Targeting the Use of Reminders and Notifications for Uptake by Populations (TURNUP): a systematic review and evidence synthesis

Findings from effectiveness review (systematic reviews and randomised controlled trials)

Free et al. 45 reported that the pooled effect on appointment attendance using text message (SMS) reminders compared with no reminder was increased, with a relative risk (RR) of 1.06 [95% confidence interval (CI) 1.05 to 1.07]. Car et al. 43 provides moderate-quality evidence showing that mobile text message reminders improved the rate of attendance at health-care appointments compared with no reminders (RR 1.10, 95% CI 1.03 to 1.17). These findings suggest that a small percentage of individuals (estimated variously at between 3% and 17%) do genuinely forget and, therefore, benefit from receipt of a reminder.

The overwhelming pattern from the systematic reviews and included trials is that reminder systems reduce non-attendance rates. However, there are differences in what investigators perceive as a meaningful difference (figures such as a 10% or 15% reduction are cited when calculating a sample size) and so some studies fail to reach statistical significance.

Stubbs et al. 52 conclude that telephone, mail and text/SMS interventions all improved attendance modestly, but at varying costs. Similarly, Jacobson Vann and Szilagyi49 found that reminding people over the telephone, sending a letter or postcard, or speaking to them in person increased vaccinations. Hasvold and Wootton47 undertook a systematic review of 33 independent studies pertaining to telephone, SMS or automated telephone reminders found that all but one reported a benefit from sending reminders and suggested a weighted mean relative change of 34% from the baseline non-attendance rate.

Bos et al. 55 tried three different types of reminder system and failed to confirm the hypothesis that a reminder would reduce the failed attendance rate. The explanation for this finding was that the clinic had a very high attendance rate with both a high perceived importance of the clinic and high levels of satisfaction with the orthodontist and the services being provided. Perron et al. 78 report that being part of a stepped approach intervention (telephone, text, mail) significantly decreased the risk of missing an appointment for all categories of patients found at risk [odds ratio (OR) 0.63, 95% CI 0.43 to 0.89]. A useful reminder function is served by whatever means is at a service’s disposal.

In their systematic review, Free et al. 45 report no difference in attendance using SMS reminders compared with other reminders; however, these findings did not reach statistical significance (RR 0.98, 95% CI 0.94 to 1.02). Car et al. 43 identified two studies of moderate quality that showed that mobile phone text message reminders and telephone call reminders had a similar impact on health-care attendance (RR 0.99, 95% CI 0.95 to 1.03). This suggests that patient compatibility/preference, not the inherent advantage of any technology per se, should determine selection of the candidate reminder system. In contrast, Bos et al. 55 found no differences between SMS, telephone and mail reminder when compared with no reminder. The authors concluded that the form of the reminder is irrelevant.

Findings from other studies

No additional studies.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

There is weak consistent evidence from two reviews that show that reminders providing additional information (e.g. orientation information, health information, etc.) over and above details of timing and location (reminder plus) are more effective than simple reminders, which provide information about time date and location of an appointment.

In their Cochrane review, Reda and Makhoul51 investigated the effects of reminders to encourage attendance at clinics for those with suspected serious mental illness. Of the four trials included in review, only one study87 (n = 120) compared a standard letter prompt with a letter ‘orientation statement’. The results of this study favoured the orientation statement approach rather than the simple letter prompting attendance, but the results did not reach conventional levels of statistical significance (RR of missed appointment 1.6, 95% CI 0.9 to 2.9). There is a chance of type II error in this particular study and simply increasing the number of participants in this study may have resulted in a more favourable result. This review provides very weak evidence that a simple orientation-type letter may be more effective than a telephone prompt at promoting attendance at mental health clinics.

A systematic review48 to assess the effectiveness of reminder systems as a means of increasing attendance and reducing ‘did not attend’ rates at all new outpatient appointments identified limited evidence, based on three studies, that ‘reminder plus’ were more effective than standard reminders. Postal reminders that offered a reward for attendance, threatened sanctions for non-attendance or provided orientation information about the clinic appeared to be more effective than standard reminders. In this review ‘warning reminders’ contained warnings that patients would fall to the bottom of the waiting list following failure to attend (FTA), ‘reward reminders’ contained the offer of rewards should patients attend and ‘orientation reminders’ contained information describing the clinic. There was an indication in this review, based on the findings of one study, that warning reminders were more effective than reward reminders. 88

Can et al. 56 describe an intervention that requested that patients return a stamped addressed postcard, so that the appointment could be confirmed. It could be hypothesised that the action of returning the postcard might be construed as a symbolic ‘contract’ between patient and provider, reducing the subsequent likelihood of non-attendance. Whatever the reasons, return of the postcard resulted in an attendance rate of 83% compared with only 48% if the card was not returned.

This evidence indicates that ‘reminder plus’ may be more effective than simple reminders; however, as a result of the small number of studies which have systematically investigated this hypothesis, it is not possible to be conclusive about this. There is no evidence to show in what way the receipt of additional information as part of the reminder will shape patients’ intentions to attend or their perceptions about the appointment. Further high-quality research is required to investigate these areas.

Findings from other studies

Several studies have utilised orientation strategies to apparent good effect. Kitcheman et al. 69 incorporated letter reminders on headed paper in their RCT explaining the time of the appointment, name of the doctor, description of the clinic and its routine, a map and a request to bring a list of medication and a friend or family member. This significantly reduced the numbers of people failing to attend a mental health clinic appointment. Booth and Bennett89 conducted a comparative study utilising telephone reminders incorporating information about appointments, directions to the clinics and other relevant information that reduced non-attendance at first appointments at a specialist alcohol treatment clinic. However, neither of these studies incorporated control groups receiving a simple reminder only and, therefore, it is not possible to determine from these studies the additive effect of orientation information.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

Strong consistent evidence from multiple high-quality studies indicates that reminders may not be accurately received by a substantial proportion of patients for whom they were intended. This will contribute to suboptimal effectiveness of reminders. The extent to which patients do not receive their reminder varies depending on the technology, with telephone reminders demonstrating the lowest success rates for making contact with the patient.

Findings from other studies

No additional studies.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

None of the included studies explicitly explored the extent to which factors such as language or the way in which a reminder is framed influences its comprehensibility and, therefore, the effectiveness of reminders. None of the included studies investigated the extent to which factors such as cognitive impairments, literacy and health literacy might influence the comprehensibility and subsequent effectiveness of the reminder. However, it is self-evident that comprehensibility of the reminder is an important determinant of the effectiveness of the reminder.

Very few trials of reminder systems specifically examine the content of the reminder. A notable exception is seen in recent interest in use of an ‘orientation statement’ specifically in the context of patients with severe mental illness [see Evidence statement (A.2–A.6): there is weak consistent evidence that ‘reminder plus’ are more effective than simple reminders at helping a patient to attend their appointment (evidence category IIIa)]. In fact, several trials exhibit a preoccupation for controlling against variation in content rather than exploring its variety. So, for example, in the trial by Chen et al. ,57 staff members making telephone calls were required by a trial protocol to follow as closely as possible the same format of words as used in the SMS group. Here we encounter a direct tension between rigorous research and pragmatic exploration of multiple alternative solutions.

Findings from other studies

No other studies investigate whether or not reminders address patient obstacles to attendance.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

No effectiveness studies investigate whether or not reminders address patient obstacles to attendance.

Findings from other studies

In the section Quality of included trials, we demonstrate that there is weak consistent evidence that reminders which provide additional information over and above details of timing and location (‘reminder plus’) are more effective than simple reminders which provide information about the time, date and location of an appointment. However, it is not known whether providing patients with additional information acts to reduce perceived obstacles to attending or by maximising enablers. This is an area that requires more research.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

A systematic review by Clough and Casey25 investigating technological adjuncts to increase client adherence to psychological therapies found, based on two RCTs, that reminders with direct client and therapist contact were more effective at increasing appointment attendance than indirect reminders, such as letters, recorded messages, or reminder calls made by the clinic secretary. Similarly, a systematic review by Hasvold and Wootton47 on the effectiveness of personal or automated telephone calls or SMS message reminders for patients attending a variety of hospital outpatient appointments found that the effect of reminders on did-not-attend rates was higher for manual reminder calls than for automated calls. Based on data from 29 studies, reporting a total of 33 estimates of reminder effectiveness, 18 of the interventions were based on manual reminders (i.e. telephone calls made by health staff) and 15 were based on automated reminders (i.e. automated telephone messages or SMS messages). Taking into account the quality of the studies, the pooled estimates of the weighted means (using study quality as the weighting factor) showed that manual reminders can achieve a reduction in the did-not-attend rate of 39% of the baseline value, while automated reminders can achieve a reduction of 29% of the baseline value. This indicated that personal reminders are more effective than those sent automatically by a computer.

Two additional RCTs66,73 investigated the differences between personal and automated reminders, but neither found any difference between the direct contact reminder and the automated reminder. In a Malaysian RCT73 comparing the effectiveness of text messaging reminder (n = 308), telephone reminder (n = 314) and control (no reminder) (n = 309) in increasing appointment attendance in patients with chronic diseases requiring long-term follow-up, the absolute non-attendance rate for telephone reminders (13.7%) was 2% lower than with the text messaging group (15.6%). This study is limited in two regards. First, the person delivering the telephone reminder message was to deliver the same standardised message that was delivered by the SMS reminder and if patients had any further queries about their appointment they were directed to contact the appointment desk, thus minimising the personalised nature of the personal contact. Second, the study was powered to detect an absolute difference of 10% between the two reminder groups. Because of the risk of a type II error, it is not possible to conclude that there is no difference in effectiveness between the direct contact reminder (telephone) and the automated reminder (SMS). An appropriately powered study with greater personalisation of the telephone reminder would be required in order to reach a conclusion.

Griffin et al. 66 undertook a three-arm RCT (n = 3610) to test whether or not an IVR system telephone call was equally as effective as a nurse-delivered telephone call at educating and preparing patients for flexible sigmoidoscopy and colonoscopy examinations. The three study conditions were nurse telephone call 7 days before the procedure, an IVR system call 7 days before the procedure and an IVR system call 3 days before the procedure. All calls included an appointment reminder, information about preparation for the examination and encouragement to prepare for and attend the examination. There was no statistically significant difference in appointment attendance between automated, interactive telephone reminders and personal telephone reminders in the context of fluoroscopy and colonoscopy clinics. Significantly more patients in IVR conditions reported neutral perceptions about the telephone calls and more patients receiving nurse calls reported very positive perceptions about the telephone calls. Of particular note is that this automated IVR system was dissimilar to many other automated systems in that it was interactive and, thus, allowed patients to request that information be repeated. The system provided patients with cues to action for appointment attendance and procedure preparation, including targeted educational information about susceptibility and severity of colorectal cancer and motivational messages that addressed risks, benefits, barriers and self-efficacy associated with preparation and procedures. The interactive nature allowing flexible access to the information about the appointment may explain its relative success compared with the personal contact provided by the nurse.

Findings from other studies

No additional studies.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

Most examples from the effectiveness review indicate that the amount of text that can be delivered in a SMS text message is fairly limited. When described, many trials simply specified participant’s name and appointment details. However, the majority of the studies did not provide any details of the message content, making it difficult to analyse their content and intent. Cited examples of actual content include:

[Name] Your apt. is at [Time] on [Day]. See you then! To end reply STOP. 75

Dear Mr/Ms [. . .] Have you taken the prescribed lipid lowering medication as scheduled? I hope to see you at your next visit on [0000 year] [00 month] [00 day], at [000 hospital], from [doctor 000]. 59

You have an appointment at [name of practice] [today/tomorrow] at [time]. Please call [number] if you can’t make it. 64

Findings from other studies

It is self-evident that the format (medium) in which a reminder message is delivered will dictate, to some extent, how long and informative the reminder message can be. By their nature, SMS messages are length restricted and, in line with guidelines around confidentiality, usually contain minimal information. 96 Accordingly, SMS reminders can serve as memory joggers, but little more. Other reminder technologies, such as e-mail, letters and personal telephone calls, lend themselves to the provision of additional information that may facilitate appointment attendance (see Quality of included trials).

Findings from effectiveness review (systematic reviews and randomised controlled trials)

None of the included studies directly explore the relationship between patient reminder preferences and the effectiveness of the reminder on patient appointment behaviour. Parikh et al. 77 conducted a survey in conjunction with their study comparing personalised or automated telephone calls with no intervention. The personalised intervention was regarded as helpful by 85% of survey respondents, compared with 48% in the automated group. Perron et al. 78 report, from a satisfaction survey with 241 patients conducted alongside their RCT, that 93% were not bothered by the reminders and 78% considered them to be useful. Bos et al. 55 reported that most interviewed participants felt positive or very positive about receiving a reminder; however, no less than 20% of interviewed participants felt negative or very negative about the reminders and considered them to be a waste of time and money.

