The ethics approval for the use of the data in this study was received from the ethics committee of the Northern Savonia Hospital District on 13th November 2012. To preserve the patient safety, individual recognizable data were not uncovered by any means to us. Information was provided only for those variables which were needed for this study rather than access to the entire dataset.
5 RESULTS
Table 4 represents the general characteristics of the study population. There are total 9288 people aged more than 20, included in our study. Among them, 47% were female and 53% were male.
Age of the population ranged from 21 to 99 years and most of the people were in between the age of 60 to 69 years (33.7 %, n = 3132). The mean age of the population was 67 (F = 69, M = 65) years.
Table 4: General characteristics of the population by gender and age group.
Age group Frequency (%) Total (%)
F M N
20 to 39 76 (1.7) 76 (1.5) 152 (3.2)
40 to 49 201 (4.6) 317 (6.4) 518 (11)
50 to 59 660 (15.1) 975 (19.8) 1635 (34.9)
60 to 69 1211 (27.7) 1921 (39.0) 3132 (33.7)
70 to 79 1229 (28.1) 1165 (23.7) 2394 (25.7)
80 to 99 989 (22.7) 468 (9.5) 1457 (15.2)
Total population 4366 (47) 4922 (53) 9288 (100)
Mean age 69 65 67
Minimum age 21 21 21
Maximum age 99 98 99
HbA1c measurement rate and management of HbA1c by gender between different age groups is presented in Table 5. A Chi-square test showed that there are statistically significant differences in HbA1c measurement rate between the age groups and the differences were statistically significant both for females (P < 0.001 in 12 and P < 0.001 in 2013-14) and males (P < 0.001 in 2011-12 and P < 0.001 in 2013-14). The best measurement rate of HbA1c was seen among the age group of 70-79 years, both in 2011-12 and 2013-14.
Table 5: Proportion of HbA1c measurement and management (2011-12 & 2013-14) by gender and age group.
(+) Included participants aged ≥ 20 and whose HbA1c measured both years. N = 6707 (F = 3225, M = 3482) Univariate analysis of variance:
Age adjusted gender difference in HbA1c measurement rate in 2011-12 (P = 0.093), and 2013-14 (P = 0.122) Age adjusted gender difference in the management of HbA1c in 2011-12 (P = 0.008), and 2013-14 (P = 0.147) One sample t-test:
Difference in HbA1c measurement rate between 2011-12 and 2013-14 (F < 0.001, M < 0.001) Difference in the management of HbA1c between 2011-12 and 2013-14 (F = 0.006, M = 0.008)
Age adjusted univariate analysis of variance was performed to observe the gender difference of HbA1c measurement rate in 2011-12 and 2013-14 (Table 5). We found that there was no difference in HbA1c measurement rate between females and males in 2011-12 (F = 79.3%, M = 76.1%, P = 0.093) or 2013-14 (F = 89.4%, M = 87.4%, P = 0.122).
It was observed that the measurement rate of HbA1c has been improved in 2013-14 compared with 2011-12 both among females (89.4% vs 79.3%, P < 0.001) and males (87.4% vs 76.1%, P < 0.001).
To observe the differences in HbA1c management between the age groups, a Chi-square test was performed (Table 5). We found that the differences were statistically significant both for females (P < 0.001 in 2011-12 and P = 0.001 in 2013-14) and males (P < 0.001 in 2011-12 and P < 0.001 in 2013-14). The best management rate was observed among the age group of 60-69 years in 2011-12 and 20-39 years in 2013-14.
Age adjusted univariate analysis of variance was performed to observe the gender difference in the management of HbA1c in 2011-12 and 2013-14 (Table 5). We found that in 2011-12 females had better management compared with males (F = 73.6%, M = 70.9%, P = 0.008) but in 2013-14 there was no difference in the management of HbA1c between females and males (F = 66.1%, M = 64.4%, P = 0.147).
It was observed that the management of HbA1c in this patient cohort deteriorated in 2013-14 compared with 2011-12 both among females (66.1% vs 73.6%, P = 0.006) and males (64.6% vs 70.9%, P = 0.008) (Table 5).
