DDHA 8503 Week 3 Causal Loop Analysis Example

Backlog, Evenings, Burnout, Fewer Clinic Hours, Fewer Visits, More Messages: One Reinforcing Loop Traced Link by Link in a Primary Care Group's Inbox

Student Name

Walden University

DDHA 8503: Strategic Management and Systems Thinking

Instructor Name

Month Day, Year

Backlog, Evenings, Burnout, Fewer Clinic Hours, Fewer Visits, More Messages: One Reinforcing Loop Traced Link by Link in a Primary Care Group's Inbox

Why One Loop

The system description of Brightwater Medical Group's patient messaging identified several feedback structures. This analysis commits to one of them and follows it all the way round, because a page of arrows drawn without argument cannot be checked, while a single circuit argued link by link can. The loop chosen connects the inbox to the clinic schedule. It matters because it explains how a messaging problem can become an access problem, and an access problem can return as a messaging problem, with delays long enough that leaders rarely see the two as one.

The Diagram

The loop has six variables, each named as a quantity that can rise or fall. Unanswered message backlog increases clinician after-hours inbox time. After-hours inbox time increases clinician burnout. Burnout decreases clinical hours offered. Clinical hours offered increase visit availability. Visit availability decreases messages sent in place of visits. Messages sent in place of visits increase the unanswered message backlog.

Table 1

The Loop's Links, Signs and Delays

In the drawn diagram the six variables sit in a circle with arrows marked plus or minus; delay marks appear on links 2 and 3, the two that hide the loop.

Argument for Each Link

Link 1 is positive. When more messages are waiting, clinicians spend more time on them outside scheduled visits, because Brightwater's clinic days are fully booked with patients and the inbox has no protected time. Sinsky et al. (2016) observed physicians spending one to two hours of personal time each night on record work, and Brightwater's own audit logs show evening inbox time rising in weeks when the Monday backlog is high.

Link 2 is positive. More after-hours inbox work raises the likelihood of burnout. Tai-Seale et al. (2019) linked heavier volumes of algorithm-generated inbox items to a higher chance that physicians described themselves as burned out. The delay is months, because burnout is a slow accumulation rather than a reaction to one heavy week.

Link 3 is negative. When burnout rises, the clinical hours clinicians offer fall. Some reduce their clinical appointment from full time to part time, some leave the group and some stop accepting new patients. Shanafelt et al. (2016), following physicians at one academic center, found that burnout was associated with a subsequent reduction in professional work effort over the following two years. This is the longest delay in the loop and the main reason the loop is hard to see.

Link 4 is positive. Fewer clinical hours mean fewer visit slots. The relationship is direct, with a delay of weeks as schedules already built run out.

Link 5 is negative. When visit availability falls, patients who cannot get an appointment soon write a message instead. They ask the question they would have asked in the room, request a refill they would have discussed, or describe a symptom in the hope that the clinician will say whether it needs to be seen. Brightwater's data show that messages from patients who were offered no appointment within two weeks are about twice as frequent as from those who were. The delay is weeks, the time it takes patients to learn that visits are scarce.

Link 6 is positive. More messages sent in place of visits raise the backlog within days, since the answering capacity of the nurse pool and clinicians does not change with demand.

Classifying the Loop

The loop contains two negative links, links 3 and 5. An even number of negative links makes a loop reinforcing (Sterman, 2000). Starting anywhere and going round, a change returns to strengthen itself: a larger backlog leads, through burnout and lost clinical hours, to fewer visits and more messages, which enlarges the backlog. The loop can also run in the favorable direction, where a smaller backlog eventually supports more clinical hours, more visits and fewer messages.

Behavior Over Time

Left alone, a reinforcing loop produces growth that accelerates, here growth in backlog and decline in clinical hours. Because of the delays on links 2 and 3, the early phase looks almost flat. For a year or more, the backlog may rise slowly while clinical hours seem stable, and leaders may treat the inbox as a workload nuisance. Then, as clinicians reduce hours or leave, visit availability drops, messages rise faster and the backlog begins to climb steeply. The pattern resembles a curve that bends upward after a long, quiet start. In practice, the growth would eventually be limited by balancing loops outside this one, such as patients leaving the group or nurses absorbing more work, but those limits come at their own cost.

What the Delays Do

The delays on links 2 and 3 do two things. First, they separate cause from effect in time, so that the clinicians who reduce their hours in one year are responding to inbox conditions from the previous year, and no meeting connects the two. Second, they make fixes look ineffective. A change that reduced after-hours inbox time today would not show up in clinical hours for a year or more, so a leader judging it after one quarter might abandon it before it could work. The delay on link 5 is shorter, which means the access-to-messages part of the loop responds quickly, and that is where the first visible signs of the loop are likely to appear.

Limits of the Analysis

This analysis follows one loop and treats the others as outside it. Link 5 rests partly on Brightwater's own data and partly on reasoning about patient behavior, and its strength is uncertain. The delays are estimates. The next step is to test the loop's prediction: if it is right, clinics with the highest after-hours inbox time two years ago should show the largest loss of clinical hours now.

References

Shanafelt, T. D., Mungo, M., Schmitgen, J., Storz, K. A., Reeves, D., Hayes, S. N., Sloan, J. A., Swensen, S. J., & Buskirk, S. J. (2016). Longitudinal study evaluating the association between physician burnout and changes in professional work effort. Mayo Clinic Proceedings, 91(4), 422-431. https://doi.org/10.1016/j.mayocp.2016.02.001

Sinsky, C., Colligan, L., Li, L., Prgomet, M., Reynolds, S., Goeders, L., Westbrook, J., Tutty, M., & Blike, G. (2016). Allocation of physician time in ambulatory practice: A time and motion study in 4 specialties. Annals of Internal Medicine, 165(11), 753-760. https://doi.org/10.7326/M16-0961

Sterman, J. D. (2000). Business dynamics: Systems thinking and modeling for a complex world. Irwin/McGraw-Hill.

Tai-Seale, M., Dillon, E. C., Yang, Y., Nordgren, R., Steinberg, R. L., Nauenberg, T., Lee, T. C., Meehan, A., Li, J., Chan, A. S., & Frosch, D. L. (2019). Physicians' well-being linked to in-basket messages generated by algorithms in electronic health records. Health Affairs, 38(7), 1073-1078. https://doi.org/10.1377/hlthaff.2018.05509