survey methods & applications in healthcare
TRANSCRIPT
Survey Methods & Applications in Healthcare
Nawanan Theera-Ampornpunt, M.D., Ph.D.Faculty of Medicine Ramathibodi Hospital
Mahidol Universitywww.SlideShare.net/Nawanan
2
2003 M.D. (First-Class Honors) (Ramathibodi)2009 M.S. in Health Informatics (U of MN)2011 Ph.D. in Health Informatics (U of MN)2012 Certified HL7 CDA Specialist
• Deputy Executive Director for Informatics (CIO/CMIO)Chakri Naruebodindra Medical Institute
• Lecturer, Department of Community MedicineFaculty of Medicine Ramathibodi HospitalMahidol University
[email protected]://groups.google.com/group/ThaiHealthITwww.SlideShare.net/Nawanan
Introduction
3
• Overview of Surveys• Survey Methodology• A Sample Survey
Outline
4
OVERVIEW OF SURVEYS
5
• An activity in which many people are asked a question or a series of questions in order to gather information about what most people do or think about something
Survey
Merriam Webster Dictionary
6
• A written set of questions that are given to people in order to collect facts or opinions about something
Questionnaire
Merriam Webster Dictionary
7
• To know something– Personal information– Knowledge, Opinions & Attitudes– Behaviors & Practice– etc.
• About someone– Individuals– Organizations
• In order to understand, create knowledge, or make decisions
Why Do a Survey?
8
Examples of Surveys
Say2KFC.com
Service Satisfaction Survey (Online)
9
Examples of Surveys
http://www.kmutt.ac.th/building/pdf/application_form_2552_1.pdf
Service Satisfaction Survey (Paper-based)
10
Examples of Surveys
Image Source: http://pixabay.com/en/agent-business-call-center-18741/
Service Satisfaction Survey (Telephone)
11
SURVEY METHODOLOGY
12
• Survey Design– Study Design– Modes of Data Collection
• Instrument Design• Sampling• Survey Conduct• Data Analysis• Reports
Survey Methodology
13
WHAT WOULD YOU DO TO INCREASE SURVEY
RESPONSES?
14
The Tailored Design Method
1978 1999 2007 2008
Dillman et al.
15
The Tailored Design Method
2014
Dillman et al.
16
• “The development of survey procedures that create respondent trust and perceptions of increased rewards and reduced costs for being a respondent, which take into account features of the survey situation and have as their goal the overall reduction of survey error”
The Tailored Design
Dillman et al. (2007)
17
• Rewards• Costs• Trust
Social Exchange Theory
Dillman et al. (2007)
Image Source: http://horsebusinessschool.com/using-stick-and-carrot-to-motivate-employees/
18
• “The likelihood of responding to the request to complete a self-administered questionnaire, and doing so accurately, is greater when the respondent trusts that the expected fix rewards of responding will outweigh the anticipated costs”
Social Exchange
Dillman et al. (2007)
19
• Providing Rewards– Show positive regard– Say thank you – Ask for advice– Support group values– Give tangible rewards– Make the questionnaire interesting– Give social validation– Inform respondents that opportunities to
respond are scarce
Survey & Social Exchange Theory
Dillman et al. (2007)
20
• Reducing Social Costs– Avoid subordinating language– Avoid embarrassment– Avoid inconvenience– Make questionnaires appear short & easy– Minimize requests to obtain personal
information– Keep requests similar to other requests to
which a person has already responded
Survey & Social Exchange Theory
Dillman et al. (2007)
21
• Establishing Trust– Provide a token of appreciation in advance– Sponsorship by legitimate authority– Make the task appear important– Invoke other exchange relationships
Survey & Social Exchange Theory
Dillman et al. (2007)
22
• Study Design– Cross-sectional– Longitudinal
Survey Design
23
• Mode of Data Collection– Self-administered survey
• Paper-based• Online• Telephone (Automated/IVR)
– Interviewer-administered survey (structured interview)
• In-person• Telephone
– Mixed-mode survey
Survey Design
24
• Choose simple over specialized words• Choose as few words as possible to pose the
question• Use complete sentences to ask questions• Avoid vague quantifiers when more precise
estimates can be obtained• Avoid specificity that exceeds respondent’s
potential for having an accurate, ready-made answer
• Use equal numbers of positive & negative categories
Instrument Design
Dillman et al. (2007)
25
Problem• Number of years lived in Idaho
Years• Your city or town
City or Town• Your county
County
Use complete sentences
Dillman et al. (2007)
26
Revision• How many years have you lived in Idaho?