LDL measurement rate and management among different age groups from the year 2011-12 to 2013-14 is presented in Table 6. Both LDL measurement rate and management appears to have enhanced by the subsequent years (Table 6). A Chi-square test showed that there are statistically significant differences in LDL measurement rate between the age groups and the differences were statistically significant both for females (P < 0.001 in 2011-12 and P < 0.001 in 2013-14) and males (P < 0.001 in 2011-12 and P < 0.001 in 2013-14). The best measurement rate of LDL was seen among the age group of 70-79 years both in 2011-12 and 2013-14.
Table 6: Proportion of LDL measurement and management (2011-12 & 2013-14) by gender and age group.
(+) included participants aged ≥ 20 and whose LDL measured both years. N = 6155 (F = 2890, M = 3265) Univariate analysis of variance:
Age adjusted gender difference in LDL measurement rate in 2011-12 (P = 0.983), and 2013-14 (P = 0.064) Age adjusted gender difference in the management of LDL in 2011-12 (P < 0.001), and 2013-14 (P < 0.001) One sample t-test:
Difference in LDL measurement rate between 2011-12 and 2013-14 (F < 0.001, M < 0.001) Difference in the management of LDL between 2011-12 and 2013-14 (F = 0.409, M < 0.001)
Age adjusted univariate analysis of variance was performed to observe the gender difference of LDL measurement rate in 2011-12 and 2013-14 (Table 6). We found that there was no difference in LDL measurement rate between females and males in 2011-12 (F = 74.3%, M = 73.6%, P = 0.983) or in 2013-14 (M = 84.3%, F = 82.5%, P = 0.064).
It was observed that the measurement rate of LDL has been improved in 2013-14 compared with 2011-12 both among females (82.5% vs 74.3%, P < 0.001) and males (84.3% vs 73.6%, P < 0.001).
To observe the differences in the management of LDL between the age groups, a Chi-square tests was performed (Table 6). We found that the differences were statistically significant both for females (P < 0.001 in 2011-12 and P = 0.001 in 2013-14) and males (P < 0.001 in 2011-12 and P
< 0.001 in 2013-14). The best LDL management rate was observed among the age group of 70-79 years in 2011-12 and 2013-14.
Age adjusted univariate analysis of variance was performed to observe the gender difference in the management of LDL in 2011-12 and 2013-14. We found that both in 2011-12 (F = 51.3%, M = 57.3%, P < 0.001) and 2013-14 (F = 53.6%, M = 61.9%, P < 0.001) management of LDL was better in males compared with females (Table 6).
It was observed that the management of LDL has improved in 2013-14 compared with 2011-12 among males (61.9% vs 57.2%, P < 0.001), but there was no statistically significant improvement in LDL management among females (53.6% vs 51.3%, P = 0.409) (Table 6).
Mean HbA1c and LDL levels by gender and different age groups in 2011-12 and 2013-14 are presented in Table 7. Analysis of variance (ANOVA) test was performed to observe the differences in mean HbA1c and LDL levels between the age groups and we found that the differences were statistically significant both for females (P = 0.003) and males (P < 0.001).
Age adjusted univariate analysis of variance was performed to observe the gender difference in mean HbA1c and LDL level in 2011-12 and 2013-14 (Table 7). We found that in 2011-12 mean HbA1c was almost similar among females and males (F = 6.60%, M = 6.66%, P = 0.310) and mean LDL level was less satisfactory in females compared with males (F = 2.58 mmol/l, M = 2.41 mmol/l, P < 0.001).
Table 7: Mean HbA1c and LDL level (2011-12 & 2013-14) by gender and age group. Included participants aged ≥ 20 and whose HbA1c or LDL measured both years.
Univariate analysis of variance:
Age adjusted gender difference in mean HbA1c level of 2011-12 (P = 0.310) and 2013-14 (P = 0.104) Age adjusted gender difference in mean LDL level of 2011-12 (P < 0.001) and 2013-14 (P < 0.001) One sample t-test:
Difference in mean HbA1c level between 2011-12 and 2013-14 (F = 0.046, M = 0.016) Difference in mean LDL level between 2011-12 and 2013-14 (F = 0.863, M = 0.009)
Table 8: HbA1c measurement and management (2011-12 & 2013-14) by area level education.
