Transcript Presentation_Moses_L..

HEALTH INFORMATION SYSTEMS FOR
DECISION MAKING
by
Moses Lemayian
Health Informatics
Data for decision Making
• Florence Nightingale invented polar-area diagrams in 1855 (below) to
show that many army deaths could be traced to unsanitary clinical
practises and were therefore preventable. She used the diagrams to
convince policy-makers to implement reforms that eventually reduced the
number of deaths
Source: (Audain 2007). (Diagram from Nightingale 1858.)
Problem statement
• Information explosion: the amount of
electronic data gathered is enormous
In fact, some experts believe that medical
breakthroughs have slowed down, attributing
this to the prohibitive scale and complexity of
present-day medical information. Computers
and data mining are best-suited for this
purpose. (Shillabeer and Roddick 2007).
data mining in the health sector
• Early detection and/or prevention of diseases.
Cheng, et al (2006) cited the use of classification
algorithms to help in the early detection of heart
disease, a major public health concern all over
the world.
• Cao et al (2008) described the use of data mining
as a tool to aid in monitoring trends in the clinical
trials of cancer vaccines. By using data mining and
visualization, medical experts could find patterns
and anomalies better than just looking at a set of
tabulated data.
Table 1. Drug
Table 2. Diet
Sr_no
Age
N
Small_n
Percentage
SE
Sr_no
Age
N
Small_n
Percentage
SE
1
15 – 24
10
3
32.2
16.2
2
25 – 34
19
6
29.9
11.5
3
35 – 44
35
23
64.3
8.8
4
45 – 54
77
62
81.3
4.8
5
55 – 64
99
90
90.8
2.6
1
2
3
4
5
15 – 24
25 – 34
35 – 44
45 – 54
55 – 64
10
19
35
77
99
3
2
21
45
52
19.7
10.8
60.4
58.7
53
13
6.8
9.5
6.6
8.5
Table 4. Smoke cession
Table 3. Weight
Sr_no
Age
N
Small_n
Percentage
SE
Sr_no
Age
N
Small_n
Percentage
SE
1
15 – 24
10
3
32.2
16.2
1
15 – 24
10
2
19.7
13
2
25 – 34
19
2
10.8
7.3
2
25 – 34
19
5
27
10.4
3
35 – 44
35
4
11.7
5.5
3
35 – 44
35
17
48.5
9.1
4
45 – 54
77
13
16.7
4.4
4
45 – 54
77
26
33.4
5.3
5
55 – 64
99
32
13.3
2.5
5
55 – 64
99
39
39.9
7.8
‘sr_no’ = serial number, (unique id -
Table 5. Exercise
Sr_no
Age
N
Small_n
Percentage
SE
1
15 – 24
10
3
32.2
16.2
2
25 – 34
19
6
33.3
10.2
3
35 – 44
35
13
36.9
7.9
4
45 – 54
77
23
29.9
5.6
5
55 – 64
99
28
27.9
5.4
Source: Abdulaziz et. al. (2010)
Data: http://www.who.int/research/en/
primary key),
‘age’ = age of patients,
‘N’ = total number of patient of each age
group,
‘small_n’ = number of patients who have
been cured with the particular type of
treatment,
percentage = percent of cured patients by
specific mode of treatment, and ‘SE’ =
Standard error.
Table 6. Comparison on predictions
Treatment
p(Y)
p(O)
Comparison of p(O)
with p(Y)
Drug
Diet
Weight
Smoke cession
Exercise
–50.616
36.4803
32.1654
12.9883
48.5004
10.1015
65.8054
61.0199
18.1215
49.0474
P(O) > p(Y)
P(O) > p(Y)
P(O) > p(Y)
P(O) > p(Y)
P(O) = p(Y) {Approx
equal}
CHALLENGES
Even if data mining results are credible, convincing the
health practitioners to change their habits based on
evidence may be a bigger problem. Ayres (2008)
Shillabeer (2009) also reported most doctors (at least in
Australia) prefer to listen to a respected opinion leader in
the medical profession, rather than to the result of data
mining.
END
THANK YOU