NICE NG89 recommendations for extended pharmacological thromboprophylaxis

The National VTE Risk Assessment Tool (available from https://www.nice.org.uk/guidance/ng89/resources) has been embedded in practice since the last version of the guideline with a high level of adherence. However, several of the committee members for NICE Guideline 89 (NG89) believed that the tool leads to over prescribing of prophylaxis in medical patients without clear evidence of benefit. In effect, the tool has become a tick box exercise where clinicians view it as an unweighted checklist of risk factors. The committee discussed the crucial need for an appropriate risk tool that will effectively reduce the number of patients being given VTE prophylaxis when they are not at high risk of VTE. A review of all the published tools revealed that none offered clear guidance on how to balance VTE risk and bleeding risk. Consequently, the committee recommended that any published tool could be used and made a recommendation for research into this area.

The National VTE Risk Assessment Tool (available from https://www.nice.org.uk/guidance/ng89/resources) has been embedded in practice since the last version of the guideline with a high level of adherence. However, several of the committee members for NICE Guideline 89 (NG89) believed that the tool leads to over prescribing of prophylaxis in medical patients without clear evidence of benefit. In effect, the tool has become a tick box exercise where clinicians view it as an unweighted checklist of risk factors. Thecommittee discussed the crucial need for an appropriate risk tool that will effectively reduce the number of patients being given VTE prophylaxis when they are not at high risk of VTE. A review of all the published tools revealed that none offered clear guidance on how to balance VTE risk and bleeding risk. Consequently, the committee recommended that any published tool could be used and made a recommendation for research into this area.

Whatever tool is used, there should be a clinical review to make sure that the initial plans for VTE prophylaxis are appropriate and the guideline recommends reassessment at the next consultant review. The correct decision may not have been made initially and circumstances may change. If the balance of VTE risk and bleeding risk has changed, prophylaxis may not always need to be continued, for example if it transpires a patient does not have reduced immobility for 3 days.

While the evidence on who to give prophylaxis is unclear, the evidence on what to give and for how long was based on the included trials. This evidence has not changed much for the general medical population since
the publication of the previous guidance (Clinical Guideline 92, CG92) in 2010 (NICE, 2010, available upon request from the corresponding author). Although no duration was specified in the CG92 recommendation, the duration of prophylaxis in the included trials for the general medical population ranged from 6 to 14 days, with an average duration of 10 days, and, for the general surgical population, the duration ranged from 6 to 11 days, with an average of 7 days. This continued to be the case in the recent guidance (NG89), where none of the included trials that informed the recommendations for the acutely ill medical patients had a duration of <6 days. There was no evidence that shorter courses of prophylaxis of <6 days are effective even though patients are being discharged from hospital earlier. For example, while acutely ill medical patients may go home after 2–3 days they still may be immobile and at risk of VTE. Thus, recommending prophylaxis for a shorter duration would not have been based on any evidence.

Lester et al also question whether this duration of prophylaxis is cost effective and the answer is yes. This is based on the results of the economic modelling work undertaken for CG92 (NICE, 2010), which informed the NG89 recommendations. In that model, all the points flagged by Lester et al have already been taken into account:

  1. The cost of district nurse visits for people unable to inject low molecular weight heparin (LMWH) was included in the model where 8% of people were estimated to require these visits (as opposed to 33% assumed by Lester et al). As explained in CG92 (NICE, 2010), this was based on two surveys, which both found that 8% of patients could not comply with administering their own LMWH prophylaxis (Spahn, 2002; Colwell et al, 2005).
  2. The average duration of prophylaxis used in the economic model was based on the trial evidence. As explained above, for general medical patients this was 10 days.

The results showed that for the general medical population, LMWH was the most cost effective prophylaxis under these assumptions. In fact, a threshold analysis run for the general medical population model also
showed that, even if every patient required district nurse visits to deliver their prophylaxis, LMWH was still cost effective (NICE, 2010).

For the general surgery population, an average duration of 7 days was used in the economic model in CG92. This was also the case in a more recent economic analysis that followed the same methods as CG92 and
informed the recommendation for the abdominal surgery (excluding bariatric surgery) in NG89 (Wade et al, 2015). This analysis found that LMWH for 7 days in combination with anti‐embolism stockings was the most cost effective in this population.

To calculate the cost‐impact of the recommended duration, it was assumed that if current practice was in line with CG92, then this duration should already be used in practice. This led to the conclusion that there will be no cost impact. If, based on Lester et al argument, hospitals have been administering prophylaxis only for the duration of in‐hospital stay, then the cost impact calculation presented by the authors is still considered to be an overestimate as they should be revised to:

  1. Include the cost of district nurse visits for only 8% of the target population rather than a third of patients, as proposed by Lester et al.
  2. Consider that some hospitals will use different risk assessment tools to the National tool [for example, using IMPROVE (https://www.outcomes-umassmed.org/improve/) and Caprini et al (2001)] risk assessment tools would classify fewer medical and surgical patients, respectively, as requiring prophylaxis.
  3. Consider the change in the definition of the target population to those who are ‘acutely ill’ rather than all ‘general’ medical patients.

Taking all the above into account will result in a lower cost impact estimate than calculated by Lester et al. However, it has to be noted that, in line with NICE methods, cost impact should not influence the committee’s recommendation in this particular case, given that both the clinical and economic evidence clearly show that LMWH given for this duration is both clinically‐ and cost‐effective.

Article [Wiley Online Library]

Leave a comment: