The requirement to respect autonomy - The Royal New Zealand ...

improving performance3. Has there been a change in patient’s meanINR? In other words, did the change in systemresult in patients sitting in a different partof the therapeutic range? Ninety-five percentconfidence intervals were used in establishingsignificance.Approval from an ethics committee was notsought because this was an audit conducted bya health provider for the purpose of qualityimprovement.Results of assessmentThirty-two patients formed the study group. Ofthose, 30 patients had AF and were on warfarinto reduce stroke risk and two were on warfarin toprevent recurrent deep vein thromboses. A summaryof the data is presented in Table 1.Frequency of testingIn the ‘ad hoc’ arm the most frequently testedpatient had an INR performed on average everyfour days and the least frequently tested patientaveraged 46 days between tests. The frequency oftesting ranged between seven and 45 days in the‘standardised’ arm. Across the entire study periodthree patients showed a statistically significant decreasein the frequency of testing after changing tothe ‘standardised’ model, and one patient showeda statistically significant increase in frequency.The remaining 28 patients showed no statisticallysignificant change in frequency of testing.The mean number of days between tests for the‘ad hoc’ arm as a whole was 18 days and for the‘standardised’ arm 17 days. This increased frequencyof testing in the ‘standardised’ arm wasnot statistically significant.Time in the therapeutic rangeIn the ‘ad hoc’ arm the patient with the poorestcontrol spent only 32.1% of the time within thetherapeutic range (of 2.0–3.0) while the best controlledpatient was in this range for 93.9% of thetime. Comparable figures for the ‘standardised’arm were 32.2% and 100%.The mean percentage time spent in the therapeuticrange for the ‘ad hoc’ arm was 65.3(SD 15.7%) and for the ‘standardised’ arm 69.3(SD 17.6%). This difference is not statisticallysignificant.Mean INRIndividual patients had their mean INR comparedover the two study periods. Two patientshad significantly different mean INRs under thedifferent systems; one higher and one lower. Theother 30 patients showed no statistically significantchange in mean INR.Strategies for quality improvementFrequency of testingIt is important to measure frequency of testingwhen examining different methods of anticoagulationmonitoring to ensure that one methoddoes not demonstrate superior results simplyby virtue of more or less regular testing. Theresults show no significant change in frequencyof testing between the two methods of INRmonitoring allowing an equitable comparison.Frequency of testing is also an important measureof patient care. Many patients find venesectionuncomfortable and inconvenient. For this reason,and to minimise cost, we would not wish to testfor good anticoagulation control more frequentlythan is necessary.Time in the therapeutic rangeThe individual variation between patients, withthe best controlled patients spending three timesas long in the therapeutic range as the poorestcontrolled, is a stark reminder of the importanceof individual patient factors (be they biological,social or behavioural) in the control of anticoagulation.Those individuals who showed poorcontrol in the first arm of the study were thesame individuals showing poor control in thesecond arm of the study. Patient to patient variabilitywas noticeably greater than any variabilitydemonstrated between the two methods ofanticoagulation monitoring.320 VOLUME 2 • NUMBER 4 • DECEMBER 2010 J OURNAL OF PRIMARY HEALTH CARE