If patients are already disposed to using mobile technology and text messaging, then they are very highly (98%) in favour of using SMS as a mechanism for delivering reminders. 70 Participants in the RCT by Bos et al. 55 demonstrated a significant preference for a reminder by mail (56.3%), followed by a telephone reminder (26.0%) and a reminder by SMS (17.7%).

Findings from other studies

No additional studies.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

There is strong consistent evidence from six studies47,48,60,66,96,97 that the timing of the appointment reminder has no bearing on the effectiveness of the reminder. This evidence indicates that a reminder can be sent any time between 1 and 7 days without obvious reduction in the effectiveness of the reminder. It is possible that reminders sent up to 2 weeks before the appointment may also be effective.

A recent good-quality systematic review47 of 29 studies providing an appointment reminder by telephone, SMS or automated telephone calls for patients attending general hospital outpatient appointments investigated the relationship between appointment attendance and the time at which the reminder was issued (varying between 1 and 7 days). There was no significant relationship between attendance and the time at which the reminder was issued (Spearman’s rank-order correlation r = 0.18). This finding is supported by a systematic review investigating the effectiveness of postal and telephone reminder systems for first-time appointments in predominantly psychiatric settings. 48 Based on two RCTs, the author found that there was no significant difference in the effectiveness of either telephone reminders or postal reminders delivered 1 day or 3 days prior to the appointment. In a literature review and cross-sectional survey describing the effect of a postal reminder system on UK adult psychiatry clinic attendance, Magnes97 reaches the conclusion that a reminder sent at 14 days seems to work as well as prompts 3 days prior to the appointment. The author provides weak evidence from his literature review and a small very low-quality cross-sectional survey to support this claim.

In their recent three-arm RCT, Griffin et al. 66 investigated whether an IVR system telephone call at 3 days or 7 days was equally effective as a nurse-delivered telephone call 7 days before the procedure at educating and preparing patients (n = 3610) for flexible sigmoidoscopy and colonoscopy examinations in a US health-care setting. All calls included an appointment reminder, information about preparation for the examination and encouragement to prepare for and attend the examination. The authors found no statistically significant differences across the three study arms in appointment attendance, indicating that timing of the IVR message did not influence appointment attendance. Christensen et al. 60 conducted a prospective and randomised controlled study to evaluate whether confirmation calls made 1 or 2 working days before scheduled appointments reduce the percentage of broken appointments in a children’s hospital dental clinic. A total of 313 patients were randomly assigned to three groups: (1) confirmation call made 1 working day before appointment, (2) confirmation call made 2 working days before appointment and (3) control group (no confirmation call). Overall, there was a 62% reduction in non-attendance at appointments among patients who received a confirmation call compared with the control group and there was no significant difference between confirmation calls placed 1 or 2 working days prior to the appointment (p = 0.51). Finally, in a non-randomised comparative study, Sims et al. 96 examined the effect of SMS, or text message, reminders on the attendance of appointments at four community mental health clinics in London. Attendance of outpatient appointments was examined between March and June of 2008 (n = 648), 2009 (n = 1081) and 2010 (n = 1088). Reminder messages were sent 7 and 5 days before an appointment in 2009 and 7 and 3 days before an appointment in 2010; patients in the 2008 sample received no reminder messages. Appointment attendance during the sample periods was compared by using multiple logistic regression analysis and adjusting for sociodemographic and clinical confounders. There was no difference in attendance behaviours between the two different reminder schedules, indicating that the 7- and 5-day and the 7- and 3-day reminder systems were similarly effective. 96

A limited number of trials explore and directly compare different notification periods; for example, Griffin et al. 66 compare the same reminders sent at 7 days and 3 days. In other cases, the comparison periods are determined not by the optimal notification period but by the inherent limitations of the technologies, e.g. assumptions regarding postal delivery times. 74,84,86 As a consequence, technology and notification period variables are frequently confounded making isolation of their effect even more challenging.

Hasvold and Wootton47 found no apparent difference in the non-attendance rate whether the reminder preceded the appointment by a day or a week. One important implication of assuming equivalence for different notification periods is that, in the event of no period favouring patient preference, the decision on when to transmit the reminder becomes bound up with considerations that relate to the capacity of the service to reschedule patients into vacant time slots. Furthermore, parity between different notification periods questions assumptions regarding the importance of reminders as a response to ‘forgetting’. The implication is that, with no direct relationship between notification period and likelihood of forgetting, other obstacles or commitments, perhaps seen by the patient as non-negotiable, are equally likely to impair the likelihood of attendance.

However, the main focus of the research agenda is on comparing the technologies with the associated implications that (1) the main task is to deliver the reminder to the patient and (2) the quest is for a single technology (a ‘magic bullet’) that caters for the needs of the average patient and, by association, for the needs of all patients. Perhaps the only trials that recognise a ‘horses for courses’ approach are the trials by Perron et al. ,78 which used a progressive escalation procedure, and by Rutland et al. ,83 which employed multiple complementary approaches. Car et al. 43 identified one low-quality study that showed that mobile text message reminders with postal reminders, compared with postal reminders alone, improved rate of attendance at health-care appointments (RR 1.10, 95% CI 1.02 to 1.19).

Findings from other studies

No additional studies.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

The aforementioned emphasis on comparison of technologies has also constrained the degree of variation in exploring whether or not the source of the message is an important variable. We might hypothesise, for reasons associated with the therapeutic relationship, that a patient is more likely to attend if invited personally by a clinician. This is apparently supported by the findings of Nelson et al. ,75 who indicated that those receiving text reminders are less likely to attend than those receiving telephone calls from clinic staff. They hypothesise that this may relate to personal relationships between clinical staff and patients; greater interactivity is offered by personal telephone calls and text messages are easier to ignore. However, such conclusions may be driven by certain assumptions that themselves need to be questioned.

Findings from other studies

Evidence from non-randomised studies presented later in the report (see Chapter 6) suggests that doctors may, perversely, give the impression that attendance at a particular appointment is not that important. In contrast, reception staff who are directly inconvenienced by non-attendance, having to revisit schedules to reallocate the cancelling patient and to seek replacements to utilise vacant slots, may view cancellation much more seriously. Adding a further layer of complexity is that, if patients anticipate an unfavourable or unfriendly reaction from administrative or clinic staff when cancelling, they may be less likely to notify the service at all, leading to arguably more uncertainty, inconvenience and expense.

Telephone reminders (evidence category Ia)

SMS reminders

Findings from effectiveness review (systematic reviews and randomised controlled trials)

None of the included studies explored the relationship between reminder systems and patient perceptions of the ease of rebooking or cancellation. The assumption is that a reminder system will facilitate cancellation and/or rebooking by offering the patient an opportunity and mechanism to cancel or reschedule the appointment.

There are suggestions that services sometimes select reminder technologies because of a perception that they will make it easier for patients to cancel or reschedule. Specifically, in two RCTs78,84 the choice of reminder method was based on the perception that this method would facilitate increased cancellation or rescheduling. Sawyer et al. 84 used telephone reminders to allow increased interaction with adolescents to allow them to cancel or reschedule within a tertiary public adolescent health-care service at the Centre for Adolescent Health, Melbourne, VIC, Australia. Clinic bookings of adolescents were randomly assigned to either a telephone reminder 1 day prior to their appointment or a routine booking (no reminder). The use of reminders (intention-to-treat analysis) significantly reduced the non-attendance rate from 20% to 8% (OR 0.35; p = 0.03). Non-attendance was three times more likely for a new appointment than for review appointments. A total of 79% of parents reported that telephone reminders were helpful at prompting attendance. It was concluded that telephone reminders greatly improved attendance at these adolescent clinics.

Similarly, Perron et al. 78 used a staged personal telephone/SMS/letter reminder design in which patients booked in a primary care or HIV clinic at the Geneva University Hospitals were sent a reminder 48 hours prior to their appointment according to the following sequential intervention: (1) telephone call (fixed or mobile) reminder; (2) if no telephone response, a SMS reminder; and (3) if no available mobile phone number, a postal reminder. This design was selected specifically to reach the maximum number of patients through different reminder systems, while giving the opportunity to a majority of patients to immediately cancel or postpone an appointment, allowing the service to reallocate consultations.

Findings from other studies

The ease of access to technology to cancel or reschedule appointments was addressed by only one qualitative study101 of attendance behaviour from the patient viewpoint at a young adult diabetes outpatient clinic in the UK. Attendance rates for 231 clinic appointments over 19 months for 102 patients were calculated. Semistructured interviews with 17 patients explored the routines, thoughts and feelings around clinic appointments. The study found that long waiting times on the telephone was a deterrent to cancellation. An audit of the cancellation system revealed that as many as 17 patients might be waiting in the queuing system on the telephone at any one time, with waiting times of more than 20 minutes, to contact an operator. In some cases, telephone calls were disconnected with no chance to wait or leave a message. Patients stated that they had also been told by friends or staff not to call the central cancellation line because of difficulties in getting through. Clearly, in this study, patients had a well-justified perception that cancellation/rebooking would not be simple and may even not be worth trying to do at all.

Interactivity of reminders

Reminder systems can be designed to facilitate patient intentions to cancel and rebook, to a variable degree, depending on the reminder technology, i.e. personal telephone reminder (maximum facility), SMS reminder with the opportunity to reply by SMS (good facility), SMS or other form of reminder providing a telephone number to call (some facility – the patient does not have to look elsewhere for the number).

Findings from effectiveness review (systematic reviews and randomised controlled trials)

There is strong evidence that a reminder will increase patient cancellation and rebook rates [see Evidence statement (E): there is strong consistent evidence that a reminder systems will promote cancellation of appointments (evidence category Ia)], indicating that reminder technology facilitates patient intentions and their subsequent cancellation and rebooking behaviour. In particular, the evidence shows that telephone reminders fare better in this regard than SMS reminders. Reti80 suggests that one dominant benefit of telephone reminders may be the ‘promotion of suitable cancellation behaviours’. The evidence on the effectiveness of SMS reminders to increase cancellation or rebooking rates is less clear, but one reason for this is that SMS reminders are frequently sent 1–2 days prior to the appointment, allowing very little time for patients to engage with cancellation or rebooking behaviours. A second reason may be that when patients receive a reminder they frequently are asked to phone a telephone number rather than simply replying to the text. Taylor et al. 85 found that the cancellation rate for patients receiving a SMS reminder that asked them to telephone a cancellation line if they were unable to attend was observed to be higher (20%) than that for patients not receiving a reminder (15%), and this observed difference in cancellation rates approached statistical significance (OR 0.70, 95% CI 0.47 to 1.05). The authors suggested that a key benefit of sending SMS reminders was in prompting patients to cancel their appointment rather than failing to attend.

Findings from other studies

In a survey of patients who had failed to attend a general surgical outpatient clinic appointment, Hogan et al. 20 report that a ‘text message reminder would have encouraged 51% (n = 49) to have either attended the appointment or contacted the hospital to cancel’. However, the way that SMS reminders are delivered varied across studies, hence influencing the way that they are used to cancel or reschedule. Most SMS reminder systems provided a telephone number for the patient to call to cancel in the body of the reminder message. 70,85 In some studies, the patient could also reply by SMS;78,98,102 however, few of these studies reported cancellation rates. Two non-randomised comparative studies98,99 did not report a statistically significant increase in cancellation rate when the reminder was implemented. In a third non-randomised comparative study,100 patients who received the reminder were able to either call a telephone number or reply to the text in order to cancel their appointment. The cancellation rates among patients who received a SMS was 13.4%, compared with 11.2% among those who did not receive a reminder. The difference between the two groups was not significant (RR of cancellation 1.26, 95% CI 0.98 to 1.61; p = 0.068), but this could be as a result of the small numbers of cancellations in each group. A larger sample size in this study may have resulted in these results reaching significance.

Although it has already been acknowledged above that the evidence for SMS as a means of promoting cancellations is not very clear [see Evidence statement (E): there is strong consistent evidence that reminder systems will promote cancellation of appointments (evidence category Ia)], none of the studies reported in this section or in the section titled Evidence statement (E): there is strong consistent evidence that reminder systems will promote cancellation of appointments (evidence category Ia) was powered to investigate the effectiveness of reminder systems to promote cancellations. Therefore, it is not possible to reach a conclusion about whether or not there would be a difference in effectiveness if a patient has to call a telephone number to cancel compared with responding by SMS.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

There is no evidence from the effectiveness review that particular systems facilitate cancellation.

