Point Of Care Testing In Danish General Practice - Part II

Point-of-care testing (POCT) is increasingly being used in general practice to assist GPs in their management of patients with diseases. An accredited external quality assessment (EQA) program and internal quality control system is recommended1. In the Copenhagen area external as well as internal quality control has been enforced by annual outreach consultant visits and by split sample EQA procedures, where POCT results have been compared with central laboratory results. However, the adherence to quality guidelines has been seen to be less than anticipated among GPs in the Copenhagen municipality and in the former county of Copenhagen.

Dissemination of guidelines alone rarely brings about improvements in clinical practice2 and even an multifaceted implementation of guidelines may not change clinical practice 3;4. Multiple strategies for implementing guidelines appear to be more effective than single ones 5;6. However, well-designed empirical research looking into various implementation strategies is still needed in this area 7.

E-mails have successfully been used in several studies to promote health behaviour change in risk populations 8;9 and our hypothesis is that electronic reminder letters (send to the GPs electronic patient records) is an efficient and inexpensive way to influence the behaviour of GP's.

Due to the low adherence, the Copenhagen General Practitioners' Laboratory (CGPL) plans to introduce electronic reminder letters (alongside the standard implementation procedures) during 2010 in order to increase adherence to the quality guideline.

The aims of this study are:

To evaluate the effect of electronic reminder letters on general practices adherence to clinical quality guidelines regarding POCT.

To evaluate the legacy effect of a potential effect of reminder letters (i.e. the impact of the intervention after it has stopped).

Participants Included in the study are all, approximately 320, GP practices conducting POCT at least four POCT (either haemoglobin or glucose) during a 4 month before study period and not conducting INR analysis (defined as four or less tests during 4 four months period). These practices were identified in the GP database of the Capital area and by the CGPL.

Data collection Data on performed split test EQA procedures is retrieved from CGPL database. These data do not contain any patient related data because all split test EQA are conducted by a constructed identification code. Process indicators (sent reminder letters) are also obtained from CGPL. KvEAP and Capital Region databases provide information on the participating practices and corresponding GPs.

Data from The Capital Region Information regarding: Sex, age, year of graduation from university, working address, type of practice, patient listed to practice and use the following tests: Hemoglobin, Glucose, INR; CRP, HbA1C were retrospectively collected 6 months before the start of the trial (tentative in order to establish a baseline). Each month in the remainder of the study period the investigators receive data from the Capital Region regarding practices in the study areas and identify practices having used Hemoglobin or blood Glucose as a POCT. These data will be compared with the CGPL database every month and those practices that have not done a split sample EQA will receive electronic reminder letters during the following 6 months.

Randomization:

Practices are stratified by area and type of organization by means of SAS (Proc PLAN) by an independent organization. Practices are allocated to usual CGPL quality guideline activities (control group) and usual CGPL implementation activities in addition to electronic reminder letters (intervention group).

Outcome:

Primary outcome:

1. Total number of split tests performed in study period for Hemoglobin and Glucose. Secondary outcomes:
2. Proportion of practices with a high quality of tests defined as 75% of the performed split tests for Hemoglobin and Glucose within the accepted interval according following the CGPL quality guidelines1 in study period.
3. Proportion of practices conducting split tests for Hemoglobin and Glucose in study period.

Power calculation:

The investigators use a 10% adherence estimate based on CGPL data from 2007 in order to ascertain the power of the study. Given a MEREDIF at 15% and a power of 90% it is estimated that 274 practices are to be included in this study.

Statistics:

Differences in the outcomes between allocation groups at baseline and at 6 months follow-up are tested by means of chi-square tests (outcomes 2 and 3) and t-tests (outcome 1).

In order to test the development of adherence relative to the intervention the investigators will in the 3 data collection points use logistic (outcome 2 and 3) and linear (outcome 1) regression where the investigators use GEE methods to account for the repeated measurements.

To identify predictors for adhering to guidelines adjusted odds ratios for the practice characteristics are estimated in multivariate logistic (outcomes 2 and 3) and linear (outcome 1) regression analysis on the outcomes at baseline.

All statistical analyses are performed using SAS, version 9.2 (SAS Institute Inc, Cary, NC).

Intervention Standard implementation The standard implementation of EQA consists of invited meetings and an annual facilitator visit in each practice. As part of the planned implementation strategy GPs were invited to meetings, received written material from the CGPL.

Electronic reminder letters

Electronic reminder letters are sent to practices not adhering to the guideline recommendations of split testing within 30 days; i.e. a reminder letter is send when the CGPL database registers that the last split test or last reminder letter was 31 days ago. Thus, all practices may receive up to 4 electronic reminder letters in each period:

After the 4 months a cross-over will be applied, so that the practices allocated to the control group receive electronic prompting while prompting ceases in the intervention group. This further follow-up acts to investigate the legacy effect of the intervention. The main hypothesis is that the first 4 months of intervention also shows some effect 4 months beyond the intervention period.