EFFECTIVENESS OF RIGHT HEART CATHETERIZATION IN …



NURS 701 - Statistical Analysis Questions – Assignment 5(Due by Midnight – Thursday, October 17th, 2019)1 – Swan-Ganz Catheterization and 30-Day Mortality of Heart PatientsBelow is the abstract from the paper “Effectiveness of Right Heart Catheterization in Critically Ill Patients” published in JAMA, 1996, by Conners et al.OBJECTIVE: To examine the association between the use of right heart catheterization (RHC) during the first 24 hours of care in the intensive care unit (ICU) and subsequent survival, length of stay, intensity of care, and cost of care. DESIGN: Prospective cohort study. SETTING: Five US teaching hospitals between 1989 and 1994. SUBJECTS: A total of 5735 critically ill adult patients receiving care in an ICU for 1 of 9 prespecified disease categories. MAIN OUTCOME MEASURES: Survival time, cost of care, intensity of care, and length of stay in the ICU and hospital, determined from the clinical record and from the National Death Index. A propensity score for RHC was constructed using multivariable logistic regression. Case-matching and multivariable regression modeling techniques were used to estimate the association of RHC with specific outcomes after adjusting for treatment selection using the propensity score. Sensitivity analysis was used to estimate the potential effect of an unidentified or missing covariate on the results. RESULTS: By case-matching analysis, patients with RHC had an increased 30-day mortality (odds ratio, 1.24; 95% confidence interval, 1.03-1.49). The mean cost (25th, 50th, 75th percentiles) per hospital stay was $49 300 ($17 000, $30 500, $56 600) with RHC and $35 700 ($11 300, $20 600, $39 200) without RHC. Mean length of stay in the ICU was 14.8 (5, 9, 17) days with RHC and 13.0 (4, 7, 14) days without RHC. These findings were all confirmed by multivariable modeling techniques. Subgroup analysis did not reveal any patient group or site for which RHC was associated with improved outcomes. Patients with higher baseline probability of surviving 2 months had the highest relative risk of death following RHC. Sensitivity analysis suggested that a missing covariate would have to increase the risk of death 6-fold and the risk of RHC 6-fold for a true beneficial effect of RHC to be misrepresented as harmful. CONCLUSION: In this observational study of critically ill patients, after adjustment for treatment selection bias, RHC was associated with increased mortality and increased utilization of resources. The cause of this apparent lack of benefit is unclear. The results of this analysis should be confirmed in other observational studies. These findings justify reconsideration of a randomized controlled trial of RHC and may guide patient selection for such a study.Research Goal: The goal of your analysis of these data is to discuss and quantify the risk of 30-day mortality of patients that had a Swan-Ganz right heart catheter used during their treatment vs. those that did not. You will first do this marginally, that is by looking only at the relationship between 30-day mortality (Y) and right heart catheterization (X). Then you will examine this relationship adjusted for other covariates that might be related to the 30-day mortality of the patient.DESCRIPTION OF VARIABLES (Datafile: RHC.JMP)Response (Y)DeathPatient died within in 30 days? (Died or Survived)Risk Factor of Primary Interest (X1)swang1Right Heart Catheterization (RHC vs. No RHC) COVARIATES USED FOR ADJUSTMENT PURPOSES (X2, ... ,X38)Demographics and Disease CategoryVariable nameVariable Definition AgeAge (yrs.)SexSex (Female or Male)RaceRace (white, black, other)EduYears of education (yrs.)IncomeIncome (Under $11k, $11k - < $25k, $25k - $50k, > $50k )NinsclasMedical insurance (No Insurance, Medicaid, Medicare, Medicare & Medicaid, Private, or Private & Medicare)CaCancer status (Yes, No, or Metastatic)Categories of Admission Diagnosis?Diagnosis variables are all coded as (Yes or No)RespRespiratory DiagnosisCardCardiovascular DiagnosisNeuroNeurological DiagnosisGastrGastrointestinal DiagnosisRenalRenal DiagnosisMetaMetabolic DiagnosisHemaHematologic DiagnosisSepsSepsis DiagnosisTraumaTrauma DiagnosisOrthoOrthopedic DiagnosisDnr1DNR status on day1 (Yes or No)Aps1APACHE