Release Notes:
Sampling Chapter TJC
Version 2012B
Population and Sampling Specifications
Introduction
Population
Defining the population is the first step to estimate a hospital’s performance. A population is generally defined as a collection of patients sharing a common set of universally measured characteristics, such as an ICD-9-CM Principal Diagnosis or Procedure Code. The Initial Patient Population and diagnosis codes meet this description for the national quality measures. For the purpose of measuring national quality core measures, the term “Initial Patient Population” is defined below:
- An “Initial Patient Population” refers to all patients (Medicare and non-Medicare) who share a common set of specified, administratively derived data elements. This may include ICD-9-CM diagnosis codes or other population characteristics such as age. For example, the population for the HBIPS discharge measures (e.g., HBIPS-1, 4, 5, 6, and 7) includes all patients having a principal or secondary psychiatric diagnosis code from Appendix A, Table 10.1.
Cases identified as being in the Initial Patient Population for the measure set, or strata (e.g., HBIPS), or sampling group (e.g., PC) are eligible to be sampled. For the definition of the Initial Patient Population(s) for each measure set, refer to the appropriate Initial Patient Population discussion in the Measure Information section of this manual.
Sampling
Sampling is a process of selecting a representative part of a population in order to estimate the hospital’s performance, without collecting data for its entire population. Using a statistically valid sample, a hospital can measure its performance in an effective and efficient manner. Sampling is a particularly useful technique for performance measures that require primary data collection from a source such as the medical record. Sampling should not be used unless the hospital has a large number of cases in the Initial Patient Population because a fairly large number of sample cases are needed to achieve a representative sample of the population. For the purpose of sampling national quality measures, the terms “sample” and “case” are defined as below:
- The “sample” is the fraction of the population that is selected for further study.
- A “case” refers to a single record (or an episode of care [EOC] or event) within the population. For example, during the first quarter a hospital may have 100 patients who had principal or secondary psychiatric diagnosis code associated to the HBIPS-1, 4, 5, 6, and 7 measures. The hospital’s Initial Patient Population would include 100 cases or 100 patient records for these measures during the first quarter.
To obtain statistically valid sample data, the sample size should be carefully determined and the sample cases should be randomly selected in such a way that the individual cases in the population have an equal chance of being selected. Only when the sample data truly represent the whole population can the sample-based performance measure data be meaningful and useful.
Each hospital is ultimately responsible that sampling techniques applied for their hospital adhere to the sampling requirements outlined in this manual. ORYX® Vendors are responsible for ensuring that the sampling techniques are applied consistently across their client hospitals.
Sampling is done by national quality inpatient measure set; however, for Perinatal Care (PC), and Hospital-Based Inpatient Psychiatric Services (HBIPS) are done by strata or sampling group. For measures requiring medical record abstraction, sampling must be done using available databases that contain all discharges for the transmission quarter.
Note:
Hospitals are NOT required to sample their data. If sampling offers minimal benefit (i.e., a hospital has 80 cases for the quarter and must select a sample of 76 cases) the hospital may choose to use all cases.
Order of Data Flow
The sampling methodology defined in the Specifications Manual for National Hospital Inpatient Quality Measures requires hospitals submitting data for any measure set that utilizes the Global Initial Patient Population to use the associated sampling methodology for all measure sets being submitted. This requirement includes the PC and HBIPS measure sets.
- If the hospital is submitting data to both the QIO Clinical Warehouse and the Joint Commission’s Data Warehouse, use sampling methodology number one.
- If the hospital is submitting data to only The Joint Commission:
- If the hospital is submitting at least one measure set that uses the Global Initial Patient Population, use sampling methodology number one.
- If the hospital is not submitting any of the measure sets that uses the Global Initial Patient Population, sample each measure set independently using sampling methodology number two.
1. Hospitals Submitting Measure Sets Under the Global Initial Patient Population to Both the QIO Clinical Warehouse and The Joint Commission’s Data Warehouse
For the submission of the Global Initial Patient Population and associated measure sets (i.e., ED, IMM, TOB, and/or SUB) the following data flow or process steps should be used to identify the data for all measure sets or stratum that are transmitted to the Joint Commission’s Data Warehouse, including PC and HBIPS. These process steps are:
Identify Global Cases To Be Abstracted (ED, IMM, SUB, TOB)
- Identify the Global Initial Patient Population. The Global Initial Patient Population is used for the ED, IMM, TOB, and SUB measure sets. This data pull utilizes administrative data such as admission date and discharge date. This identification process must be completed prior to the application of data integrity filter, measure exclusions, and the application of sampling methodology. For specific Global Initial Patient Population definitions, refer to the Global Initial Patient Population discussion in the Measure Information section of the Specifications Manual for National Hospital Inpatient Quality Measures. This data pull is completed once for each hospital. This is not performed for each measure set that utilizes the Global population.
