Enhancing resource utilisation in outpatient clinics: impact of telephone and text reminders on did not attend and short notice cancellation . Authors: HP Priyantha Siriwardana, Frances Kiernan, Abhishek Kapila, Kimberly Ann Cardoso, David Agunede, Mahima Pillai, Russel Emamdee, Sritharan Kadirkamanathan Published: 2026 ABSTRACT AbstractDid not attend (DNA) and short notice cancellation (SNC) in outpatient (OP) clinics waste scarce capacity and delay care. We undertook a quality improvement project in general surgery OP clinics with the aim to reduce DNA below the baseline 12.2% and SNC below the baseline 5% using telephone and text reminders to increase OP appointment (OPA) slot utilisation from around 83% towards 85%–90%.We studied DNA and SNC in four consecutive phases in a general surgery clinic. Phase 1: a retrospective baseline OPA attendance; Phase 2: a prospective diagnostic where non-attenders were telephoned within 3 days to complete a questionnaire on reasons for DNA/SNC; Phase 3: Preappointment telephone reminders 72 hours before OPAs; Phase 4: one-way text reminder.Across four phases, 1500 OPA slots were analysed. Baseline utilisation was 82.8% in phase 1 (360 slots; DNA 12.2%, SNC 5.0%) and 83.3% in phase 2 (420 slots; DNA 11.7%, SNC 5.0%). Questionnaire responses: most non-attendees were forgetting their appointment or unable to get through to the hospital to cancel. Phase 3 (360 slots): 286 patients (79.4%) were successfully contacted, 251 confirmed attendance and 35 wanted to cancel, creating potentially reusable slots; DNA fell to 9.4% and SNC to 0.6%, but only 20 of the 35 freed slots were re-booked due to limited booking staff (utilisation; 85.8%). Phase 4: 360 slots, utilisation was 84.7% (DNA 8.1%, SNC 7.2%). Statistical process control charts suggested a favourable trend in DNA and SNC over time but did not demonstrate consistent special cause variation.Telephone reminders led to fewer SNCs and created opportunities to reallocate appointments, but impact was constrained by limited staff. One-way text reminders appeared to reduce DNA but higher SNC, suggesting that reminder systems alone are insufficient. Sustainable improvement in OPA utilisation requires reliable contact information, simple two-way routes for patients to cancel or rearrange, and adequate administrative capacity.Keywords: Outpatients, Quality improvement, Reminders, Communication WHAT IS ALREADY KNOWN ON THIS TOPIC Patient non-attendance occurs through did not attends (DNAs) or short notice cancellations (SNCs), both representing widespread issues that compromise outpatient appointment (OPA) resource utilisation. Patient non-attendance occurs through did not attends (DNAs) or short notice cancellations (SNCs), both representing widespread issues that compromise outpatient appointment (OPA) resource utilisation. WHAT THIS STUDY ADDS While telephone reminders may decrease non-attendance (DNA and SNC), their efficacy depends on sufficient booking staff availability to maximise OPA time utilisation. Text reminders can reduce DNA, but one-way text reminders may paradoxically increase SNC. While telephone reminders may decrease non-attendance (DNA and SNC), their efficacy depends on sufficient booking staff availability to maximise OPA time utilisation. Text reminders can reduce DNA, but one-way text reminders may paradoxically increase SNC. HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY DNA rates are significantly high among patients unreachable by phone. One-way text reminders alone may prove insufficient for optimising OPA attendance. Further research is warranted to examine the effect of two-way reminder platforms and real-time cancellation tracking to determine whether these systems produce durable improvements. DNA rates are significantly high among patients unreachable by phone. One-way text reminders alone may prove insufficient for optimising OPA attendance. Further research is warranted to examine the effect of two-way reminder platforms and real-time cancellation tracking to determine whether these systems produce durable improvements. Problem Non-attendance at hospital outpatient appointments (OPA) is a persistent challenge in many healthcare systems and is associated with wasted clinical capacity, longer waiting times and delayed care.1 2 Non-attendance in this context includes patients who do not attend without prior warning, commonly documented as did not attend (DNA), while those who cancel at short notice [short notice cancellation (SNC)] leave insufficient time to reallocate the appointment. Mid and South Essex National Health Service (NHS) Foundation Trust provides a wide range of clinical services for approximately 1.2 million people across Essex. Care is provided through three main hospital sites and their affiliated community hospitals, serving a mixed urban and semirural population with varying transport links and travel distances. This project focused on general surgery outpatient (OP) clinics at Broomfield Hospital (an urban hospital site) and its affiliated Braintree Community Hospital (situated in a semiurban area, nine miles from Broomfield hospital, which offers OP services and day case operating theatres but has free parking). At the time of this project, average waiting times for a first general surgery OPA exceeded 34 weeks, meaning non-attendance represented significant capacity loss, with substantial implications for timely access, particularly for urgent surgical referrals. Our primary focus was on studying general surgery OPA attendance rather than examining the entire hospital. A preliminary retrospective study (phase 1 or baseline) for 3 months from September 2023 to November 2023 of a single consultant’s general surgery clinics (360 OPA slots) showed that 17.2% of scheduled