以兒科加護病房護理表單為基礎之護理診斷決策輔助系統
A Nursing Diagnosis Decision Support System Based On Pediatric Intensive Care Unit Nursing Forms
中文摘要
國內醫療院所自民國74 年開始推行 NANDA 護理診斷，藉此將護理過程的各個
環節整合至臨床病患照顧工作中，以提升護理品質；目前各家醫院所導入的護 理診斷資訊系統，多數僅將護理診斷內容及格式標準化，主要目的是取代人工 書寫作業，但護理問題相關資料之收集與整合仍需以人工方式實現，護理診斷 的確立也是基於護理人員自身的經驗判斷，較缺乏護理診斷推論的過程及精神。
因此，本研究期望能利用臨床護理表單提供的表單欄位內容，找出相對應的護 理診斷定義性特徵及導因，藉由系統自動轉換及對應機制，進行病患表單資料 的自動判斷及歸納，推論出與病患現況相符之定義性特徵，產生護理診斷建議 清單。
本研究對象為某醫學中心兒童加護病房，收集該單位三張紙本臨床護理表單，
及二十八個兒科常用護理診斷及定義性特徵。首先依據表單欄位特性將欄位重整 及編碼，接著擴展為表單欄位之臨床判斷，與意義相符之定義性特徵和導因進 行人工對應，形成護理診斷推論知識庫，最後以該知識庫為核心，建立虛擬之
護理診斷決策輔助系統。系統評估則輸入35 位兒童加護病房病患之真實數據，
經系統進行分析及推論，得到各病患之護理診斷建議清單，並請護理專家檢視 後以五分法進行護理診斷適當性評分，另列出各病患需以手動方式加入之護理 診斷，計算自動推論與手動加入之護理診斷的比例。
本護理診斷決策輔助系統之功能包括: (1) 對病患表單欄位資料進行分析並提出 可成立之臨床判斷； (2) 依據成立之臨床判斷列出相對應之定義性特徵及護理 診斷；(3)利用系統條件限制產生護理診斷建議清單；且使用者可彈性調整建議 清單內容並決定是否接受。系統評估由三位護理專家進行，護理診斷適當性評分
同意以上得分的平均有91%，不同意以下得分的平均有 4%；而系統經由自動推
論可得到95%的護理診斷，5%須透過手動方式加選。因此，本研究之護理診斷
決策輔助系統利用護理診斷推論知識庫可以推論出護理診斷建議清單，更可以 提供護理診斷成立的確切證據。
英文摘要
Since 1985，Taiwan’s hospitals started to carry out the NANDA nursing diagnosis, and linked the diagnosis into the workflow of nursing care to promote the quality of care. However, most hospitals implementing their nursing information systems only standardized forms and formats of nursing diagnoses for substitution of manually
hand-writing work. They collect data manually, and make nursing diagnosis based on their own experience, not based on the deduction rules as those in the NANDA specification. Therefore, the goal of this research is to produce a nursing diagnose using the field data in nursing forms used to collect data routinely by clinical nurses.
In the approach presented here we map nursing form fields into the corresponding nursing diagnosis definition characteristics, and then map the definition characteristics and other patient present conditions to a suggestion list of plausible nursing diagnoses for user selection.
In this thesis we uses a pediatric intensive care unit in a medical center as a base for field research. We collect three clinical nursing forms and 28 nursing diagnoses and their definition characteristics commonly used in the pediatric department. After form item rallying and code, extends the form item becomes the clinical judgment, then manual corresponds the clinical judgment and the definition characteristic that have the same significance. These correspondences will form the nursing diagnosis decision knowledge base; finally, take it as the core to implementing the nursing diagnosis decision support system. The system evaluation inputs 35 case clinical nursing form data, and analysis those data by the system, to obtain nursing diagnosis suggestion list for each case; next, the nursing expert evaluate suggestion list to be whether suitable by five-level Likert scale, and records the nursing diagnosis by the manual way for computation automatic deduction to manual joins proportion.
There support system was composed three functions: 1. Transforms clinical form data become the clinical judgment; 2. lists the definition characteristic and the nursing diagnosis which corresponds with the clinical judgment; 3. proposes the nursing diagnosis suggestion list. The user can adjust suggestion list and decided whether to accept. The system evaluation by three nursing experts; suggestion list suitability was 91% above agrees, 4% under disagree; the system can obtain 95% nursing diagnosis by way of the automatic deduction, 5% by the manual way. Therefore, this system using the nursing diagnosis decision knowledge base, can effectively assist nursing diagnosis decision-making, and also provide the evidence for the patient’s nursing diagnosis.
