Over the past decade there has been significant improvement in control of cardiovascular risk factors (lipid control, blood pressure (BP) control) among Veterans with diabetes. This improvement has been driven at least in part by performance measurement efforts that focused on attainment of specific thresholds for these risk factors. However, current performance measures for BP and lipids focus only on achievement of a specified level of the intermediate outcome (BP <100/40 mm Hg; LDL<100 mg/dL), which may promote overtreatment leading to adverse events.
We had a unique opportunity to work with the VA Office of Analytics and Business Intelligence (OABI), formerly the Office of Quality and Performance, to adapt our previously proposed "tightly-linked" clinical action measures for assessing blood pressure and hyperlipidemia management in diabetes. The tightly-linked measure gives credit for appropriate clinical action even when a target threshold is not achieved.
The primary aims of the study were to: 1) To work with OABI and with our Diabetes Quality Enhancement Research Initiative (QUERI) Workgroup on Clinical Action Measures in order to finalize specification of tightly-linked clinical action measures (BP and Lipid) for diabetes; and 2) To use data from the VA Corporate Data Warehouse (CDW) to examine performance on the measures, including variation across facilities.
This project examined the following research questions: 1) Do current sources of automated data, including data from the CDW, provide all the information required for full specification of the BP and lipid clinical action measures?; and 2) What is the level of performance achieved on the resulting measures and how much variation is there across facilities?
We worked with the Diabetes QUERI Workgroup on Clinical Action Measures to develop tightly-linked action measures that were consistent with current literature, to specify medication classes and dosing levels, and to define time periods and intermediate outcome targets for measurement. We used a retrospective cohort study design to examine 2009-2010 patient data contained in CDW. Eligible patients included active VA primary care patients, 18 years and older, with an established diagnosis of diabetes mellitus based on related ICD-9 codes and medication fills.
Meeting the BP Clinical Action Measure
An eligible patient age 18-75 was determined to have met the measure if the index SBP was <140 mmHg and the DBP was <90; or if the index SBP was <150 and the DBP was <65; or if the index SBP was <150 and the patient was on 3 or more moderate dose antihypertensive medications or if appropriate action (e.g., increased dose of existing BP med, started a new BP med, repeat BP <140/90) occurred within 90 days.
Potential BP Overtreatment
A patient 18 years or older was considered to have received possible overtreatment if their index SBP was <140 and DBP was <65 and they were receiving 4 or more BP medications or medication intensification occurred in the face of the low diastolic BP.
Meeting the Lipid Clinical Action Measure
An eligible patient age 50-75 was determined to have met the measure if the index LDL (defined as the last LDL value recorded in the measurement period, July 1, 2009-June 30, 2010) was <100 mg/dL; or they were on at least a moderate dose statin at the time of the index LDL; or appropriate clinical action (e.g., on a moderate dose statin, increase/start a statin, repeat DL <100) occurred within 90 days following the index LDL; or there was no Index LDL, but there was a prescription for a moderate or higher dose statin within the last 120 days of the measurement period.
Potential Lipid Overtreatment
A patient 18 years or older was considered to have received potential overtreatment if they did not have a diagnosis of Ischemic Heart Disease (IHD) and filled a prescription for high dose statin at the time of the index LDL or 90 days following the index LDL, or, for those with no index LDL, if they filled a prescription for a high dose statin within the last 120 days of the measurement period.
For each measure, we calculated the number of patients who passed the clinical action measure and identified the reasons for meeting the measure. Next, using multilevel logistic regression models, we examined the predicted rates of passing the action measure across the almost 900 facilities of care. Similarly, we calculated the proportion of patients with potential overtreatment and the reasons for overtreatment. Using multilevel models, we examined the predicted rates of potential overtreatment across all sites of care. Finally, for each measure, we divided all sites into quartiles based on meeting the currently employed dichotomous threshold measure (BP <140/90 or LDL <100) and examined the association between facility quartile of meeting the current measure and overtreatment using a multilevel logistic model. We examined what proportion of facilities in each quartile were also in the highest quartile of overtreatment.
For the BP measure, 713,790 patients were 18-75 years old and thus eligible for the action measure. 94% passed the measure: 82% because they had a BP<140/90 at the visit; and an additional 12% with BP>140/90 and appropriate clinical actions. Facility pass rates varied from 76.6% to 98.7% (p<0.001). Among all diabetics, 249,109 (25.5%) had a BP<140/65; of these, 76,766 (7.9% of all diabetics) had potential overtreatment. Facility rates of overtreatment varied from 3.3% to 17.8% (p<0.001). Facilities with higher rates of meeting the current threshold measure (<140/90 mm Hg) had higher rates of potential overtreatment (p<0.001).
