Effects of Observer Variation on Performance in Probabilistic Diagnosis of Jaundice

1981 ◽  
Vol 20 (03) ◽  
pp. 163-168 ◽  
Author(s):  
G. Llndberg

A system for probabilistic diagnosis of jaundice has been used for studying the effects of taking into account the unreliability of diagnostic data caused by observer variation. Fourteen features from history and physical examination were studied. Bayes’ theorem was used for calculating the probabilities of a patient’s belonging to each of four diagnostic categories.The construction sample consisted of 61 patients. An equal number of patients were tested in the evaluation sample. Observer variation on the fourteen features had been assessed in two previous studies. The use of kappa-statistics for measuring observer variation allowed the construction of a probability transition matrix for each feature. Diagnostic probabilities could then be calculated with and without the inclusion of weights for observer variation. Tests of system performance revealed that discriminatory power remained unchanged. However, the predictions rendered by the variation-weighted system were diffident. It is concluded that taking observer variation into account may weaken the sharpness of probabilistic diagnosis but it may also help to explain the value of probabilistic diagnosis in future applications.

1982 ◽  
Vol 21 (03) ◽  
pp. 137-142 ◽  
Author(s):  
C. Helmers ◽  
G. Lindberg

Four observers interviewed and examined 19 jaundiced patients, recording sixteen common symptoms and ten clinical signs that had been defined in advance. Kappa statistics were used for evaluating the data. All studied symptoms and seven signs showed agreement between observers significantly greater than expected by chance. The clinical significance of inter-observer variation was studied in a set of 144 jaundiced patients. The diagnostic value of studied symptoms and signs was calculated before and after correction for inter-observer variation. Only five symptoms: itching, loose bowels, ache, ache/pain description and alcohol intake, and three signs: spider naevi, palmar erythema and ascites retained more than 60% of their diagnostic value after correction for inter-observer variation.


1982 ◽  
Vol 56 (5) ◽  
pp. 628-633 ◽  
Author(s):  
Kenneth W. Lindsay ◽  
Graham Teasdale ◽  
Robin P. Knill-Jones ◽  
Lilian Murray

✓ The management of individual patients with subarachnoid hemorrhage depends greatly on assessment of the patient's clinical condition. Difficulty in applying current grading systems prompted the authors to conduct studies of observer variability and to attempt to identify sources of inconsistency. Observers graded 15 patients by both the Hunt and Hess and Nishioka systems. Considerable observer variability was found, with up to four different grades being selected for the same patient. Kappa statistics were used to evaluate the data. This method determines observer agreement occurring in excess of chance. Kappa values for each grading system showed observer agreement to be significantly better than chance, yet revealed marked observer variation. Most variation occurred when Grade 3 was selected, irrespective of the system used. In a further study where observers graded clinical summaries, similar variation occurred; therefore, inconsistency was due mainly to difficulty in matching patients with levels described in the grading system, rather than to fluctuation in the patients' clinical condition or difference in the observers' examination technique. Variability was high when patients with systemic disease or vasospasm on angiography were graded with the Hunt and Hess system. The studies show that a simpler and more reliable grading system is required, and emphasize the need for caution when interpreting the results from different published series.


Author(s):  
Pawan Gupta

A significant number of patients attending the ED are those who are often referred to as ‘minors’, ‘streamers’, ‘walking wounded’, etc. These include patients with minor injuries, wounds, fractures or other soft tissue injuries. Therefore, a basic knowledge of anatomy and its application in various circumstances is mandatory. The injuries mentioned above are rarely life-threatening, but they may be limb-threatening and severely disabling. So it is extremely important to avoid errors in diagnosis and management, and to know when to ask for help at the appropriate time. By following the key principles listed below, you will be able to avoid many problems with such patients: • In the history, a detailed description of the mechanism of injury and the patient’s complaint will help in predicting the type of injury sustained. • A careful and thorough physical examination can point to the site and type of injury, on the basis of which appropriate radiological images can then be requested. • A neurovascular examination must be completed and documented in every limb injury, before and after any reductions, and before and after immobilization. • Appropriate radiological imaging, accompanied by a thorough physical examination, can pick up injuries with a high degree of accuracy. Inadequate radiographic films should not be accepted. • Immobilize the patient if a fracture is clinically suspected even if the X-rays are negative. • In cases of dislocations or subluxations, X-rays should be done before and after reductions, except when a delay could be potentially harmful to the patient (for example, when a severe traumatic deformity of a joint threatens to jeopardize the viability of the overlying skin). • The patient should be able to mobilize safely before being discharged from the ED. • Patients should be given proper aftercare instructions before leaving the ED, including how to look after themselves and to recognize limb-threatening features, the follow-up arrangement, and to return if things go wrong. • Ask for senior help if you are not sure about an injury or its management.


