Attendance Behavior and Rebooking Scenarios

Overview

Once appointments are generated, not all patients attend.
Some cancel, others miss their visit entirely, and some are rebooked.
The parameters status_rates and rebook_category jointly control these post-booking behaviors, shaping the realism of appointment outcomes over time.

You will explore three practical scenarios:

1. A balanced system with typical attendance and partial rebooking.

2. A no-rebooking system representing rigid scheduling (lost cancellations).

3. A highly adaptive system with full rebooking and low non-attendance.

Each scenario shows how these parameters modify attendance outcomes and day-to-day clinic activity.

---

Example 1 – Balanced outpatient attendance (default behavior)

A standard clinic using NHS-derived attendance proportions and 50% rebooking intensity.

from medscheduler import AppointmentScheduler
from medscheduler.utils.plotting import (
    plot_appointments_by_status,
    plot_appointments_by_status_future,
    plot_status_distribution_last_days,
    plot_status_distribution_next_days,
)
import pandas as pd

sched_default = AppointmentScheduler(
    date_ranges=[("2024-01-01", "2024-12-31")],
    ref_date="2024-11-25",
    fill_rate=0.85,
    rebook_category="med"
)

slots_df, appts_df, patients_df = sched_default.generate()

# Visualizations
plot_appointments_by_status(appts_df, scheduler=sched_default)
plot_appointments_by_status_future(appts_df, scheduler=sched_default)
plot_status_distribution_last_days(appts_df, scheduler=sched_default, days_back=21)
plot_status_distribution_next_days(appts_df, scheduler=sched_default, days_ahead=21)

Output preview:
Below are the four main outputs for this configuration:

1. Past appointment outcomes – Proportion of attended, cancelled, and missed visits before the reference date.
   

2. Future appointment outcomes – Expected proportions for upcoming appointments.
   

3. Recent daily activity (past 21 days) – Trends in attended vs. missed appointments.
   

4. Upcoming daily bookings (next 21 days) – Near-term expected activity.
   

Interpretation:
Attendance follows a balanced pattern: most patients attend, a smaller share cancel, and few miss appointments. About half of the cancelled visits are rebooked, keeping utilization stable over time.

---

Example 2 – No-rebooking system (low adaptability)

A rigid scheduling model where cancelled appointments are never rebooked.
This setup mimics systems with poor patient follow-up or limited administrative capacity.

sched_norebook = AppointmentScheduler(
    date_ranges=[("2024-01-01", "2024-12-31")],
    ref_date="2024-11-25",
    fill_rate=0.85,
    status_rates={"attended":0.75,"cancelled":0.15,"did not attend":0.09,"unknown":0.01},
    rebook_category="min"
)

slots_df, appts_df, patients_df = sched_norebook.generate()

plot_appointments_by_status(appts_df, scheduler=sched_norebook)
plot_appointments_by_status_future(appts_df, scheduler=sched_norebook)
plot_status_distribution_last_days(appts_df, scheduler=sched_norebook, days_back=21)
plot_status_distribution_next_days(appts_df, scheduler=sched_norebook, days_ahead=21)

Output preview:
Below are the four main outputs for this configuration:

1. Past appointment outcomes – Shows lower overall completion rates due to unrebooked cancellations.
   

2. Future appointment outcomes – Demonstrates reduced forward occupancy and more open capacity.
   

3. Recent daily activity (past 21 days) – Displays visible dips following cancellation spikes.
   

4. Upcoming daily bookings (next 21 days) – Shows empty days ahead due to missing reschedules.
   

Interpretation:
With no rebooking, cancellations directly reduce throughput.
Future occupancy remains lower, making this configuration representative of clinics that cannot replace missed visits quickly.

---

Example 3 – Fully adaptive scheduling (maximum rebooking)

A highly responsive service where all cancelled appointments are immediately rebooked, and non-attendance rates are minimal.

sched_rebookmax = AppointmentScheduler(
    date_ranges=[("2024-01-01", "2024-12-31")],
    ref_date="2024-11-25",
    fill_rate=0.9,
    status_rates={"attended":0.85,"cancelled":0.10,"did not attend":0.04,"unknown":0.01},
    rebook_category="max"
)

slots_df, appts_df, patients_df = sched_rebookmax.generate()

plot_appointments_by_status(appts_df, scheduler=sched_rebookmax)
plot_appointments_by_status_future(appts_df, scheduler=sched_rebookmax)
plot_status_distribution_last_days(appts_df, scheduler=sched_rebookmax, days_back=21)
plot_status_distribution_next_days(appts_df, scheduler=sched_rebookmax, days_ahead=21)

Output preview:
Below are the four main outputs for this configuration:

1. Past appointment outcomes – Almost all historical slots are attended or rebooked.
   

2. Future appointment outcomes – Shows a nearly filled forward calendar with minimal cancellations.
   

3. Recent daily activity (past 21 days) – Displays consistently high attendance and rapid recovery from cancellations.
   

4. Upcoming daily bookings (next 21 days) – Maintains high occupancy across future days.
   

Interpretation:
This system maximizes capacity by instantly recycling cancellations into new bookings. Attendance remains strong, future utilization near complete, and daily activity stable — typical of high-efficiency environments.

---

Summary

Scenario	rebook_category	Attendance pattern	Utilization outcome
Balanced	"med"	Typical mix of attended/cancelled/DNA	Stable utilization
No rebooking	"min"	Lower throughput, cancellations lost	Declining occupancy
Full rebooking	"max"	Almost all cancellations recovered	Near-max efficiency

---

Next Steps

• Examine how attendance influences patient flow in Patient flow.

• Explore Appointments table for how outcomes are stored.

• See Booking dynamics for how fill rate and scheduling affect these behaviors.

• Continue with Appointment Timing and Daily Flow to link attendance outcomes with punctuality and visit timing.