A chest trauma scoring system

to predict outcomes
Jennifer Chen, MD,a Elan Jeremitsky, MD,b Frances Philp, MS,a William Fry, MD,c and
R. Stephen Smith, MD,c Pittsburgh, PA, Lowell, MA, and Columbia, SC

Background. Rib fractures (RIBFX) are a common injury and are associated with substantial morbidity
and mortality. Using a previously published RIBFX scoring system, we sought to validate the system by
applying it to a larger patient population. We hypothesized that the RIBFX scoring system reliably
predicts morbidity and mortality in patients with chest wall injury at the time of initial evaluation.
Methods. A 3-year, registry-based, retrospective study involving 1,361 trauma patients was performed.
Patients were divided into two groups with a Chest Trauma Score (CTS) < 5 and $5 (n = 724 and
637, respectively). Each cohort was analyzed for specific outcomes (mortality, pneumonia, acute respiratory failure). CTS was defined by age, severity of pulmonary contusion, number of RIBFX, and the
presence of bilateral RIBFX with a maximum score of 12. Receiver operating characteristics were used to
determine the use of CTS $5 cut point.
Results. Patients with a CTS of 5 or more were (P # .05) older (61 vs 50 years), had greater Injury
Severity Scores (21.6 vs 16.2), and had a greater prevalence of pneumonia (10.1 vs 3.5%), tracheostomy (7.4 vs 2.9%), and mortality (9.0 vs 2.2%). Patients with CTS $ 5 had nearly 4-fold
increased odds of mortality (odds ratio 3.99, 95% confidence interval 1.92–8.31, P = .001) compared
with those who had CTS < 5.
Conclusion. A CTS of at least 5 is associated with worse patient outcomes. Increased vigilance is needed
with trauma patients who present with RIBFX and a CTS $ 5 at initial presentation. This simple
RIBFX scoring system may improve early identification of vulnerable patients and expedite therapeutic
interventions. (Surgery 2014;156:988-94.)
From the Division of Trauma, Surgical Critical Care and Acute Care Surgery, Department of Surgery,a Allegheny
General Hospital, Pittsburgh, PA; Department of General Surgery,b Lowell General Hospital, Lowell, MA; and
Department of Trauma and Critical Care,c University of South Carolina School of Medicine, Columbia, SC

RIB FRACTURES (RIBFX) are a common problem in
blunt trauma, with 10% of injured patients sustaining RIBFX.1 Factors associated with increased
morbidity and mortality from RIBFX are varied
and include age, number of RIBFX, and the development of pneumonia. Pre-existing conditions
such as congestive heart failure and renal and liver
dysfunction have also been found to be associated
with poorer outcomes.2,3 Interventions to implement protocols for elderly trauma victims with
RIBFX reported improved outcomes and
decreased hospital length of stay.4 Several chest
trauma scoring systems have been published to
Presented at the Central Surgical Association 2014 Annual
Meeting in Indianapolis, IN, March 8, 2014.
Accepted for publication June 23, 2014.
Reprint requests: Jennifer Chen, MD, Allegheny General Hospital, 320 E. North Ave., Pittsburgh, PA 15212. E-mail: jchen@
0039-6060/$ - see front matter
Ó 2014 Elsevier Inc. All rights reserved.


identify at risk patients, but these systems have
been evaluated in small volume patient cohorts.5
The rationale for this study is to develop and eventually to use a chest trauma scoring system that is
easy to calculate by a medical provider so medical
decisions and or interventions can be made in an
expedited fashion.
This work uses a RIBFX scoring system that was
previously published by Pressley et al,6 and we
sought to validate that chest trauma scoring system
in a larger dataset. The Chest Trauma Score (CTS)
was derived from several factors identified previously to be associated with worse outcomes,
including age, number of RIBFX, pulmonary contusions, and bilaterality of injury.6 We hypothesized
that a CTS system may be used to predict outcomes
in blunt trauma patients with RIBFX.
