Leveraging Clinical Digitized Data to Understand Temporal Characteristics and Outcomes of Acute Myocardial...

Few data exist on trends in acute myocardial infarction (AMI) patterns spanning recent epidemiological shifts in low middle-income countries (LMICs). To understand temporal disease patterns of AMI characteristics and outcomes between 1988-2018, we used digitized legacy clinical data at a large tertiary care centre in Pakistan.

We reviewed digital health information capture systems maintained across the Aga Khan University Hospital and obtained structured elements to create a master dataset. We included index admissions of patients >18 years that were discharged between January 1, 1988, and December 31, 2018, with a primary discharge diagnosis of AMI (using ICD-9 diagnoses). The outcome evaluated was in-hospital mortality.Clinical characteristics derived from the electronic database were validated against chart review in a random sample of cases (k 0.53-1.00).

The final population consisted of 14,601 patients of which 30.6% (n = 4,470) were female, 52.4% (n = 7,651) had ST elevation MI and 47.6% (n = 6,950) had non-ST elevation MI. The median (IQR) age at presentation was 61 (52-70) years. Overall unadjusted in-hospital mortality was 10.3%. Across the time period, increasing trends were noted for the following characteristics: age, proportion of women, prevalence of hypertension, diabetes, proportion with NSTEMI (all p_trend < 0.001). In-hospital mortality rates declined significantly between 1988-1997 and 2008-2018 (13.8% to 9.2%, p < 0.001).

The patterns of AMI have changed over the last three decades with a concomitant decline in in-hospital mortality at a tertiary care centre in Pakistan. Clinical digitized data presents a unique opportunity for gaining insights into disease patterns in LMICs.