Introducing Glucometrics

Glucometrics may be defined as the systematic analysis of blood glucose data. This section assumes that you now have a stream or sets of glucose measures data available for analysis as outlined above. Even if this is not so, the principles presented will apply to the more limited data you can obtain using chart or laboratory database sampling strategies.

There are currently no official standards or guidelines for formulating metrics on the quality of inpatient glycemic control. In the absence of official standards, we present some principles for characterizing and trending your glucometrics in a manner that is robust and meaningful. We conclude with a tiered summary of recommendations for practical glucometrics that we hope will be useful to individual improvement teams and will enable benchmarking and a more valid comparison of hospital data on glycemic control efforts with those of published reports. True standards for glucometrics must await the American Association of Clinical Endocrinologists (AACE), American Diabetes Association (ADA), Society of Hospital Medicine (SHM), and others reaching a formal consensus (and perhaps for more evidence), but we believe these practical recommendations are a useful guide in the interim.

Getting Started

Reliable metrics for assessing glycemic control and the frequency of hypoglycemia are a must to assess whether your interventions are resulting in more improvement than harm. Hypoglycemia metrics must be especially convincing because fear of hypoglycemia remains a major source of clinical inertia, impeding efforts to improve glycemic control. We advocate direct measurement of hypoglycemia, rather than using markers of hypoglycemia, such as the amount of D50W used, which depends on many factors unrelated to glycemic control.

A. Key considerations

1. There are many ways to characterize glycemic control data for analysis and presentation. This is an evolving field, so we recommend retaining all data so you can “slice and dice” it differently as the definitions and standards evolve over time. You may also find you want to look at the data in different ways as your initiative matures internally, so this flexibility is important.
2. Sometimes getting the stream of data and analysis to flow can take weeks to months. Don’t let this discourage you, and don’t wait for perfect data and metrics to initiate and/or move forward on initiatives to improve the insulin-prescribing patterns, safety, or education for the inpatient with diabetes or hyperglycemia. Data can always be collected retrospectively if necessary, that is, once your analytic capabilities are in place

Defining the Target Patient Population

B. Inclusion and exclusion criteria

Your team needs to decide which patients and glucose values to include in your analysis. There should be separate/independent analyses for critical care and non–critical care units, for example, because their glycemic targets and their considerations in glucose management differ.

1. Define your target population precisely so you can be confident you are analyzing data from the patients most likely to be affected by your interventions. You might accomplish this by including all patients in your target area who are given insulin while hospitalized or who are discharged on insulin; all patients who have a certain number of point-of-care (POC) glucose readings; all patients meeting glycemic criteria for diabetes (eg, 2 random glucose values > 180 mg/dL or preprandial glucose values ≥ 126 mg/dL); all patients with a diagnosis of diabetes (using DRG and 250.xx ICD9 codes or chart review) plus all patients with hyperglycemia (eg, BG > 180 mg/dL); or other predefined criteria. In critical care units, you might want to include all patients who would meet criteria for an intravenous insulin infusion.
   You may also want to exclude some patients who might not be representative of your target group. For example, if you are targeting subcutaneous insulin use in general hospitalized patients, you might eliminate those patients who are admitted specifically as the result of a diabetes emergency (eg, diabetic ketoacidosis [DKA] and hyperglycemic hyperosmolar state HHS]), as their marked and prolonged hyperglycemia will skew BG data. Pregnant women will generally be excluded from broad-based hospital unit BG analyses and considered as a discrete category because they have very different targets for BG therapy.
2. You may wish to focus on a certain period of hospitalization, such as the day of a procedure and the next 2 days in assessing the impact of the quality of perioperative care or the first 10–14 days of a non–critical care stay to keep outliers for length of stay (LOS) from skewing your data. You may also want to exclude some periods from your analysis, such as the first day of hospitalization, as early BG control may not realistically be affected by your interventions and is likely to be impacted by multiple variables beyond the efforts of the glycemic control team.
3. When possible, work with your IT department to obtain standardized reports of glycemic control and hypoglycemia rates that are service- and unit specific.
4. Some programs eliminate hypoglycemic values from calculations of the glycemic mean to prevent giving false assurance of good glycemic control gained by inducing hypoglycemia.
5. You may only want to examine patients with several (4–8 or more) bedside glucose meter readings to ensure that a mean calculation is meaningful for a given patient.
6. If feasible to do so with your data collection methods, you may wish to select only the regularly scheduled (qAC and qHS, or q6 hour) glucose readings for inclusion in the summary data of glycemic control, thereby reducing bias caused by repeated measurements around extremes of glycemic excursions. An alternative is to exclude glucose readings within 60 minutes of a previous reading (in non–critical care units).

C. Units of analysis

You can examine glycemic control using different units of analysis. Each method has some advantages and disadvantages, and a blend of reporting methods often provides the most balanced and accurate picture.