With institutional review board approval, we used
trauma registry data to identify 1,361 patients with
blunt torso trauma from a single busy Level 1 Trauma
Center during a 3-year period (2009–2011). A CTS

and tracheostomy procedure. which was similar to the previous study by Pressley et al6. and a positive sputum culture was also used to define the presence of pneumonia.2%). Stata SE 12 (College Station. Number 4 was calculated for each of the individual cohorts.001. and VII). The registry codes are assigned to the definitions to enter these complications into the registry. Specific patient outcomes were examined for acute respiratory failure (ARF). 3. or culture. 3–5 = 2. acute pneumonia and respiratory failure). and ARF). Our cohort was then divided into two groups defined by having a CTS $ 5.Chen et al 989 Surgery Volume 156. 45–65 = 2. pneumonia. had greater mortality (9. KaplanMeier survival analysis and multivariable logistic analysis were performed to determine the association of CTS with mortality. All three separate outcomes by ISS and CTS.4%. TX) was used for analysis. Chest scoring system Age score <45 y 45–65 y >65 y Pulmonary contusion score None Unilateral minor Bilateral minor Unilateral major Bilateral major Rib score <3 RIBFX 3–5 RIBFX >5 RIBFX Bilateral RIBFX No Yes 1 2 3 0 1 2 3 4 1 2 3 0 2 RIBFX. When multivariable logistic regression was used. Univariate analysis was conducted using 2 sample t test or Wilcoxon rank sum test for continuous Table I. All tests were two-sided. however. ARF). Rib fracture. bilateral major = 4). Computed tomography (CT) images were reviewed by the authors (W. patients with a CTS $ 5 were older (61. greater rate of hospital complications (ie. Figure 1 illustrates the Kaplan Meier survival curve for patients with CTS $ 5 and < 5 over the course of hospitalization. bilateral minor = 2. and greater duration of stay (Table IV). pneumonia. Table III illustrates the % mortality by CTS. The CTS is composed of four different components with a point system assigned: age (<45 years = 1.C. were not different by ROC test of equality. Radiographic documentation of a pulmonary infiltrate or consolidation. pneumonia. The CTS ranges from 2 to 12 (Table I). the CTS appeared to have leas area under the curve for mortality and pneumonia but greater area under the curve for ARF. Injury Severity Score (ISS). Compared with ISS. and ARF as separate outcomes.F.6 vs 50.1).361 patients was 5. Compared with patients who had a CTS < 5. Specific pulmonary complications. duration of stay. Standard definitions. RESULTS Overall mortality for the cohort of 1. ARF was defined as need for ventilator support after a period (>48 hours) of nonassisted breathing or requirement for reintubation. sepsis. and 4) were calculated for the different outcomes separately (mortality. A CTS $ 5 was associated with greater mortality (P < .) to assign pulmonary contusion scores and create the individual CTS for each patient in the cohort. In contrast. ROC curves (Figs 2. Hospital complications included deep-vein thrombosis and pulmonary embolism. and the presence of bilateral RIBFX = 2. Retrospective trauma registry data were collected for each patient. development of pneumonia. however. from the Pennsylvania Trauma Systems Foundation’s 2011 Pennsylvania Trauma Outcomes Study Manual. >65 = 3). unilateral major = 3. including: demographics. unilateral minor = 1. Table II demonstrates the different cut points for the different outcome markers by CTS. number of RIBFX (<3 = 1. and J. for these complications were used as follows: Pneumonia required documentation by physical examination of chest pathology as well as presence of sputum. such as pneumonia and respiratory failure. we further delineated the patient cohort by using different cut points identified with sensitivity and specificity for the CTS with the outcomes being evaluated (mortality. abbreviated injury scale for chest (AIS chest). >5 = 3). tracheostomy. variables based on the distribution and chi-square analysis was used for categorical variables. and mortality. pneumonia. were chosen because of the nature of the study involving blunt chest trauma with RIBFX as these carry substantial morbidity. . VI. CTS $ 5 continued to be an important independent predictor for all three outcomes separately (mortality. ISS and AIS chest lacked significance for the three outcomes in some of the multivariable regression models (Tables V. and ARF). Receiver operating characteristics (ROCs) were used to compare the ISS and CTS for the different outcomes. test for equality by log rank).0 vs 2. Definitions for PNX and ARF are given in the Pennsylvania Trauma Systems Foundation 2011 manual to standardize capture of these entities in the registry. pulmonary contusion (none = 0. AIS 98 and 2005 were used to assess rib fractures and pulmonary contusions.

