Prognostication--the lost skill of medicine

Evolution of Making Clinical Predictions in Neurosurgery

Prediction of clinical outcomes is an essential task for every physician. Physicians may base their clinical prediction of an individual patient on their intuition and on scientific material such as studies presenting population risks and studies reporting on risk factors (prognostic factors). A relatively new and more informative approach for making clinical predictions relies on the use of statistical models that simultaneously consider multiple predictors that provide an estimate of the patient's absolute risk of an outcome. There is a growing body of literature in the neurosurgical field reporting on clinical prediction models. These tools have high potential in supporting (not replacing) neurosurgeons with their prediction of a patient's outcome. If used sensibly, these tools pave the way for more informed decision-making with or for individual patients. Patients and their significant others want to know their risk of the anticipated outcome, how it is derived, and the uncertainty associated with it. Learning from these prediction models and communicating the output to others has become an increasingly important skill neurosurgeons have to master. This article describes the evolution of making clinical predictions in neurosurgery, synopsizes key phases for the generation of a useful clinical prediction model, and addresses some considerations when deploying and communicating the results of a prediction model. The paper is illustrated with multiple examples from the neurosurgical literature, including predicting arachnoid cyst rupture, predicting rebleeding in patients suffering from aneurysmal subarachnoid hemorrhage, and predicting survival in glioblastoma patients.

A technical solution to a professional problem: The risk management functions of prognosticators in the context of prognostication post-cardiac arrest

Cardiac arrest (CA) is a major cause of mortality and morbidity globally. Two-thirds of deaths among patients admitted to intensive care units following out-of-hospital CA are due to neurological injury, with most as a consequence of withdrawing life-sustaining treatment, following prognostication of unfavorable neurological outcome. Given the ramifications of prognosis for patient outcome, post-cardiac arrest (P-CA) guidelines stress the importance of minimizing the risk of falsely pessimistic predictions. Although prognosticator use is advocated to this end, 100% accurate prognosticators remain elusive, therefore prognostication P-CA remains pervaded by uncertainty and risk. Bioethical discourse notwithstanding, when located within a wider socio-cultural context, prognostication can be seen to present risk and uncertainty challenges of a professional nature. Such challenges do not, however, subvert the medical profession's moral and ethical prognostication obligation. We interpret prognosticator use as an attempt to manage professional risk presented by prognostication P-CA and demonstrate how through performing "risk work," prognosticators serve professional functions, mediating tension between the professional duty to prognosticate, and risk presented. We draw on sociological analyses of risk and uncertainty, and the professions to explicate these (hitherto less enunciated) professional risk management functions of prognosticators. Accordingly, the use of prognosticators is conceived of as a professional response - a technical/scientific solution to the problem of professional risk, inherent within the P-CA prognostication process.

The clinical frailty scale predicts 1-year mortality in emergency department patients aged 65 years and older

OBJECTIVE

To validate the Clinical Frailty Scale (CFS) for prediction of 1-year all-cause mortality in the emergency department (ED) and compare its performance to the Emergency Severity Index (ESI).

METHODS

Prospective cohort study at the ED of a tertiary care center in Northwestern Switzerland. All patients aged ≥65 years were included from March 18 to May 20, 2019, after informed consent. Frailty status was assessed using CFS, excluding level 9 (palliative). Acuity level was assessed using ESI. Both CFS and ESI were adjusted for age, sex and presenting condition in multivariable logistic regression. Prognostic performance was assessed for discrimination and calibration separately. Estimates were internally validated by Bootstrapping. Restricted mean survival time (RMST) was determined for all levels of CFS.

RESULTS

In the final study population of 2191 patients, 1-year all-cause mortality was 17% (n = 372). RMST values ranged from 219 days for CFS 8 to 365 days for CFS 1. The adjusted CFS model had an area under receiver operating characteristic of 0.767 (95% confidence interval [CI]: 0.741-0.793), compared to 0.703 (95% CI: 0.673-0.732) for the adjusted ESI model.

