In order to ensure the accurate collection of all survey data, hospitals/survey vendors administering the HCAHPS survey must develop, implement, and document quality control procedures for all survey administration activities. The HCAHPS decision rules and coding guidelines were developed to address situations in which survey responses are ambiguous, missing or incorrectly provided, and to capture appropriate information for data submission.
Hospitals/Survey vendors must adhere to the following guidelines to ensure valid and consistent coding of such instances.
Mail Surveys
A common problem in mail surveys is ambiguity of responses on returned questionnaires. In order to ensure uniformity in data coding, hospitals/survey vendors must strictly apply the following guidelines. Hospitals/Survey vendors that scan or key-enter mail surveys must employ the following decision rules for resolving common ambiguous situations.
¾ If a mark falls between two response options but is obviously closer to one than the other, then select the choice to which the mark is closest
¾ If a mark falls equidistant between two response options, then code the value of the item as “M – Missing/Don’t Know”
¾ If a value is missing, code as “M – Missing/Don’t Know.” Hospitals/Survey vendors must not impute a response.
¾ When more than one response option is marked, code the value as “M – Missing/Don’t Know”
• Exception: Question 26, “What is your race? Please choose one or more.” For Question 26, enter responses for ALL of the categories that the respondent has selected.
Skip Patterns
There are several items in the HCAHPS survey that can and should be skipped by certain patients. These items form skip patterns. The following decision rules are provided to assist in the coding of patient responses to skip pattern questions.
Four questions in the HCAHPS survey serve as screener questions (Questions 10, 12, 15, and 18) that determine whether the following dependent questions require an answer.
Decision Rules for Screener and Dependent Questions
Decision rules for coding screener questions 10, 12, 15, and 18:
¾ Enter the value provided by the patient. Do not impute a response based on the patient’s answers to the dependent questions.
¾ If the screener question is left blank, code it as “M – Missing/Don’t Know.” Do not impute a response based on the patient’s answers to the dependent questions.
Decision rules for coding dependent questions 11, 13, 14, 16, and 17:
¾ If the corresponding screener question is answered “Yes” and the dependent question is left blank, then code as “M – Missing/Don’t Know”
¾ If the screener question is answered “Yes” and the dependent question is not left blank, then enter the value provided by the patient
¾ If the screener question is answered “No” and the dependent question is left blank, then code as “8 – Not Applicable”
then enter the value provided by the patient
Decision rules for collecting data from dependent questions 19 and 20:
¾ If screener Question 18 is answered “1 – Own home” or “2 – Someone else’s home” and the dependent question is left blank, then code as “ M – Missing/Don’t Know”
¾ If Question 18 is answered “1 – Own home” or “2 – Someone else’s home” and the dependent question is not left blank, then enter the value provided by the patient
¾ If Question 18 is answered “3 – Another health facility” and the dependent question is left blank, then code as “8 – Not Applicable”
¾ If Question 18 is answered “3 – Another health facility” and the dependent question is not left blank, then enter value provided by the patient
In summary, dependent questions that are appropriately skipped are coded as “8 – Not Applicable.” In instances where the patient made an error in the skip pattern, dependent questions are coded with the response provided by the patient. That is, hospitals/survey vendors must not “clean” or correct skip pattern errors returned by a patient. For further information on screener and dependent questions, see Appendix L.
Note: For the telephone and IVR survey modes, skip patterns should be programmed into the electronic telephone interviewing /IVR system. If screener questions 10, 12, 15, and 18 are either “No,” “Another Health Facility” or “Missing/Don’t Know,” then the corresponding dependent questions should be skipped. In these instances, appropriately skipped dependent questions should be coded as “8 – Not Applicable.” For example, if a respondent answers “No” to questions 10 of the HCAHPS questionnaire, the program should skip question 11, and go to question 12. Question 11 should then be coded as “8 – Not Applicable.” Coding may be done automatically by the telephone interviewing/IVR system or later during data preparation.
Header Record
¾ All fields in the Header Record must have a valid value entered with the exception of “NPI,” “DSRS Strata Name,” “DSRS Eligible,” and “DSRS Sample Size” fields. The DSRS fields are required only when “Type of Sampling” is “3 – Disproportionate Stratified Random Sample.”
¾ Once the “Survey Mode” field has been defined for the first month in a quarter, the survey mode for the quarter can be changed by resubmitting this file ONLY if the data files for another month in the quarter have not yet been submitted to My QualityNet
• The “Survey Mode” field must be coded with the approved survey mode for the hospital. For example, if the hospital is using IVR survey mode and have patients who opt to complete the survey by telephone, the “Survey Mode” field must still be coded as “4 – IVR.”
