以醫療器材生命週期所建立之單臂式ECG與PPG技術在非加壓睡眠血壓量測研究

A Study of Single Arm ECG and PPG Wearable Device for Cuff-less Sleep Blood Pressure Measurement Based on Medical Device Life Cycle

陳美芬(Mei-Fen Chen)

指導教授 ： 林康平

中原大學/電機資訊學院/電機工程研究所/博士(2021年)

https://doi.org/10.6840/cycu202100660

摘要

本論文基於睡眠血壓監測需求，提出一種以ECG與PPG量測技術為基礎的單臂式非加壓睡眠血壓量測裝置設計，同時以醫療器材生命週期為框架，以本論文所提出的技術為標的，探討醫療器材在技術開發過程中應建立的重要確效任務，期望以實際技術為案例，更具體的演繹醫療器材，從技術發展到產品上市的轉化路徑。
本論文克服單臂式ECG與PPG訊號微弱之困難，藉由適當的電路匹配與高增益設計，將現有臂式ECG測量從三電極降為雙電極，在單一手臂式上實現雙電極ECG測量與PPG測量，進而完成基於PAT原理之睡眠血壓量測裝置，同時驗證了單臂式ECG與PPG訊號對於PAT血壓估測理論的適用性。
對於非加壓式血壓裝置而言，校正程序是決定血壓數值表現的關鍵，因此在系統驗證部分，本論文以本實驗室已建立的多項式血壓回歸公式，搭配五種校正模型(校正模型I~III、IA、IB)，在以市售血壓計量測值為參考依據下(參考血壓)，分別進行靜態血壓與長時間睡眠血壓二種估測實驗。
就靜態血壓估測比對結果，共收錄10位受測者資料，與參考血壓相較，可以達到收縮壓3.5±3.8 mmHg、舒張壓3.0±3.3mmHg的平均絕對差，R2分別為0.94與0.89，顯示所使用回歸公式與校正模型對於靜態血壓估測具有良好的配適度。
就長時間睡眠血壓估測部分，共收錄8位受測者資料，使用5種校正模型進行分析，結果顯示，以校正模型III對於參考血壓的估測一致性最佳，但對於以靜態參考血壓數值為基礎的校正模型IA，透過適當的調整，可以將平均絕對差進一步改善至收縮壓7.8 mmHg、舒張壓5.2mmHg，接近IEEE 1708規範所訂之要求，顯示此種校正模型後續應用於無袖式血壓校正程序的潛力。
“確效”是醫療器材產品發展過程中重要的任務之一，也是主管機關對於醫療器材上市審查重要的依據。本論文依據醫療器材生命週期，以本論文所進行之睡眠血壓量測技術進行產品假設，並依據ISO 14971、IEC 62304以及ISO 27005等國際規範，就風險分析、軟體確效驗證與資訊安全分析三項確效任務進行討論，同時產出報告範本。整體風險分析、軟體確效內容與資訊安全內容，事實上須相互連結且支持，本論文依照假定設計規格，共列出21項測試內容(V-A01~ V-E02)、鑑定出18項風險(R1-R18)以及5項資訊安全風險來源。雖然整體範本文件是以假定的產品規格進行撰寫，未有真實測試數據，但藉由本論文的整理，將醫療器材法規要求以具體文件呈現，對於醫療器材技術實務發展具有重要參考價值。

關鍵字

無袖式血壓 ； 單臂式心電圖 ； 醫療器材生命週期 ； 醫療器材確效

並列摘要

The aim of this study was to design a cuff-less sleep blood pressure (BP) measurement device, called SArm-BP, based on single-arm electrocardiography (ECG) and photoplethysmography (PPG) technology. The necessary validation steps based on the medical device development life cycle are discussed.
This study develops a circuit that allows the number of single-arm ECG electrodes to be reduced from three to two. The ECG and PPG signals are detected simultaneously on the signal arm for cuff-less sleep BP monitoring.
Calibration is important for the accuracy of cuff-less BP devices. To evaluate the performance of SArm-BP, a static test and a sleep test were carried out. A cuff-based reference BP monitor was placed on a subject’s right arm and SArm-BP was placed on the left arm. Systolic BP (SBP) and diastolic BP (DBP) for SArm-BP were estimated separately using a polynomial regression formula developed in our lab and the five calibration models (Models I-III, IA, and IB) presented in this paper. The mean absolute difference (MAD) of SBP and DBP measurements between SArm-BP and the reference BP device was calculated. According to the cuff-less BP standard (IEEE 1708), the MAD should be within 7 mmHg for both SBP and DBP measurements. The coefficient of determination (R2) was also used to evaluate the performance of SArm-BP.
Ten subjects participated in the static test. The MAD values were 3.5±3.8 and 3.0±3.3 mmHg and the R2 values were 0.94 and 0.89 for SBP and DBP, respectively. The results show that good static BP estimation can be achieved with the polynomial regression formula and the correction model used for the static test.
Eight subjects participated in the sleep test. The estimation results for five correction models were compared. The graphs show that Model III gives the best BP estimation performance based on the sleep reference BP monitor, followed by Model IA based on the static reference BP monitor. The MAD values between measurements taken using SArm-BP with Model IA and the reference BP device were 7.8 mmHg for SBP and 5.2 mmHg for DBP, which are close to the requirements of IEEE 1708. This comparison shows the potential of using the static test for cuff-less BP calibration.
Validation is an important step in the medical device development life cycle for regulatory approval. This study proposes an example medical product based on SArm-BP technology. ISO 14971, IEC 62304, and ISO 27005 are used as medical device validation standard to complete three document templates, including software report, risk management report, and cybersecurity management report.

This study completed a cuff-less sleep BP measurement device prototype and discussed the necessary validation steps based on the medical device life cycle. These preliminary results provide important information for further BP medical device development.

並列關鍵字

Cuff-less Blood Pressure ； Single-arm ECG ； Medical Device Life Cycle ； Medical Device Validation

參考文獻

[1] Collaboration, P.S., Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. The Lancet, 2002. 360(9349): p. 1903-1913.

[2] Organization, W.H., The top 10 causes of death. 2020.

[3] Association, A.H. What is High Blood Pressure? ; Available from: https://www.heart.org/en/health-topics/high-blood-pressure/the-facts-about-high-blood-pressure/what-is-high-blood-pressure.

[4] Association, A.H., Understanding Blood Pressure Readings.

[5] Dolan, E., et al., Superiority of ambulatory over clinic blood pressure measurement in predicting mortality: the Dublin outcome study. Hypertension, 2005. 46(1): p. 156-161.