OPA slots were not utilised, due to DNA (12.2%) and SNC (5%). When a substantial proportion of such slots are lost to DNA or SNC, accommodating ‘suitable OPAs’ for urgent non-cancer general surgical referrals becomes challenging without either cancelling some already booked non-urgent appointments or deliberately overbooking clinics in anticipation of non-attendance.1 In this context, ‘suitable OPAs’ refers to appointments with appropriate specialty, clinician and timing that align with the urgency and clinical needs of referred patients. Cancelling already booked OPAs is not fair. The current booking systems lack the flexibility to address the core issues of DNA or SNC. Moreover, DNAs and SNCs create a domino-effect of inconvenience, as minimal non-attendance (DNA and SNC) rates have been estimated at 15% meaning considerably more appointments than necessary would need booking to accommodate this unpredictability,1 2 which is poor use of resources. Additionally, overbooking could cause increased waiting times in the clinic when there are minimal DNAs and would lead to breach of the Patients’ Charter.3 Background Non-attendance across the NHS varies considerably. Published reports indicate that of the 103 million NHS OPAs booked in 2021/2022, 7.6% resulted in a ‘DNA’ outcome, equivalent to approximately 650 000 monthly appointments4 (which includes general practitioner (GP) appointments), though this figure excludes OPA underutilisation due to patient-initiated SNCs. Non-attendance rate had been reported as approximately 15%2 5 and 12% across different studies.6 7 Published estimates suggest that each missed hospital OPA costs the NHS £165 wasted clinical time and additional administrative burden, totalling approximately £1.2 billion annually.8 9 Beyond the financial impact on the NHS, these missed or cancelled appointments carry significant clinical consequences, affecting both non-attending patient and others awaiting timely care access. This compromises efficient clinical resource utilisation, leading to delays in diagnosis and treatment, adversely affecting patient outcomes and reducing overall system productivity.10 Understanding the root causes of non-attendance is essential for developing effective and targeted interventions.11–14 Research identifies multiple contributing factors, including forgetfulness, transportation difficulties, inconvenient scheduling, and patients experiencing symptom improvement prior to their scheduled visit.13 14 Additionally, socio-demographic variables including age, gender and socioeconomic status have been associated with attendance patterns.14 The Model for Improvement, Plan-Do-Study-Action (PDSA) cycles are widely used as a pragmatic quality improvement methodology for testing changes in real clinical environments, learning from outcomes, and refining interventions iteratively15 16 with measures for improvement.17 This quality improvement project in general surgery OP clinics aims to characterise patterns of non-attendance and to use telephone and text reminders to reduce DNA and SNC below baseline to improve OPA utilisation from approximately 83% towards 85%–90%. The project team comprised of surgical doctors ranging from Foundation Year trainees through Clinical Fellows to Consultant. Measurement For this quality improvement project, we used a family of measurements that were classified as baseline measures, outcome measures, process measures, balancing measures and additional measures. Baseline measures Baseline measures were obtained through a retrospective study (Phase 1) examining DNAs, SNCs and OPA slot utilisation for 3 months of one consultant’s general surgical clinic from September 2023 to November 2023 (360 patients) DNA was defined as not attending the OPA without successfully contacting the hospital in advance to cancel. We defined SNC as any OPA cancelled by patients less than 72 hours before the scheduled time, excluding cancellations agreed during pre-appointment telephone reminders in phase 3. Baseline rates were 12.2% (DNA), 5% (SNC), 5% and 82.8% of OPA slot utilisation. Outcome measures DNA rate, SNC rate and overall OPA slot utilisation rate with variation over the time period on SPC (Statistical Process Chart). The outcome categories were ‘attended’, ‘DNA’, ‘cancellation but re-booked’ (used), ‘cancellation and not utilised’ (SNC with vacant slot). DNA and SNC rates were calculated as total DNAs or SNCs divided by the total number of appointments. True appointment utilisation rates, calculated as the sum of attended appointments including those that were successfully rebooked following timely appointment cancellations divided by the total number of appointments. Phase 1 study measuring baseline provided limited insight into the underlying reasons for appointment cancellations, particularly SNC. Obtaining comprehensive understanding of missed patient appointment causation yields practical benefits, enhancing the safety, coherence and quality of healthcare delivery.17 Therefore, the project team proposed to contact all DNA and SNC patients by telephone to identify underlying reasons for non-attendance or cancellation, leading to the decision to conduct a prospective study (diagnostic phase or phase 2) over 3 months from December 2023 to March 2024. In phase 2, we studied 420 patients (appointment slots) that included 49 (11.7%) DNAs, 21 (5%) SNCs and 350 (83.3%) OPA slot utilisation, comparable to baseline. Detailed results are shown in online supplemental table 1. 