Six ED Visits in Six Months: Diabetes, Homelessness, and No Refrigerator
Mr. D is a 52-year-old man with type 2 diabetes (A1c 12.4%, BMI 34) who has had six emergency department visits in six months: two for hypoglycemia, one for cellulitis of the foot, and three for hyperglycemia. He has no primary care provider, lives in a men's shelter where he sleeps in a dormitory and has no refrigerator or secure storage, and is covered by Medicaid. His insulin glargine prescription has gone unfilled twice; he says he "didn't have a way to keep it cold" and missed a clinic appointment because he was working a day-labor job. He screens positive for food insecurity and reports eating mostly from a convenience store. He is alert, oriented, and independent in self-care, and he says he is "done with hospitals" but agreed to meet with the case manager. The shelter has a visiting nurse once a week and a community health worker program through the local federally qualified health center.
1. How should the case manager prioritize this patient's problems, and what should the assessment cover?
Read the worked answer
Prioritize by risk of immediate harm and by what blocks everything else. First, the acute clinical hazards: recurrent hypoglycemia on insulin without reliable food or glucose monitoring, and a foot cellulitis with no one inspecting his feet. Second, the structural barriers, because no education plan survives without them - refrigeration and secure storage for insulin, food, and dependable access to a prescriber. The assessment should cover medical status (A1c, renal function, hypoglycemia history, foot exam), function (he is independent in self-care, so this is not a custodial issue), behavior and knowledge (food choices shaped by access rather than ignorance), environment (shelter rules, no kitchen, dormitory theft), financial and benefit status (Medicaid, pharmacy options), and psychosocial factors (distrust of hospitals, no PCP, isolation). Use a structured social-determinants screen. The judgment: a carbohydrate-counting class would fail here; the actionable priorities are a place to keep insulin, reliable food, foot checks, and a clinic he will actually attend.
2. Build the interprofessional care plan, including how to handle insulin storage and the lack of a primary care provider.
Read the worked answer
Build the plan around the shelter, not the hospital. Team: the FQHC community health worker, the shelter's visiting nurse, a pharmacist, a diabetes educator, a social worker, and the shelter case manager. Concrete actions: simplify the regimen to something he can store and use - for example, a prefilled pen and, if clinically appropriate, a basal regimen whose storage requirements the pharmacist confirms for unrefrigerated use; arrange pharmacy delivery to the shelter or a lockbox; enroll him in medically tailored meals or a food pantry with diabetes-appropriate options; set foot checks at the weekly nurse visit and teach daily self-inspection; book a PCP visit at the FQHC with transportation arranged; and give him a written hypoglycemia crisis plan. Assign an owner and a date to each element. Frame the plan around his goals - keeping the day-labor job and staying out of the ED - and use the community health worker for in-person follow-up. Confirm he agrees with each element; a plan he did not help build will not survive the first week.
3. Leadership wants to know whether the program works. Which outcome measures should be used, and what are their limits?
Read the worked answer
ED visit counts are a useful utilization measure but a poor sole outcome: they are low-frequency, shift with where he seeks care, and do not show whether his diabetes is actually safer. Use a small set. Process: insulin in hand and correctly stored, PCP visit completed within 30 days, foot exam documented, meals program enrollment. Outcome: HbA1c trend at 3 and 6 months (comprehensive diabetes care, with HbA1c above 9% as poor control), hypoglycemia episodes, ED visits and admissions, and patient-reported confidence or diabetes distress. Track them on a run chart with a stated denominator. Be honest about attribution: six ED visits in six months is a tiny baseline, so a drop to three could easily be chance. Report the full picture to leadership and watch for gaming - an improved ED count alongside a worsening A1c is not success.