乳癌核心測量在病人層次的指標遵從度分析及其與病患存活之相關性研究

Abstract

背景：國民健康局於民國92年開始蒐集六大癌症之診療明細於「癌症診療資料庫(TCDB)」，並於93至95年委由台大公衛學院的研究團隊(95年成立健保研究中心)發展完成六大癌症(子宮頸癌、乳癌、大腸直腸癌、口腔癌、肺癌、肝癌) 之核心測量指標，目的讓各癌症照護機構藉此工具達到內部品質改善及外部監測。而乳癌為目前研究文獻較豐富、國內診療資料蒐集較齊全的癌症，但透過品質測量工具所監測的數據多以一群人為測量單位，調查該群研究對象當中符合所建議之照護品質指標的比例，屬於以人口為基礎(population-based)的總體測量方式，雖可看出整個照護系統的服務品質但卻無法看出每個獨立個體所接受到的整體照護情形，且鮮少文獻探討乳癌核心測量組合與病患結果之間的關係，本研究欲透過以個案為基礎(case-based)的測量方式，藉乳癌核心測量指標了解目前國內乳癌照護服務提供情形，並探討病人層級核心測量組合遵從度的影響因素、及其與病患存活之間的關連性。 方法：本研究為一回溯性世代研究，利用癌症診療資料庫佐以健康保險資料庫進行研究資料之蒐集，包括病人人口學/臨床特性、申報醫院特性/服務量，及乳癌12項分為治療前、治療及追蹤三階段之核心測量指標適用與符合人數。研究對象為91年至93年29家申報醫院所申報10579位乳癌新發確診病患，追蹤至94年底之存活情形，透過個案為基礎之測量方法了解每位患者從治療前至治療後追蹤一系列照護過程中所接受到的照護情形。本研究以SAS9.1版軟體進行統計分析，以次數分配、百分比呈現乳癌個案基本資料、臨床特性、申報醫院特性、及核心測量項目遵從百分比，以卡方檢定、複迴歸檢定病人與醫院特性與核心測量項目遵從度相關性，及以Log-rank檢定和Cox比例危害模式(Cox proportional hazards model)探討病人及醫院特性與病人存活之相關性。 結果：(一)申報醫院特性部分：僅醫學中心(17家)及區域醫院(12家)兩類醫院參與申報，依健保分局地區別分類以北市醫院居多(12家, 41.4%)；每年手術量中位數95件(範圍為9~476件)；每年放射線治療量中位數35件(3~244件)；每年化學治療服務量中位數68件(1~382件)；每年荷爾蒙治療服務量中位數81件(14~280件)； (二)病患人口學部份：發病年齡平均51.0歲，以40~50歲族群最多；大多數患者發病後僅在固定一家醫院接受治療，且多在每月乳癌門急診服務量介於1000~2000人次醫院就診，追蹤至94年底存活者共9810人(92.7%)，平均追蹤25.4個月；(三)病患臨床特性：病理期別以二期最多(4170人, 43.3%)，大多以組織或病理學診斷為依據(10528人, 99.6%)，腫瘤大小以2公分以下患者居多(4149人, 39.2%)，淋巴結侵犯呈陰性患者多於陽性患者，侵襲癌多於原位癌，治療情形以手術為主，其次依序為化學治療、荷爾蒙治療與放射線治療。12項核心測量項目平均適用率52.9%，平均符合率67.4%，依照病理期別建議之核心測量組合遵從度平均45.37%呈左偏分布，第一期患者多達到50%~75%的遵從度，第零、二、三期患者所接受照護多以25~50%遵從度居多，第四期遵從度最低，大多數僅接受低於25%核心測量項目所建議照護服務。(四)影響核心測量遵從度之因素包括：病人特性之發病年齡、就醫機構數、病理期別、腫瘤大小、淋巴侵犯情形、及腫瘤性態六項；醫院特性之評鑑等級、地區別、乳癌每月門急診服務量、手術服務量、放射線治療服務量、化學治療服務量、及荷爾蒙治療量七項。(五)與病患存活情形有關的預測因子包括：病人特性之發病年齡、病理期別、腫瘤大小、淋巴結侵犯情形及核心測量組合遵從度，在醫院特性部分與病人存活未達顯著差異。 結論：乳癌核心測量指標可以預測病患存活情形，利用以個案為基礎之測量方法，可以補足總體測量下只能看到某一小部份照護服務的情形，進而協助研究者了解每位患者從發病開始到治療結束後續追蹤一系列接受照護服務情形，本研究結合總體測量及個案為基礎之測量方式，發現核心測量項目遵從率高低不一及部分核心測量項目對患者存活達顯著影響，相關單位可針對較低核心測量項目加以宣導及改善，並加強醫療服務偍供者對核心測量指標之認知及態度，未來核心測量指標可改善病患健康的結果將指日可待。

