Development of a primary care research network focused on chronic disease: is this feasible for both practices and networks?

Problem

When researching true population health issues, pooled general practice (GP) data should be a key resource. The challenges faced by researchers working with primary care data are the lack of tools to explore a narrative record that contains no coded or structured data.

Despite the advantages that computers offer with respect to note-keeping and prescribing, clinical coding of consultations remains an area that is far from universal in countries as diverse as, Australia and Ireland. The limitations of Irish general practice coding have impacted on the ability to conduct trials.

The overall aim was to assess the feasibility and effectiveness of a program of training and feedback to improve the quality of coding for chronic conditions in general practice information systems. Specific objectives were to assess:

1. How feasible is it to deliver this programme and collect the data?

2. How reliable is the data generated by this activity?

Approach

Searches were developed for GP software. Gross prevalence of type 2 diabetes mellitus (T2DM), asthma, chronic obstructive pulmonary disease (COPD), ischaemic heart disease (IHD), heart failure (HF), atrial fibrillation (A/Fib), transient ischaemic attack (TIA) and stroke was studied. Average monthly prevalence was compared with national and international estimates.

Findings

16 practices were recruited, with 65.5 full time equivalent GPs and 36,327 patients.

Average prevalence of all conditions varied greatly, being lowest for T2DM and A/Fib (3.6 and 5.4 fold difference respectively) and highest for IHD and HF (34.0 and 20.0 fold respectively).

The average reported prevalences were as follows: T2DM 6.8%; Asthma 9.2%; COPD 4.1%; HF 1.7%; IHD 3.3%; Stroke 2.2%; TIA 1.1%; A/Fib 4.3%. These rates were similar for December. Some showed large variation with estimated Irish figures (e.g. COPD 4.1 times estimated figure) and others were broadly similar (e.g. HF 1.1 times estimated figure). T2DM (6.8%) and IHD (3.3%) figures were broadly similar to English rates (6.3% and 3.2% respectively), while A/Fib (4.3%) and COPD (4.1%) were substantially different (1.7% and 1.8% respectively). Search accuracy was improved by automation.

Consequences

It is feasible to deliver training, collect monthly data and issue reports.

There is considerable variation of known prevalence of the conditions studied, both between practices and with national and international rates. Coding accuracy should be confirmed by audit.