Risk of fragility fractures in older women appears to be under-recognized and under treated. Analysis of a national sample of older US women reveals that over 5 million are at high risk of fracture; only one third of these report being told they have osteoporosis and one quarter are receiving appropriate treatment.

To assess the impact of the aging population on the occurrence of fragility fractures, we examined hospital discharges for hip fracture among U.S. women and men aged 45 years and older from 1993 through to 2003. The number of hospitalizations declined by 5%, and age-adjusted rates fell by over 20% for both women and men during this period.

OBJECTIVE:To develop models in the Mortality Probability Model (MPM II) system to estimate the probability of hospital mortality at 48 and 72 hrs in the intensive care unit (ICU), and to test whether the 24-hr Mortality Probability Model (MPM24), developed for use at 24 hrs in the ICU, can be used on a daily basis beyond 24 hrs.DESIGN:A prospective, multicenter study to develop and validate models, using a cohort of consecutive admissions.SETTING:Six adult medical and surgical ICUs in Massachusetts and New York adjusted to reflect 137 ICUs in 12 countries.PATIENTS:Consecutive admissions (n = 6,290) to the Massachusetts/New York ICUs were studied. Of these patients, 3,023 and 2,233 patients remained in the ICU and had complete data at 48 and 72 hrs, respectively. Patients < 18 yrs of age, burn patients, coronary care patients, and cardiac surgical patients were excluded.OUTCOME MEASURE:Vital status at the time of hospital discharge.RESULTS:The models consist of five variables measured at the time of ICU admission and eight variables ascertained at 24-hr intervals. The 24-hr model demonstrated poor calibration and discrimination at 48 and 72 hrs. The newly developed 48- and 72-hr models--MPM48 and MPM72--contain the same 13 variables and coefficients as the MPM24. The models differ only in their constant terms, which increase in a manner that reflects the increasing probability of mortality with increasing length of stay in the ICU. These constant terms were adjusted by a factor determined from the relationship between the data from the six Massachusetts and New York ICUs and a more extensive data set, from which the ICU admission Mortality Probability Model (MPM0) and MPM24 were developed. This latter data set was assembled from ICUs in 12 countries. The MPM48 and MPM72 calibrated and discriminated well, based on goodness-of-fit tests and area under the receiver operating characteristic curve.CONCLUSIONS:Models developed for use among ICU patients at one time period are not transferable without modification to other time periods. The MPM48 and MPM72 calibrated well to their respective time periods, and they are intended for use at specific points in time. The increasing constant terms and associated increase in the probability of hospital mortality exemplify a common clinical adage that if a patient's clinical profile stays the same, he or she is actually getting worse.

OBJECTIVE:To revise and update models in the Mortality Probability Model (MPM II) system to estimate the probability of hospital mortality among 19,124 intensive care unit (ICU) patients that can be used for quality assessment within and among ICUs.DESIGN AND SETTING:Models developed and validated on consecutive admissions to adult medical and surgical ICUs in 12 countries.PATIENTS:A total of 12,610 patients for model development, 6514 patients for model validation. Patients younger than 18 years and burn, coronary care, and cardiac surgery patients were excluded.OUTCOME MEASURE:Vital status at hospital discharge.RESULTS:The admission model, MPM0, contains 15 readily obtainable variables. In developmental and validation samples it calibrated well (goodness-of-fit tests: P = .623 and P = .327, respectively, where a high P value represents good fit between observed and expected values) and discriminated well (area under the receiver operating characteristic curve = 0.837 and 0.824, respectively). The 24-hour model, MPM24 (developed on 10,357 patients still in the ICU at 24 hours), contains five of the admission variables and eight additional variables easily ascertained at 24 hours. It also calibrated well (P = .764 and P = .231 in the developmental and validation samples, respectively) and discriminated well (area under the receiver operating characteristic curve = 0.844 and 0.836 in the developmental and validation samples, respectively).CONCLUSIONS:Among severity systems for intensive care patients, the MPM0 is the only model available for use at ICU admission. Both MPM0 and MPM24 are useful research tools and provide important clinical information when used alone or together.