1. Unit of analysis is glucose value, denominator is all glucose values
   This is the simplest measure and the one with the most statistical power. All glucose values for all patients of interest are in the denominator. The numerator is the number of glucose values in a given range. You might report, for example, that 1% of the 1000 glucose values were less than 70 mg/dL for a certain period or that the mean of all glucose values collected for the month from patients in non–critical care areas of the wards was 160 mg/dL. The obvious disadvantages of this approach are that these analyses are less clinically relevant than patient-level analyses and that patients with many glucose readings and long hospitalizations may skew the data.
2. Unit of analysis is patient (or patient stay)
   All patients who are monitored make up the denominator. The numerator may be the percentage of patients with any hypoglycemia during their hospital stay or the percentage of patients achieving a certain mean glucose during their hospitalization, for example. This is inherently more clinically meaningful than using glucose value as a unit of analysis.
   Not controlling for LOS effects is a major shortcoming of this approach. For example, a hospitalized patient with a long length of stay is much more likely to be characterized as having at least one hypoglycemic value read than is a patient with a shorter length of stay. Another shortcoming is that uneven distribution of testing is not corrected for. A patient’s mean glucose might be calculated on the basis of 8 glucose values on the first day of hospitalization, 4 on the second day, and 1 on the third day. In addition, as already noted, repeat testing around the time of an outlier glucose value can skew results. Despite all these shortcomings, reporting by patient remains a popular and valid method of giving glycemic control results, particularly when complemented by other views and refined to control the number of readings per day and the number of days in the hospital beyond a certain point.
3. Unit of analysis is monitored patient-day
   The denominator in this setting is the total number of days patient glucose level is monitored. The benefits of this method have been described and advocated in the literature (Goldberg PA, Bozzo JE, Thomas PG, et al. “Glucometrics” — assessing the quality of inpatient glucose management. Diabetes Technol Ther. 2006;8:560–569). As with patient-level analyses, this measure will be more rigorous and meaningful if the BG measures evaluated have been standardized. Typical reports might include parameters such as percentage of monitored days with any hypoglycemia, percentage of monitored days with all glucose values in the desired range, or percentage of days with extreme values. This unit of analysis is probably the most difficult to analyze and interpret. On the other hand, it’s clinically relevant and less biased by length-of-stay effects. This method provides a good balance when presented with data organized by patient.
   One way to express patient-day glycemic control that deserves special mention is the day-weighted mean. A mean glucose is calculated for each patient-day, and then the mean is taken across all patient-days. The advantage of this approach is that it corrects for variation in the number of glucose readings each day; all hospital days are weighted equally.
   The following example of using all three units of measurement, in this case to determine the rate of hypoglycemia, demonstrates the different but complementary information that each method provides and highlights the importance of precisely defining glucometrics when analyzing and comparing data across time and with those of other institutions.




In 1 month, 3900 POC glucose measurements were obtained from 286 patients repre­sent­­ing 986 monitored patient-days. With hypoglycemia defined as POC BG ≤ 60 mg/dL, the results showed: 50 of 3900 measurements were hypoglycemic                         1.4% 22 of 286 patients had ≥1 hypoglycemic episodes                     7.7% 40 of 986 monitored days had ≥1 hypoglycemic episodes          4.4%

Some have argued that in this particular case, the patient-day metric gives the most balanced view of hypoglycemic events (Goldberg et al. Diabetes Technol Ther. 2006;8:560–569).

4. Analysis describing change in glycemic control over time in the hospital

In the critical care setting, this unit of analysis may be as simple as the mean time to reach the glycemic target on your insulin infusion protocol.
On non–critical care wards, it is a bit more challenging to characterize the improvement (or clinical inertia) implied by failure of hyperglycemia to lessen as an inpatient stay progresses. One method is to calculate the mean (or percentage of glucose values in a given range) for each patient on hospital day (HD) 1, and repeat for each HD (up to some reasonable limit, such as 5 or 7 days). See examples of this form of analysis. Alternatively, just the first preprandial glucose value of each hospital day could reasonably be used to assess the same process.

D. Measures of control

In addition to deciding the unit of analysis, your team also needs to decide which measures of control to use. These could include rates of hypo- and hyperglycemia, percentage of glucose readings within various ranges (eg, <70, 70–150, 70–180, >180 mg/dL), mean glucose value (with or without exclusion of hypoglycemic values), percentage of patients or patient-days during which mean glucose is within various ranges, or the “in control” rate (ie, when all glucose values are within a certain range).

As with the various units of analysis, each of these measures of control has various advantages and disadvantages. For example, mean glucose is very easy to report and understand, but masks extreme values. Percentage of glucose values within a certain range (eg, per patient, averaged across patients) presents a more complete picture but is a little harder to understand and will vary depending on the frequency of glucose monitoring; as mentioned above, this latter problem can be corrected by including only certain glucose values. The important point is that you choose a few, but not all, measures of control in order to get a complete picture of glycemic control at your institution.

