This chapter discusses the risk factors for the death of patients with cardiovascular diseases in Ethiopia as well as the analysis and findings of the data collected. The study used EDHS data to include 4712 CVD samples in total. 2952 (62.6%) of them are uncensored overall, while 1760 (37.3%) are censored. The patient's functional status is documented under two headings: Not Alive (death or event) and Alive (censored). Out of 4712 patients, 1760 (37.4%) were alive during the study period, and 2952 (62.6%) were either dead or not alive. This represents a descriptive output of the patients. According to this median value, there was a 0.5 probability of death for half of the cardiac patients.
Table 4.1 above shows that 4712 patients in total were taken into consideration for the analysis. The median time of these cases—which had a 95% confidence interval with lower bound of 5 and upper bound of 6—was 6 months after their cases; of them, 62.6% were uncensored and 37.4% were still alive (censored). 2577 (54.6%) and 2135 (45.4%) of the 4712 cardiac patients were female. Of those patients, the percentage of female patients who die (57.31%) is higher than the percentage of male patients (42.69%), at 1692.
Approximately 695 patients (14.74%) came from the Tigray region, 288 patients (6.11%) from the Afar region, 504 patients (10.69%) from the Amhara region, and 3235 patients (68.46%) from other regions. Based on the entire sample, the Tigray region appears to have a higher death rate (14.74%) than the Amhara and Afar regions (10.69% and 6.11%, respectively). In terms of educational attainment, 840 patients (28.4%) were primary, 243 patients (8.2%) were higher, and 1591 patients (53.8%) had no formal education.
This indicates that the mortality rate was higher for patients with no formal education (1591–53.8%) and higher for primary and secondary cardiac patients (840–28.4%, 278–9.4%, and 243–8.2%), respectively. Cardiovascular patients 30 years of age and older have the highest death rate, 2436 (82.5%), compared to 20–30 and 10–20 years old, who have death rates of 41 (1.3%) and 475 (16%), respectively. Among the total number of cardiovascular patients in this study, 1513 patients (32.21%) were non-smokers and 3199 patients (67.89%) were smokers. Patients who did not smoke had a death proportion of 1346 (45.6%) versus smokers, who had a death proportion of 1606 (54.4%).
Of all the patients with cardiovascular disease, 3636 (77.1%) had alcohol use disorders, while 1076 (22.8%) did not use alcohol. The mortality rate for patients who did not use alcohol was lower—659 (22.3%)—than for those who did, with 2293 (77.6%) representing the death rate for alcohol users.
Survival Analysis
Researchers employed estimation techniques like the Kaplan Meier curve function to examine each person's estimated survival time.
Table 4. 3 Kaplan Meier survival estimate
Here there is another output up to n, since n = 4712
As we have seen from the above Table 4.3 the survival time of cardiovascular patients risk becomes decreasing order with starting one.
The Kaplan-Meier curve (KM) graph of the survival times for patients with cardiovascular disease (CVD) is shown in Fig. 4.1, guaranteeing censored (alive) or uncensored (death) data in the research. It demonstrates that the majority of cardiac deaths happened during the initial months of dots and that they decreased during the subsequent follow-up month. The majority of the deaths happened in the earlier month, and the estimate of the overall Kaplan-Meier Survivor function showed a decline in the following months of follow-up. This indicates that the most extensive observations are unfiltered. This suggests that there is no zero convergence of the survival time.
Log Rank Test
A test of equality was conducted along the probabilities among the various groups using the Log Rank test. The idea that there is no difference in the odds of an event happening for any population at any given time point is the null hypothesis that needs to be tested. If the Log Rank test yields a p-value of less than 0.05, it is assumed that the predictor would be included in a model. It is extremely unlikely that a predictor with a p-value greater than 0.05 would add anything to a model that already contains other predictors. The differences in survival experiences between two or more levels of the covariates are indicated by the p-values of the log-rank test. Each variable's following hypotheses are being investigated:
H0: There is no difference between survival curves
H1: There is a difference between survival curves
Comparison of Survival Curves Among Covariates (Log- Rank Test)
The following represents the findings of the log-rank test for survivor function equality.
Log-rank test table 4.4 various covariate groups.
Covariates
|
Chi-square value
|
Df
|
P value
|
Sex
|
25.6
|
1
|
4e-07
|
Age
|
0.2
|
2
|
0.9
|
Educational status
|
11.3
|
3
|
0.01
|
Economic level
|
6.7
|
4
|
0.2
|
Diabetes
|
13.7
|
1
|
2e-04
|
BMI
|
2.3
|
2
|
0.3
|
Pulse rate
|
6.2
|
3
|
0.1
|
Region
|
23.6
|
11
|
0.051
|
Family history of cardiac
|
0.2
|
1
|
0.7
|
Smoking
|
317
|
1
|
2e-16
|
Alcohol use
|
1.2
|
1
|
0.03
|
Blood pressure
|
49.1
|
2
|
2e-11
|
To investigate the significance of variations in survival experience among various factors, the researcher employed the log-rank test. For each categorical covariate, the null hypothesis that needed to be tested was that there would be no difference in the odds of an event occurring at any given time.