- If the hospital is sampling, use the Global Initial Patient Population identified above and pull the sample of medical records for the ED, IMM, TOB, and/or SUB measure sets using the Sample Size Requirements defined in the Global Initial Patient Population Information section of the Specifications Manual for National Hospital Inpatient Quality Measures. Note: This is completed once for each hospital. This is not performed for each measure set that utilizes the Global population.
- Collect or abstract from the identified medical records the general and measure specific data elements that are needed for the measure set. Run the data through the algorithms for the measure sets under the Global Initial Patient Population (ED, IMM, SUB and/or TOB). The count of the number of cases used in this step is collected in the Global Initial Patient Population and Sample Size data elements.
- If the hospital is only submitting the measure sets under the Global Initial Patient Population (i.e., ED, IMM, SUB or TOB), the process is complete.
Identify Cases To Be Abstracted For The Remaining Measure Sets, Strata, and Sub-populations (PC, HBIPS, AMI, CAC, HF, PN, SCIP, STK, VTE)
- Identify the Initial Patient Population for the other measure sets (AMI, HF, PN, STK) strata or sub-populations (PC, HBIPS, SCIP, CAC, VTE). This data pull utilizes administrative data such as ICD-9-CM diagnosis and procedure codes, admission date, and birthdate. All ICD-9-CM diagnosis and procedure codes included in the Initial Patient Population definition must be applied. This identification process must be completed prior to the application of data integrity filter, measure exclusions, and the application of sampling methodology. For specific PC and HBIPS measure set definitions, refer to the Initial Patient Population discussion in the Measure Information section of this manual. For all other measures, refer to the Initial Patient Population discussion in the Measure Information section of the Specifications Manual for National Hospital Inpatient Quality Measures. The number of cases in the Initial Patient Population of each measure set, strata, and sub-population are collected in the appropriate Initial Patient Population Size data elements.
- If the hospital is not sampling, collect or abstract from the identified medical records the general and measure specific data elements that are needed for the measure set(s), strata or sub-populations. The count of the number of cases used in this step is collected in the Sample Size data elements.
- If the hospital is sampling, use the Initial Patient Population (N) identified above and pull the sample of medical records for the measure set, strata or sub-population using the “Sample Size Requirements” in the appropriate sampling discussion in the Measure Information section of this manual for PC and HBIPS. For all other measures, refer to the “Sample Size Requirements” in the appropriate sampling discussion in the Measure Information section of the Specifications Manual for National Hospital Inpatient Quality Measures.
- Using the Global Initial Patient Population identified above, identify and count the number of cases that are also in the other Measure Sets (e.g., AMI, HF, PN, and STK), strata, or sub-populations (e.g., PC, HBIPS, SCIP, CAC, or VTE) Initial Patient Population(s).Determine the number of cases that need to be sampled (n) from the cases in the other measure set(s) or stratum(s) Initial Patient Population (N). Use the “Sample Size Requirements” in the appropriate sampling discussion in the Measure Information section of this manual for PC and HBIPS. For all other measures, refer to the “Sample Size Requirements” in the appropriate sampling discussion in the Measure Information section of the Specifications Manual for National Hospital Inpatient Quality Measures.
- If there are enough Initial Patient Population cases in the Global sample pull to meet the specific initial patient population and sampling requirements for the measure set(s), strata, or sub-populations, then no additional sampling is required. Collect or abstract from the identified medical records the general and measure specific data elements that are needed for the measure set(s), strata, or sub-populations. The count of the number of cases used in this step is collected in the Sample Size data elements.
- If there are not enough Initial Patient Population cases in the Global sample pull to meet the specific initial patient population and sampling requirements for the measure set(s), strata or sub-populations, complete the sample by pulling additional cases from the other measure set(s), strata or sub-populations Initial Patient Population(s). Use the “Sample Size Requirements” in the appropriate Sampling discussion in the Measure Information section of this manual for PC and HBIPS. For all other measures, refer to the “Sample Size Requirements” in the appropriate sampling discussion in the Measure Information section of the Specifications Manual for National Hospital Inpatient Quality Measures. Collect or abstract from the identified medical records the general and measure specific data elements that are needed for the measure set(s). The count of the number of cases used in this step is collected in the Sample Size data elements.