Factors related to hospital resource use by intensive care unit (ICU) patients, including severity of illness at admission and intensity of therapy during the first 24 ICU hours were explored in this study. Analysis was based on 2,749 patients admitted to the general medical-surgical ICU at Baystate Medical Center, Springfield, Massachusetts, between February 1,1983 and January 10, 1985. Resource use was indexed by hospital length of stay (LOS) adjusted for differences between ICU and other hospital days. Severity of illness was measured by the Mortality Prediction Model (MPM0), a validated predictor of outcome but not previously used to analyze resource consumption. Intensity of therapy was measured using the Therapeutic Intervention Scoring System (TISS). The 10% of patients with longest ICU stays were significantly different from the other 90% with respect to previous ICU use, MPM probability, and TISS score. Variability in resource use was analyzed using four diagnosis-related groups (DRGs) accounting for large numbers of ICU patients. The relationship between severity of illness and resource use was nonlinear: as severity increased from low levels, resource use increased at a decreasing rate, reached a plateau, and eventually declined. Within each DRG, MPM0 explained a statistically significant percentage of the variability in resource use.

Six months after hospital discharge, we followed up 1545 patients who had received care in the general medical-surgical intensive care unit (ICU) of a tertiary care hospital. Vital status could not be ascertained for 200 of these patients. Of the 1345 former ICU patients for whom a determination of vital status could be made, 1261 (94%) were alive and 84 (6%) had died. Of those known to be living, 887 (70%) responded to a questionnaire regarding employment, functional, and social status. A large proportion of survivors less than 40 years of age had returned to work. Younger patients admitted to the hospital for elective surgery reported as much compromise of physical and psychological activity as did older patients admitted for emergency reasons. Older survivors reported an increase of interaction with family members and a decrease of social interaction with those other than family.

Statisticians are being asked with increasing frequency to develop models for occurrences in medical environments. Until recently, only subjective models were available to predict mortality for patients in an intensive care unit (ICU). These models were based on variables and associated weights determined by panels of medical "experts." This article shows how multiple logistic regression (MLR) can be used to develop an objective model for prediction of hospital mortality among ICU patients. An MLR model to be applied when a patient is admitted to the ICU was developed on 737 ICU patients. The final model is based on the following variables: presence of coma or deep stupor at admission, emergency admission, cancer part of present problem, probable infection, cardiopulmonary resuscitation (CPR) prior to admission, age, and systolic blood pressure at admission. To validate this model, a new cohort of 1,997 consecutive ICU patients was entered into the study. Information was collected for the variables in the MLR model and, in addition, the variables necessary to evaluate the "subjective" models. The admission mortality prediction model [MPM0(CPR)] was validated on this new cohort of patients using goodness-of-fit tests. It was found that this model had excellent fit (p = .74). In addition, the overall correct classification for this model in the validation data set was 86.1%. The predictive values for dying and surviving, sensitivity, and specificity were 71.3%, 88.5%, 50.2%, and 95.0%, respectively. The direct comparison of MPM0(CPR) and the subjective systems based on the new cohort demonstrated that although all methods considered demonstrated comparable sensitivity, specificity, predictive values, and total correct classification rates, goodness-of-fit tests suggest that the probabilities of hospital mortality as produced by the statistical model best fit the observed mortality experience. One of the commonly used subjective models tended to overestimate the probabilities of hospital mortality, and the other tended to underestimate these probabilities. To be useful, a severity index should be based on simple calculations using readily available data, and be independent of medical treatment in the ICU. Given the widespread availability of microcomputers, the newly developed statistical model seems to satisfy these criteria. Key Words: Multiple logistic regressionMaximum likelihood estimateMortality prediction modelGoodness-of-fit testSensitivitySpecificityRelative risk

Teres, Daniel; Rapoport, John; Lemeshow, Stanley; Haber, Russell; Gage, Robert W.; Avrunin, Jill Spitz

Estimating prognosis is potentially useful as a measure of ICU performance and as a guide for the clinical care of individual patients. In this study, mortality prediction models (MPMs) for patients in an adult general medical-surgical ICU were derived from data gathered at ICU admission and after 24 and 48 h of ICU care. A predictive model was developed which incorporated a sequence of probabilities collected over time in the ICU. The results of this study suggest that using serial observations may enhance substantially the usefulness of the MPM as a vehicle for helping families anticipate the patients' likely outcome.

