Which Patient Needs Your Team's Attention Today?

Ask a care team a simple question on a Monday morning. Of everyone on your list right now, who needs you today?

Most organizations cannot answer it. Not because the people are careless, and not because the data is missing. They cannot answer it because nothing they measure was built to produce an answer.

The numbers we can produce

What a good team can put in front of you is impressive. Last month's discharge volume. The 30 day readmission rate, risk adjusted, trended against the same quarter last year. A benchmark position against peer hospitals, with the gap to median calculated to one decimal place.

Those numbers are real, and they matter. They are also, every one of them, descriptions of something that already finished happening.

There is a small tell buried in how we talk about them. Ask around the industry what the readmission rate is, and you will hear one in five, said confidently, by people running programs built on it. AHRQ's analysis of national data from 2016 through 2020 puts it at 13.9 readmissions per 100 index admissions overall, and 17.0 for Medicare. Closer to one in six. Nobody is lying. The number has been repeated so often that it stopped being a measurement and became a kind of folk figure, which tells you something uncomfortable about how closely we look at the thing we are supposedly managing.

An average describes a crowd

The structural problem is simpler than it looks. A rate is a property of a population, and the question is about people.

Seventeen per hundred describes every Medicare discharge in the country. It does not tell you which of the patients your hospital sent home yesterday are drifting toward a return, and it never will, because that is not what an average is for. You can risk adjust it, trend it, break it out by service line and by payor, and at the end of all that work you have a beautifully characterized crowd and not one name.

A forecast would name names. A rate is a report card.

Risk does not spread out

The second thing worth holding onto is that risk is not evenly distributed, and it is not close.

AHRQ's most recent expenditure analysis found that in 2022 the top 5 percent of people ranked by healthcare spending accounted for 49.7 percent of all spending. The top 1 percent alone accounted for 21.7 percent. Half the money, five percent of the people.

That shape turns up everywhere anyone bothers to look for it. A small group generates a disproportionate share of admissions, of cost, and of exactly the events a care management program exists to prevent. It is close to the most reliable finding in population health, and nearly every program nods at it and then organizes itself as though it were not true.

The half of this nobody writes about

Concentration is the part that gets published. The other half is capacity, and capacity is what actually decides what happens on Monday.

A coordinator has a finite number of hours and a decidedly non-finite list. Published time and motion work gives a sense of what a single contact really costs. A study of telephone depression care management in the American Journal of Managed Care measured every component of a care manager's day and found that one successful follow up call consumed 51 to 60 minutes of total staff time once preparation, documentation, and communication back to the provider were counted. The conversation itself was the smaller part. An initial assessment ran 75 to 95 minutes. Unsuccessful attempts, which are most attempts on most afternoons, cost 9 to 11 minutes each and produce nothing at all.

That same study put sustainable panel size at roughly 143 to 165 patients per care manager per quarter.

Hold those two figures next to each other. Roughly an hour of real work per meaningful contact, against a ceiling near 150 patients across thirteen weeks. That is not a team that reaches everyone. That is a team that reaches each person on the list two or three times in three months, assuming the list is the right length and nothing goes sideways. The program studied was specific, and your numbers will be different. The shape will not be.

Now go count how many people are on your list.

What blanket outreach actually does

Put the two halves together, and the failure sitting inside most care management programs comes into focus.

If risk concentrates and outreach does not, the program spends its scarcest resource at a flat rate across a population where need is anything but flat. Everyone discharged gets the same call in the same window from the same script. That sounds equitable. In practice, it means the one call that would have changed an outcome went out carrying precisely the same urgency as the ninety that would not have.

The team is not failing here. The team is faithfully executing a plan that mistakes an average for a distribution.

Prioritization is arithmetic, not preference

This is the part people resist, and I think they should sit with it anyway.

When the panel is larger than the team, you are already prioritizing. You do it by admission date, or by whichever chart surfaced first, or by which referral happened to arrive with a phone call attached. Those are rankings. Nobody chose them, and nobody would defend them out loud, but they are running your program right now.

The only real question is whether the ranking is deliberate.

That is not a philosophical position, it is division. A finite number of contacts spread across a larger number of patients produces an order whether or not anyone writes it down. Choosing not to rank is still a ranking, and it is reliably worse than one you would have chosen on purpose.

Two numbers worth more than the rate

A program serious about this should watch two things that say more about it than the readmission rate ever will.

Time to first contact for the highest risk decile. Not the average time to contact across everyone, which blends the urgent with the routine and hides the only thing you need to see. Just the top ten percent, measured from the moment they were identified.

And the share of high risk patients actually reached within seven days. Reached, not attempted. Those 9 to 11 minutes per failed attempt are real time the program spent, and counting attempts as outreach is how a team ends up looking productive while the patients who mattered most were never spoken to.

Neither measure needs data you do not already have.

The list is the deliverable

Ranking a panel well is harder than it sounds, and I will say plainly that this is the difficult part. It takes more than a risk score, because a score tells you who is likely to have an event, not who you can still do something about. A patient with advancing disease and a patient with four new medications and no scheduled follow up can carry an identical number and represent completely different opportunities. Your team can help one of them this week.

That gap is what we are building toward. Preventra is the one member of the team who has read everything: every physician's note, every medication change, every follow up that was or wasn't scheduled, pulled together around one patient. From that, with as much precision as the data allows, it surfaces who needs attention at discharge and in the weeks after, and why. Not a verdict. A clearer starting point for the clinicians who make the call.

Last week was National Case Management Week, and this year's theme was Driving Outcomes. So here is what I would offer. The people doing that work are not short on commitment, and they are not short on skill. They are short on hours, and they have been handed a panel that arithmetic says cannot be carried evenly. The most useful thing an organization can give them is not another dashboard. It is a shorter list, in a better order.

Which patient needs your team's attention today? Someone on your list is going to find out the hard way whether you could answer that.