Example:
For 4th quarter the Global Initial Patient Population is 1550, 100 for AMI, 40 for HBIPS discharges, 250 for PC-Mothers, and 300 for PC-Newborns with Breast Feeding. If the hospital is sampling, the minimum number of cases that would be required to be sampled would be 306 for Global (ED, IMM, TOB, and/or SUB), 78 for AMI, 40 for HBIPS discharges, 75 for PC-Mothers, and 37 for PC-Newborns with Breast Feeding.
The hospital would pull 306 cases for the Global sample. From those 306 cases the hospital would determine how many of those cases were also AMI, HBIPS discharges, PC-Mothers, PC-Newborns with Breast Feeding cases that met the initial patient population criteria for the specific measure set. If there are enough AMI, HBIPS discharges, PC-Mothers, and PC-Newborns with Breast Feeding cases in the Global sample pull to meet the minimum sampling requirements for those measure sets, then no additional sample pull is needed.
If there are not enough cases in the Global sample pull to meet the other measure sets minimum sampling requirements then an additional sample pull is needed. For example, from the Global sample pull there were 72 AMI, 40 HBIPS discharges, 20 PC-Mothers, and 5 PC-Newborns with Breast Feeding cases identified that met the initial population criteria for the specific measure set. As the minimum sample requirements for AMI is 78, 6 additional AMI cases would need to be pulled from the AMI Initial Patient Population. No additional cases would need to be pulled for HBIPS discharges as the minimum number of cases required is 40. As the minimum sample requirements for PC-Mothers is 75, an additional 55 cases would need to be pulled from the PC-Mothers Initial Patient Population. As the minimum sample requirements for PC-Newborns with Breast Feeding is 37, an additional 32 cases would need to be pulled from the PC-Newborns with Breast Feeding Initial Patient Population.
Note: PC-Newborns with BSI and HBIPS restraint and seclusion events are not eligible for sampling and will use the entire Initial Patient sampling group, as appropriate, for reporting.
2. Hospitals Not Submitting the Measure Sets Under the Global Initial Patient Population to The Joint Commission Only
For hospitals not submitting the measure sets under the Global Initial Patient Population to The Joint Commission only, an independent sample pull should be used to pull the sample for the applicable measure sets (i.e., PC, HBIPS, AMI, CAC, HF, PN, SCIP, STK and VTE).
Each measure set, stratum, or sub-population has a unique definition of Initial Patient Population and sample size requirement. However, the same data flow or process steps can be used to identify the data that is transmitted to the Joint Commission's Data Warehouse. These process steps are:
- First, identify the Initial Patient Population for the measure set. An Initial Patient Population is defined for each measure set, stratum, and sampling group and the count is collected in the Initial Patient Population Size data elements.
All data elements in the appropriate Initial Patient Population definition, including ICD-9-CM Diagnosis Codes when appropriate, must be applied. This identification process must be completed prior to the application of data integrity filter, measure exclusions, and the application of sampling methodology.
For specific measure set, stratum, and sampling group definitions, refers to the appropriate Initial Patient Population discussion in the Measure Information section of this manual for PC and HBIPS. For all other measures, refer to the “Sample Size Requirements” in the appropriate sampling discussion in the Measure Information section of the Specifications Manual for National Hospital Inpatient Quality Measures.
- Second, if the measure allows sampling and the hospital is sampling, use the Initial Patient Population identified above and pull the sample of medical records for each measure set, stratum, or sub-population using the Sample Size Requirements defined in the appropriate Measure Information section of this manual for PC and HBIPS. For all other measures, refer to the “Sample Size Requirements” in the appropriate sampling discussion in the Measure Information section of the Specifications Manual for National Hospital Inpatient Quality Measures.
- Third, collect or abstract from the identified medical records the general and measure specific data elements that are needed for the measure set. The count of the number of cases used in this step is collected in the Sample Size data elements.
- If the hospital is not sampling, use the medical records identified in the first data pull.
- If the measure allows sampling and the hospital is sampling, use the medical records from the cases in the identified sample.
Sample Size Requirements
Hospitals that choose to sample have the option of sampling quarterly or sampling monthly. The sample size requirements for each of these options are described in turn. Hospitals need to use the next highest whole number when determining their required sample size. See below for rounding examples. For each measure sets sample size requirements, refer to the appropriate measure set’s Measure Information section in this manual.
