A° Patient and Method
1° Study population
Our study population consists of patients taken care of at the CRSF for severe dysmenorrhoeic pain, and who received an injection of Botulinum Toxin (BT) after failure of first-line CT.
The main inclusion criteria in our study were: 1) Having received the injection of BT at the Women's Health Research Center (CRSF); and 2) Have a medical file and questionnaires completely completed and returned. Thus, a total of 20 patients were analysed in each group.
All women, aged ³18 years old, having received uterine injection of TB for chronic pelvic pain in French gynecological center l’AVANCEE in Aix-en-Provence, and having necessary data for our analysis in their medical records were enrolled.
2° Study design
This was a retrospective Before-After study using the database of a pilot study assessing the Effectiveness of uterine injection of BT in women with chronic pelvic pain (+/- dysmenorrhea and dyspareunia) after failure of standard treatment (hormonal and analgesic) [14]. Efficacy data: Quality-adjusted life years (QALY), Patient Global Impression Improvement (PGI-I) and Visual Analogic Scale (VAS) as well as healthcare consumption data: Medical treatments; Paraclinical acts (imaging and biological examination); Paramedical acts (physiotherapy and central therapy) were collected prospectively and analysed over a period of 12 months (M0 to M12) during the Before phase, during which the patients received CTs, and during the After phase during which they underwent BT injection. The design of our study is schematized as follows (fig. 1).
The main objective of our study was to evaluate the efficiency of BT in addition to CT compared to the conventional strategy.
3° Effectiveness and Costs assessment
a) Effectiveness assessment
On the clinical level, the scale PGI-I1 was submitted to the patients to assess their satisfaction, and the scales VAS2 and EQ-5D-5L to assess their overall quality of life in relation to the treatments and their health states. The PGI-I scale is graded as follows: much better = 1; better = 2; slightly better = 3; no change = 4; slightly worse = 5; worse = 6; much worse = 7, and has already been used in gynecology [24]. When the patients reported a score between 1 and 2 inclusive, they were considered “ satisfied ” or treated successfully, otherwise “ dissatisfied ”. The VAS scale is scaled from 0 “worst imaginable state of health” to 100 “best imaginable state of health” [25]. The proportions of satisfied patients and the means of the VAS scores as well as the standard deviations were calculated.
The Effectiveness endpoint used to assess efficiency was the QALY, measured using the EQ-5D-5L Health-Related Quality of Life questionnaire validated within the French population [24]. He includes 5 items (Mobility; Autonomy; Current activity; Pain/Discomfort; Anxiety/Depression) with 5 response levels for each: No problem; Mild problems; Moderate problems; Severe problems; Extreme Problems, [24].
Indeed, the QALY is a composite criterion, which results from the weighting of the duration of life by the quality of life associated with this state of health. It is used in health economics to capture the effect on quality of life of various health states [ 26, 27] The quality of life or utility score (U) was measured at M1 and at M12 and calculated according to the following formula: U = (1- U1 –U2 – U3 - U4 – U5 ). This score varies between [0 (worst state); 1 (best state)] and translates the patient's preference or not for his state of health [ 28].
In practice, the effectiveness (QALY) is obtained by multiplying the utility score (U) calculated at M1 = t1 ; M12 = t12 by the phase duration (Before, and After) which is 12 months in our study, according to the following formula: QALY(t1 - t12 ) = [(U(t1) + U(t12)) /2] ´(t12– t 1). Where: t = time duration in a health state (in years); ex: t 1 – t 12 =12 Months = 1 year; U = utility score.
b) Cost assessment
The resources consumed by patients were identified, and quantified (measured) retrospectively from patient files, over a time horizon of 12 months, then monetarily valued (evaluated) from the rates reimbursed by health insurance which generally correspond to 70% of the basic tariff, acts and medications.
The main cost identified in this study are as follows: direct costs medical: Medical consultations and paramedical acts (physiotherapy, central therapy), paraclinical acts (imaging, biological, examinations), pharmacy (drugs and medical devices), hospitalization (outpatient); non-medical direct costs: transport; indirect costs: daily allowances.
