Title to go here - PowerPoint PPT Presentation

1 / 22
About This Presentation
Title:

Title to go here

Description:

Ipsos MORI, part of the Ipsos Group, is a leading UK research company with global reach. Ipsos MORI: Report Title. ... CATI. From CATI we get around 50 responses. – PowerPoint PPT presentation

Number of Views:90
Avg rating:3.0/5.0
Slides: 23
Provided by: pete3314
Category:
Tags: cati | here | research | title

less

Transcript and Presenter's Notes

Title: Title to go here


1
Title to go here
The GP Patient Survey - Managing the largest
healthcare study in the UK
  • By Peter Deacon Mel Parekh

2
Introduction
  • Ipsos Mori
  • Who are Data Delivery
  • What is the GP Patients Survey
  • The work we do
  • Reporting
  • Close
  • Q A

3
Ipsos Mori
  • Ipsos MORI, part of the Ipsos Group, is a
    leading UK research company with global reach.

4
Who are Data Delivery?
5
The GP Patients Survey
  • 1.3 million Questionnaires are sent out each wave
  • It is currently run twice a year
  • Each wave has two sets of reminders
  • The current questionnaire contains 58 questions
  • Its broken into 12 Sections
  • Response Rate of roughly 36
  • Assess the performance of practices

6
GP Patients Survey Team
7
Data Collection
  • Scanned
  • From the scanned we get around 450,000
    questionnaires back
  • which is approximately 95.5 of the total
    response
  • Online
  • From Online we get around 20,000 responses
  • which is approximately 4.5 of the total response
  • CATI
  • From CATI we get around 50 responses
  • which is approximately 0.001 of the total
    response

8
How we use SPSS to solve the data issues
  • What we do?

9
What we do?
  • Data cleaning
  • Merge
  • Weighting
  • Reporting Variables
  • Aggregation
  • Reports

10
How do we make three different files the same?
  • The Raw Data

11
Issues with the Raw data
  • The Data files are all different
  • We need to remove any incorrect responses
  • There may be multiple responses from the same
    respondent
  • We need to make sure all the variables and values
    are present

12
Raw Data Files
  • The structure of the files
  • Three data files
  • Scanned
  • CATI
  • Online
  • Each need to be cleaned separately to be matched
    to each other
  • Error Codes, when are they used
  • Duplicate records
  • Frequencies and checking

13
How do we bring the files together with the
metadata?
  • Merging the data

14
Problems faced when merging
  • The size of the files, each wave we have 500,000
    cases
  • A years data is nearly 2 million cases, with over
    200 variables
  • 1 gig worth of data
  • We have to bring in meta-data from 6 different
    files

15
Sample Data
  • SHA level
  • Names of the SHAs, old codes and new ONS codes
  • GOR data
  • Government Office Region codes
  • Names of the GORs, the population within each
    and the amount of questionnaires we distributed
    in each region
  • LA data
  • Local Authority names and codes, as well as the
    populations of each and the number of
    questionnaires distributed within each
  • We bring in sample data in the merge from several
    different files
  • Patient Level
  • Postcode, practice code they apply to, PCT codes
    and SHA codes as well as IMD scores
  • Practice level
  • Name of the practices the population of their
    catchment area, the number of people signed up to
    attend the practice
  • PCT level
  • Names of the PCT, old codes and new ONS codes

16
Merging Files
  • Merge, bringing all three files together
  • Matching in Sample data
  • Using Break files, to save data at certain
    points
  • Weights
  • Master Patient Level Data

17
We cant report on the data as it is
  • Preparing the data for reporting

18
Issues with reporting
  • We need to show the data at different levels of
    order
  • We need to able to produce percentages for each
    question
  • We need to be able to calculate confidence
    intervals
  • We need to put the data into excel templates

19
Preparing data for reports
  • We create Reporting Variables where we break all
    the variables down again
  • Creating base values
  • Creating top 2 figures
  • Distpop syntax files
  • Bring patient and practice sample files
  • Create aggregation files that contain the number
    of questionnaires distributed and the population
    sizes at the National, SHA, PCT and practice
    levels

20
Aggregating the data
  • Aggregation files we can create multiple excel
    datasets based on what level of order we require
  • Percentage and CI files
  • Practice and PCT
  • Un-weighted data and weighted data files
  • Export to Excel for reporting
  • Excel data is then loaded into the reports

21
Reporting
  • SPSS Dataset at the Patient level
  • Excel Data files
  • National
  • PCT
  • Practice
  • Checked and signed off by the exec team
  • Data gets uploaded to the website

22
Thank you
  • peter.deacon_at_ipsos.com mel.parekh_at_ipsos.com
Write a Comment
User Comments (0)
About PowerShow.com