Findings from other studies

Good evidence from other types of studies indicates that whether or not a patient cancels/rebooks will be influenced by system factors that facilitate/hamper patient intentions to cancel/rebook. Four studies18,101,103,104 acknowledge potential system difficulties hampering cancellation by telephone, outside the context of a reminder system. Problems included:

• difficulties faced by patients accessing the central booking line (long waiting times, call disconnecting with no option to wait or leave a message, phone engaged)18,101,103,104

• poor reputation regarding access to the central booking number for the reasons outlined above, discouraging patients from making contact101

• internal hospital communication problems resulting in cancellations and changes of address not being passed on to the clinic. 101

A systematic review on primary care appointment attendance behaviour (i.e. not on appointment systems per se)18 cites several studies indicating that busy telephone lines or administrative problems around the cancellation process are significant issues. This is acknowledged in a survey of primary care staff in Yorkshire: ‘When asked specifically how easy it was for patients to cancel . . . It was acknowledged that phone lines could be busy, making it difficult to cancel . . .’. 104 Similarly, in another study that explored patients reasons for failure, 14 out of 75 participants stated that they had telephoned or written to cancel their appointment already, indicating possible difficulties in the cancellation system. 103 The convenience of e-mail for cancellation, in so far as there is no need to wait to get through to a receptionist, is proposed by Atherton et al. 42

Findings from effectiveness review (systematic reviews and randomised controlled trials)

There is no evidence from the effectiveness review that structural factors may hinder rebooking.

Findings from other studies

Only one study91 explicitly examined the structural barriers to rebooking and cancellation. This study looked at the effectiveness of a telephone reminder on attendance and cancellation for dental appointments in a deprived community in Sheffield, UK. The authors observed that many patients in the community did not have telephones, or their telephones had been disconnected. This limited access to reminders and the ability of the patient to cancel, but many patients who were contacted did cancel. The authors assumed it likely that, if other patients had been contactable, they may have also cancelled if given the opportunity. Self-evidently, the requirement to telephone and cancel an appointment might present difficulties for those populations with higher levels of mobility or deprivation and, thus, less likely to have access to a telephone.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

There is no evidence to explore whether or not receiving an appointment reminder will increase patient perceptions that cancelling the appointment is the morally right thing to do, if the patient has no intention of taking up the offer of the appointment.

Findings from other studies

The issue of the morality of cancellation and rebooking was not explored directly. The issue was explored peripherally in one qualitative study105 included in one systematic review18 and two surveys. 103,104 Two included studies suggest that cancellation is not given a high priority by patients, suggesting that they forget to cancel or do not think to do so in the first place. A study105 exploring dental appointment attendance reported ‘Well, a person has a lot more to think about than . . . just sitting around making a bunch of phone calls, so much is going on in my life, . . . you must have kids, and schools to deal with, and parent’s meetings and work . . .’.

Primary care staff surveyed by Hussain-Gambles et al. 104 did not perceive patients to have a sense of moral duty to cancel. In this study, some staff suggested that some patients may be too embarrassed or ashamed to cancel, many staff felt that patients were disrespectful or irresponsible. Similarly, in a study of appointment attendance for a surgical clinic serving a deprived population,103 the authors speculate that the patients may have a negative public attitude towards the NHS, with a corollary that this may influence whether or not patients are prepared to cancel appointments.

Findings from effectiveness review (systematic reviews and randomised controlled trials)

In a recent systematic review, Guy et al. 44 performed subgroup analyses by age. They found no significant subgroup differences in reminder effectiveness by broad target age groups (i.e. paediatric, adult, older) (Table 8). However, their age classification was a relatively crude construct based on the median age of patients receiving SMS reminders, or specification of clinic type, e.g. as paediatric.

In six papers94,95,98,106,108,109 in this systematic review,44 the SMS reminders targeted patients attending paediatrics clinics. In paediatric settings, the SMS reminder was sent to the young patient or their parent/carer; in the non-randomised study by Stott and Gray,109 SMS reminders were sent directly to the young people. In two RCTs, the target population were adults (mean ages of 38 years64 and 33 years72). In four papers (three RCTs57,59,73 and one non-RCT107), the target group was older (mean age of 58 years57 or 50 years,59 aged 50–70 years3 or 68% > 50 years107), and in five others the age of the target group was not specified.

The systematic review of RCTs by Guy et al. 44 also demonstrated that SMS reminders are effective in a wide range of age groups from paediatric to older clients. Mobile phone usage data demonstrate that in many countries more than 90% of the population own mobile phones, but uptake is higher in younger people. 109 Neal et al. 110 postulate that because younger patients have higher non-attendance rates at clinical services, SMS reminders may be more beneficial in this group; however, this is not supported by any empirical findings.

The source of the original appointment

Only two studies25,80 were identified that explored the differential impact of reminders delivered by different sources and these produced equivocal findings. In addition, limited and inconsistent evidence was yielded on the effect of source of appointment on attendance behaviour. Just two UK papers89,121 reviewed included quantitative analyses addressing this hypothesised relationship and these related to contrasting settings and employed inconsistent categorisation of sources.

A UK-based study in a gynaecology clinic121 compared 105 non-attenders with 105 attenders at the same index clinics and concluded that referrals from primary care were more likely to be missed than interspecialty or referrals from the accident and emergency (A&E) department. In contrast, a study of attendance at a specialist alcohol treatment centre in the UK89 undertook multivariate analysis of patient characteristics of non-repliers (n = 100), non-attenders (n = 100) and attenders (n = 100) and found no statistically significant association between source of referral and attendance. A study in a paediatric dermatology clinic in Hong Kong122 found that non-attendance was significantly higher among referrals from the emergency department than among referrals by private practitioners (p = 0.05).

A qualitative study exploring attendance for breast screening by women in Turkey123 provided some support for the hypothesis that invitations/recommendations for screening received from family doctors might be important in encouraging attendance. Respondents in the study reported ‘if we face a genial approach, we wouldn’t hesitate going to a doctor, even if it is a male doctor’, ‘we might be ignorant, but our doctor should enlighten us in detail, we need information’. 123

Despite the limited information in the review, a scan of the wider literature (including citations in our reviewed papers) indicates that the hypothesis that the source of the appointment/invitation can influence attendance has received some attention in both research and practice, particularly in relation to screening interventions. A key contrast of interest has been between GP/primary care referral and referral from a hospital clinic or central administrative unit. For instance, in relation to attendance for breast screening, some UK interventions have involved letters of endorsement or telephone calls from the woman’s local GP practice in addition to standard screening invitations and leaflets, but the reported impact on uptake has been disappointing. 124,125

Timing of the appointment

The review identified several studies that had explored patterns of attendance by appointment timing, looking at both the day of the week and time of day. Explanation for higher non-attendance at particular times or on particular days has centred on the incompatibility of the appointment with employment commitments, with some discussion of caring commitments also.

Time to wait to the appointment

The review generated strong consistent evidence that a longer time between the date of the appointment being made and the date of the appointment taking place is associated with higher non-attendance. Multiple primary quantitative comparative studies (across a range of contexts and patient groups) found a statistically significant association between time between appointment scheduling and appointment date and non-attendance. No studies were found that reported a lack of association between the time between booking and appointment and attendance behaviour.

In a UK RCT study based in a respiratory clinic,82 authors employed linear regression to explore predictors of non-attendance. They demonstrated that each increase in wait time for initial patient visits of 30 days resulted in a 25% increase in the no-show rate. Similarly, a study in a UK eating disorders clinic serving young adults and adults132 reported that the adjusted OR of attending an appointment was 2.4 times greater for those with short waiting times (≤ 4 weeks) than for those with long waiting times (≥ 5 weeks). A logistic regression model in which waiting time (in weeks) was entered as a continuous variable suggested that the odds of attendance at an appointment falls by 15% per week (OR 0.85, 95% CI 0.770 to 0.945; p = 0.002). A UK study121 in a gynaecology clinic serving an ethnically diverse and relatively deprived population also reported that the longer the interval between hospital appointment letter and the actual appointment date, the more likely patients were to default on their appointments (p = 0.001). Another UK study,89 based in a specialist alcohol treatment centre, also found that attendance was less likely with longer delays between referral and an appointment invitation being sent by the clinic. However, in this context the authors also noted that delays could result from both patient- and agency-originated factors contributing to the total waiting time and that patient delays might indicate greater ambivalence about their alcohol problem.

A review paper133 employing hand searches and forward citation searching focused on mental health services in the UK and also reported that waiting time to appointments was a factor that is consistently linked to higher non-attendance.

In addition to the UK studies, numerous studies from other settings report a negative relationship between attendance and time to wait for the appointment. A US study based at a veterans outpatient clinic134 found that missed appointments increased with time between scheduling and appointment up to 13 days, with the rate of missed appointments being between 10% and 20% when the interval was 1 day and between 16% and 25% for an interval of 13 days; however, they found no further rise thereafter. In contrast, for cancellations, the authors found that the rate continues to rise with time interval up to and beyond 1 month. A study of adult primary care in Geneva79 reported that having a follow-up appointment of more than 1 year was significantly associated with missed appointments. Similarly, an Australian study in a general hospital setting93 reported that those with a waiting time of more than 21 days until their appointment were significantly less likely to attend than those with a shorter waiting time (OR 2.8; p = 0.002).

Patient involvement in the initiation and scheduling of the appointment

Several studies support the hypothesis that when patients make their own choice of appointment time and date, attendance is better. A US-based study of elderly patients attending a veterans clinic138 explored the effects of advanced clinic access on patient appointments. Patients were sent a letter advising them to call and make an appointment a month before their next anticipated visit. Advanced clinic access (or open-access scheduling system) worked for elderly veteran patients and indicated that they would prefer to schedule their next appointment than to use the previous system. A UK study in a paediatric hospital108 discusses ‘partial booking’, whereby patients are sent a letter inviting them to call and make an appointment. The authors report that the rate of non-attendance for first appointments was significantly lower for those who experienced partial booking than those who did not regardless of whether or not a reminder was sent (5.2% vs. 15.3% among those with no SMS reminder and 3.4% vs. 9.8% among those who did receive SMS). Murdoch et al. 12 refer to an earlier study139 in which they achieved a reduction in non-attendance by asking patients to make their own appointments and confirm their intention to attend.

Closely linked to the above discussion is the issue of whether the appointment is patient or provider initiated. Service-initiated appointments will often be screening-type appointments for which most patients will not recognise any current symptoms or prompts. However, service-initiated appointments may also be referrals between departments. In both these cases, some studies suggest lower levels of attendance than when patients initiate the appointment themselves in response to some perceived need (see First versus follow-up appointments for a discussion of evidence on this).

First compared with follow-up appointments

As discussed above, a small number of studies48,89 have examined whether reminder effectiveness varies according to whether or not an appointment is a first or a follow-up, but findings are rather inconsistent. The review also identified several studies that explored attendance behaviour more generally in relation to first compared with follow-ups and, again, the findings were equivocal.

Five quantitative papers74,98,108,121,135 reported higher no-show rates for follow-up appointments in comparison with initial appointments, with most speculating that this is because initial appointments are more likely to be prompted by the patient rather than the provider (see Patient involvement in the initiation and scheduling of the appointment).

However, four quantitative comparative studies59,82,84,85 reported higher no-show rates for initial rather than for follow-up appointments. Suggested explanations related to patient–provider relationships and perceived value of the appointment (see Patient–provider ‘alliance’, communication and ‘engagement’). A UK study based in a respiratory clinic82 found that the overall no-show rate was significantly higher for new patients than for established patients. A study of a general adolescent clinic in New Zealand84 reported that 22% of new appointments were not attended compared with 9.5% of review appointments (OR 2.7, 95% CI 1.1 to 6.8; p = 0.034), although this no longer reached significance in a multivariate analysis.

Furthermore, some studies63,112 report no significant differences between first and follow-ups in non-attendance.

Explanations and implications for reminders

The studies highlighted above suggest that, when there is a poor relationship between the patient and the service (or particular providers within the service), the patient will tend to perceive the likely costs of attending an appointment to outweigh any anticipated benefits, for instance because interactions with staff will be unpleasant or stressful or because the care/treatment on offer is perceived to be inappropriate, unhelpful or even risky/harmful.

These findings suggest several implications for reminder systems and wider intervention to support attendance. There may be some, relatively simple, measures that can be taken to ensure that reminder systems do not have a negative effect on patient engagement and patient–provider relationships, for instance ensuring that patients receive reminders in the format of their preference and do not receive reminders that they find irritating or stressful. 55 It is worth noting that several of the reviewed studies highlighted the fact that many patients view regular communication from their health professional positively, so that reminders may actually contribute to a more positive patient–service relationship.

High levels of non-attendance are likely to reflect obstacles to attendance over and above forgetfulness and may be reflective of poor patient experiences and low levels of therapeutic alliance, such that reminder systems alone are unlikely to be sufficient to achieve high levels of attendance.

