This the second of three posts based on a presentation I did for noon Oncology Rounds.
1) The Background Article: Why do Patients Miss Appointments?
2) The Practical Article: 15 Ways to Reduce Missed Appointments.
3) Advise from our Lawyers: implement a color coded system to track missed appointments
SERIES SUMMARY: Patients who miss appointments have worse outcomes. Missed appointments can be predicted and reduced with deliberate interventions.
IV. 15 Ways to Reduce Missed Appointments
Summary
A) Improve Communication
1. Keep accurate records of patient contact information.
2. Make it easy for patients to contact the clinic
3. Get the appointment into your patient's calendars
4. Remind patients of their appointment
5. Network hospital databases
B) Remove Barriers to Attendance
6. ** Identify patients who are at high risk for not showing **
7. Use patient navigators for high risk patients
8. Involve Social Work
9. Make clinics easier to attend.
10. Educate your patients or use a carrot
C) Optimize the Clinic
11. Predict aggragate no-show rates
12. Use a computerized system and real time check in times
13. Have a standby list
14. Have adult-pediatric transition clinics
15. **Track and follow up missed appointments **
As this list was compiled using Oncology literature it misses the contribution of interventions researched in other specific populations such as heart failure, HIV, rural, and general practice.
A) Improve Communication
1. Keep accurate records of patient contact information.
A remarkable number of patients were uncontactable in follow up studies. Studies were frequently unable to contact at least 25% of their patients. In some studies they were even unable to contact a next of kin in 50% of cases.
2. Make it easy for patients to contact the clinic
Some patients do not even have the phone number of the clinic. In other cases, the clinic makes it very difficult for patient to contact the office. Offices have long answer machine messages or put patients on hold for indefinite periods of time. Patients may put up with this when they call 'requiring' something from the clinic. However, if they are calling with the curtesy to inform staff they are unable to attend, or will be late for an appointment, the willpower wait on a busy line is very limited.
Clinics should implement multiple channels for communication: phone, text, email, and web chat portal. All these methods can be implemented easily, affordably, and securely. Do away with keeping patients on the telephone on hold by implementing an automated "call-back service" that allows patients to leave their phone number and provides them with an estimated time frame for when the office will call back.
3. Get the appointment into your patient's calendars
Clinic staff are frustrated when patients say "they never knew they had an appointment" or "they thought it was next week." Especially when there is a clear note in the chart saying the patient was phoned, told of their appointment date, and even read back the day of their appointment. This sometimes happens if the appointment did not make its way into the patient's calendar. When I am called at work in the middle of something I may not have time or remember to enter the appointment into my schedule. Also, manually entering dates has the risk of placing it at the wrong time or day.
Clinics can send links in text message or e-mail that automatically add themselves to a patient's cell phone or computer calendar. For patients who still use paper calendars, it may help to send a sticker that can be placed on the calendar.
4. Remind patients of their appointment
Since the 1990s randomized trials have demonstrated the benefits of reminder telephone calls over mailouts in reducing no-show rates. Is it better to personally call patients or use an automated service? A 2007 study randomized 10,546 outpatients to receive three days before their appointment (1) no reminder, (2) an automated phone call, or (3) a phone call by the Clinical Staff. No-show rates were 23.1%, 17.3%, and 13.6% respectively. Amusingly, when patients were asked in clinic who they spoke to they couldn't remember accurately if it was a clinical staff person or an automated machine (Parikh. The Effectiveness of Outpatient Appointment Reminder Systems in Reducing No-Show Rates. Am J Med. 2010).
The optimal time to remind a patient before an appointment remains a matter of debate. As the time between booking an appointment and its date lengthens the no-show rate increases. Appointments made in a primary care practice of over 9000 patients showed that when there were more than 61 days between booking and the appointment the odds ratio (OR) of a no-show was 15.26 (95% CI 6.75-34.5). Appointments with a lag of 31-60 days (OR 3.46), 15-30 days (OR 1.92), 9-14 days (OR 2.50), 3-7 days (OR 2.12), and 2 or less days (OR 2.12). (Steiner, J. Reducing Missed Primary Care Appointments in a Learning Health System: Two Randomized Trials and Validation of a Predictive Model. Medical Care 2016).