Critical Care Services, Baystate Medical Center, Springfield, Massachusetts 01199 and Division of Public Health, University of Massachusetts, Amherst, Massachusetts 01003

This paper presents results of the first study explicitly designed to compare three methods for predicting hospital mortality of ICU patients: the Acute Physiology Score (APS), the Simplified Acute Physiology Score (SAPS), and the Mortality Prediction Model (MPM). With respect to sensitivity, specificity, and total correct classification rates, these methods performed comparably on a cohort of 1,997 consecutive ICU admissions. In these patients from a single hospital, the APS overestimated and the SAPS underestimated the probability of hospital mortality. The MPM probabilities most closely matched the observed outcomes. Each method holds considerable promise for assessing the severity of illness of critically ill patients. The MPM should be particularly useful for comparing ICU performance, since it is independent of ICU treatment and can be calculated at the time a patient is admitted.

We tested recently developed admission and 24-h models of hospital mortality on 1,997 consecutive admissions to a general medical/surgical ICU. This study population was independent of the group used to develop the models. The admission prediction model estimated each patient's probability of hospital mortality based on seven routinely collected admission variables. The 24-h model utilized seven variables routinely available at 24 h in the ICU. The admission model accurately described the mortality experience of the new cohort, while the 24-h model did not. Advantages of the admission model are that it is evaluable at the time of ICU admission, is independent of ICU treatment, and can be used to stratify patients by severity of illness, thereby making ICU comparisons possible. Its excellent goodness-of-fit, correct classification rate, sensitivity, and specificity suggest that this model is now ready for multihospital testing.

Baystate Medical Center, Springfield, MA 01199 and School of Health Sciences, University of Massachusetts, Amherst, MA 01003.

School of Health Sciences, University of Massachusetts, Amherst, MA 01003 and Baystate Medical Center, Springfield, MA 01199.

This paper illustrates how a microcomputer spreadsheet package can be used by epidemiologists to facilitate the computation of multiple logistic regression (MLR) probabilities, as well as odds ratios and associated confidence intervals, given the coefficients of the MLR model. By formatting a spreadsheet, data entry is greatly simplified, and computations are accomplished without any arithmetic manipulations on the part of the user. This approach makes it feasible for clerical support staff to assist in the computation of seemingly complex expressions. The increasing availability of microcomputers in clinical and research settings suggests that numerous analytic applications are amenable to this approach, thereby decreasing reliance on mainframe computers and desk-top calculators.

Data at ICU admission and after 24 h in the ICU were collected on 755 patients, to derive multiple logistic regression models for predicting hospital mortality. The derived models contained relatively few and easily obtained variables. The weight associated with each variable was determined objectively. There were seven admission variables, none of which were treatment dependent, and seven 24-h variables reflecting treatments and patients' conditions in the ICU. Predicted outcomes using these two models were closely correlated with actual outcome. Theoretically, a predictive model would be useful to physicians for triage decisions as well as determining aggressiveness of care through discussions with families, determining utilization of ICU facilities, and objectively comparing different ICUs. This research represents an initial attempt to develop models that are not based on subjectively determined weights.