Although many patients indicate that they are happy to receive a reminder (see Chapter 6, Other reminder characteristics), there is a risk of antipathy from a sizeable percentage of patients who feel negative or very negative about reminders. Those who were disturbed by the reminder gave the following reasons: telephone call too early in the morning, they were waiting for other important results, they were contacted at work, felt that were being treated as senile or felt that the reminder was unnecessary. 79

Service location and transport difficulties

The review indicated that service location and transport difficulties have been implicated in non-attendance in both research and practice interventions across a variety of settings. However, the available evidence regarding the impact of service location on attendance levels appears somewhat limited and equivocal, suggesting that the importance of transportation issues is likely to vary importantly between service settings and patient groups.

Several descriptive studies identified transport difficulties as a reason given for non-attendance in diverse settings. For example, a US study118 reported that the main reasons associated with missed appointments in paediatric clinics were effort (waiting time, distance, transportation, parking), communication (misunderstanding about appointment time and date) and forgetting. Hamilton and Gourlay148 conducted 120 face-to-face survey interviews with women who had missed maternity appointments in the London Boroughs of Richmond and Newham. Almost 70% considered transport the main reason for their missing an appointment and more than half the women had dependent children and reported difficulties travelling with their children to hospital, especially where this travelling was on public transport. In a qualitative Turkish study123 focused on breast screening attendance, respondents noted that receiving the screening procedures in a familiar institution would be more convenient. Those women who discussed transportation facilities as something they were concerned about shared lower socioeconomic status.

However, in contrast, one descriptive study130 of diabetes patients in New Zealand (n = 142) reported that transport difficulties/location issues affect only a small proportion of patients and are not a major reason for non-attendance, with just 2 out of 142 patients giving this as the reason for missing their appointment. Furthermore, the available quantitative studies in the review produced equivocal evidence on the relationship between travel distance and appointment attendance.

Three quantitative studies explored the association between travel distance and attendance. One quantitative study89 indicated that location of the clinic may affect attendance, but only for patients who live very close to the clinic. This UK study based in a specialist alcohol treatment centre undertook multivariate analysis of patient characteristics of non-repliers (n = 100), non-attenders (n = 100) and attenders (n = 100) and found that patients who had a very short distance to travel were more likely to attend their appointment. The distance variable was calculated from patient postcodes. These were divided into four bands on the basis of distance from the clinic (with band 1 being the immediate surrounding area). Patients living in distance band 1 were more likely to attend than patients living at all other distances [OR vs. distances 2, 3 and 4: 4.41 (p < 0.001), 3.12 (p < 0.01) and 3.32 respectively (p < 0.05)]. Further from the clinics, there was no difference between attenders and non-attenders and there was also no effect of distance on the likelihood of replying to the appointment invitation. The authors speculated that, although the clinics were easily accessible by public transport, the cost of travelling probably reduced attendance by some patients. They cited previous observations, from a US treatment manual,149 that problems arising from lower socioeconomic status, including lack of transportation and a telephone (which would, in addition, make it unlikely that patients would cancel their appointments), may lead to decreased chances of attendance specifically for treatment of alcohol problems. Other research150 had previously demonstrated that distance travelled is also correlated with non-attendance at a specialist alcohol clinic.

One small RCT in the USA,61 that sought to explore whether or not a series of interventional supports could improve attendance levels among substance abuse patients, found no significant effect overall, but reported that van transportation was the most frequently used element of the package of support that was on offer. However, two quantitative comparative studies60,82 concluded that service location is not related to attendance; and one study113 suggested that patients living closer to the clinic were more likely not to attend.

A US study60 of attendance at a paediatric dental clinic serving a deprived population reported that 85% or more of patients travelled < 1 hour to the appointment and that only 3 of the 28 long-distance patients (> 1 hour) failed to show up for their scheduled appointment. The authors also reported that the long-distance patient group demonstrated better punctuality than the study group as a whole. They speculated that this may reflect the greater degree of planning required for the long-distance drive or the practice of co-ordinating multiple health care. Similarly, a UK study82 of respiratory outpatients, employing a randomised controlled design and involving 504 patients, reported that travel distance was not associated with attendance.

Interestingly, findings from a UK service evaluation113 of a genitourinary clinic involving telephone interviews with 182 non-attenders contradict the hypothesis that easier/closer appointments are more likely to be attended. Patients were significantly more likely not to attend if they were unemployed, resident within the local catchment population area (i.e. Southampton Primary Care Trust area) and had previously accessed genitourinary medicine (GUM) services. The authors speculate that as such patients do not have competing employment demands, live close to the clinic, have previously made and kept GUM appointments and do not cite clinic opening hours, or similar, as a reason for not attending, their non-attendance does not reflect problems of accessibility.

A brief scan of the wider literature also suggests some varying degrees of success for interventions that have sought to overcome transport difficulties and, thereby, improve attendance at appointments. For example, Ogilvie and Mayhew151 report on a carefully conducted audit in the UK and conclude that holding appointments for school-aged children at schools rather than clinics was an effective way to increase attendance by parents and children. In contrast, Bell et al. 125 evaluated an intervention through which private transport was provided to minority ethnic women, but found no increased uptake of breast screening appointments.

Seniority/identity of the health-care provider

Findings from the review were equivocal in relation to the effect of health-care provider identity/seniority on attendance. Three descriptive quantitative studies with adequate sample sizes reported a positive association between seniority/qualification level of the provider and attendance rate.

A UK study based in a dentistry clinic98 reported that significantly more patients failed to attend their appointment with non-specialist dentists and a similar pattern was observed at clinic level, with the non-attendance being low at specialist staff clinics but high at non-specialist staff clinics, although the difference did not attain statistical significance. Similarly, a study135 in Switzerland of adolescent attendance at outpatient clinics explored the characteristics of non-attenders (n = 2193, women = 1873) and found that the intended provider being a member of the paramedical staff (vs. a physician) significantly increased the risk of missing appointment for females with a somatic diagnosis but not for those in the psychiatric model. The risk of missing an appointment was also increased among those who had consulted a paramedical provider at the last appointment. The most prominent factor predicting non-attendance was the intended provider, with a non-weighted probability of non-attendance of 6% for physician appointments compared with 17% for paramedics. In another Swiss study,79 the authors examined attendance at a primary care and ambulatory HIV clinic in a deprived Genevan population and studied the characteristics of 2123 patients included in a RCT and found that being cared for by a junior doctor rather than a senior doctor was significantly associated with missed appointments. Analysis of predictors of non-attendance showed an OR of 0.5 (senior vs. junior; 95% CI 0.27 to 0.93).

In addition to these primary studies, the review identified a study by Magnes97 that included a limited review of earlier work on psychiatry outpatient attendance. This study identified a paper by McIvor et al. 152 in which non-attendance rates were examined by cross-sectional survey of 482 patients seen by psychiatrists of different grades and a consultant clinical psychologist over a 21-month period. The clinical psychologist’s patients had the lowest rate of non-attendance (7.8%), followed in turn by those of consultant psychiatrists (18.6%), specialist registrars (34%) and senior house officers (37.5%).

However, in contrast to the findings reported above, a study of physiotherapy in Ireland112 reported higher non-attendance for appointments with senior grade staff and suggested various reasons why this might be the case. This prospective study found that non-attendance rates were significantly higher for senior grade (12.7%) than staff grade (8.9%) physiotherapists. A survey by Murdock et al. 12 found that fear of being seen by a junior doctor rated fifth on a list of reasons for non-attendance alongside more common explanations.

Explanations and implications for reminders

Whether and how societal/community-level norms, attitudes and understandings shape attendance behaviours among different groups of patients deserves greater investigation and there is greater scope to consider how these factors might be relevant to the design of reminder interventions.

Age

Our review indicated that age has been found to be associated with attendance behaviour in a large number of studies across varied geographical and clinic settings. Despite the fact that differences in the categorisation of age groups complicates comparisons across studies, a largely consistent pattern is reported among adult patient populations of higher levels of non-attendance among younger patients than older patients.

Sharp and Hamilton162 provide an informal review of the evidence on non-attendance and report that youth is frequently found to be associated with hospital non-attendance. Parikh et al. ,77 in their study of outpatients from 10 clinics in the USA, used linear regression modelling to demonstrate that, for every 1-year increase in age, the absolute no-show rate decreased by 2.4% (p < 0.0001). Similarly, Costa et al. 63 found a decreasing level of non-attendance rate with increasing age in their study of outpatient attendance at a district hospital in Portugal. In their study in primary care of patients receiving lipid-lowering therapy in the Republic of Korea, Cho et al. 59 noted age differences in attendance, with older adults being more likely to attend than younger adults. In a study of primary care attendance in urban Geneva,79 missed appointments were significantly associated with younger age. Similarly, Booth and Bennett89 found that older patients were more likely than younger patients to attend their assessment appointments at an alcohol treatment clinic. Waller and Hodgkin163 examined attendance patterns across nine general practice clinics in the UK and reported that non-attendance rates were highest among young adults. The investigation by Moore et al. 126 of over 4000 appointments at a family practice centre in the USA found that older patients were less likely to fail to show than younger patients and that appointment keeping increased with age. Thus, an older person was 1.18 times more likely to keep an appointment than a person 10 years younger (OR 1.18, 95% CI 1.13 to 1.23). Neal et al. 108 also report higher non-attendance rates at clinical services among younger patients in the UK general practice context.

Patterns of attendance for children across ages or in comparison with adults are less clear and, as in the UK context they are often attending specialist paediatric clinics, it is difficult to draw direct comparisons. In their study of appointment attendance in a Scottish general dental clinic, Patel et al. 164 reported the highest rate of non-attendance among the 21–30 year age group, followed by 31–40 years and then 11–20 years. Clearly, this study focused on younger patients and grouped children with young adults, which may explain why the observed pattern was not similar to the patterns presented in previous studies. Bos et al. 55 examined attendance at an orthodontic clinic in the Netherlands and found similar levels of non-attendance in patients aged under 12 years and 12–18 years (around 3.4%) but higher non-attendance in those aged greater than 18 years (8.1%), although numbers included in the study were small.

Gender

Our review indicated that there is inconsistent evidence regarding differential attendance behaviour between male and female patients, but that researchers often confidently state that male patients are more likely than female patients to default on appointments.

Several studies reported higher levels of non-attendance among male patients than females. Sharp and Hamilton162 provide an informal review of the evidence on non-attendance and report that being male is one of the main associations with hospital non-attendance. Moore et al. 126 investigated over 4000 appointments at a family practice centre in the USA and found that females tended to be less likely to miss appointments than males, although the effect was of borderline statistical significance in a multivariate analysis (OR 0.8, 95% CI 0.7 to 1.0; p = 0.06). Hon et al. 122 report on a study of non-attendance at a paediatric dermatology clinic in Hong Kong and found that 67% of non-attenders were male and that males were more than twice as likely as females to fail to attend clinic (p = 0.010).

However, in contrast, Sims et al. 96 found that gender was not a significant predictor of attendance at four adult community mental health outpatient clinics in London. Similarly, Bos et al. ,55 in their study of attendance at an orthodontic clinic in the Netherlands, including both adult and child patients, reported no significant difference between males (n = 109) and females (n = 153). Meanwhile, Can et al. 56 reported that females were twice as likely as males to not attend their appointment in their study of 232 new patients referred to an orthodontic clinic in England. Furthermore, Waller and Hodgkin163 examined attendance patterns across nine general practice clinics in the UK and reported that women were disproportionately represented among the group of patients who were frequent defaulters (defined as missing an appointment five or more times per year).

Deprivation

Our review indicates that there is weak but consistent evidence to suggest that deprivation status is a factor in attendance. Although the use of a variety of measures of deprivation or low socioeconomic status makes aggregation of evidence across studies difficult, as does variation in the economic context across geographical settings, there is nevertheless evidence from a range of studies that patients with higher levels of deprivation are more likely to miss appointments than those who are more affluent/advantaged.

Several studies employing individual measures have reported a negative association between deprivation or low socioeconomic status and the likelihood of appointment attendance. Sharp and Hamilton162 provide an informal review of the evidence on non-attendance. They report that deprivation is one factor that is commonly associated with hospital non-attendance. Ramm et al. 166 compared attenders and non-attenders at a cardiac rehabilitation programme in New Zealand using simple univariate statistics and found lower levels of attendance among patients categorised as having a low socioeconomic status on the basis of their occupational grouping. In a study of colposcopy attendance in Cornwall, Oladipo et al. 76 reported that younger, unmarried women of lower socioeconomic status were less likely to attend than more affluent patients. Another colposcopy clinic study,86 in Manchester, reported a similar association with deprivation. In their study of 232 new patients referred to an orthodontic clinic in England, Can et al. 56 reported that patients living in an area of high social deprivation were 2.7 times more likely to fail to attend an appointment than people who were resident in more affluent areas (95% CI 1.1 to 6.5). A study167 (as cited by Dixon-Woods et al. 168) of management of women with cancer found that significantly more patients from deprived areas failed to attend hospital appointments, although the difference was small.

