After 6 month period, 29 experimental group members and 26 control group members stay throughout the program. The participant demographic statistics is listed below.
The body mass index (BMI) of experimental group is lower than that of control group is the beginning. Other than BMI, all other factors are the similar between 2 groups.
Table 4-1. Participant demographic statistics
Participant demographic statistics
people/(average) %(SD) people/(average) %(SD)
Age 68.68 4.868 70.42 4.178 1.61(0.21)
Years in community 29.1 16.236 28.38 15.095 0.03(0.87)
Body Mass Index(BMI) 23.24 2.572 25.08 3.752 6.07(0.02*)
Gender 2.64(0.11)
Male 10 34.5% 4 15.4%
Female 19 65.5% 22 84.6%
Education 1.09(0.78)
Elementary or less 6 20.7% 8 30.8%
Junior High 6 20.7% 5 19.2%
Senior High 10 34.5% 9 34.6%
College or above 7 24.1% 4 15.3%
Marriage 0.86(0.36)
Married 17 58.6% 12 46.2%
Others (Divored, Separate, Widowed,
Living together, Single) 12 41.4% 14 53.8%
Occupation 4.46(0.22)
Military, Government, Teacher 5 17.9% 2 8.0%
Engineer, Business 10 35.7% 7 28.0%
Need hypnotic or tranquilizer 2 7.7% 8 30.8%
Regular Exercise 0.44(0.51)
doi:10.6342/NTU201801947
18
4.1 Physical Fitness
We conducted the physical fitness focusing on 3 dimensions: body composition analysis, muscular endurance and flexibility. By looking at the pretest and posttest data in experimental group, we noticed some improvements. In the body composition analysis category, body mass index (BMI) went from 23.45 to 23.29. In muscular endurance category, 30 second chair rise jumped from 18 to 23.07. 2 min leg lifting increased from 119.48 to 137.75. In the flexibility category, back scratch test: right hand on top went from -6.83 to -4.15 while back scratch test: left hand on top also improved from -14.34 to -11.21. On the other hand, there is no statistically significant improvement in the control group in all categories.
Table 4-2. Physical Fitness Test: Pretest vs Posttest
Physical Fitness
Variable Average SD Average SD
Body Mass Index (BMI) 23.45 2.59 23.29 2.56 -2.35 (-0.29, -0.02) 0.026
Body Fat Percentage (BFP) 28.18 6.70 28.04 6.46 -0.79 (-0.51, 0.23) 0.439
Grip Strength 27.14 8.43 27.04 8.53 -0.27 (-0.92, 0.71) 0.790
30 second chair rise 18.00 7.14 23.07 6.41 6.38 (3.44, 6.69) 0.000
2 min leg lifting steps 119.48 21.69 137.75 18.13 3.87 (8.60, 27.94) 0.001 Back scratch test: right hand on top -6.83 16.62 -4.15 12.08 2.35 (0.34, 5.00) 0.026 Back scratch test: left hand on top -14.34 17.19 -11.21 11.75 2.09 (0.06, 6.21) 0.046 1. Body
Body Fat Percentage (BFP) 33.68 7.50 33.19 8.21 -0.93 (-1.57, .60) .36
Grip Strength 24.89 6.68 25.93 6.30 1.68 (- .23, 2.32) .11
30 second chair rise 20.00 7.33 22.27 7.35 2.06 (- .00, 4.54) .05
2 min leg lifting steps 127.85 28.47 132.73 23.02 1.15 (-3.90, 13.67) .26 Back scratch test: right hand on top -2.94 11.86 -2.78 10.89 0.23 (-1.34, 1.68) .82 Back scratch test: left hand on top -7.33 10.09 -6.56 9.21 1.01 (- .85, 2.41) .33 1. Body
We also did the trend analysis for the experimental group. On "30 second chair rise",
"2 min leg lifting", "back scratch test: right hand on top", "back scratch test: left hand on top" items, we found a statistically significant linear trend.
Table 4-3. Experimental group physical fitness trend analysis
Fig 4-1. Experimental group 6 month + baseline physical fitness statistics chart
Trend Analysis
Parameter Estimate Standard
Error Z Pr > |Z|
Intercept 18.2178 1.2716 15.7255 20.7102 14.33 <.0001 Time 1.1782 0.2008 0.7847 1.5717 5.87 <.0001 Intercept 122.2137 3.847 114.6738 129.7537 31.77 <.0001
Time 3.6696 1.1209 1.4727 5.8666 3.27 0.0011
Intercept -7.6896 3.2228 -14.0061 -1.3731 -2.39 0.017
Time 0.9187 0.3563 0.2204 1.6171 2.58 0.0099
Intercept -16.0818 3.6035 -23.1444 -9.0191 -4.46 <.0001
Time 1.2937 0.5258 0.2631 2.3242 2.46 0.0139
Back scratch test: right hand on top
Back scratch test: left hand on top
Analysis Of GEE Parameter Estimates
doi:10.6342/NTU201801947
20
4.2 Quality of Life
We used SF8 health survey to measure the perceived quality of life. The lower the score in the survey, the better the perceived quality of life. In the experimental group, physical component summary (PCS) showed improvement, particularly on bodily pain domain. In the control group, neither physical component summary (PCS) nor mental component summary (MCS) demonstrate changes statistically. Only vitality domain improved but not enough to lead to overall MCS change.
