This paper presents primary outcome data from 9,372 Queensland adults who participated in the My health for life program from July 2017 to December 2019. Tables 1 and 2 presents baseline study modality, and socio-demographic characteristics by healthy lifestyle indices grouped into quintiles (Quintile 1 represents unhealthy lifestyle behaviours; Quintile 5 represents greatest number of healthy lifestyle behaviours).
Bivariate comparisons of the healthy lifestyle index showed that healthy lifestyle was associated with age (45 years or older; χ2(4) = 285.15, p < .01), gender (female; χ2(4) = 22.34, p < .01), retirement (χ2(16) = 328.41, p < .01), higher educational attainment (χ2(16) = 79.10, p < .01), and greater relative advantage (IRSAD Quintiles 4 and 5; χ2(16) = 124.93, p < .01). Socio-demographic characteristics by HLI quintile are outlined in Table 1.
Table 1
Baseline characteristics by healthy lifestyle index (HLI) quintiles a
| Quintile 1 | Quintile 2 | Quintile 3 | Quintile 4 | Quintile 5 | Total |
| n (%) | n (%) | n (%) | n (%) | n (%) | n (%) |
Mode | | | | | | |
THC | 74 (8.1) | 126 (8.8) | 78 (8.3) | 80 (8.9) | 143 (8.0) | 501 (8.4) |
GBP | 841 (91.9) | 1,302 (91.2) | 858 (91.7) | 819 (91.1) | 1,645 (92.0) | 5,465 (91.6) |
Employment status | | | | | | |
Employed | 519 (60.0) | 719 (52.9) | 426 (47.1) | 414 (48.1) | 747 (43.6) | 2,825 (49.6)* |
Home duties | 53 (6.1) | 86 (6.3) | 52 (5.8) | 37 (4.3) | 72 (4.2) | 300 (5.3) |
Retired | 105 (12.1) | 333 (24.5) | 307 (34.0) | 320 (37.2) | 728 (42.5) | 1,793 (31.5) |
Not working | 111 (12.8) | 117 (8.6) | 64 (7.1) | 43 (5.0) | 78 (4.6) | 413 (7.2) |
Other | 77 (8.9) | 105 (7.7) | 55 (6.1) | 46 (5.3) | 87 (5.1) | 370 (6.5) |
Gender | | | | | | |
Female | 656 (71.7) | 1,067 (74.7) | 739 (79.0) | 715 (79.5) | 1,380 (77.2) | 4,557 (76.4)* |
Male | 259 (28.3) | 361 (25.3) | 197 (21.0) | 184 (20.5) | 408 (22.8) | 1,409 (23.6) |
Age bracket | | | | | | |
<45 years | 265 (29.0) | 215 (15.1) | 97 (10.4) | 69 (7.7) | 135 (7.6) | 781 (13.1)* |
45 or older | 649 (71.0) | 1,211 (84.9) | 838 (89.6) | 830 (92.3) | 1,651 (92.4) | 5,179 (86.9) |
First Nations People | | | | | | |
No | 825 (90.2) | 1,373 (96.1) | 898 (95.9) | 879 (97.8) | 1,753 (98.0) | 5,728 (96.0)* |
Yes | 90 (9.8) | 55 (3.9) | 38 (4.1) | 20 (2.2) | 35 (2.0) | 238 (4.0) |
Educational attainment | | | | | | |
Primary education | 28 (3.1) | 43 (3.1) | 38 (4.1) | 28 (3.1) | 51 (2.9) | 188 (3.2)* |
Secondary education | 321 (35.6) | 462 (32.9) | 302 (32.9) | 281 (31.6) | 522 (29.8) | 1,888 (32.2) |
Certificate/diploma | 381 (42.3) | 534 (38.0) | 324 (35.3) | 322 (36.2) | 584 (33.3) | 2,145 (36.6) |
Bachelor/postgraduate | 157 (17.4) | 336 (23.9) | 241 (26.2) | 239 (26.9) | 562 (32.1) | 1,535 (26.2) |
Other | 14 (1.6) | 31 (2.2) | 14 (1.5) | 19 (2.1) | 34 (1.9) | 112 (1.9) |
CALD | | | | | | |
No | 899 (98.3) | 1,383 (96.8) | 910 (97.2) | 876 (97.4) | 1,688 (94.4) | 5,756 (96.5) |
Yes | 16 (1.7) | 45 (3.2) | 26 (2.8) | 23 (2.6) | 100 (5.6) | 210 (3.5) |
IRSAD quintile | | | | | | |
Quintile 1 | 136 (14.9) | 190 (13.3) | 113 (12.1) | 92 (10.2) | 154 (8.6) | 685 (11.5)* |
Quintile 2 | 220 (24.1) | 261 (18.3) | 148 (15.8) | 133 (14.8) | 229 (12.8) | 991 (16.6) |
