Background A "proper" definition of POPULIS has been a long time coming. Various perspectives exist regarding what POPULIS is; many people, including some centre staff, feel that POPULIS is a complete integrated software package -- press a button and out comes a result. Here's the (current!) "final" word on the subject.

POPULIS has gone through several growth stages starting before the inception of MCHP, before POPULIS existed, with a series of modules on the University of Manitoba mainframe created by Pat Nicol. In the first medical care supplement done by MCHP, POPULIS was described as an approach and as software (Black et al, 1995). This article described the software as "... based on two front-end, easy-to-use SAS commands..." and 'A newly developed graphic interface ("point and click") supports even inexperienced users in producing most types of runs.' POPULIS Prime was the graphical front end that was developed using some of the generalized portions of the SAS software.

In both cases we found that supporting the software became quite difficult. New, and old, researchers would ask new questions or request new analyses in different ways. The existing software was modified to account for new requests to the point that it was becoming difficult to maintain all of the different options and methods that were requested. Basically the software was becoming more complicated to use than going back to the original data sources and writing the code for the specific question based on a standard set of exclusions and criteria that were outlined in the fledgling concept dictionary.

More recently the whole approach to population health research at MCHP has been described by Roos et al (1999). This publication has moved from describing POPULIS as a front end or of software program to a framework or approach. The description of establishing a population data-based policy unit provides a good outline of the kinds of data and approach that has been taken but the paper still leaves the reader believing (or feeling) that POPULIS is single system for answering or dealing with the outlined issues. Roos et al (1999) have also described the conceptual underpinnings of POPULIS and the nature of the data this system generates. This last paper comes the closest to actually answering the question 'What is POPULIS?'.

POPULIS Defined

• POPULIS stands for POPULation health Information System
• There is no POPULIS database per se with a single program or programming environment.
• POPULIS is an approach to population health analysis developed at MCHP using a consistent set of methods, procedures, and databases intended for both cross sectional and longitudinal studies.

The POPULIS Approach

1. POPULIS uses a population model for measuring the provision of health care services, health indicators, and determinants of health grouped at a geographic or social level (e.g. RHAs or Income Quintiles). This is different than using individual or institutional measures.
   • Click here for a pictorial explanation of the POPULIS approach.

2. It is dynamic in nature with additions and modifications of methods and information as they are developed or obtained.

3. It uses a consistent set of procedures for defining and calculating measures. Many of these are outlined in the MCHP Concept Dictionary.
   • Exclusions/inclusions (e.g. day surgery, residence status)
   • Definitions of indicators (e.g. hip fracture, diabetes)
   • Use of population rates of services (e.g. number of visits to physicians per 1000 population)
   • GEE modeling (e.g. significant changes in hospital utilization over the years 1985 to 1999)

4. The data resources are administrative and population based. These consist of the MCHP Population Health Research Data Repository (PHRDR) The data is representative of the use and/or access for all individuals in a population not just a sample.
   • Administrative data created during the provision of health services
   • Population data routinely collected (e.g. census)
   • Examples include:
     • Health insurance registry
     • Hospital
     • Physician billing
     • Drug
     • Immunization
     • Mortality
     • Census data

5. The databases are organized around a standardized data model - including linkages between and within databases that follow a clear set of rules. The data model includes the ability to identify and follow an individual through time and space.

• Black CB, Burchill CA, & Roos LL (1995). The Population Health Information System: Data Analysis and Software. Medical Care, 33(12), DS127-131.

• Roos NP. (1999) Establishing a Population Data-Based Policy Unit. Medical Care, 37(6), JS15-JS26.

• Roos NP, Black C, Roos LL, Frohlich N, DeCoster C, Mustard C, Brownell MD et al (1999). Managing Health Services How the Population Health Information System (POPULIS) Works for Policymakers. Medical Care, 37(6), JS27-JS41.