|
|
|
|
|
|
||
|
|
|
|
Due: 11/18/99 |
|
|
|
|
|
|
||
|
|
|
|
Due: 11/18/99 |
Instructions: For this assignment, forms groups of 2-4 students. Meet to plan an outline of the various steps in completing the assignment. Prepare a program that is aimed at addressing as many questions as possible for the 11/15/99 or 11/16/99 class. Your program should be posted on a WEB page for someone in your group as an ASCII file if you have program specific questions.
Complete programs that jointly enable question g to be answered, with a permanent data set created. Using the permanent data set, write individual programs that answer question h. Compare your answers among the group, and construct a single group report that contains the results of question g and h, with the programs listed in an appendix (and on a WEB page). Print a paper copy of the report to hand in.
Midwifery Transfer Study
It is increasingly common for births to be delivered by a midwife instead of a physician. Certified nurse midwives have physician back-up. In some cases, women are transferred from midwife care to physician care during the course of the pregnancy. A pilot study was conducted to describe the reasons for these transfers. Data were collected using a form that is attached. The form reported the patient ID, medical record number, date of transfer and date of birth. The patient's name has been removed from the data to protect confidentiality. The site and gestational age is also recorded.
Following this basic descriptive information, each patient was categorized as to one of four possible "reason for transfer" groups. For each "reason for transfer", some follow-up information was collected that further defined the reason. Finally, there was an indication of whether or not patients returned to certified nurse midwife care. Data for this study are contained in a D-BASE data set called HW10.dbf. A copy of the study instrument is given for reference.
a. Response to some questions was not-applicable, since the reason for transfer did not apply. A not-applicable response was different from a response of "unknown". Response for other questions was missing, since a response was not recorded. For other questions, a response may have been recorded for a question, but the response may have conflicted with other data in for the subjects.
Assume that the response reported for "REASON FOR TRANSFER" is entered correctly. Recode response for questions 9-12 to "not applicable" if they were not appropriate according to response for "REASON FOR TRANSFER". For questions 9-12, assign a special missing value code of :
b. Create a variable representing patient's age groups at the time of transfer. Let the groups be defined as <15, 15-17, 18-19, and 20+.
c. Create a variable representing trimester of pregnancy at transfer from the variable for gestational age. Define this variable as <14 wks in 1st trimester, 14-27 weeks in 2nd trimester, and 28-42 weeks in third trimester. Values of gestational age greater than 42 weeks are suspect, and should be treated as "missing".
d. Add labels to your SAS data set, and to all variables in your SAS data set.
e. Create formats for each nominal and ordinal variable in the data set. Save these formats in a permanent SAS data set using the CNTLOUT option.
f. Create a temporary data set that links the formats to the midwifery data. Obtain the contents of this data set.
g. Obtain the frequency of response for questions 8-12 using your formats.
h. Obtain additional tables as necessary to answer the following questions:
What percent of patients who suffered an early loss were over 20 years old?Of those transferred for medical or obstetric indications, what percent returned to certified nurse managed care (CNM)?
Of those self transferred or transferred as a result of early loss, what percent returned to certified nurse managed care (CNM)?