Quintile 3 | 200 (21.9) | 321 (22.5) | 230 (24.6) | 201 (22.4) | 417 (23.3) | 1,369 (23.0) |
Quintile 4 | 184 (20.1) | 310 (21.7) | 188 (20.1) | 211 (23.5) | 441 (24.7) | 1,334 (22.4) |
Quintile 5 | 174 (19.0) | 346 (24.2) | 255 (27.3) | 262 (29.1) | 546 (30.6) | 1,583 (26.6) |
THC, Telephone health couching; GBP, Group-based program; CALD, Culturally or Linguistically Diverse; IRSAD, Index of Relative Socio-economic Advantage and Disadvantage. |
a Highest quintile represents greatest number of healthy lifestyle indices while the lowest represents most unhealthy lifestyle behaviours |
* p < .01 |
Overall, three-quarters of participants were female, most were aged 45 years or older (> 80%), around two-thirds reported a secondary school or certificate/diploma level education, and half were employed outside the home. Some modest but statistically significant differences were noted with attrition highest in men (χ2(4) = 16.41, p < .01) aged 45 years or less (χ2(2) = 67.36, p < .01) with primary or secondary school education (χ2(8) = 16.93, p = .03).
Table 2
Percentage of healthy behaviours among complete cases at Sessions 1, 5 and 6
| Session 1 | Session 5 | Session 6 |
| n (%) | n (%) | n (%) |
Diet index |
Daily fruit intake a | | | |
<2 serves | 5,032 (53.7) | 1,716 (29.2) | 1,097 (26.5)* |
2 or more serves | 4,340 (46.3) | 4,168 (70.8) | 3,047 (73.5) |
Daily veg. intake a | | | |
<5 serves | 8,447 (90.1) | 4,522 (76.8) | 3,081 (74.3)* |
5 or more serves | 925 (9.9) | 1,363 (23.2) | 1,064 (25.7) |
Sugar-sweetened drinks | | | |
More than weekly | 1,531 (16.3) | 559 (9.1) | 410 (8.2)* |
Once a week | 1,071 (11.4) | 715 (11.6) | 497 (9.9) |
Less than weekly | 6,770 (72.2) | 4,869 (79.3) | 4,103 (81.9) |
Takeaway | | | |
More than weekly | 29 (0.3) | 14 (0.2) | 6 (0.1)* |
Once a week | 3,289 (35.1) | 1,416 (23.0) | 993 (19.8) |
Less than weekly | 6,054 (64.6) | 4,726 (76.8) | 4,014 (80.1) |
Physical activity index |
Physical activity b | | | |
Sedentary | 1,803 (19.2) | 383 (6.5) | 549 (10.9)* |
Insufficient for health | 4,370 (46.6) | 2,104 (35.5) | 1,802 (35.8) |
Sufficient for health | 3,199 (34.1) | 3,433 (58.0) | 2,687 (53.3) |
Alcohol and smoking index |
Alcohol quantity | | | |
5 or more | 3,574 (38.1) | 2,371 (38.5) | 1,913 (38.2)* |
1–4 drinks | 1,584 (16.9) | 943 (15.3) | 650 (13.0) |
None | 4,214 (45.0) | 2,840 (46.1) | 2,450 (48.9) |
Alcohol frequency | | | |
Daily | 203 (2.2) | 58 (0.9) | 45 (0.9)* |
Weekly or less | 9,169 (97.8) | 6,096 (99.1) | 4,967 (99.1) |
Smoking status | | | |
Current | 752 (8.0) | 348 (3.9) | 301 (3.3)* |
Former | 2,248 (24.0) | 2,334 (25.9) | 2,350 (26.1) |
Never | 6,372 (68.0) | 6,339 (70.3) | 6,362 (70.6) |
Weight index |
Body mass index | | | |
≥ 40 kg/m² | 1,011 (16.0) | 710 (14.6) | 426 (13.0)* |
30–39.9 kg/m² | 3,134 (49.5) | 2,409 (49.4) | 1,515 (46.4) |
25–29.9 kg/m² | 1,621 (25.6) | 1,318 (27.0) | 942 (28.8) |
< 25 kg/m² | 569 (9.0) | 436 (8.9) | 384 (11.8) |
Waist circumference c | | | |
Greater increased risk | 5,154 (84.8) | 3,892 (83.3) | 2,357 (78.4)* |
Increased risk | 612 (10.1) | 535 (11.4) | 417 (13.9) |
Normal | 314 (5.2) | 248 (5.3) | 234 (7.8) |
a Current dietary guidelines recommend a minimum of 2 fruit per day and 5 serves of vegetables [27] |