Author(s):  
Norbert Lameire ◽  
Raymond Vanholder ◽  
Wim Van Biesen

The prognosis of acute kidney injury (AKI) depends on early diagnosis and therapy. A multitude of causes are classified according to their origin as prerenal, intrinsic (intrarenal), and post-renal.Prerenal AKI means a loss of renal function despite intact nephrons, for example, because of volume depletion and/or hypotension.There is a broad spectrum of intrinsic causes of AKI including acute tubular necrosis (ATN), interstitial nephritis, glomerulonephritis, and vasculitis. Evaluation includes careful review of the patient’s history, physical examination, urinalysis, selected urine chemistries, imaging of the urinary tree, and eventual kidney biopsy. The history should focus on the tempo of loss of function (if known), associated systemic diseases, and symptoms related to the urinary tract (especially those that suggest obstruction). In addition, a review of the medications looking for potentially nephrotoxic drugs is essential. The physical examination is directed towards the identification of findings of a systemic disease and a detailed assessment of the patient’s haemodynamic status. This latter goal may require invasive monitoring, especially in the oliguric patient with conflicting clinical findings, where the physical examination has limited accuracy.Excluding urinary tract obstruction is necessary in all cases and may be established easily by renal ultrasound.Distinction between the two most common causes of AKI (prerenal AKI and ATN) is sometimes difficult, especially because the clinical examination is often misleading in the setting of mild volume depletion or overload. Urinary chemistries, like calculation of the fractional excretion of sodium (FENa), may be used to help in this distinction. In contrast to FENa, the fractional excretion of urea has the advantage of being rather independent of diuretic therapy. Response to fluid repletion is still regarded as the gold standard in the differentiation between prerenal and intrinsic AKI. Return of renal function to baseline or resuming of diuresis within 24 to 72 hours is considered to indicate ‘transient, mostly prerenal AKI’, whereas persistent renal failure usually indicates intrinsic disease. Transient AKI may, however, also occur in short-lived ATN. Furthermore, rapid fluid application is contraindicated in a substantial number of patients, such as those with congestive heart failure.‘Muddy brown’ casts and/or tubular epithelial cell casts in the urine sediment are typically seen in patients with ATN. Their presence is an important tool in the distinction between ATN and prerenal AKI, which is characterized by a normal sediment, or by occasional hyaline casts. There is a possible role for new serum and/or urinary biomarkers in the diagnosis and prognosis of the patient with AKI, including the differential diagnosis between pre-renal AKI and ATN. Further studies are needed before their routine determination can be recommended.When a diagnosis cannot be made with reasonable certainty through this evaluation, renal biopsy should be considered; when intrarenal causes such as crescentic glomerulonephritis or vasculitis are suspected, immediate biopsy to avoid delay in the initiation of therapy is mandatory.


2017 ◽  
Vol 18 (01) ◽  
pp. 78-85
Author(s):  
Dony Permana

Customer Lifetime Value, familiar as CLV is valuability a customer in marketing system. High CLV has a meaning that the customer will bring in a big return for a firm. CLV is determined by some factors, as retention rate, acquisition rate, some costs, product price, and interest rates. Markov Chain is one of model that used to determine CLV.  In Markov Chain, a customer is assumed some state. Transition inter states are assumed Markovian. Here, we make CLV model using Markov Chain with four states. There are four type of model that have four states. Each type have different transition chart and of course have different probability transition matrix. Here, we describe every type of CLV model using Markov Chain.