Two extremes to avoid are sending reminders too early or too late. Many clinics mail their reminder the day the appointment is made. In specialty clinics this may mean a follow up in 6 months or a wait of over a year and a half. People may not even have a calendar for the year of their appointment to write on when they receive the letter. Contacting too late is another problem. Clinics may mail their reminders the week prior to appointment. In practice this means getting a letter the morning of (or after) the appointment.
It is difficult to know how the 1990s studies on mail-outs and telephones apply in 2016. Despite an increase in ways to contact patients, each method has its own limitations. Cell phones have caller ID and often people do not answer calls from unknown or blocked numbers. Community mailboxes are not checked on a daily basis. Some people check them weekly, monthly, or even "once when they move in." Groups such as Kaiser have implemented text message reminders with high patient acceptability. Only 1.8% of people asked to be removed from the service.
As 'conversational bots' (artificial intelligence that can hold a communication) become more sophisticated I anticipate hospitals will be able to purchase automated texting services that communicate in natural language to book or cancel appointments. Based on preliminary uptake of such 'chat bots' in online retail, there likely will be high acceptability of such a product.
5. Network hospital databases
In a study of 1352 patients with advanced cancer who were referred to a supportive care clinic at University of Texas Medical Centre 16% missed their first appointment. This appointment was 7-9 days from the referral date.
Where were the absent patients? 18% were at another appointment with their primary oncologist, 9% were at another appointment, 9% were in the ER, 8% were admitted to hospital, and 2% were dead. Reasons for the other 54% was unknown (Delgado et al. Characteristics and outcomes of advanced cancer patients who miss outpatient supportive care consult appointments. Supportive Care in Cancer. 2014). Although this patient population is an extreme case, it demonstrates that nearly half of the missing patients in the study were elsewhere in the hospital system. By networking the clinics, ER, and hospital databases together scheduling conflicts could be avoided, and clinicians could be aware when their patients are in acute care.
B) Remove Barriers to Attendance
6. ** Identify patients who are at high risk for not showing **
When clinicians are provided in advance a list of their clinic schedule they can identify which patients are at high risk for not attending. This is based off their past experience with the patient. Several studies have show how this intuition can be quantified into a predictive algorithm.
*** In 2014 a 10 variable prediction model for missed appointments was designed for Kaiser Permanente Colorado's primary care network. The original data set was based on 8,804 adults, and it was verified on 7,497 patients at a second clinic. The accuracy of the model was very good, a c-statistic of 0.90.
The strongest predictor of a no-show was having missed one or more appointments in the past six months. Other factors included, being younger, male, having longer time between scheduling and date of appointment, and non-white patients. Those patients in the lowest quartile of the algorithm only missed 0.4% of appointments, whereas those in the highest quartile missed 24.1-28.9%.
The clinic implemented an intervention using phone, text, and email to improve attendance. It was statistically significant at reducing no-shows. They estimated the clinic would have "14 fewer missed visits each week and have 3 additional open appointments due to cancellations if all adult patients received a single reminder (Steiner, J et al. Reducing Missed Primary Care Appointments in a Learning Health System: Two Randomized Trials and Validation of a Predictive Model. Medical Care. 2016).
*** Newer techniques such as using Natural Language Processing can read the text of electronic medical record looking for key phrases and sentiments that suggest a patient has a history of not attending or taking medications. When this factor was combined with classic risk factors Bluthenthal generated a model with 92% specificity (although only 33% sensitivity) (Blumenthal, D et al. Predicting Non-Adherence with Outpatient Colonoscopy Using a Novel Electronic Tool that Measures Prior Non-Adherence. Journal of General Internal Medicine. 2015).
7. Use patient navigators for high risk patients
Patient Navigators have become popular in institutions, such as Cancer Care, to help guide patients through complex care systems. Navigators provide consistency and familiarity. They help answer questions for the patient, remind them of appointments, coordinate care, and remove barriers.
I did not review in depth all the literature surrounding Patient Navigators. However, it seems this resource intensive method can be effective. As an example, a tertiary care colposcopy referral clinic reduced their no-show rate from 49.7% to 29.5% with the use of patient navigators in their study population of 4,199 patients. The authors concluded,
We found that 45% of patient no-shows were anticipated or a result of patient misunderstanding and could be mediated with targeted education by the patient navigator.
