Diabetic audits: measuring quality with EMRs

My most recent diabetic audit has just been completed. The results are:

2004: BP<=140/90: 65%;

LDL less than 2.6: 50%

A1C <= 8.4%: 74%

2006: BP<=140/90: 82%; <=130/80: 66%;

LDL less than 2.6: 63%, <=2: 40%

A1C <=8.4%: 81%; <=7%: 48%

Missing: BP 19 patients, LDL 26 patients, A1C 20 patients

2008: BP<=140/90: 90%; <=130/80: 83%

LDL less than 2.6: 79%; <=2: 52%

A1C<=8.4%: 81%; <=7%: 42%

Missing: BP 0 patients, LDL 5 patients, A1C 3 patients

The averages for my practice are:

2006:

BP 125.8 / 76.8;

LDL 2.32;

A1C 7.37

2008:

BP 121.7 / 71.4;

LDL 2.17;

A1C 7.42

I am happy with those results. I can see that I have steadily improved my results for BP and cholesterol control; however, my A1C results for 2008 are slightly worse than for 2006.

Before I discuss this, I would like to talk about how I track these diabetic results. The EMR auditing process is not as good as I would like, but it is significantly better than what I was doing on paper.

The 2004 audit was done prior to the EMR. I had a co-op student from the local high school, and she did the audit. I used billing data from my old billing/scheduling program to get a list of my diabetic patients. At that time, I put in an initial quality improvement step: the student put a yellow sticky note with a reminder (example: LDL) in the progress notes of all patients not at goal while she was doing my audit. I wrote a short article about this process for Canadian Family Physician.

The 2006 audit was done by my resident (all residents have to do a quality improvement project), and entered in Excel. She did this in August 2006; this was four months into the EMR, so not all of the diabetic data was in electronic form yet. That is why there are so many missing results. I had already started using diabetic flowsheets; the time required for audits was significantly shorter than on paper. At that time, I put in reminders for myself to check the flowsheet every 3 months, and started sending Actions to my front staff more aggressively when parameters were not at goal (example: cholesterol high, call patient and ask her to double up on lipitor). I also started using pop-up alerts.

The 2008 data is much more complete. All patients now have flowsheets. My secretary did the audits for me, and I have asked our new resident to do the initial audit for my practice partner.

The EMR audit process is not perfect; I would like, for example, to have a program that automatically extracts all of the last BP results for every diabetic (as well as all LDLs, and all A1Cs). This is not available, so I generate a list of diabetics from the EMR, and the data is manually extracted from the flowsheets and entered in Excel. It is still much faster than using paper charts: all the data is in the flowsheets, so once the chart is loaded, you access everything in a single area. The data is clear, legible, and easy to find.

I would like to have this data available for all of my colleagues in my FHN; I have asked residents to do the audits for others in my group (with permission from each physician), so we will get additional practices done. If I had automated audits, like I do for our preventive services, I could get this done more often and for all of us. I think that it is important to track your results, so you can set goals and see if your Quality Improvement program works. You can see what I am currently doing to try to improve diabetes care here.

I should start talking to my Family Health Team about spreading this to all FHT physicians. It would certainly be possible to do audits for all 40 physicians; we likely will need a bit of funding dedicated to such a project.

I now have an interesting quandary. I don’t know if I can improve on BP or LDL anymore; I can see that I need to maintain my current efforts. However, my A1C average is getting a bit worse. The recent ACCORD study randomized diabetics to tight (aim for A1C <6.0%). The group in the tight control had an average A1C of 6.4%, while the standard group's was 7.5%. The results: "As compared with standard therapy, the use of intensive therapy to target normal glycated hemoglobin levels for 3.5 years increased mortality and did not significantly reduce major cardiovascular events."

My average A1C is 7.42%, which is very similar to the standard group's results in ACCORD; based on this study, I am not convinced that I should change my management. Perhaps guideline developers should consider changing the A1C target to between 7 and 7.9%.

Michelle