Hi,
The article referred to above does not make any recommendation regarding MUAC or weight based discharge criteria but merely looks at the relationship between MUAC and weight changes during recovery.
A soon to be published article in the same journal reviews the safety and practicability of MUAC discharge > 12.5cm.
In short the more malnourished the child is on admission then the longer it will take to recover. This time can be shortened by investing in good case finding and early admission to treatment. The median length of stay for all cases we found in the Malawian context was approximately 49 days which is not so different from other criteria and other contexts.
As a rough approximation, a child with a MUAC > 11cm will recover in a median time of about 6 weeks while if they have a MUAC <10cm it will be approximately double that.
In terms of 'success' it depends what you mean. The MUAC > 12.5cm criterion has been demonstrated to be more appropriate than others (e.g. proportional weight gain) since the most severely malnourished children get the most treatment, see:
Dale NM, Myatt M, Prudhon C, Briend A. Using mid-upper arm circumference to end treatment of severe acute malnutrition leads to higher weight gains in the most malnourished children. PLoS One. 2013;8:e55404.
There has also been some concern that the criterion may be unsuitable for short children (i.e. < 65cm on admission) but we did not find that to be the case.
Relapse rates were comparable to other discharge criteria used in CMAM programmes and all of the children with negative outcomes at 3 months following discharge as cured (i.e. they died or relapsed) also had weight for height well above -1 z-score on discharge.
If you invest in good case finding and early admission you are more likely to achieve shorter lengths of stay and higher cure rates. If you have good follow up of absentees this will improve recovery rate since attendance rate was found to be linked to chances of non-recovery. The 'success' of a criterion is not only to be judged by the threshold but also by other operational factors.
I hope this helps,
Paul