Abstract
Introduction:
St. Thomas (STCHC) has a long history of serving an urban, uninsured, and underinsured (Medicaid) population in the greater New Orleans area. Beginning in 2010, STCHC faced financial danger when Louisiana did not expand Medicaid coverage. Partnering with the nonprofit Institute for Healthcare Optimization (IHO), we increased patient throughput by balancing the clinic’s resources between two modalities of care: urgent (same/next day) appointments in which patients were seen by whichever provider was available; and less urgent, in which patients saw their preferred PCP at the provider’s availability. Assessing patient’s appointment preferences was crucial in directing this innovation.
Methods:
We analyzed 8913 appointments and patients’ priorities between being cared for by a preferred provider or being seen on the same/next day. Appointment outcomes (i.e., occurred, no-show, canceled, or rescheduled) were matched to each preference group and no-show rates were calculated.
Results:
When 80% to 90% of same-day requests were satisfied, the proportion of no-shows declined significantly, and patients’ satisfaction with access to care reached 97%. Clinic appointment kept rate became 87% (95% CI [0.869975, 0.870025]).
Conclusions:
Disseminating these results nationwide would result in many lives and millions of dollars saved. Large-scale educational and implementation efforts are needed.
Introduction
Nearly all primary care practices experience “no-shows.” For underserved populations, fixed appointment scheduling is often a deterrent, given competing demands/obligations, including family responsibilities and employment, resulting in limited degrees of freedom for attendance.1 -3 The impact of missed appointments are well known. In one recent study, for example, patients were 40% less likely to experience the care they needed within a year after just one no-show. 4 Another analysis found that even one no-show increased the likelihood of patients abandoning a practice site by 70%. 5
An expensive way that some clinics lower no-show rates is through the “concierge care” model that offers patients maximally convenient appointment times, in exchange for a membership/retainer fee. At a glance this approach does not seem to be applicable to low-income populations. Can we offer access to “on-demand” appointments without requiring additional fees? The answer is “yes.”
In 2007, St. Thomas Community Health Center (STCHC) changed its status to a Federally Qualified Health Center (FQHC), a designation that allowed it to collect higher levels of Medicaid reimbursement and purchase drugs at a discount in exchange for treating all patients that arrive, regardless of income or insurance status. That higher reimbursement arrangement allows the physicians there to provide low cost, high quality primary care services. This business model, operating via a sliding fee schedule for patients, depended primarily on grant support and CMS funding. After the 2010 passage of the Affordable Care Act, Louisiana refused to expand its Medicaid program, and the clinic lost some funding. In parallel, the uninsured rate among the St. Thomas’ patient base increased from 18% to 35% in less than 2 years.
Most STCHC patients live close to or under the national poverty guideline which, in 2022, is an annual income of $27 750 for a family of four. 6 Sixty percent of our patients’ incomes are below 200% Federal Poverty Level, with 46% below 100% Federal Poverty Level. Eighty-seven percent of its patients have some type of health insurance, the majority having either Medicaid (62%) or Medicare (12%). Thirteen percent of STCHC patients have commercial insurance through their employer or a family member’s employer. The remaining 13% of patients are uninsured. In 2021, St. Thomas provided health care to 18 337 patients from 314 unique zip codes. 65% of the patient population is female and 35% is male. Its patients’ ethnicities are: 67% African American; 12% is Hispanic; 16% is White; and 1% is Asian. The remaining 4% of the population’s race/ethnicities are unreported.
St. Thomas struggled to accommodate the increased proportion of uninsured patients, leading to a dire position: shutdown, or find another solution. Shutting down would mean the cessation of care for thousands of patients, and an increased burden on local emergency departments, thereby exacerbating Emergency Department overcrowding.
To stay afloat, any FQHC must maintain sustainable margins. To do so, increasing patient throughput is a necessity. On one hand, to increase patient throughput, physicians’ utilization should be as high as possible, leaving little or no room for unexpected demand. On the other hand, with high utilization, there is little flexibility for urgent requests to be met; to provide those patients with high quality care means lowering the utilization rate, allocating capacity for drop-in patients.
For St. Thomas, meeting patient needs has always been the main priority, which required a lowered utilization rate. This choked the system’s margins and prompted STCHC to seek collaboration with the IHO, a nonprofit education and consulting organization. With IHO’s guidance, St. Thomas was able to split its provider pool by assigning some providers to urgent demand and some to scheduled demand.
This subdivision made it unlikely that patients could simultaneously choose same-day appointments and their own/preferred physician, because their physician might be serving previously scheduled patients. There is a deeply rooted notion in healthcare that patients prefer to be seen by their own PCP, even at the expense of not seeing them on the same day that a particular need arises. St. Thomas decided to reassess this preconceived notion by examining patient no-show rates.
Methods
To better learn the needs of the patient population, appointment scheduling calls through the St. Thomas call center were collected and analyzed retrospectively, as well as actual appointment outcomes. Along the lines of inquiry listed below, call center information was matched with appointment outcome data on a per-patient basis. 8913 such records were analyzed from across a two-month period.
Patients calling into the STCHC call center to schedule appointments were asked their preference for:
(1) the appointment date (same/next day or future date);
(2) the appointment type (In person or telemedicine); and
(3) their preferred provider.
Every effort was made to meet the patient’s preferences. The appointment request would be considered “met” if the scheduled appointment met the first priority the patient had for an appointment.
Appointments were analyzed in two groups. The first group prioritized same-day appointments, and the second group was for requests to be seen by the preferred provider, at a future date (greater than 4 days hence). For both groups, the patient’s actual experiences were matched to whether their preferences were met.