If, at the five percent significant levels, the p-value from the log-rank test is less than 0.05, the variables are included in the study. Sex, educational status, smoking, alcohol use, blood pressure, and diabetes mellitus are significant covariates at the 5% level of significance, according to the log-rank test results in Table 4.4 above. The corresponding p-values are 4e-07, 0.01, 2e-16, 0.03, 2e-11, and 0.0002, all of which are less than 0.05 at the five percent significant level. On the other hand, at the 5% level of significance, age, family history of heart disease, region, economic status, pulse rate, and BMI have no significant impact. In order to make comparisons between categorical covariates, the researcher can estimate survivor function among various categorical covariates using the Kaplan-Meier estimator survival curve.
Overall Fit For The Coxph Regression Model
There are several different tests of fit of the Coxph regression model:
Table 4.5
Several different tests of fit of the Coxph regression model:
Method
|
Primary Use
|
1. Cox-Snell residuals
|
Overall Fit
|
2. Martingale residuals
|
Form of covariates (should covariates be transformed)
|
3. Deviance residuals
|
Outliers
|
4. Schoenfeld residuals
|
Proportional Hazards assumption
|
5 Score residuals
|
Influential points
|
The p-value illustrates the proportional hazard assumption, which is tested at the 5% significant level. Therefore, if the p-value for any given result is less than 0.05, the proportionality assumption is broken.
Table 4.6
Output of Proportional Hazard assumptions
Covariates
|
Chi-square
|
Df
|
P-value
|
Sex
|
0.32296
|
1
|
0.0570
|
Smoking
|
0.48474
|
1
|
0.486
|
Diabetes
|
0.00225
|
1
|
0.962
|
Blood pressure
|
0.43318
|
2
|
0.510
|
Alcohol use
|
2.56724
|
1
|
0.109
|
Region
|
4.47760
|
3
|
0.064
|
Economic level
|
0.12068
|
4
|
0.728
|
BMI
|
2.69895
|
2
|
0.100
|
Educational status
|
1.42419
|
2
|
0.233
|
Pulse rate
|
1.25357
|
1
|
0.263
|
Family history of cardiac disease
|
1.23254
|
1
|
0.267
|
Age
|
4.56917
|
1
|
0.053
|
GLOBAL
|
19.21640
|
12
|
0.083
|
The findings demonstrated that neither the global test nor the test for any one of the covariates are statistically significant. The Cox Proportional Hazard assumptions were not broken by any of the covariates. Thus, the Cox Proportional Hazard model is applied in this investigation.
Table 4.6 above displays the results, which indicate that the p-value of the rho statistic and its corresponding covariates are greater than the 0.05 level of significance. This suggests that the assumption of proportionate hazard is satisfied for those variables, and that we cannot reject the null hypothesis of the proportionality of the Cox proportional hazard model.
Results of Cox Proportional Hazards Model
Table 4.7
Results of Univarate cox PH regression Analysis
Covariates
|
Coef
|
exp(coef)
|
se(coef)
|
Z
|
P
|
Smoking
|
0.664774
|
1.944052
|
0.037695
|
17.635
|
< 2e-16
|
Region
|
-0.003323
|
0.996683
|
0.005965
|
-0.557
|
0.57752
|
Sex
|
-0.150281
|
0.860466
|
0.037431
|
-4.015
|
5.95e-05
|
Economic level
|
-0.010898
|
0.989161
|
0.011572
|
-0.942
|
0.34632
|
Diabetes
|
0.117661
|
1.124863
|
0.038951
|
3.021
|
0.00252
|
Educational status
|
-0.039460
|
0.961308
|
0.019381
|
-2.036
|
0.04174
|
Alcohol use
|
-0.062099
|
0.939790
|
0.044392
|
-1.399
|
0.016185
|
Pulse rate
|
-0.021591
|
0.978641
|
0.019998
|
-1.080
|
0.28031
|
Family CVD
|
0.006588
|
1.006610
|
0.041113
|
0.160
|
0.87269
|
Age of patients
|
-0.014778
|
0.985330
|
0.043536
|
-0.339
|
0.73427
|
Blood pressure
|
0.181830
|
1.199410
|
0.032673
|
5.565
|
2.62e-08
|
Body mass index
|
0.021786
|
1.022025
|
0.020215
|
1.078
|
0.28116
|
The covariates Sex, Educational status, Blood Pressure, Alcohol, Smoking use, and Diabetes mellitus are statistically significant at the 20–25% level, as shown by the univarate Cox PH regression analysis Table 4.7 above. hen, based on univariate cox PH regression analysis, the covariates Sex, Educational status, Blood Pressure, Alcohol use, Smoking, and Diabetes mellitus were chosen as significant risk factors for the death of cardiovascular patients. These covariates are statistically significant at the 20–25% level of significance. However, at 20–25% of the significant level, the other four covariates—region, economic level, age, pulse rate, and BMI—are not significant.
Multivariable Cox PH Regression Analysis
All possible risk factors with a P-value of less than 20–25% at the significant level in the univariable Cox PH regression analysis were included in the multivariable Cox PH regression analysis. The stepwise method was used to choose the best subgroup of covariates for our model.
