健康行為群聚類別與自覺健康狀況及醫療利用之關係—以臺灣18~25歲成年初顯期族群為例

Abstract

研究目的：探討18~25歲成年初顯期族群之健康行為是否有群聚現象，並分析群聚類別與其自覺健康狀況以及急診、西醫門診以及住院等醫療利用之間的相關性。 資料來源：本研究運用2009年國民健康訪問調查，以18~25歲成年初顯期族群為研究對象，刪除各子題所使用之研究變項為遺漏值之樣本後，子題一樣本數為2,679人，子題二為2,636人，子題三為2,633人，子題四之急診及住院醫療利用樣本數為2,631人，西醫門診醫療利用樣本數則為2,628人。 方法：本研究使用二階段群聚分析法 (Two-Step Cluster Analysis)依成年初顯期族群之健康行為型態進行群聚分析並加以命名，並以多元羅吉斯迴歸分析探討社會人口學變項與健康行為群聚類別之關係；為瞭解群聚類別與自覺健康狀況之關係，乃使用Ordered Probit Regression分析；在探討群聚類別與醫療利用情況，本研究乃使用羅吉斯迴歸分析法 (Logistic regression)以及零膨脹負二項迴歸模型 (Zero-inflated negative binomial regression model)分析群聚類別與急診、西醫門診以及住院等醫療利用與否以及利用次數之間的相關性。 結果：成年初顯期族群可分為三個群聚類別：菸酒不離族共637人，佔23.78%、飲食不均族有947人、佔35.35%，以及健康保健族，共1,095人，佔40.87%。在自覺健康狀態方面，相較於健康保健族，菸酒不離族 (Coef.= -0.26, P-value=0.00)與飲食不均族 (Coef.= -0.41, P-value=0.00)的自覺健康狀況較差。在有無利用急診醫療方面，相較於健康保健族，菸酒不離族較會使用急診醫療 (OR=1.59; 95% CI=1.14-2.21)；在急診醫療次數方面，飲食不均族使用較多次數的急診醫療 (IRR=1.49; 95% CI=1.08-2.05)，而健康行為群聚類別與西醫門診、住院與否或其利用次數皆無顯著相關。 結論與建議：成年初顯期族群可以分為菸酒及飲食之健康行為型態之區別。菸酒不離族與飲食不均族均有較差的自覺健康狀況。在醫療利用方面，菸酒不離族較可能會使用急診醫療。而飲食不均族較可能使用較多次數的急診醫療。由於本研究所探討的年齡層屬於較健康的族群，在醫療利用上，雖看出群聚與急診醫療利用之相關性，但受限於資料並無法得知就診之主要原因，且因受限於橫斷性資料，也無法看出群聚類別對後續西醫門診或住院醫療利用狀況之關係，儘管如此，成年初顯期族群身處在青少年以及成人之間的轉捩點，培養健康之生活型態仍為重要之課題。本研究建議，未來在政策上可強化在大專校院及職場的成年初顯期族群之健康促進並設計多重模式的介入方案，針對不同族群進行客制化之健康促進活動。Objective To examine the relationship between the health behavior clustering, self-rated health and health utilization including emergency, inpatient and outpatient services among emerging adulthood aged 18-25 in Taiwan. Data Sources The 2009 National Health Interview Survey was used in this study. After deleting the cases with missing value in four different subjects, the sample sizes are 2,679, 2,636, and 2,633 in the first three subjects respectively. In the fourth subject, the sample size are 2,633 and 2,631 for analyzing the relationships between the groups of clustering and emergency services and inpatient care utilizations and 2,631 for analyzing the relationships between the groups of clustering and outpatient care.

Methods This study used the Two-step Cluster Analysis to identify groups of emerging adulthood with similar behavior patterns among emerging adulthood. Logistic regression was used to analyze the association between sociodemographic factors and groups of clustering. We used the ordered probit regression model to analyze the relationships between groups of clustering and self-rated health. As for healthcare utilization, logistic regression was used to analyze the relationships between groups of clustering and the probability of using emergcy services, outpatient services or inpatient care. Zero-inflated negative binomial regression model were used to investigate the association between groups of clustering and the number of visits of emergency services, outpatient healthcare services, as well as inpatient care.

Results Three clusters were identified including the “smoking-alcohol” (n=637, 23.78%), “unbalance diet”(n=947, 35.35%), and “healthy lifestyle”(n=1,095, 40.87%) groups. When compared to the "healthy lifestyle” group, the "cigarette-alcohol” group and the " unbalance diet" group are more likely to report having worse health. Those who were classified in the “unbalance diet” group were more likely to use emergency services (OR=1.59; 95% CI=1.14-2.21) and those who were classified in the “unbalance diet” group were more likely to use more emergency services when compared to those in the “healthy lifestyle” group (IRR=1.49; 95% CI=1.08-2.05). This study found no relationships between groups of clustering and outpatient healthcare services, no matter whether in terms of usage or not, or in terms of the number of visits.

Conclusions The groups of clustering can be generated based on smoking-alcohol and diet habit among the emerging adulthood. Those who were in the “smoking-acohol” and “unbalance diet” groups were more likely to self-report worse health. In addition, unhealth behavior clusters were found to be associated with a higher risk of using emergency care and more visits of emergency services. As the generation we discussed belongs to the the healthier group, and although we have found the relationships been groups of clustering and emergency services, we cannot know the main reason why these people used emergency service due to lack of detailed data. In addition, this cross-sectional study could not analyze the effects of clustering groups on outpatient or inpatient medical care utilizations. All in all, emerging adulthood is at a critical turning point from life-course perspective, therefore, it is an important task to develop a healthy lifestyle. Our study suggests that whether or not in college, university, or the workplace, health promotion programs could adopt an integrated approaches and tailor-made programs, when taking into account the common determinants across different types of health behaviors.