Data Cart

Your data extract

0 variables
0 samples
View Cart
Number of nights in hospital, past 12 months (pre-1997)

Codes and Frequencies

Can't find the category you are looking for? Try the Detailed codes


For persons with at least one overnight hospital stay in the past 12 months (HOSPNGHTD), HOSPNITED reports the number of nights these persons spent in the hospital during the past 12 months.

This variable is a generated variable, which was constructed by NCHS from responses to questions that probed for specific information regarding hospitalizations. Respondents were first asked to report when they entered the hospital and provide dates or the best possible estimate for dates. As is described in detail in the variable description for HOSPNGHTD, interviewers were encouraged to probe for as accurate information on admission and release dates as possible. The Field Representative's Manuals directed interviewers, "You should always be able to make some sort of write-in entry to the question on date of entry. For statistical purposes, a date must always be assigned for each hospital entry, so it is essential that you obtain the maximum amount of information available." If the respondent could not provide adequate information on dates, then a number for HOSPNITED was not generated.


More detail on whether a hospital night was "generated" can be learned from the Field Representative's Manuals, which note that, for each hospitalization, the interviewers should verify the date of entry into the hospital and the number of nights, footnote "date verified," and fill in answers for the remaining hospitalization questions (such as the type of hospital). The Manuals state that "any necessary deletions will be handled when the questionnaires are processed."


Question wording for HOSPNITED was consistent for 1969-1981 and from 1982-1996, but some changes occurred between these periods.


For 1969-1981, for each hospitalization episode in the past year, the interviewer initially asked, "When did -- enter the hospital (nursing home)?" and filled in the admission month, date, and year. After collecting the name and address of the institution, the interviewer asked, "How many nights was -- in the hospital (nursing home)?" Three additional questions were asked if the answers were not evident from prior entries: "How many of these -- nights were during the past 12 months?" "How many of these -- nights were in the past 2 weeks?" and "Was -- still in the hospital (nursing home) last Sunday night for this hospitalization (stay)?"

For 1982-1996, for each hospitalization episode in the past year, the interviewer asked, "On what date did -- enter the hospital?" (recording admission month, date, and year) and then immediately asked, "How many nights was -- in the hospital?" In the latter period, the additional questions for clarification were dropped, and there was no reference to nursing home stays. This last change parallels the dropping of follow-up questions about stays in convalescent homes and nursing homes in the initial probe about any hospital stays in the past year. (See HOSPNGHTD for further details.)

Despite the aforementioned changes, the general definition of what counted as "a night in the hospital" was largely consistent throughout the survey.


For example, as discussed in greater detail for HOSPNGHTD, nights spent in the emergency room, staying with a sick family member, or admission to and release from the hospital on the same date did not count as overnight stays.

Substantively, HOSPNITED is similar to the variable HOSPNITE, available from 1997 forward. There were, however, some differences in how the information was collected before and after 1997.


After 1997, interviewers did not ask about the dates the individual was hospitalized, nor ask the number of hospital night stays for each hospitalization episode. Rather, respondents were asked a summary question: "Altogether, how many nights were you in the hospital during the past 12 months?" For HOSPNITE, data are based directly on responses to a question, rather than generated by the NCHS, based on information reported to interviewers. Researchers should thus be cautious in combining results from HOSPNITED, for the pre-1997 years and from HOSPNITE, for 1997 forward.

As the 1997 NHIS Survey Description states:

Health care utilization estimates based on the 1997 NHIS may differ from those for earlier years of the NHIS due to changes in the questions and/or the context of the questions. Thus, 1997 estimates of health care utilization may not be comparable to estimates from previous years. ... Users are advised to compare 1997 NHIS questionnaire items pertaining to health care utilization to those used in previous NHIS surveys.


  • 1963-1996: All persons.


  • 1963-1996