8. Involve Social Work
Multiple studies identify the cost of transportation as a contributing factor for non-attendance. In addition, surveys performed on underserved minority patients in New York City as part of the Immigrant Cancer Portal Project show a correlation between 'unmet needs' and self reported missed appointments. Unmet housing was one with the highest odds ratio. (Costas-Muniz, R et al. Association of socioeconomic and practical unmet needs with self-reported nonadherence to cancer treatment appointments in low-income Latino and Black cancer patients. Ethnicity & Health. 2016. & Gany, F et l. Targeting social and economic correlates of cancer treatment appointment keeping among immigrant Chinese patients. Journal of Urban Health. 2011.)
Although it is unrealistic for clinic social workers to address every unmet need for all their patients, they could help advocate and co-ordinate affordable transportation. Literature from no-shows for colonoscopies identifies the inability to find a driver / chaperone to be a significant reason patient do not attend (Bhise, V et al. Patient-Reported Attributions for Missed Colonoscopy Appointments in Two Large Healthcare Systems. Digestive Diseases and Sciences. 2016). Applying this concept to oncology, patients attend appointments often with another person given the logistics of parking. Coordinating transport may help patients attend.
9. Make clinics easier to attend.
Improve parking: One of our Oncology centres has reserved oncology patient parking 25 feet in front of the door. Patients greatly prefer this site for treatment over sites where the parkade is blocks away from the hospital and is sometimes full.
Have local clinics: Research shows that implementing community sties for screening helps reduce no-shows. A study of high risk Canadian women suggested implementing local colposcopy programs reduced their no-show rates from 17.2% to 1.3%. (Ogilvie, G et al. Access to colposcopy services for high-risk Canadian women: can we do better? Canadian Journal of Public Health. 2004).
10. Educate your patients or use a carrot
Patients are more inclined to attend when there are direct incentives. An incentives to attend is when that the patient feels unwell and wants to get better. Many conditions do not have significant symptoms, and helping educate and involve patients in their care will enable them to understand their disease and treatments better, and in turn be more adherent with appointments.
Patients with a history of substance abuse have increased attendance when their opioid prescription lasts the precise interval between appointments (Pillet & Eschiti. Managing chronic pain in patients with cancer who have a history of substance abuse. Clinical Journal of Oncology Nursing. 2008).
C) Optimize the Clinic
11. Predict aggregate no-show rates
In 1999 Beitler, Garcia, and Vikram used 11 months of data (21,265 radiation appointments) to model the exact no show rate for a radiation oncology clinic. They calculated it to be 11% with 95% confidence interval of 1%. (Beitler et al. Experience-based demand scheduling: a more efficient model for radiation oncology. American Journal of Clinical Oncology. 1999). With this type of accuracy, one can 'overbook' a clinic so that it runs closer to 100%, rather than 89% capacity.
Seasonal variations and weather are known to have an impact on attendance. An extreme example is that 23% of all no-shows in an El Salvidor clinic were during tropical storms in October (Salaverria). IBM purchased in 2016 The Weather Channel. This data will be very useful to its predictive analytics software, Watson. Hopefully this type of data will be one day easy to integrate into clinic scheduling software.
12. Use a computerized system and real time check-ins
The current booking system has two problems. First, a clinic may not find out their 3:00pm patient is a no-show until 3:00. This does not leave time to react to the missed appointment. Second, clinics rarely run on time. In many clinics patients may wait hours.
An automated reminder system could texted or call patients the day prior to their appointment and confirm their attendance. Instead of booking a time, patient's could book a time window in which their appointment would occur, such as early afternoon. As this time window approaches they receive text notifications of what time their appointment will be. These notifications can be timed based on the real time flow of the clinic that day. This reduces the time patients spend in the waiting room, and increases flexibility in clinic scheduling.
13. Have a standby list
Empty and missed appointment slots can be filled with patients who are in hospital and have an upcoming appointment. Alternatively, a list of patients who are able to come into clinic on short notice can be used. If a system such as that proposed in Item 12 above is used, the missed appointment slots can be collapsed, and the appointment time of patients later in day advanced. This creates time to notify patients on the standby list to fill the end of day appointments.
14. Have adult-pediatric transition clinics
Evidence suggests that joint transitional care oncology clinics for patients moving from pediatrics to adult has better success (Frey & van der Pal. Transitional care of a childhood cancer survivor to adult services: facilitating the process of individual access to different models. Current Opinion in Supportive and Palliative Care. 2013).
15. ** Track and follow up missed appointments **
Most clinics I have been to do not have a system to track patients who miss appointments and figure out why they missed. Without recording this type of data it is difficult for a clinic to decrease their no-show rate. The studies also show that even when clinics do record no-shows, the data often is inaccurate because it may join no-shows, with cancelations, with patients the physician phoned and told they do not need to come into clinic.