The perspective adopted in this study is that of French HI. in the main analysis, the total cost is equal to direct costs (medical and non-medical) added to indirect costs ; thus, the resources under consideration were not restricted to the production of care (direct costs), as suggested by the High Authority of Health (HAH ; or HAS in France), but extended to indirect costs, due to numerous illness-related work stoppages. However, in secondary analysis, total cost are limited to Direct cost (medical and non-medical).
c) Costs valorisation
In practice, the costs have been valued as follows:
- direct medical costs: hospitalization was valued from the Homogeneous Stay Group (HSG)3 = 300.76 euros; the cost of acts (medical, paraclinical, paramedical) or medication equal to basic tariff of (the procedure or medication) multiplied by the reimbursed rate by HI; for example MRI = 69 euros x 70% = 48.3 euros reimbursed by HI;
- non-medical direct costs: We have valued transport costs (TP) for hospitalization by the transport tariff reimbursed/ HI. Which equal to distance travelled in kilometers multiplied by the reimbursed rate by HI, following formulae: = 2*((distance in km * 0,3) *0,65)
- indirect costs: the loss of income was valued by the Daily Allowance (DA) calculated using the following formulae:
DA = Basic Daily Salary (BDS) x 50%; BDS = 3 x gross salary / 91.25. Example for a gross salary of 2000 euros; BDS = 3 x 2000 / 91.25 = 65.75 ; IJ = 65.75 x 50% = 32 euros .
As the time horizon is not greater than 12 months, no updating has been carried out on our results, according to HAS recommendations. All costs are given in euros, relate to year 2021, and all formulae are extracted of website Ameli.fr.
4° Cost-Effectiveness assessment
To assess the efficiency of the innovative treatment, we calculated the incremental cost-effectiveness ratio ( ) of introducing BT in addition to CTs compared to CTs alone, by establishing the ratio of the mean cost difference (denoted ΔC) and the mean effectiveness difference (denoted ΔE) between both the strategies.
Otherwise, the ICER confidence interval was estimated using Fieller 's theorem [Proto, 52;50]. The ceiling ratio, that is the maximum ratio that community is willing to pay for a gain in effectiveness (or Willingness to Pay) was set at €30,0004 per QALY gained. Then, the strategy was deemed efficient when the RCEI was less than €30,000 per additional QALY gained [29, 30 ].
To help institutional decision-making, we also estimated the incremental monetary net benefit (IMNB) which is equal to the difference between the mean effectiveness difference (ΔE), multiplied by the ceiling ratio (λ) and the mean cost difference (ΔC) as follows:
Finally, a cost-effectiveness acceptability curve (CEAC) was developed to show the probability of efficiency of the innovative strategy with respect to standard strategy according to community's willingness to pay defined by the decision maker.
5° Statistical Analysis
Quantitative variables are described in terms of effective, mean, standard deviation and 95% confidence interval of the mean, median, range and interquartile range. Mean costs, mean QALYs and mean days off work were compared for matched subjects (Before-After) by Student's or Wilcoxon's test for paired series, depending on whether the distribution of variables is normal or not.
Qualitative variables are described in terms of number, percentage and 95% confidence interval according to the exact binomial distribution. The proportions of satisfied or dissatisfied patients (PGI-I) were compared for the matched subjects (Before-After) using Mc Nemar 's chi-squared test. Descriptive analysis was performed using R studio software, Version 4.0.
The significance level was considered at 5% for the various statistical tests.
Then, sensitivity analyses was carried out on the parameters likely to impact the ICER to handle the uncertainty on the hypotheses of the model parameters. Thus, an univariate sensitivity analysis on costs and on QALYs was performed by varying different parameters. A Tornado diagram was also created to identify the most influential variables (cost) on the ICER [31,32].
Finally, the statistical uncertainty (due to sampling fluctuations) around the ICER was captured by a probabilistic sensitivity analysis, using, the Monte Carlo simulation method [Article Acupuncture]. Therefore, the initial sample was simulated 1000 times, in order to obtain 1000 mean cost difference and mean QALY difference, the associated incremental cost-effectiveness ratios (ICERs) and the cost-effectiveness acceptability curve that capture the associated uncertainty. Sensitivity analysis was performed using Excel 2017. The 95% confidence region around the ICER was also calculated using the truncated Fieller method ( Siani , C. and Moatti , JP, 2003,) [33].