Using a Mobile App to Manage Type 1 Diabetes ... - ACM Digital Library

3 downloads 48 Views 143KB Size Report
We present TreC Diabetes a system consisting of a mobile diary and a web dashboard. The Android app allows patients to record. Blood Glucose Values ...
Using a Mobile App to Manage Type 1 Diabetes: The Case of TreC Diabetes Francesco Miele

Claudio Eccher

Enrico Maria Piras

Bruno Kessler Foundation Via Sommarive 18, Trento, Italy +39 0461314 129

Bruno Kessler Foundation Via Sommarive 18, Trento, Italy +39 0461314 161

Bruno Kessler Foundation Via Sommarive 18, Trento, Italy +39 0461314 126

[email protected]

[email protected]

[email protected]

carbohydrate content, and physical activity. Moreover, the application provides additional supporting functionalities, as the computation of the insulin bolus, and a statistical and graphical section that provides an overlook of diabetes management through indexes and charts. All data are synchronized with a central database and available for the consultation by means of a web dashboard. The system has a built-in rule-based alarm module that notifies healthcare professionals of risky conditions (e.g., a pattern of consecutive BG values higher than a threshold) sending an e-mail. Through the web dashboard, the diabetologist can visualize the alarm and the parameters that caused it, to promptly intervene and contact the patient if the situation requires doing so.

ABSTRACT Type 1 diabetes is chronic condition due to the autoimmune destruction of the insulin-producing beta cells in the pancreas. We present TreC Diabetes, a system consisting of a mobile diary and a web dashboard allowing patients to record disease-related information, which can be visualized by diabetologists through a dedicated dashboard. We also present the results of an observational study to assess the user acceptance of the system.

Categories and Subject Descriptors H.0 Information Systems

Keywords

3.

Personal Health Record; telemonitoring; diabetes type 1; information management; mixed methods; log analysis.

1.

INTRODUCTION

Type 1 diabetes is chronic condition due to the autoimmune destruction of the insulin-producing beta cells in the pancreas. To keep their blood sugar values within an acceptable range, patients have to take daily insulin therapy and have a healthy diet and life style. Self-management in Type 1 diabetes is vital to avoid complications (e.g., heart disease, stroke). In last decades, diabetes management has been more and more delegated to patients and their caregivers, who must develop self-management skills to cope with the disease and have a satisfying quality of life.

15 young patients were initially enrolled in the study. The age ranged from 4 to 12 years (median value of 8). One patient dropped out for reasons not related to the experimentation. Patients were provided with a smartphone and a glucometer able to communicate with it to upload data. Alternatively, patients were free to use their own glucometers and manually input the BGVs. To assess the application use, we used a mixed method approach combining the analysis of the log files and semistructured interviews with the users (generally the parents). The analysis was quantitatively driven. Firstly, we identified and interpreted significant patterns of use through the log analysis by means of descriptive statistics. Secondly, we used the interviews to understand the perspective of the users for the quantitative data.

A multitude of diabetes-related m-health (mobile health) apps have been developed [1]. As Arnhold and colleagues [2] recently argued analysing more than 600 apps, m-health apps for diabetes have generally one (54,1%) or two functions (28,2%), do not have data forwarding/communication functions (68,9%) and do not offer an interface to a measurement device (95,4%). Moreover, they are rarely evaluated [3].

2.

THE STUDY

An observational study (Ministero della Salute DGDFSC 0032830-P-22/04/2014, I.5.i.m 2/2014/953), aimed at assessing the user acceptance of the system, started the 27th September 2014, and ended the 27th December 2014 (91 days).

4.

RESULTS

From the logs of the database, we extracted a set of values regarding the use of the feature to store BGVs and the additional features provided by the application.

THE SYSTEM

We present TreC Diabetes a system consisting of a mobile diary and a web dashboard. The Android app allows patients to record Blood Glucose Values (BGVs), meal composition and

As regards the management of blood glucose measurements, during the trial, the 14 patients entered 8584 BGVs, an average of about 613 values per patient (min 432, max 1011), 6.7 measurements per patient per day (min 4.7, max 11). Only six patients entered at least the 70% of the values within two days from the measurement. Three patients entered the information indicating the moment of the day in a sufficient number of measurements (>90%, see discussion). If we integrate these quantitative results with the data of the qualitative interviews, we can notice that TreC Diabetes was used mainly in two ways. Firstly, a minority of users, during experimentation, began to use the digital application with the (approximately) same frequency of paper-based logbook. These users entered information

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for components of this work owned by others than ACM must be honored. Abstracting with credit is permitted. To copy otherwise, or republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee. Request permissions from [email protected]. DH '15, May 18–20, 2015, Florence, Italy. Copyright© 2015 ACM. ISBN 978-1-4503-3492-1/15/05…$15.00. DOI: http://dx.doi.org/10.1145/2750511.2750541

113

information only parents have. This interpretation is supported by the scant attention paid by many user to integrate their BGVs with other information or the delay in synchronizing the data. These data are mostly considered “personal” [5] and useful to take decisions autonomously rather than information to be shared with doctors to have their support.

continuously and precisely. Secondly, there is the majority of users that seemed to use TreC Diabetes as an integration of the paper-based logbook, filling in the electronic diary sporadically. As regards the modality of glycaemia data input, eight patients entered all the data using the Bluetooth glucometer provided for the study, while three of them manually entered a small part of data. Three patients continued to use their own device and entered all the values manually. The choice was motivated by a greater familiarity of the parents with their own glucometer, a supposed greater fidelity of their device, or the presence of small features (like the screen light) useful during the night.