Years• In what city or town do you live?
City or Town• In what Idaho county do you live?
Idaho County
Use complete sentences
Dillman et al. (2007)
27
Problem• How often did you attend religious services
during the past year? Never Rarely Occasionally Regularly
Avoid vague quantifiers
Dillman et al. (2007)
28
Revision• How often did you attend religious services
during the past year? Not at all A few times About once a month Two to three times a month About once a week More than once a week
Avoid vague quantifiers
Dillman et al. (2007)
29
Problem• About how many books have you read for
leisure during the past year?Number of books
Avoid too much specificity
Dillman et al. (2007)
30
Problem• About how many books have you read for
leisure during the past year? less that 10 11-25 26-50 51-75 76 or more
Avoid too much specificity
Dillman et al. (2007)
31
• Distinguish undecided from neutral by placement at the end of the scale
• Avoid bias from unequal comparisons• State both sides of attitude scales in the
question stems• Eliminate check-all-that-apply question formats
to reduce primacy effects• Develop response categories that are mutually
exclusive
Instrument Design
Dillman et al. (2007)
32
Problem• Which one of the following do you feel is most
responsible for recent outbreaks of violence in American’s schools? Irresponsible parents School policies Television programs
Avoid bias from unequal comparisons
Dillman et al. (2007)
33
Revision• Which one of the following do you feel is most
responsible for recent outbreaks of violence in American’s schools? The way children are raised by parents School policies Television programs
Avoid bias from unequal comparisons
Dillman et al. (2007)
34
Problem• From which one of these sources did you first
learn about the tornado in Derby? Radio Television Someone at work While at home While traveling to work
Mutually Exclusive
Dillman et al. (2007)
35
• Use cognitive design techniques to improve recall
• Provide appropriate time referents• Be sure each question is technically accurate• Choose question wordings that allow essential
comparisons to be made with previously collected data
• Avoid asking respondents to say yes in order to mean no
• Avoid double-barreled questions
Instrument Design
Dillman et al. (2007)
36
Problem• Should the city build a new swimming pool that
includes lanes for swimming laps that is not enclosed for winter use? Yes No
Double Barreled
Dillman et al. (2007)
37
• Soften the impact of potentially objectionable questions
• Avoid asking respondents to make unnecessary calculations
Instrument Design
Dillman et al. (2007)
38
• Choose simple over specialized words• Choose as few words as possible to pose the
question• Use complete sentences to ask questions• Avoid vague quantifiers when more precise
estimates can be obtained
Instrument Design
Dillman et al. (2007)
39
• Survey Design– Study Design– Modes of Data Collection
• Instrument Design• Sampling• Survey Conduct• Data Analysis• Reports
Survey Methodology
40
• Target Population• Sample• Inclusion & Exclusion Criteria• Sampling Frame• Sampling Method/Technique
Sampling Design
41
• Census• Random (probability) sampling
– Simple random sampling– Systematic random sampling– Stratified random sampling– Cluster random sampling– Multi-stage random sampling
• Non-probability sampling– Purposive sampling– Convenience sampling– Snowball sampling– etc.