For the Lipid measure, our current results show 601,908 patients were 50-75 years old and thus eligible for the clinical action measure. 492,944 (81.9%) passed the clinical action measure: 64.1% with an index LDL <100 mg/dL; 6.6% with an index LDL >=100 mg/dL and on a moderate dose statin; 6.3% with an index LDL >=100 mg/dL and appropriate clinical action; and 4.9% with no index LDL but prescribed a moderate dose statin. Facility pass rates varied substantially from 33% to 94% (p < .001). 123,343 patients without IHD were on high dose statins and thus potentially overtreated - representing 14.1% of all diabetic Veterans. Rates of potential overtreatment by facility varied substantially, ranging from 6% to 25% (p < 0.001). Facilities with higher rates of meeting the current threshold measure (LDL <100 mg/dL) had higher rates of potential overtreatment (p < .001).
We demonstrated use of two tightly-linked clinical action measures among VA patients with diabetes seen in nearly 900 facilities of care. Our findings illustrate that some patients receiving care in the VA may be potentially overtreated; providers and facilities may be using high dose statins - or intensifying BP meds even for those with a low diastolic blood pressure - in order to achieve the current threshold dichotomous targets. Use of the tightly-linked clinical action measures has the potential to reduce overtreatment and the potential harms associated with it.
External Links for this Project
- Vimalananda VG, Miller DR, Hofer TP, Holleman RG, Klamerus ML, Kerr EA. Accounting for clinical action reduces estimates of gender disparities in lipid management for diabetic veterans. Journal of general internal medicine. 2013 Jul 1; 28 Suppl 2:S529-35. [view]
- O'Connor PJ, Bodkin NL, Fradkin J, Glasgow RE, Greenfield S, Gregg E, Kerr EA, Pawlson LG, Selby JV, Sutherland JE, Taylor ML, Wysham CH. Diabetes performance measures: current status and future directions. Diabetes Care. 2011 Jul 1; 34(7):1651-9. [view]
- Kerr EA, Lucatorto MA, Holleman R, Hogan MM, Klamerus ML, Hofer TP, VA Diabetes Quality Enhancement Research Initiative (QUERI) Workgroup on Clinical Action Measures. Monitoring performance for blood pressure management among patients with diabetes mellitus: too much of a good thing? Archives of internal medicine. 2012 Jun 25; 172(12):938-45. [view]
- Kerr EA. Blood Pressure Control in a High Performing Health Care System: Are We Overtreating? Paper presented at: AcademyHealth Annual Research Meeting; 2011 Jun 14; Seattle, WA. [view]
- Kerr EA. Clinical Action Measures for Diabetes: Moving toward Second Generation eMeasures. Paper presented at: VA Office of Analytics and Informatics National Performance Measurement Workgroup; 2011 May 18; Washington, DC. [view]
- Kerr EA. Developing More Clinically Meaningful Performance Measures: What we can learn and where do we have to go? Paper presented at: VA Northwest Center for Outcomes Research in Older Adults Colloquium; 2011 Jun 15; Seattle, WA. [view]
- Kerr EA, Pogach LM, Krein SL, Hofer TP. Developing more meaningful performance measures for VA: a partnership between OQP and QUERI. Poster session presented at: VA “One Team toward One Dream” National Quality, Patient Safety and Systems Redesign Conference; 2010 Jun 20; San Francisco, CA. [view]
- Kerr EA. Improving Quality of Care through Health Services Research. Paper presented at: RWJ Foundation Clinical Scholars Program Annual Meeting; 2011 Nov 9; Crystal City, VA. [view]
- Kerr EA. Moving Toward Patient Centered Performance Measurement : how VHA can lead. Paper presented at: VA Leadership Board Annual Meeting; 2012 May 15; Washington, DC. [view]
- Kerr EA. Partnership Research in Performance Measurement. Paper presented at: VA HSR&D / QUERI National Meeting; 2012 Jul 17; National Harbor, MD. [view]
- Kerr EA. Performance Measures for Hypertension: Are We Overtreating? Paper presented at: VA HSR&D National Meeting; 2011 Feb 17; Washington, DC. [view]
- Kerr EA. Quality Measurement in the Era of Meaningful Use: Pitfalls and Opportunities. Paper presented at: Society of Internal Medicine Annual Meeting; 2012 May 10; Orlando, FL. [view]