A very impressive study published in 2015 from El Salvador demonstrated the effects of diligently following up with patients who miss appointments. In their case they aimed to decrease 'treatment abandonment' among pediatric oncology patients. Treatment abandonment was defined as missing "4 consecutive weeks of treatment or not starting therapy after a cancer diagnosis". The country's treatment abandonment rate among pediatric oncology patients had remained at 13% for over a decade.
In the study they followed 1627 patients with 44,894 appointments (58% physician visits and the rest chemotherapy treatments). There were a total of 1111 missed appointments (2% of total). These were from 30% of patients (491 people).
The country initiated a Time Sensitive Adherence Tracking Procedure. The system alerted clinics to patients who missed an appointment. These patients were promptly followed up with by the clinic. If the clinic couldn't reach the patient the assistance of local health clinics and municipalities was use. Lastly, if the patient still couldn't be reached, law enforcement was used.
The reasons for not attending the appointment were varied.
- 23% financial need: lack of funds for travel on the day of appointment, transportation aid not available that day
- 16% unforeseen barriers: natural disaster, public transportation problem, labour strike
- 12% domestic needs: caregiver was ill or had to care for another member of the family, or there was no one able to bring them to hospital, caregiver had to work
- 10% caregiver decided against treatment that day: non-compliant with treatment, attended another appointment (eg school)
- 8% caregiver forgot appointment date
- 7% ongoing demands: stresses and demands of living conditions
- 5% reason not obtained / missing information
- 2% caregiver disagreed with treatment: fearful, alternative treatment opted
- 1% unclassifiable
By using this aggressive follow up program with the intention of contacting every patient who no-showed, the treatment abandonment rate fell from 13% to 3% during the two years of the trial. In fact there was only one abandonment case in the last year of the trial. (Salaverria, C. et al. Interventions targeting absences increase adherence and reduce abandonment of childhood cancer treatment in El Salvador. Pediatric Blood & Cancer. 2015)
Pediatric oncology appointments are 'free' in El Salvador. However, 13% of patients were not completing treatment. The dramatic ability to drop the treatment abandonment rate in this study strongly suggests that the reason treatment is abandoned is not because patients do not want to continue, but in fact because they fall through the cracks in the system.
V. Future Research
1) New studies are required that use modalities such as text message and e-mail, and compare it against current rates of answering calls on cell phones and land lines or mailouts. It would be interesting to see this data with sub-group analysis based on economic status and age.
2) What is the effect of tele-health on medication adherence or attendance at treatments such as chemotherapy or radiation therapy? A study compared the no-show rate of patients scheduled to either meet their colonoscopist at an appointment prior to booking the procedure, or at the procedure itself. Patients who met the doctor at an appointment ahead of time were more likely to have the procedure 66.9% vs 45.5%, and had near half the number of no shows 7.3% vs 13.5%, or cancelations 19% vs 31.5%. (Ghaoui, R et al. Open access colonoscopy: Critical appraisal of indications, quality metrics and outcomes. Digestive and Liver Disease. 2016). Is there an 'adherence effect' or 'therapeutic effect' of meeting the clinician in person. Is this 'effect' lost via tele-health or mobile medicine?
3) What about missed appointment deterrence fees? These are common in clinics such as dentistry. How do they work in medical clinics? Many patients who do not show already have financial difficulties. Do deterrence fees perpetuate a cycle of being unable to pay, and not seeking care?
4) Do some physicians have higher or lower cancelation rates? When all other factors are held constant such as staff, clinic, and reminder policies. What can be learned from those who have low cancelation rates.
5) Behavioral Psychology Research: marketing is consistently testing multiple campaigns against each other to determine which is most effective. Healthcare could apply the same rigor to understand what is the most effective way to help patients to keep their appointments, understand their health, and take medications better.
VI. The Future of Appointments
We know medical appointments take a significant amount of time and energy to attend for very little time spent with the clinician.
The ratio of 'time-wasted' to 'time-with-clinician' must improve.
Being able to deliver efficient, effective, empathic, and economical "appointments" via mobile devices will be an important step forward in medical care.
continue onto part 3/3: Our Lawyers Propose a Color Coded System for Missed Appointments
return to part 1/3: Why do Patients Miss Appointments?