scoreScoma1Glasgow Coma ScoreWtkilo1Weight Temp1TemperatureMeanbp1Mean blood pressureResp1Respiratory rateHrt1Heart ratePafi1PaO2/FIO2 ratioPaco21PaCo2Ph1PHWblc1WBCHema1HematocritSod1SodiumPot1PotassiumCrea1CreatinineBili1BilirubinAlb1Albumin wtkilo1Weight (kg)QUESTIONS AND ANALYSESa) Use appropriate methods to examine the relationship between right heart catheterization (swang1) and 30-day mortality (Death). The methods you use should provide achieve the following analysis goals:Determine if a significant relationship between right heart catheterization and 30-day mortality exists.Estimate the risk difference associated with right heart catheterizationEstimate the risk in multiplicative terms.Summarize all your findings in regards to these three analysis goals. (10 pts.)b) Now consider adjusting for all of the covariates listed above. Use backward elimination or forward selection (using the default settings in JMP) to develop a logistic regression model that contains the most important covariates along with the right heart catheterization predictor (swang1), then calculate the OR for 30-day mortality associated with the Swan-Ganz right heart catheterization procedure adjusted for the other included covariates. Summarize your findings and be sure to explain how you arrived at your final multiple logistic regression model. (15 pts.)c) Using the table below, show the OR associated with each of the covariates in the final model including Swan-Ganz line status. For continuous/numeric covariates report the Unit Odds Ratio, i.e. the change in odds for a one unit change in that predictor. You will notice that JMP has combined categories on the cancer status variable (ca) to create a single term Metastatic & Yes vs. No. This term compares those with those with metastatic or yes classification to those with no cancer in terms of odds for death. (10 pts.)Factor or TermOR95% CI for ORp-valueHint: The table provided above has the same number of terms as the final model you should get using stepwise selection methods.Body Fat & Body Density (Datafile: Bodyfat.JMP)In problem 2 you look at the simple linear regression of height on weight. In problems 3 & 4 you will look at regression models for body density and percent body fat using other body measurements for potential predictors. These data were collected on n = 252 adult males and were originally used in the article “Generalized body composition prediction equation for men using simple measurement techniques”, published in Medicine and Science in Sports and Exercise by Penrose et al. (1985). Description of VariablesDensity - body density determined from weighing underwater (Response for Prob. #4)PCTBF - percent body fat from Siri’s equation (1956) (Response for Prob. #3)Age- age (yrs.)Weight- weight (lbs.) (Response for Prob. #2)Height- height (in.)Neck- neck circumference (cm)Chest- chest circumference (cm)Abdomen- abdomen circumference (cm)Hip - hip circumference (cm)Thigh- thigh circumference (cm)Knee- knee circumference (cm)Ankle - ankle circumference (cm)Biceps- bicep circumference (cm)Forearm - forearm circumference (cm)Wrist - wrist circumference (cm)2 - SIMPLE LINEAR REGRESSION PROBLEMWhat is the correlation between Height (in) and Weight (lbs.)? Give a 95% CI for this correlation and interpret. (3 pts.)Look at the scatterplot of Weight (lbs.) vs. Height (in.). Are there any extreme outliers? If so, find the subject number corresponding to this outlier. If you had to describe this outlier what would you say? (3 pts.)Fit the simple linear regression of Weight (lbs.) on Height (in.) using Analyze > Fit Y by X in JMP then answer the following questions based upon the results this regression analysis.What is the estimated y-intercept, be sure to give the proper units? If possible, give the interpretation of this quantity. (3 pts.)What is the estimated slope, be sure to give the proper units? If possible, give the interpretation of this quantity. (3 pts.)What is estimated mean weight (lbs.) for all adult men who are 70 inches tall? (1 pt.)I am 71.5 inches tall, how much do you think I weigh? (1 pt.)What is the R-square for the model? Explain in words what this quantity tells you specific to this model? (2 pts.)How is the R-square related to the correlation between height (in.) and weight (lbs.)