Documentation of Immunizations from Billing Data and Chart Abstract: Implications for a Large Health

1
Documentation of Immunizations from Billing Data
and Chart Abstract Implications for a Large
Health System Reporting Billing Data to a State
Immunization Registry

• Lee R. Kallenbach, Gordon Jacobsen, Richard
  Bassous.
• Wayne State University and Henry Ford Health
  Sciences Center, Detroit Michigan

2
Background

• The State of Michigan mandated reporting of
  childhood immunizations as of 1/1/98.
• The Michigan Department of Community Health
  provides a client software program for data entry
  and lookup (MCIR-Link).
• A large health system with a clinical information
  system (CIS) did not want to add additional
  software, hardware, and data entry requirements
  in its clinics.

3
Background (cont.)

• A plan to submit billing information to document
  the provision of immunizations was adopted during
  CIS implementation.
• The health system in question maintains a
  corporate data warehouse assembled from encounter
  and billing information.
• The accuracy and completeness of the billing
  records was not fully assessed before data
  submission to the State registry began.

4
Purpose

• To assess the completeness and accuracy of
  billing data as a means to document the provision
  of immunizations.
• Chart abstracted immunization histories served as
  the gold standard to which billing data was
  compared.

5
Methods Chart abstract

• 5 chart abstracting sessions at 4 clinics
• 907 individual medical charts
• Children lt36m of age at time of chart review
• Information Collected
• Type of immunization
• Date administered

6
Methods Billing Data

• Complete billing histories drawn from corporate
  data warehouse for 907 patients.
• Immunizations documented prior to date of chart
  abstract were reviewed.
• Internal billing code converted to CDC vaccine
  code
• Type of immunization / Date administered
• Location, insurance type, encounter history

7
Methods Matching

• Exact- match criteria
• 1) Medical record number
• 2) Date of service
• 3) CDC vaccine code
• Lifelong patient subgroup
• Clinical encounter within 45 days of birth
• Age appropriate recent encounter (3-12 months)

8
Results Matching
9
(No Transcript)
10
Initial Matching

• Much lower than expected - some problems
• Methodological (readily explainable)
• Data entry errors
• Difference in dates
• Combination vaccines (DTP-Hib)
• Procedural (other explanations)
• Some vaccines not billed (VFC)
• Vaccines provided elsewhere
• Incomplete billing documentation(?)

11
Number of Vaccines Documented from Each
Source (No Matching Performed)
12
Next Steps

• Before performing matching procedure in more
  detail need to know more about billing data
  procedures.
• Encounter forms
• Data entry and coding procedures
• Data storage and retrieval (data warehouse)
• Compare to MCIR
• Especially for lifelong patients