70 non-attendees (49 DNA and 21 SNC patients) were contacted by the project team via telephone within three working days following their OP clinic appointment to complete a questionnaire survey proforma (online supplemental appendix 1) exploring various factors potentially associated with non-attendance. Telephone calls were made by a Clinical Fellow and Foundation Year 2 doctor using hospital landlines to contact patients between 9.00 and 17.00 on weekdays. Contact details available on the electronic health record (EHR) and referral letters were used to reach patients (mobile telephone number and/or landline number). Verbal consent was obtained from patients at the beginning of telephone calls, explaining the call’s purpose and the study. Up to three attempts per telephone number (mobile and landline) were made at different times to reach patients. A standardised script (online supplemental appendix 2) was used during telephone calls, with each caller taking an average 8 min to complete the questionnaire. Process measures Included proportion of patients successfully contacted by telephone in phase 2 (with reasons for DNA or SNC) and telephone reminder stage (phase 3), respectively, number of contact attempts required, and proportion of freed slots communicated to the OPA booking team. Balancing measures Administrative staff workload, patient contactability rates as a proxy for equity of access, and SNC rate in phase 4 relative to phase 3, which was monitored for unintended consequences of one-way text reminders. Additional measures Included demographic data (patient age and sex) and appointment-related data such as day of the week, session timing (morning or afternoon), appointment time, appointment type (new or follow-up), mode of delivery (face-to-face or telephone), booking method (Choose and Book (C&B) or traditional), and time interval between booking date and appointment date. In phase 2, details of DNA and cancellation including documented comments, non-attendee characteristics, stated reasons and cancellation timing were recorded. Questionnaire survey results All 21 SNC patients were successfully contacted but only 45 DNA patients could be contacted to complete the questionnaire survey, yielding an overall response rate of 94% (66 out of 70). Five of 45 contactable DNA patients were unaware of their appointment. The remaining 40 DNA patients and 21 SNC patients provided various reasons for non-attendance or cancellation as shown in the Pareto chart (figure 1). 10 (25%) of DNA patients and four (19%) of SNC patients had simply forgotten about their appointment. 12 (30%) of contactable DNA patients reported attempting to contact the hospital to cancel but were unsuccessful. Non-attenders expressed frustration with phrases such as ‘Had to ring multiple times to change or cancel’, Horrible—holding on 10 min, 30 min, 45 min trying to cancel’, ‘not enough credit (in the phone) to call to cancel’. Some of the responders said that they preferred an email communication about appointments while others stated preference for telephone calls or text reminders. Nine (20%) out of 45 contactable DNA patients no longer required the appointment as three patients had already undergone surgery elsewhere, five were pursuing private healthcare, and one patient’s problem had resolved while awaiting the appointment. Four ‘C&B’ patients were among them. Pareto chart—reasons for DNA and SNC. DNA, did not attend; OPA, outpatient appointment; SNC, short notice cancellation. Design The study was conducted in Broomfield Hospital and its affiliated Braintree Community Hospital within the general surgical OP clinics of a senior clinician. The results of phases 1 and 2 were presented by the project team at the local governance meeting, and recommendations to reduce DNA and SNC with improving OPA slot utilisation were discussed prior to designing phase 3 interventions. Suggestions for improvements were; Contact patients booked to the OP clinic by sending a telephone text reminder. Contact patients booked to the OP clinic by sending a telephone text reminder. Telephone the patient in advance as a reminder as well as to find out if they are attending or not; cancel the OPA if necessary and use the slot for a different patient. Telephone the patient in advance as a reminder as well as to find out if they are attending or not; cancel the OPA if necessary and use the slot for a different patient. Telephone patients in advance but cancel the appointment if unable to contact anyone after three attempts at different times. Telephone patients in advance but cancel the appointment if unable to contact anyone after three attempts at different times. Implement changes to ways of contacting patients for OPA such as email. Implement changes to ways of contacting patients for OPA such as email. Implement changes to facilities for patients to contact hospital such as dedicated telephone numbers or an email. Implement changes to facilities for patients to contact hospital such as dedicated telephone numbers or an email. Implement a partial booking system where OPAs are booked provisionally, and the patients are advised to contact hospital to confirm the appointment. Implement a partial booking system where OPAs are booked provisionally, and the patients are advised to contact hospital to confirm the appointment. Ultimately, it was agreed that the Trust would commence sending pre-appointment text reminders. In the interim, the proposed communication method was telephone reminders at least 72 hours prior to OPAs (phase 3) and follow-up cycle (phase 4) following text reminders once the Trust implemented the text reminder system. The project team designed the intervention for phase 3 (telephone reminders) that was conducted over 3 months from August 2024 to October 2024, with data collected prospectively. Telephone call reminders were made at least 72 hours before appointments using hospital landlines between 09.00 and 17.00 on weekdays, employing a standardised script (online supplemental appendix 3) for reproducibility. Up