(+) Included participants aged ≥ 20 and whose HbA1c measured both years. N = 6707 (F = 3225, M = 3482) Table 9: LDL measurement and management (2011-12 & 2013-14) by area level education.
Area level
(+) Included participants aged ≥ 20 and whose LDL measured both years. N = 5880 (F = 2778, M = 3102), missing data (n = 275)
In 2013-14 there was no difference in mean HbA1c levels between females and males (F = 6.82%, M = 6.86%, P = 0.104) and LDL level was higher in females compared with males (F = 2.54 mmol/l, M = 2.35 mmol/l, P < 0.001) (Table 7). It was observed that the mean HbA1c level deteriorated in 2013-14 from 2011-12 males (6.86% vs 6.66%, P = 0.016) (Table 7). The mean LDL level improved in 2013-14 from 2011-12 among males (2.35 mmol/l vs 2.4 mmol/l, P = 0.009) but remained almost the same among females (2.54 mmol/l vs 2.58 mmol/l, P = 0.863) (Table 7).
HbA1c measurement rate and achievement in the target level of HbA1c according to the proportion of educated citizens in the postal code area (at least high school graduate or vocational training) is presented in Table 8. A Chi-square test showed that there is statistically significant difference in HbA1c measurement rate between the postal code areas by the proportion of educated citizens and the differences were statistically significant for females (P < 0.001) but non-significant for males (P = 0.106) in 2011-12, in 2013-14 result was significant for both females (P = 0.029) and males (P < 0.001). The best measurement rate of HbA1c was seen in the postal code area where the proportion of educated citizens is 60 - 69.9 % both in 2011-12 and 2013-14.
To observe the differences in the management of HbA1c between the postal code areas by the proportion of educated citizens, a Chi-square tests was performed and we found that the differences were statistically significant for females (P = 0.009 in 2011-12 and P = 0.002 in 2013-14) in both years but for males the differences were significant only in 2011-12 (P = 0.004). The best management rate was observed in the postal code area where the proportion of educated citizens is more than or equals to 70% both in 2011-12 and 2013-14 (Table 8).
LDL measurement rate and management according to the proportion of educated citizens in the postal code area (at least high school graduate or vocational training) from the year 2011-12 to 2013-14 is presented in Table 9. A Chi-square test showed that there are statistically significant differences in LDL measurement rate between the postal code areas by the proportion of educated citizens and the differences were statistically significant for females only in 2011-12 (P = 0.004) but for males these differences were significant in both years (P = 0.013 in 2011-12 and P < 0.001 in 2013-14). The best measurement rate of LDL was seen in the postal code area where the proportion of educated citizens is 60 - 69.9% both in 2011-12 and 2013-14.
Table 10: Mean HbA1c and LDL level (2011-12 & 2013-14) by area level education.
Included participants aged ≥ 20 and whose HbA1c or LDL measured both years.
Table 11: HbA1c measurement and management (2011-12 & 2013-14) by area level unemployment.
Area level
(+) Included participants aged ≥ 20 and whose HbA1c measured both years. N = 6623 (F = 3193, M = 3430), missing data (n = 84)
To observe the differences in the management of LDL between the postal code areas by the proportion of educated citizen’s, a Chi-square tests was performed and we found that the differences were statistically significant only for males (P = 0.020) in 2011-12 and in 2013-14 there was no significant differences among females (P = 0.881) or males (P = 0.681) (Table 9).
The mean HbA1c and LDL level in postal code areas by the proportion of educated citizen’s in 2011-12 and 2013-14 are presented in Table 10. Analysis of variance (ANOVA) test was performed to observe the differences in mean HbA1c and LDL levels and we found that there are statistically significant differences in mean HbA1c in 2011-12 between the postal code areas with different education level for females (P = 0.001) and males (P < 0.001). Similar result was observed in mean HbA1c in 2013-14 (F < 0.001, M = 0.002). For the mean LDL, the differences between the areas were not statistically significant in 2011-12 or in 2013-14 neither among females (P = 0.511 in 2011-12 and P = 0.917 in 2013-14) nor males (P = 0.135 in 2011-12 and P = 0.864 in 2013-14).
