Introduction

Determining the number of hospital beds in use (people) at any particular point in time is an important part of any study that looks at changes in usage patterns over short periods of time (e.g. seasonal changes due to specific diseases such as the flu). The following method has been used to calculate an overnight census using hospital discharge abstracts data.

Methodology

1. Choose a date which is at least 1 year (6 months?) before the end of the last date for which you have hospital claims (MCHP Hospital claims data is separation based). This ensures that, for your chosen date, you will have more complete date of admission data.

   E.g. if you choose to count patients who occupied beds on Feb.1, 1998 and you have hospital claims only until 31 March 1998 (end of fiscal year 97/98), then you will not yet have data for persons who were admitted on or before Feb 1 and stayed longer than 60 days - their hospital claims will be in the 98/99 fiscal year dataset.

2. Flag all records with an admission date LESS THAN or EQUAL TO your chosen date AND whose separation date is GREATER THAN your chosen date.

   Caution : do NOT flag records with an admission date AND separation date EQUAL TO your chosen date - these persons did not occupy a bed overnight.

3. Count all flagged records - this is the number of persons in hospital overnight on your chosen date.
   
   
4. For each subsequent day, add (or subtract) to this initial count each day's difference between number of admissions and number of separations. This will give you the count of patients in hospital each night from your chosen date forward.

   Caution : the count will become LESS ACCURATE as you near the end of the last fiscal year for which you have hospital data (for the reasons mentioned in point 1 above).