手術室為特殊醫療單位,排班除了考量單位政策、人力需求配置,同時必須顧慮護理師個人需求。手術室護理排班問題限制條件多且其複雜性高。手術室和一般病房護理排班不同,手術室有多性專科手術和班別不同限制等問題,手術變數多護理師無法逐一安排。目前排班為人工排班表,對於排班者而言相當耗時且繁瑣,最後排班結果仍無法完全符合護理人員需求,因人工排班表未檢定無法確保公正性、公平性和正確性。 本研究探討手術室護理排班問題之目的,建置排班系統確保排班表之公正性、公平性和正確性,符合護理師需要,提升排班之滿意度及護理品質;在護理主管快速管控人力配置,有效縮短排班時間,提升排班管理效能。 本研究首要先蒐集相關文獻,依護理師之專長能力進行分群,進而設計層次分析法調查護理師班別認知、工作負擔、值班津貼之權重;排班前設定專科分群和班別權重值,再加入排班軟硬限制式,建置基因演算法模型進行求解最佳班表。本研究驗證結果證明基因演算法班系統的求解速度快;加入班別權重值之基因演算結果排班表滿意度調查,人員滿意程度較優於人工排班表。
Nurse scheduling is a difficult task. It not only has to meet the requirements of departmental policy and staff allocation, but also has to put personal needs into consideration. There are many rules, prerequisites and limitations in Operation Room nurse scheduling, and these requirements are quite complex. Nurse scheduling for Operation Room is more challenging than that for general wards because emergent surgeries need to be conducted whenever is needed. Currently, nurse shifts are scheduled manually in most medical institutions. It appears to be quite a time-consuming and verbose process for any scheduler because when drafting the duty roster, she or he needs to attend to a lot of detailed aspects, such as continuous working hours, manpower demand, vacation allotments, equality of all kinds of shift, personal preferences to particular shifts, professional specialty considerations, crisis management capabilities of personnel, and allowances. And yet the first few versions are far from finalized. It still has a lot to be further adjusted, because the roster is not fully in line with the demand of nurses. Most importantly it is difficult or even possible to guarantee fairness and reliability of the schedule, since there is no effective way to verify the manually generated roster. This study aims to build a nurse scheduling system for Operation Room to ensure the fairness and reliability of the duty roster, meet the personal needs of individual nurses, and furthermore indirectly enhance nurses’ performance and their service quality. Since the system generates the duty roster automatically, it will help supervisors save a lot of time usually spent up to 8 hours when they go through a lengthy process to schedule a roster by hand. In this way, they can quickly and effectively manage staff allocation, which also means more time and energy to fulfill their management duty. A genetic algorithm model was utilized to obtain the optimal schedule. Nurses were first divided into four distinct groups based on their professional expertise. Three main factors were weighted into the model, namely perception, workload and duty fee, and Analytic Hierarchy Process (AHP) was applied to determine their weighs. In addition, several soft and hard constraints suggested from literature were also included in the model. The model was then validated with a field test and the superiority of the automatically generated schedule over its manual opponent was also confirmed by a satisfaction survey.