Cost-effectiveness of cell salvage and donor blood transfusion during caesarean section: results from a randomised controlled trial

Objectives To evaluate the cost-effectiveness of routine use of cell salvage during caesarean section in mothers at risk of haemorrhage compared with current standard of care. Design Model-based cost-effectiveness evaluation alongside a multicentre randomised controlled trial. Three main analyses were carried out on the trial data: (1) based on the intention-to-treat principle; (2) based on the per-protocol principle; (3) only participants who underwent an emergency caesarean section. Setting 26 obstetric units in the UK. Participants 3028 women at risk of haemorrhage recruited between June 2013 and April 2016. Interventions Cell salvage (intervention) versus routine care without salvage (control). Primary outcome measures Cost-effectiveness based on incremental cost per donor blood transfusion avoided. Results In the intention-to-treat analysis, the mean difference in total costs between cell salvage and standard care was £83. The estimated incremental cost-effectiveness ratio (ICER) was £8110 per donor blood transfusion avoided. For the per-protocol analysis, the mean difference in total costs was £92 and the ICER was £8252. In the emergency caesarean section analysis, the mean difference in total costs was £55 and the ICER was £13 713 per donor blood transfusion avoided. This ICER is driven by the increased probability that these patients would require a higher level of postoperative care and additional surgeries. The results of these analyses were shown to be robust for the majority of deterministic sensitivity analyses. Conclusions The results of the economic evaluation suggest that while routine cell salvage is a marginally more effective strategy than standard care in avoiding a donor blood transfusion, there is uncertainty in relation to whether it is a less or more costly strategy. The lack of long-term data on the health and quality of life of patients in both arms of the trial means that further research is needed to fully understand the cost implications of both strategies. Trial registration number ISRCTN66118656.


62
• Study strengths of this model-based economic evaluation include that it was based 63 on a rigorously conducted RCT and it benefitted from significant clinical and 64 statistical input throughout its design and development. 65 • The analyses were conducted from a healthcare perspective, and the cost and 66 outcome data measures incorporated into the model were collected prospectively 67 during the RCT using forms filled out during the pre-intra-and postoperative phases 68 and at the time of discharge from hospital. 69 • A limitation of the study is that not all potential outcomes have been included 70 because of the limited time scale in the model and the lack of long-term data. 71 • A further limitation of the evaluation is that outcomes were expressed in terms of 72 clinical effectiveness rather than in terms of a standard unit of benefit such as the 73 QALY (quality-adjusted life year).  Full details of the randomised controlled trial (RCT) are reported elsewhere [11]. Briefly, the 117 study was performed in 26 obstetric units. The sample comprised women who were 118 admitted to the labour ward for delivery by emergency or elective caesarean section, with 119 an identifiable increased risk of haemorrhage. Participating women were randomised to 120 either caesarean section with cell salvage (intervention group), or to caesarean section  Intra-operative costs were estimated for each item to arrive at a mean cost per caesarean 157 procedure for each treatment pathway in the model. To estimate the cost of a caesarean 158 section procedure some costs were calculated at the patient level e.g. swab washing and 159 some at the procedural level e.g. drugs used in the caesarean section procedure (see Table   160 2). Postoperative costs were estimated for each item based on their occurrence in each 161 brand of the model to arrive at a mean cost per patient for each branch (see Table 3). Given the objectives of the trial and the duration of follow up, a within trial economic 165 analysis was carried out. The analysis took the perspective of the NHS following current 166 recommendations from NICE [12]. The main economic analysis was a cost-effectiveness 167 analysis (CEA) with results expressed as cost per donor transfusion avoided.
168 169 We carried out three main analyses on the trial data. Analysis 1, referred to as the base-170 case, was based on the intention-to-treat (ITT) principle. In this method, patients are 171 compared within the treatment groups to which they were originally randomised 172 irrespective of the treatment received [13]. This method of analysis allows the estimates to 173 follow real-life scenarios in which patients may not always receive the planned treatment. 174 ITT is recommended to ensure that the benefits of a given intervention are not exaggerated 175 [13]. Analysis 2 was based on the treatment actually received by patients irrespective of 176 randomisation (the 'per protocol' (PP) analysis). Within the SALVO trial patients were 177 randomised to cell salvage or standard care in a 1:1 ratio. However, because some clinicians 178 managing women in the control arm had access to a cell salvage machine, it was possible 179 that women in the control arm could receive cell salvage in place of a donor blood 180 transfusion. A PP analysis was carried out to look at the effect of treatment received on the 181 outcome estimates. Therefore, in Analysis 2 all patients who received cell salvage were 182 compared with those who received standard care, irrespective of the treatment to which 183 they were randomised. Analysis 3 considered only patients who underwent an emergency 184 caesarean section. This analysis was considered necessary as the SALVO trial found that 185 numerically, there was a greater reduction in the rate of transfusion within the emergency 186 F o r p e e r r e v i e w o n l y 10 patient group compared to the elective patient group. This analysis followed the same 187 methodology as the ITT and PP analyses. Deterministic and probabilistic sensitivity analyses 188 (PSAs) were carried out for each analysis to explore the effects of the inherent uncertainty 189 in parameter estimates on model results (summarised in Table 4). 190 191 Patient and public involvement 192 The UK National Childbirth Trust collaborated in the project by providing patient and public 193 input through involvement in trial design and protocol development. Prior to the SALVO 194 trial, a survey was conducted among women who received cell salvage, showing that they 195 perceived the interventions as reassuring, safe, and preferable to donor blood transfusion. A 196 patient representative was a member of the trial steering committee to provide oversight 197 and advice regarding recruitment, dissemination, and general trial management. We are 198 planning to disseminate finding to participants in the form of a newsletter following primary 199 publication of these results.

