Background: Heart attack risk calculators are readily accessible on the world wide web, offering potentially powerful means of health education and risk awareness. Laypersons may be unaware of differences in applicability, risk calculation algorithms and output formats among such calculators. This study assesses the impact of basic web searching terms on type of calculator accessed and on the resulting risk score.
Design: Observational study.
Methods: Seventy-two notional individual risk factor profiles were constructed, based on six combinations of presence or absence of smoking habit, hypercholesterolaemia, mixed hyperlipidaemia, hypertension and family history of premature coronary disease among males and females in age groups 30, 40, 50, 60, 70 and 80 years. The term heart attack risk calculator was entered into the Google, Yahoo, MSN, AltaVista and Excite search engines.
Results: The first five web pages purporting to contain heart attack risk calculators were included. Subpages of URLs leading to duplicate calculators were excluded. All search engines provided similar "hits" for the same search term. Framingham or PROCAM risk prediction models were the templates for all calculators. Different calculators often gave different absolute percentage risk scores for the same notional risk factor profiles. Differences were clinically insignificant in most cases when comparisons were made between bracketed risk scores within 5% of one another. One calculator gave disproportionately high risk estimates for women compared to men with the same risk factor profile and compared to other calculators into which identical risk profiles were entered.
Conclusions: Simple search terms resulted in appropriate "hits". All calculators were based on reputable risk assessment models. There was broad agreement across different calculators for the range of risk factor profiles entered, but one calculator gave inconsistent risk scores.