Injury Severity Score.4 5.043 <. patients 2 3 4 5 6 7 8 9 10 11 12 2. % ICU admission. Table IV.0 21. not significant.1 85. Abbreviated injury scale.0 54.3 98. Clinical pathways also have been developed to expedite the most appropriate treatment for patients with blunt torso trauma. ROC. it illustrates the importance of developing a system to triage patients.4%) Mortality.9 36.4 99.0 85.7 69.6 ± 19.9 0.001 <.6 2.2 4 IQR (2–11) 7 IQR (4–13) 3 IQR (3–4) 7.7 28.7 23.4 15.001 NS NS <.9 10.59–0.001.3 28.3 15 12. mean ± SD Mortality.2 1. Todd et al8 evaluated trauma patients with age greater than 45 years with four or more RIBFX. y.9 271 9.2 0 99.7 92.5 7 0 2 *P < . Sensitivity/specificity ROCs for chest trauma score (n = 1.70) Specificity % 100 0 100 10.9 61.0 5. Although this article was published more than a decade ago.72 (CI 0.8 0. with 73 mortality or 5.7 92. there is an unmet need for a simple scoring system to help categorize patients into greater-risk groups. They created a thoracic trauma severity score using several patient variables and found the thoracic trauma severity score was better at predicting complications compared to ISS and AIS chest alone.3 96. Confidence interval.001 AIS.64–0.0 155 10.7 6. they found that patients who were enrolled in an aggressive .65–0.6 96.8 79.1 9.77) Acute respiratory failure Specificity % Sensitivity % 100 0 95. lung contusions. % CTS < 5 (n = 724) CTS $ 5 (n = 637) P value 50. % Deep-vein thrombosis. CTS.4 1.5 29.7 4.4 1. CI.1 46.1 86. Clinical characteristics of patients by CTS cutpoint Age.001 <.8 74.71 (CI 0.0* 56.5 10.5 27.8 85.1 55. receiver operating characteristic.7 98. DISCUSSION There is no standardized method to assess the severity of blunt chest trauma immediately after the traumatic event.990 Chen et al Surgery October 2014 Table II.1* 55.3 133 0.5 1. ISS. Chest Trauma Score.7 3 IQR (1–6) 4 IQR (2–8) 3 IQR (2–3) 2. mean ± SD Pneumonia.2 29. Therefore.3 73. % ICU duration of stay (median) Hospital duration of stay (median) AIS chest (median) Tracheostomy (n = 68).9 55.2 0 99.4 78. % Pulmonary embolus. Table III.65 (CI 0. and number of RIBFX were associated with poor outcomes.2 341 5.6 ± 10.001 <.4 96 15.2 9.2 71.1 ± 18.1 93.1 23.2 74. % ISS.3 92.0 43.001 .361) Mortality Cutpoint $2 $3 $4 $5 $6 $7 $8 $9 $10 $11 $12 ROC Sensitivity % Pneumonia Specificity % Sensitivity % 100 0 95.001 <.001 <.3 0 99. % Abdominal trauma.79) *Statistical significance.4 <.2 ± 9.361. % No.1 94.8 36.001 <.6 3.1 2.7 2. interquartile range. Mortality by Chest Trauma Score* (total n = 1. IQR. % Acute respiratory failure. ICU.0 96.2 16. intensive care unit.4 26.8 0.5 1.2 10. NS.4 65 20.8 24 6. Pape et al7 found that bilateral chest injury.7 98.2 0 99.4 9.4 0 99.001 <.8 250 3.