to three contact attempts per number (mobile or landline) were made at different times to reach patients. No action was taken if any patient remained unreachable after three attempts. When contacted via telephone reminder, patients were asked if they planned to attend. If any patient was unable to attend, the project team member would arrange rescheduling of the appointment for that patient and identify a different patient from the OPA waiting list to be booked into that slot by emailing the appointment booking staff, flagging the email as ‘high priority’. Phase 4 was conducted when the Trust introduced automated telephone text reminders, with data collected prospectively from January 2025 to March 2025. The project team discontinued telephone reminders in phase 4 to assess the effect on telephone text reminders alone on OP clinic attendance. The text reminders were sent to patients both 5–7 days, and one day prior to the appointment. Patients were not involved in the design of this quality improvement project but DNA and SNC patients in phase 2 contributed indirectly through responding to brief post-appointment telephone questionnaires exploring reasons for non-attendance. The Trust’s EHR system was used to identify the study population, and to apply clinic codes to obtain clinic list of patients booked to each clinic and facilitate data collection. All data were anonymised and stored according to NHS data governance guidelines. Descriptive data were presented as absolute numbers and percentages for categorical variables, while continuous variables were presented as mean (SD) and/or median (range) unless otherwise noted. SPC of DNA, SNC and OPA utilisation (outcome measures) were used to monitor the stability and performance of the quality improvement process over time. Statistical analyses were performed using the R statistical software for the evaluation of additional measures. Categorical variables were compared using Pearson χ2 tests with Yates continuity correction for 2×2 tables. Continuous variables were compared using Wilcoxon rank sum tests. The statistical significance was set to p<0.05. Strategy We adopted Plan-Do-Study-Act (PDSA) quality improvement methodology in four phases with Iterative Learning. Our SMART (Specific-Measurable-Achievable-Relevant-Time based) aim was to reduce the OPA DNA rate from baseline of 12.2% and SNC rate from baseline of 5% and thereby increase OPA slot utilisation from 82.8% to ≥85% in one consultant’s general surgery clinics at Broomfield Hospital and Braintree Community Hospital by March 2025, through implementation of telephone and text reminder interventions. The initial aim of phase 1 was to obtain an estimate of baseline OPA utilisation with any identifiable pattern and reasons for DNA and cancellation. The key learning from phase 1 was that routine EHR data provided insufficient insight into the reasons for non-attendance, particularly for SNC. This gap directly informed the decision to conduct phase 2 as a prospective diagnostic study, in which all non-attendees would be contacted by telephone within three working days following their OP clinic appointment to complete a questionnaire survey proforma (online supplemental appendix 1) on various factors that could be associated with non-attendance (as described in measurements). The questionnaire findings from phase 2 revealed two actionable system failures: patients forgetting appointments (25% of DNA patients) and patients being unable to get through to the hospital to cancel (30% of contactable DNA patients). These findings directly shaped the design of the phase 3 intervention—proactive 72-hour telephone reminders that simultaneously served as a reminder and an accessible cancellation channel. Phase 2 also identified inaccurate contact information as a systemic barrier, leading us to supplement EHR data with referral letter details during phase 3 calls. In phase 3, we planned a structured intervention involving pre-appointment telephone call reminders to reach our target of improvement of OPA utilisation by reducing DNA and SNC. We attempted to contact each patient booked to the clinic at least 72 hours prior to the scheduled OPA to ascertain if they were attending. If any patient was unable to attend, the plan was made to cancel the OPA and use the slot for an urgent patient on the OPA waiting list. Telephone calls were made using hospital landlines similar to phase 2. Phase 3 revealed two system constraints limiting impact: 20.6% of patients remained unreachable despite three contact attempts, and only 20 of 35 freed slots were reallocated due to insufficient booking staff capacity. These findings informed phase 4, which tested the Trust’s newly implemented automated text reminders as a lower-resource alternative. In phase 4, we used preappointment text reminders to improve OPA utilisation to reach our target. Telephone reminders were deliberately discontinued in phase 4 to independently assess the effect of text reminders alone. Appointment booking staff were advised (following learning points from phase 3) to allocate a dedicated person to answer phone calls or emails. Data were collected prospectively similar to phase 3. Results Population characteristics A total of 1500 patients (OPA slots) were analysed across four phases (September 2023–March 2025). The median patient age was 61 years (range 18–92) (IQR 41–68), and 881 (59%) were female. Most patients (63%) lived within 15 km (9 miles) of the hospital; 652 (44%) were new referrals and 848 (56%) follow-up appointments. There were no significant demographic differences between phases. Population characteristics in each phase are shown in online supplemental table 1. Baseline (phase 1) The baseline measurement included 360 OPA slots (30 