Table 1.
Proposed (hypothetical) factors influencing nonusage attrition and dropout attrition in eHealth trials
| Factor | Impact on Nonusage Attrition Rate | Impact on Dropout Attrition Rate |
| Quantity and appropriateness of information given before the trial, expectation management | Inappropriate information leads to unrealistic expectations which in turn leads to disenchantment discontinuance | Indirectly through nonusage (usage discontinuance leads to drop out) |
| Ease of enrolment (eg, with a simple mouseclick as opposed to personal contact, physical examination etc), recruiting the “right” users, degree of pre-enrolment screening | If the “wrong” participants are enrolled, ie, those who are less likely to use it, and willing to invest time, and for whom the intervention does not “fit” | The easier it is to enroll, the more users will later drop out if they realize that filling in questionnaires, etc creates more work than they thought. Also indirect via nonusage. |
| Ease of drop out / stop using it | The easier it is to stop using the application, the higher the nonusage attrition rate will be (and indirectly through dropouts) | The easier it is to leave the trial, the higher the attrition rate will be (and indirectly through nonusage) |
| Usability and interface issues | Usability issues obviously affect usage | Indirectly through nonusage (usage discontinuance leads to drop out) |
| “Push” factors (reminders, research assistants chasing participants) | Participants may feel obliged to continue usage if reminded (cave external validity) | Participants may feel obliged to stay in trial |
| Personal contact (on enrolment, and continuous contact) via face-to-face or phone, as opposed to virtual contact | Mainly indirectly via dropout | The more “virtual” the contact with the research team is, the more likely participants will drop out |
| Positive feedback, buy-in and encouragement from change agents and (for consumer health informatics applications) from health professionals / care providers | Participants may discontinue usage without buy-in from change agents. In particular, patients may stop using eHealth applications if discouraged (or no actively encouraged) by health professionals | Indirectly through nonusage (usage discontinuance leads to drop out) |
| Tangible and intangible observable advantages in completing the trial or continuing to use it (external pressures such as financial disadvantages, clinical/medical/quality of life/pain) | Yes | Yes |
| Intervention has been fully paid for (out-of-pocket expense) | If individuals have paid for an innovation upfront they are less likely to abandon it (as opposed to interventions paid on a fee-per-usage basis) | Indirectly through nonusage (usage discontinuance leads to drop out) |
| Workload and time required | Yes | eg, to fill in the follow-up questionnaires may create such a burden that participants drop out |
| Competing interventions | For example similar interventions on the web or offline can lead to replacement discontinuance | Indirectly through nonusage (usage discontinuance leads to drop out) |
| External events (9/11 etc) | These may lead to distractions and discontinuance, especially if the intervention is not essential | Indirectly through nonusage (usage discontinuance leads to drop out) |
| Networking effects/peer pressure, peer-to-peer communication, and community building (open interactions between participants) | Communities may increase or slow the speed with which an innovation is abandoned. | Communities may increase or slow dropout attrition. |
| Experience of the user (or being able to obtain help) | As most eHealth applications require an initial learning curve and organizational change, users have to overcome initial hurdles to make an application work. Experience/external help can contribute to overcoming these initial hurdles and help to see the “light at the end of the tunnel” | Indirectly through nonusage (usage discontinuance leads to dropout) |