This was prompted by some new information that’s been released by ACC under the Official Information Act, regarding their funding of acupuncture treatments.
ACC reports spending over $25 million per year on acupuncture, even though ACC’s reviews of the evidence for acupuncture have been largely inconclusive or negative. There were only three types of injury for which they have concluded acupuncture may be effective:
- Frozen shoulder: The ACC guideline titled, ‘Diagnosis and Management of Soft Tissue Shoulder Injuries and Related Disorders’ (2004) states that there is “some evidence that exercise and acupuncture, compared with exercise alone, may lead to better outcomes.”
- Chronic neck and shoulder pain: The ACC review, ‘Pragmatic Evidence Based Review: The efficacy of acupuncture in the management of musculoskeletal pain’ (2011) concludes that “the evidence for the effectiveness of acupuncture is most convincing for the treatment of chronic neck and shoulder pain. In terms of other injuries, the evidence is either inconclusive or insufficient.”
Until recently, the only available breakdown of ACC’s spending on acupuncture treatments was categorised by “injury diagnosis”. Unfortunately, this breakdown is not very useful because it lumps 94% of acupuncture spending into a single treatment category:
|Amputation / Enucleation||$3,798|
|Concussion / Brain Injury||$62,738|
|Foreign body in Orifice / Eye||$4,517|
|Fracture / Dislocation||$662,598|
|Inhalation / Ingestion||$907|
|Laceration / Puncture Wound||$317,251|
|Mental Injury / Nervous Shock||$170|
|Soft Tissue Injury||$24,788,178|
Earlier this year, I met with someone from ACC to discuss what data is available that might help me answer the question of whether or not ACC’s funding of acupuncture is supported by the conclusions of their evidence-based reviews. They suggested that I ask what the top read codes are that are used for acupuncture treatment in ACC claims.
In ACC’s terminology, a ‘read code’ is a five character code that denotes a specific injury type. For example, “S572.” denotes a lumbar sprain, whereas “TE532” means a toxic reaction to a bee sting.
Following this meeting, I sent another Official Information Act request to ACC. I asked for the number of accepted claims and cost of treatment of acupuncture in 2014/15 categorised by read code, and for any significant confounding factors that would make the data difficult to interpret. That was something that had been discussed at my meeting with ACC earlier, so I knew the best I was going to be able to get was an estimate, and wanted to make sure I knew just what the information I’d be given would and would not mean.
To answer the question of confounding factors, ACC explained in their response that they had categorised claims by their primary read code, and that this information isn’t able to tell me exactly how acupuncture was used in individual claims:
The read code information provided in this response records the primary read code of every claim that has received a payment for acupuncture treatment. As you [are] aware, there can be more than one read code under a single claim.
The read code information alone does not indicate how acupuncture was used in individual claims, because it is not possible to determine whether acupuncture was used in relation to the primary read code or some other read code on the claim. This would only be possible by reviewing individual claims. This is also the case with the primary body site and primary diagnosis information provided. Please take this into account when considering the data provided.
Response to Mark Hanna (19 April 2016) | ACC
The response also had a pleasant surprise, in that ACC had supplied some extra data I hadn’t asked for, in case it would assist me. This contrasts somewhat with some of the frustration I’ve felt in the past with delayed and denied requests, but I’m very happy with how they responded this time.
The extra information they provided is a breakdown of acupuncture spending by primary injury site. Unlike the injury type breakdown I’d been provided in the past, this could be very helpful in determining how much of ACC’s funding of acupuncture treatments is aligned with the findings of their own reviews of the evidence.
Since their reviews only found positive conclusions for two injury sites – neck and shoulder – it seems like it should be a reasonable first estimate to look at the proportion of ACC’s spending on just these injury sites, allowing for the charitable assumption that these were all treating chronic neck or shoulder pain, or frozen shoulder. Allowing for some amount of error because of the caveats ACC mentioned, ideally this would come pretty close to 100%.
|Primary injury site of claim||Claims Paid Count||Cost ($) Ex GST|
|Back Except Head Vertebrae||<4*||$2,043|
|Head (except Face)||426||$142,220|
|Hip, Upper Leg, Thigh||2,511||$894,522|
|Neck, Back Of Head, Vertebrae||8,262||$2,982,805|
|Other Internal Organ||9||3,127|
|Shoulder (incl Clavicle/blade)||9,454||$3,640,599|
|Upper And Lower Arm||2,293||$863,645|
*Small numbers were reported as <4 or <$500 in order to protect privacy
Top acupuncture claims
Although shoulder and neck are in the top three primary injury sites for acupuncture, together they made up just 25% of the cost of acupuncture claims to ACC. This leaves just under $20 million for claims involving other primary injury sites.
I hadn’t expected to see such a strong trend toward a single injury site that was neither shoulder nor neck, but there were more claims with the lower back as the primary injury site than there were for neck and shoulder combined.
Looking at the data for individual read codes, I found that 33% of all ACC’s spending on claims involving acupuncture had a primary read code of “S572.”, which indicates a lumbar sprain.
Because of the caveats mentioned earlier, it’s likely that not all of the $8,652,237 spent on these 20,409 claims was for acupuncture used to treat a lumbar sprain. But it certainly indicates that ACC spends a large amount of money on ACC for lumbar sprain – large enough to be measured in the millions.
ACC has evaluated the evidence for acupuncture used to treat lower back pain. Its 2004 New Zealand Acute Low Back Pain Guide* categorised acupuncture as having “evidence of no improvement in clinical outcomes”.
Their more recent (2011) review on acupuncture for musculoskeletal pain concluded that:
- The evidence for the use of acupuncture in (sub)acute lower back pain is inconclusive
- There is limited evidence to support the use of acupuncture for pain relief in chronic lower back pain in the short term (up to 3 months)
- The evidence is inconclusive for the use of acupuncture for long term (beyond 3 months) pain relief in chronic lower back pain
- There is no evidence to recommend the use of acupuncture for lumbar disc herniation related radiculopathy (LDHR)
This is hardly the sort of ringing endorsement that I’d expect to back up the spending of millions of dollars of public money each year on a treatment for lower back pain.
Guidelines in the UK
Until recently, the National Health Service (NHS) in the UK would pay for acupuncture to treat lower back pain. But the Guardian reported in March that acupuncture for lower back pain is no longer recommended for NHS patients. The latest draft guidelines for lower back pain, which will replace the previous guidelines from 2009, involved a thorough review of the evidence and recommended not offering acupuncture at all for treating lower back pain. Its summary for acupuncture notes that:
Comparison with sham acupuncture showed no consistent clinically important effect, leading to the conclusion that the effects of acupuncture were probably the result of non-specific contextual effects.
“Non-specific contextual effects” is just a more descriptive way of saying “placebo effect”.
Continued spending despite a lack of evidence
In the last year, the New Zealand government has been under intense criticism for spending $26 million over three years on a referendum for changing the flag. More recently, the importance of funding evidence-based treatment has been emphasised in the media when reporting on Pharmac’s decision not to fund the effective, yet extraordinarily expensive, melanoma drug pembrolizumab (branded as Keytruda), estimated to cost $30 million annually.
In this context, it seems increasingly bizarre that ACC continues to spend $25 million or more each year on a treatment that they themselves have found is not supported by evidence for at least three quarters of the injuries it’s used to treat.
*ACC’s website notes that “due to the age of this guideline, some sections may have been superseded by more recent evidence”, although as far as I can tell they haven’t published an updated guideline.
Featured image: CC Pixabay