2021 ◽  
Vol 108 (Supplement_7) ◽  
Author(s):  
B Sebastian ◽  
B Mirshekar-Syahkal ◽  
T Athisayaraj ◽  
N Ward

Abstract Background With the increased awareness and push for earlier diagnosis of colorectal cancer, the number of patients undergoing colonoscopy is increasing. Being a common condition, a number of these patients will have herniae. We describe a rare complication during a Sigmoidoscopy in a patient with an Inguinal hernia. The case A 75-year-old man was booked for a flexible sigmoidoscopy for rectal bleeding. The scope was successfully inserted to the Splenic flexure. During withdrawal, the scope stopped moving. It was still possible to advance the scope, but not to withdraw. The patient confirmed the presence of a left inguinal hernia. Physical examination and the position on magnetic scope imager confirmed the loop in the hernia. Various manoeuvres to withdraw the scope were unsuccessful. A colleague was called for a second opinion. We came up with a plan to maintain a ‘long loop‘ position in the hernia, by holding the scope through the scrotum and allowing it to slide on withdrawal. This was successful and the patient was discharged. Discussion Incarceration of the scope in the hernia occurs when the hernial defect permits entry and exit of the scope, leaving a loop in the hernia, in a long loop position. During withdrawal, the configuration changes to a short loop, crowding the hernial defect, preventing the scope from sliding. On searching literature, we found that this technique has been described by Koltun et al and is known as the “Pulley” technique. We suggest that colonoscopists are familiar with this technique.


1982 ◽  
Vol 21 (03) ◽  
pp. 114-116
Author(s):  
D. J. Spiegelhalter

Lindberg [2] proposes an adjustment of a probabilistic system for diagnosis in the light of known observer variation in the eliciting of symptoms and signs. He finds that his proposal leads to somewhat diffident predictions. We show that such an adjustment is only appropriate when the system is implemented with observers who are less reliable than those who created the data-base. In this case, with certain assumptions, an adjustment may be made that is less radical than that of Lindberg. A simple numerical solution for binary symptoms is provided.


2014 ◽  
Vol 14 (01) ◽  
pp. 1550003 ◽  
Author(s):  
Liu Yang ◽  
Kai-Xuan Zheng ◽  
Neng-Gang Xie ◽  
Ye Ye ◽  
Lu Wang

For a multi-agent spatial Parrondo's model that it is composed of games A and B, we use the discrete time Markov chains to derive the probability transition matrix. Then, we respectively deduce the stationary distribution for games A and B played individually and the randomized combination of game A + B. We notice that under a specific set of parameters, two absorbing states instead of a fixed stationary distribution exist in game B. However, the randomized game A + B can jump out of the absorbing states of game B and has a fixed stationary distribution because of the "agitating" role of game A. Moreover, starting at different initial states, we deduce the probabilities of absorption at the two absorbing barriers.


1993 ◽  
Vol 78 (6) ◽  
pp. 884-890 ◽  
Author(s):  
Lee A. Kearse ◽  
Dean Martin ◽  
Kathleen McPeck ◽  
Maria Lopez-Bresnahan

✓ The purpose of this prospective study was twofold: 1) to determine the sensitivity and specificity of computer-derived density spectral array in detecting analog electroencephalographic (EEG) ischemic pattern changes during carotid artery cross-clamping in patients undergoing carotid endarterectomy; and 2) to assess the ability of density spectral array to identify such changes in comparison with the degree and type of change seen in the analog EEG ischemic pattern. Sixteen channels of anteroposterior bipolar and two to four channels of referential electroencephalography with four channels of density spectral array were monitored simultaneously during carotid endarterectomy in 103 patients under general anesthesia. Two “observers” interpreted the density spectral array and the analog electroencephalograms, one during and immediately after the operations and the other 6 months after completion of all surgery. Analyses were conducted to establish both the number of patients with analog EEG ischemic changes and the number of ischemia events during carotid artery cross-clamping. Observer A indicated that the density spectral array identified analog EEG ischemic changes in 21 of 29 patients, for a sensitivity of 72% (specificity 99%), whereas Observer B's results showed that the density spectral array identified analog EEG ischemic changes in 16 of 27 patients, for a sensitivity of 59% (specificity 96%). Density spectral array detection of analog EEG ischemic changes based on severity classifications were 61% and 18% in the mild group, 70% and 71% in the moderate group, and 95% in the severe group, indicating a relationship between density spectral array sensitivity and severity of analog EEG ischemic change, with p = 0.02 and p = 0.004 for the two observers. The kappa statistics for observer reproducibility were highly significant, with k = 0.95 for analog EEG ischemic changes and 0.85 for density spectral array changes. It is concluded that density spectral array does not reliably detect mild analog EEG pattern changes of cerebral ischemia and is not a reliable substitute for 16-channel analog EEG monitoring of cerebral ischemia during carotid endarterectomy.


Sign in / Sign up

Export Citation Format

Share Document