Glucometrics (defined as the systemic analysis of blood glucose data) can be used as a benchmarking tool/ quality indicator of glycemic control in hospital or a patient care unit of the hospital.1,2 Metrics generated are utilized to compare the quality of care provided to patients along with other indicators like rates of pressure sores and catheter related blood stream infections.2 The study of glucometrics entails use of automated glucometers3 and automated data management systems.4 In hospitals with limited internal information technology (IT) resources and capability, accessing and assessing the large amount of glucose data generated, remains a challenge. In our hospital with resource limited IT capability, we wanted to evaluate if chart abstraction technique,5 can be used to effectively generate glucometrics. We also studied use of the generated metrics to evaluate in-patient glycemic control in ICU patients and its use as a benchmarking tool, to compare it with other hospitals and centers of excellence.
Indoor records of all patients admitted to ICU of our hospital from July 2011 to September 2011 were evaluated. Data regarding blood glucose level (BSL), its timing, its relation to feeding, as well as insulin dose and type given at that time, along with patient characteristics were noted. All this data was entered into a spreadsheet and then analyzed to generate metrics. A Glucometrics Report (Table 1) was prepared for the period of assessment. The glucometrics data was compared with data from other hospitals.4
Table 1.
Glucometrics Report of Our Study.
| Serial no. | Metric | Result | Remarks |
|---|---|---|---|
| 1 | Measures of glycemic exposure | ||
| 1.1 | Mathematical mean BSL | 179.3 mg/dl (no. of BSLs used for calculations = 676) | Within recommended acceptable BSL range for hospitalized patients (140-180 mg/dl) |
| 1.2 | Mean BSL (not including hypoglycemic values) | 179.78 mg/dl (no. of BSLs used for calculation = 669) | Mean BSL ranged from 119 to 227 mg/dl in the RALS study4 |
| 1.3 | Median BSL | 188 mg/dl | |
| 2 | Efficacy of control (percentage of glucose readings within a specified range) | ||
| 2.1 | Values < 70 mg/dl | 1.03% | 3.11% in the RALS study4 |
| 2.2 | Values > 300 mg/dl | 4.88% | |
| 2.3 | Values within 80-139 mg/dl | 23.22% | |
| 2.4 | Values within 140- 180 mg/dl | 21.44% | |
| 2.5 | Values within 80-200 mg/dl | 52.8% | |
| 3 | Units of analysis | ||
| 3.(a) | Patient | ||
| 3.(a).1 | % of patients with hypoglycemia | 15.36% | |
| 3.(a).2 | % of patients with extreme hyperglycemia | 50% | |
| 3.(b) | Patient-stay | ||
| 3.(b).1 | Average glucose per patient-stay | 173.29 mg/dl | Most practical metric ofhyperglycemia associated risk |
| 3.(b).2 | % of patient-stays with average BSL within target range (140-180 mg/dl) | 50% | |
| 3.(b).3 | % of patient-stays with any BSL < 70 mg/dl | 15.38% | |
| 3.(b).4 | % of patient-stays with any BSL >300 mg/dl | 30.77% | |
| 3.(b).5 | Average number of glucose readings per patient for entire hospitalization (patient-stay) | 26 | 4.2 to 7.9 in a study3 |
| 3.(c) | Patient-day | ||
| 3.(c).1 | Average glucose per patient-day, ie, patient-day weighted mean BSL | 175.42 mg/dl | |
| 3.(c).2 | % of patient-days with average BSL within the target range (140-180 mg/dl) | 37.19% | |
| 3.(c).3 | % of patient-days with any BSL<70 mg/dl | 4.26% | |
| 3.(c).4 | % of patient-days with any BSL > 300 mg/dl | 14.63% | |
| 3.(c).5 | Average number of glucose readings per patient-day | 4.12 | 0.67 to 1.5 in a study3 |
| 4 | Patterns of insulin use | ||
| 4.1 | Regular human insulin was used in all patients | ||
| 4.2 | Sliding scale insulin was used for treatment of 22 patients (84.6%) | ||
| 4.3 | Insulin infusion was used in 4 patients (15.38) and for 14 patient-days (8.53%) | ||
| 4.4 | After starting insulin infusion it took an average of 4.4 hours for BSL to reach target range (140-180 mg/dl) | ||
Total number of blood glucose levels (BSLs) included for calculation = 676; total number of included patients = 26; total number of included patient-days = 164.
The data required to generate glucometrics for a hospital or unit of the hospital can be collected by the following methods:
Analysis of the actual medical records (chart abstraction technique):5 Pros: BSL values can be correlated with meal timings and with treatment given. Cons: labor-intensive.
Analysis of data from automated lab information systems:3-5 Pros: more efficient, large amount of data analyzed in a much shorter time, data from all patients can be analyzed. Cons: need for automated glucometers and laboratory information systems; cannot correlate BSL with meals and action taken.
Hospitals with limited IT resources can connect to and import their data to external data management systems.4 For hospitals with limited IT resources and who are unable to connect to these external data management systems due to their geography or other limitations, the chart abstraction technique remains the only option. As the process is labor-intensive, only a convenience sample of records can be analyzed. But, this technique offers a chance for institutes to assess their in-patient glycemic control. The glucometrics report offers a glimpse into the world of glycemic control at that hospital during the period of the study and to compare it with results from other hospitals and centers of excellence.4 The limitations of our study are that it was a single unit study, with a small sample size.
Thus, in hospitals with limited IT support, chart abstraction technique can be used to generate these metrics. Even in the United States, more than half (59%) of the hospitals do not have an automated capability to extract and analyze glucose data.6 Our study can be useful for all these hospitals.
Footnotes
Abbreviations: IT, information technology; BSL, blood glucose level; ICU, intensive care unit.
References
- 1. Society of Hospital Medicine. Glycemic Control Resource Room. Available at: https://www.hospitalmedicine.org. Accessed May 5, 2014.
- 2. Goldberg PA, Bozzo JE, Thomas PG, et al. “Glucometrics”—assessing the quality of inpatient glucose management. Diabetes Technol Ther. 2006;8(5):560-569. [DOI] [PubMed] [Google Scholar]
- 3. Buchs AE, Rapoport MJ. Institutional glucometrics to determine glucose control as practised by general medicine wards. IMAJ. 2010;12:463-467. [PubMed] [Google Scholar]
- 4. Anderson M, Zito D, Kongable G. Benchmarking glucose results through automation: the 2009 remote automated laboratory system report. J Diabetes Sci Technol. 2010:4:1507-513. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Cook CB, Wellik KE, Kongable GL, Shu J. Assessing inpatient glycemic control: what are the next steps? J Diabetes Sci Technol. 2012;6(2):421-427. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Cook CB, Elias B, Kongable GL, Potter DJ, Shepherd KM, McMahon D. Diabetes and hyperglycemia quality improvement efforts in hospitals in the United States: current status and barriers to implementation. Endocr Pract. 2010;16(2):219-230. [DOI] [PubMed] [Google Scholar]