Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID- 19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active. diabetes research and clinical practice 169 (2020) 108396 Contents available at ScienceDirect Diabetes Research and Clinical Practice journal homepage: www.elsevier.com/locat e/dia bre s Brief Report Managing gestational diabetes mellitus using a smartphone application with artificial intelligence (SineDie) during the COVID-19 pandemic: Much more than just telemedicine Lara Albert a,*, Ismael Capel a, Gema Garcı́a-Sáez b,c, Pablo Martı́n-Redondo b, M. Elena Hernando b,c, Mercedes Rigla a a Endocrinology and Nutrition Department, Parc Taulı́ Hospital Universitari, Institut d’Investigació i Innovació Parc Taulı́ I3PT, Universitat Autònoma de Barcelona, Parc Taulı́ 1, 08208, Sabadell (Barcelona), Spain b Bioengineering and Telemedicine Group, Centro de Tecnologı́a Biomédica, ETSI de Telecomunicación, Universidad Politécnica de Madrid, Avda. Complutense 30, 28040, Madrid, Spain c CIBER-BBN: Networking Research Centre for Bioengineering, Biomaterials and Nanomedicine, Ciudad Universitaria, s/n, 28040, Madrid, Spain A R T I C L E I N F O A B S T R A C T Article history: We describe our experience in the remote management of women with gestational dia- Received 10 June 2020 betes mellitus during the COVID-19 pandemic. We used a mobile phone application with Received in revised form artificial intelligence that automatically classifies and analyses the data (ketonuria, diet 4 August 2020 transgressions, and blood glucose values), making adjustment recommendations regarding Accepted 26 August 2020 the diet or insulin treatment. Available online 3 September 2020 Ó 2020 Elsevier B.V. All rights reserved. Keywords: Gestational diabetes mellitus Telemedicine eHealth Mobile phone application Artificial intelligence Gestational diabetes mellitus (GDM) is a frequent disease of perinatal complications. Its management involves frequent and if the glucose control is inadequate, it increases the risk hospital visits to check self-monitoring data (glycemia levels, * Corresponding author at: Endocrinology Department, Parc Taulı́ University Hospital Parc Taulı́ 1, 08208 Sabadell (Barcelona), Spain. E-mail addresses: lalbert@tauli.cat (L. Albert), icapel@tauli.cat (I. Capel), gema.garcia.saez@upm.es (G. Garcı́a-Sáez), pablo.martin. redondo@alumnos.upm.es (P. Martı́n-Redondo), mariaelena.hernando@upm.es (M.E. Hernando), mrigla@tauli.cat (M. Rigla). https://doi.org/10.1016/j.diabres.2020.108396 0168-8227/Ó 2020 Elsevier B.V. All rights reserved. 2 Table 1 – Summary of Sinedie app data received from GDM women. diabetes research and clinical practice Start date Id Number of days using the appa Total BG uploaded (BG/day) Glucose levelsb FBG / 1hBG App therapy change proposals First 3 days Last 3 days Diet changes Start insulin therapy Adjust insulin 31/03 1 43.3 168 (3.9) 84 / 115 87 / 103 1 – – 06/04 2 3.5 23 (6.5) 103 / 125 – – – – 08/04 3 36.6 147 (4.0) 82 / 119 73 / 104 – 1 – 08/04 4 36.6 146 (4.0) 94 / 123 96 / 120 – 1 4 08/04 5 12.0 52 (4.3) 85 / 130 92 / 134 – 1 – 08/04 6 37.1 164 (4.4) 84 / 129 84 / 112 – 1 2 08/04 7 36.0 144 (4.0) 89 / 132 91 / 131 – 1 6 08/04 8 37.0 151 (4.1) 93 / 136 96 / 141 – 1 6 09/04 9 22.2 89 (4.0) 83 / 126 92 / 120 – – – 16/04 10 29.0 115 (4.0) 87 / 132 93 / 118 1 1 – 22/04 11 22.6 91 (4.0) 85 / 127 93 / 126 2 1 – 28/04 12 16.5 36 (2.2) 83 / 115 78 / 117 – – – 28/04 13 16.5 71 (4.3) 97 / 124 95 / 138 – 1 1 29/04 14 15.4 71 (4.6) 92 / 128 89 / 122 – 1 – 169 (2020) 108396 29/04 15 4.6 17 (3.7) 96 / 122 – – 1 – 29/04 16 15.5 55 (3.6) 86 / 123 89 / 120 1 – – 06/05 17 8.6 36 (4.2) 71 / 128 71 / 111 – – – 06/05 18 8.6 37 (4.3) 91 / 122 86 / 124 – 1 – 12/05 19 2.5 12 (4.7) 82 / 96 – – – – 12/05 20 2.1 10 (4.8) 64 / 104 – – – – BG, blood glucose measurement; FBG, fasting blood glucose; 1hBG, 1-h postprandial blood glucose a From 31 March to 14 May 2020 b Mean glucose levels (mg/dL) during the first 3 full days vs. last 3 full days. If there is not enough data, the field is blank. diabetes research and clinical practice 169 (2020) 108396 3 fasting ketonuria, adherence to the diet). Throughout the condition with a code of colors (normal, altered or signifi- follow-up