202
The results of the ITT analysis (Analysis 1) suggest that routine cell salvage is more costly 203 than standard care with the average cost per patient estimated at £1,244. The Incremental 204 Cost-Effectiveness Ratio (ICER) representing the additional cost of routine cell salvage during 205 caesarean section in women at risk of haemorrhage compared to standard care for these 206 women was estimated to be approximately £8,110 per donor blood transfusion avoided. 207 The overall result, based on the PP analysis (Analysis 2), was an ICER of £8,252 per 208 transfusion avoided for cell salvage compared to standard care. The emergency caesarean 209 analysis (Analysis 3) resulted in an ICER of £13,713 per transfusion avoided. See Table 5. The results of the corresponding PSAs are presented using cost-effectiveness acceptability 212 curves (CEACs) to graphically represent uncertainty in the appropriate threshold cost-213 effectiveness value. The CEAC for the ITT analysis ( Figure 2) shows that the probability that 214 routine cell salvage is cost-effective increases as the willingness to pay for a donor blood 215 transfusion avoided increases. If the maximum willingness to pay for a donor blood 216 transfusion avoided was £50,000, the probability of routine cell salvage being cost-effective 217 would be 62%. For Analysis 2 (the PP analysis - Figure 3) the probability that routine cell 218 salvage was cost-effective at a willingness to pay of £50,000, would be 63%. The CEAC for 219 the emergency caesarean analysis (Analysis 3 - Figure 4) shows that the probability that cell 220 salvage is cost-effective remains between 47% and 55% as the willingness to pay for a donor 221 blood transfusion avoided increases.  Table   225 4): 226 (i) The main analyses used costs for consumables based on a particular model of the 227 cell saver machine. To assess the difference that variation in these estimates would 228 make, the unit costs were replaced with unit costs obtained from the NICE costing 229 statement for blood transfusion [14]. This had a marginal impact on the ICER in each 230 analysis. The impact of the inclusion of acquisition costs for a cell salvage machine 231 was also explored and found this had only a minimal impact on the ICER in each 232 analysis. In the trial, 202 centres used a cell salvage machine that did not require 233 consumables. Probabilistic Sensitivity Analysis (PSA) suggests that at an arbitrary willingness to pay 272 threshold of £50,000, the probability of routine cell salvage being cost-effective is 62%. 273 These results suggest that cell salvage is not likely to be considered a cost-effective 274 alternative to donor blood whilst donor blood is provided freely by individuals and supply is 275 not typically restricted even though it is not a costless provision to the NHS.  A per protocol analysis produced an ICER of £8,252 per transfusion avoided, but this result   285 should be considered with caution as the population in this analysis is a subset of the ITT 286 population who completed the study without any major protocol violations [13]. In clinical 287 practice uptake of cell salvage is unlikely to be "per-protocol". In a third analysis, looking at 288 emergency caesareans only, cell salvage appears to be more effective than standard care for 289 avoiding a donor blood transfusion but the resulting ICER of £13,713 is driven by the 290 increased probability that these patients will require a higher level of post-operative care 291 and additional surgeries.