A study that employed analysis at the aggregate clinic level also documented this association. Waller and Hodgkin’s UK study,163 in 2000, found that levels of non-attendance at general practice level showed a high correlation (Pearson’s r = 0.72; p = 0.028) with the score at each practice on the Townsend Index of Deprivation (attributed to practices according to the practice population in each enumeration district).

In contrast, a smaller number of studies report no difference in attendance levels between patients categorised by measures of deprivation or socioeconomic status, for example a study by Conduit et al. 169 of attendance at a university-based psychology clinic in Australia. However, when no difference was reported, the results are often compromised by small sample sizes, questionable measures of socioeconomic status or low levels of socioeconomic diversity within the patient population.

Ethnicity

Race and ethnicity are complex, contingent and contested biosocial concepts that have been operationalised in diverse ways in health-related research, making any attempt to compare and synthesise findings across studies fraught with difficulties. 180 It is not, therefore, surprising to find that our review yielded rather inconsistent evidence on the association between (minority) ethnicity and attendance behaviour. Furthermore, compared with measures of deprivation, race/ethnicity appears to have been less often considered in attendance research, particularly outside the USA.

Patel et al. 164 explored attendance at dental clinics in the UK and reported that Asian patients had higher non-attendance rates than English patients in an outpatient setting. In a study130 of non-attendance at Care Plus and diabetes clinics in New Zealand, the ethnic breakdown showed that New Zealand European people were under-represented as non-attendees in proportion to the enrolled population, while Pacific people were over-represented, accounting for 13% of non-attendances while constituting 3% of the enrolled population. Maori non-attendance (13%) was slightly higher than the enrolled population of 11%. An audit study181 of paediatric outpatients in a hospital in the West Midlands of England found differences in the appointment-keeping behaviour of patients by ethnicity. An original 1995 study had shown that non-attendance for clinic visits was around 33% for Europeans compared with 50% for Asian patients. 182 A subsequent follow-up study in 1998, following interventions, showed a fall to 12.0% and 13.5%, respectively, with an overall rate of 12.3%. 183

There is a much wider body of evidence that reports on the health-care experiences of minority ethnic patients in different settings, often highlighting issues of cultural (in)competence, lack of trust and in some cases inappropriate or abusive treatment. Some of this work has made direct connections to low levels of attendance and poor uptake of health-care interventions among minority ethnic groups as compared with majority populations (see, for example, Greer184).

Substance abuse

Our review yielded indicative evidence to suggest that substance abuse can be a factor in non-attendance. A study188 of patients with schizophrenia found that those who missed 20% or more of their appointments were more likely to abuse drugs and alcohol and manifested lower levels of community functioning. Clearly, the proportion of a patient population who are substance abusers will vary greatly between service contexts and geographical areas, with mental health services being the most likely to have large numbers of patients with this characteristic.

Mental health

Our review yielded indicative evidence to suggest that mental health status can be a factor in attendance. One study188 showed that patients who missed more than 20% of appointments in mental health clinics were more likely to abuse drugs and alcohol and exhibited lower levels of community functioning. Sims et al. 96 reported percentages of missed appointments of 36% (in 2008), 26% (in 2009) and 27% (in 2010). The reduction in RR of failed attendance following a text message intervention was 28% between the 2008 and 2009 samples and 25% between the 2008 and 2010 samples.

Donaldson and Tayar189 conducted a feasibility study of use of SMS messages in mental health outpatients. They encountered some unexpected difficulties, for example, despite mobile phone ownership matching that expected for the national average (76%), only 74% of these could remember their telephone number and only 53% were agreeable to being contacted by text message. The authors concluded that although SMS appointment reminders appear to be a potentially useful and cost-effective method of improving psychiatric outpatient clinic attendance rates, it is ‘unlikely that psychiatric clinics would provide as impressive results as those reported in other settings’. 189

Pennington and Hodgson190 conducted a non-randomised study of non-attendance to assessment for clients referred for psychological therapy in relation to invitation type. The study concluded that telephone invitations followed by a telephone prompting reminder significantly reduced non-attendance to initial assessment appointments. Limitations of the study included key variables being omitted or not being controlled which may have biased the findings, together with the small effect size.

Physical illness/comorbidity

Our review yielded evidence to suggest that poor health status or perceived poor health status can be a factor in attendance.

Roberts et al. 111 used a questionnaire-based study to explore reasons for non-attendance among 204 outpatients at neurology clinics in Ireland and found that, after simply ‘forgetting’, being ‘too ill’ to attend was one of the most frequent explanations for missed appointments.

Killaspy et al. 191 is one of several studies8,192,193 that demonstrates that non-attendance at psychiatric services is related to increased pathology and greater need, and serves as a corrective to suggestions that non-attenders may self-select because they do not need to be seen.

Alexandre et al. 8 conducted an exploratory prospective cohort study in New Yorkers with low back pain. Part of their study investigated predictors of attending. Following multivariate analysis, the authors identified that the presence of other comorbidities, such as diabetes and hypertension, was associated with poor attendance at scheduled physical therapy sessions. As part of their RCT, Resnick et al. 192 evaluated attendance at treatment visits among older women post hip fracture. At 6- and 12-month follow-ups, physical health (measured by evidence of comorbidities or overall physical health status) and mental health (measured by evidence of depressive symptoms or mental health status) directly influences attendance at treatment visits. Whitmarsh et al. 193 investigated psychological variables predicting poor attendance or non-attendance at cardiac rehabilitation. Using multivariate analysis they found that poor/non-attenders at cardiac rehabilitation were distinguished by illness representations, distress and usage of coping strategies.

Symptomatology and severity

In contrast to the tendency for some patients to be ‘too ill’ to attend their appointment, the review highlights the fact that attendance at appointments may also be less likely when patients do not recognise symptoms that warrant the appointment on offer. Several studies suggest, perhaps not surprisingly, that patients are less likely to attend when asymptomatic (for instance, when being called for a screening check) or when the symptoms they have are perceived to be minor or unconcerning.

Oladipo et al. 76 report that patients referred to colposcopy are mainly symptom free and, therefore, exhibit some inertia to keep their appointments. Booth and Bennett89 similarly observed that patient contact with the referrer may be initiated after concerns about physical symptomatology or other crises, specifically within the context of alcohol disorders. Roberts et al. 111 reported that when patients were ‘happy’ with their neurological symptoms they were more likely to be non-attenders. In her study of respiratory outpatient clinics in the UK, Roberts et al. 82 cited Grufydd-Jones et al. ,194 who showed that the main reason for non-attendance in an asthma clinic in primary care was low perception of severity. Parikh et al. 77 report, within the context of the US health system, that if a patient thought an appointment was unnecessary (either the condition had resolved or the patient was able to self-manage the symptoms), he/she was less likely to show. As discussed above under Time to wait to the appointment, the higher levels of non-attendance with longer wait to appointment have been explained in part by the resolution of symptoms in the interim period. A study20 in a general surgical clinic in Ireland revealed resolution of symptoms to account for almost one-third of missed appointments. The authors questioned the appropriateness of referral, suggesting that many patients might be managed in the primary care setting. It may also be a function of delays in obtaining a specialist opinion. GPs may be forced to institute treatment measures as intermediate solutions, and often these prove sufficient to alleviate symptoms. This could be viewed as a successful outcome provided that the patients (or GP) contact the hospital to let them know of the resolution of symptoms in sufficient time to reallocate the appointment slot.

However, as Martin et al. 142 observe, ‘one cannot assume that a did not attend is a medical condition resolved’. They further point out that ‘patients still present, but at less convenient times, in less appropriate care settings, with the additional health and financial implications that frequently accompany a worsened condition’. Being too ill to attend was discussed in the previous section (see Physical illness/comorbidity).

Previous patterns of non-attendance

Our review indicated that whether non-attendance at health-care appointments can be predicted on the basis of past attendance patterns has been the subject of some interest in past research. For instance, several studies categorise patients into ‘high non-attenders’ or ‘frequent defaulters’ and try to isolate the characteristics of these patients. However, several studies suggest that patterns of attendance behaviour are unpredictable and that patients do not necessarily display consistent patterns of poor or good attendance over time.

Snow and Fulop101 reported that diabetic patients in their study could not be divided into ‘attenders’ and ‘non-attenders’ as many showed a complex record of attendance, non-attendance and cancellations. Non-attendance can thus be viewed as ‘the manifestation of a critical level of unsuitability in the agreed arrangements for an access episode, either from the outset or as a result of change in circumstances’. 101 Within the study sample, participants could be grouped into those who made a cost–benefit analysis of the obstacles and benefits of going to clinic and those who did not think about it at all. However, some patients moved from one group to another over time. Waller and Hodgkin163 similarly concluded that it was difficult to predict non-attendance on the basis of past behaviour. They noted that most people who defaulted missed one appointment and did not default again. In addition, they found that, among those categorised as ‘frequent defaulters’ (missing five or more appointments in 1 year), most ceased to be frequent defaulters in the following year. The authors suggest that many of these patients may be ‘experiencing a life crisis of some form, or living a chaotic phase of their lives. They will have more pressing problems to deal with than their tendency to default.’163 They further suggest that prioritising the clinical management of the underlying health problem may be more effective in reducing non-attendance than the application of administrative processes to attempt and alter their behaviour.

However, Chariatte et al. 135 found that in an adolescent outpatient clinic, the probability of missing an appointment was explained in their statistical models in part by having missed or cancelled the next to last appointment and having missed the antepenultimate appointment. Having missed the next to last appointment was found to be the most important of these factors, providing 39% of the total explanation.

At least two of the RCTs64,83 examined in review 2 were targeted at defaulters rather than at a general clinic population. The study by Rutland et al. 83 included all patients aged from 16 to 30 years who did not attend a booked GUM appointment during a 6-month study period. In the study by Fairhurst and Sheikh,64 only patients who failed to attend two or more appointments in preceding 12 months were included. Rutland et al. 83 recorded a reduction in non-reattendance of 3.7%, while Fairhurst and Sheikh64 demonstrated a reduction in the non-attendance rate of 5.3%. These rates appear fairly modest compared with some of those achieved in the general clinic population. This may suggest that, for this particular group, it is factors other than ‘forgetting’ that have acted as barriers to attendance.

Simple reminders

All reminders are effective at improving attendance at appointment. Simple reminders, which provide details of timing and location of appointments, are effective at increasing attendance at appointment [see Chapter 4, Evidence statement (A.1): there is strong consistent evidence that simple reminders which provide details of timing and location of appointments are effective at helping a (forgetful) patient to attend their appointment (evidence category Ia)]. Forgetfulness is reported by patients and by health professionals as the most common reasons for not attending a health-care appointment and accounts for between 8% and 44.8% of patients who did not attend their appointments, indicating that this is a substantial problem (see Chapter 5, Reported reasons for non-attendance). Forgetfulness or confusion over date, time or location of the appointment appears to act as the theoretical basis for many reminder systems identified in this review. 59,63,70,76,78,79,98,99,102,106,198–200 It has been suggested that simple reminders may be effective because they act as cues to counteract prospective memory failures. 25 Numerous studies suggest that as many as 80–90% of patients have a positive attitude to receiving a reminder that jogs the memory. 55,79,145,200 The evidence presented indicates that the use of simple reminders is both acceptable and feasible across a wide array of health-care settings. Although very few studies have investigated the influence of preference on reminder effectiveness [see Chapter 4, Evidence statement (B.4): very few studies investigated whether or not the reminder preferences of the patient may influence the impact a reminder has on the patient's appointment behaviour (evidence category VIIa)], it may provide some indication about a patient’s ability to access the reminder, which may ultimately influence the effectiveness of the reminder [see Chapter 4, Evidence statement (B.1): there is strong consistent evidence that the reminder may not be received by the patient (evidence category Ia)]. Therefore, the choice of reminder system remains an important consideration for health services. In an acquired immunodeficiency syndrome (AIDS) clinic serving young adult patients, a high percentage of whom were mobile phone users, SMS was the preferred option. 200 In a different study of adolescents attending dental clinics, participants reported a significant preference for letter reminders (56.3%), followed by telephone reminders (26.0%) and SMS reminders (17.7%). 55 Patient preferences for reminders appear to vary considerably, and we hypothesise that this may be related to sociodemographic factors such as age, employment status, etc. or possibly in relation to the health setting; patients attending sensitive health-care appointments may prefer a very confidential reminder, although this needs to be confirmed through research.

