Overview

For patients treated by a selected physician or hospital who died, we break that population into the following three categories:

• Early Hospice - Patients who were admitted to hospice care prior to their last thirty days of life.
• Late Hospice - Patients admitted to hospice during the last 30 days of their life.
• No Hospice - Patients who were not admitted to hospice.

You can see these three categories on the left side of the pictured table.

For each of these categories, this table provides overall patient counts for each category, and projected counts (per 1000 patients) for each of four different hospital levels of care.

Patients included in the counts for either Early Hospice or Late Hospice could have potentially revoked hospice care, and also been re-admitted to hospice or not. The variety of possible hospitalization scenarios are used to calculate the metrics in the table, but those events do not impact the categorization of Early vs. Late. In this table we only count hospice stays that are the last hospice admission prior to the patient's death.

The difference between Physicians and Hospitals

This table is available for both hospitals and physicians. The metrics contained in the table are the same for either, and so is the value of the insights the table provides. The details for each are in the following table:

Really important note

There are two types of metrics in this table:

• Distinct Patients - Each row has a single column of counts of distinct patients who fit the hospice timing criteria. Any specific patient can only be in one row. These patients are identified by death records from the Social Security Administration.
• Visits per 1000 - All other metrics are calculated to show visits "per 1000" patients. These are not counts of actual visits. We use visit counts from Medicare claims to calculate these metrics.

Usage

Although the image shows metrics for a hospital, the Hospice Timing table for a physician would provide the same use case.

The value of hospice

In light of the cost savings and benefits to patients and families that hospice care provides over hospitals for patients near the end of life, this table verifies the value of hospice care. For each type of hospital visit, the table shows that early admission into hospice care is better than later admission, and that any hospice care is better than no hospice care at all. We will look at the numbers below.

This table makes it almost too easy to prove that getting hospice appropriate patients admitted to hospice care early is preferable. If you want to use these metrics in this way, most of the time you can prove this point using either the physician's metrics or the state averages. 

State metrics for comparison

In addition to metrics specific to the selected provider, the table provides state averages for similar providers. The comparative nature of these metrics allows you to see how your selected provider performs against peers. If your selected provider underperforms or overperforms against peers, you will want to do some careful evaluation of the demographics of the patient population to find out why.

You will want to be careful relaying these metrics to a physician. Although for a high performer you could communicate, "Hey, you're doing great!" it is important to dig in on why the physician's metrics are so good. If the metrics show poor performance against the state averages, you would want to only provide these insights in a context of trust and with a persuasive message of how you can help.

Keep in mind that these metrics also reflect on the care provided by all hospices that treated the provider's patients. This means that, for this table, the metrics do not provide insights into the success of your agency.

 Patient mix between categories

The patient counts are a breakdown of the selected provider's patients who died during the reporting period. 

The sample image of a provider's metrics (above) demonstrates the need. This patients break down thus:

• 114  - early hospice (8.3% of the total)
• 538  - late hospice (39.2% of the total)
• 720 - no hospice (52.5% of the total)

Considering the value of hospice care, these numbers are upside down. The message is clear. This provider needs help to identify hospice appropriate patients and get them admitted to hospice earlier.

Hospital events

There are four hospital events that we track for hospice patients in this table.

• Inpatient Visit - A patient is admitted to a hospital for inpatient care.
• ER-to-Inpatient Visit - a patient is admitted to inpatient care through the emergency room.
• ER Outpatient Visit - a patient is taken to the emergency room, but not admitted for other additional care.
• Observation Visit - a patient stays overnight for observation but is not admitted to inpatient care.

The metrics in this table are based on visit counts, not patient counts. That is, a single patient could be counted multiple times for any level of hospital care, and counted for more than one type of hospital care.

Understanding the Metrics

For the image above, we have moved some pieces around to demonstrate how to understand the metrics. Although we are only viewing one level of care, Inpatient Visits per 1000, the metrics for the other three levels of care work the same way.

Hospice timing

Three timing categories:

• Early Hospice - Patients who were admitted to hospice care prior to their last thirty days of life.
• Late Hospice - Patients admitted to hospice during the last 30 days of their life.
• No Hospice - Patients who were not admitted to hospice.

The metrics in each row pertain to patients for which the listed timing applies. In this case, there is no overlap of patients in multiple rows. A specific patient will only be counted in one row.

