外科醫師學習效果再探討─工作地點移轉、先前經驗與醫師學習適應期之研究

Abstract

此論文包含三篇探討外科醫師學習效果的實證文章： 第一篇文章：(本文第二章) 外科醫師學習效果之分析－兼論工作地點移轉對於醫師學習之影響 本研究使用1997-2008年之全民健保資料，以罹患子宮肌瘤並進行剖腹式子宮全切除手術之台灣婦女作為研究對象，探討外科醫師手術量與手術績效之關聯，並著眼於轉院醫師與非轉院醫師的差異性探討，以及層級別與移轉前後之學習速度變化。研究結果顯示轉院醫師與非轉院醫師皆呈現顯著之正向學習，且兩類醫師之學習速度在移轉之前並無顯著差異。此外，不僅外科醫師之學習速度會因醫院層級不同而有所差異，轉院醫師在進行層級間的移轉時，層級變化對於醫師的學習速度亦產生增強的效果；不過，移轉之後呈現顯著學習效果差異者，僅侷限於向上層級移轉(區域醫院轉至醫學中心，抑或是地區醫院轉至區域醫院或醫學中心)的醫師。 第二篇文章：(本文第三章) 外科手術績效影響因素分析─醫師先前經驗與醫院層級之考量 本研究援用學習型組織中個人技術精進的概念，並以人力資源發展理論為基礎，透過外科醫師技能精進的角度來檢視醫師樣本期間以外之先前經驗以及醫院層級等因素對於手術績效的影響。使用全民健保剖腹式全子宮切除手術資料所得之實證結果顯示，不論資歷深淺，醫學中心醫院的醫師皆能透過手術次數的累積改善其手術效果；就先前經驗的影響而言，則只有資淺醫師方能在不同層級的醫院皆呈現顯著的學習效果。此結果說明醫師之手術經驗與其執刀醫院的類型確實為影響其手術績效的重要關鍵。子宮切除手術效果的衡量與比較，很早便受到國外醫學界的重視，相關實證研究並證實該項手術之手術效果存在跨國家別、跨州別、跨醫院服務區域別以及跨醫院之差異，而過去相關研究多由手術量與醫療效果的關係來解釋這些差異的存在。本研究之實證結果除了說明醫師先前經驗與醫院規模對於外科醫師學習與醫療效果的影響，亦可進一步為不同組織所呈現的學習速度差異提供一個合理的解釋。 第三篇文章：(本文第四章) 外科醫師學習適應期之研究─以剖腹式與腹腔鏡全子宮切除手術為例 本篇論文有別於過去衡量醫師學習效果之實證研究，嘗試提供一個新的衡量方式來探討醫師的動態學習效果，並以執業於醫學中心之方取得專科執照醫師為研究對象，運用滾動迴歸估計，嘗試估計醫師動態學習過程，並驗證醫師學習適應階段以及S型學習曲線是否存在。本研究主要使用1997-2008年之全民健保資料，以罹患子宮肌瘤並進行剖腹式子宮全切除手術(TAH)以及腹腔鏡全子宮切除手術(LH)為研究對象，刻劃工作於醫學中心、醫學中心之公立醫院、以及醫學中心之法人與私立醫院醫師的動態學習曲線，並比較兩類手術於不同層級別之學習適應期收斂速度消長。實證結果指出，無論TAH抑或LH手術，醫師學習皆存在一段為期頗長的學習適應期，換言之，有別於過去文獻L型學習曲線設定(僅具有成長期與成熟期)，本研究成功的證實了S型學習曲線真實存在於醫師學習行為中。此外，透過兩組手術動態學習曲線之比較，研究結果指出LH手術於適應期之收斂速度短於TAH手術，此結果亦與醫學文獻之結果一致。

This thesis includes three empirical studies on the learning effect of surgeons: Essay 1: (Chapter 2 in this study) An Analysis of Surgeons’ Learning Effect Accounting for the Impact of Workplace Switching Drawn from Taiwan’s National Health Insurance Database (NHID), data with inpatients underwent the uterine-leiomyoma-causing hysterectomy surgery during 1997-2008 are used to evaluate the association between surgical volume and outcome. We focus on examining the volume-outcome relationship from two distinct perspectives. The first involves the learning differences between switching and non-switching surgeons. The second concerns switching surgeons’ mean differences in learning rate during the pre- and post-switching periods. Our result confirms the persistence of performance improvement as surgeons’ trials accumulate, but find no significant learning differences between the non-switching and switching surgeons evaluated at the pre-switching period. Furthermore, in addition to hospital accreditation, an upward switching such as from the regional hospitals into the medical centers and from the district hospitals into the regional hospitals or medical centers, are found to explain the variations in learning rates among hysterectomy surgeons. Essay 2: (Chapter 3 in this study) Determinants of Surgical Performance: Taking Surgeons’ Prior Experiences and Hospital Accreditation into Account Adapting the “personal mastery” essence of the learning organization and based on the theory of human resource development, this study examines the determinants of the volume-outcome relationship for the abdominal total hysterectomy with a special emphasis on surgeons’ out-of-sample experience and hospital-specific capital. It is found that regardless of out-of-sample prior experience, hysterectomy surgeons in large hospitals on average exhibit improvement in surgical outcome with the increase in accumulated volume, and only those starting out as a junior physician during the sample period experience a positive learning effect in spite of affiliated hospital accreditation. The result not only demonstrates the significance of prior experience and hospital specialty on hysterectomy surgeons’ learning and surgical outcome, it also offers a plausible explanation to the organizational learning differences observed in previous research. Essay 3: (Chapter 4 in this study) Hysterectomy Surgeons’ Adaptive Learning Phase: Evidence from Abdominal and Laparoscopic Surgeries This study examines the adaptive learning phases of the dynamic learning process in hysterectomy surgeries. The data used in this study is taken from Taiwan’s National Health Insurance Database (NHID). Only inpatients underwent the hysterectomy surgery in medical center as a result of uterine-leiomyoma during the time period of 1997-2008 were selected into our final sample. Rolling-window OLS (ordinary least squares) estimation with consecutive windows of 25, 30, 35, 40, 45, and 50 patients, respectively, is used to illustrate surgeons’ S-shaped learning pattern. Taking the TAH and LH surgeries as an example, this study provides evidences supporting the presence of the adaptive phase in hysterectomy surgeons’ dynamic learning process. Moreover, it is found that the duration for slow learning is consistently long for the two types of hysterectomy surgery. A comparison of the abdominal as well as laparoscopic surgeries suggests that the duration of the adaptive phases for LH is shorter than that of the TAH.