HbA1c measurement rate and achievement in the target level of HbA1c according to the percentage of unemployed citizens in the postal code area is presented in Table 11. A Chi-square test showed that there is statistically significant difference in HbA1c measurement rate between the postal code areas by the percentage of unemployed and the differences were statistically significant for both females (P = 0.005) and males (P < 0.001) in 2011-12, but in 2013-14 significant differences was seen only among males (P = 0.004). The best measurement rate of HbA1c was seen in the postal code areas where the proportion of unemployed citizens is < 6.0 % in 2011-12 and in 2013-14 measurement rate was the highest among those whose postal code area level unemployment rate is 6.0% - 6.9%.
To observe the differences in the management of HbA1c between the postal code areas by the percentage of unemployed, a Chi-square tests was performed and we found that the differences were not statistically significant among females (P = 0.121 in 2011-12 and P = 0.153 in 2013-14) or males (P = 0.526 in 2011-12 and P = 0.576 in 2013-14) (Table 11).
Table 12: LDL measurement and management (2011-12 & 2013-14) by area level unemployment.
(+) Included participants aged ≥ 20 and whose LDL measured both years. N = 6073 (F = 2860, M = 3213), missing data (n = 82)
Table 13: Mean HbA1c and LDL level (2011-12 & 2013-14) by area level unemployment.
Included participants aged ≥ 20 and whose HbA1c & LDL measured both years.
Table 14: HbA1c measurement and management (2011-12 & 2013-14) by area level median income of the citizens.
Area level median income (€) of citizens
HbA1c Measured (2011-12) *
(%)
HbA1c Measured (2013-14) *
(%)
HbA1c Measured both year *
(%)
HbA1c management (2011-12) +
(%)
HbA1c management (2013-14) +
(%)
Female Male Female Male
F M F M F M ˂ 7 7- 8.9 ≥ 9 ˂ 7 7- 8.9 ≥ 9 < 7 7- 8.9 ≥ 9 < 7 7- 8.9 ≥ 9
≤ 15000 75.9 73.4 87.2 88.1 68.8 68.3 72.2 22.1 5.8 68.4 23.9 7.8 64.7 28.6 6.7 67.2 25.5 7.3 15001 - 16000 81.4 79.6 91.5 89.0 76.8 75.0 72.9 22.2 4.9 71.4 22.3 6.3 65.9 26.1 8.0 64.7 27.8 7.5 16001 - 17000 80.8 73.5 88.8 87.5 74.9 68.4 72.7 22.2 5.1 69.4 24.8 5.8 63.5 28.6 8.0 60.6 31.2 8.3 17001 - 19000 78.1 69.2 89.3 82.3 72.1 61.8 77.2 19.0 3.8 73.5 21.6 4.9 70.6 24.1 5.4 66.8 26.5 6.7
≥19001 77.5 78.3 88.3 86.1 72.8 72.3 76.0 18.7 5.3 73.3 21.0 5.7 69.0 25.9 5.1 65.1 28.3 6.7 Total 79.3 76.1 89.5 87.4 73.9 70.8 73.7 21.3 5.0 71.0 22.7 6.3 66.2 26.7 7.0 64.7 27.9 7.4 Chi square
P value 0.013 < 0.001 0.014 0.002 0.001 < 0.001 0.614 0.451 0.188 0.504 (*) Included participants aged ≥ 20. N = 9165 (F = 4321, M = 4844)
(+) Included participants aged ≥ 20 and whose HbA1c measured both years. N = 6623 (F = 3193, M = 3430), missing data (n = 84)
The measurement rate and management of LDL according to the percentage of unemployed in the postal code area from the year 2011-12 to 2013-14 is presented in Table 12. A Chi-square test showed that there are statistically significant differences in LDL measurement rate between the postal code areas by the percentage of unemployed and the differences were statistically significant both for females (P = 0.040 in 12 and P = 0.001 in 2013-14) and males (P = 0.006 in 2011-12 and P = 0.008 in 2013-14) (Table 2011-12). To observe the differences in management of LDL between the postal code areas by the percentage of unemployed, a Chi-square tests was performed and we found that the differences were statistically significant for females (P = 0.009 in 2011-12 and P = 0.040 in 2013-14), and males (P = 0.008 in 2011-12 and P = 0.049 in 2013-14). The best LDL management rate was observed among the citizens of the postal code area where unemployment rate is between 6.0-6.9% both in 2011-12 and 2013-14 (Table 12).
