To our knowledge, this study is the first large study to report the effectiveness of text-messaging versus telephone reminders in a primary care setting in Europe. Our findings confirm the non-inferiority of text messaging versus telephone reminders to reduce the rate of missed appointments in an academic primary care clinic and the cost-effectiveness of text-message reminders compared to manual phone reminders. Both types of reminders were well accepted by all patients, but text-messages were preferred, particularly among patients consulting for a tobacco, alcohol or other substance abuse.
Rates of attendance were slightly higher following telephone reminders than text-message reminders among general primary care patients, indicating that direct personal contact with the patient may be more effective than a machine-generated message. This is consistent with another study which showed that telephone reminders by a staff member were also slightly more effective than automated phone calls to reduce non-attendance in an academic outpatient practice . However, these findings were not confirmed in another study assessing the effectiveness of automated versus staff phone reminders for colonoscopies . Patients’ perceptions may differ according to the type of health care setting and care provision. The difference in the rate of attendance between both groups was smaller than 2%, our pre-defined non-inferiority margin. The benefits of a higher attendance with telephone reminders can thus be considered minimal.
The fact that a higher percentage of patients in the substance abuse unit cancelled appointments following a text-message rather than a telephone reminder suggests that this particular group of patients may be more receptive to text-messaging than telephone reminders. It is well known that non-attendance rate among patients with mental disorders is high (up to 40%) [18, 19], predominantly among patients with alcohol and drug abuse (18-36%) . This is probably linked to the fact that patients suffering from substance abuse or mental disorders in general may have increased socio-economic difficulties and impairment complicating their regular access to care [18, 21]. Also, motivation for substance abuse care can fluctuate between the time of scheduling and the time of appointment. We found no comparative data on the effectiveness of telephone or text-messaging reminders in this particular population of primary care attenders. Contrary to our expectations, no slots were reallocated after appointments had been cancelled in response to the reminder. This might be due to the short delay (24 h) between the reminder and the consultation itself. In a previous study, 48 hours delay allowed reallocation of 28% of the spaces made vacant after cancellation . Increasing the reminder delay and improving the ability to identify vacant spaces in our electronic appointment system may also help reallocate these free appointments more efficiently in the future. However, although both reminders were equally effective in reducing the rate of missed appointments, only the text-message system was cost-effective, because of the absence of additional administrative work. Further research should explore the non-inferiority of text-messaging reminder compared to automated phone call reminders.
As in previous studies on telephone and text-message reminders, both types of reminders were well accepted by a large majority of patients [13, 17, 22]. Only very few patients declared that they had been disturbed by the intervention, although those who opted out might have done so because they judged the intervention would be disturbing in the first place. However, their number was minimal as shown in the flow chart. We are unable to explain why substance abuse patients showed a preference for a text-message over a phone reminder. However, such preference should be taken into consideration since text-messaging was associated with a higher rate of cancelled appointments.
One negative effect of reminders, whether phone or text-message, is that they shift the responsibility of attendance away from the patient to the organization . Alternative effective interventions include incentives, such as free treatment for patients with regular attendance , or compulsory involvement in an educational program in order to continue receiving care after several missed appointments . Given the high number of missed appointments and the difficulty to define a specific profile of non-attenders, large scale implementation of such interventions does not appear to be realistic, at least in our context.
Our study had some limitations. It was a single-center study conducted in an academic primary care clinic providing care to patients with low socio-economical status with a rapid turnover of physicians. Results may not be generalizable to other settings although they are consistent with other studies. Since the main outcome variable was the appointment and not the patient, we did not collect information about the number of reminders sent to each patient during the main study. The automated software was not set to provide information on whether patients had received the text message or not. Therefore, we did not collect information about the estimated rate of patients reached by the text-message. The fact that text messages were sent in French only whereas phone reminders occurred in four languages may have enhanced the effectiveness of telephone reminders over text messages. The satisfaction survey also had several limitations: patients’ acceptance of reminders was assessed in a limited sample of them, and only among those who could be reached by phone during the survey; we only asked patients which type of reminder they had received, but did not ask them about the number of reminders received; we did not record the number of patients who reported not having received a text message or a phone call reminder. Costs in our cost-effectiveness analysis represent charges. A more detailed cost-effectiveness analysis including purchase and maintenance costs was not possible, since we could not assess the costs corresponding to the purchase and to the maintenance of the central telephonic switchboard of our division and of our hospital, and of an automated text messaging system. Nevertheless, maintenance costs of both systems are probably equivalent. Finally, we did not collect data about misreading/misinterpretation of written or oral information, issues of privacy and disclosure .