CONCLUSION

The CFS predicts 1-year all-cause mortality for older ED patients and predicts survival time in a graded manner. The CFS is superior to the ESI when adjusted for age, sex, and presenting condition.

Validation of a Simple Score for Mortality Prediction in a Cohort of Unselected Emergency Patients

BACKGROUND

Prognostication is an important component of medical decision-making. A patients' general prognosis can be difficult to measure. The Simple Prognostic Score (SPS) was designed to include patients' age, mobility, aggregated vital signs, and the treating physician's decision to admit to aid prognostication. Study Aim. Our study aim is to validate the SPS, compare it with the Emergency Severity Index (ESI) regarding its prognostic performance, and test the interrater reliability of the subjective variable of the decision to admit.

METHODS

Over a period of 9 weeks all patients presenting to the ED were included, routinely interviewed, final disposition registered, and followed up for one year. The C-statistics of discrimination was used to compare SPS and ESI predictions of 7-day, 30-day, and 1-year mortality. Youden J Statistics and Odds ratio, using logistical regression, were calculated for the Simple Prognostic Score. In a subset, a chart review was performed by senior physicians for a secondary assessment of the decision to admit. Interrater reliability was calculated using percentages and Cohens Kappa.

RESULTS

Out of 5648 patients, 3272 (57.9%) had a low SPS (i.e., ≤ 1); none of these patients died within 7 days, 2 (0.1%) died within 30 days after presentation and 19 (0.6%) died within a year. The area under the curve for 1-year mortality of the Simple Prognostic Score was 0.848. Secondary analysis of the interrater agreement for the decision to admit was 92%.

CONCLUSION

In a prospective study of unselected ED patients, the Simple Prognostic Score was validated as a reliable predictor of short- and long-term mortality.

Top Ten Tips Palliative Care Clinicians Should Know About Prognostication in Oncology, Dementia, Frailty, and Pulmonary Diseases

Prognostication has been described as "Medicine's Lost Art." Taken with diagnosis and treatment, prognostication is the third leg on which medical care rests. As research leads to additional beneficial treatments for vexing conditions like cancer, dementia, and lung disease, prognostication becomes even more difficult. This article, written by a group of palliative care clinicians with backgrounds in geriatrics, pulmonology, and oncology, aims to offer a useful framework for consideration of prognosis in these conditions. This article will serve as the first in a three-part series on prognostication in adults and children.

A simple prognostic score predicts one-year mortality of alert and calm emergency department patients: A prospective two-center observational study

STUDY OBJECTIVE

To derive and validate a prognostic score to predict 1-year mortality using vital signs, mobility and other variables that are readily available at the bedside at no additional cost.

METHODS

Post hoc analysis of two independent prospective observational studies in two emergency departments, one in Denmark and the other in Switzerland.

PARTICIPANTS

Alert and calm emergency department patients.

MEASUREMENTS

The prediction of mortality from presentation to 365 days by vital signs, mobility and other variables that are readily available at the bedside at no additional cost.

RESULTS

One thousand six hundred and eighteen alert and calm patients were in the Danish cohort and 1331 in the Swiss cohort. Logistic regression identified age >68 years, abnormal vital signs, impaired mobility and the decision to admit as significant predictors of 365-day mortality. A simple prognostic score awarded one point to each of these predictors. Less than two of these predictors were present in 45.6% of patients, and only 0.4% of these patients died within a year. If two or more of these predictors were present, 365-day mortality increased exponentially.

CONCLUSION

Age >68 years, the decision for hospital admission, any vital sign abnormality at presentation and impaired mobility at presentation are equally powerful predictors of 1-year mortality in alert and calm emergency department patients. If validated by others these predictors could be used to discharge patients with confidence since nearly half of these patients had less than two predictors and none of them died within 30 days. However, when two or more predictors were present 365-day mortality increased exponentially.

Factors associated with physicians' predictions of six-month mortality in critically ill patients

Background

Physician's estimates of a patient's prognosis are an important component in shared decision-making. However, the variables influencing physician's judgments are not well understood. We aimed to determine which physician and patient factors are associated with physicians' predictions of critically ill patients' six-month mortality and the accuracy and confidence of these predictions.