293
Strengths and limitations of the study 294 The strength of this model based economic evaluation is that it was based on a rigorously 295 conducted Randomised Controlled Trial (RCT). The cost and outcome data measures that 296 were incorporated into the model were collected prospectively during the RCT using forms 297 filled out at the pre-, intra-, and post-operative phase and at the time of discharge from 298 hospital. In addition the economic evaluation benefitted from significant clinical and 299 statistical input throughout its design and development. All assumptions used in the model 300 were agreed with the trial team before the analysis was carried out and without knowledge 301 of how these assumptions would affect the results. In terms of limitations, it was not possible to account for long-term implications relating to 304 maternal fetal exposure as data relating to this wasn't available from the trial. In addition, 305 information relating to the clinical status and care of the infant was not included in the 306 analysis. A further limitation of the study is that outcomes were expressed in terms of 307  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  Meaning of the study 322 The results of the economic evaluation suggest that while routine cell salvage is a marginally 323 more effective strategy than standard care in avoiding a donor blood transfusion there is 324 uncertainty in relation to whether it is a less or more costly strategy. Under the conditions 325 reported here, for a high income country (HIC) such as the UK, where donor blood is 326 typically available, cell salvage is unlikely to be considered a cost-effective alternative to the 327 provision of donor blood by the health service. However, in lower and middle income 328 countries (LMIC) where the provision of a safe and secure blood supply may be more 329 challenging the relative cost-effectiveness may be very different. In addition, the lack of long 330 term data on the health and quality of life of patients in both arms of the trial means that 331  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39          Adverse events 3 0 n/a n/a n/a Assumption that transfusion would be discontinued in the event of    Results: In the intention-to-treat analysis the mean difference in total costs between cell 46 salvage and standard care was £83. The estimated incremental cost-effectiveness ratio 47 (ICER) was £8,110 per donor blood transfusion avoided. For the per-protocol analysis the 48 mean difference in total costs was £92 and the ICER was £8,252. In the emergency 49 caesarean section analysis the mean difference in total costs was £55 and the ICER was 50 £13,713 per donor blood transfusion avoided. This ICER is driven by the increased 51 probability that these patients would require a higher level of post-operative care and 52 additional surgeries. The results of these analyses were shown to be robust for the majority 53 of deterministic sensitivity analyses.        Williamstown, MA, USA) to represent the alternative strategies. The pathways of the model 142 (see Figure 1) represent, as far as possible, the clinical procedures carried out in the study.

144
Clinical data used in the model 145 The model combines the probability of a woman following a particular path and the 146 associated costs. Probabilities, detailed in Table 1 some at the procedural level e.g. drugs used in the caesarean section procedure (see Table   169 2). Postoperative costs were estimated for each item based on their occurrence in each 170 brand of the model to arrive at a mean cost per patient for each branch (see Table 3).       The results of the corresponding PSAs are presented using cost-effectiveness acceptability 231 curves (CEACs) to graphically represent uncertainty in the appropriate threshold cost-232 effectiveness value. The CEAC for the ITT analysis ( Figure 2) shows that the probability that 233 routine cell salvage is cost-effective increases as the willingness to pay for a donor blood 234 transfusion avoided increases. If the maximum willingness to pay for a donor blood 235 transfusion avoided was £50,000, the probability of routine cell salvage being cost-effective 236 would be 62%. For Analysis 2 (the PP analysis - Figure 2) the probability that routine cell 237 salvage was cost-effective at a willingness to pay of £50,000, would be 63%. The CEAC for 238 the emergency caesarean analysis (Analysis 3 - Figure 2) shows that the probability that cell 239 salvage is cost-effective remains between 47% and 55% as the willingness to pay for a donor 240 blood transfusion avoided increases.   A number of deterministic sensitivity analyses were conducted in each analysis (see Table   248    Probabilistic Sensitivity Analysis (PSA) suggests that at an arbitrary willingness to pay 297 threshold of £50,000, the probability of routine cell salvage being cost-effective is 62%. 298 Based on the results of this trial, in a UK setting the results suggest that cell salvage is not 299 likely to be considered a cost-effective alternative to donor blood whilst donor blood is 300 provided freely by individuals and supply is not typically restricted even though it is not a 301 costless provision to the NHS.   practice uptake of cell salvage is unlikely to be "per-protocol". In a third analysis, looking at 314 emergency caesareans only, cell salvage appears to be more effective than standard care for 315 avoiding a donor blood transfusion but the resulting ICER of £13,713 is driven by the 316 increased probability that these patients will require a higher level of post-operative care 317 and additional surgeries.         Table 1 Probabilities used in the model 490 Table 2 Intraoperative resource use and costs  491  Table 3 Postoperative resource use and costs  492  Table 4 Results 493 Table 5 Deterministic sensitivity analysis 494 495   1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59   Clinical data used in the model 143 The model combines the probability of a woman following a particular path and the 144 associated costs. Probabilities, detailed in Table 1  some at the procedural level e.g. drugs used in the caesarean section procedure (see Table   165 2). Postoperative costs were estimated for each item based on their occurrence in each 166 brand of the model to arrive at a mean cost per patient for each branch (see Table 3).  practice uptake of cell salvage is unlikely to be "per-protocol". In a third analysis, looking at 287 emergency caesareans only, cell salvage appears to be more effective than standard care for 288 avoiding a donor blood transfusion but the resulting ICER of £13,713 is driven by the 289 increased probability that these patients will require a higher level of post-operative care 290 and additional surgeries. In terms of limitations, it was not possible to account for long-term implications relating to 303 maternal fetal exposure as data relating to this wasn't available from the trial. In addition, 304 information relating to the clinical status and care of the infant was not included in the 305 analysis. A further limitation of the study is that outcomes were expressed in terms of 306 clinical effectiveness rather than in terms of a standard unit of benefit such as the quality-307 adjusted life-year (QALY). Finally, the use of platelets and other blood products has not been 308 included in the study. However, the results of the sensitivity and threshold analyses           Figure 1 Decision tree structure 483 Figure 2 Cost-effectiveness acceptability curve for donor blood transfusion avoided 484  1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60