Abstract
Despite improvements in diagnosis, prevention and treatment, heart disease remains a common and troublesome burden nationally and globally. We will describe how linked administrative data (admissions, ED, death, PBS, MBS) have been used to evaluate healthcare services and practice at the population
level. Applications to coronary heart disease, heart failure and atrial fibrillation will be presented showing: improvements in measuring disease by accounting for hospital transfers and readmissions, analytical methods such as landmark analysis, use of restricted cubic splines in medication adherence, and machine learning methods and issues to consider (missing data, class imbalance, boosting algorithms, feature selection)
level. Applications to coronary heart disease, heart failure and atrial fibrillation will be presented showing: improvements in measuring disease by accounting for hospital transfers and readmissions, analytical methods such as landmark analysis, use of restricted cubic splines in medication adherence, and machine learning methods and issues to consider (missing data, class imbalance, boosting algorithms, feature selection)
| Original language | English |
|---|---|
| Publication status | Published - Dec 2019 |
| Event | 11th Health Services and Policy Research Conference - Pullman Hotel Auckland, Auckland, New Zealand Duration: 4 Dec 2019 → 6 Dec 2019 Conference number: 11 http://www.healthservicesconference.com.au/hsraanz2019/ |
Conference
| Conference | 11th Health Services and Policy Research Conference |
|---|---|
| Abbreviated title | HSRAANZ 2019 |
| Country/Territory | New Zealand |
| City | Auckland |
| Period | 4/12/19 → 6/12/19 |
| Internet address |