Table 4-4. Quality of Life SF8 Health Survey: Pretest vs Posttest (The lower the number, the better the quality of life.)
SF8 Health Survey
Variable Average SD Average SD
Physical Functioning (PF) 2.11 0.83 1.93 0.90 1.04 (- .17, .53) .31
Role Physical (RF) 1.66 0.81 1.72 0.80 -0.32 (- .51, .37) .75
Bodily Pain (BP) 2.48 1.27 1.97 0.94 2.73 ( .13, .91) .01
General Health (GH) 2.90 0.72 2.86 0.92 0.20 (- .32, .39) .85
Vitality (VT) 2.76 0.83 2.38 0.90 2.09 ( .01, .75) .05
Social Functioning (SF) 1.59 0.68 1.66 0.72 -0.47 (- .37, .23) .65
Role Emotional (RE) 1.46 0.64 1.54 0.74 -0.40 (- .44, .29) .69
Physical Functioning (PF) 1.91 0.73 2.00 0.80 -0.49 (- .45, .28) .63
Role Physical (RF) 1.96 0.96 1.71 0.91 1.30 (- .15, .65) .21
Bodily Pain (BP) 2.35 0.75 2.23 0.91 0.62 (- .27, .50) .54
General Health (GH) 2.77 0.91 3.08 0.84 -1.55 (- .72, .10) .13
Vitality (VT) 2.85 0.78 2.42 0.76 2.19 ( .02, .82) .04
Social Functioning (SF) 1.72 0.79 1.52 0.59 1.16 (- .16, .56) .26
Role Emotional (RE) 1.72 0.81 1.56 0.71 0.72 (- .30, .62) .48
4.3 Social Network
We ask the experimental group whether they know one another before and after the 6 month intervention. Before the program started, 5 participants knew no one in the group and most of them know 1 to 6 people. Only 3 participants knew 7 or more people. After 6 month intervention, the survey revealed that all of the participants have at least 4 friends and 11 in average. Most of them have 8 to 14 friends. Some (4 people) even got to know 15 to 17 friends.
Table 4-5. Number of acquaintances
Moreover, the cluster analysis chart also showed the participants' network increased noticeably after the 6 month test.
Fig 4-2. Cluster Analysis Chart
Social Network
Acquaintance Number % Acquaintance Number %
0 5 17% 0 0 0%
1~3 11 38% 4~5 3 10%
4~6 10 34% 8~11 15 52%
7 or more 3 10% 12~14 7 24%
15~17 4 14%
Experimental Group (n=29)
Pretest Posttest
After Before
doi:10.6342/NTU201801947
5.1 Reach: How do we reach our target population?
35 experimental group and 35 control group members were recruited. Some experimental group participants were referrals from local clinics in Wanhua district and NTUH (National Taiwan University Hospital) Bei-Hu Branch while some control group members were volunteers from Wanhua district health center. This may compromise the randomness of the samples. If budget and time allows, we may improve on advertising and outreach. For example, we may show up in local events to promote our program. Recruit seniors as our program team members. Utilize radio, use display in the elderly hang-out locations, such as senior community centers.
Also, when we pick seniors to join experimental group, we simply ask them to get the answers without verifying it. We were not 100% certain that they were really our target population (for example, having hyperglycemia, hyperlipidemia or hypertension). Additional efforts could be added for ensuring that the right population are selected.
Plus, at the end of 6 month period, only 29 experimental group (83%) and 26 control group members (74%) still stayed with the program. This showed that we have room for improvement in terms of participant retention. The attrition (churn) rate is especially high in control group. This might have something to do with the fact that the program lasted 6 long months. If we could also conduct regular monthly physical test on control group, it might potentially reduce the churn rate.
5.2 Effectiveness or Efficacy: How do we know our intervention is effective?
Firstly, the physical condition of participants did improve. All categories (body composition analysis, muscular endurance, and flexibility) in the physical fitness test of experimental group showed improvement comparing posttest and pretest data. To