The mean HbA1c and LDL levels of postal code areas by unemployment in 2011-12 and 2013-14 is presented in Table 13. Analysis of variance (ANOVA) test was performed to observe the differences in mean HbA1c and LDL level between the postal code areas by unemployment and we found that there are statistically significant differences in mean LDL levels only among males in 2011-12 (P = 0.005) and 2013-14 (P = 0.035).
HbA1c measurement rate and achievement in target level of HbA1c according to the median income of the citizens of the postal code area is presented in Table 14. A Chi-square test showed that there are statistically significant differences in HbA1c measurement rate by the different levels of median income of the citizens and the differences were statistically significant both for females (P = 0.013 in 2011-12 and P = 0.014 in 2013-14) and males (P < 0.001 in 2011-12 and P = 0.002 in 2013-14). The best measurement rate of HbA1c was seen in the postal code areas with citizens whose median income is within 15001 – 16000 euros both in 2011-12 and 2013-14.
To observe the differences in the management of HbA1c between the postal code areas by the level of median income of the citizens, a Chi-square tests was performed and we found that the differences were statistically non-significant both for females (P = 0.61 in 2011-12 and P = 0.188 in 2013-14) and males (P = 0.450 in 2011-12 and P = 0.504 in 2013-14). The best management rate was observed among the citizens whose median income of the postal code area is within 15001 – 16000 euros both in 2011-12 and 2013-14 (Table 14).
Table 15: LDL Measurement and management (2011-12 & 2013-14) by area level median income of the citizens.
(+) Included participants aged ≥ 20 and whose LDL measured both years. N = 6073 (F = 2860, M = 3213), missing data (n = 82)
Table 16: Mean HbA1c and LDL level (2011-12 & 2013-14) by area level median income of the citizen.
Included participants aged ≥ 20 and whose HbA1c or LDL measured both years.
Table 17: HbA1c measurement and management (2011-12 & 2013-14) by municipality.
Municipality HbA1c Measured (2011-12) *
(%)
HbA1c Measured (2013-14) *
(%)
HbA1c Measured both year *
(%)
HbA1c management (2011-12) +
(%)
HbA1c management (2013-14) +
(%)
Female Male Female Male
F M F M F M ˂ 7 7- 8.9 ≥ 9 < 7 7- 8.9 ≥ 9 ˂ 7 7- 8.9 ≥ 9 ˂ 7 7- 8.9 ≥ 9 Ilomantsi 77.2 74.7 92.6 87.5 73.5 69.0 73.8 21.1 5.1 72.6 22.3 5.1 64.0 27.3 8.6 56.2 29.4 14.4 Joensuu 77.2 73.0 86.9 83.8 70.5 66.3 67.3 24.5 8.2 70.5 22.3 7.3 66.8 27.4 5.7 64.1 31.1 4.8 Juuka 84.6 87.1 91.0 92.7 79.1 82.8 72.2 22.2 5.6 65.2 28.1 6.7 62.9 29.6 7.5 61.7 29.0 9.3 Kitee 89.0 82.5 93.1 91.0 85.2 77.8 69.9 25.3 4.8 77.2 18.8 4.1 65.3 25.0 9.7 61.8 30.0 8.2 Kontiolahti 79.5 79.6 91.8 85.0 75.8 71.9 77.5 18.5 4.0 78.8 15.3 5.8 62.7 29.5 7.8 64.5 26.4 9.1 Outokumpu 89.5 83.8 91.9 90.8 84.1 79.2 69.2 24.4 6.3 68.6 24.0 7.4 71.9 22.9 5.2 72.3 21.5 6.2 Lieksa 74.0 70.1 91.3 87.2 70.0 66.5 79.1 16.0 4.9 71.6 21.9 6.5 59.7 28.9 11.4 64.7 25.6 9.6 Liperi 87.0 81.4 89.8 90.7 81.4 77.9 74.7 20.4 4.9 72.2 19.8 8.0 65.8 28.9 5.3 63.2 28.1 8.7 Nurmes 67.7 69.9 89.4 90.3 61.6 66.5 83.2 13.9 3.0 68.5 28.3 3.3 70.4 22.2 7.4 71.2 22.2 6.6 Polvijärvi 80.3 70.1 87.6 86.4 73.7 62.6 64.3 27.1 8.6 60.2 24.1 15.7 73.3 21.8 5.0 65.2 28.3 6.5 Rääkkylä 82.2 81.7 90.0 95.2 77.8 79.8 78.5 19.8 1.7 73.9 17.0 9.1 54.3 37.1 8.6 60.2 28.9 10.8 Tohmajärvi 82.4 84.5 90.2 90.8 79.1 80.1 78.7 21.3 0.0 57.4 34.0 8.5 68.6 23.1 8.3 67.9 23.6 8.5 Valtimo 67.1 61.4 90.4 83.0 64.4 53.4 73.6 21.4 5.1 70.9 22.8 6.3 74.5 21.3 4.3 66.0 29.8 4.3 Total 79.3 76.1 89.4 87.4 73.9 70.7 73.8 21.1 5.1 72.6 22.3 5.1 66.1 26.9 7.0 64.6 28.0 7.5 Chi square