Hospitals selecting sample cases for measure sets that are not stratified must ensure that it's Initial Patient Population(s) and sample size(s) meet the conditions stated in the measure set’s Sample Size Requirements.
For hospitals selecting sample cases for stratified measure sets or measure sets with sampling groups (e.g., HBIPS and PC), a modified sampling procedure is required. Hospitals selecting samples cases for these sets must ensure that each individual stratum's Population/sampling group and sample size meets the conditions stated in the measure set's Sample Size Requirements.
Regardless of the option used, hospital samples must be monitored to ensure that sampling procedures consistently produce statistically valid and useful data. Due to exclusions and contraindications, hospitals selecting sample cases MUST submit AT LEAST the minimum required sample size. The sample size tables for each option automatically build the number of cases needed to obtain the required sample sizes.
Hospitals that sample, should sample by their Joint Commission’s Health Care Organization Identifier. All data that are sampled must be transmitted to The Joint Commission.
A hospital may choose to use a larger sample size than is required. Hospitals whose Initial Patient Population size is less than the minimum number of cases per quarter/month for the stratum cannot sample. For the Discharge measures (e.g., HBIPS-1, 4, 5, 6, 7, and PC), hospitals that have five or fewer discharges (both Medicare and non-Medicare combined) are not required to submit patient level data to the Joint Commission's Data Warehouse. For the event measures (e.g., HBIPS-2, and 3), hospitals must submit patient level data to The Joint Commission regardless of the number of discharges or events they have each quarter. Refer to the Sample Size Requirement tables provided in the Measure Information section to determine the minimum number of cases that need to be sampled for each HBIPS measure set.
Quarterly Sampling Examples
Quarterly Example 1: Measure set is Not Stratified
Hospitals selecting sample cases for measure set ABC, which is not stratified, must ensure that its Initial Patient Population and quarterly sample size meet the following conditions:
Quarterly Sample Size
Based on Initial Patient Population for the ABC Measure set
Examples
- A hospital’s ABC Initial Patient Population is 77 patients during the first quarter. Using the above table, no sampling is allowed – 100% of the population is required.
- A hospital’s ABC Initial Patient Population is 100 patients during the second quarter. Using the above table, the required sample size is seen to be a minimum of 78 ABC patients for this quarter.
- A hospital’s ABC Initial Patient Population is 401 patients during the third quarter. Using the above table, the required sample size is seen to be 20% of the population, or 81 cases for the quarter (twenty percent of 401 equals 80.2 rounded to the next whole number = 81).
- A hospital’s ABC Initial Patient Population is 5 patients during the first quarters. Using the above table, submission of patient level data is not required. If the hospital chooses to submit patient level data, the required quarterly sample size would be 100% of the patient population or 5 cases for the quarter.
Quarterly Example 2: Measure set is stratified
For hospitals selecting sample cases for measure set XYZ which contains 8 strata, a modified sampling procedure is required. Hospitals selecting sample cases for these sets must ensure that each individual stratum’s population and quarterly sample size meets the following conditions.
- Select within each of the seven individual measure stratum and the 8th XYZ stratum.
Quarterly Sample Size
Based on Initial Patient Population for the XYZ measure set
Example
- The XYZ Initial Patient Population sizes for a hospital is 5, 50, 15, 140, 35, 201, 3, and 481 patients respectively per stratum for the quarter. Since the total Initial Patient Population for XYZ is 930, the hospital must submit patient level data. The required quarterly sample sizes for each stratum would be 5, 16, 15, 16, 16, 21, 3, and 48.
- The 1st, 3rd, and 7th strata are less than the minimum required quarterly sample size, so 100% of each of these strata are sampled.
- The 2nd, 4th, and 5th strata each require 16 cases to be sampled.
- The 6th stratum has 201 patients per quarter, which requires a 10% sample size, or 21 cases (twenty percent of 201 equals 20.1 rounded to the next whole number = 21).
- The 8th stratum is more than the maximum required quarterly sample size, so this stratum requires 48 cases to be sampled.
- The XYZ Initial Patient Population sizes for a hospital 1, 1, 0, 0, 1, 0, 1, and 1 patients respectively per stratum for the quarter. Since the total Initial Patient Population for XYZ is 5, the hospital may choose to not submit patient level data. If the hospital chooses to submit patient level data, the required quarterly sample sizes for each stratum would be 1, 1, 0, 0, 1, 0, 1, and 1.
- The 1st, 2nd, 5th, 7th, and 8th strata are less than the minimum required quarterly sample size, so 100% of each of these strata are sampled.