clinics with 12 slots per clinic), of which 298 (82.8%) were attended, 44 (12.2%) were DNAs, and 18 (5%) were SNCs leaving no time for utilisation of those 18 slots. Of 298 used OPA slots, 241 patients attended without changing the original booking while 57 patients had their appointment cancelled in advance (35 by patient and 22 by hospital) and different patients were booked to the respective slots. On EHR, data were very sparse regarding reasons for cancellation. Diagnostic (phase 2) The diagnostic study included 420 OPA slots (35 clinics with 12 slots per clinic) with a DNA rate (11.7%), SNC rate (5%), and utilisation rate (83.3%) were similar to phase 1. Please see online supplemental table 1. Reasons for DNA and SNC are displayed in figure 1. Phase 3 (following telephone reminders) There were 360 OPA slots, however, it was only possible to contact 286 (79.4%) patients prior to their OPA at the telephone reminder stage, while the remaining 74 were non-contactable, despite attempting at least three telephone calls at different times. None of the 74 non-contactable patients had contacted the hospital following our attempts to call them, but 40 of these patients did attend. 34 (9.4%) patients DNA. Of the 286 patients contacted, 251 confirmed attendances at OP clinic and 12 patients wanted to change the face-to-face appointment to a telephone appointment. However, 2 (out of 251) patients, despite confirming attendance at telephone reminder, cancelled OPA “last-minute” (SNC). The remaining 35 patients, at the telephone reminder stage, requested to cancel OPA and rebook on a different day. This was conveyed to the hospital appointment booking team, advising them to use those slots for patients waiting for urgent OPAs. However, only 20 slots were used, leaving 15 slots unused due to lack of appointment booking staff. This resulted in actual utilisation of 85.8% (309 out of 360) OPA slots (see figure 2). Phase 3—contactability of patients at telephone reminder and attendance at outpatient clinics. DNA, did not attend; GP, general practitioner; OPA, outpatient appointment; SNC, short notice cancellation. Phase 4 (following text reminders) There were 360 OPA slots (30 OP clinics), and each patient was sent a preappointment unidirectional text reminder both 5–7 days, and one day prior to their OPA. The total OPA slot utilisation was 305 (84.7%) of 360 with 29 (8.1%) DNA and 26 (7.2%) SNC. The results are displayed in online supplemental table 1. Data analysis/SPC/time-series summary SPC charts of monthly analysis for DNA, SNC and OPA utilisation across the four phases from August 2023 to March 2025 (figure 3) showed modest downward trends in DNA rates and a marked reduction in SNC during phase 3, but no sustained run or shift meeting standard rules required for special-cause variation. This indicates the interventions did not produce a detectable, stable shift in system behaviour—a distinction important in QI, where progressive stabilisation without special-cause signals reflects improvement in process consistency rather than a step-change in performance. Statistical process charts on DNA, SNA and OPA utilisation. DNA, did not attend; OPA, outpatient appointment; SNC, short notice cancellation. CL, Center Line; LCL, Lower Control Limit; UCL, Upper Control Limit A temporary rise in SNCs during winter months (phase 2) corresponded with local influenza activity affecting patients and staff sickness, which may have affected clinic scheduling. Overall, the patterns suggested progressive stabilisation of attendance. There was a substantial 34% relative reduction of DNA rates from the preintervention (phases 1 and 2) to the final phase (phase 4). The SNC rate significantly reduced from 5% in preintervention (phases 1 and 2) to 0.6% in phase 3 following telephone reminders (χ²=11.571; p<0.001). This represents an 88% relative reduction with a difference in proportions of 0.044 (95% CI 0.019 to 0.069). However, in phase 4, SNC increased to 7.2% following text reminders alone. Non-attendance of patients who could be contacted at telephone reminder was significantly lower than that of patients who were not contactable (χ²=118.3; p<0.001). Patient age demonstrated a significant association with DNA status using both categorical and continuous analysis. Pearson’s χ2 test on age groups (<30 yo, 30–49 yo, 50–69 yo, 70+ years) yielded χ²=16.545 with 4 DoF and p=0.001, providing strong evidence against the independence of age category and DNA status. DNA rates decreased across most age groups in Phase 3 following telephone reminders. The 30–49-year age group showed the largest improvement, with DNA rates falling from 17.5% to 6.1%. However, younger patients (<30 years) showed a slight increase in DNA rates (30.0% to 33.3%). Several demographic and scheduling variables were examined for associations with DNA rates, with most showing non-significant relationships. Gender showed no significant association with DNA status using Pearson’s χ2 test with Yates’ continuity correction (χ²=2.734, p=0.098). Most DNAs occurred in face-to-face appointments, with telephone-based OPAs representing a minority. Appointment type (new vs follow-up) approach did not achieve statistical significance (χ²=3.695, p=0.055), suggesting that this aspect may warrant further investigation despite falling short of conventional significance levels. Clinical site location (main hospital or community hospital) showed no meaningful association with DNA rates (χ²=0.887, p=0.346), indicating consistent performance across sites. Day-of-the-week analysis using Fisher’s exact testing showed no significant weekly patterns in non-attendance. Likewise, the time of day of appointments did not show significant associations using Pearson’s χ2 testing with Yates continuity