P value < 0.001 < 0.001 0.020 < 0.001 < 0.001 < 0.001 0.088 < 0.001 0.058 0.001 (*) Included participants aged ≥ 20. N = 9288 (F = 4366, M = 4922)
(+) Included participants aged ≥ 20 and whose HbA1c measured both years. N = 6707 (F = 3225, M = 3482)
LDL measurement rates and management according to the postal code area level median income of the citizens in the year 2011-12 and 2013-14 is presented in Table 15. A Chi-square test showed that there are statistically significant differences in LDL measurement rate between the postal code areas with different level of median income of the citizens and the differences were statistically significant for females in both time periods (P < 0.001 in 2011-12 and P = 0.001 in 2013-14). In males, the differences were significant only in 2011-12 (P < 0.001 in 2011-12 and P = 0.063 in 2013-14).
To observe the differences in management of LDL between the postal code areas with different level of median income of the citizens, a Chi-square tests was performed (Table 15). We found that the differences were statistically significant for females (P = 0.009 in 2011-12 and P = 0.014 in 2013-14) but in males the differences were significant only in 2011-12 (P = 0.008 in 2011-12 and P = 0.191 in 2013-14).
The mean HbA1c and LDL levels of postal code areas by median income of the citizens in 2011-12 and 2013-14 is presented in Table 16. Analysis of variance (ANOVA) test was performed to observe the differences in mean HbA1c and LDL level between the areas by median income of the citizens. We found that there are significant differences in mean HbA1c by median income of the citizens among females in 2013-14 (P = 0.039). It was also observed that there are differences in mean LDL level between areas in females both in 2011-12 (P = 0.004) and 2013-14 (P = 0.019).
HbA1c measurement rate and management according to the municipality is presented in Table 17.
A Chi-square test showed that there were statistically significant differences in HbA1c measurement rate between the different municipalities and the differences were statistically significant both for females (P < 0.001 in 2011-12 and P = 0.020 in 2013-14) and males (P < 0.001 in 2011-12 and P < 0.001 in 2013-14). Females living in Outokumpu showed the best measurement rate of HbA1c in 2011-12 and in 2013-14 HbA1c measurement rate was the highest among males living in Rääkkylä. To observe the differences in the management of HbA1c between the different municipalities, a Chi-square tests was performed and we found that there are no statistically significant differences in the management of HbA1c among females in 2011-12 (P = 0.088) or in 2013-14 (P = 0.058). However, there were differences in the HbA1c management levels among males living in different municipalities both in 2011-12 and 2013-14 (Table 17).