- There is no data to sample for the 3rd, 4th, and 6th strata.
Quarterly Example 3: Measure set has sub-populations
For hospitals selecting sample cases for measure set DEF which contains 3 independent sub-populations a modified sampling procedure is required. The three sub-populations must be sampled independently from each other.
1-Hospitals selecting sample cases for sub-population 1 must ensure that the Initial Patient Population and sample size for the sub-population 1 meet the following conditions:
Quarterly Sample Size
Based on Initial Patient Population for the Patient Sub-Population 1
2 - Hospitals selecting sample cases for sub-population 2 must ensure that the initial Patient Population and sample size for sub-population 2 meet the following conditions:
Quarterly Sample Size
Based on Initial Patient Population Size for the
Patient Sub-Population 2
3 - Sub-population 3 is not eligible for sampling and will use the entire Initial Patient Population for reporting.
Example
1. Quarterly sampling for sub-population 1:
- A hospital's sub-population 1 is 752 during the second quarter. Using the quarterly sampling table for sub-population 1, the sample size required is 20% of this sub-population, or 151 cases for the quarter (twenty percent of 752 equals 150.4 rounded up to the next whole number = 151).
- A hospital's sub-population 1 is 5 during the first quarter. Using the quarterly sampling table for sub-population 1, the sample size is less than the minimum required quarterly sample size, so 100% of this sub-population is sampled.
- A hospital's sub-population 1 is 99 during the third quarter. The required quarterly sample is 45 cases.
2. Quarterly sampling for sub-population 2:
- A hospital's sub-population 2 is 511 during the second quarter. Using the quarterly sampling table for sub-population 2, the sample size required is 20% of this sub-population, or 103 cases for the quarter (twenty percent of 511 equals 102.2 rounded up to the next whole number = 103).
- A hospital's sub-population 2 is 3 during the first quarter. Using the quarterly sampling table for sub-population 2, the sample size is less than the minimum required quarterly sample size, so 100% of this sub-population is sampled.
- A hospital's sub-population 2 is 300 during the third quarter. The required quarterly sample is 90 cases.
3. Quarterly sampling for sub-population 3:
- Sub-population is not eligible for sampling and will use the entire initial Patient Sub-Population for reporting.
Quarterly Example 4: Measure set has Sampling Groups
For hospitals selecting sample cases for measure set HGI which contains 3 independent sampling groups a modified sampling procedure is required. The three sampling groups are sampled independently from each other. A patient falls into multiple sampling groups but may not actually be sampled for all the groups for chich the patient is eligible.
1-Hospitals selecting sample cases for sampling group 1 must ensure that the Initial Patient Population and sample size for the sampling group 1 meet the following conditions:
Quarterly Sample Size
Based on Initial Patient Population Size for the
Patient Sampling Group 1
2 - Hospitals selecting sample cases for sampling group 2 must ensure that the initial Patient Population and sample size for the sampling group 2 meet the following conditions:
Quarterly Sample Size
Based on Initial Patient Population Size for the
Patient Sampling Group 2
3 - Hospitals selecting sample cases for sampling group 3 must ensure that the Initial Patient Population and sample size for the sampling group 3 meet the following conditions:
Quarterly Sample Size
Based on Initial Patient Population Size for the
Patient Sampling Group 3
Example
- A Hospital's sampling group 1 size 347 during the second quarter. The required sample size is 20% of the patient population or 70 cases for the quarter (twenty percent of 347 equals 69.4 rounded up to the next highest whole number is 70.)
- A Hospital's sampling group 2 size is 250 patients during the second quarter. The required sample size is seen to be 100 patients for this quarter.
- A Hospital's sampling group 3 size is 700 patients during the second quarter. The required sample size is seen to be 140 patients for this quarter.
Monthly Sampling Examples
Monthly Example 1: Measure set is Not Stratified
Hospitals selecting sample cases for ABC measure set must ensure that its Initial Patient Population and effective monthly sample size meet the following conditions:
Monthly Sample Size
Based on Initial Patient Population for the ABC measure set
Examples
- A hospital’s ABC Initial Patient Population is 25 patients during January. Using the above table, no sampling is allowed – 100% of the population is required.
- A hospital’s ABC Initial Patient Population is 130 patients during February. Using the above table, the required sample size is seen to be a minimum of 26 ABC patients for this month.