Five hundred fifty-eight patients admitted to a general/medical surgical intensive care unit were studied 2 years after hospital discharge to determine whether they were still alive, were able to perform daily activities, and had returned to work. The overall 2-year survivorship (hospital and long-term) was 63.5%. Two-year survival was considerably lower for patients with certain condition or treatment characteristics than for others. This ranged from 14% 2-year survival for patients with 48 or more hours of coma to 82.2% for patients with no condition or treatment characteristics recorded. Once a patient was discharged alive, the 2-year cumulative survival of surgical ICU patients (84.6%) was significantly better than that of medical ICU patients (76.5%). Among ICU survivors responding to a follow-up survey, 85% were able to perform daily activities, but only 66% were working. Of the 44 patients experiencing a change in ability to perform daily activities at time of follow-up compared with pre-ICU admission, functional status of 34 (77%) improved, while 10 (23%) got worse. By comparison, of the 45 patients experiencing a change in working status, only 7 patients (16%) who did not work prior to ICU admission had returned to work, whereas the remaining 38 patients (84%) who worked prior to ICU admission were not working at time of follow-up study.

The effect of the duration of hypothermic (T = 15°C) potassium cardioplegic arrest and ischemia on the heart was determined by measuring the response of the isolated in situ pig heart to 180 min of perfusion (n = 12) to provide appropriate control values for the study of 30 (n = 25) or 120 (n = 27) min of ischemia, followed by 60 min of reperfusion. In some of these animals, myocardial tissue samples were obtained for measurement of adenosine triphosphate (ATP) and creatine phosphate (CP), (6 in the perfusion group, 7 in the 30 min of ischemia and 60 min of reperfusion group and 15 in the 120 min of ischemia and 60 min of reperfusion group). In the remaining animals, measurements of either left ventricular performance (LVP), myocardial oxygen metabolism (MV̇O2) or plasma creatine kinase (CK) were obtained (6 in the prolonged perfusion group, 12 in the 120 min of ischemia and 60 min of reperfusion group, [6 LVP and MV̇O2 and 6 CK] and 18 in the 30 min of ischemia and 60 min of reperfusion group [13 LVP, 17 MV̇O2 and 6 CK]). During prolonged perfusion, left ventricular performance, expressed as developed pressure, ΔP, fell from an initial value of 175 ± 36 to 128 ± 19 mm Hg at 30 min of perfusion, followed by a more gradual decline to a final value of 113 ± 8 mm Hg at 180 min of perfusion. These decreases were not significantly lower than the initial value. The percentage of myocardial extraction declined in a similar manner, but coronary blood flow was constant over this interval. The primary effect of 30 or 120 min of ischemia was to reduce left ventricular developed pressure, ΔP, during reperfusion to more than 70% of the corresponding value in the control group (these differences were statistically significant) which suggests that prolonging the period of ischemia did not cause further deterioration of cardiac performance. The plasma concentration of CK rose in the control group of hearts subjected to prolonged perfusion from an initial value of 35 ± 6 to a final value of 59 ± 8 IU/liter (P < 0.05). While plasma CK increased during reperfusion in both ischemia/ reperfusion groups, these values were not significantly higher from prearrest values. Thus hypothermic cardioplegic ischemia of this duration did not appear to result in tissue necrosis, but there was a significant reduction in left ventricular performance which was independent of the duration of ischemia between the limits of 30 and 120 min.

Dinitrotoluene (DNT) and toluene diamine (TDA) are intermediates in the production of toluene diisocyanate and polyurethane plastics. Some reproductive effects in rodents have been reported; and the National Institute for Occupational Safety and Health reported probable reproductive toxic effect to humans after a preliminary survey. Accordingly, 84 workers exposed to DNT/TDA (classified by intensity and recency of exposure) and 119 nonexposed workers were studied at Olin's chemical complex at Lake Charles, La. Each worker was the subject of a physician's urogenital examination, a reproductive and fertility questionnaire, an estimation of testicular volume, an assessment of serum follicle-stimulating hormone, and an analysis of semen for sperm count and morphology. No differences were found between the exposed and control groups among any of these variables. Although both TDA and DNT are readily absorbed (percutaneous, inhalation, ingestion), they did not present detectable reproductive hazard to the workers.