correction (χ²=0.723, p=0.395) although there were more DNAs at morning OP clinics than afternoon OP clinics. There was no correlation between OPA time and DNA. Time interval between the booking and OPA has no correlation with non-attendance. There was no significant difference in DNA rates between patients booked via the ‘C&B’ and traditional methods in phase 1 and phase 2 (preintervention). However, in phase 4 after text reminders, there were no DNAs among 33 C&B patients; in fact, 7 out of 33 C&B patients had cancelled their OPA in advance and rescheduled following text reminders. There were no C&B patients in phase 3 (online supplemental table 1). Unexpected problems or challenges encountered during the study Several operational barriers were identified. Approximately one-fifth of DNA patients in the diagnostic phase (phase 2) and non-contactable patients in phase 3 lacked up-to-date contact details as noted while trying to contact them for the questionnaire survey and telephone reminder. This would reduce the effectiveness of both telephone and text reminders. In addition, limited booking staff availability prevented real-time reallocation of cancelled slots created at telephone reminders in phase 3, reducing potential gains in utilisation. No patient safety incidents or confidentiality breaches occurred during the reminder implementation. Lessons and limitations Summary of main findings and comparison with prior evidence/literature The baseline DNA rates in our surgical clinic at the main hospital and the community hospital in this study were found within the range in previous reports.2–7 True appointment utilisation rates showed modest improvement in phases 3 and 4 (postintervention). Our data support previous findings that such interventions have only a modest impact18 and telephone reminders have positively influenced non-attendance in line with previous research that identifies communication as a key factor in reducing non-attendance.19 20 Our findings are consistent with previous improvement projects showing that appointment reminder systems are effective but not optimal. For example, a large NHS meta-analysis by McLean et al5 found that SMS reminders typically reduce DNAs by 20%–25%, although results vary depending on contact accuracy and message interactivity such as two-way communication. Phase 3 of our study, telephone reminders provided two-way communication, but phase 4 text reminders were just one-way. In our study, there was a downward trend in DNA rate over the time following preappointment telephone or text reminders. This was reflected in what non-attendees stated in the questionnaire survey, where they indicated that a telephone or text reminder would make them more likely to attend. The survey revealed that patients simply forgetting their appointment was the most common reason for non-attendance (figure 1). A telephone or text reminder would have a positive impact on such patients’ non-attendance. Some of the patients did not need the OPA (eg, already had surgery, going or gone private, and problems solved while waiting) while some patients needed to cancel the appointment at the last minute due to various reasons (eg, transport issues, childcare concerns, location or time not convenient). Such cases could have been identified at a telephone reminder, and the hospital could have used those OPA slots for some other patients. Telephone reminders are hampered by lack of accurate details and the time of telephone calls. We attempted to find out up-to-date contact details by reviewing referral letters and contacting GP administration staff when telephone contact details were not available in referral letters or EHR. If all patients had been contactable, telephone reminders may have had a significant impact and therefore it is essential to have accurate contact details if we were to implement a policy of telephone reminders. We suggest that referrals be returned to the referrer if patient contact details are not available on the referral letter. The suggestion at the governance meeting to cancel OPA if a patient is not contactable at telephone reminders is not advisable as more than half of such patients did attend OPA as noted in phase 3. We did not attempt to contact patients out of hours but a telephone call after 18:00 may be answered; however, it involves employing someone after hours to contact patients. It is evident from the questionnaire survey in phase 2 that some patients were not aware of the missed appointment, consistent with Sung et al21 and many patients attempted to contact the hospital to cancel their OPA but were unsuccessful. Hospital Trusts must address this issue ensuring accurate contact details for patients and providing a dedicated phone number or email address mentioned in the appointment letter for patients to contact the hospital should they wish to cancel. Hospitals must ensure regular checking of their voicemail and/or emails, and if a message is received by a team member regarding a cancellation, it should be passed to the out-patient booking team without delay. These simple measures are key to reducing non-utilisation of cancelled OPA slots, but there should be adequate OP booking team members to act by booking another patient to use the vacant slots. Lack of OPA booking staff resulted in 15 (4.2%) of 360 preventable non-utilisation of OPA slots in phase 3, where potential utilisation would have been 324 (90%) of 360 following 34 (9.4%) DNA and 2 (0.6%) SNC. Increased SNC rate in phase 4 probably suggests that some patients may have forgotten about their appointment but found they were unable to attend following text reminder. Alternatively, they may have realised they no longer needed the appointment after receiving text reminders and would have