CICU OUTCOMES RESEARCH-Methodology Section Draft
 
Research Methods in FCS
 
Elizabeth Pash Penniman RD, LD
April 2006
 
            Outcomes research is likely to provide the gold standards for evaluating the effectiveness of health care interventions. To provide optimal care to each individual patient is the goal of all clinical practice. It is my professional obligation to understand quality of care and its improvement for the benefit of patients nutritionally speaking, in the coronary intensive care unit (CICU) as the unit dietitian.  As the unit dietitian I have noticed many patients’ nutritional status and care are often overlooked by the medical staff. It is my role as the professional to alert the physicians of patient’s nutritional status and recommendations for nutritional treatment. The main issue I encounter in the CICU is delayed feeding of patients mechanically ventilated or also referred to as intubated. The goal of my study is to promote a policy change for the CICU, based on the results of the study.  
This issue prompted me to use Wallace’s Wheel for the research process and I quickly began a review of the literature. Upon reviewing the literature, many conclusions pointed in the direction that early enteral nutrition (tube feeding) support improves patient outcomes. From here I developed my hypothesis; early enteral nutrition within 72 hours of mechanical ventilation in CICU patients admitted with myocardial infarctions, cardiac arrest, endocarditis, or congestive heart failure, will have decreased length of ICU stay, decreased hospital length of stay, decreased days on vent, and increased albumin levels, compared to patients feed after 72 hours of mechanical ventilation.
The next step is to operationalize each concept in the hypothesis. These operations are necessary to accurately and precisely measure each concept. The sampling method will be based on a post-hoc experimental design.  There is a need to draw a sample size in order to generalize findings to larger CICU populations.  Based on the number of 4 variables a 200 patient sample size will be sufficient. Sampling will be gathered from existing medical records of the patient. There will be four dependent variables. These variables are the interest in which I am exploring. The four dependent variables are:
1.      Decreased days on ventilator
2.      Decreased length of stay in ICU
3.      Decreased length of stay in hospital
4.      Increased albumin levels.
 
A hypothesis for each o the four variables is listed for clarity of each step of the research process.
 
1.      Among mechanically vented CICU patients admitted with myocardial infarction, cardiac arrest, endocarditis, and congestive heart failure, patients who receive early enteral nutrition (within 72 hours of intubation), will have decreased days on the ventilator, when compared to those who do not receive early enteral nutrition.
 
2.      Among mechanically vented CICU patients admitted with myocardial infarction, cardiac arrest, endocarditis, and congestive heart failure, patients who receive early enteral nutrition (within 72 hours of intubation), will have decreased length of stay in the ICU, when compared to those who do not receive early enteral nutrition.
 
3.      Among mechanically vented CICU patients admitted with myocardial infarction, cardiac arrest, endocarditis, and congestive heart failure, patients who receive early enteral nutrition (within 72 hours of intubation), will have decreased length of stay in the hospital, when compared to those who do not receive early enteral nutrition.
 
4.      Among mechanically vented CICU patients admitted with myocardial infarction, cardiac arrest, endocarditis, and congestive heart failure, patients who receive early enteral nutrition (within 72 hours of intubation), will have increased albumin levels, when compared to those who do not receive early enteral nutrition.
 
When gathering the data it is important to focus only on the four variables we are studying. When reviewing the patient charts, it will be imperative to look for type of illness of patient or diagnosis, timing of the feeding, which is found on nursing medication sheets that are timed and documented each hour for tube feeding administration. Also the number of days on the ventilator, which are documented in nursing assessment form, number of days in the ICU and hospital, can look at admission date, transfer orders to regular nursing floors and discharge dates. Lab values will be noted at the first albumin draw closet to admission ad discharge dates. A table will be made into an excel spreadsheet, so the data can easily be transferred. A codebook would be used to quickly organize and reference the data.
 
Codebook:
Column 1         Patient identification number
                        001-999
Column 2         Enteral Nutrition Timing
1.      Within 72 hours of intubation
2.      After 72 hours of intubation
Column 3         Admitting Diagnosis
1.      Myocardial Infarction
2.      Cardiac Arrest
3.      Endocarditis
4.      Congestive heart failure
Column 4         Days on Ventilator
                        0-999
 
Column 5         Days in ICU
                        0-999
Column 6         Days in Hospital
                        0-999
Column 7         Albumin Level near/on admission date
                        .5-6.5mg/dl
Column 8         Albumin level near/on discharge date
                        .5-6.5mg/dl
Column 9         Reasons patients did not complete study
1.      None
2.      Expired
3.      Transferred to LTAC (long term acute care facility)
4.      Extubated within 72 hours without tube feeding
5.      Other reasons
 
After using the codebook as an organizational tool to promptly track data, it can then be entered into a table/spreadsheet.
(LOS-length of stay)
 
Table/Spreadsheet with 2 “sample” patients

Pt. ID #	TF Timing	Admitting Dx	LOS Vent	LOS ICU	LOS Hospital	Admission Albumin	Discharge Albumin	Reasons uncompleted
001	2	1	6	8	14	2.8	3.2	1
002	1	3	13	17	23	2.4	2.9	1

 Once the data has been collected and entered into the spreadsheet with 200 patients completing the study, the data can interpreted into statistics. A T-test would be the best statistical method to use, as it determines the difference between 2 means measuring the same variable. This could also lead to an ANOVA, which is analysis that estimates difference between groups after a posttest, which would be the difference between patients who received enteral nutrition within 72 hours of mechanical ventilation and after 72 hours of mechanical ventilation. The T-Test will include all patients in the sample, to address each hypothesis, and establish whether or not early enteral nutrition in mechanically ventilated CICU patients will have the hypothesized effect.
 T-Test to be filled with data

TF within 72H (TF-tube feeding)	Mean of days on vent	Mean of days in ICU	Mean of days in hospital	Mean of albumin difference

 
 Characteristics of the sample to be filled with calculated data

Variables	Mean	Standard Deviation
Vent days
ICU days
Hospital days
Albumin Difference

 T-Test for effects of early enteral feeding to be filled with calculated data

Variables	Timing	Mean	Standard Deviation	t	df	Sig. (2-tailed)
Vent Days	w/in 72 H
	after 72 H
ICU Days	w/in 72 H
	after 72 H
Hosp Days	w/in 72 H
	after 72 H
Alb. Difference	w/in 72 H
	after 72 H

 
Significant at p<.05
 
            After the study completion, data entry, calculations and statistical analysis I will be able to write my findings based on how the comparative T-Tests were used. Examples in the writings might include, “ Patients feed within 72 hours on intubation had an average length of ICU stay of 4.5 days compared to patients feed after 72 hours on intubation had and average length of stay of 8.25 days in the ICU.”  From here one might conclude that early enteral feeding within 72 hours of mechanically ventilation decreases length of ICU stay which is beneficial to the patient. Further conclusions can be drawn using more T-Tests or other statistical methods and comparing the differences between each admitting diagnosis, tube feeding time from intubations and the other 4 dependent variables as well. Conclusions can be drawn that a possible diagnosis may have better outcomes, pending the diagnosis and treatment applied. Upon completion of the data collection and analysis of statistics, recommendations will be made for the benefits of early enteral nutrition benefits in the CICU setting. Long-term plans would include initiating a CICU policy of when to start enteral nutrition in the CICU population with the aforementioned diagnosis.