? Explain. (2 pts.)Examine the set of residual plots given by select Plot Residuals from the Fit Line pull-down menu located below the scatterplot and answer the following questions. Do these plots suggest any model violations? If so, what are they? (3 pts.)How big is the residual corresponding to the largest outlier you identified in part (b)? (1 pt.)3 - MULTIPLE LINEAR REGRESSION PROBLEM (Percent Bodyfat)Using variables Age – Wrist as potential predictors in the list of the variable descriptions above and percent body fat (PCTBF) as the response answer the following questions.Which three variables have the strongest correlation with percent body fat? What are these correlations and are they statistically significantly different from zero? (4 pts.)Which variable has the weakest correlation with percent body fat? What is this correlation and is it statistically significantly different from zero? (2 pts.)Using all available predictors in your model (Age – Wrist) fit a multiple regression for PCTBF and answer the questions in parts (c) – (g). Use Minimal Report as the fitting Emphasis in the Fit Model dialog box.Which predictors are significant at the ????????level? (3 pts.)Give and interpret the R-square for this model. (2 pts.)Complete the following sentence: Holding all other predictors constant we estimate that the change in the percent body fat corresponding to a 1 cm increase in the abdominal circumference is between _________ % and ________% with 95% confidence. (2 pts.)Using the default settings from JMP, perform stepwise model selection to arrive at a “final” model for percent body fat (PCTBF). What predictors are in your final model? (3 pts.)What is the R-square for your “final” model and how does it compare to the R-square from part (d)? (2 pts.)Save the Predicted Values by selecting that option from Save Columns menu. Then find the correlation between the Predicted Values and the actual percent body fat (PCTBF) and square that correlation. How does this compare to your answer from part (g)? (2 pts.)In your “final” model you should find two covariates that have negative estimated coefficients, what are they? (2 pts.)Look at the pairwise correlations between percent body fat and the covariates with negative coefficients in part (i), does something seem strange to you? If so, what? Explain. (2 pts.)Here is the reason for the strange result described in part (j). The plot below is scatterplot of percent body fat vs. wrist circumference. In the histogram of Abdomen circumference I have highlighted only those with individuals with an abdomen circumference between 90-100 cm. A line fit to only those highlighted points (in blue) in the percent body fat vs. wrist plot shows a negative trend! This is why the coefficient of wrist circumference is negative when abdomen is included in the model also.4 – MULTIPLE REGRESSION MODEL PROBLEM (Body Density) Using Y = Body Density as the response and every variable except percent body fat (PCTBF) as potential predictor, use stepwise methods to develop a final model for predicting body density. In terms of the predictors chosen how does this model compare that for percent body fat? (5 pts.)Use the residuals from your “final” model from part (a) to check model assumptions. Do the residuals suggest any problems with model assumptions? Explain. (4 pts.)5 – MULTIPLE REGRESSION (ANCOVA) – (Datafile: Low Birth Study.JMP)Description of VariablesPrev?- History of premature labor (None or History)Hyper - Hypertension during pregnancy (Normal or HT)Smoke- Smoking during pregnancy (No Cig or Cig)Uterine- Uterine irritation during pregnancy (None or Irritation)Minority - Minority status (Nonwhite or White)Age - Age of mother (yrs.)Lwt- Mothers weight at last menstrual cycle (lbs.)Y = Birth Weight (g), birth weight of infant in gramsUsing Y = Birth Weight (g) as the response develop a regression model using all available predictors. Use stepwise selection methods to choose a “final” model. What predictors are in your final model? (5 pts.)Estimate the change in mean birth weight associated with smoking adjusted for predictors in your “final” model. Give a 95% CI for this change. Discuss. (4 pts.) ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download