Research reveals reasons underlying patient no-shows
By Stacey Butterfield
It's the $19 million-dollar question: How do you get your patients to show up for their appointments? Howard Houghton, MD, and Patricia Alafaireet believe they have the answer.
The two began working on the problem of no-shows a few years ago at the University of Missouri, where Dr. Houghton is an assistant professor of clinical psychiatry and Ms. Alafaireet is the director of applied health informatics.
“The no-show rate was between 19% and 22% of appointments in the outpatient psychiatric clinic. That's not unusual. It also represented between $11 and $19 million in unrealized revenue. I don't know any department that can afford that kind of loss,” explained Ms. Alafaireet during a session at the Medical Group Management Association's annual meeting.
Using electronic health records and billing systems, they collected information about the characteristics of missed medical visits. Their main focus was the psychiatric department where Dr. Houghton practices, but data was also gathered from the university's continuity care clinic.
“What we really want to do is project ahead of time whether the patient is going to show up, and if they're not, we want to change something to get them to come in.”
A statistical analysis of the collected data revealed a number of factors that strongly affect whether a patient makes a visit, some of which had not been uncovered in previous research.
According to their research of 11,000 scheduled visits, some socioeconomic factors do affect the likelihood that a visit will take place. For example, Medicaid recipients had a higher rate of no-shows. But several of the most important factors related to the logistics of the visit, rather than the characteristics of the patient.
For example, one strong predictor was who scheduled the appointment for the patient. “We were really, really shocked the first time we went through the data and found out that the maker of the appointment controlled as much as one-third of the probability of whether the patient would show up,” said Ms. Alafaireet.
Schedulers with higher proportions of no-shows were following practice protocol and scheduling the first caller for Monday at 8 a.m., the second for Monday at 8:30, etc. More successful schedulers, however, started the conversation by asking patients, “When would you like to come in?”
These intuitive, relatively easy changes could be used to improve attendance rates. “For example, scheduling an appointment after 3 o’clock in the afternoon for a Medicaid patient was a 100% no-show, because public transportation ceases [in our area]. We started seeing some very obvious changes that needed to occur,” said Dr. Houghton.
Logistical changes also could be used to counteract the demographic factors that predict no-shows, they concluded. Young, unmarried males had the highest rate of no-show, but it wasn't consistent throughout the week. Scheduling this group mid-mornings Tuesday through Thursday could improve attendance.
Similar accommodations could be made for long-distance travelers. The study revealed patients who lived five to 10 miles from the practice were very likely to show, while those who lived between 19 and 60 miles were more likely to miss their appointments, but anyone coming from more than 60 miles away almost always showed up.
A combined analysis of the relevant factors makes it even clearer whether a visit will be a no-show, the researchers said. “Scheduling an 8 a.m. appointment for somebody who has to come 100 miles on public transportation, that's a 100% no show,” said Dr. Houghton.
The next, tougher step will be using that knowledge to reduce no-show rates, Dr. Houghton said.
“What we really want to do is project ahead of time whether the patient is going to show up, and if they're not, we want to change something to get them to come in. That's sort of our end goal, to create something like the Gail model that's used for breast cancer risk prediction.”
That model is likely to include many logistical factors along with the few demographic characteristics (marital status, gender) that were found to matter. “A staggering number of times it was transportation, insurance—logistical issues,” he said.
He and Ms. Alafaireet expect that practices will, for the most part, have to adapt the model by conducting a similar analysis of no-show records. “We really do think it's probably practice-specific,” she said. (It's not too hard to do using Excel, Ms. Alafaireet said.)
“You want to look for the factors that can be controlled,” advised Dr. Houghton. Of course, with some factors, the benefits of controlling for them will have to be balanced against the costs.
For example, the Missouri study found that no-shows were strongly related to the type of provider scheduled to see the patient. Residents got far more no-shows than attendings, while clinical social workers, who mostly had long-standing relationships with the patients, had very low no-show rates.
Obviously, not every patient can be scheduled to see a senior provider with whom they have a long-standing relationship, but if accommodations can be made patients who are particularly high-risk for no-shows, it benefits physicians, patients and every other patient.
“You lose a lot of productivity. That person who no-shows bumps somebody else out of a slot,” said Dr. Houghton.
Internist Archives Quick Links
Internal Medicine Meeting 2015 Digital Presentations
Choose from over 170 recorded Scientific Program Sessions and Pre-Courses. Available in a variety of packages and formats so you can choose the combination that works best for you.