Patient Count

These are counts of patients who fit the following criteria for each timing category:

• For Physicians - The count of patients who died during the one-year reporting period and were treated by the physician within six months prior to the death date
• For Hospitals - The count of patients deceased during the most recent 4 quarters, who were admitted to the selected facility up to 6 months prior to death.

The metrics in the rest of the table are based on these counts and counts of visits to the different hospital levels of care. Those visit count are not shown in the table. (However, you can calculate them - see below.)

Why do we use "Visits per 1000"

The metrics in this table are the easiest to use and yet challenging to understand. As a result, when someone asks, "Why use Visits per 1000," there is a simple answer and a complex really mathy answer.

Simple Answer

Trust Trella Health.

Our data scientists spend a lot of time to assure that our metrics are accurate and useful. So it is wise to focus on the purpose of this table. This table shows the value of getting patients into hospice care as early as possible.

If you look at the image below - Inpatient Visits per 1000 for the selected facility (This Facility). We see some numbers; 15, 217, and 371.

If you forget about the "per 1000" for a second and just look at the numbers, we see:

• Deceased patients from the hospital who were admitted to hospice late are almost 15 times as likely to enter inpatient care than those who entered hospice early. (15 X 15 = 225, which is close to 217)
• Deceased patients from the hospital who were not admitted to hospice care at all are 25 times as likely to require inpatient care than those who get early hospice. (15 X 25 = 375, which is close to 371)

Bottom line - Early Hospice is better for the patients, and the numbers prove it!

Keep in mind that the "per 1000" counts don't represent actual patients, they are numbers that we calculate in order to provide a comparison of the three hospice timing scenarios for the selected hospital or physician. Read on if you want to understand why and how we do this.

Lotsa Math

Now let's jump into why we do this.

Imagine that we wanted to figure out the average number of times a group of people go to the grocery store each month. We will start with a small group.

From the image we see the number of trips each person in our group took. We add the trips together and then divide by the number of people in the group.

3 + 2 + 4 + 5 = 14 total trips for groceries in the month

14 / 4 = 3.5 trips - the average number of trips for our group.

If we do the same for a bigger group, the process is the same

3+3+3+2+5+2+4+2+5+3+3+4+4+3+2+4+2+2+3+5+3+5+3+4 = 79

79 / 24 = 3.3 trips/month (rounded from 3.2916667)

Some observations:

• No matter how many people are in our group, the process is the same.
• If one or two or even a few of our shoppers never went to the grocery store, we would still get an average that makes sense.

But consider what happens when we change what we are looking for.

For the same 24 people from the last calculation, let's get an average number of times those same people gave blood at Red Cross in the same month:

Only three people gave blood, so the average is:

3 / 24 or .125 donations per month, or, the average number of donations in a month is 1/8 (one eighth) of a donation. Although the math works, conceptually that is a hard way to think about averages.

Historical Note: For the first pass of this table we included the averages, but often the metrics were so small that most users had difficulty making sense of them. For example, what is an average of .02 visits?

The Solution

An easier way to show the comparative value of this metric is to display an "out of 1000" count. 

So, looking at our table again, we can see that for the 65 patients represented, we have a 15 visits per 1000 patients.

What does this mean?

The way to read this is, based on the patient count and the number of visits (admitted to inpatient care), IF we had 1000 patients in our early hospice group, we would expect that 15 of them would be admitted to inpatient care within the last 30 days of their life.

Finally, some Calculations

The final mystery is, how many actual visits are we talking about. Oddly, we can figure this out.

The "Visits per 1000" count is 15. So we divide 15 by 1000

15 / 1000 = .015

Multiply .015 times 65 - the number of patients from the group of those who would have been admitted to inpatient care. The answer is .975, which we round to one.

Obviously, to calculate the 15 per 1000, we went the other way.

The calculation is:

Here are the steps we followed.

1. Get a count of deceased patient who were admitted to hospice early - 65
2. Count how many of those patients were admitted to Inpatient care within the last 30 days of life - 1 (which we don't show in the table.)
3. Divided the 1 visit by 65 patients 1 visit / 65 patients = .01538 visits per patient.
4. Multiply by 1000 .01538 X 1000 = 15.38
5. Round to the nearest whole number 15

And that is the number we see in the column for visits per 1000 patients.