Background: The Bureau of Health Promotion (BPH) in Taiwan started to collect six cancer treatment data included breast cancer, cervical, colorectal, oral, lung, and liver cancer at Taiwan Cancer Data Base (TCDB) from hospitals since year 2003. In order to improve care quality and benchmark among cancer care hospitals, the BHP supported a project from year 2004 to 2006 for quality improvement mechanism that included establishing six cancer core measure sets by research team of College of Public Health (It was the pre-phase of the Center for Health Insurance Research), National Taiwan University. There are more breast cancer researches than other kinds of cancers. However, most researches using quality measures are population-based to investigate the proportion of received eligibility care over total number of patients. However, it can’t point out each case whether he/she gets the eligible care or not. There are limited searches to discuss the relationship between the core measure set compliance rate with patient survival. This study attempt to using breast cancer core measure set to realize the care service offered at present, to explore the factors of patient-level compliance, and to explore the relationship between core measure compliance rate and patient survival. Method: Using TCDB and NHI database to collect patients’ demographic and clinical characteristics, hospital characteristics and volume, as well as the eligibility and compliance numbers of patients among 12 breast cancer core measures included pre-treatment (2), treatment (9) and follow-up (1) items. This study performed a retrospective cohort study of all 10,579 breast cancer patients registered from 29 hospitals in Taiwan between 2002 and 2004, and connected to these cases survival data between 2002 and 2005. To complement the population-based measurement, this study use case-based approach to find out the care services each patient received in the whole care process from pretreatment to follow-up. The log-rank test and chi-square are used to compare the survival contribution and compliance rate. The regression model is used to predict the factors of the core measure set compliance at each patient. To adjust for potentially confounding variables, this study used a Cox proportional hazards model as multivariate analysis to examine relationships between patient-level compliance and survival. Result: (I) Hospital characteristics: These cases are registered from 17 medical centers and 12 regional hospitals, and most registered hospitals are located in Taipei metropolitan (12, 41.4%). Among this hospitals, the medium surgery volume per year is 95(range from 9 to 476); the medium radiation therapy volume per year is 35(range from 3 to 244); the chemotherapy volume per year is 68(range from 1 to 382); the mean hormone therapy volume if 81(range from 14 to 280). (II)Patient characteristics: most patients’ diagnosed age is between 40 to 50, with a mean age of 51; most of them received cancer care in one hospital after diagnosed and most of these hospitals breast cancer ambulatory and emergency volume is between 1000 to 2000 cases per month. During the research period, the mean survival follow-up time is 25.4 month and the survival rate is 92.7% (n=9810) in the end of 2005. (III)Clinical characteristics: Most patients are stage II (43.3%); most diagnostic confirmation used positive histology or cytology (99.6%); most tumor size under 2cm (n=4149, 39.2%); most node negative is more than positive; invasive carcinoma is more than carcinoma in situ; the most treatment type is surgery, followed by chemotherapy, hormone therapy and radiation therapy. The core measure items mean eligibility rate is 52.9% and the mean compliance rate is 67.4%. The core measure set mean compliance rate is 45.37% according to pathology staging. Most stage I patient compliance rate is 50% to 75%; most Stage0, II, III patient compliance rate is 25% to 50%; most stage IV patient compliance rate is under 25%. (IV) The factors related to patient-level compliance rate included diagnosed age, the number of hospital patient visits, pathology stage, tumor size, node involvement, tumor behavior, hospital accreditation level, location, ambulatory and emergency volume, as well as the volume of surgery, radiation therapy, chemotherapy, and hormone therapy. (V) The factors related to patient survival are diagnosed age, pathology stage, tumor size, node involvement, and core measure set compliance. The relationship between hospital characteristics and patient survival is not significant in this study. Conclusion: Breast cancer core measure set can predict patient survival that the higher the patient-level compliance rate, the better the survival. This study combines population-based and case-based measurement to explore the breast cancer care at present and find the well performed items needed to maintain and deficiency items needed to be improved. It also finds some core measure items are related to the patient survival. The health authorities and providers may follow these recommendations according to the results and the patient survival will be enhanced in the future.