Sociodemographic and clinical characteristics of older patients with hypertension
Table 1 and Figure 2 present the sociodemographic and clinical characteristics of older hospitalized patients with hypertension who were enrolled in this study. In terms of quantity, the number of hospitalized patients increased from 4,368 to 18,691 in 2020, a 4.3-times increase. The sex ratio was relatively stable in the study’s cohort, with a slightly higher proportion of men than of women; the average proportion of men was 52.75%. Regarding age, the overall average age of patients was 72.12 years, of which those in the 60–69-year category constituted the highest proportion of the study population; the average age of patients decreased yearly, from 72.96 years in 2015 to 71.71 years in 2021. Concerning the number of hospitalizations, patients who were hospitalized once constituted the highest proportion of our cohort, and the average number of hospitalizations for patients increased from 1.43 in 2015 to 3.36 in 2020. In China's tertiary general hospitals, a typical single hospitalization period is approximately 4 days, and the highest proportion of hospitalization duration was 5–8 days. The average length of hospitalization continuously decreased from 11.31 days in 2015 to 8.79 days in 2021. The number of comorbidities increased from 2015 to 2020; the proportion of older inpatients with hypertension and two or more diseases increased from 95.01% to 99.69%. From 2015 to 2020, the main modes of discharge for older inpatients with hypertension were ranked in descending order as follows: discharge against medical advice, transfer by medical advice, discharge without medical advice, other, and death, with average proportions of 93.70%, 4.56%, 0.88%, 0.44%, and 0.41%, respectively.The overall proportion of older hospitalized patients with hypertension who suffered from two or more diseases increased from 94.73% to 99.76%, and the average number of diseases increased from 4.13 to 4.86. Table 2 presents significant increasing trends in the number of hospitalizations and the average number of comorbidities (P<0.01 for all). Additionally, significant decreasing trends were observed in the average age and the average length of hospitalization (P<0.01 for both).
[Table 1 near here]
[Figure 2 near here]
[Table 2 near here]
Disease combination characteristics in older patients with hypertension
Figures 3 and 4 show the proportions and numbers of the top five disease combinations under the ICD-10 disease classification criteria that frequently affected older hospitalized patients with hypertension. The top five most common diseases in patients with hypertension were "diabetes," "cerebrovascular disease," "ischemic heart disease," "other types of heart disease," and "other diseases of the respiratory system." The proportion of "hypertension + diabetes" combination decreased from 73.85% in 2015 to 50.56% in 2020; the proportion of "hypertension + cerebrovascular disease" combination increased from 30.72% in 2015 to 35.35% in 2020, and the proportion of other disease combinations remained relatively stable. Concerning quantity, the number of patients with each disease combination increased rapidly from 2015 to 2020, and among them, the number of patients with "hypertension + diabetes," "hypertension + cerebrovascular disease," "hypertension + ischemic heart disease," "hypertension + other types of heart disease," "hypertension + ischemic heart disease," and “hypertension + other diseases of the respiratory system” increased 3.09, 5.20, 4.54, 5.70, 5.70, and 4.33 times, respectively.
[Figure 3 near here]
[Figure 4 near here]
Financial burden of hospitalization
Figure 5a and 5b shows the composition and changes of single hospitalization expenses for older inpatients with hypertension in this study cohort, respectively. Hospitalization expenses mainly included medicine, diagnosis, treatment, service, and nursing expenses; among these, medicine, diagnosis, and treatment accounted for relatively high proportions (average: 35.45%, 28.49%, and 21.27%, respectively). The proportions of service and nursing expenses were relatively low (average: 3.63% and 2.55%, respectively). From 2017 to 2020, the proportion of drug and nursing expenses of older inpatients with hypertension continued to decrease, and the proportion of diagnosis, treatment, and service expenses continued to increase. Regarding the cost changes, the average single hospitalization cost for older inpatients with hypertension has continually increased from 15,557.55 yuan in 2015 to 19,407.66 yuan in 2020. When the analysis combined our data with that of the medical insurance reimbursement data in Jiangsu Province in 2020, the average single hospitalization cost for older patients with hypertension was 2,500–4,000 yuan higher than that for adult patients with hypertension. Concerning data distribution, the median cost of single hospitalization for the older adult with hypertension was less than the mean, which means that more than half the patients had single hospitalization costs that were lower than the average cost, and the single hospitalization cost showed a right-skewed distribution.
[Figure 5a and 5b near here]
As the single hospitalization cost for older patients with hypertension in this study had a right-skewed distribution, to achieve the optimal effect of normal distribution and minimize the error caused by uneven sample distribution, the square root values of the hospitalization costs were taken, and variance analysis and multiple linear regression analysis were performed. As shown in Table 3, the total cost of a single hospitalization (square root) for older inpatients with hypertension significantly differed with regard to sex, age, number of hospitalizations, and days of hospitalization. All P-values were <0.01.
[Table 3 near here]
A multiple linear regression analysis was conducted by using the square root of the total cost of a single hospitalization as the dependent variable, and sex, age, number of hospitalizations, days of hospitalization, and the number of illnesses as independent variables. According to the results of the analysis presented in Table 4, the Durbin–Watson value of this research was 1.21, which is far from the standard value of 2 and indicated a slight autocorrelation between the data, which was barely acceptable. This could be because, on the one hand, the data on the front page of medical records were mainly from tertiary general hospitals, the older patients who were admitted to the hospital had been screened by primary and secondary hospitals, and there was a certain similarity among the patients; on the other hand, some patients in the 6-year medical records who were identified from front page data were of repeated hospitalizations, and there could have been multiple cases of similar data. The variance inflation factor values of the independent variables were all <5, indicating the absence of collinearity; however, the significance level was <0.01, which indicates that each variable has a significant impact on the total cost of a single hospitalization. From the standardized coefficient, it can be inferred that the number of days of hospitalization had the greatest impact on the total cost of a single hospitalization (coefficient 0.49), followed by sex and number of hospitalizations (coefficient −0.05 and −0.04, respectively), whereas the number of illnesses and the age had little effect on the total cost of a single hospital stay (coefficient 0.03 and −0.03, respectively).
[Table 4 near here]
The total cost of a single hospital stay for women in this cohort was lower than that for men and, with an increase in the days of hospitalization and the number of comorbidities, the total cost of a single hospital stay increased significantly. Furthermore, the total cost of a single hospital stay decreased significantly with the increasing number of hospitalizations and age. As shown in Table 2, age was positively correlated with the total cost of hospitalization, which contradicts the results of the multiple linear regression analysis. This may be attributed to the positive correlation between age and the number of comorbidities and days of hospitalization, which resulted in a negative coefficient in the overall model.