- A hospital’s ABC Initial Patient Population is 301 patients during March. Using the above table, the required sample size is seen to be 20% of the population, or 61 cases for the month (twenty percent of 301 equals = 60.2 rounded to the next whole number = 61.
- A hospital’s ABC Initial Patient Population is 516 patients during April. Using the above table, the required sample size is seen to be a minimum of 104 ABC patients for this month.
Monthly Example 2: Measure set is Stratified
For hospitals selecting sample cases for the XYZ measure set, a modified sampling procedure is required. Hospitals selecting sample cases for this set must ensure that each individual strata population and effective monthly sample size meets the following conditions:
- Select within each of the seven individual measure stratum and the 8th XYZ stratum.
Monthly Sample Size
Based on Initial Patient Population for the XYZ measure set
Example
- The XYZ Initial Patient Population sizes for a hospital is 5, 50, 15, 141, 35, 201, 3, and 481 patients respectively in June. The required monthly sample sizes would be 5, 6, 6, 15, 6, 16, 3, and 16.
- The 1st and 7th strata are less than the minimum required monthly sample size, so 100% of each of these strata are sampled.
- The 2nd, 3rd, and 5th strata each require 6 cases to be sampled.
- The 4th stratum has 141 patients per month, which requires a 10% sample size, or 15 cases (twenty percent of 141 equals 14.1 rounded to the next whole number = 15).
- The 6th and 8th strata are each more than the maximum required monthly sample size, so this stratum requires 16 cases to be sampled.
Monthly Example 3: Measure set has sub-populations
For hospitals selecting sample cases for measure set DEF which contains 3 independent sub-populations a modified sampling procedure is required. The three sub-populations must be sampled independently from each other.
1 - Hospitals selecting sample cases for sub-population 1 must ensure that the Initial Patient Population and sample size for sub-population and sample size for sub-population 1 meet the followind conditions:
Monthly Sample Size
Based on Initial Patient Population Size for the
Patient Sub-Population 1
2 - Hospitals selecting sample cases for sub=population 2 must ensure that the Initial Patient Population and sample size for sub-population and sample size for sub-population 2 meet the following conditions:
Monthly Sample Size
Based on Initial Patient Population Size for the
Patient Sub-Population 2
3 - Sub-population 3 is not eligible for sampling and will use the entire Initial Patient Sub-Population for reporting.
Example
1. Monthly sampling for sub-population 1:
- A hospital's sub-population 1 is 81 during March. Using the monthly sampling table for sub-population 1, the sample size required is 20% of this sub-population, or 17 cases for the month (twenty percent of 81 equals 16.2 rounded up to the next whole number = 17).
- A hospital's sub-population 1 is 5 during February. Using the monthly sampling table for sub-population 1, the sample size is less than the minimum required monthly sample size, so 100% of this sub-population is sampled.
- A hospital's sub-population 1 is 45 during January. The required monthly sample is 15 cases.
2. Monthly sampling for sub-population 2:
- A hospital's sub-population is 387 during March. Using the monthly sampling table for sub-population 2, the sample size required is 20% of this sub-population, or 78 cases for the month (twenty percent of 387 equals 77.4 rounded up to the next whole number = 78).
- A hospital's sub-population 2 is 3 during February. Using the monthly sampling table for sub-population 2, the sample size is less than the minimum required monthly sample size, so 100% of this sub-population is sampled.
- A hospital's sub-population 2 is 47 during January. The required monthly sample is 30 cases.
3. Monthly sampling for sub-population 3:
- Sub-population 3 is not eligible for sampling and will use the entire initial Patient Sub-Population for reporting.
Monthly Example 4: Measure set has Sampling Groups
1 - Hospitals selecting sample cases for sampling group 1 must ensure that the Initial Patient Population and sample size for the sampling group 1 meet the following conditions:
Monthly Sample Size
Based on Initial Patient Population Size for the
Patient Sampling Group 1
2 - Hospitals selecting sample cases for sampling group 2 must ensure that the initial Patient Population and sample size for the sampling group 2 meet the following conditions:
Monthly Sample Size
Based on Initial Patient Population Size for the
Patient Sampling Group 2
3 - Hospitals selecting sample cases for sampling group 3 must ensure that the Initial Patient Population and sample size for the sampling group 3 meet the following conditions:
Monthly Sample Size
Based on Initial Patient Population Size for the
Patient Sampling Group 3
Example
1. Monthly sampling for sampling group 1:
- A hospital's sampling group 1 is 81 during March. Using the monthly sampling table for sampling group 1, the sample size required is 20% of this sampling group, or 17 cases for the month (twenty percent of 81 equals 16.2 rounded up to the next whole number = 17).