contacted the hospital to cancel. A two-way text reminder would have helped those patients. The exact number of such patients is not known as we did not conduct any questionnaire survey of non-attendees in phases 3 and 4. In qualitative research, reasons for non-attendance are categorised into two main types: patient-centred and hospital-specific. Among patient-centred reasons, forgetfulness is a primary reason, similar to our study, that is, frequently highlighted in various international studies,12–14 although the complexity of non-attendance suggests that other social and cultural factors may also be influential.13 As for age comparison, in our study, the less than 30-year-old group was consistently the worst for not attending. This compares with already-published literature which illustrates that younger patients are more prone to missing appointments, potentially due to external commitments or better health status.12–14 This suggests that intervention for improvement may be focused on younger patients. In some studies, gender was identified as a significant factor in non-attendance, with male patients more likely to miss appointments13 14 whereas the female-to-male ratio among DNAs in our study showed no statistical significance. The inclusion of telephone reminder before appointments appears to have helped, particularly among male patients in our study. This supports studies by Sawyer et al,20 who found that personalised communication methods improved engagement in OP settings. Sharma et al22 reported an appointment notice of more than 2 months resulted in a significantly higher proportion of patients who did not attend (DNA) their appointment than for those who did, as this possibly led to patients forgetting about their appointments. However, waiting times was not an important factor in non-attendance in our study supporting the study by Zykienė and Kalibatas.23 Automated text reminders from no-reply numbers are just one-way communication, unlike telephone reminders. An alternative to telephone reminders is to send a text message with an automated system at least 5 days prior to the appointment asking to respond ‘Yes’ or ‘No’ about attendance. However, this is complicated by the high prevalence of phone-text scams and is likely to result in hesitancy for some patients to reply. Previous studies have found that reminder systems, whether via phone or SMS, significantly improve clinic attendance rates.19 20 While telephone reminders improved attendance, 21% of patients were unreachable, potentially excluding those without reliable phone access (eg, older adults or low-income groups). More recent studies24 have shown that multimodal reminders (telephone plus text) outperform single interventions. However, cost and staffing constraints often limit adoption in busy NHS services. Unlike controlled trials, our work reflects real-world constraints within a hospital, where resource and workflow limitations can dilute potential improvement effects. Henderson25 reported providing information about the clinic appears more effective than standard reminders especially for those patients who have multiple problems attending numerous OPAs of various specialties. Although this was not assessed in our study, authors noted in phase 4 that patients were not sure which text they received for which appointment. The same applied to the OPA letter that patient had received. Some studies have revealed that high DNA rates can be addressed in part by offering patients more choice around their appointments.26 Jenkins27 reported the importance of flexibility and opt-in initiatives with patient choice in reducing DNA rates, but in our study, there was no difference in non-attendance between ‘C&B’ to traditional booking. Also, there are significant gaps in the literature regarding the precise impact of opt-in procedures. Partial booking system where patients must contact the hospital to confirm the appointment could reduce DNA rate but requires a dedicated phone number or an email address for patients to contact the hospital and adequate staff available to retrieve messages and act on them accordingly. Implementing changes to contact patients via email could save postage fees accrued by sending appointment letters and it is essential that the hospitals have accurate up to date email addresses of patients incorporated into the system. There may be some patients who do not have an email address. However, providing a dedicated email address would benefit most patients in contacting the hospital should they wish to cancel appointments; this in turn could have a positive effect on DNA rate, provided a dedicated team of hospital staff is allocated to check the emails and act on them accordingly. Imposing financial penalties for DNA patients had been proposed by some authors in previous studies22; however, this raises similar administrative problems discussed above, as some patients may have unsuccessfully tried to cancel the OPA and would be unjustly charged. Others2 had suggested charging a nominal fee for OPAs that could be refunded once attended. However, a deposit-based system would risk discouraging patients from lower income groups from booking or attending OPAs in the first place, as even a refundable charge may represent a meaningful financial barrier. This could exacerbate existing health inequalities by disproportionately reducing access among already-underserved groups and would unfairly penalise patients who attempted but were unable to cancel—a situation our phase 2 data show affected 30% of contactable DNA patients. Consistency in OPA non-attendance patterns of patients across phases suggests that communication and administrative strategies can play a vital role in addressing systemic inefficiencies.18 Hospitals