Methods

Prospective cohort study evaluating physicians' predictions of six-month mortality. Using univariate and multivariable generalized estimating equations, we assessed the association between baseline physician and patient characteristics with predictions of six-month death, as well as accuracy and confidence of these predictions.

Results

Our cohort was comprised 300 patients and 47 physicians. Physicians were asked to predict if patients would be alive or dead at six months and to report their confidence in these predictions. Physicians predicted that 99 (33%) patients would die. The key factors associated with both the direction and accuracy of prediction were older age of the patient, the presence of malignancy, being in a medical ICU, and higher APACHE III scores. The factors associated with lower confidence included older physician age, being in a medical ICU and higher APACHE III score.

Conclusions

Patient level factors are associated with predictions of mortality at six months. The accuracy and confidence of the predictions are associated with both physician and patients' factors. The influence of these factors should be considered when physicians reflect on how they make predictions for critically ill patients.

The Israeli Terminally Ill Patient law of 2005

Israel, like many other countries, is struggling with numerous bioethical dilemmas due to its cultural and religious diversity. Until recently there was no legal guidance for how to deal with end-of-life issues. However, in 2005 a law was passed regulating the treatment of dying patients. Its most controversial aspect is the distinction it makes between withholding therapy (which is allowed) and withdrawing continuous therapy (which is not allowed). In this formulation, the law attempted to strike a balance between respecting the autonomy of the patient and respecting the sanctity of life. The law respects autonomy by establishing the right of the patient to refuse treatment; it respects the sanctity of life by prohibiting active euthanasia and physician-assisted suicide. However, this compromise was not acceptable to all members of the public advisory body that framed the law. Some argued that there was no moral basis for the distinction between withholding and withdrawing treatment.

The PARIS score can reliably predict 7-day all-cause mortality for both acute medical and surgical patients: an international validation study

BACKGROUND

We believe errors in the risk assessment of acutely ill patients occur because only vital signs without concurrent functional capacity are considered. We, therefore, developed the PARIS risk score based on blood pressure, age, respiratory rate, loss of independence and oxygen saturation.

AIM

Validation of the PARIS score in four independent cohorts from three countries.

METHODS

Retrospective cohort study of acutely ill patients admitted to hospitals in Denmark, Ireland and Uganda. Vital signs and functional capacity (registered as ability to stand or walk or get into bed unaided) was recorded upon arrival. Patients were followed up for 7 days (Denmark and Ireland) or until discharge (Uganda) and mortality recorded. The discriminatory power (ability to identify patients at increased risk) was determined using area under the receiver operating characteristics curve (AUROC) and calibration (precision) using Hosmer-Lemeshow goodness of fit test.

RESULTS

Out of 14 447 patients, 327 (2.3%) died within 7 days: median age was 59 (39-75) years and 7458 (51.8%) were female. Seven-day mortality increased from 0.3% with a score of 0-26.7% with a score of 5. The score's AUROC as 0.833 [95% confidence interval (95% CI) 0.811-0.856], 0.817 (95% CI 0.792-0.842) and 0.894 (95% CI 0.813-0.974) for all patients, medical patients and surgical patients, respectively. However, except for surgical patients, calibration of the score was poor.

CONCLUSION

The PARIS score can identify both high and low risk acutely admitted medical and surgical patients, but calibration was poor for medical patients.

Truth disclosure on prognosis: Is it ethical not to communicate personalised risk of death?

Predicting risk of death based on personalised and objective clinical indicators is an improvement over intuition and clinical judgement. Risk assessment can benefit clinicians by improving prognostic certainty, and truth disclosure helps patients and families by preventing futile management. Some argue that consent should be obtained before a patient is given an estimate of their prognosis as disclosure of bad news can overburden patients. In this article, we argue that it is unethical not to use existing person-specific information to guide diagnosis and shared decision making on management in partnership with well-informed patients. Disclosure of a poor prognosis should be normalised in personalised medicine, performed incrementally and with sensitivity so that it is acceptable to patients, and only occur if patients want to know it. However, a requirement of consent for truth disclosure should not be mandatory. Despite some level of imprecision, personalised risk estimations can be used to tailor management to the patient's informed wishes and ensure that healthcare providers and families are acting ethically in the patient's best interest.

