Throughout Marketscape you will find some counts and performance metrics that do not include actual numbers. This article explains the alternative entries.
Our data is derived from the CMS Virtual Research Data Center. In order to maintain the levels of patient confidentiality required by CMS, all counts and percentages that could reveal a patient count less than 11 will be indicated by <11 in that cell in the table.
- For counts, <11 is used for any patient count less than eleven (11).
- For percentages, this would represent any instances where the calculation includes a patient count that is less that eleven (11) in the denominator. In these cases, the percentage is replaced with "ins." See below.
For entries where the table is sorted and multiple <11 entries are found, the order of those entries does not reveal anything about the relative value of the hidden <11 metrics. We do not receive numbers back from CMS that we then hide; numbers less than eleven in the CMS database are delivered to us with no identifiable number. As far as usage is concerned, the order of <11 entries in a table should not be construed with any conclusion about what the numbers are that are hidden.
"ins" stands for insignificant or insufficient.
Sometimes when we calculate a performance statistic, the final number is too small, or too small statistically to be meaningful. Also, the size of the data pool used in calculating some numbers is too small to provide a meaningful final number. Finally, any calculation calculated from a <11 can't have a value. In all of these cases, we present the metric as "ins."
See Statistical Significance for Solution Display, (below).
In order to maintain privacy suppression, we sometimes need to hide counts in a specific column. Imagine that you have an annual patient count of 100 patients. Then, in the same table, we break out those counts into quarters. Depending on the breakout, we might need to "hide" a count by using "ins". See examples in the following table.
In Example 1, we have no problem showing the quarterly counts since all counts are 11 or more
In Example 2, we need to hide the most recent quarter. If we told you the Q4 count was 20 patients, you would be able to calculate Q3 as 5 patients. In order to maintain patient privacy, we need to hide the count in the most recent quarter.
If you see a "-", this means that the metric in the cell was either not available or not calculated. The hyphen is a null entry.
If you have been using Trella Health products for a while, you might remember that the hyphen was used for zero. That is no longer the case.
If you see a zero (0), this means that the metric was calculated and the count or percent is a zero.
Statistical Significance for Solution Display
Calculated fields exist in Marketscape for which the denominator of the patient population is sometimes known, and sometimes unknown. Per CMS rules, no field can be displayed if the cell value is 10 or less. However, when calculating certain measures, Trella Health increases the limits on the patient denominator to ensure statistical significance. If the statistical significance threshold is not met, a value of “ins” is displayed to denote “insignificant” patient population.
Average Length of Stay (ALOS)
Since ALOS is a metric calculated for a patient population that is solely within the confines of the agency to control, Trella Health displays this measure when the population of patients is >10 patients. Thus, if the number of patients is not <11, the ALOS will be displayed. Otherwise, “ins” is displayed.
Performance metrics are measures that typically involved another provider (e.g., hospitalizations and readmit rates). To ensure anomalies in the data are smoothed and a significant patient population is available, Trella creates these measures using a two-year population of patients. In this case, the threshold is 30 patients, otherwise “Ins” is displayed. This enables 30 patients for a denominator when creating two-year rates, and for one-year (rare) the threshold would be 15 patients. By calculating these rates in this way, we can display low rates by mandating a large denominator for the patient population, and calculating the measure over two years, ensuring the calculation cannot be applied to a patient count of one year.