The following subdivisions were also made:
(1)
(2)
Each combination resulted in four outcomes. Combination 1 outcomes (i.e., requested date and type of appointment) were:
(1) The appointment met both the requested date and appointment type (2/2 requests met).
(2) The appointment satisfied the requested date, but NOT the appointment type (1/2 requests met).
(3) The appointment did NOT meet the requested date, but did meet the requested appointment type (1/2 requests met).
(4) The patient’s appointment had neither the requested date nor appointment type (0/2 requests met).
The combination 2 outcomes (i.e., preferred appointment time and preferred provider) were:
(1) The appointment made met both the requested date and provider (2/2 requests met).
(2) The appointment met the requested date, but the appointment was not with the requested provider (1/2 requests met).
(3) The appointment was not on the requested date, but with the requested provider (1/2 requests met).
(4) The patient’s appointment had neither the requested date nor provider (0/2 requests met).
The appointment outcome (i.e., occurred, no-show, canceled, or rescheduled) was determined for each of the conditions above and no-show rates were calculated and compared.
The analytical focus of this project is on the appointment date and provider variables. The reason for denying appointments at a patient’s desired time with a particular provider was the availability of those providers, regardless of whether the appointment was conducted in-person or another way. Therefore, analyzing appointment completion across providers and appointment type was not within the scope of this study.
Results
During the study period, 39% of the 8913 scheduling calls were for same-day appointments, while the remaining 61% of requests were for future appointment dates. As per Table 1 below, the clear preference for patients requesting same-day appointments was the date of the appointment: patients who requested, and received, a same-day appointment, failed to show up 12% to 13% of the time while those that requested, but did not receive a same-day appointment, failed to show up 44% to 47% of the time. This is more than a tripling of no-show rates, indicating the patients’ strong likelihood to keep same-day appointments, even if not with their preferred provider. Table 2 compares no-show rates pertaining to visit type fulfillment and time preference fulfillment, again showing a strong likelihood of patients keeping appointments that align with their time preferences (13% no-show rate when date requests were made; 44% no-show rate when date requests were not met).
No-Show Rates (Preferred Appointment Date vs Provider).
No-Show Rates (Preferred Appointment Date vs Visit Type).
Patients requesting same-day appointments prioritize being seen on the same-day over all other factors (70% of the time). Even a 1-day delay in the patient being seen more than doubled the patient’s no-show rate (from 13% to 33%). Patients requesting same-day appointments at the beginning of the week (Monday) and end of the week (Thursday and Friday), and not receiving same day appointment were less likely to keep appointments made for the next day (44%-48% no-show rate) as compared to the patients requesting same-day appointments in the middle of the week (Tuesday-Wednesday) (29%-31% no-show rate).
For future appointments, despite being given every preference requested 90% of the time (i.e., date, provider, and type), the no-show rate was 43%. This rate was similar for appointments initiated by the clinic staff (e.g., providers scheduling the next appointment at the end of a visit, health coaches suggesting follow-up according to care plans, etc.). STCHC patients readily accept appointments when suggested and report every intention of keeping the appointment.
Discussion
The success of St. Thomas’ achievement is extrapolatable to any FQHC setting—especially those struggling to stay open. Such change requires a strong organizational desire for improvement, strong leadership, and a mission-driven commitment; as well as the vision necessary to reach out and recruit the needed services for operational improvement, and awareness of internal data structures.
St. Thomas exemplified these characteristics throughout the change process. For example, STCHC endeavored not only to meet patient’s preferences, but to also study and assess the trends behind patients’ priorities.
Our results do not support the long-held belief that continuity of care by the same provider is so important to patients, that they are willing to wait. Instead, the data suggest that patients prefer to see someone who can take care of their needs urgently. Increasing the practices’ ability to make on-demand appointments possible would positively impact both patient health and patient throughput, thereby increasing practice margins.
Originally, St. Thomas booked appointments at a 70% to 80% utilization rate. Actual utilization rates at the time did not come close to this figure. By splitting the patient flows into two categories, catering to urgent (same/next day appointments) and scheduled appointments (non-urgent), lead to a reduction in wait time for urgent patients (to a degree that is competitive with the concierge medicine industry) 7 coupled with a greater throughput for scheduled patients. The proportion of no-shows declined significantly, and according to a 3rd party assessment, patients’ satisfaction with access to care reached 97%.
Overall, 25% more patients were seen the same year of the implementation, accompanied by an increase in capacity to accommodate an additional 10% increase in demand. Learning from the trends in patients’ preferences, we began offering a concierge-level of convenience that is affordable for patients. The financial implications of this intervention were vital to St. Thomas’ sustainability: $16 million in annual practice revenue. The provider workforce increased from 12 to 26 physicians and physician-extenders (PAs, NPs, etc.). This intervention resulted in positive cost savings for patients—the extra revenue enabled St. Thomas to subsidize the patients’ treatment for little or no reimbursement.
This project had one notable limitation, in that the applicability of these results to patient populations outside of the urban underserved population needs to be studied further.
Conclusion
By assessing our patient’s needs and preferences, we were able to fine tune and deliver an intervention that benefits them massively through increased access to care. We strongly encourage FQHCs and FQHC-lookalikes nationwide to assess their patients and communities and utilize those results when implementing patient-centric interventions like the one that saved us from the red ink. We believe that implementing this intervention nationwide could result in many lives saved, major health improvements, and millions of saved dollars. To achieve this goal, large-scale educational and implementation efforts would be required. HRSA has the unique ability to champion these innovations, benefiting the communities nationwide that are served by 1400 FQHCs. 8
