Morten Hasselstrøm Jensen
Department of Health Science and Technology, Aalborg University, Denmark/Center for Information Technology Research In the Interest of Society, UC Berkeley, CA, USA
Jenna Hua
School of Public Health, UC Berkeley, CA,USA
Mette Dencker Johansen
Department of Health Science and Technology, Aalborg University, Denmark
Jay Han
Department of Physical Medicine and Rehabilitation, UC Davis School of Medicin, Sacramento, CA, USA
Gnangurudasan Prakasam
Center of Excellence in Diabetes & Endocrinology, Sacramento, CA, USA
Ole Hejlesen
Department of Health Science and Technology, Aalborg University, Denmark/Department of Health and Nursing Science, University of Agder, Norway/Department of Computer Science, University of Tromsø, Norway
Edmund Seto
Center for Information Technology Research In the Interest of Society, UC Berkeley, CA, USA/School of Public Health, UC Berkeley, CA, USA
Ladda ner artikel
Ingår i: Scandinavian Conference on Health Informatics 2013; Copenhagen; Denmark; August 20; 2013
Linköping Electronic Conference Proceedings 5:10, s. 45-49
Publicerad: 2013-08-21
ISBN: 978-91-7519-518-6
ISSN: 1650-3686 (tryckt), 1650-3740 (online)
Persons with Type 1 diabetes need continuous exogenous insulin
supply throughout their life. Determining the optimal
insulin treatment in relation to diet and physical activity is one
of the main goals of diabetes management; but is difficult;
especially for vulnerable populations; such as adolescents.
Erroneous treatment may result in both repeated and severe
low blood glucose events. Continuous glucose monitoring
(CGM) may help in avoiding these events; but is inaccurate
compared to traditional glucose monitoring. Models have
been developed to significantly improve C’s detection of
insulin-induced events by using information from the CGM
signal itself. Additional temporal data on insulin doses; diet
and physical activity may improve hypoglycaemia prediction
models. In this research; we present and pilot test a study in
which a smartphone was used to obtain these data. Data from
one female was obtained over a period of two days. CGM and
continuous physical activity accelerometry data were collected with minimum and no dropouts; respectively. The collection
of diet; insulin and blood glucose data; also; proceeded without
problems. These results indicate that it is possible to collect
glucose; diet; insulin and physical activity data of high
quality. These data will facilitate further development of models
for the detection and prediction of low blood glucose.
Persons with Type 1 diabetes need continuous exogenous insulin
supply throughout their life. Determining the optimal
insulin treatment in relation to diet and physical activity is one
of the main goals of diabetes management; but is difficult;
especially for vulnerable populations; such as adolescents.
Erroneous treatment may result in both repeated and severe
low blood glucose events. Continuous glucose monitoring
(CGM) may help in avoiding these events; but is inaccurate
compared to traditional glucose monitoring. Models have
been developed to significantly improve CGM’s detection of
insulin-induced events by using information from the CGM
signal itself. Additional temporal data on insulin doses; diet
and physical activity may improve hypoglycaemia prediction
models. In this research; we present and pilot test a study in
which a smartphone was used to obtain these data. Data from
one female was obtained over a period of two days. CGM and
continuous physical activity accelerometry data were collected
with minimum and no dropouts; respectively. The collection
of diet; insulin and blood glucose data; also; proceeded without
problems. These results indicate that it is possible to collect
glucose; diet; insulin and physical activity data of high
quality. These data will facilitate further development of models
for the detection and prediction of low blood glucose.
Hypoglycaemia; detection; prediction; continuous
glucose monitoring; study design.
Inga referenser tillgängliga