This indicates that only covariates with a P-value of less than or equal to 0.05 will be tested in the model; additionally, a covariate's P-value must be less than or equal to 0.05 in order for it to be retained in the model. The outcome is displayed in Table 4.8 below.
Table 4.8
Results of the Multivariable Cox PH regression Analysis
Covariates
|
Categories
|
Coef
|
Exp(coef)
|
Se(coef)
|
P-value
|
95%CI
|
Lower
|
Upper
|
Sex
|
Female
|
|
|
|
-
|
|
|
Male
|
0.1321165
|
0.8762389
|
0.038058
|
0.000518
|
0.8133
|
0.9441
|
Education
|
No education
|
|
|
|
|
|
|
Primary
|
0.0198685
|
1.0200672
|
0.043110
|
0.644892
|
0.9374
|
1.1100
|
Secondary
|
-0.0653261
|
0.936762
|
0.065646
|
0.319680
|
0.8237
|
1.0654
|
Higher
|
0.1569656
|
0.8547335
|
0.069355
|
0.023624
|
0.7461
|
0.9792
|
Alcohol use
|
No
|
-
|
-
|
-
|
-
|
|
|
Yes
|
-0.0544140
|
0.9470400
|
0.044911
|
0.022567
|
0.8672
|
1.0342
|
Blood pressure
|
Low
|
-
|
-
|
-
|
-
|
-
|
-
|
Normal
|
-0.0427697
|
0.9581320
|
0.052371
|
0.052371
|
0.8647
|
1.0617
|
High
|
0.33306
|
1.39523
|
0.06224
|
8.77e-08
|
1.2350
|
1.5763
|
Diabetes
|
No
|
-
|
-
|
-
|
-
|
-
|
-
|
Yes
|
0.1231410
|
1.1310439
|
0.039479
|
0.001814
|
1.0468
|
1.2220
|
Smoking
|
No
|
|
|
|
|
|
|
Yes
|
0.68681
|
1.9873678
|
0.0378
|
2e-16
|
1.8451
|
2.1406
|
These variables are the most factors of time to death for cardiovascular patients in Ethiopia.
A positive sign implies a higher risk of cardiovascular events, whereas a negative sign denotes a lower risk of cardiovascular events. As we've seen, the variables sex, educational attainment, blood pressure, alcohol consumption, diabetes, and smoking are included in the final model at the 5% significance level. Therefore, we concluded that since these covariates increase the risk of a cardiovascular patient's death, they also significantly affect the patient's chance of survival.
The multivariable Cox PH model results are shown in the above table, and the hazard ratios were the basis for discussion. For the categorical covariates, comparisons are made within groups and with the reference category. For the sex, the reference category is female. For men, the estimated hazard ratio is 0.876. This suggests that male patients had a survival rate that was approximately 0.876 times higher than that of female patients when it came to the risk of cardiovascular disease.
This indicates that, of all the covariates under control, the risk of cardiovascular disease rises by approximately 0.876 times for the covariate males. According to the 95% confidence interval, the hazard ratio can range from 0.8133 to 0.9441.
"No education" is the reference category for the educational status group. For primary school education, the estimated hazard ratio is 1.020. This suggests that the risk of cardiovascular disease (CVD) rises by approximately 1.020 times for cardiovascular patients in the primary group compared to those who receive no education.
The 95% confidence interval indicates that there may be as little as 0.9374 and as much as 1.1100 times more CVD cases as a result of primary educational status categories.
For secondary school, the estimated hazard ratio is 0.936. Conversely, the secondary estimate showed a negative value, suggesting that the hazard ratio falls with time. It can be discussed that the risk of CVD drops by approximately 0.936 times when secondary schooling is included as a covariate compared to not attending any schooling.
The 95% confidence interval indicates that there may be as little as 0.8237 times and as much as 1.0654 times as much of an increase in the rate of CVD related to secondary educational status categories. This suggests that there is a difference in the CVD survival rate in this variable between the primary, secondary, and uneducated groups. In the case of normal blood pressure, the estimated hazard ratio is 0.951. According to the confidence interval (0.8647, 1.0617), the rate may potentially be as high.
This suggests that people with low blood pressure and those with normal blood pressure have different chances of surviving cardiovascular disease. Those who don't smoke make up the reference category for the smoking group. For smokers, the estimated hazard ratio is 1.987. Conversely, the smoke users' positive value showed that the hazard ratio rises with time.
For patients with diabetes, the estimated hazard ratio is 1.1310. This suggests that individuals with diabetes die at a rate that increases their risk of cardiovascular disease (CVD) by approximately 1.13 times when compared to individuals without diabetes.
The time to death associated with alcohol use has an estimated hazard ratio of 0.947. The estimate for alcohol consumption, on the other hand, showed a negative value, suggesting that the hazard ratio drops with time. This suggests that patients in the alcohol use group die at a rate where their chance of developing cardiovascular disease (CVD) drops by roughly 0.947 times when compared to patients in the no alcohol use group.