Hypo-RESOLVE researchers have developed a new app to study the daily impact of low blood glucose
By Dr Uffe Søholm
In the Hypo-RESOLVE study, we are exploring how hypoglycaemia (low blood glucose) affects the daily life of people with diabetes. To do this, we are using a new Smartphone App called ‘Hypo-METRICS’.
Apps provide great opportunities for new types of research. They enable people to share what is happening in their daily life, by responding to questions on their phone in real-time and real-life settings. But what steps are needed to develop high quality apps for use in research? Here are the steps we took and what we considered when developing the ‘Hypo-METRICS app’.
1. Collaboration: Find the right team of experts. We included researchers and clinicians from various professional backgrounds. Importantly, we also included people with lived experience of diabetes.
2. Literature review: Identify areas of life relevant to the daily impact of the condition, by looking at past research studies. We used existing concepts (such as ‘quality of life’) to guide our search and to inform the process.
3. Selecting relevant constructs: Combine the literature review findings with feedback from the expert team. We discussed which life areas might be suitable to measure daily (or multiple times per day).
4. Conceptual framework: Use the findings (above) to build a conceptual framework. This is simply a plan for what you want to assess, and how each element relates to another. For our app, we agreed on the concept of ‘daily functioning’. Our framework was built around aspects of ‘quality of life’ that can be assessed daily (or multiple times per day).
5. Develop items and scales: We considered many issues when developing the questions and responses to be used in the App:
- Should items be phrased as questions or statements?
- Should items be specific to the condition or general, e.g., ‘How did [condition] impact your mood?’ or ‘How is your mood right now?’
- Frequency of each item: considering sensitivity vs. participant burden
- Order of items: consider mixing it up to avoid ‘automatic responding’
- Number of items per ‘concept’: Is one enough or are more needed?
- Recall period: To what period of time do you want participants to refer, e.g. to right now, today, since last ‘check-in’?
- Length of the response scales: 7 to 11 points is usually appropriate
6. Test it: Ask an independent group of people with the condition to test and give feedback on your questions and scales. For example, are they relevant and understandable? Also, ask if any important issues have been missed. We did this and then implemented the changes.
7. Translation: If other language versions are needed, build a structured translation plan. We used this resource as a guide.
8. Implement into a smartphone platform: Work with technicians who have data security policies that are right for your study. We also considered the design and function of the app. For example:
- Should time of assessment be random or at specific times?
- Timing of reminders
- Is it ok for people to skip some questions?
- Text size, colours etc.
This was a brief overview of some of the issues that we considered. For more details, see our Hypo-METRICS app design paper here.
Søholm U, Broadley M, Zaremba N, Divilly P, Nefs G, Mahmoudi Z, … Hendrieckx C, Speight J, Choudhary P, Pouwer F for the Hypo-RESOLVE Consortium. Investigating the day-to-day impact of hypoglycaemia in adults with type 1 or type 2 diabetes: design and validation protocol of the Hypo-METRICS application BMJ Open, 2022; 12(2), e051651.
To read more of our research about hypoglycaemia check out our previous blogs on this topic.