Am I too old for this, Doctor? Using population life expectancy to guide clinical decision-making

OBJECTIVE

Prognostication is important in clinical decision-making, especially for older people. The aim was to present estimates of life expectancy for older people in New Zealand.

METHODS

Statistics New Zealand age-sex-specific death rates were used to derive quartiles of expected years of life remaining in people aged over 65 years.

RESULTS

Given current patterns and trends in New Zealand death rates, 50% of women reaching age 80 years in 2016 can expect to live at least another 10.5 years, 25% will live over 14.7 years, and 25% will die within 6.2 years. Comparable results for men reaching age 80 years in 2016 are 8.5 years, 12.7 years and 4.6 years, respectively. Of those reaching age 90 years in 2016, median expected years of life left is 4.2 years for women and 3.4 years for men.

CONCLUSION

Demographic norms are useful as a guide when specific predictive tools are unavailable.

Death is not the only healthcare outcome important to patients

Unfortunately throughout history there have been wide variations in the way death has been handled by the medical profession in different times and places, and even today within the same hospital there are big difference between what doctors say they do and what actually happens. It is not currently possible to determine when severely ill patients become irreversibly unsalvageable and when attempts at resuscitation after death are futile. Without this knowledge it is impossible to honestly advise patients and their loved ones. There is little data available to show what proportions of patients are less sick or feel better on discharge from hospital than they were on admission and no robust systems for predicting such outcomes. Death is not the only healthcare outcome important to patients. Regaining or preserving health is the ultimate goal for patients, yet most hospital outcomes are reported only in terms of mortality. Developing models that predict a good clinical outcome may be more clinically useful than those that predict death. Patients are more likely to want to know their chances of getting better than their chances of dying. Also expressing treatment options in terms of its benefits (i.e. the chance of getting better) versus the risks (i.e. the chances the treatment will kill you) may be far more acceptable to patients than providing their risks versus the chance that they are going to die anyway.

Nurses and physicians in a medical admission unit can accurately predict mortality of acutely admitted patients: a prospective cohort study

BACKGROUND

There exist several risk stratification systems for predicting mortality of emergency patients. However, some are complex in clinical use and others have been developed using suboptimal methodology. The objective was to evaluate the capability of the staff at a medical admission unit (MAU) to use clinical intuition to predict in-hospital mortality of acutely admitted patients.

METHODS

This is an observational prospective cohort study of adult patients (15 years or older) admitted to a MAU at a regional teaching hospital. The nursing staff and physicians predicted in-hospital mortality upon the patients' arrival. We calculated discriminatory power as the area under the receiver-operating-characteristic curve (AUROC) and accuracy of prediction (calibration) by Hosmer-Lemeshow goodness-of-fit test.

RESULTS

We had a total of 2,848 admissions (2,463 patients). 89 (3.1%) died while admitted. The nursing staff assessed 2,404 admissions and predicted mortality in 1,820 (63.9%). AUROC was 0.823 (95% CI: 0.762-0.884) and calibration poor. Physicians assessed 738 admissions and predicted mortality in 734 (25.8% of all admissions). AUROC was 0.761 (95% CI: 0.657-0.864) and calibration poor. AUROC and calibration increased with experience. When nursing staff and physicians were in agreement (±5%), discriminatory power was very high, 0.898 (95% CI: 0.773-1.000), and calibration almost perfect. Combining an objective risk prediction score with staff predictions added very little.

CONCLUSIONS

Using only clinical intuition, staff in a medical admission unit has a good ability to identify patients at increased risk of dying while admitted. When nursing staff and physicians agreed on their prediction, discriminatory power and calibration were excellent.

