L19 - secondary analysis of data

Question 1

What do we mean by secondary analysis?

Answer 1

learning from data that was originally collected for another purpose

using data generated as a result of health care proceses or transactions

not using prospectively-specified data collection

Question 2

What are some types of routine data?

Answer 2

medicare billing data

pharmacy dispensing data

computerised pathology report

the national inpatient minimum data set

Question 3

How is routine hospital data created?

Answer 3

recording of:

principal disgnosis
secondary diagnoses (co-morbidities)

principal procedure
hospital-acquired complications

past history

medications

Question 4

What happens to all the hospital data?

Answer 4

medical e-sheet send to patients GP. Hopefully given to patients family and aged care facilities

Question 5

True or false

There is an international classification of diseases so these codes form the e-sheet

Answer 5

true

Question 6

What is a DRG (used for research)?

Answer 6

Diagnosis-related groups - works off the principle diagnosis and principle procedure

Independent hospital pricing authority oversees classification for use in activity-based funding in Aus

Question 7

DRG assignment logic runs through an ____

Answer 7

algorithm

Question 8

Complicaions in a previous hospitalisation increase the risk of a

Answer 8

Complicaions in a previous hospitalisation increase the risk of a

Question 9

What are the strengths of clinical/health serices research?

Answer 9

whole population of interest

consistent and audited data extraction

documented coding standard

no reruitment delay

low cost

hypothesis generation

Question 10

What are the weaknesses of clinical/health serices research?

Answer 10

sample may be biased - particularly with only one hospital

retrospective; no opportunity

• intervene
• collect specific variables

Statistical but not ‘randomised’ control of confounding

clinician suspicion of coding process

Question 11

What is meant by ‘‘Big Data’’?

Answer 11

'’data mining’’ ‘‘text mining’’

may consist of:

• met-data
• content data

generally required ‘‘big’ hardware capacity and sophisticated data management software

Question 12

We can look at which DRGs and how they cluster arund certain _______

Answer 12

complications e.g

• CV
• post op
• obstetric
• neonatal