The Use of Cyclic Processes in Medical Decision Making: An Application of the Markov Model

Objectives: We aimed to explain the conceptual basis of the Markov model and to show the use of this model by an example application in medical decision making and medical predicting. Study Design: An example model regarding the effectiveness of St. Jude Total Therapy XIIIB protocol in Acute Lymphoblastic Leukemia (ALL) was hypothesised to evaluate the Markov model concept. The expected remission probabilities in 10 cycles were calculated in a cohort simulation with 10,000 trials, in a cohort in remission in the initial state. Results: Markov models are effective prediction models when the timing of events is important, when the decision problem involves risk over time and when events may happen more than once (as in recurrence). Markov models can be used in estimating such events. As a result of derived model, the remission probability without relaps of any case treatrd with St. Jude Total Therapy XIIIB protocol in ALL disease in the second cycle was found as 43% and it was sharply reduced after this cycle. Conclusion: Cost, effectiveness, and health-related quality of life criteria of clinical strategies can be synthesised by the help of Markov models and used in the calculation of life expectancy, quality adjusted life expectancy and lifetime cost. Turkish Baslik: Saglik Alaninda Karar Vermede Dongusel Sureclerin Kullanimi: Bir Markov Model Uygulamasi Anahtar Kelimeler: Markov model; dongusel surec; karar; kestirim Amac: Calismada Markov modelin kavramsal yapisina aciklik getirilerek, saglik alaninda karar verme ve kestirimde bulunmada kullaniminin ornek bir model ile gosterilmesi amaclandi. Calisma Plani: Markov model kuramini irdelemek icin hipotetik bir calismayla, St. Jude toplam terapi XIIIB protokolunun akut lenfoblastik losemideki etkinligine iliskin ornek bir model kurgulandi. Baslangic durumu remisyon olan 10000 denemelik bir kohort benzetim cozumlemesi duzenlenerek 10 dongu surecinde olgularin remisyonda kalma beklentisi hesaplandi. Bulgular: Olaylarin zamanlamasi onemli oldugunda, zaman asimi karar probleminde risk yaratiyorsa ve olaylar birden fazla kez tekrar edebiliyorsa (nuks gibi), Markov model soz konusu olaylari tahmin etmede kullanilabilir bir modelleme turudur. Kurguladigimiz ornek modelin sonucu olarak, St. Jude toplam terapi XIIIB protokolu ile tedavi goren herhangi bir olgunun 2. dongude relapssiz remisyonda kalma olasiliginin %43 oldugu ve bu donguden sonra artan bir hizda dustugu bulundu. Sonuc: Markov modeller yardimiyla klinik stratejilerin maliyet, etkinlik ve yasam kalitesi olcutleri sentez edilerek, bunlarin sonucunda yasam beklentisi, kalite duzeltmeli yasam beklentisi ve yasam maliyeti hesaplanabilir.