clinicians decide whether any treatment adjust- cantly altered) and points out the patients with treatment ment is necessary. proposals awaiting for validation. When a patient is selected, Telemedicine is commonly defined as health services pro- the professional has access to all the data registered, treat- vided by healthcare professionals using remote communica- ment details, and proposals about treatment changes. When tion technologies and it includes a wide selection of the system generates a recommendation about insulin ther- modalities. In the field of GDM, telemedicine can reduce the apy, a proposal of daily insulin prescription is generated: need for hospital outpatient visits, allowing remote patient amount and type of insulin, and moment of administration. monitoring and requiring face-to-face visits only if insulin The clinicians can accept, modify, or reject the proposals. Ini- treatment is initiated [1]. A recent meta-analysis of 32 ran- tiation of insulin prescription requires to schedule a face-to- domized controlled trials (5108 patients) showed that com- face visit to give advice about insulin therapy management, pared to standard care, the use of telemedicine could while subsequent modifications are managed through the improve glycemic levels of women with GDM more effectively app after approval of the clinician. When a treatment change and reduce the risk of pregnancy complications [2]. is made, it is automatically notified to the patient with a pop- Some limitations of telemedicine include usability, real- up message. time feedback, a more significant workload (as it generates From 31 March to 14 May 2020, 20 pregnant women with a greater amount of data), and decision-making capabilities GDM started using the Sinedie app. They were followed dur- (both for patients and clinicians). Decision support tools inte- ing a median of 16.5 days (8.6–36.2; Q25-Q75). Mean ± SD grated into telemedicine systems can help clinicians with the number of daily glucose measurements were 4.2 ± 0.8 and analysis of the information generated. In 2016 we developed a data transfer frequency was 0.9 ± 0.2 upload/24 h (342 uploads web-based telemedicine platform (SineDie) that operated as a in 7 weeks). Mean ± SD fasting blood glucose and 1- clinical decision support system designed to manage GDM hour postprandial blood glucose were 89 ± 12 mg/dL and treatment. SineDie automatically prescribed diet therapy 122 ± 23 mg/dL (breakfast), 123 ± 21 mg/dL (lunch), modifications, identified the need for insulin treatment and 122 ± 22 mg/dL (dinner). Table 1 includes mean of fasting proposed insulin dose changes to doctors. We performed a blood glucose and 1-hour postprandial the first three days randomized clinical trial that showed an 88.6% reduction of using the app and the last three days, for patients followed the face-to-face visits and a 27.4% reduction of the time for more than 8 days. During the follow-up, the system gener- devoted by clinicians to patients’ evaluation. The system ated diet adjustments in 20% of the patients, 12 patients ini- detected all situations that required therapy adjustment, gen- tiated insulin treatment and 41.7% of them required therapy erating safe recommendations [3]. However, when using the adjustment afterwards (Table 1). 45.2% of insulin proposals web-based system some patients experienced downloading generated by the system were accepted, 29.0% postponed problems with the glucose meter, patients had to be trained and 25.8% rejected. Before publication, we reviewed maternal very well before using it and they needed access to the com- and neonatal outcomes: 18 women gave birth at a median of puter for data registration. Lately, we have developed the 39 weeks of gestation (38–40; Q25-Q75). Two women had mobile version (only available on Android smartphones) to planned caesarean section delivery, one for breech presenta- make the tool more user-friendly and easy to use, and we tion and the other due to intrauterine growth restriction are about to start the clinical study. (IUGR). Four women had c-section delivery after failure of The current COVID-19 pandemic has brought telemedi- labor induction. None had preeclampsia. None of the new- cine in the spotlight [4,5]. In such an exceptional situation, borns were large for gestational age (median percentile we started using the SineDie app for the remote evaluation weight was 42 [32.5–58; Q25-Q75]. Newborns did not present of all GDM patients visited in our Endocrinology department hypoglycemia however, six of them had jaundice and three (after they signed an informed consent form). Women with of them were admitted to neonatal intensive care unit due GDM using the SineDie app follow the same self-monitoring to preterm birth (IUGR in a twin pregnancy and a premature guidelines as in standard care, but instead of going to the rupture of membranes). weekly outpatient visit, they send their data to the app. Currently, we might have to wait a long time before the Their glycemia data is uploaded through Bluetooth connec- routine outpatient clinic follow-up is restored. We must adapt tivity from the glucose meter while ketonuria, lack of com- to the new situation and leverage the potential of telemedi- pliance to dietary, and insulin treatment can be entered cine resources. The use of telemedicine in the usual care manually. The system automatically analyzes the data and entails a paradigm shift and its use could reduce the impact the patients receive instant feedback after each data of a pandemic to healthcare. SineDie app may be an excellent upload: Nutritional advice about carbohydrates intakes is tool to prevent unnecessary hospital visits while keeping the prescribed to the patient in the therapy scenario of the best quality healthcare and reducing clinicians’ workload for app and if insulin treatment is needed, the physician is GDM management. notified. When no data is received during the last three days, the app sends a reminder to insist on the importance of frequent data transmission. Declaration of Competing Interest The clinicians can access the professional SineDie app as many times as needed and can contact by telephone, send a The authors declare that they have no known competing text message with the app, or schedule a face-to-face visit if financial interests or personal relationships that could have necessary. The app shows the patients’ current metabolic appeared to influence the work reported in this paper. 4 diabetes research and clinical practice 169 (2020) 108396 Funding Diabetes Mellitus after a Telecare Approach Are Not Inferior to Traditional Outpatient Clinic Visits. Int J Endocrinol 2010;2010. This work is supported by Instituto de Salud Carlos III (DTS https://doi.org/10.1155/2010/386941. [2] Xie W, Dai P, Qin Y, Wu M, Yang B, Yu X. Effectiveness of 16/00010) and a Menarini scientific grant. telemedicine for pregnant women with gestational diabetes mellitus: an updated meta-analysis of 32 randomized Author Contributions controlled trials with trial sequential analysis. BMC Pregnancy Childbirth 2020;20:198. https://doi.org/10.1186/s12884-020- L.A. made clinical visits where GDM patients were taught 02892-1. about Sinedie app and wrote the manuscript. M.R. con- [3] Caballero-Ruiz E, Garcı́a-Sáez G, Rigla M, Villaplana M, Pons B, Hernando ME. A web-based clinical decision support system tributed to the development of the Sinedie app and critically for gestational diabetes: Automatic diet prescription and revised the manuscript. I.C. made critical revisions. G.G-S, P. detection of insulin needs. Int J Med Inform 2017;102:35–49. M R and M.E.H contributed to the development of the Sinedie https://doi.org/10.1016/j.ijmedinf.2017.02.014. app, analyzed the data and revised the document. M.R. is the [4] Hollander Judd E, Carr Brendan G. Virtually Perfect? guarantor of this work and, as such, had full access to all the Telemedicine for Covid-19. N Engl J Med 2020;382:1679–81. data in the study and takes responsibility for the integrity of https://doi.org/10.1056/NEJMp2003539. the data and the accuracy of the data analysis. [5] Garg SK, Rodbard D, Hirsch IB, Forlenza GP. Managing New- Onset Type 1 Diabetes During the COVID-19 Pandemic : Challenges and Opportunities. Diabetes Technol Ther 2020;22:1–9. https://doi.org/10.1089/dia.2020.0161. R E F E R E N C E S [1] Pérez-Ferre N, Galindo M, Dolores Fernández M, Velasco V, Runkle I, José de la Cruz M, et al. The Outcomes of Gestational