Assessment of electronic medical records. Relationship between process indicators measured using electronic records and intermediate health outcomes in patients with hypertension

Objective: To study relationship between institutional process indicators (measured using electronic records) and intermediate outcomes of patients with hypertension. Design: Cross-sectional epidemiological study. Setting: Primary Care Health District 1. Madrid. 2010. Patients: All patients with hypertension. n = 80,306. Main measurements: Variables. Independent. Institutional process indicators. Dependent. Intermediate outcomes: blood pressure within target limits, LDL-cholesterol, tobacco and weight and detected complications. Confounding. Age, gender, co-morbidity, drugs and professional variables. Results: The BP of 55.1% (SE 0.2%) of patients was within target limits. Bivariate analysis and multivariate logistic regression showed that the recording of some process indicators was associated with an increase in the probability to achieve targets in intermediate outcomes: smoking advice (OR: 1.69, 95% CI: 1.61 - 1.77), reviewing personal history (OR: 1.54, 95% CI:1.42-1.68), increase was less or biased: BP (OR: 1.19, 95% CI: 1.14-1.25), sodium and potassium (OR: 1.14, 95% CI:1.09-1.19), BMI (OR 1.08, 95% CI:1.04-1.12); also diabetes, edema, and creatinine, but there was timing bias. The relationship between other indicators (those oriented to lifestyle, family history, classification, urine examination, reviewing of drug therapy, LDL, electrocardiogram and cardiac auscultation) and a higher probability to achieve targets was not found. Conclusions: In hypertension, some institutional process indicators measured on electronic records were associated with an increase in the probability to achieve targets in intermediate outcomes. No relationship was found between other indicators. This suggests maintaining process and outcome measurement, to include the impact of interventions, to prioritize improvements in process indicators that show low performance and high impact and to remove or to change process indicators where no relationship is found. (c) 2011 Elsevier Espana, S.L. All rights reserved.