Sampling Design
42
Example sequence• Pre-notice Letter• 1st Questionnaire Mailout• Reminder + 2nd Questionnaire Mailout
Survey Conduct
Simplified from Dillman et al. (2007)
43
• Sampling error• Coverage error• Measurement error• Nonresponse error
– Questionnaire effects– Data collection mode effects– Interviewer effects– Respondent effects
• Processing error
Errors in Surveys
Dillman et al. (2007) & Office of Management and Budget (2001)
44
• Sampling error• Coverage error• Measurement error• Nonresponse error• Processing error
– Data entry error– Pre-edit coding errors– Editing errors– Imputation errors
Errors in Surveys
Dillman et al. (2007) & Office of Management and Budget (2001)
45
A SAMPLE SURVEY
46
NAWANAN’S DOCTORAL STUDY
47
Unknown State of IT Adoption in Thai Hospitals
Need to improve theoretical knowledge
This Study
Dual Opportunities
48
• To describe current state of IT adoption in Thai hospitals nationwide
• To test proposed conceptual framework & explore relationships between organizational characteristics, IT management, and IT adoption
Study Objectives
49
Hypothesized Model
50
• Study Design: Nationwide cross-sectional mail survey• Sample: All hospitals in Thailand except pilot (N =
1,302)• Pilot: 5 hospitals (10 respondents each)• Sampling Frame: List of hospitals from Ministry of
Public Health’s Web site• Subjects: Hospital’s staff responsible for managing
information systems (CIO/IT manager or equivalent; hospital director if N/A)
• Data Collection Period: 16 weeks
Design & Population
51
• Modified from original instrument• Face & content validity established
(Theera-Ampornpunt, 2009)
• Further modified based on pilot findings
• Translated to Thai
Section 1 Hospital ProfileSection 2 IT Adoption & Use ProfileSection 3 Respondent’s Information
English version
Survey Instrument
52
• Managerial: To what extent do you agree or disagree with each statement? e.g.,
• Those who will use the information systems are fully involved in hospital IT development
• Functional: How much is each activity supported by computerized information systems in your hospital?
• Technological: To what extent is each technology made available in your hospital?
• Integration: To what extent is information shared or transmitted among information systems within/outsideyour hospital?
Sample Questions
53
• 150-baht (~US$5) incentive if completed• Endorsed by President of the Thai Medical Informatics Association• Funded by a leading medical school with known informatics focus• Anonymous unless contact information provided for incentive & results
mailing
Survey Methodology (Nationwide)
54
• 64% response rate• Some items problematic
– Differing within-hospital responses on total & IT budgets, No. of IT staff, quality accreditation status
– Poor interrater reliability for some dimensions• Quality accreditation status dropped• Item wording revised & survey shortened• Integration sophistication items restructured
Pilot Study Findings
55
Hospital Characteristic Site 1 Site 2 Site 3 Site 4 Site 5Response rate 40% 50% 70% 70% 90%Hospital beds
Authoritative source
30 ± 0
30
120.2 ± 0.4 (120-121)
120
360 ± 0
335
303.1 ± 9.4 (282-307)
305
1,058.1 ± 187.1(863-1,500)
938PublicPrivate
100%0%
0%100%
100%0%
100%0%
100%0%
Accreditation statusNot accredited &
without planNot accredited,
with plan but no significant progress
Not accredited, with plan and significant progress
Accredited
25%
75%
0%
0%
0%
40%
40%
20%
0%
0%
0%
100%
0%
14%
86%
0%
0%
0%
0%
100%Number of IT staff
None1-56-2021-5051 or more
0%75%25%0%0%
0%80%20%0%0%
0%100%
0%0%0%
0%43%57%0%0%
0%0%
22%11%67%
Pilot Study Findings
56