appear to have substantial limitations. we were able to validate their scoring system and incorporate the CTS $ 5 into the data analysis. Because no other general marker exists for injuries over the course of the hospitalization. Injury Severity Score. Using this Chen et al 991 Fig 3. including pneumonia and mortality. Injury Severity Score. ISS. Number 4 Fig 1. does. Injury Severity Score. ISS. although comprehensive. Receiver operating characteristic (ROC)-derived mortality.001. Their study suggested a benefit to accurate early identification of patients with chest trauma to improve outcomes. clinical pathway had better outcomes. CTS. Bergeron et al9 found that patients older than 65 with three or more RIBFX had much greater mortality. Fig 4. Receiver operating characteristic (ROC)-derived pneumonia. however. Chest Trauma Score. as well as to simplify the coding by using one score to identify patients at risk rather than have several different scores. The ISS. ISS was used in the analysis and compared with CTS. Bulger et al10 found that older patients compared to younger patients with similar numbers of RIBFX had more complications. and mortality compared with historic control patients. Chest Trauma Score.Surgery Volume 156. The CTS was not different from ISS in the outcomes studied by ROC analysis. RIBFX scoring system but with a different CTS cutpoint ($5) based on ROC. CTS. They found that patients with CTS $ 7 were associated with greater duration of stay and ventilator days and greater mortality. Fig 2. By using CTS cutpoints defined by ROC. Chest Trauma Score. It is apparent from these studies that both age and the number of RIBFX are important risk factors that may be used to predict outcomes. AIS chest and ISS were not predictors in all the multivariable models. P < . We compared the CTS with ISS by ROC analysis. we corroborated their findings with similar predictors for mortality and morbidity in a larger data set. including decreased duration of stay. Pressley et al6 created the RIBFX scoring system we used and modified. CTS. pneumonia. ISS. Receiver operating characteristic (ROC)-derived acute respiratory failure. but the CTS was . Survival analysis by Chest Trauma Score (CTS).

001 2.92–8.70–2. Table VI.91 3.003 . The CTS provides an expedited way identify patients at risk of morbidity and mortality from RIBFX which the ISS and AIS chest are unable to duplicate.09–1.99 .22 0.04y P value Confidence interval 0.44 <.11 0.71.97–1.13–2.16–3. whereas in previously published papers investigators primarily used chest x-rays.74y 1.05 0.74 <0. and morbidity). The ROC for CTS and mortality is 0.001 0.45–0.45–12.043 <. of note.001y 1.22–2. predictive.05 0.94 1.31y 1.79 1.47–1.53 0.027 0.26 *Referent: Chest Trauma Score <5.001 .04 2.79 0. Abbreviated injury scale. while survivors were more likely to undergo a tracheostomy. AIS.11 0.67 Table VII.59 0.03 0.05 0.026 0. *Referent: Chest Trauma Score <5. AIS. none had actually died.92 2.58 0.10–5.49 3.48 0.03–7. Because of the heterogeneous characteristics of our trauma population with patients who present Odd ratio Chest Trauma Score $5* Injury Severity Score AIS chest Duration of stay Mortality Pulmonary embolus Deep-vein thrombosis Acute respiratory failure Tracheostomy Abdominal trauma 2.006 . AIS. Logistic regression multivariable analysis for pneumonia and chest trauma There are many limitations to this study.42 1.01 0.004y .51 0.57 .02 P value . whereas the CTS is easily calculated. yStatistical significance.38 0.08 *Referent: Chest Trauma Score <5.67–1.47 1.97–12.992 Chen et al Surgery October 2014 Table V. CT is more sensitive in detecting injuries.15 0.60 0.001 .014 0.12 0.02–16.96 Confidence interval 1.058 .001 .50 1.04 0.00–1.04 7.95 2.03–27.04 6. Abbreviated injury scale. The ability of the CTS to predict outcomes is clearly limited.43y 1.06 0. This clearly impacts duration of stay and tracheostomy as outcome variables because mortality occurred earlier in the hospital course of for some of these patients. Although it is beyond the scope of this study to address all the potential ramifications and clinical implications of using the CTS in daily practice.10 0. ISS is also a score that is calculated after discharge and all the charts have been abstracted ISS is not readily available on admission.01 0.99 0.99y <.60 .06 0.86–0. We did not assess the impact of subsequent therapeutic interventions on outcomes.62 <0.32 1.07 2. It was not possible in this retrospective study to control for patients who had care withdrawn rather than face a futile ICU course.36 1. This permitted the detection of more negative outcomes (mortality.21 0. it is still somewhat skewed toward the lesser scores greater than 5.30 .095 0. and.20 . yStatistical significance. This also explains why the greatest scores had at times discordant mortality rates.67 0.08 3.63–12.47–10.02–1.07y 0.05–1.67 1. Logistic regression multivariable analysis for acute respiratory failure and chest trauma Odd ratio Chest Trauma Score $5* Injury Severity Score AIS chest Duration of stay Pneumonia Pulmonary embolus Deep-vein thrombosis Mortality Tracheostomy Abdominal trauma 4. yStatistical significance.46–2. We used CT to assess blunt chest trauma.96 1. when divided by a CTS score of 5. . There were only two patients in the CTS 12 group. even though the cohorts are fairly equal in patient size.27 . It is also apparent that when Table III is evaluated further.88–5. and because of the low number of patients.09–3.021y 1. but increased the number of falsepositive results. especially pulmonary contusions. our overall mortality was 5%.91 4.14 0.69 5.055 1.52–7. Abbreviated injury scale.96 1.46 2. the CTS clearly allows early identification of patients that are at high risk for complications so that they may be considered for early interventions such as epidural analgesia or operative management for stabilization of the RIBFX.001 <.48 . Logistic regression multivariable analysis for mortality and chest trauma Odd ratio Chest Trauma Score $5* Injury Severity Score AIS chest Duration of stay Pneumonia Pulmonary embolus Deep-vein thrombosis Acute respiratory failure Tracheostomy Abdominal trauma P value 95% confidence interval 3.11–3.95–22. The CTS was weighted purposely toward greater sensitivity at the expense of specificity. Injured patients were managed by a mature trauma program that decreased variation in care through the use of standard management protocols.68–6.35 .30 1.090–1.

