the condition of concern. The incremental cost-of-illness measure for claims data has substantial advantages in terms of enhancing our understanding of the specific cost impact of SUI.

You have accessJournal of UrologyDiscussed Poster, Tuesday, May 11, 2004, 1:00 - 5:00 pm1 Apr 20041620: The Comparison of Hypertension Prevalence Rates Between Men with and without Erectile Dysfunction: Evidence From A Large National Managed Care Claims Database Peter Sun, and Ralph Swindle Peter SunPeter Sun More articles by this author , and Ralph SwindleRalph Swindle More articles by this author View All Author Informationhttps://doi.org/10.1016/S0022-5347(18)38828-1AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail "1620: The Comparison of Hypertension Prevalence Rates Between Men with and without Erectile Dysfunction: Evidence From A Large National Managed Care Claims Database." The Journal of Urology, 171(4S), p. 428 © 2016 by American Urological AssociationFiguresReferencesRelatedDetails Volume 171Issue 4SApril 2004Page: 428 Advertisement Copyright & Permissions© 2016 by American Urological AssociationMetricsAuthor Information Peter Sun More articles by this author Ralph Swindle More articles by this author Expand All Advertisement Loading ...

PURPOSE:We quantified the prevalence of diagnosed hypertension, hyperlipidemia, diabetes mellitus and depression in male health plan members with erectile dysfunction (ED).MATERIALS AND METHODS:We used a nationally representative managed care claims database that covered 51 health plans with 28 million lives for 1995 through 2002. Based on 272325 identified patients with ED population and age specific prevalence rates were calculated for the same period.RESULTS:The crude population prevalence rates were 41.6% for hypertension, 42.4% for hyperlipidemia, 20.2% for diabetes mellitus, 11.1% for depression, 23.9% for hypertension and hyperlipidemia, 12.8% for hypertension and diabetes mellitus, and 11.5% for hyperlipidemia and depression. The crude age specific prevalence rates varied across age groups significantly for hypertension (4.5% to 68.4%), hyperlipidemia (3.9% to 52.3%), and diabetes mellitus (2.8% to 28.7%), and significantly less for depression (5.8% to 15.0%). Region adjusted population prevalence rates were 41.2% for hypertension, 41.8% for hyperlipidemia, 19.7% for diabetes mellitus and 11.9% for depression. Only 87163 patients with ED (32%) had no comorbid diagnosis of hypertension, hyperlipidemia, diabetes mellitus or depression.CONCLUSION:Hypertension, hyperlipidemia, diabetes mellitus and depression were prevalent in patients with ED. This evidence supported the proposition that ED shares common risk factors with these 4 concurrent conditions. Therefore, as a pathophysiological event, ED could be viewed as a potential observable marker for these concurrent diseases. This finding suggests that clinicians could include ED in the assessment profile of these concurrent conditions for earlier detection and treatment.

You have accessJournal of UrologyDiscussed Poster, Monday, May 10, 2004, 8:00 am - 12:00 pm1 Apr 2004869: The Prevalence of Four Diagnosed Concurrent Conditions Among Male Health Plan Members with Erectile Dysfunction Peter Sun, Allen D. Seftel, and Ralph Swindle Peter SunPeter Sun More articles by this author , Allen D. SeftelAllen D. Seftel More articles by this author , and Ralph SwindleRalph Swindle More articles by this author View All Author Informationhttps://doi.org/10.1016/S0022-5347(18)38118-7AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookLinked InTwitterEmail "869: The Prevalence of Four Diagnosed Concurrent Conditions Among Male Health Plan Members with Erectile Dysfunction." The Journal of Urology, 171(4S), p. 230 © 2016 by American Urological AssociationFiguresReferencesRelatedDetails Volume 171Issue 4SApril 2004Page: 230 Advertisement Copyright & Permissions© 2016 by American Urological AssociationMetricsAuthor Information Peter Sun More articles by this author Allen D. Seftel More articles by this author Ralph Swindle More articles by this author Expand All Advertisement PDF DownloadLoading ...