One size does not fit all. Even at a superficial inspection, log file analysis shows that there is not a single killer functionality. No one involved in the clinical study used all the functions provided in the application and nobody used (or even tried) only a single functionality. Another look at the stats reveals, though, a systematic use of the functions a patient uses. This is consistent with the highly personalized strategy adopted by each family in diabetes management and its reflection in their different information need. It also brings to reflect on the opportunity to pursue a design strategy aimed not at identifying a few features universally used but rather at offering a wider set of options selected by users according to their needs.

All the risks are calculated each Sunday at midnight on the BGVs of the preceding week, if there are at least five days with three measures per day. Alarms fire if the risk is >5 for LBGI, >15 for HGBI and >40 for ADRR. Hyper GL is the recurrence of BGVs>300 in any moment of the day for two consecutive days. The interviews show that, in spite of the generated alarms, users do not contact doctors and, on the other hand, doctors do not contact users. The explanation is quite simple: patients and their caregivers generally know how to manage emergencies without involving the diabetologists. To contrast hypoglycaemia they know they can eat food with sugars and carbohydrates, to lower BGVs patients can do physical activity or inject an additional dose of insulin. For these reasons, we argue that patients or parents manage emergencies effectively in real time, without the help of doctors and, consequently, of the alarm feature of our tool.

5.

To sum up, the application under analysis has been tested in a well-defined patient-doctor relationship, which is satisfying for both parties. In particular, parents do not seem to have considered the application as a tool to redefine the ways they interact with the hospital, rather as a tool for personal health information management and, in few cases, to support decision making process. In this context, TreC Diabetes is positively considered by patients to fill some information management gaps not addressed by current tools offered by the department. Our evaluation suggests that its use could be beneficial for healthcare professionals as well, should the tool be better integrated in the current patient-doctor relationship. More specifically, we argue that TreC Diabetes should be considered as a counselling means rather than a curing one.

DISCUSSION

The management of a child with type 1 diabetes requires the family a significant workload. It entails activities ranging from food preparation to carbohydrate counting, from blood glucose level measuring to insulin dosing, not to mention the constant attention to detect early signs of hyper/hypoglycaemia. Each of these activities must be skilfully mastered and have to become a part of the everyday routine since people with diabetes (or parents) are responsible for almost the 99% of the decisions regarding care [5]. In this scenario, it is not surprising that families came up with personalized strategies somehow reflected in the way people use the mobile application provided.

6.

ACKNOWLEDGMENTS

This work is a part of a larger research project (TreC – Cartella Clinica del Cittadino) [6] funded by the Department of Health and Social Politics of the Autonomous Province of Trento (Italy).

7.

Moreover, the use of the application reveals more. We shall try to synthetize the main lessons learned.

REFERENCES

[1] Alberti, KG., Zimmet PZ., 1998. Definition, diagnosis and

Inadequacy of current paper based logbooks. Despite healthcare professionals educate the patients to manage diabetes taking into account many different aspects, paper-based logbooks only allow to keep track of BGVs and insulin intake. This reflects the limited time doctors can spend analysing patient-recorded data; hence, the logbook is limited to the two most relevant parameters to make an assessment during the visits. Parents’ choice of recording other information in the digital application, however, suggests that the daily management of the condition requires a much more fine-grained data collection strategy.

[2]

[3] [4]

The complexities in information sharing among patients and doctors. The above-described need of more information by parents signals an even greater difference between them and doctors. On the one hand, patient’s education provides parents with tools to manage the vast majority of the challenges posed by diabetes. On the other hand, a careful analysis of patientgenerated data would require healthcare systems to dedicate resources they do not seem to have. Moreover, many of these data would not be interpretable by doctors alone without the contextual

[5] [6]

114

classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med. 15(7), 539-553. Arnhold, M., Quade, M., Kirch, W., 2014. Mobile Applications for Diabetics: A Systematic Review and Expert-Based Usability Evaluation Considering the Special Requirements of Diabetes Patients Age 50 Years or Older. Journal of Medical Internet Research, 16(4):e104. Demidowich, A. P., Lu, K., Tamler, R., & Bloomgarden, Z., 2012. An evaluation of diabetes self-management applications for Android smartphones. Journal of telemedicine and telecare, 18(4), 235-238. Piras E. M., Zanutto A., 2014. “One day it will be you who tells us doctors what to do!”. Exploring the “Personal” of PHR in paediatric diabetes management, Information Technology & People, 27 (4), 421-439. Anderson R.M., Funnell M.M, 2010. Patient empowerment: myths and misconception. Patient education and counselling, 79 (3), 277282. Eccher C., Piras E.M., Stenico M., 2011. TreC - a REST-based regional PHR.. Studies in health technology and informatics, 169, 108-112.

Suggest Documents