REFERENCES 1. this is a follow-up study to a previous publication looking at the scoring system. Lavoie A. Hackett Renner C. Number 4 with polytrauma and multiple comorbidities. the CTS was developed to gear toward greater sensitivity at the expense of specificity to avoid patients who may potentially have worse outcomes. Although no single scoring system is perfectly predictive of outcomes. The authors thank Amy H. Pulmonary contusion is a very challenging thing to quantify. J Trauma Nurs 2012. J Trauma Nurs 2012. Brasel KJ. that kind of calls into question whether you should be validating your scoring system with the same patients that you used to develop the scoring system. Mock CN. J Trauma 2003. Implementation of a rib fracture triage protocol in elderly trauma patients. 2. A couple of questions: one. Our definition of the pulmonary contusion was based off of the AIS chest coding system. Mion LC. 6. DISCUSSION Dr Brian Harbrecht (Louisville. Predicting outcome of patients with chest wall injury. Maxwell CA. 3. CTS is not able to identify every outcome. In your mortality analysis. Patients with rib fractures. So they were entirely different patient populations. et al. Sahr SH. however. Webb ML. Bergeron E. or did they just pull a pneumonia diagnosis because the fourth-year medical student said they might have pneumonia in the medical record? The other question hinges on mortality. 8. which is obviously a very important end point.Surgery Volume 156. 5. MS. we quantified that as being involved in less than one lobe. and thus. Apprasial of early evaluation of blunt chest trauma: Development of a standardized scoring system for initial clinical decision making. did the registrars use standard definitions? Was there a quality control to make sure that they were really coding pneumonia correctly. because many of the authors are the same and the institution is the same. Walters MR. Fry WR. and our study had about 1. this particular scoring system may help. Injury 2012. et al. the CTS provides a method to categorize chest wall injury and potentially intervene earlier in the hospital course of individual patients. it was done radiographically. J Trauma 2000. 10. McNally MM. or was this reviewed by the physicians? If it was registrar coding. and some patients with five don’t seem to turn a hair. Pape HC. yes could say there is some subjectiveness because it is based off two surgeons. Was this simply pulled from the registry based on registrar coding. Elderly trauma patients with rib fractures are at greater risk of death and pneumonia. Rice J. Am J Surg 2012.20:172-5. .48:1040-6. et al. Rib fractures in the elderly. I agree completely with your conclusion that people who have bad chest trauma are going to do worse. Bulger E. in many trauma patients. Todd SR. for her invaluable assistance in editing the manuscript.192:806-11.300 patients. Arneson MA. people who died of exsanguinating hemorrhage. A multidisciplinary clinical pathway decreases rib fracture-associated infectious morbidity and mortality in high-risk trauma patients. Was there a standard definition for applying the grading score for pulmonary contusion? Was this performed radiographically? Was this performed off the radiology report? Was it performed by the same individual? What was the standardization for that? Similarly for pneumonia: as you know.49:496-504. patients are usually overtriaged in an effort to avoid missed injuries. MPH. are the patients from the same data set? They are not. The other is that when you are dealing with some subjective variables. Pressley CM. et al. Rib fractures: relationship with pneumonia and mortality. Evans PA. I’m always humbled in terms of trying to predict who exactly is going to do better and worse. The CTS is an easy and quick method to assess the relative severity of blunt chest injury in a patient. Philp AS.54:3. et al. 4. Guse CE. Battle CE. And so for minor pulmonary contusions. So. because of the nature of trauma management. Crit Care Med 2006. Hospitalized injured older adults. Dietrich MS. They are actually analyzed from two different institutions.19:89-91. Am J Surg 2006.43:8-17. it was greater than one lobe. If it was a major pulmonary contusion. yes. et al. In addition.19:168-74. J Trauma Nurs 2013. 