Hospital Characteristic Site 1 Site 2 Site 3 Site 4 Site 52009 total budget (million baht)
22.0[n=1]
300.0[n=1]
578.0[n=1]
368.4 ± 93.7[n=3]
7,000.0 ± 1,414.2 [n=2]
2009 IT budget (million baht)
0.4[n=1]
10.0[n=1]
2.1 ± 1.6 [n=3]
5.5± 0.7[n=2]
93.0± 40.0[n=3]
Number of computers in hospital
23.8 ± 4.8 (20-30)
106.7 ± 90.2(20-200)
170.0 ± 108.9(10-300)
207.1 ±82.2
(100-290)
2,350.0 ± 1,332.3
(100-4,000)Calculated percentage of 2009 IT budget according to provided amount
1.8%[n=1]
3.3%[n=1]
0.5%[n=1]
1.4%[n=2]
1.3%[n=1]
Subjective estimated percentage of 2009 IT budget (if amount not provided above)
Below 1%1-4%5-8%Above 8%
0%75%25%0%
0%40%40%20%
20%60%20%0%
0%86%14%0%
0%17%50%33%
Pilot Study Findings
57
Construct Overall Site 1 Site 2 Site 3 Site 4 Site 5Managerial Sophistication
3.6 ± 0.4 3.2 ± 0.2 4.2 ± 0.4 3.9 ± 0.5 3.8 ± 0.4 3.2 ± 0.9
Technological Sophistication
3.5 ± 0.3 3.1 ± 0.2 3.4 ± 0.6 3.7 ± 1.0 3.4 ± 0.5 3.8 ± 0.5
Functional Sophistication
4.0 ± 0.3 3.5 ± 0.5 4.4 ± 0.4 4.1 ± 0.6 4.2 ± 0.5 4.0 ± 0.3
Integration Sophistication(Within Hospital)
3.8 ± 0.3 3.8 ± 0.3 4.2 ± 1.2 3.8 ± 0.6 3.7 ± 0.5 3.4 ± 0.7
Integration Sophistication (Outside Hospital)
2.3 ± 0.9 1.1 ± 0.04 2.5 ± 1.4 3.6 ± 0.8 2.0 ± 0.9 2.1 ± 0.5
Overall IT Sophistication
3.4 ± 0.4 2.9 ± 0.2 3.7 ± 0.7 3.9 ± 0.6 3.4 ± 0.3 3.3 ± 0.4
Pilot Study Findings
58
• IT Sophistication ItemsConstruct Intraclass
CorrelationCronbach’s
AlphaManagerial Sophistication 0.26* 0.91Technological Sophistication 0.04 0.81Functional Sophistication 0.20 0.93Integration Sophistication(Within Hospital)
0.00 0.89
Integration Sophistication(Outside Hospital)
0.50* 0.97
Overall IT Sophistication 0.30* 0.96*p < 0.05 on F-test.
Pilot Study Findings
59
• 4 of 1,302 hospitals ineligible• Response rate 69.9%
Characteristic Overall Responding Hospitals
Non-Responding
HospitalsN of eligible hospitals 1,298 908 390Bed size** 106.9 117.5 82.9Public status**
PrivatePublic
24.0%76.0%
17.4%82.6%
39.2%60.8%
Geography*CentralEastNorthNortheastSouthWest
33.4%7.5%11.1%27.1%15.3%5.6%
31.1%7.8%13.5%26.9%14.9%5.8%
39.0%6.7%5.4%27.7%16.2%5.1%
*p < 0.01, **p < 0.001.
Nationwide Findings
60
Characteristic Number of Responses Statistic†Public status
PrivatePublic
908158750
17.4%82.6%
Teaching statusNon-teachingTeaching
901716185
79.5%20.5%
Total employees 890 368.2 ± 573.5 (10-5269)IT employees 901 4.3 ± 5.3 (0-60)Total budget (million baht) 443 146.67 ± 313.60 (0.25-3,067)IT budget (million baht) 598 2.77 ± 8.79 (0-100)Ratio of IT budget to total budget‡
< 1%1-4%5-8%> 8%
4161352184023
2.7% ± 4.6% (0-43.3%)32.5%52.4%9.6%5.5%
Extent of overall IT utilizationVery lowLowModerateHighVery high
9055
35169454242
0.6%3.9%
18.7%50.2%26.7%
Total PCs in use 883 126.1 ± 218.6 (0-3,000)
Nationwide Findings
61
Estimate (Partial or Complete Adoption) NationwideBasic EHR, outpatient 86.6%Basic EHR, inpatient 50.4%Basic EHR, both settings 49.8%Comprehensive EHR, outpatient 10.6%Comprehensive EHR, inpatient 5.7%Comprehensive EHR, both settings 5.3%order entry of medications, outpatient 96.5%order entry of medications, inpatient 91.4%order entry of medications, both settings 90.2%order entry of all orders, outpatient 88.6%order entry of all orders, inpatient 81.7%order entry of all orders, both settings 79.4%
Adoption Estimates
62
Final Model
63
• High IT adoption rates• Drastic changes in adoption landscape• Local context might play a role
– Supply Side– Demand Side
• International Comparison– Relatively higher adoption
Discussion
64
• Overview of Surveys• Survey Methodology• A Sample Survey
Recap
65
• Survey Design– Study Design– Modes of Data Collection
• Instrument Design• Sampling• Survey Conduct• Data Analysis• Reports
Recap