Physicians' perceptions of the value of prognostic models: the benefits and risks of prognostic confidence

BACKGROUND

The communication of prognosis in end-of-life (EOL) care is a challenging task that is limited by prognostic uncertainty and physicians' lack of confidence in their prognostic estimates. Clinical prediction models (CPMs) are increasingly common evidence-based tools that may mitigate these problems and facilitate the communication and use of prognostic information in EOL care; however, little is known about physicians' perceptions of the value of these tools.

OBJECTIVE

To explore physicians' perceptions of the value of CPMs in EOL care.

DESIGN

Qualitative study using semi-structured individual interviews which were analysed using a constant comparative method.

SETTING AND PARTICIPANTS

Convenience sample of 17 attending physicians representing five different medical specialties at a single large tertiary care medical centre.

RESULTS

Physicians perceived CPMs as having three main benefits in EOL care: (i) enhancing their prognostic confidence; (ii) increasing their prognostic authority; and (iii) enabling patient persuasion in circumstances of low prognostic and therapeutic uncertainty. However, physicians also perceived CPMs as having potential risks, which include producing emotional distress in patients and promoting prognostic overconfidence in EOL care.

DISCUSSION AND CONCLUSIONS

Physicians perceive CPMs as a potentially valuable means of increasing their prognostic confidence, communication and explicit use of prognostic information in EOL care. However, physicians' perceptions of CPMs also indicate a need to establish broad and consistent implementation processes to engage patients in shared decision making in EOL care, to effectively communicate uncertainty in prognostic information and to help both patients and physicians manage uncertainty in EOL care decisions.

Comparison of clinician-predicted to measured low vision outcomes

PURPOSE

To compare low-vision rehabilitation (LVR) clinicians' predictions of the probability of success of LVR with patients' self-reported outcomes after provision of usual outpatient LVR services and to determine if patients' traits influence clinician ratings.

METHODS

The Activity Inventory (AI), a self-report visual function questionnaire, was administered pre-and post-LVR to 316 low-vision patients served by 28 LVR centers that participated in a collaborative observational study. The physical component of the Short Form-36, Geriatric Depression Scale, and Telephone Interview for Cognitive Status were also administered pre-LVR to measure physical capability, depression, and cognitive status. After patient evaluation, 38 LVR clinicians estimated the probability of outcome success (POS) using their own criteria. The POS ratings and change in functional ability were used to assess the effects of patients' baseline traits on predicted outcomes.

RESULTS

A regression analysis with a hierarchical random-effects model showed no relationship between LVR physician POS estimates and AI-based outcomes. In another analysis, kappa statistics were calculated to determine the probability of agreement between POS and AI-based outcomes for different outcome criteria. Across all comparisons, none of the kappa values were significantly different from 0, which indicates that the rate of agreement is equivalent to chance. In an exploratory analysis, hierarchical mixed-effects regression models show that POS ratings are associated with information about the patient's cognitive functioning and the combination of visual acuity and functional ability, as opposed to visual acuity or functional ability alone.

CONCLUSIONS

Clinicians' predictions of LVR outcomes seem to be influenced by knowledge of patients' cognitive functioning and the combination of visual acuity and functional ability-information clinicians acquire from the patient's history and examination. However, clinicians' predictions do not agree with observed changes in functional ability from the patient's perspective; they are no better than chance.

Identifying admitted patients at risk of dying: a prospective observational validation of four biochemical scoring systems

OBJECTIVES

Risk assessment is an important part of emergency patient care. Risk assessment tools based on biochemical data have the advantage that calculation can be automated and results can be easily provided. However, to be used clinically, existing tools have to be validated by independent researchers. This study involved an independent external validation of four risk stratification systems predicting death that rely primarily on biochemical variables.

DESIGN

Prospective observational study.

SETTING

The medical admission unit at a regional teaching hospital in Denmark.

PARTICIPANTS

Of 5894 adult (age 15 or above) acutely admitted medical patients, 205 (3.5%) died during admission and 46 died (0.8%) within one calendar day.

INTERVENTIONS

None.