Although simple reminders are effective memory joggers for people who might otherwise forget, this does not explain why, in settings where reminders have been used, forgetfulness continues to be a commonly cited reason for non-attendance. 74,136 Forgetfulness may also be a reason that is offered by patients, as it may be more convenient and easier than the truth. Murdock et al. 12 have suggested that the underlying reason may be apathy; however, the truth may be considerably more complex and uncomfortable than that. Patients may be uncomfortable stating that they did not find the treatment helpful or that they did not like their doctor, etc. The implication of this is that ‘true forgetfulness’ may be over-rated, whereas other reasons may be under-rated or even unaccounted. Numerous other reasons have also been reported (see Chapter 5 for an account of these reasons). It has been suggested that, for patients who receive, or expect to receive, some benefit from attending their appointment, reminders will be effective in increasing appointment attendance. 25 Reminders will be effective in increasing appointment attendance provided the clients receive some benefit from their attendance, such as useful information, alleviation of symptoms or emotional support. 201 If patients no longer find the treatment sessions rewarding, then the prompts will eventually be ignored by the client;201 therefore, clients may fail to attend after the initial appointment because the treatment is simply not meeting their expectations or needs. 25 This suggest that reminders, even simple reminders, have a more complex mechanism than simply ‘jogging the memory’.

Reminder plus

There is currently weak evidence (category IIIa) that ‘reminder plus’, which provides additional information over and above the date, time and location of the appointment, are more effective than simple reminders at helping a patient to attend their appointment [see Chapter 4, Evidence statement (A.2–A.6): there is weak consistent evidence that ‘reminder plus’ are more effective than simple reminders at helping a patient to attend their appointment (evidence category IIIa)]. Many authors have suggested that the provision of additional information may reduce the number of perceived obstacles to patient attendance and, in this way, increase patient intentions to attend, thereby enhancing the effectiveness of a reminder system.

Age

There is weak evidence that age does not have a differential impact on the effectiveness of reminders [see Chapter 4, Evidence statement (F.1): there is sufficient weak equivocal evidence to suggest that age does not have a differential effect, over and above pre-existing appointment behaviour, on reminder effectiveness (in terms of attendance, cancellations or rebooking) (evidence category IVa)], indicating that reminder systems can be employed across all age groups. Further investigation of wider attendance literature (see Chapter 5, Age) indicates that there is largely consistent evidence, across health and geographical settings, that higher rates of non-attendance is related to younger age, although this may not hold for paediatric patients who have greater parental involvement in the health care. However, older people tend to have multiple health problems or conditions that may be less likely to resolve and therefore take up the bulk of outpatient appointments. There is speculation surrounding the reasons why younger patients are at greater risk of not attending; however, there has been little solid exploration of the reasons. Therefore, there is little theoretical basis to inform reminder selection so we contend that there is no reason not to employ reminder systems equitably across all age groups. However, as discussed under Accessibility, there are important considerations to be made in relation to the type of reminder technology that would be most appropriate or acceptable for different age groups. Accessibility is a key issue in selection of reminder technology, and older populations are likely to have lower levels of accessibility to mobile phones and, therefore, SMS than younger people108 (see Accessibility). Patient preferences for reminders are another key issue as it may be a proxy for acceptability, with patients unlikely to prefer reminder systems that they cannot access (see Other reminder characteristics).

Gender

There are no studies indicating that gender has a differential impact on the effectiveness of reminders [see Chapter 4, Evidence statement (F.2–F.11): few studies investigate whether or not a range of distal/proxy individual attributes have a differential effect, over and above pre-existing appointment behaviour, on reminder effectiveness (in terms of attendance, cancellations or rebooking) (evidence category Va or VIIa)], suggesting that reminder systems can be employed regardless of gender (see Chapter 5, Gender). There is conflicting evidence, across health and geographical settings, that gender predicts non-attendance at appointments. Half of the studies searched find that women are more likely to miss appointments and the other half find that women are more likely to attend their appointment. On this basis, we can see no reason for employing reminder systems differentially for male and female patients.

Deprivation

There are no studies indicating that deprivation has a differential impact on the effectiveness of reminders [see Chapter 4, Evidence statement (F.2–F.11): few studies investigate whether or not a range of distal/proxy individual attributes have a differential effect, over and above pre-existing appointment behaviour, on reminder effectiveness (in terms of attendance, cancellations or rebooking) (evidence category Va or VIIa)]. There is weak but consistent evidence that deprivation is a factor in non-attendance (see Chapter 5, Deprivation). This section reveals an array of suggested reason why patients from deprived backgrounds may not attend appointments and reveals a potentially complex set of circumstances that could make it difficult to engage with health-care systems in general, or appointment and reminder systems in particular. Although there has been little exploration of these circumstances in relation to appointment attendance behaviours, these circumstance explain why this group of patients may struggle to engage with complex appointment systems that require a relatively high degree of health literacy;207 they may also not receive appointment information or reminders and subsequently may find it difficult to attend, cancel or reschedule appointments. The available literature suggests that deprived populations face multiple obstacles to health-care appointment attendance. 86,89,123,148,166 Health services providing care to deprived populations have a wide range of complex issues to consider in addition to thinking about strategies for improving attendance. Clinics that achieve high levels of attendance despite serving patients populations with high levels of socioeconomic deprivation would warrant close investigation, as they may have introduced ways of working that could be transferable to other settings. Reminder systems are likely to be indicated although, as discussed above, reminder systems are very dependent on accurate data from patients and so deprived populations with unstable contact details may be disadvantaged. 103 Health services will need to give careful attention to having robust procedures for keeping patient contact details up to date and for providing easy to use systems for patients to make, cancel and reschedule appointments.

Ethnicity

There are no studies indicating that ethnicity has a differential impact on the effectiveness of reminders [see Chapter 4, Evidence statement (F.2–F.11): few studies investigate whether or not a range of distal/proxy individual attributes have a differential effect, over and above pre-existing appointment behaviour, on reminder effectiveness (in terms of attendance, cancellations or rebooking) (evidence category Va or VIIa)]. Our evidence suggests that there may be a link between ethnicity and non-attendance, although this is not completely consistent. On the whole, studies in this review have failed to examine ethnicity and to understand the causal pathways that link ethnicity to non-attendance. Ethnicity is itself an umbrella term that covers multiple potentially relevant and inter-related factors that, by themselves, may be linked to non-attendance, e.g. socioeconomic or deprivation status (see Deprivation), language barriers (see Other reminder characteristics), health literacy (see Deprivation), issues of trust, mutual respect and cultural competency (see How do the perceptions and beliefs of patients, their carers and health professionals regarding specific types of reminder systems, and patient/carer resources and circumstances, influence their effectiveness?). Several factors could interact to create difficulties with patient engagement and therapeutic relationships (see Characteristics of the appointment system). Ethnic groups face multiple potential obstacles to health-care appointment attendance. Health services providing care to ethnic groups have a wide range of complex issues to consider in addition to thinking about wider strategies for improving attendance in this group. Reminder solutions are one possible way forward; however, owing to the likelihood of concurrent deprivation status, similar issues around accuracy of contact details and general accessibility of patients to reminder technologies are likely to exist. In addition, health services need to give consideration to having reminder systems, cancellation procedures and rescheduling procedures that suit the language requirements of their particular ethnic groups and are easy for this group of patients to understand and engage with.

Substance abuse/mental health/comorbidity and physical illness

There are no studies indicating that substance abuse, mental health or comorbidity has a differential impact on the effectiveness of reminders [see Chapter 4, Evidence statement (F.2–F.11): few studies investigate whether or not a range of distal/proxy individual attributes have a differential effect, over and above pre-existing appointment behaviour, on reminder effectiveness (in terms of attendance, cancellations or rebooking) (evidence category Va or VIIa)]. High levels of non-attendance are found in patients who abuse drugs and alcohol (see Chapter 5, Substance abuse), patients with mental health problems (see Chapter 5, Mental health) and patients with comorbidities and illness (see Chapter 5, Physical illness/comorbidity). Although forgetfulness may occur, it is likely that other, more substantial, problems may be acting as the obstacle to attendance. Non-attendance at the appointment is an indicator that there may be a deterioration of the mental health, drug and alcohol problems. In such cases, patients may need active follow-up in order to support them through their relapses and to re-engage them in their treatment programmes. Patients with other co-morbidities or illness may find themselves too ill to attend, too ill to cancel or may even be admitted to hospital. Health services serving large numbers of these patients need to be aware that this group are at high risk of not attending and to be aware that non-attendance may indicate an increased health need and will have a wide range of complex issues to consider in relation to finding wider strategies for improving attendance in this group. Simple reminders and automated reminders to attend may be ignored or overlooked, particularly when patients are experiencing an increased level of illness or substance abuse. In fact, the use of simple reminders may put these patient groups at a disadvantage compared with general outpatient populations. Reminders with direct personal contact might be appropriate in these groups (see Other reminder characteristics, Format of the reminder). To facilitate attendance, a sequential reminder intervention such as that described by Perron et al. 78 could be initiated: first a phone call to either landline or a mobile phone; second, a SMS message if participants do not answer the telephone after three attempts and have a mobile phone; and, finally, a postal reminder if participants do not answer the telephone, have no mobile phone for SMS, or had no phone at all. Such a design, although labour intensive, would reach the maximum number of participants and may increase attendance rates while still being cost-effective. A further consideration is how to re-engage patients with treatment after they have missed their appointment. Intensive approaches, such as ‘stepped reminders’46,195 and patient navigators,196,197 have been effective at increasing attendance at screening and immunisation programmes in disadvantaged and vulnerable populations and might also be effective at re-engaging similar groups of patients who have dropped out of treatment.

Symptomatology and severity

There are no studies indicating that symptoms or severity of a condition have a differential impact on the effectiveness of reminders [see Chapter 4, Evidence statement (F.2–F.11): few studies investigate whether or not a range of distal/proxy individual attributes have a differential effect, over and above pre-existing appointment behaviour, on reminder effectiveness (in terms of attendance, cancellations or rebooking) (evidence category Va or VIIa)]. Our review does suggest that patients who have no symptoms, improved symptoms or low perception about the severity of the condition are more likely to not attend a scheduled appointment (see Chapter 5, Symptomatology and severity). However, there will be patients in this bracket who do not have a life- or function-threatening condition who can manage their health condition satisfactorily for many years and for whom continued attendance is unnecessary. In this case, a reminder system could be used to promote cancellation without rescheduling leading to a negotiated discharge. Conversely, there may be patients or people in this bracket who do have a potentially serious health problem but who are either unaware of its potential severity and impact if it remains untreated or who do not know it because they have not attended a screening appointment. There is a huge amount of literature, which has not been investigated in this review, investigating strategies for increasing uptake of screening appointments. Some of these strategies, including stepped reminders and patient navigation, are labour intensive but might be relevant to consider for re-engaging patients with health-care procedures if they have dropped out (see Substance abuse/mental health/comorbidity and physical illness). The end point of these kinds of processes is either rescheduling of further health-care appointments or a negotiated discharge.

Previous patterns of non-attendance

Previous patterns of attendance are often cited as indicators of future attendance; however, the evidence in this review presents a more unpredictable picture than this (see Chapter 5, Previous patterns of non-attendance). In some studies, patterns of attendance and non-attendance can be variable, with no clear picture emerging of those patients who are ‘attenders’ or those who are ‘non-attenders’. In this review we have already identified that people may not attend for many different reasons and these may be variable over time. For anyone at any time, the appointment can be unsuitable from the outset or can become unsuitable because of a change of circumstances,101 indicating that all patients are likely, at some point, to be unable to attend an appointment. We recommend sending all patients reminders that include a message about cancelling an unwanted appointment and rescheduling for a future appointment if treatment is still required.

Accessibility

There is strong consistent evidence that the reminder may not be received by the patient [see Chapter 4, Evidence statement (B.1): there is strong consistent evidence that the reminder may not be received by the patient (evidence category Ia)]. It is largely intuitive that if the patient does not receive the reminder, this may contribute to suboptimal effectiveness of the reminder. Several reasons for non-receipt of reminders are discussed in the literature. There appears to be a consistent amount of either technology error or human error associated with all reminder systems, and consequently many patients report not receiving a reminder or receiving it late. The reminder systems will be most effective when contact details are accurate and patients receive the reminder. It is therefore important for the health-care organisation to have systems in place to ensure that patient contact details are current and that reminders are successfully received by the patient. 95,208

Timing of the reminder

There is strong evidence that the timing of a reminder, between 1 and 7 days prior to the scheduled appointment, has no effect on patient attendance behaviour [see Chapter 4,Evidence statement (B.5): there is strong consistent evidence that the timing of a reminder, between 1 and 7 days prior to the scheduled appointment, has no effect on patient attendance behaviour (evidence category Ia)]. SMS or telephone reminders are typically sent either the day before or on the day of the health-care appointment. 64,70,71,73,76,78,84,89,100,108 Reminders of this nature appear to be sent to benefit those patients who have genuinely forgotten about their appointment [see Chapter 4, Evidence statement (A.1): there is strong consistent evidence that simple reminders which provide details of timing and location of appointments are effective at helping a (forgetful) patient to attend their appointment (evidence category Ia)]. For this reason, the reminder is not sent too far in advance as this may allow that patient to forget again. 102 It might be supposed that if a reminder mainly serves as a memory prompt for the genuinely forgetful patient, then it is more likely to be effective at increasing attendance behaviour the closer it is to the date of the appointment. Although the evidence from our included studies does support the use of reminders for supporting forgetfulness [see Chapter 4, Evidence statement (A.1): there is strong consistent evidence that simple reminders which provide details of timing and location of appointments are effective at helping a (forgetful) patient to attend their appointment (evidence category Ia)], it further indicates that reminders can be sent any time between 1 and 7 days, with no diminished effectiveness of the reminder at facilitating attendance at appointments. The evidence does not determine whether or not effectiveness of reminders is diminished beyond 7 days, but presumably a 1- to 7-day window is plenty of time in which to operate any kind of reminder system, e.g. SMS, telephone, letter, etc., with any kind of outcome in mind, e.g. attendance, cancellation, rescheduling, information sending, etc.