- A hospital's sampling group 1 is 5 during February. Using the monthly sampling table for sub-population 1, the sample size is less than the minimum required monthly sample size, so 100% of this sampling group is sampled.
- A hospital's sampling group 1 is 45 during January. The required monthly sample is 10 cases.
2. Monthly sampling for sampling group 2:
- A hospital's sampling group is 2 is 387 during March. Using the monthly sampling table for sampling group 2, the sample size required is 20% of this sampling group, or 78 cases for the month (twenty percent of 387 equals 77.4 rounded up to the next whole number = 78).
- A hospital's sampling group 2 is 3 during February. Using the monthly sampling table for sampling group 2, the sample size is less than the minimum required monthly sample size, so 100% of this sampling group is sampled.
- A hospital's sampling group 2 is 47 during January. The required monthly sample is 25 cases.
3. Monthly sampling for sampling group 3:
- A hospital's sampling group 3 is 125 during January. The required monthly sample is 25 cases.
Sampling Approaches
As previously stated in this section, hospitals have the option to sample from their population, or submit their entire population. Hospitals that choose to sample must ensure that the sampled data represent their Initial Patient Population by using either the simple random sampling or systematic random sampling methods and that the sampling techniques are applied consistently within a quarter. For example, monthly samples for a measure set, stratum, or sampling group must use consistent sampling techniques across the quarterly submission period.
- Simple random sampling - selecting a sample size (n) from a population of size (N) in such a way that every case has the same chance of being selected.
- Systematic random sampling - selecting every kth record from a population of size N in such a way that a sample size of n is obtained, where k ≤ N/n. The first sample record (i.e., the starting point) must be randomly selected before taking every kth record. This is a two-step process: a) Randomly select the starting point by choosing a number between one and k using a table of random numbers or a computer-generated random number; and a) Then select every kth record thereafter until the selection of the sample size is completed.
Each hospital is ultimately responsible that sampling techniques applied for their hospital adhere to the sampling requirements outlined in this manual. ORYX Vendors are responsible for ensuring that the sampling techniques are applied consistently across their client hospitals.
Sampling Approach Examples
For a hospital with an Initial Patient Population size of 350 ABC measure set discharges per quarter, the sample size would be 78. To select a random sample of 78 ABC patients:
- Simple random sampling:
- Generate random numbers for individual ABC patient records from a random number function using a statistical software package or computer programming language.
- Sort data by the random numbers either in an increasing or decreasing order.
- Select the first 78 ABC patient records as the random sample.
- Systematic random sampling:
- In this example, the hospital’s Initial Patient Population size= 350 and the sample size = 78. Divide the Initial Patient Population size by the sample size and take the quotient (i.e., the integer portion) as the sampling interval k. The sampling interval k = 350/78 = 4.5. Thus, every 4th ABC patient record will be selected from the Initial Patient Population until 78 cases are selected.
- To ensure that each ABC patient has an equal chance of being selected, the “starting point” must be randomly determined before selecting every 4th ABC patient record. This can be done using a computer random number generator or a random number table to randomly choose a number between 1 and 4 as the starting point.
Transmission of Initial Patient Population and Sample Data Elements
The Joint Commission requires transmission of Initial Patient Population and sample count data. Transmission of Initial Patient Population and sample count data elements are used to assist in evaluating completeness of submission in accordance with The Joint Commission sampling requirements.
The Initial Patient Population Size refers to all patients (Medicare and non-Medicare) who share common payment sources which can be identified by utilizing administrative data such as the UB-04. All ICD-9-CM diagnosis and procedure codes included in the appropriate Initial Patient Population definition must be applied. This identification process must be completed prior to the application of data integrity filter, measure exclusions, and the application of sampling methodology. For specific measure set and strata definitions, refer to the appropriate Initial Patient Population discussion in the Measure Information section of this manual.
The Initial Patient Population and sample data elements are:
- ICD Population Size *
- Initial Patient Population Size – Medicare Only **
- Initial Patient Population Size – Non-Medicare Only **
- Sample *
- Sampling Frequency **
- Sample Size – Medicare Only **
- Sample Size – Non-Medicare Only **
Sample indicates whether or not the hospital has sampled data for the specified time period. Sampling Frequency indicates if the hospital has sampled using the monthly or quarterly methodology, or whether the entire population was used for the specified time period or the hospital had five or fewer discharges for the discharge quarter and did not submit patient level data.