b Physical activity was defined according to the Australian Physical Activity Guidelines [28] denoting the accumulation of at least 150 minutes of activity over one week |
c Waist circumference (WC) ≥ 94 cm in men and WC ≥ 80 cm in women denoted increased risk and WC ≥ 102 cm in men and WC ≥ 88 cm in women denoted greater increased risk |
* p < .01 |
Table 2 presents the descriptive health behaviours for complete cases at Sessions 1, 5 and 6. The proportion of participants consuming recommended daily serves of fruit (Session 1, 46.3%; Session 5, 70.8%; Session 6, 73.5%, p < .01) and vegetables increased over time (Session 1, 9.9%; Session 5, 23.2%; Session 6, 25.7%, p < .01) while the frequency takeaways also decreased. Similarly, the proportion of participants who were sufficiently active for health according to the Australian Physical Activity Guidelines [28] increased from 34.1% at Session 1 to 53.3% at Session 6. Risky alcohol intake (i.e., daily drinking or having more than 4 standard drinks on any one day [29]) was largely unchanged over the program period though current cigarette smoking decreased from 8.0% at Session 1 to 3.3% at Session 6 (p < .01 but percentage differences < 10% [31]). Finally, there were small incremental decreases in BMI and WC with increased tendency towards healthy body mass (Session 1, 9.0%; Session 6, 11.8%, p < .01) and normal WC (Session 1, 5.2%; Session 6, 7.8%, p < .01).
However, while there were general trends towards healthy lifestyle behaviours over the program period, significant attrition might have influenced prevalence and therefore data were imputed. To assess the robustness of imputation, the original and imputed healthy lifestyle indices summary statistics are provided. Table 3 shows the point estimates for the HLI (range 0–18) which did not change at each time point with the average HLI at Session 1 being 10.9 (SD = 2.5), 11.2 (SD = 2.4) at Session 5 and 11.6 (SD = 2.4) at Session 6.
Table 3
Summary statistics for the original and imputed healthy lifestyle indices (with and without weight)
| Session 1 | Session 5 | Session 6 |
| Original | Imputed | Original | Imputed | Original | Imputed |
Healthy lifestyle index a | | | | | | |
n | 5,858 | 202,808 | 3,928 | 126,828 | 2,568 | 132,618 |
M(SD) | 10.1 (2.5) | 10.0 (2.5) | 11.2 (2.4) | 11.2 (2.4) | 11.4 (2.4) | 11.6 (2.4) |
Median [IQR] | 10.0 [8.0, 12.0] | 10.0 [8.3, 11.7] | 11.0 [10.0, 13.0] | 11.3 [9.6, 12.9] | 11.0 [10.0, 13.0] | 11.7 [10.0, 13.2] |
Minimum | 1 | -1 | 3 | 1 | 3 | -1 |
Maximum | 18 | 22 | 18 | 21 | 18 | 22 |
a Healthy lifestyle index computed as the sum of dietary, physical activity, alcohol and smoking and weight indices |
The results of Gaussian Generalized Estimating Equations which incrementally adjusted for program characteristics (Model 2) and personal background (Model 3) are shown in Table 4. Over the program period, the average HLI increased by around 1-point at Session 5 (Model 1: β = 0.93, 95% CI = 0.87, 1.00, p < .01; Model 2: β = 0.93, 95% CI = 0.87, 1.00, p < .01; Model 3: β = 0.94, 95% CI = 0.87, 1.01, p < .01) and this was sustained at Session 6 (Model 1: β = 1.29, 95% CI = 1.10, 1.29, p < .01; Model 2: β = 1.19, 95% CI = 1.10, 1.28, p < .01; Model 3 β = 1.18, 95% CI = 1.09, 1.28, p < .01).