There are several advantages of sending out earlier reminders. First, there is strong evidence that reminders, particularly telephone reminders, increase cancellation and rescheduling of appointments [see Chapter 4, Evidence statement (E): there is strong consistent evidence that reminder systems will promote cancellation of appointments (evidence category Ia)]. Earlier discussion also suggests that other types of reminders have the potential to increase cancellation and reallocation of appointments to other patients with the commensurate benefits that this can bring (see Are there any systems which effectively support the cancellation of appointments?). Because the timing of appointment reminders makes no appreciable difference to the scale of attendance behaviour, it is sensible to send out reminders that are received by patients a minimum of 2–3 days prior to the appointment in order to take advantage of the efficiencies that can be gained through patient cancellations and health service reallocation of appointments. Second, earlier reminders allow for the patients to be provided with information leaflets about appointments or procedures, orientation information such as maps, etc. that patients can read prior to attending their appointment, which may also help to increase the effectiveness of the reminder [see Chapter 4, Evidence statement (A.2–A.6): there is weak consistent evidence that ‘reminder plus’ are more effective than simple reminders at helping a patient to attend their appointment (evidence category IIIa) and Chapter 6, Reminder plus for more detailed about provision of additional information].

Other reminder characteristics

There is either no evidence, weak evidence or conflicting evidence that a range of other reminder characteristics influence the effectiveness of the reminder. This includes language and framing of information [see Chapter 4, Evidence statement (B.2): there are no studies investigating whether or not reminder factors (such as language, information provided, framing of information) influence the accessibility/comprehensibility of the reminder message for particular patient groups (evidence category VIIa)], content of the reminder [see Chapter 4, Evidence statement (B.3): there are no studies to show that the content of the reminder may effectively address the obstacles/enablers experienced by the patient in question (evidence category VIIa)], personal reminders compared with automated reminders [see Chapter 4, Evidence statement (B.4): there is controvertible evidence, based on multiple good-quality studies, that personal reminders have a greater impact on attendance rates than automated reminders (evidence category IIa)], the format of the reminder [see Chapter 4, Evidence statement (B.4): the format of the reminder will compromise the amount, and type, of information that can be delivered to the patient (self-evident)] and patient reminder preferences [see Chapter 4, Evidence statement (B.4): very few studies investigated whether or not the reminder preferences of the patient may influence the impact a reminder has on the patient's appointment behaviour (evidence category VIIa)].

Administrative errors

There is strong evidence that reminders are not always received by patients [see Chapter 4, Evidence statement (B.1): there is strong consistent evidence that the reminder may not be received by the patient (evidence category Ia)]. In Accessibility we focused primarily on patients’ ability to access reminder technology; however, health-care systems frequently have incorrect or out-of-date contact details for their patients. 95,208 It is largely intuitive that incorrect patient contact details will lead to suboptimal reminder effectiveness and that the reminder system will be most effective when contact details are accurate and patients receive the reminder. This is a greater problem for health services that are dealing with less geographically stable communities, e.g. populations consisting mainly of students, young adults or socioeconomically deprived groups, who may frequently change address, mobile phones or landline telephone numbers. 103 However, this population is also greatly at risk of non-attendance at appointments (see Chapter 5, Demographic and socioeconomic characteristics). Although the use of reminders can be justified in all health services, there appears to be a particular need in health services dealing with geographically unstable populations. Nevertheless, the use of reminders could be considered redundant unless there are also robust processes in place that regularly check to ensure the accuracy of their patient contact details.

Characteristics of the appointment system

Although several characteristics of appointment systems were discussed in detail in Chapter 5, Characteristics of appointment systems, three main issues affecting appointment attendance are discussed because they are potentially modifiable through the use of reminder systems.

Patient–provider alliance and engagement

There is good evidence that the relationship between the patient and the health service may impact on patient attendance at appointments (see Chapter 5, Patient–provider ‘alliance’, communication and ‘engagement’). Patients are less likely to attend appointments when they feel that staff are being judgemental or lack empathy. 101,145 One interesting difference between patients who seek structured health care and those who do not is the apparent need for the support and reassurance that can be offered by health-care professionals. 145 The provision of orientation information or ‘navigational support’, which is information and guidance regarding the way that the health service works and what to expect from treatment/care, is viewed by many patients to be a positive interaction that would encourage attendance at appointments. 145 This has direct implications for utilisation of reminder systems and the simple act of sending a reminder may increase patient perceptions of the value of the appointment and of the interaction with the health service. 80,145 The provision of information through automated systems is an opportunity to build on that (see Reminder plus). It could be argued that the provision of individualised information to patients through direct regular patient contact would be specifically beneficial for some patients, such as those with long-term health problems and more vulnerable patients. 61 The findings presented here would suggest that reminder messages which are gain-framed would be more welcome than those which are loss-framed and that may be perceived as pressurising, stressful or irritating (see Reminder plus for a brief discussion about message framing).

Clinic accessibility

The available evidence regarding the impact of service location and transport on attendance levels is equivocal and suggests clinic accessibility is not simply a function of travelling time or distance to the clinic (see Chapter 5, Service location and transport difficulties). Some of the reported difficulties faced by patients trying to access clinics were difficulties with getting transport, parking, availability of public transport and cost of travelling, and the impact of these factors is likely to vary between service settings and within patient groups. The impact of transport issues is potentially greater for patients who are more socioeconomically deprived because they may be more reliant on public transport, which may be relatively and prohibitively expensive. Transport difficulties may also be more pertinent to patients with young children or for the elderly who may have problems negotiating public transport systems. Many of the issues discussed here require a broader solution to tackle the problems of patient accessibility in order to increase attendance. However, health services should consider whether or not they can include useful information that could aid the patient journey to an appointment within any reminder, e.g. availability and cost of parking, bus routes, etc., that may encourage patients to make the journey (see Reminder plus for information regarding orientation information). In many cases, patients may know in advance that they will not be able to attend their appointment because of transport difficulties and a reminder that is framed to support cancellation of appointments and rescheduling to a more convenient time may reduce a proportion of travel-related non-attendance (see Are there any systems which effectively support the cancellation of appointments?). It is unlikely to make much difference for patients who encounter travel difficulties while they are en route to their appointment.

Accuracy of patients contact details/clerical errors

Reminder systems rely on accurate and stable data from their patients. Our review has highlighted that contact details of patients are frequently not up to date, incorrectly entered into the database or wrong contact details were given by patients [see Chapter 4, Evidence statement (B.1): there is strong consistent evidence that the reminder may not be received by the patient (evidence category Ia)]. In addition, there can often be clerical errors in the sending out of appointments whereby a patient may receive a reminder intended for another patient. This raises possible implications for patient confidentiality, appointment non-attendance and increased cost and time implications of dealing with errors. Health services need to work hard to ensure that they have robust procedures for updating patient records in order to optimise attendance and cancellation/rebooking systems. 103 An effective reminder system with appropriate cancellation and reminder systems will almost certainly increase attendance and utilisation of health-care appointments (see Are there any systems which effectively support the cancellation of appointments?), which will increase the workload of health-care clinicians. 69 Health services will also need to consider the impact on staff that frequently use unattended appointments as an opportunity to catch up on other health-care-related activities (see Are there any systems which effectively support the cancellation of appointments?).

Inequitable access to reminder technologies

With any reminder system there is likely to be a level of unsuccessful contact; however, some patient groups, e.g. deprived populations and homeless groups, may not have equitable access to landline systems or mobile phone technology. Other groups may not answer the phone and may let the phone run to the answerphone, e.g. patients with mental health problems, deafness or those with poor mobility. Our review illustrates that reminder interventions are not being developed specifically to address inequalities in access to services between population subgroups, nor is consideration being given to whether or not interventions may actually serve to increase disparities (so-called intervention-generated inequalities). Henderson48 observes that the impact of implementing initiatives on inequalities in access to services was not considered by any of the studies included in his review. Health services need to bear in mind the sociodemographic status of the population that they are serving and consider issues of accessibility to reminder technologies of all their patients otherwise they run the risk of increasing inequalities in access.

Reminder preferences

Although many patients indicate that they are happy to receive a reminder (see Other reminder characteristics), there is a risk of antipathy from a sizeable percentage of patients who feel negatively or very negatively about reminders. 55 Those who were disturbed by the reminder gave the following reasons: telephone call too early in the morning, they were waiting for other important results, they were contacted at work, felt that were being treated as senile or felt that the reminder was unnecessary. 78 Evaluations of reminders do not typically examine any inherently harmful effects of reminder systems and it is unclear how reminder preference might influence the effectiveness of the reminder [see Chapter 4, Evidence statement (B.4): very few studies investigated whether or not the reminder preferences of the patient may influence the impact a reminder has on the patient's appointment behaviour (evidence category VIIa)]. We have postulated that how comfortable the patient is with the reminder technology used may have an impact on how well the reminder message is ‘received’ and, thus, whether or not the patient chooses to act on it. If the reminder is irritating, demeaning or badly done, then this could lead to a loss of faith or trust in the health-care service and for some patients, the effect of the reminder could be counter-productive. Potential negative impacts of reminders and reminder preferences should be considered in future evaluations.

Ceiling effects

Several studies suggest the presence of a ceiling effect, i.e. when the baseline attendance rates are already high the potential for improvement is correspondingly reduced. 63 In such cases, the use of complex reminders may simply not be required. Recognising that all patients across all health services will occasionally forget or find that the appointment is no longer convenient, it is likely that simple reminders that encourage cancellation of unwanted appointments and rescheduling to a more convenient time may be all that is warranted in health services with an already high attendance rate.

Conceptual framework: implications for practice

We examined existing conceptual frameworks to identify models that explain reasons for FTA appointments. No existing models explored the complexity of the different interactions between the patient, the service, the reminder system and the factors that support or inhibit attendance. Consequently, we developed a conceptual framework exploring FTA. This framework places patients at the centre, to convey their centrality in forgetting or deciding whether or not they will attend (or at least intend to attend). The framework examines patient status when they receive the appointment notification in terms of obstacles or enablers to their attendance, ‘distal/proxy’ attributes that could characterise a patient group to predict the effectiveness of different types of reminder systems and the characteristics of the reminders.

Our study proposes five possible outcomes relating to appointment attendance, cancellation, rebooking and FTA (see Appendix 2):

• outcome 1 – patient attends the original appointment

• outcome 2 – patient does not attend the original appointment but rebooks and attends

• outcome 3 – patient does not attend the original appointment, cancels, but does not rebook and attend

• outcome 4 – patient does not attend the original appointment and does not cancel

• outcome 5 – patient does not attend the original appointment and does not cancel, but is rebooked by the clinic.

The conceptual framework was further developed to propose a series of hypotheses with respect to the interaction between the patient and the reminder system (see Appendix 3). These explore the patient–reminder interaction [A] and the attributes associated with increasing the likelihood of attendance, including reminding the forgetful patient to attend [A.1], providing information that attending the appointment will positively affect his/her well-being [A.2], providing information that non-attendance will negatively affect his/her well-being [A.3], that attendance will not be difficult [A.4], the experience will be positive [A.5], attendance is compatible with other commitments [A.6] and is morally the right thing to do [A.7].

The hypotheses further explored the relationship between the accessibility of the reminder [B] and the outcomes for the patient. In particular, the outcomes are more likely to be detrimental for the service if the reminder is not received [B.1] or understood [B.2] by the patient, does not overcome potential obstacles experienced by the patient [B.3], is not formatted appropriately [B.4] or timed appropriately to allow cancellation or rebooking [B.5].

Several components of the reminder can be varied to impact on the effectiveness of the reminder, specifically, the medium (e.g. telephone, SMS, letter, multiple modalities), timing, content, interactivity and intensity.

Further, we hypothesised that the impact and effectiveness of reminder systems vary by health-care settings [C], including the source of the original appointment [C.1], service location [C.3] and initial or follow-up appointment [C.5].

