Article type
Year
Abstract
Introduction/Objective: To undertake a pilot study using decision analysis to assess the value of different medical treatments for menorrhagia (heavy menstrual bleeding). This method enables the 'ranking' of all treatments in terms not only of efficacy, but also the probability of side effects and the 'inconvenience' of the intervention.
Methods: All publications which described randomised controlled trials of medical treatments for menorrhagia were obtained using the search strategy developed by the Menstrual Disorders and Subfertility Review Group, handsearching of specialist journals and searches of a number of electronic databases. Citation lists of included trials and relevant review articles were searched and drug companies were approached for unpublished and original data. The identified trials were then assessed for quality. From each of the included trials, data were extracted for efficacy and side effects. Values were assigned for the probabilities associated with 'response' to treatment (reduction of menstrual blood loss to 'normal' ie <80ml/cycle) and side effects. The probability values used in the decision analysis for response and side effects for each treatment were derived from the median values from these trials. Patients' own value judgments about treatments (utility values) were measured in a pilot study of 20 patients with menorrhagia. In a structured interview, these patients rated each separate treatment in terms of the tradeoff between probability of response, side effects and the inconvenience of the dosage regimen on a direct rating scale. The final utility value for each treatment was then derived from the median patient response. The response and side effects probability values and utility scores for each treatment were then recorded on a decision tree and the expected value calculated at the root of the decision tree. Sensitivity analysis was performed on the results of the decision analysis to assess the validity of the suggested ranking.
Results: The LNG-1US (levonorgestrel releasing intrauterine system) had the highest expected value compared to other medical therapies when efficacy, side effect profile and patient acceptability were all taken into account. The treatment with the highest expected value at the root of the decision tree represents the 'best' option. The ranking of medical therapies from highest to lowest were: LNG-1US, tranexamic acid, NSAIDs, progestagen therapy (21 day course), oral contraceptive pill, danazol, progestagen therapy (luteal phase).
Discussion: Decision analysis has enabled the simultaneous assessment of efficacy, side effect profile and patient acceptability of treatments. The relative ranking of all medical therapies for the treatment of menorrhagia in this way represents another useful way of presenting the evidence.
Methods: All publications which described randomised controlled trials of medical treatments for menorrhagia were obtained using the search strategy developed by the Menstrual Disorders and Subfertility Review Group, handsearching of specialist journals and searches of a number of electronic databases. Citation lists of included trials and relevant review articles were searched and drug companies were approached for unpublished and original data. The identified trials were then assessed for quality. From each of the included trials, data were extracted for efficacy and side effects. Values were assigned for the probabilities associated with 'response' to treatment (reduction of menstrual blood loss to 'normal' ie <80ml/cycle) and side effects. The probability values used in the decision analysis for response and side effects for each treatment were derived from the median values from these trials. Patients' own value judgments about treatments (utility values) were measured in a pilot study of 20 patients with menorrhagia. In a structured interview, these patients rated each separate treatment in terms of the tradeoff between probability of response, side effects and the inconvenience of the dosage regimen on a direct rating scale. The final utility value for each treatment was then derived from the median patient response. The response and side effects probability values and utility scores for each treatment were then recorded on a decision tree and the expected value calculated at the root of the decision tree. Sensitivity analysis was performed on the results of the decision analysis to assess the validity of the suggested ranking.
Results: The LNG-1US (levonorgestrel releasing intrauterine system) had the highest expected value compared to other medical therapies when efficacy, side effect profile and patient acceptability were all taken into account. The treatment with the highest expected value at the root of the decision tree represents the 'best' option. The ranking of medical therapies from highest to lowest were: LNG-1US, tranexamic acid, NSAIDs, progestagen therapy (21 day course), oral contraceptive pill, danazol, progestagen therapy (luteal phase).
Discussion: Decision analysis has enabled the simultaneous assessment of efficacy, side effect profile and patient acceptability of treatments. The relative ranking of all medical therapies for the treatment of menorrhagia in this way represents another useful way of presenting the evidence.