MAIN OUTCOME MEASURES

The main outcome measure was the ability to identify patients at an increased risk of dying (discriminatory power) as area under the receiver-operating characteristic curve (AUROC) and the accuracy of the predicted probability (calibration) using the Hosmer-Lemeshow goodness-of-fit test. The endpoint was all-cause mortality, defined in accordance with the original manuscripts.

RESULTS

Using the original coefficients, all four systems were excellent at identifying patients at increased risk (discriminatory power, AUROC ≥0.80). The accuracy was poor (we could assess calibration for two systems, which failed). After recalculation of the coefficients, two systems had improved discriminatory power and two remained unchanged. Calibration failed for one system in the validation cohort.

CONCLUSIONS

Four biochemical risk stratification systems can risk-stratify the acutely admitted medical patients for mortality with excellent discriminatory power. We could improve the models for use in our setting by recalculating the risk coefficient for the chosen variables.

The average lifespan of patients discharged from hospital with heart failure

BACKGROUND

There are no life-tables quantifying the average life-spans of post-hospitalized heart failure populations across various strata of risk.

OBJECTIVE

To quantify the life-expectancies (i.e., average life-spans) of heart failure patients at the time of hospital discharge according to age, gender, predictive 30-day mortality heart failure risk index, and comorbidity burden.

DESIGN

Population-based retrospective cohort study.

SETTING

Ontario, Canada.

PATIENTS

7,865 heart failure patients discharged from Ontario hospitals between 1999 and 2000.

MEASUREMENTS

Data were obtained from the Enhanced Feedback for Effective Cardiac Treatment EFFECT provincial quality improvement initiative. All patients were linked to administrative data, and tracked longitudinally until March 31, 2010. Detailed clinical variables were obtained from medical chart abstraction, and death data were obtained from vital statistics. Average life-spans were calculated using Cox Proportion Hazards models in conjunction with the Declining Exponential Approximation of Life Expectancy (D.E.A.L.E) method to extrapolate life-expectancy, adjusting for age, gender, predicted 30-day mortality, left ventricular function and comorbidity, and was reported according to key prognostic risk-strata.

RESULTS

The average life-span of the cohort was 5.5 years (STD +/- 10.0) ranging from 19.5 years for low-risk women of less than 50 years old to 2.9 years for high-risk octogenarian males. Average life-spans were lower by 0.13 years among patients with impaired as compared with preserved left ventricular function, and by approximately one year among patients with three or more as compared with no concomitant comorbidities. In total, 17.4 % and 27 % of patients had died within 6 months and 1 year respectively, despite having predicted life-spans exceeding one-year.

LIMITATIONS

Data regarding changes in patient clinical status over time were unavailable.

CONCLUSIONS

The development of risk-adjusted life-tables for heart failure populations is feasible and mirrored those with advanced malignant diseases. Average life span varied widely across clinical risk strata, and may be less accurate among those at or near their end of life.

Accuracy of Neurovascular Fellows' Prognostication of Outcome After Subarachnoid Hemorrhage

Background and Purpose—

The purpose of this study was to determine the accuracy and optimal timing of physician prognostication in patients with subarachnoid hemorrhage, a prototypical neurological disease characterized by variable outcomes and frequent disability.

Methods—

From October 2009 to April 2010, treating neurologists at a tertiary care academic medical center made daily predictions of the modified Rankin Scale at 6 months for consecutive patients with subarachnoid hemorrhage. Actual functional outcomes at 6 months were determined by phone interview and dichotomized into good (modified Rankin Scale 0–2) and poor (modified Rankin Scale 3–6) outcomes. Descriptive statistics were used to assess the accuracy of prognostications. Multiple logistic regression and generalized estimating equations were used to assess changes in prognostication accuracy over time and the relationship between prognostication accuracy and clinical factors.

Results—

Physicians made 648 prognostications for 66 patients. Overall accuracy ranged from 78% to 88%. Among patients predicted to have a good outcome, 81% (95% CI, 71%–92%) actually had a good outcome, whereas 88% (95% CI, 77%–99%) of patients predicted to do poorly had poor outcomes. No significant trends were seen in prognostication accuracy over time during the hospital course ( P =0.72). Increasing age, infection, mechanical ventilation, hydrocephalus, and seizures all significantly worsened physician accuracy.