The study also hypothesised that the wider social systems in which the appointment is situated impacts on the likelihood of attendance [D] and that obstacles to attendance and rebooking impact on attendance [E].

Clinical scenarios: implications for practice

In response to the findings from this study, we have developed a series of clinical scenarios relating to the use of reminder systems (see Appendix 7) that synthesise the evidence presented into this report to address specific scenarios for service providers, outlined below:

1. if you want to optimise the likelihood of the patient attending the appointment

2. if you want to optimise the likelihood of attendance at an initial appointment

3. if you want to increase the rates of attendance at screening appointments

4. if you want to optimise the likelihood of attendance at follow-up appointments

5. if you want to optimise the likelihood of attendance of specific population groups

6. if you want minimise the number of unfilled, non-attendances on any one day (i.e. minimise unused capacity)

7. if it is a high priority that patients cancel then it is important that health-care services make it as easy as possible for patients to cancel

8. if it is a high priority that patients reschedule their unwanted appointment in order to receive priority treatment

9. if you want to make it easy for the patient to cancel or reschedule

10. if the health service has to reschedule an appointment.

Reasons for non-attendance

In addition, we have summarised below the major reasons proposed for missing appointments (see Appendix 8). In particular, we propose solutions to address each of these issues based on the evidence presented in the report.

1. problems with appointment system

2. patient illness

3. patient perceptions and fears regarding the importance or value of appointment

4. timing and incompatibility of the appointment: planned (avoidable)

5. timing and incompatibility issues: unplanned (unavoidable)

6. issues to do with patient: practitioner relationship or interaction

7. issues to do with the accessibility of the appointment

8. patient attributes.

Advantages and disadvantages of different reminder systems

The advantages and disadvantages of the range of reminder systems are presented in Appendix 9, including a discussion of the costs and implications for cancellation and rebooking.

Practice guidelines

In our original proposal, we identified that one of the outcomes of the project would be a stand-alone practice guide based on the information provided in Appendices 7–9. Owing to the complexity of the information, we propose to create a web-based practice guide which will allow health service managers to negotiate the information more readily. A brief outline proposal of this is provided in Appendix 10.

Intensive reminders

We have identified that high levels of non-attendance are found in patients who abuse drugs and alcohol (see Chapter 5, Substance abuse), patient with mental health problems (see Chapter 5, Mental health) and patients with comorbidities and illness (see Chapter 5, Physical illness/comorbidity). For these groups, simple reminders and automated reminders to attend may be ignored or overlooked particularly when experiencing an increased level of illness or substance abuse. In fact, the use of simple reminders may put these patient groups at a disadvantage compared with general outpatient populations. Reminders with direct personal contact might be appropriate in these groups (see Other reminder characteristics, Format of the reminder). To facilitate attendance, a sequential reminder intervention such as that described by Perron et al. 78 could be initiated. First, a phone call to either landline or mobile; second, a SMS if participants do not answer the phone after three attempts and have a mobile phone and, finally, a postal reminder if participants did not answer the phone, had no mobile phone for SMS, or had no phone at all. Such a design, although labour intensive, would reach the maximum number of participants and may increase attendance rates while still being cost-effective. A further consideration is how to re-engage these groups with treatment after they have missed their appointment. Intensive approaches, such as ‘stepped reminders’46,195 and patient navigators,196,197 have been effective at increasing attendance at screening and immunisation programmes in disadvantaged and vulnerable populations, and might also be effective at re-engaging similar groups of patients who have dropped out of treatment.

E-mail reminders

E-mail has been proposed as a suitable medium for sending appointment reminders42 (see Accessibility). Bespoke secure e-mail programmes can incorporate special features such as standard forms guiding the use and content of the e-mail sent and the capacity to show read receipts (in order to confirm the patient has received the correspondence). However, they are costly to set up and may require a greater degree of user skill than standard unsecured mail. 42 E-mail may also be a suitable medium for many patients to cancel appointments that are no longer convenient or no longer required. It is unlikely to be suitable for all patients, but does provide an automated mechanism for quick and easy cancellation. Future research is required to consider the likely usefulness/availability/accessibility and practical implications of e-mail reminders.

Booking/cancellation over the internet

Derby Hospitals NHS Foundation Trust uses online cancellation forms to make it easier for patients who do need to cancel to get in touch with the trust. This addresses difficulties, such as those identified by Martin et al. ,142 in accessing, or indeed providing, staffing for a cancellation telephone line. A Department of Health press release222 indicates that because of projects to reduce missed appointments, the number of missed appointments in Derby between April and December 2011 fell by over 12,000 (2%) compared with the same period the previous year. A 2011 survey of 1700 patients by www.patient.co.uk showed that 85% wanted to book GP appointments online. 223 Nearly half of patients surveyed (47%) were over 55 years of age and, therefore, challenged assumptions about the demographics for users of such technologies. 223

Integrated care including appointment booking

University College London Partners has been working on an internet-based solution that allows a pregnant woman to automatically receive all the information she needs about her forthcoming 12-week scan. Appointments can be made online with the content being tailored to individuals (e.g. according to age group or language spoken). 224

Charging for non-attendance

In their qualitative study, Martin et al. 142 discuss the ramifications of charging for non-attendance. They describe how some patients held the view that that financial penalties for patients who did not attend appointments was justified. If this measure had an effect on attendance then it might seem justified; however, some patients indicated that this would not have prevented the non-attendance in their particular case. They further indicated that, in this context, a fine would not be an intervention to reduce non-attendance but would simply provide a financial reimbursement for the NHS or a penance for patients who did not attend.

They go on to compare this with the USA, where most appointments actually do involve a financial cost to the patient, and non-attendance rates are similar. They report an association with lack of insurance cover and Medicaid for which individuals often have to make some contribution to costs and are, by definition, poor. They conclude, as does Bech,215 that financial penalties would have some effect in reducing non-attendance, but that it would not be substantial.

Resistance to a system of financial penalties came from medical staff who were not only opposed to the system but who emphasised that implementation would prove difficult. 142 A contrary view comes from the Developing Patient Partnerships Survey and Hussain-Gambles et al. 104 qualitative study of primary care team members. However, there was common recognition of the administrative pitfalls and approval of fining was not universal. Martin et al. 142 conclude by proposing that an evaluation of such a policy on health service efficiency, administrative staff burden, costs and patient–clinician relationships would be an important first stage that should precede any implementation of monetary penalties. We were unable to establish whether or not such a pilot has actually taken place.

The effectiveness of charging a fine for non-attendees was been debated in Denmark and the UK in the middle of the last decade, but has been investigated in very few studies. 215 These studies225,226 reveal that a fine will, as most economists expected, reduce the non-attendance rate. The literature of non-attendance discloses ‘a significant need for studies addressing the non-attendance problem applying economic theory and standards of analysis’. 215

Open access systems

A recent innovation, adopted by many primary care clinics, is the advanced or open-access system, credited to Murray and Tantau. 227 In this approach, physicians attempt to accommodate patient requests for appointments on the day they call; this is not to be confused with walk-ins who do not call in advance. Future appointments at a time that is more convenient for the patient are also permitted. Service providers vary available capacity to meet the demand of each day. The ability of a patient to book an appointment on the day he or she calls is no longer a function of his/her medical condition and, in contrast, clinics that do not offer advanced access often employ a triage nurse to assess the urgency of medical need of a caller who requests an appointment without delay. Only callers whose need is deemed urgent are offered a slot reserved for urgent requests.

The impetus for adopting advanced access comes from the desire to make clinic practices more patient focused, to accommodate faster access for patients with urgent needs and to gain competitive advantage. St Levan Surgery in Plymouth won the Health Service Journal (HSJ) Award for a patient access system whereby patients phone the surgery and receptionists record all requests for doctor contact on the computer screens of the GPs. 228 The GP phones the patient back, usually within an hour, to agree a plan. Options may include an appointment at a time convenient to the patient (70% of patients choose to be seen on the day), phone advice, investigations, appointments with other practice staff or directions to other services. Follow-ups are also often carried out by phone. However, the implementation of advanced access systems remains a challenge.

First, providers are limited in their capacity to absorb variations in daily demand. It may not be possible to accommodate all appointment requests on the day they originate. This leads to demand spill over to a future day, limiting the clinics ability to meet the demand of that future day.

Second, true demand for same-day service is not captured by the appointments data because it is difficult to tell whether a patient actually preferred to book an appointment on a future date or (s)he did so because a same-day appointment was not available. Same day services may prove particularly challenging for those managing appointment services who seek to determine approximately how much capacity should be available at the start of each day for the demand.

Third, in many clinics, different physicians’ patient list compositions and sizes are significantly different, with the result that some physicians have fewer available slots to accommodate same-day demand. 228,229

The open-access policy that calls for ‘meeting today’s demand today’ has been demonstrated to be a reasonable choice when the patient load is relatively low. 230

Luo et al. 231 have shown that appointment scheduling policies that ignore interruptions for emergency requests perform quite badly in relation to patient waiting times, especially when the number of appointments to be scheduled is also a decision variable. They also found that policies requiring equally spaced appointments perform reasonably well when the interruption rate is constant; however, their performance worsens significantly when the interruption rate is time dependent.

Choose and Book

An important component of enhancing patient experience of care is to provide more flexibility to patients regarding how, when and where to receive treatment. In pursuit of this objective, the NHS in the UK launched its electronic booking system, Choose and Book, for outpatient appointments in January 2006. 232 However, there is some evidence, mainly anecdotal, to suggest that this system will impair outpatient practice233 and, in particular, patient attendance is worse under a Choose and Book service. 234 Commentators have suggested that part of the problem is that patients are given only one chance for booking with one phone call with the implication that a three contact system, at different times, of the day would prove more successful. 235

Patient-initiated appointments

Patient-initiated appointments offer patients the opportunity to self-manage their disease by requesting specialist reviews at times of need instead of clinician-scheduled appointments. Claimed benefits for patient-initiated care, in comparison with traditional clinician-driven care, include reduction of unnecessary medical reviews, greater satisfaction to patients and staff and maintenance of the patient’s physical and psychological status. 236 In a 2-year RCT, subsequently extended to 6 years, 209 consecutive patients with rheumatoid arthritis were assigned to either a direct access group or a control group receiving regular clinician-initiated care. 236 Patients in the direct access group expressed significantly higher satisfaction and confidence in the system and also had significantly fewer (38%) hospital appointments (median 8 vs. 13 hospital appointments; p < 0.0001) than patients in the control group. This approach is regarded as particularly promising for other chronic diseases.

Systematic overbooking

Each service faces a challenging dilemma, i.e. whether to build in assumptions of ‘non-attendance without opportunity to reassign’ into their appointment schedules (i.e. overbooking) or whether to expect clinical staff to redeploy the freed time (with a consequent risk of underutilisation). Flexibility to redeploy the time to the benefit of other patients will also be limited by factors relating to those patients themselves, i.e. if a patient turns up early (or is invited to turn up early) for their appointment, he or she may be able to occupy a previous slot. However, such patterns then drive future appointment behaviour with being seen early on one occasion being likely to modify expectations for future appointments and, conversely, perceived ‘idle time’ experienced by turning up early but being seen at the appointed time being likely to lead to presentation closer to the scheduled time on subsequent occasions. While service providers such as airlines respond to uneven attendance patterns by offering stand-by opportunities, they risk losing significant good will when overbooking flights. 237 They similarly face a vicious cycle whereby a customer’s previous experience of the convenience of standby may modify their expectations that this option will then be available on a repeat occasion, thereby increasing demand and making such provisions less likely for the particular individual. In an award-winning paper, LaGanga and Lawrence237 show that a simple solution is to shorten the time between appointments relative to the average rate of no-shows. For example, if the average no-show rate is 30% and the average time allotted per visit is 15 minutes, a practice could reduce that 15 minutes by 30% and allow only 10.5 minutes per appointment, resulting in more appointment slots.

Teleconsultation

Newham University Hospital NHS Trust has reportedly started seeing diabetes patients who do not need a physical examination via Skype™ (Skype Ltd, Rives de Clausen, Luxembourg). Results from a pilot suggest that this has caused missed appointments to fall by 11% because of the time saved in travelling, waiting and fewer patients attending A&E. Importantly, feedback from patients suggests that the quality of care was the same as with face-to-face appointments. 222 The project had a significant impact on patients, particularly those who have difficulty accessing care because of their busy lifestyles or multiple commitments and those with limited mobility or who are housebound. For staff, it encouraged more focused consultations and better use of face-to-face time for clinical activities. 238

Patient decision aids

There are also early indications that patient decision aids, given at an initial meeting with a nurse and followed up by a telephone consultation, may reduce the need for face-to-face appointments. A follow-up appointment is required only when the patient is struggling to make a decision. However, the need to respect patient choice means that such provision can only be offered as an alternative and not as a standard default. 239