Data Collection and Processing

2000 Data Collection Procedures

The data collection for the 2000 NHHCS began with a letter sent to all sampled agencies informing the administrator of the authorizing legislation, purpose, and content of the survey. Each agency was then contacted by an interviewer to discuss the survey and to arrange an appointment with the administrator.

Three questionnaires and two sampling lists were used to collect the data. The Agency Questionnaire was completed with the administrator or a person designated by the administrator. The interviewer then constructed the Current Patient Sampling List and the Discharged Patient Sampling List. These lists were used to select the sample patients and discharges. Sampling was accomplished by using tables showing sets of sample line numbers for each possible count of current patients and discharges in the agency. Up to six current patients and up to six discharges were selected.

After the samples had been selected, the Current Patient Questionnaires and the Discharged Patient Questionnaires were completed for each sampled person by interviewing the staff member most familiar with the care provided to the patient. The respondent referred to patient medical and other records as necessary. No patient was interviewed directly.

After the data had been collected it was converted into machine-readable form by NCHS. Extensive editing was then conducted by computer to assure that all responses were accurate, consistent, logical, and complete. The medical information recorded on the patient questionnaires was coded by NCHS staff according to the International Classification of Diseases, 9th Revision, Clinical Modification. Up to 12 diagnostic codes (a maximum of 6 at admission, and a maximum of 6 at the time of survey or discharge) and up to 2 procedure codes were assigned for each sample patient or discharge.

Imputation Procedures

In most cases, missing data are coded as “unknown.” Unknown codes were not allowed and imputations were made for the following data items: date of interview, date of admission, date of discharge, and sex. Any imputations that were made are indicated by a “1” in the appropriate field.

Date of interview, date of admission, date of discharge, and date of birth are not included in these data files. Therefore, the following items were computed for the user: length of service in days, age at admission in years, current age in years (age at the time of the interview for current patients), and age at discharge (for discharges). If imputations were made to any of the items involved in these computations, this is indicated by a “1” in the appropriate field.

The following imputation procedures were used:

• If date of birth was unknown and age was not reported, age was imputed as follows: If any source of payment was Medicare, age was imputed as 65. Otherwise, the average age with the same sex and type of care category was used.
  • Date of admission was imputed based on the average length of service of patients or discharges within the same sex and type of care category.
• Since the month and year of discharge was assigned for each agency, day of discharge was the only imputation made. Day of discharge was imputed as 15 unless this was incompatible with date of admission.

1998 Data Collection Procedures

The data collection for the 1998 NHHCS began with a letter sent to all sampled agencies informing the administrator of the authorizing legislation, purpose, and content of the survey. Each agency was then contacted by an interviewer to discuss the survey and to arrange an appointment with the administrator.

Three questionnaires and two sampling lists were used to collect the data. The Agency Questionnaire was completed with the administrator or a person designated by the administrator. The interviewer then constructed the Current Patient Sampling List and the Discharged Patient Sampling List. These lists were used to select the sample patients and discharges. Sampling was accomplished by using tables showing sets of sample line numbers for each possible count of current patients and discharges in the agency. Up to six current patients and up to six discharges were selected.

After the samples had been selected, the Current Patient Questionnaires and the Discharged Patient Questionnaires were completed for each sampled person by interviewing the staff member most familiar with the care provided to the patient. The respondent referred to patient medical and other records as necessary. No patient was interviewed directly.

After the data had been collected it was converted into machine-readable form by NCHS. Extensive editing was then conducted by computer to assure that all responses were accurate, consistent, logical, and complete. The medical information recorded on the patient questionnaires was coded by NCHS staff according to the International Classification of Diseases, 9th Revision, Clinical Modification. Up to 12 diagnostic codes (a maximum of 6 at admission, and a maximum of 6 at the time of survey or discharge) and up to 2 procedure codes were assigned for each sample patient or discharge.

Imputation Procedures

In most cases, missing data are coded as “unknown.” Unknown codes were not allowed and imputations were made for the following data items: date of interview, date of admission, date of discharge, and sex. Any imputations that were made are indicated by a “1” in the appropriate field.

Date of interview, date of admission, date of discharge, and date of birth are not included in these data files. Therefore, the following items were computed for the user: length of service in days, age at admission in years, current age in years (age at the time of the interview for current patients), and age at discharge (for discharges). If imputations were made to any of the items involved in these computations, this is indicated by a “1” in the appropriate field.

The following imputation procedures were used:

• Date of interview was imputed for 20 agency records. Imputation was made by examining other forms received from the same agency and using the dates reported on these forms.
• If date of birth was unknown and age was not reported, age was imputed as follows: If any source of payment was Medicare, age was imputed as 65. Otherwise, the average age with the same sex and type of care category was used. Age was imputed for 13 current patient records and for 15 discharge records.
• Sex was imputed as female. Sex was imputed for 1 current patient record and for 2 discharge records.
• Date of admission was imputed based on the average length of service of patients or discharges within the same sex and type of care category. Date of admission was imputed for 37 current patient records and for 32 discharge records.
• Since the month and year of discharge was assigned for each agency, day of discharge was the only imputation made. Day of discharge was imputed as 15 unless this was incompatible with date of admission. Day of discharge was imputed for 10 discharge records.

1996 Data Collection Procedures

Data collection for the 1996 NHHCS began with a letter sent to all sampled agencies informing the administrator of the authorizing legislation, purpose, and content of the survey. Each agency was then contacted by an interviewer to discuss the survey and to arrange an appointment with the administrator.

Three questionnaires and two sampling lists were used to collect the data. The Agency Questionnaire was completed with the administrator or a person designated by the administrator. The interviewer then constructed the Current Patient Sampling List and the Discharged Patient Sampling List. These lists were used to select the sample patients and discharges. Sampling was accomplished by using tables showing sets of sample line numbers for each possible count of current patients and discharges in the agency. Up to six current patients and six discharges were selected.

After the samples had been selected, the Current Patient Questionnaires and the Discharged Patient Questionnaires were completed for each sampled person by interviewing the staff member most familiar with the care provided to the patient. The respondent referred to patient medical and other records as necessary. No patient was interviewed directly.

After the data had been collected, it was converted into machine-readable form by NCHS. Extensive editing was then conducted by computer to ensure that all responses were accurate, consistent, logical, and complete. The medical information recorded on the patient questionnaires was coded by NCHS staff according to the International Classification of Diseases, 9th Revision, Clinical Modification. Up to 12 diagnostic codes (a maximum of 6 at admission, and a maximum of 6 at the time of survey or discharge) and up to 2 procedure codes were assigned for each sample patient or discharge.