Conclusions—

Neurologists were generally but not perfectly accurate in their prognostications of functional outcomes. The accuracy of prognoses did not correlate with the hospital day on which they were made but was affected by clinical factors that can cloud the neurological examination.

Flexible modeling improves assessment of prognostic value of C-reactive protein in advanced non-small cell lung cancer

Background:

C-reactive protein (CRP) is gaining credibility as a prognostic factor in different cancers. Cox's proportional hazard (PH) model is usually used to assess prognostic factors. However, this model imposes a priori assumptions, which are rarely tested, that (1) the hazard ratio associated with each prognostic factor remains constant across the follow-up (PH assumption) and (2) the relationship between a continuous predictor and the logarithm of the mortality hazard is linear (linearity assumption).

Methods:

We tested these two assumptions of the Cox's PH model for CRP, using a flexible statistical model, while adjusting for other known prognostic factors, in a cohort of 269 patients newly diagnosed with non-small cell lung cancer (NSCLC).

Results:

In the Cox's PH model, high CRP increased the risk of death (HR=1.11 per each doubling of CRP value, 95% CI: 1.03–1.20, P=0.008). However, both the PH assumption (P=0.033) and the linearity assumption (P=0.015) were rejected for CRP, measured at the initiation of chemotherapy, which kept its prognostic value for approximately 18 months.

Conclusion:

Our analysis shows that flexible modeling provides new insights regarding the value of CRP as a prognostic factor in NSCLC and that Cox's PH model underestimates early risks associated with high CRP.

Ability of physiological parameters versus clinical categories to predict mortality on admission to an internal medicine ward

BACKGROUND

The prediction of mortality in internal medicine departments may help in taking diagnostic and therapeutic decisions. We analyzed the usefulness of two mortality prediction models, one physiological and the other mainly clinical, and determined whether one approach is better than the other to predict mortality at admission.

METHODS

This is a prospective observational cohort study in patients admitted to an acute internal medicine ward in a tertiary care, urban, university teaching hospital in Spain. Five hundred consecutive patients either electively admitted or coming from the emergency department from May to December 2008 were analyzed. Medical history, physical examination and routine clinical laboratory tests were performed on admission. At discharge, diagnosis and dead or survived status was recorded. Logistic regression analyses were used to test variables that emerged as independent predictors of mortality. The area under the curve was used to determine which model best predicted mortality.

RESULTS

Mortality in the ward was 13.0%. Age, chronic respiratory failure, creatinine, mean arterial pressure, respiratory rate and Glasgow coma scale independently predicted mortality. ROC curves showed that the physiological model was superior to the clinical model, but differences were not statistically significant. The predictive capacity improved when the two models were combined but the improvement was not significant.

CONCLUSIONS

Both models are satisfactory predictors of in-hospital mortality for management purposes but neither proved to be a useful tool for individual predictions. Complementary approaches need to be considered.

Hospital medicine (Part 2): what would improve acute hospital care?

There are so many obvious delays and inefficiencies in our traditional system of acute hospital care; it is clear that if outcomes are to be improved prompt accurate assessment immediately followed by competent and efficient treatment is essential. Early warning scores (EWS) help detect acutely ill patients who are seriously ill and likely to deteriorate. However, it is not known if any EWS has universal applicability to all patient populations. The benefit of Rapid Response Systems (RRS) such as Medical Emergency Teams has yet to be proven, possibly because doctors and nurses are reluctant to call the RRS for help. Reconfiguration of care delivery in an Acute Medical Assessment Unit has been suggested as a "proactive" alternative to the "reactive" approach of RRS. This method ensures every patient is in an appropriate and safe environment from the moment of first contact with the hospital. Further research is needed into what interventions are most effective in preventing the deterioration and/or resuscitating seriously ill patients. Although physicians expert in hospital care decrease the cost and length of hospitalization without compromising outcomes hospital care will continue to be both expensive and potentially dangerous.