¾ In calculating the “Eligible Discharges” field, the number of eligible discharges in the sample frame in the month must not include patients who are determined to be ineligible or excluded, regardless of whether they are selected for the survey sample
• “Sample Size” can therefore be larger than the number of “Eligible Discharges.” For example, if a patient was selected for the survey sample and later determined to be ineligible (i.e., “Final Survey Status” code of “3 – Ineligible: Not in eligible population”), the patient must be subtracted when reporting the “Eligible Discharges” field (number of eligible discharges in sample in the month). However, this does NOT apply to “Final Survey Status” codes of “2 – Ineligible: Deceased,” “4 – Ineligible: Language barrier,” or “5 – Ineligible: Mental/Physical incapacity.”
• If a patient was not selected for the survey sample and later determined to be ineligible (i.e., received an update with an ineligible MS-DRG code for the patient), the patient must be subtracted when reporting the “Eligible Discharges”
• The “Eligible Discharges” field must include the count of patients who are eligible for the HCAHPS survey, even if the patient’s information is received from the hospital with discharge dates that are beyond the 42-day initial contact period.
Note: A Discrepancy Report must be filed to account for patient information received beyond the 42-day initial contact protocol.
¾ Once the “Type of Sampling” field has been defined for the first month in a quarter, the sample type for the quarter can be changed by resubmitting this file ONLY if the data files for another month in the quarter have not yet been submitted to My QualityNet ¾ When using DSRS as “Type of Sampling,” at least two strata must be defined, with a
minimum of 10 sampled patients per stratum. Once the strata names are defined, they cannot be changed until the beginning of the next quarter.
¾ When small hospitals sample 100% of the eligible discharges (i.e., a census), in order to obtain as close to 300 completes as possible, the “Type of Sampling” must be coded as “1 – Simple Random Sample”
Note: Hospitals with five or fewer eligible HCAHPS patient discharges in a month, including “zero cases,” need to submit an HCAHPS Header Record (Survey Month Data) online via My QualityNet.
Patient Administrative Data Record
¾ All fields in the Patient Administrative Data Record must have a valid value. Use code “M – Missing/Don’t Know” for all missing fields, with the following exceptions:
• When - “Point of Origin for Admission” is missing, it is coded as “9 – Information not available”
• When - “Survey Language” is missing, it is coded as “8 – Not applicable” • When - “Lag Time” is missing, it is coded as “888 – Not applicable”
¾ Patient administrative information must be submitted for all patients selected for the survey sample, including patients found to be ineligible prior to survey administration • If a patient is found to be ineligible or excluded after the sample is drawn, the patient
should be assigned a “Final Survey Status” code of “3 – Ineligible: Not in eligible population”
• If the patient is selected for the HCAHPS survey and the discharge date is beyond the 42-day initial contact period, then the patient should be assigned a “Final Survey Status” code of “3 – Ineligible: Not in eligible population”
setting, not the discharge date from the swing bed
¾ The “Lag Time” field is the number of days between the patient’s discharge from the hospital and the return of the mail survey, or the final disposition of the telephone or IVR survey
• Surveys that receive a “Final Survey Status” code of “1 – Completed survey” or “6 – Non-response: Break-off” must contain the actual lag time.
o These surveys should NOT be coded “888 – Not Applicable” for lag time • Surveys that receive a “Final Survey Status” code of 2, 3, 4, 5, 7, 8, 9, 10, or M (that
is, any “Final Survey Status” code OTHER THAN 1 or 6) need not contain the actual lag time
o Such surveys MAY use either the actual lag time or “888 – Not Applicable”
Note: Even though the lag time can be up to 84 days, survey administration must be completed within 6 weeks (42 days) of initial contact (first questionnaire mailing or telephone/IVR attempt).
¾ Patient administrative information must be submitted for all patients selected for the survey sample, including patients found to be ineligible prior to survey administration
Patient Response/Survey Results Record
¾ Enter all survey responses as provided by the patient for each survey item
¾ All survey questions must have a valid value. For “Final Survey Status” of “1 – Completed survey” or “6 – Non-Response: Break-off,” code missing answers as “M – Missing/Don’t Know” or “8 – Not Applicable.”
¾ It is possible to select more than one response category in Question 26, “What is your
race? Please choose one or more.” Enter all of the categories that the patient has selected. For any categories not selected, enter “0.” If no categories are selected, enter “M – Missing/Don’t Know” for all categories.
¾ If the same patient completes two surveys for the same hospital visit, the hospital/survey vendor uses the first HCAHPS questionnaire received