Chapter 2. Survey Administration Statistics
2009 Comparative Database Report
This chapter presents descriptive information on the 2009 database hospitals regarding how they conducted survey administration.
The 2009 database consists of survey data from 622 hospitals with a total of 196,462 hospital staff respondents. Participating hospitals administered the hospital survey to their staff between October 2004 and July 2008 and voluntarily submitted their data for inclusion in the database.
Hospitals do not necessarily administer the hospital patient safety culture survey every year. They may administer it on an 18-month, 24-month, or other cycle. Therefore, the comparative database is a "rolling" indicator. Data from prior years are retained in the database when a hospital does not have new data to submit; older data are replaced with more recent data when available; and data are added from hospitals submitting for the first time.
Overall statistics for the hospitals included in the 2009 database are shown in Table 2-1, broken down according to when the data were submitted. The 2009 database includes 395 hospitals carried over from the 2008 report and new data submissions from 227 hospitals. Of the 395 hospital submissions carried over from the 2008 database, 314 hospitals submitted data only once, and 81 hospitals submitted data more than once. Of the 227 new hospital submissions, 104 hospitals submitted data for the first time, and 123 hospitals submitted new data based on a readministration of the survey. Old data from hospitals that submitted more than once were replaced by data from their readministration, so the database reflects their most recent survey data.
Table 2-2 presents data on the number of surveys completed and administered, as well as the response rate.
Most hospitals administered only paper surveys (44 percent), followed by Web (33 percent) and mixed-mode administration involving both paper and Web surveys (23 percent) (Table 2-3).
Table 2-4 shows average response rate by survey mode. Paper survey administration had a considerably higher average response rate than Web or mixed mode. It is therefore still an overall recommendation that hospitals conduct the hospital survey as a paper survey. But each hospital should consider its prior experience with survey modes and response rates when determining which mode is best.
Most hospitals (463, or 74 percent) administered the survey to a census of all hospital staff, or a sample of staff, from all hospital work areas/units. Fewer hospitals (105, or 17 percent) administered the survey to a subset of selected staff or work areas/units. Fifty-four hospitals (9 percent) administered the survey to a subset of selected staff and selected work areas/units (Table 2-5). Twelve hospitals did not administer the entire survey; they excluded one or more of the nondemographic survey items. Those 12 hospitals were excluded from composite calculations if they omitted one or more of the items within a particular composite, but were included in item-level calculations for the items they retained.