Initial Patient Population Size – Medicare Only includes all patients that are billed under Medicare or Title 18. Medicare can be listed as a primary, secondary, tertiary or lower on the list of payment sources for the patient. In addition, patients who are participating as a member of a Medicare HMO/Medicare Advantage are included in the Medicare counts, e.g., Medicare Blue, Humana gold, Secure Horizons, AARP, Coventry Advantra, etc.
* Transmitted in the aggregate data file. Refer to the ORYX Technical Implementation Guide for more information.
** Transmitted in the Hospital Initial Patient Population data file. Refer to the Hospital Initial Patient Population Data XML File Layout in the Transmission section of this manual.
Initial Patient Population and Sample Size Examples
Example 1 – Hospital does not sample
A hospital uses the Initial Patient Population(s) for the ABC measure set to identify 120 cases in the ABC Initial Patient Population during the second quarter. The hospital does not sample the ABC measure set, so data for all 120 cases are collected and used to calculate the hospital’s rate for each ABC measure. 40 of the 120 cases in the ABC Initial Patient Population are Medicare patients.
Note: Sampling Frequency = 3 (not sampling) is the only valid value for HBIPS event measures (HBIPS-2 and 3).
The breakdown of data by month and Medicare / Non-Medicare is:
The following is transmitted for each month in the quarter:
Example 2 – Hospital samples monthly
A hospital uses the Initial Patient Population(s) for the ABC measure set to identify 120 cases in the ABC Initial Patient Population during the second quarter. From these 120 cases, the hospital uses the monthly sample size requirements and randomly selects a sample of 26 cases for each month. Data for these 26 cases are collected and used to calculate the hospital’s rate for each ABC measure. 40 of the 120 cases in the ABC Initial Patient Population are Medicare patients and 24 of these cases were included in the sample.
Note: Sampling Frequency = 1 (sampling data monthly) is not valid for HBIPS event measures (HBIPS-2 and 3).
The breakdown of data by month and Medicare / Non-Medicare is:
The following is transmitted for each month in the quarter:
Example 3 – Hospital samples quarterly
A hospital uses the Initial Patient Population(s) for the ABC measure set to identify 120 cases in the ABC Initial Patient Population during the second quarter. From these 120 cases, the hospital uses the quarterly sample size requirements and randomly selects a sample of 78 cases. Data for these 78 cases are collected and are then used to calculate the hospital’s rate for each ABC measure. 40 of the 120 cases in the ABC Initial Patient Population are Medicare patients and 20 of these cases were included in the sample.
Note: Sampling Frequency = 2 (sampling data quarterly) is not valid for HBIPS event measures (HBIPS-2 and 3).
The breakdown of data by month and Medicare / Non-Medicare is:
The following is transmitted for each month in the quarter:
Example 4 – Hospital has five or fewer discharges and chooses not to submit patient level data
Note: This example is only valid for the HBIPS Discharge measures (HBIPS-1, 4, 5, 6, 7, and 8). This is not valid for the HBIPS Event measures (HBIPS-2 and 3) since all data must be submitted for these measures regardless of the number of discharges or events that occur during the quarter.
A hospital uses the Initial Patient Population(s) for the ABC measure set to identify 5 cases in the ABC Initial Patient Population for the entire measure set during the second quarter. Since the total Initial Patient Population for ABC is 5, the hospital chooses to not submit patient level data.
Note: Sampling Frequency = 4 (N/A, submission of patient level data is not required) is not valid for HBIPS event measures (HBIPS-2 and 3).
The breakdown of data by month and Medicare / Non-Medicare is:
The following is transmitted for each month in the quarter:
Example 5 – Hospital has five or fewer discharges and chooses to submit patient level data
Note: This example is only valid for the HBIPS Discharge measures (HBIPS-1, 4, 5, 6, 7, and 8). This is not valid for the HBIPS Event measures (HBIPS-2 and 3) since all data must be submitted for these measures regardless of the number of discharges or events that occur during the quarter.
A hospital uses the Initial Patient Population(s) for the ABC measure set to identify 5 cases in the ABC Initial Patient Population for the entire measure set during the second quarter. Even though the total Initial Patient Population for ABC is 5, the hospital chooses to submit patient level data.
Note: Sampling Frequency = 4 (N/A, submission of patient level data is not required) is not valid for HBIPS event measures (e.g., HBIPS-2 and 3).
The breakdown of data by month and Medicare / Non-Medicare is:
The following is transmitted for each month in the quarter: