Arthroplasty: Effects of a Patient Blood Management Program
Arthroplasty: Effects of a Patient Blood Management Program
This quality improvement report describes the design of a patient blood management programme in response to a combination of local data, national guidelines, and international opinion. The resulting systematic approach to optimizing patients' red cell mass and limiting Hb loss perioperatively was associated with lower ABT rates, shorter LOS, and a reduction in re-attendance after elective arthroplasty.
The relationships between anaemia, the factors commonly associated with anaemia (e.g. age, systemic disease), and ABT are complex. Their relative contributions to outcome are thus difficult to specify though preoperative anaemia is common, easy to identify, and treatable. There is mounting evidence that preoperative anaemia and ABT are additive risk factors for poor postoperative outcome, including mortality. The relationship which we found between anaemia, ABT, and poor outcomes is not new, although it has not to our knowledge been previously described in UK practice. Two Scottish groups found preoperative Hb levels <12 gdl and low body weight to predict ABT needs, but considered no other outcomes.
Despite this relative paucity of UK outcome data, the Department of Health endorses blood management strategies. Donor blood is a limited resource. The size of the donor pool may also decrease or become precarious because of changes in donor characteristics or infective pandemics. Donation is frequently sought on the basis that the gift is used in cases where medical correction of anaemia is impossible. Blood management may thus be said to be worthwhile in itself for the sake of donors and for public health reasons, even if further work on the risk:benefit balance of ABT for recipients in some situations is still needed.
A key consideration in our programme to implement the national guidance was that it started with an analysis of the possible relationships between anaemia and outcomes in our own hands. Overall, our baseline outcomes for transfusion, LOS, and re-admission were comparable with those from elsewhere in England and Europe. The median nadir Hb value associated with transfusion at baseline was also >1 g dl lower than in the OSTHEO study, a large multicentre investigation of transfusion for arthroplasty across Europe. Even though we cannot be sure that the nadir Hb was the value that triggered transfusion, this was always the last pre-transfusion Hb when we had prospective data available. In our view, this indicates that our baseline transfusion practice was not inappropriately liberal.
Our uniform overall outcomes masked a significant effect of preoperative anaemia and perioperative ABT. We therefore felt obliged to implement the NHSBT and DOH recommendations on blood management, but could find no published algorithm directly applicable to UK practice. Observational studies and one cluster-randomized trial, respectively, from Europe and Canada have shown encouraging results. However, these made a widespread use of autologous pre-donation which is no longer offered routinely in the UK. We therefore had to design our own algorithm, using our local baseline data to make the case for change and to modify previously published algorithms in a rational fashion.
Another consideration was that any programme should be practicable and acceptable to our colleagues. We therefore purposely left much to their discretion, rather than prescribe to them how to manage their own patients. This included decisions on whether to cancel arthroplasty for investigation of any anaemia, aspects of surgical practice (e.g. wound drains), and, for example, the use of cell salvage or intraoperative anti-fibrinolytics. In order to limit the up-front cost of the project (i.e. cash spent on drugs in the hope of clinical improvement and savings further on), we placed emphasis on treating anaemia only where our own baseline data suggested that an outcome improvement was likely. Our criteria for the use of iron were consequently more liberal than those traditionally used for the diagnosis of iron deficiency anaemia per se. The rationale for this is increasing recognition that joint inflammation-related iron sequestration may make serum ferritin values falsely high even in the presence of iron deficiency at the bone marrow level. Iron is also relatively cheap. Most patients had oral iron because we reserved the parenteral formulation for patients intolerant of, non-responsive to oral iron, or to avoid postponement of surgery. We also reserved the more costly treatment (EPO) for patients with preoperative Hb<12 g dl in whom iron deficiency could be excluded. This approach is relatively conservative compared with previous studies.
The Network for Advancement of Transfusion Alternatives (NATA) published blood management guidelines for orthopaedic surgery after the start of our programme. The authors cite extensive evidence from mixed surgical populations and recommend a standardized approach to the evaluation and management of elective orthopaedic patients. Our algorithm is different in some respects. NATA recommend that WHO criteria for anaemia diagnosis also be used as treatment thresholds. These thresholds are gender-specific, but transfusion triggers suggested by, for example, the British Committee for Standards in Haematology or used in randomized trials are not. In our series, women and men lost similar amounts of Hb (and hence proportionally similar blood volumes) perioperatively. However, women presented with lower preoperative Hb values and subsequently had higher transfusion rates. Gender-specific treatment thresholds may therefore expose women to disproportionate risk of transfusion and possible consequent morbidity. We rather based our treatment thresholds on the scheduled procedure, since THR was substantially more likely to lead to ABT than TKR in our series. NATA also recommends that the initial Hb level be performed as close as possible to 28 days before surgery and that further investigation proceeds only if the patient is anaemic. We preferred on practical grounds to submit patients to a standard battery of tests including Hb and ferritin as soon as the surgical decision was made to proceed to arthroplasty, in order to fit in with our existing pathways.
On this background, our algorithm delivered good results, being associated with >50% improvements in our primary process and outcome measures, without leading to many patients having their surgery postponed (seven in total). Preoperative anaemia prevalence decreased from 26% at first assessment to 10% after treatment, on an intention-to-treat basis. The modest (0.5–0.7 g dl) observed decrease in Hb loss would be consistent with an increased use of intraoperative blood conservation measures and/or more careful surgery, but we did not specifically audit individual practice, in accordance with our agreement with colleagues when we started the programme. The combination of less preoperative anaemia and less Hb loss leads to a two-thirds reduction in ABT for THR and effective cessation of ABT for TKR, without a decrease in patients' postoperative Hb levels.
In addition to reducing the burden we placed on blood donors, we observed improved measures of patient outcome, namely LOS and re-admission. Since these improved concurrently, it is likely that this was a real outcome change, rather than simply artifact related to the inverse relationship between LOS and re-admissions in the context of elective surgery. The observed improvements occurred, despite the fact that the post-implementation group was significantly older than the baseline group, which in itself would be a risk factor for prolonged LOS and complications.
The drug cost of the programme was £16 695 over ~8 months. Of this, £12 625 was offset by savings on the purchase of red cells, making the net 'cash cost' of implementing the algorithm ~£4070 for a cohort of 281 patients. These measures further do not allow for nursing time, consumable equipment (e.g. blood giving sets vs cheaper fluid sets), repeat blood tests, or the treatment of any complications related to transfusion. While these additional activity-based costs of transfusion are difficult to quantify, recent data indicate that they may be as much as four times the product cost. We did not quantify the cost of enteral iron or intraoperative tranexamic acid as both are inexpensive drugs. We also specified the use of relatively expensive formulations of EPO and parenteral iron for reasons of convenience while the pathway was still under evaluation. It is therefore likely that implementing the algorithm was cost-saving once total in-hospital costs are accounted for, and it would be possible for hospitals to design algorithms that are cost-saving even in 'cash terms' by the use of cheaper drug formulations. Although they are surrogate markers of outcome, LOS and re-admission rates are also widely used as a discriminator between hospitals, and proposals have been made to penalize hospitals financially for re-admissions, both in the NHS and USA. Our observed LOS and re-admission reductions may therefore represent an additional cost-saving from blood management.
This audit has important weaknesses. Our data are observational. The observed improvements may constitute a Hawthorne effect (whereby outcomes are improved by close observation) rather than an effect of the blood management algorithm itself. We also cannot comment on what the relative contributions of iron therapy, EPO, and the intraoperative measures were. We collected no data on medical co-morbidities, intraoperative blood loss, patient weight, fluid management, surgical approaches, complications, or the reasons for health-care resource use after discharge, since these are not coded into our hospital databases. Confounding and bias is therefore possible.
However, our audit cycle also has strengths. We collected a substantial baseline data set from records subject to regular external audit, in a way which is more accurate than notes review. Our exclusion criteria ensured high specificity and the resultant series was similar in size and ratio of THR:TKR to that which we reported to a national register, that is, its sensitivity was also good. We could control statistically for potential confounders when we determined the baseline predictors of outcome. We could therefore accurately describe our outcomes and use the baseline group as a control for evaluating the programme's effects. It is possible that our results underestimate the magnitude of change that can be achieved, since (due to having other commitments) we were unable to personally see and start treatment for a substantial number of anaemic patients as our protocol required. We conducted our before-and-after comparisons on the basis of intention to treat.
We conclude that preoperative Hb is an easily identifiable predictor of poor outcome after arthroplasty in routine UK practice and that preoperative anaemia treatment is feasible in this setting. The introduction of a patient blood management algorithm was associated with a reduction in the demand we made on blood donors and improvements in patient outcome; LOS and re-admissions both decreased. Randomized trials are necessary to determine whether blood management programmes improve patient outcomes other than ABT rate, and what the optimal algorithm is in terms of cost and patient outcome.
Discussion
This quality improvement report describes the design of a patient blood management programme in response to a combination of local data, national guidelines, and international opinion. The resulting systematic approach to optimizing patients' red cell mass and limiting Hb loss perioperatively was associated with lower ABT rates, shorter LOS, and a reduction in re-attendance after elective arthroplasty.
The relationships between anaemia, the factors commonly associated with anaemia (e.g. age, systemic disease), and ABT are complex. Their relative contributions to outcome are thus difficult to specify though preoperative anaemia is common, easy to identify, and treatable. There is mounting evidence that preoperative anaemia and ABT are additive risk factors for poor postoperative outcome, including mortality. The relationship which we found between anaemia, ABT, and poor outcomes is not new, although it has not to our knowledge been previously described in UK practice. Two Scottish groups found preoperative Hb levels <12 gdl and low body weight to predict ABT needs, but considered no other outcomes.
Despite this relative paucity of UK outcome data, the Department of Health endorses blood management strategies. Donor blood is a limited resource. The size of the donor pool may also decrease or become precarious because of changes in donor characteristics or infective pandemics. Donation is frequently sought on the basis that the gift is used in cases where medical correction of anaemia is impossible. Blood management may thus be said to be worthwhile in itself for the sake of donors and for public health reasons, even if further work on the risk:benefit balance of ABT for recipients in some situations is still needed.
A key consideration in our programme to implement the national guidance was that it started with an analysis of the possible relationships between anaemia and outcomes in our own hands. Overall, our baseline outcomes for transfusion, LOS, and re-admission were comparable with those from elsewhere in England and Europe. The median nadir Hb value associated with transfusion at baseline was also >1 g dl lower than in the OSTHEO study, a large multicentre investigation of transfusion for arthroplasty across Europe. Even though we cannot be sure that the nadir Hb was the value that triggered transfusion, this was always the last pre-transfusion Hb when we had prospective data available. In our view, this indicates that our baseline transfusion practice was not inappropriately liberal.
Our uniform overall outcomes masked a significant effect of preoperative anaemia and perioperative ABT. We therefore felt obliged to implement the NHSBT and DOH recommendations on blood management, but could find no published algorithm directly applicable to UK practice. Observational studies and one cluster-randomized trial, respectively, from Europe and Canada have shown encouraging results. However, these made a widespread use of autologous pre-donation which is no longer offered routinely in the UK. We therefore had to design our own algorithm, using our local baseline data to make the case for change and to modify previously published algorithms in a rational fashion.
Another consideration was that any programme should be practicable and acceptable to our colleagues. We therefore purposely left much to their discretion, rather than prescribe to them how to manage their own patients. This included decisions on whether to cancel arthroplasty for investigation of any anaemia, aspects of surgical practice (e.g. wound drains), and, for example, the use of cell salvage or intraoperative anti-fibrinolytics. In order to limit the up-front cost of the project (i.e. cash spent on drugs in the hope of clinical improvement and savings further on), we placed emphasis on treating anaemia only where our own baseline data suggested that an outcome improvement was likely. Our criteria for the use of iron were consequently more liberal than those traditionally used for the diagnosis of iron deficiency anaemia per se. The rationale for this is increasing recognition that joint inflammation-related iron sequestration may make serum ferritin values falsely high even in the presence of iron deficiency at the bone marrow level. Iron is also relatively cheap. Most patients had oral iron because we reserved the parenteral formulation for patients intolerant of, non-responsive to oral iron, or to avoid postponement of surgery. We also reserved the more costly treatment (EPO) for patients with preoperative Hb<12 g dl in whom iron deficiency could be excluded. This approach is relatively conservative compared with previous studies.
The Network for Advancement of Transfusion Alternatives (NATA) published blood management guidelines for orthopaedic surgery after the start of our programme. The authors cite extensive evidence from mixed surgical populations and recommend a standardized approach to the evaluation and management of elective orthopaedic patients. Our algorithm is different in some respects. NATA recommend that WHO criteria for anaemia diagnosis also be used as treatment thresholds. These thresholds are gender-specific, but transfusion triggers suggested by, for example, the British Committee for Standards in Haematology or used in randomized trials are not. In our series, women and men lost similar amounts of Hb (and hence proportionally similar blood volumes) perioperatively. However, women presented with lower preoperative Hb values and subsequently had higher transfusion rates. Gender-specific treatment thresholds may therefore expose women to disproportionate risk of transfusion and possible consequent morbidity. We rather based our treatment thresholds on the scheduled procedure, since THR was substantially more likely to lead to ABT than TKR in our series. NATA also recommends that the initial Hb level be performed as close as possible to 28 days before surgery and that further investigation proceeds only if the patient is anaemic. We preferred on practical grounds to submit patients to a standard battery of tests including Hb and ferritin as soon as the surgical decision was made to proceed to arthroplasty, in order to fit in with our existing pathways.
On this background, our algorithm delivered good results, being associated with >50% improvements in our primary process and outcome measures, without leading to many patients having their surgery postponed (seven in total). Preoperative anaemia prevalence decreased from 26% at first assessment to 10% after treatment, on an intention-to-treat basis. The modest (0.5–0.7 g dl) observed decrease in Hb loss would be consistent with an increased use of intraoperative blood conservation measures and/or more careful surgery, but we did not specifically audit individual practice, in accordance with our agreement with colleagues when we started the programme. The combination of less preoperative anaemia and less Hb loss leads to a two-thirds reduction in ABT for THR and effective cessation of ABT for TKR, without a decrease in patients' postoperative Hb levels.
In addition to reducing the burden we placed on blood donors, we observed improved measures of patient outcome, namely LOS and re-admission. Since these improved concurrently, it is likely that this was a real outcome change, rather than simply artifact related to the inverse relationship between LOS and re-admissions in the context of elective surgery. The observed improvements occurred, despite the fact that the post-implementation group was significantly older than the baseline group, which in itself would be a risk factor for prolonged LOS and complications.
The drug cost of the programme was £16 695 over ~8 months. Of this, £12 625 was offset by savings on the purchase of red cells, making the net 'cash cost' of implementing the algorithm ~£4070 for a cohort of 281 patients. These measures further do not allow for nursing time, consumable equipment (e.g. blood giving sets vs cheaper fluid sets), repeat blood tests, or the treatment of any complications related to transfusion. While these additional activity-based costs of transfusion are difficult to quantify, recent data indicate that they may be as much as four times the product cost. We did not quantify the cost of enteral iron or intraoperative tranexamic acid as both are inexpensive drugs. We also specified the use of relatively expensive formulations of EPO and parenteral iron for reasons of convenience while the pathway was still under evaluation. It is therefore likely that implementing the algorithm was cost-saving once total in-hospital costs are accounted for, and it would be possible for hospitals to design algorithms that are cost-saving even in 'cash terms' by the use of cheaper drug formulations. Although they are surrogate markers of outcome, LOS and re-admission rates are also widely used as a discriminator between hospitals, and proposals have been made to penalize hospitals financially for re-admissions, both in the NHS and USA. Our observed LOS and re-admission reductions may therefore represent an additional cost-saving from blood management.
This audit has important weaknesses. Our data are observational. The observed improvements may constitute a Hawthorne effect (whereby outcomes are improved by close observation) rather than an effect of the blood management algorithm itself. We also cannot comment on what the relative contributions of iron therapy, EPO, and the intraoperative measures were. We collected no data on medical co-morbidities, intraoperative blood loss, patient weight, fluid management, surgical approaches, complications, or the reasons for health-care resource use after discharge, since these are not coded into our hospital databases. Confounding and bias is therefore possible.
However, our audit cycle also has strengths. We collected a substantial baseline data set from records subject to regular external audit, in a way which is more accurate than notes review. Our exclusion criteria ensured high specificity and the resultant series was similar in size and ratio of THR:TKR to that which we reported to a national register, that is, its sensitivity was also good. We could control statistically for potential confounders when we determined the baseline predictors of outcome. We could therefore accurately describe our outcomes and use the baseline group as a control for evaluating the programme's effects. It is possible that our results underestimate the magnitude of change that can be achieved, since (due to having other commitments) we were unable to personally see and start treatment for a substantial number of anaemic patients as our protocol required. We conducted our before-and-after comparisons on the basis of intention to treat.
We conclude that preoperative Hb is an easily identifiable predictor of poor outcome after arthroplasty in routine UK practice and that preoperative anaemia treatment is feasible in this setting. The introduction of a patient blood management algorithm was associated with a reduction in the demand we made on blood donors and improvements in patient outcome; LOS and re-admissions both decreased. Randomized trials are necessary to determine whether blood management programmes improve patient outcomes other than ABT rate, and what the optimal algorithm is in terms of cost and patient outcome.
