探討醫師使用中央健康保險局健保資訊網服務系統之影響因素─以糖尿病指標回饋為例

Abstract

資訊回饋(information feedback)是一種給予醫師關於其執業行為、病患照護與其他醫師相較之結果，或是外部標準等資訊的過程(Eisenberg, 1993)；也可定義為報告病患過去的照護活動給醫師，以影響未來醫療臨床決策的一種方式(Balas et al., 1996)。中央健康保險局自2005年開始，透過健保資訊網服務系統(Virtual Private Network, VPN)作為資訊回饋之媒介，針對各級醫療院所實施醫療品質資訊的指標回饋。本研究旨在探討醫師參與糖尿病醫療給付改善方案之程度，對於其使用健保局VPN查詢糖尿病指標回饋情況之影響，也探討指標對於醫師行為之影響，有助於了解醫師行為決定因素以提升糖尿病治療之品質。 本研究之研究架構有二，第一個架構探討醫師參與糖尿病方案之程度，對於使用VPN查詢糖尿病指標回饋行為之影響，自變項為醫師參與糖尿病方案收案狀況，依變項為醫師查詢VPN糖尿病指標，控制變項為醫師個人特性以及參與健保局政策情況，對於曾使用VPN查詢指標之醫師，再放入第二個架構進行分析。研究架構二運用Ajzen(1991)提出之計劃行為理論(Theory of Planned Behavior)，探討醫師在面對健保局以VPN作為糖尿病醫療品質提升之回饋機制時，是否使醫師產生增加病患抽血檢查以提升病患受檢率的意願與行為，以了解醫師因資訊回饋而自覺受影響之程度。 本研究以2007年7月至9月健保局台北分局轄下照護糖尿病之西醫基層共1211家診所醫師作為母群體，研究工具為研究者自行設計的結構式問卷，採用郵寄自填式問卷進行研究，總計發出510份問卷，回收151份有效問卷，有效問卷回收率為30%。 本研究以卡方檢定和Fisher真實檢定分析自變項與依變項間之相關性；最後再以多項羅吉斯迴歸分析(Multinomial Logistic Regression Analysis)，在控制醫師個人特性與參與健保局政策情況之後，對於研究架構一之假說進行驗證。研究架構二受限於目前使用VPN醫師之樣本數不足之故，僅能以無母數檢定之Kruskal-Wallis檢定法以及Spearman等級相關分析，進行保守性相關關係之探索，提出可能影響醫師行為之因素。 本研究主要研究發現與建議有四： 一、醫師知道VPN也有查詢有佔15.2%、知道VPN但沒查詢佔60.9%，而不知道VPN也沒查詢佔23.8%。針對不知道VPN的醫師，建議健保局增強政策宣導，以提升VPN的能見度與使用率；而針對已經知道VPN的醫師則可進行管理上的支持，透過激勵行為之誘因，可降低醫師對於使用網路服務須登錄個人資料的心理風險，以提高系統之使用率。 二、醫師參與糖尿病醫療給付改善方案、參與家庭醫師制度、曾經收到健保局所寄發之「糖尿病診療概況報告表」(品質報告卡)以及高診次所造成之執業上對於查詢VPN指標之需求，皆是促使醫師主動上VPN擷取自己指標表現之重要驅動因素，建議健保局重視並結合多重驅動因素以擴大資訊回饋的效果與力量。 三、本研究發現糖尿病病患收案率達50%以上之醫師知道VPN也有查詢指標回饋的機會比沒有加入方案的醫師高，目前糖尿病醫療給付改善方案的品質獎勵措施當中，病患收案門檻目標獎勵值為30%，此門檻能達到獎勵性質，有助於促使醫師上VPN查詢指標。 四、本研究調查使用健保局VPN查詢糖尿病指標回饋的醫師，有86.9%傾向於會依照「糖尿病指標回饋」來增加對病患的抽血檢查。在指標適當性方面，有91.3%的醫師持中立傾向認同的態度，認為指標能適當衡量糖尿病照護品質；而在指標是否符合真實狀況的看法方面，則有82.6%的醫師持中立傾向認同的意見。從本研究從相關分析當中得知，醫師對糖尿病指標的正確性與適當性的看法，是醫師是否會依指標而增加對病患抽血檢查行為的重要相關因素，因此持續提升醫師對糖尿病指標回饋資訊的信任程度，是促使醫師增加對病患抽血檢查的行為不可或缺的相關要素，有助於促進糖尿病的醫療照護品質。

Information feedback involves giving physicians information about how their practices or patient outcomes compare with those of other physicians or with an external standard(Eisenberg, 1993).It is also a reporting on past patient care activities future clinical decisions(Balas et al., 1996).The Bureau of National Health Insurance use VPN information system as a medium to feedback the quality indicator performance to medical institution since 2005. The objectives of this study are to research how the physician’s participation in diabetes mellitus disease management program impacts their use of VPN information system inquiring DM indicators, and to explore the influence of indicators feedback on physician’s practice. There are two frameworks in the study. The first one is the factors affecting physician’s using of VPN. The independent variable is the extent of participation in DM disease management program, the dependent variable is the situation of using VPN, and the control variables are demographic characteristics, working condition, participation of family physician system, and reception of DM report card. Then put the physicians who had used the VPN for inquiring DM indicators performance to the second framework for analysis. The second framework derived from the Theory of Planned Behavior (Ajzen, 1991), exploring the relevant factors influencing physician’s intention to change practice according to the DM indicators feedback. The population of this study is 1211 clinics which care the diabetes patients under the Bureau of National Health Insurance Taipei Branch’s authority during July to September 2007. This study used a structured self-evaluated questionnaire to collect data. The survey subjects were 510 clinic presidents. There are 151 efficient returned questionnaires, and the response rate is 30%. In the first research framework, Chi-square test and Fisher Exact test have been used to examine the correlation between dependent and independent variables. Then Multinomial logistic regression analysis has been used to exam the hypotheses. In the second research framework, Kruskal-Wallis and Spearman rank-correlation analysis have been used to explore the correlation between the variables. The main findings of this study are as follows: 1.There are 15.2% of respondents knew the VPN and also inquired the implementation rate of DM indicators, 60.9% knew the VPN but didn’t inquire the DM indicators performance, and 23.8% didn’t know the VPN. 2.The participation of diabetes mellitus disease management program and family doctor system, the reception of diabetes mellitus report card, and high frequency of admission influence the physician taking positive action to use VPN to the DM indicator information. 3.Physician’s rate of diabetes patients in the DM disease management program become a significant factor for inquiring DM indicators information on VPN when it is over 50%. 4.Among the 23 physicians who use the VPN to inquire the implementation rate of DM indicators, there are 86.9% tend to enhance patient’s blood tests, 91.3% tend to take the indicators as proper measurements of diabetes care quality, 82.6% tend to consider the outcome of indicators could reflect their real practice situation. According to the result of correlation analysis, the more positive the physician’s attitude toward the DM indicators, the more they enhance the implementation rate of patients’ blood tests. According to the study results, some suggestions are given to Bureau of National Health Insurance Taipei Branch: 1.For the physicians don’t know VPN, enhancing the visibility and utilization of VPN information system by increasing policy advocacy; for the ones already know VPN, reducing physician’s phobia about needing to sign personal data in using the Web service by encouraging and releasing incentives. 2.Pay attention to the physician’s multiple drivers of using VPN to strengthen the effect of information feedback to physician’s practice. 3.Continuously put emphasis on the validity of DM indicators information could conduce to physician’s trust on the feedback. It is the essential elements which contribute to the quality of diabetes care.