Model 2 examined the additive effect of program characteristics. In Models 2 and 3, adjustment suggested that neither study modality (THC vs. GBP), nor number of sessions attended, significantly influenced HLI scores (p > .01). Model 3 also adjusted for background socio-demographic factors. Findings showed being retired (β = 0.73, 95% CI = 0.61, 0.85, p < .01), aged 45 or older (β = 0.97, 95% CI = 0.81, 1.13, p < .01), and having a certificate or diploma (β = 0.47, 95% CI = 0.19, 0.76, p < .01) or bachelor’s degree or higher (β = 1.05, 95% CI = 0.76, 1.34, p < .01) conferred a higher average HLI while being male, Aboriginal or Torres Strait Islander background, or not currently working conferred lower average HLI scores (p < .01 for all).
Table 4
Longitudinal modelling of a HLI using Gaussian Generalized Estimating Equations with an Exchangeable structure
| Model 1 a | Model 2 b | Model 3 c |
| β (95% CI) | β (95% CI) | β (95% CI) |
Constant | 10.23 (10.16, 10.30)* | 9.88 (9.43, 10.32)* | 8.17 (7.64, 8.70)* |
Sessions | | | |
Session 1 | - | - | - |
Session 5 | 0.93 (0.87, 1.00)* | 0.93 (0.87, 1.00)* | 0.94 (0.87, 1.01)* |
Session 6 | 1.20 (1.10, 1.29)* | 1.19 (1.10, 1.28)* | 1.18 (1.09, 1.28)* |
Mode | | | |
THC | | - | - |
GBP | | -0.07 (-0.19, 0.06) | -0.14 (-0.27, -0.02) |
No. sessions (range 1–6) | | 0.09 (0.00, 0.18) | 0.06 (-0.03, 0.15) |
Employment status | | | |
Employed | | | - |
Home duties | | | -0.12 (-0.36, 0.12) |
Retired | | | 0.73 (0.61, 0.85)* |
Not working | | | -0.54 (-0.75, -0.33)* |
Other | | | -0.29 (-0.50, -0.08)* |
Gender | | | |
Female | | | - |
Male | | | -0.23 (-0.35, -0.11)* |
Age bracket | | | |
<45 years | | | - |
45 or older | | | 0.97 (0.81, 1.13)* |
Educational attainment | | | |
Primary education | | - |
Secondary education | | | 0.26 (-0.02, 0.55) |
Certificate/diploma | | | 0.47 (0.19, 0.76)* |
Bachelor/postgraduate | | | 1.05 (0.76, 1.34)* |
Other | | | 0.59 (0.13, 1.05) |
First Nations People | | | |
No | | | - |
Yes | | | -0.55 (-0.84, -0.27)* |
IRSAD quintile | | | |
Quintile 1 | | - |
Quintile 2 | | | -0.02 (-0.19, 0.16) |
Quintile 3 | | | 0.37 (0.20, 0.54)* |
Quintile 4 | | | 0.53 (0.35, 0.71)* |
Quintile 5 | | | 0.68 (0.51, 0.85)* |
THC, Telephone health couching; GBP, Group-based program; CALD, Culturally or Linguistically Diverse; IRSAD, Index of Relative Socio-economic Advantage and Disadvantage. |
a Model 1, unadjusted relationship between HLI and time (sessions 5 and 6) |
b Model 2, adjusted for program characteristics (delivery mode and no. sessions attended |
c Model 3, adjusted for program characteristics and personal background (employment status, gender, age bracket, educational attainment, First Nations People, and IRSAD quintile) |
* p < .01 |
To assess the changes of each health behaviour individually, the predicted probabilities for each health behaviour were estimated using nominal logistic models, with results showing consistent trends towards healthier lifestyle behaviours over the program period. Overall, obesity decreased from 65% at Session 1 to 57% at Session 6, while the proportion of people in the normal weight range increased from 9–13%. Similar trends were noted for waist circumference, though to a lesser extent. Dietary indices also showed a shift towards recommended dietary guidelines with 25% meeting the guidelines for daily vegetable intake, 70% meeting the guidelines for daily fruit intake, and ~ 82% consuming take-away and sugar sweetened drinks less than weekly.
At baseline, 19% of participants reported being sedentary and 46% were insufficiently active for health. Over the program period, the proportion of people meeting physical activity guidelines increased, though at Session 6, only 53% reported being sufficiently active for health. Finally, few participants consumed alcohol daily (< 1%) though around one-third (37%) of participants consumed an average of 5 or more alcoholic drinks in one session and this was largely unchanged over the program period. Percentage changes using predicted probabilities in individual health behaviours from Session 1 to Session 6 are illustrated in Fig. 1.