Establishing Prognosis System For The Critically Ill Obstetric Patients And Evaluating It's Effectiveness

BackgroundMaternal mortality is defined as the death of a woman while pregnant or within 42 days of termination of pregnancy, regardless of the site or duration of pregnancy, from any cause related to or aggravated by the pregnancy or its management. China has one of the highest maternal mortality ratio in the world. The implementation of effective measures is an effective treatment to reduce the maternal death. Severe illness prognosis scoring system applied in critically ill obstetric patients has some limitations. The critically ill obstetric patients of evaluation and prognosis system established according to physiological and pathological characteristics during pregnancy allow an assessment of the severity of disease and evaluation the risk of death in-hospital. It is significant in improving the quality of treatment and lowering maternal mortality.PartⅠEstablishing of the evaluation and prognosis system for the critically ill obstetric patientsMaterials and Methods :1. clincal data: A retrospective medical record review was conducted involving 311 cases critically ill obstetric patients, who were admitted to the ICUs of Third Affiliated Hospital of Guangzhou Medical College form January 2000 to july 2008.2. selecting parameters: The number of variables was reduced from 35 to 18. This was accomplished by eliminating infrequently variables and redundant variables,the smallest number of variables that reflected physiological derangement maintained statistical precision was found to be 12. 3. assignment: Relative weights of 18 variables were assigned according to the physiological and pathological characteristics during pregnancy, and referred the APACHE system of the parameters of the assignment.4. Prediction equation for the risk of death: three hundred and eleven critically ill obstetric patients were enrolled.After score was calculated, then analysis the relationship between mortality and prognosis,and establish the regression mode.5.statistical methods: Screening indicators using t test,measurement data are the number of people using the standard deviation (±s), a logistic regression model was applied fo setting the equation of mortality prediction.Result:1.The critically ill obstetric patients evaluation and prognostic system involves 18 physiologic variables include Body temperature, Bloodpressure, Respiratory rate,Heartrate, PaO2/FiO2 , arterial PH, Serum sodium, Serum potassium,Blood glucose concentration , Serum creatinine, Bilirubin total, Serum albumin, ALT transaminas, Blood platelets count, fibrinogen, White blood count, Hematocrit, Systema nervosum. The relative weights of each variable (0-4)were assigned and the maximum total score is 78.2.Prediction equation for the risk of death : log eY/(1-Y)=-7.275+0.245×总评分Part II The comparisons of critical ill obstetric patient assessment system and APACHEⅡMaterials and Methods :1.collecting cases: This was a retrospective medical record review that collected clinical information of 371 cases critically ill obstetric patients transferred in Third Affiliated Hospital of Guangzhou Medical College form January 2000 to February 2009, the worst value of each variable within the first 24 hr after admission was used.2. comparison: To compare and evaluate the prognostic value of the newly system and APACHEⅡon predicting mortality rate in critically ill obstetric patients, the calibration and discrimination of calculation of expected risk of death were used.3. statistical methods: The overal1 goodness-of-fit of two prognostic systems were assessed by using the areas under the receiver operating characteristic〔ROC curve cut-off point and the area under the curve (AUROCC)〕curves and classification tables.The standard mortality rate (SMR) is computed for each scoring system, which calculated the ratio of predicted death and observed death. The value of the Hosmer-Lemeshow statistic was obtained.There is significance if p is less than 0.05.Results1. Two hundred and ninty six cases of survival, seventy five cases of death; A signlficant difference was found between survivors and non-survivors groups on the score of patients scored by APACHEⅡand Evaluation system for critically ill obstetric patients, respectively (P<0.000;P<0.000).2. There were no significant differences in the c-index for APACHE II(0.910) and for evaluation system in critically ill obstetric patients( 0.906), both ROC area under the curve were greater than 0.80.3. There were significant differences in the predicted mortality and the actual mortality for APACHEⅡ(P = 0.000). And the predicted mortality in different clusters (≤10 score's group, 11-20 score's group,21-30 score's group,≥31 score's group) were significantly higher than the actual death (P =0.011,0.000,0.000,0.000).4. There were no significant differences in the actual mortality and the predicted mortality for Evaluation system for critically ill obstetric patients (P = 0.287). And there was no significant in different clusters (≤10 score's group, 11-20 score's group ,21-30 score's group,≥31 score's group) (P =1.000,0.748,0.756,0.679).5.There was no significant differences in the actual mortality and the predicted mortality of the evaluation system for critically ill obstetric patients by the Hosmer-Lemeshow goodness-of-fit test(x2 = 10.66, P-value = 0.28 > 0.05). There was significant in the actual mortality and the predicted mortality of APACHEⅡ(x2 = 19.08, P-value = 0.016 <0.05). Conclusions:1.The Evaluation and Prognosis system for the critically ill obstetric patients composed of 18 physiological variables has been established.2. The prediction equation for risk of mortality in the critically ill obstetric patients has been established: log eY / (1-Y) = 7.275 +0.245×total score.3. The newly system has been showed as a good predictor of prognosis and assessment of severity for the critically ill obstetric patients.4. The risk of mortality prediction equation has been showed that it was more accurate than APACHE II in predict the risk of death in the crowd .