should audit patient communication preferences at registration. Future interventions could integrate SMS/email used by 78% of UK adults.28 Telephone reminders require approximately 2 full-time equivalent staff hours per clinic. If scaled, automated SMS systems with two-way communication could offer a more sustainable alternative while reducing administrative burden.8 These improvements not only enhance service efficiency but also reduce economic waste, echoing findings from broader NHS analyses.9 System-level interpretation/PDSA reflection On reflection, non-attendance is complex, and to secure maximum attendance, targeted interventions are required by healthcare facilities to ensure patient’s needs are met. This quality improvement project allowed for iterative testing and reflection. The diagnostic phase (phase 2) identified the need for a reminder process and accurate contact data. The telephone reminder intervention (phase 3) tested that change and revealed administrative bottlenecks. The follow-up phase (phase 4) tested automation (automated text reminders) but exposed the limitation of one-way communication. Each phase thus generated practical learning about the interaction between patient behaviour, administrative processes and system design. This demonstrated that improvements require changes among people, processes and technology, rather than focusing solely on patient behaviour. Considerable effort was invested in telephone calls and reminders, yet the achievable improvement was modest. The modest overall improvement observed in this project reflects not a failure of the interventions per se; rather, it reflects the constraining effect of three upstream system failures—reliability of contact details, administrative capacity, lack of two-way communication—that prevented the process from operating as intended. First, inaccurate or absent contact information—affecting approximately one-fifth of patients—prevented reminders from reaching those most likely to be at risk of non-attendance. Second, inadequate booking staff capacity meant that only 57% (20 of 35) of slots freed by Phase 3 telephone cancellations were reallocated to waiting patients; had all been filled, utilisation would have approached 90%. Third, the one-way nature of the automated text reminders in Phase 4 prompted patients to recognise they could not attend, but provided no embedded mechanism for cancellation, likely driving the rise in SNC from 0.6% to 7.2%. These findings reframe the problem: sustainable improvement in OPA utilisation is not primarily a reminder design problem, but a system design problem requiring reliable contact data, two-way communication channels, and adequate administrative capacity to act on freed slots in real time. Limitations Several limitations should be acknowledged. First, data were collected from a single specialty and within one Trust, which may limit generalisability. Second, although SPC charts illustrated favourable trends, the small number of data points and short duration of each phase limited our ability to demonstrate special-cause variation. Third, patient questionnaires relied on self-reporting, which may underrepresent individuals who could not be contacted. Fourth, operational data such as staff time and reminder costs were not measured, preventing cost-effectiveness analysis. Lastly, due to limited capacity, freed appointment slots were not always reappointed, meaning the observed improvement reflects both behavioural and operational influences rather than purely patient factors. Sustainability and next steps For sustained impact, reminder interventions must be integrated into routine workflows and supported by adequate infrastructure. Automating reminders can reduce staff burden but effectiveness depends on reliable patient contact data and mechanisms for two-way communication, such as text replies or web portals. Embedding reminders within electronic booking systems could allow automatic slot reallocation when patients cancel, maximising resource use. Engaging administrative and clinical staff early in design is essential to ensure processes are feasible within existing capacity. Future PDSA cycles could test two-way reminder platforms and real-time cancellation tracking to determine whether these systems produce durable improvements. Conclusions Reminder systems are valuable tools but cannot independently resolve OP non-attendance, a finding underscored by this project’s modest overall improvement, despite an 88% reduction in SNCs from telephone reminders. The limiting factors were not the interventions themselves, but three upstream system failures: inaccurate or absent contact information, which prevented reminders from reaching roughly one-fifth of patients; inadequate booking staff capacity, which meant only 57% of freed slots were reallocated when full utilisation could have approached 90%; and the one-way nature of automated text reminders, which prompted patients to recognise they could not attend but offered no embedded cancellation mechanism, contributing to a rise in SNC from 0.6% to 7.2%. Together, these findings demonstrate that sustainable improvement in OPA utilisation is not primarily a reminder-design problem but a system-design problem requiring reliable contact data, accessible cancellation routes, bidirectional communication channels, and real-time administrative capacity to act on freed slots. Without addressing these operational dependencies, even effective reminders yield limited gains. These insights are transferable across NHS OP services seeking efficiency within existing resources, and future work should evaluate integrated digital platforms enabling two-way communication and automated rebooking to achieve meaningful, scalable impact. Supplementary material