NHDS 83-C justification for the addition of 2 items

NATIONAL HOSPITAL DISCHARGE SURVEY

0920-0212 (expiration date Aug. 31, 2008)

This request is to add two variables to an approved data collection, the National Hospital Discharge Survey (NHDS), (OMB No. 0920-0212), beginning with the 2007 data collection year. The two new data items are admitting diagnosis and a Present on Admission (POA) indicator for the discharge diagnoses currently collected.

A. Justification

1. Circumstances Making the Collection of Information Necessary

The National Hospital Discharge Survey (NHDS), which has been conducted continuously since 1965, is the Nation’s principal source of data on inpatient utilization of short-stay, non-Federal hospitals and is the only annual source of nationally representative estimates on the characteristics of discharges, the lengths of stay, diagnoses, surgical and nonsurgical procedures, and the patterns of use of care in hospitals in various regions of the country. It is the benchmark against which special programmatic data sources are compared.

The data collected for the NHDS conform to the Uniform Hospital Discharge Data Set (UHDDS). The Uniform Hospital Abstract Subcommittee of the U.S. National Committee on Vital and Health Statistics selected the UHDDS as a basic core of variables that would be of value to many potential users and that could be readily and reliably obtained. The U.S. National Committee on Vital and Health Statistics recommended in 1972 that UHDDS constitute the minimum basic data set for all hospital discharge abstracts. In 1974, the Secretary of the (then) Department of Health, Education and Welfare approved the collection of the UHDDS for Medicare, Medicaid, and other Department programs, including the NHDS. The NHDS collects most of the items contained in the UHDDS.

The National Uniform Billing Committee (NUBC) was formed to develop a single uniform bill and standard data set that could be used nationwide by institutional providers and payers for handling health care claims, particularly Medicare claims. The current bill is the UB-92, which includes UHDDS items.

In February 2005, the NUBC approved the UB-04 as the replacement for the UB-92. The UB-04 will be phased in nationally from March 1 to May 22, 2007 and becomes mandatory on May 23, 2007. The UB-04 will provide for the collection of two new clinically relevant variables: admitting diagnosis and a diagnosis indicator to flag diagnoses that were present on admission. Through the NHDS, the National Center for Health Statistics (NCHS) is requesting to add these two variables to the survey beginning with the 2007 data collection. Because definitions for the new variables have been established for the UB-04 and the variables are required for Medicare reporting and other payments, we anticipate that the NHDS hospitals will be able to provide the new information to us in a uniform manner.

The addition of these variables to the NHDS data collection also must be approved by the NCHS Ethics Review Board (ERB) before it can be fielded. The ERB request to add these variables to the NHDS will be submitted in December 2006.

2. Purpose and Use of Information Collection

The collection of admitting diagnosis will provide increased information over the span of the hospitalization. It will allow analysis of the symptoms which people present at the hospital and whether these symptoms are indicative of their final diagnosis (i.e., admitting diagnosis of appendicitis which later turns out to be gastroenteritis). Information on admitting diagnosis also enables analyses of whether different groups present with different symptoms for certain conditions (e.g., males and females diagnosed with an acute myocardial infarction).

The present on admission (POA) indicator will be collected on the seven discharge diagnoses (the principal diagnoses and up to six secondary diagnoses) the NHDS currently collects. The POA indicator can help distinguish between pre-existing conditions and those that were developed, or were first recognized during the hospitalization. Both the California and New York health data systems already collect POA and it has proven valuable for both risk adjustment and outcome assessment.

12. Estimates of Annualized Burden Hours and Costs

A. Burden Hours

The current average time for hospital personnel to complete the NHDS abstract is 5 minutes per record, 4 minutes for abstracting and 1 minute for pulling, refilling and other activities. Because the two data items being added to the medical abstract beginning with the 2007 NHDS (admitting diagnosis and present on admission indicator) will be located near the other NHDS data items, it is estimated that one additional minute is required for a total of 6 minutes to complete the abstract form (line 1 of burden table).

There is no change to line 2 of the burden table, alternate procedure hospitals, in which Bureau of the Census personnel complete the NHDS medical abstract record.

For hospitals which submit their abstract data via printout or in-house tape/CD-ROM (line 3 of the burden table), it is estimated that they currently spend 12 minutes per month. For those hospitals which were already submitting electronic or printout data for NHDS prior to 2007, it is estimated that about 2 hours of staff time is required to program and verify the two additional data items added to the NHDS. Prorated over 10 years of participation in the NHDS, one minute per month is added to the table for a total of 13 minutes.

Lines 4 and 5 of the burden table remain unchanged.

Respondent burden is summarized in the following table:

	Number of responding hospitals	Number of responses per hospital	Hours per response	Response burden (hours)
Medical Record Abstracts Primary Procedure Hospitals	62	250	6/60	1,550
Alternate Procedure Hospitals	124	250	1/60	517
In-House Tape or Printout Hospital	80	12	13/60	208
Induction	15	1	2	30
Non-response Study	50	1	2	100
TOTAL				2,405

B. Burden Costs

The average annual response burden cost for the NHDS is estimated to be $53,789 for each survey year. The hourly wage estimate for medical coders is based on the 2004 American Health Information Management Association (AHIMA) Membership Profile. The hourly wage rate for hospital executives is based on the 2004 Hay Group’s Hospital Compensation Survey. The following table shows how the respondent cost was calculated:

Type of Respondent	Response burden hours	Hourly Wage Rate	Respondent Cost
Medical coder	2,275	$ 17.45	$ 39,699
Hospital CEO/CFO	130	$108.00	$ 14,040
TOTAL			$ 53,739

Hospitals are reimbursed for their participation.

15. Explanation for Program Changes or Adjustments

The increase in burden hours (274 hours) from 2,131 to 2,405 is due to a program change with the addition of two variables to the NHDS abstract form: 1) admitting diagnosis and 2) present on admission indicators for the principal diagnosis and six secondary diagnoses.

16. Plans for Tabulation and Publications and Project Time Schedule

The two variables will be added to the 2007 NHDS data collection, which starts January 2007. However, since the two new variables do not become mandatory for the hospitals to collect until May 23, 2007, it is uncertain how complete the 2007 records will be for this information. It is not until the 2008 NHDS data collection that we expect to have annual estimates for these two new variables. If the new variables have appropriate response rates and meet NCHS levels of reliability, the data will be released with the 2008 public use file.

B. Collections of Information Employing Statistical Methods

2. Procedures for the Collection of Information

Data Items

The revised Medical Abstract Form is shown at Attachment A. The only changes are to add item number 17, admitting diagnosis and to add a check-box for item 18 (numbered 18.a.) for present on admission options.

4