Telemedicine for Surgical Site Infection Diagnosis in Rural Rwanda: Concordance and Accuracy of Image Reviews

Abstract

Background In rural Africa where access to medical personnel is limited, telemedicine can be leveraged to empower community health workers (CHWs) to support effective postpartum home‐based care after cesarean section (c‐section). As a first step toward telemedicine, we assessed the sensitivity, specificity, and interrater reliability of image‐based diagnosis of surgical site infections (SSIs) among women delivering via c‐section at a rural Rwandan Hospital. Methods Women ≥18 years who underwent c‐section from March to October 2017 at Kirehe District Hospital (KDH) were enrolled. On postoperative day 10 at KDH, participants underwent a physical examination by a general practitioner, who provided a diagnosis of SSI or no SSI. Trained CHWs photographed patients’ incisions and the collected images were shown to six physicians, who upon review, assigned one of the following diagnoses to each image: definite SSI, suspected SSI, suspected no SSI, and definite no SSI, which were compared with the diagnoses based on physical exam. We report the sensitivity and specificity and assessed reviewer agreement using Gwet's AC1. Results 569 images were included, with 61 women (10.7%) diagnosed with an SSI. Of the 3414 image‐reviews, 49 (1.4%) could not be assigned diagnoses due to image quality. The median sensitivity and specificity were 0.83 and 0.69, respectively. The Gwet's AC1 estimate for binary classification was 0.46. Conclusions We demonstrate decent accuracy but only moderate consistency for photograph‐based SSI diagnosis. Strategies to improve overall agreement include providing clinical information to accompany photographs, providing a baseline photograph for comparison, and implementing photograph‐taking processes aimed at improving image quality.

mHealth-community health worker telemedicine intervention for surgical site infection diagnosis: a prospective study among women delivering via caesarean section in rural Rwanda

“…The telemedicine SSI diagnosis had a low sensitivity but a high specificity, which was similar to findings when evaluating the wound photography for remote postoperative SSI diagnosis in high-income settings 42. This low sensitivity can be attributed to poor quality of wound image or the ability and confidence of the GP to consistently evaluate the wound status based on images alone, which can change from one to another and from time to time 43 44. While the sensitivity is low, the high negative predictive value could help reduce unnecessary postoperative clinic visits.…”

“… 42 This low sensitivity can be attributed to poor quality of wound image or the ability and confidence of the GP to consistently evaluate the wound status based on images alone, which can change from one to another and from time to time. 43 44 While the sensitivity is low, the high negative predictive value could help reduce unnecessary postoperative clinic visits. This could relieve financial and travel hardships for these vulnerable patients and workload for clinicians, as has been shown in other studies.…”

Prevalence of post‐caesarean section surgical site infections in Rwanda: A systematic review and meta‐analysis

“…Nine studies were cohort studies, 1,9,15–17,19–21,24 six were randomized controlled trials 10,13,18,22,25,26 and two were cross‐sectional studies 14,23 . Five studies had sample sizes above 650 participants 9,13,16,21,26 while 12 studies had sample sizes below 650 participants 1,10,14,15,17–20,22–25 . Eight studies were published from 2021 to 2023 9,13,14,16,17,20,21,24 while nine studies were published before 2021 (from 2020 to 2014) 1,10,15,18,19,22,23,25,26 .…”

“…The included studies were published from 2014 to 2023. A total of 10 studies were conducted at Kirehe District Hospital, 1 , 9 , 13 , 15 , 16 , 17 , 18 , 19 , 20 , 21 two studies at University Teaching Hospital of Kigali, 10 , 22 and one study at University Teaching Hospital of Butare, 23 Kibungo District Hospital, 24 Ruhengeri Referral Hospital, 25 Kabgayi Hospital 14 and Bushenge Provincial Hospital each. 26 Nine studies were cohort studies, 1 , 9 , 15 , 16 , 17 , 19 , 20 , 21 , 24 six were randomized controlled trials 10 , 13 , 18 , 22 , 25 , 26 and two were cross‐sectional studies.…”

Image-based surgical site infection algorithms to support home-based post-cesarean monitoring: Lessons from Rwanda

“…Over the last eight years, our research team has explored several strategies for sustainable, home-based, CHW-supported care for women after cesarean delivery in rural Rwanda. We have demonstrated that home-based follow-up by CHWs is feasible and acceptable [5,6], but our attempts for simple clinical screenings or sending pictures to general practitioners has resulted in low SSI diagnostic accuracy [5][6][7][8]. In contrast, our machine learning image-based diagnostic algorithms are yielding higher accuracy: 81.3% sensitivity and 65.3% specificity for visible image algorithms [9], increasing to 97% sensitivity and 87% specificity when using computer vision techniques and color calibration in the image processing pipeline, and 95% specificity and 84% specificity for thermal image algorithms [10].…”