*2.2.3 Patient must enter the data himself or herself*

We minimized data entry tasks and burden (aka response burden) by asking patients to identify what was important to them. Although the questionnaire is relatively short at only 8-questions, not all questions are relevant to all patients. In a clinical trial setting, there is no option but to use the entire scale because missing data can easily invalidate a study. However, for an app that tracks the individual effectiveness of clinical treatment, focusing on what's most important to the patient can go a long way to decreasing response burden and increasing compliance to data capture; some data is always better than no data. The purpose of the app is not to capture data, but to help the patient manage their disease and to track the trajectory of their disease. If the patient rarely or never experiences a particular symptom, there is no need to track that information. If patients experience their worst symptoms in a particular part of their body, then tracking that symptom may be an excellent sentinel event for tracking overall disease state. Using this approach, we were able to decrease the number of questions that the patient had to answer.

Another way to reduce response burden is to ask patients to fill out the questionnaire randomly spread out over a longer period. By 'sampling' the patient's disease experience over a longer period, it is possible to get a good sense of the patient's disease trajectory and then sample more or less frequently to optimize the amount of information compared to the response burden for the patient.

The magnitude and/or the magnitude of change in patient responses can also be used to make data entry more meaningful to the patient. If the patient's symptoms are low and/or stable, the app can take longer to notify the patient to enter data. If the patient's symptoms are worse or changing, then capturing data more frequently makes more sense and is likely to be more acceptable to the patient.

#### *2.2.4 Patient cannot use information in a meaningful way*

Patients generally have little control over their clinical condition. They cannot order diagnostic testing, prescribe medications or refer themselves to specialists. They are dependent to a great extent on their HCPs for those clinical actions. However, it does not mean that they have no control over their condition.

In myasthenia gravis, the patient has 1 major concern on a day-to-day basis: when my symptoms worsen on any one day, am I just having a 'bad day' or is it a worsening of my underlying condition? The uncertainty of symptom worsening can be a great source of anxiety [88]. The app could help patients see the trends of their disease and help them better and more quickly distinguish between a 'bad day' and a worsening of their disease, thereby giving them greater control over their disease experience.

The other way that patients can be empowered to use the app is to provide an easy-to-use report from the app for their HCP. This report, if useful and used by the HCP, would benefit the patient directly. Because of the uncertainty and daily variation in symptoms experienced by patients with MG, recalling their symptoms and experience of their disease during a visit with their HCP can be challenging. This was confirmed by our panel of clinical experts who indicated that history taking in MG is challenging and time-consuming. HCPs are likely to benefit from and value a simple, easy to read, visual report of the patient's symptoms overlaid with information about compliance to medication and exacerbating factors experienced at the time of symptom worsening. Since, this is the most beneficial information for HCP, we need only capture medication compliance and exacerbating factor information at the time of symptom worsening, thereby minimizing response burden further.

Patients are influenced by non-verbal cues from their HCP. If the HCP does not use the information in the report, they are likely to stop using the app. However, if patients perceive that HCPs value the reports, they will continue using the app. HCPs have a strong influence on patient app usage and can reinforce app usage by requesting the report and by viewing the information and acting on in it, where appropriate.

#### *2.2.5 Lack of incentives like cost saving or social approval*

There are few external incentives for using an app in MG. MG is not a lifestyle disease, so getting empathy from other people is rather difficult. MG is also a stigmatizing disease (see below), so social sharing of progress may not be appropriate or desired by the patient. However, there is a useful behavior change theory called the Self-determination Theory (SDT) [90] which was determined to be appropriate for patients with MG and could provide a significant incentive to use the app. How SDT was incorporated into the app is described below in Section 2.3.

#### *2.2.6 Daily use of the app is not required and therefore patient does not get into the habit of using it*

Although patients with MG can experience symptoms at any time, the lived reality is that exacerbations and improvements play out over weeks and months.

#### *Designing Disease-Specific mHealth Apps for Clinical Value DOI: http://dx.doi.org/10.5772/intechopen.99945*

Collecting data everyday can become tedious and lead to response fatigue. How do we design an app that maintains a good balance between infrequent data collection and loss of skill over time?

Through Expert Usability reviews, we simplified the data collection so that patients only reported data about the symptoms that are most troublesome to them. We also simplified the task loops so that they could get feedback from their data entry in smaller cycles of data entry, instead of getting one big report after answering a long series of questions. Encouraging and educational messages are spread throughout the data entry tasks so that patients do not get fatigued with answering many questions. Messages that explain the rationale for capturing specific information can help the patient connect the data they are entering to the outcomes they are hoping to achieve.

Notifications can help remind the patient to track information. To improve response rates, the notification could be developed to enable data entry directly into the notification rather than having to open the app and enter the data there. This is particularly useful where a single response is required, such as 'Did you take your medications as prescribed yesterday?' It is better to ask a question like that about a completed time unit rather than asking it for an incomplete time unit, since there is a finite chance that the patient may not have completed a task that the app is asking about.