7. et al. In terms of defining pulmonary contusions. are these patients the same as that earlier data set? Because if they are. the definitions of those variables are key. people who had severe traumatic brain injuries? Or did you just simply quantify all-cause mortality? Dr Jennifer Chen: In terms of your questions. Some patients with one RIBFX are in Chen et al 993 excruciating agony. Risk factors that predict mortality in patients with blunt chest wall trauma: a systematic review and meta-analysis. Layde P. pneumonia can be a very challenging diagnosis.204:910-4. not the radiologist. The first study looked at about 600 patients at a separate institution. KY): As Dr Chen mentioned. did you exclude people whose death would be unlikely to be contributed to by the blunt chest trauma. Clas D. with many of the same coauthors from that institution. 9. Remmers D. Holcomb JB. Kao MD. Hutchings H. J Trauma 2000. Brown SD. We looked at CT scans by two separate surgeons. 34:6.

yes. I would like to think that for our institution. of associated injuries. we did not have a chance to exclude all of our mortality. The break point from one to five was 5% mortality at one RIBFX up to approximately 10% at five. And it was based off of coding. We did also report that age was an independent predictor of outcome. in the patients where age was not a predominant factor in that chest score. Dr Tim Hayward (Indianapolis. dramatically improve the data in one’s model to predict these things and would eliminate some of these people who weren’t going to survive from their other injuries and give you a much more accurate presentation. did you still have the similar findings that you did based on what you presented? Dr Jennifer Chen: When we looked at our four factors in our chest trauma scoring system. but. however. Did we look to see separately if age was a factor? We didn’t do that. the mortality went through the ceiling that when you got to eight. And in terms of any exclusion for mortality. . from the way that our study was conducted. we found that age and number of RIBFX were the most significant factors. Dr Frederick Luchette (Maywood. such as brain trauma. we didn’t exclude those who died for other reasons. it was pulled from the trauma registry. there’s a significant increase in mortality. KY): My question is that you included age as part of your scoring criteria. it was 30%. we actually presented a paper here using the National Trauma Data Bank.994 Chen et al In terms of how pneumonia was found. IL): Several years ago. My question to you is quite simple. When you get above five. it is something that we Surgery October 2014 would like to look into and see whether patients with a greater score would benefit from fixation. it would bring more vigilance to those patients who are at risk. we would be more apt to having them stay in the intensive care unit for vigilance and for early adaptation of epidural catheters for pain relief. the standard definitions. pneumothorax. Dr Martin Zielinski (Rochester. We did exclude some who showed up in our trauma bay and quickly died and for whom we didn’t have a chance to get a CT scan. Is this a scoring system that could potentially be applied to predict who needs it? Dr Jennifer Chen: We didn’t look to see in this study whether these patients ultimately required operative fixation. statistically significant. I didn’t see any data up there if any of your patients did get it. with each successive RIBFX. Otherwise. as you know. 40%. knowing that the patients have a greater CTS. What do I do with the CTS when I go home? How do I use that in my management of the patients? Dr Jennifer Chen: I would like to think that with our CTS. And there was no way to tell whether the coders were coding it accurately or if they were being pulled randomly from the chart because a medical student had written it. there’s an increased incidence. Dr Jason Smith (Louisville. Are we simply seeing a reproduction of older patients with a greater score having greater morbidity and mortality related to chest trauma? When you look at your scoring. pulmonary contusion. We validated that with each successive RIBFX and chest trauma. What also occurs. hemothorax. the training that your trauma registrars have to be put through on a monthly basis. MN): The American Association for the Surgery of Trauma scoring system developed back in the early 1990s and the AIS are terrible at predicting who needs rib fracture fixation. IN): Are you using the Trauma Quality Improvement Program at your trauma center? Because using that and looking at the probability of survival that those things calculate.