In critical care and perioperative settings, interest in glycemic control is often more intense around the time of a particular event such as major surgery or after admission to the ICU. Some measures commonly used in performing such analyses are:

1. All values above a specified target or outside a target range within a designated crucial period
   For example, the University Healthcare Consortium and other organizations use a simple metric to gauge perioperative glycemic control. They collect the fasting glucose on post-op days 1 and 2 and then calculate and trend the percentage of post-op days with any fasting glucose > 200 mg/dL. Of course, this is a very liberal target, but it can always be lowered in a stepwise fashion once it is regularly being reached.
2. Three-day blood glucose average
   The Portland group uses the mean glucose of each patient for the period that includes the day of coronary artery bypass graft (CABG) surgery and the following 2 days. The 3-day blood glucose average (3-BG) correlates very well with patient outcomes and can serve as a well-defined target (Endocr Pract. 2004;10[suppl 2]:1–33). Portland has progressively lowered the 3-BG target over the years, as mentioned above, and continues to obtain very good outcomes associated with improved glycemic control (current upper limit set at 110 mg/dL). It is likely that use of the 3-BG would work well in other perioperative/trauma settings and could work nicely in the MICU as well, with admission to the ICU as the starting point for calculation of the 3-BG.
3. Hyperglycemic index Measuring the hyperglycemic index is a validated method (Vogelzang M, van der Horst ICC, Nijsten MWN.Hyperglycaemic index as a tool to assess glucose control: a retrospective study Crit Care. 2004;8[3]:R122–R127) of summarizing glycemic control of ICU patients. It is designed to take into account the sometimes uneven distribution of patient testing. Time is plotted on the X axis and glucose values on the Y axis. Vogelzang et al. calculated the area under the curve (AUC) of glycemic values more than the upper limit of normal in millimoles per liter divided by the length of stay in hours. Glucose values in the normal or hypoglycemic range were not included in the AUC. Mortality correlated well with this glycemic index.
   Once monitoring has been in place for a while, you should be able to get a sense of changes in the time it takes to achieve acceptable glycemic control in your inpatients.
   

E. Establishing definitions

Your steering committee needs to define terms. For example, many centers define hypoglycemia as ≤60 mg/dL, whereas the ADA definition, based on physiologic changes that may take place, defines hypoglycemia as ≤70 mg/dL. You may further define hypoglycemia by severity, with any glucose ≤40 mg/dL defined as severe hypoglycemia.

In a similar vein, as you create reports that reflect good glycemic control, you need to decide what to report. Will you characterize a patient as in optimal control only if all preprandial glucose values are <130 mg/dL and no glucose value is >180 mg/dL, or will you pick a more liberal range of 61–180 mg/dL, without sorting out which glucose values represent preprandial readings? Given the difficulties of determining fasting versus prandial readings, the latter may be more practical.

Introducing optimal BG targets in a stepped fashion over time should also be considered. Furnary has done this in the Portland Project, which tracks glycemic control in cardiac surgery patients receiving intravenous insulin therapy. The initial BG target for this project was <200 mg/dL; it was subsequently lowered stepwise over several years to 150 mg/dL, then to 120 mg/dL, and most recently to 110 mg/dL. This approach allows the safe introduction of targeted glycemic control and promotes acceptance of the concept by physicians and the allied nursing and medical staff.

F. Other considerations relative to glucometrics

Yale Glucometrics Web site
The Yale Informatics group has put together a Web-based resource (Yale Glucometrics Web site) that describes glucometrics in a manner similar to the discussion here and in an article by group members (Goldberg PA, Bozzo JE, Thomas PG, et al. Diabetes Technol Ther. 2006;8: 560–569).

The Web site allows downloads of deidentified glucose data, with which it can automatically prepare reports on glucose control. Currently analysis of the following measures can be included in the reports:

1. General results
   
   1. Number of included and number of excluded patients.
      
   2. Average number of glucose readings per patient for the entire hospitalization (ie, patient-stay).
      
   3. Average number of readings per patient-day.
2. Glucose reading
   
   1. Average of all glucose readings from all patients.
      
   2. Percentage of glucose readings within specified ranges: <60, >300, and 80–139 mg/dL.
3. Patient stay
   
   1. Average glucose per patient stay.
      
   2. Percentage of patients whose average glucose is in the target range (80–139 mg/dL).
      
   3. Percentage patients with any result < 60 or > 300 mg/dL.
4. Patient-day
   
   1. Average glucose per patient-day, ie, patient-day weighted mean.
      
   2. Percentage of patient-days with average glucose in the target range (80–139 mg/dL).
      
   3. Percentage of patient-days with any result < 60 or > 300 mg/dL.