Table 18: LDL Measurement and management (2011-12 & 2013-14) by municipality.
(+) Included participants aged ≥ 20 and whose LDL measured both years. N = 6155 (F = 2890, M = 3265)
Table 19: Mean HbA1c and LDL level (2011-12 & 2013-14) by municipality. Included participants aged ≥ 20 and whose HbA1c or LDL measured both years.
LDL measurement rate and management according to the municipality is presented in Table 18.
Chi-square test showed that there was statistically significant difference in LDL measurement rate and management of LDL between people living in different municipalities and the differences were statistically significant both for females and males in both years.
Mean HbA1c and LDL levels in different municipalities in 2011-12 and 2013-14 are presented in Table 19. Analysis of variance (ANOVA) test was performed to observe the differences in mean HbA1c and LDL levels between municipalities. We found that there were statistically significant differences in mean HbA1c and LDL levels between municipalities among females and males both in 2011-12 and 2013-14.
Table 20: Multivariate logistic regression model explaining the effect of background variables in the improvement of HbA1c and LDL follow-up.
Variables Improvement in HbA1c follow-up (0 = not improved, 1 = improved)
Municipality < 0.001 < 0.001
A multivariate logistic regression was performed to observe the effects of age, gender, proportion of educated, median income, unemployment rate and municipality on the likelihood that participants have improvement in HbA1c follow-up (Table 20). The logistic regression model was statistically significant, χ2(17) = 192.6, P < 0.001. The model explained 3.5% of the variance in the improvement in HbA1c follow-up and correctly classified 83.9% of cases. Increasing age was associated with a decreased likelihood of improvement in HbA1c follow-up. It was also observed that being in a certain municipality may affect the likelihood of improvement in HbA1c follow-up.
Similar analysis was performed to observe the effects of age, gender, proportion of educated, median income, unemployment proportion and municipality on the likelihood that participants have improvement in LDL follow-up (Table 20). The logistic regression model was statistically significant, χ2(17) = 91.16, P < 0.001. The model explained 1.6% of the variance in the improvement in HbA1c follow-up and correctly classified 82.8% of cases. Increasing age was found to be associated with a decreased likelihood of improvement in LDL follow-up. Municipality was also found to have effect on the likelihood of improvement in LDL follow-up.
Table 21: Multivariate logistic regression model explaining the effect of background variables and baseline level of HbA1c and LDL in the improvement of HbA1c and LDL management.
Variables Improvement in HbA1c management (0 = not improved, 1 = improved) We also wanted to observe if the background variables and baseline level of HbA1c or LDL predict the likelihood of improvement in the management of HbA1c or LDL. A multivariate logistic regression was performed to observe the effects of age, gender, proportion of education, median income, unemployment rate, municipality and baseline HbA1c level (2011-12) on the likelihood that participants have improvement in the management of HbA1c (Table 21). The logistic regression model was statistically significant, χ2(17) = 192.6, P < 0.001. The model explained 3.5%
of the variance in the improvement of the management of HbA1c and correctly classified 83.9% of
cases. Increasing age was associated with a decreased likelihood of improvement in the management of HbA1c. It was also observed that being in a certain municipality may affect the likelihood of improvement in the management of HbA1c. Baseline HbA1c was found to have positive association with the outcome, meaning that higher baseline HbA1c levels were associated with increased likelihood of improvement in the management of HbA1c.
Similar analysis was performed to observe the effects of age, gender, proportion of educated, median income, unemployment rate, municipality and baseline LDL level on the likelihood that participants have improvement in the management of LDL (Table 21). The logistic regression model was statistically significant, χ2(18) = 714.4, P < 0.001. The model explained 15.2% of the variance in the improvement in the management of LDL and correctly classified 63.6% of cases.
Increasing age was found to be associated with an increased likelihood of improvement in the management of LDL. Males were 1.25 times more likely to show improvement in the management
Increasing age was found to be associated with an increased likelihood of improvement in the management of LDL. Males were 1.25 times more likely to show improvement in the management