An alternative methodology for the prediction of adherence to anti HIV treatment

摘要: Successful treatment of HIV-positive patients is fundamental to controlling the progression AIDS. Causes failure are either related drug resistance and/or insufficient levels in blood. Severe side effects, coupled with intense nature many regimens, can lead fatigue and consequently periodic or permanent non-adherence. Although non-adherence a recognised problem HIV treatment, it still poorly detected both clinical practice research often based on unreliable information such as self-reports, setting, Medication Events Monitoring System caps prescription refill rates. To meet need for having objective adherence, we propose method using viral load genome sequence data identify amongst patients. With operationally defined sharp increase absence mutation, hypothesised that periods be identified retrospectively observed relationship between changes mutation. Spikes (VL) from time over which VL rises above undetectable level point at decreases by threshold amount. The presence mutations established comparing each reference sequences pairs taken sequentially time, order within around 'treatment change events'. Observed spikes measurements without mutation corresponding then serve proxy indicator It envisaged validation approach will first step road practice. inferred adherence would crucially important feature prediction tools provided practitioners aid daily In addition, distinct characteristics biological markers routinely used assess state disease may adherent non-adherent groups. This latter directly help clinicians differentiate non-responding