Cost-effectiveness of monthly follow-up for the treatment of uncomplicated severe acute malnutrition: An economic evaluation of a randomized controlled trial

Severe acute malnutrition (SAM) is a major source of mortality for children in low resource settings. Alternative treatment models that improve acceptability and reduce caregiver burden are needed to improve treatment access. We assessed costs and cost-effectiveness of monthly vs. weekly follow-up (standard-of-care) for treating uncomplicated SAM in children 6–59 months of age. To do so, we conducted a cost-effectiveness analysis of a cluster-randomized trial of treatment for newly-diagnosed uncomplicated SAM in northwestern Nigeria (clinicaltrials.gov ID NCT03140904). We collected empirical costing data from enrollment up to 3 months post-discharge. We quantified health outcomes as the fraction of children recovered at discharge (primary cost-effectiveness outcome), the fraction recovered 3 months post-discharge, and total DALYs due to acute malnutrition. We estimated cost-effectiveness from both provider and societal perspectives. Costs are reported in 2019 US dollars. Provider costs per child were $67.07 (95% confidence interval: $64.79, $69.29) under standard-of-care, and $78.74 ($77.06, $80.66) under monthly follow-up. Patient costs per child were $21.04 ($18.18, $23.51) under standard-of-care, and $14.16 ($12.79, $15.25) under monthly follow-up. Monthly follow-up performed worse than standard-of-care for each health outcome assessed and was dominated (produced worse health outcomes at higher cost) by the standard-of-care in cost-effectiveness analyses. This result was robust to statistical uncertainty and to alternative costing assumptions. These findings provide evidence against monthly follow-up for treatment of uncomplicated SAM in situations where weekly follow-up of patients is feasible. While monthly follow-up may reduce burdens on caregivers and providers, other approaches are needed to do so while maintaining the effectiveness of care.


Itemized response to reviewer comments "Cost-effectiveness of monthly follow-up for the treatment of uncomplicated severe acute malnutrition: an economic evaluation alongside a randomized controlled trial" (PGPH-D-22-00813)
Thank you for the opportunity to revise our manuscript. Below we provide itemized responses to the editor's and reviewers' comments. We have renumbered these for the purpose of this response.
Editor's comments 1. In the online submission form, you indicated that "The deidentified dataset supporting this research can be made available following a submitted request as per Epicentre and General Data Protection Regulation (EU) 2016/679 data sharing policy." All PLOS journals now require all data underlying the findings described in their manuscript to be freely available to other researchers, either 1. In a public repository, 2. Within the manuscript itself, or 3. Uploaded as supplementary information. This policy applies to all data except where public deposition would breach compliance with the protocol approved by your research ethics board. If your data cannot be made publicly available for ethical or legal reasons (e.g., public availability would compromise patient privacy), please explain your reasons by return email and your exemption request will be escalated to the editor for approval. Your exemption request will be handled independently and will not hold up the peer review process, but will need to be resolved should your manuscript be accepted for publication. One of the Editorial team will then be in touch if there are any issues. 2. Thank you for submitting your paper to PGPH for publication. Two independent reviewers have assessed your manuscript and found it to have merit. However, they raised substantial methodological issues that will need your attention.
The issues include: -Clarity on the quantity of therapeutic food per participant, and the extended period of visits. -They also felt there is a need for you to look at the definition of nutritional recovery.
-The stage at which participants were discharged. -Justification of sample size used in the study.
-The assumptions made in the sample size estimation.
-There is also a need for clarity on the presentation of the results.
Authors' response: Thank you for these comments. We have responded to each of these issues raised by the reviewers, as detailed below.
4. There is one major methodological issue which I would like to better understand before I can determine the validity of this work, however. On line 158, the authors define their outcome of nutritional recovery as being free from complications and with MUAC of at least 125 mm for 2 consecutive visits. But it seems to me that since the treatment variable itself is the time between successive visits, this criteria might inadvertently lead to the treatment arm appearing less efficacious than the control arm, where there are successive visits closer together towards the end of the study. It appears possible, at least in theory, that this requirement could produce a bias the classification of recovery for the monthly treatment arm. As a thought experiment, consider if instead, caregivers had taken at-home measurements and reported MUAC at least 125 mm for two consecutive weeks to define recovery. If there are only three follow-up visits for the monthly follow-up arm, children must have a time-to-recovery of less than 8 weeks to be classified as recovered in this arm but could take up to 11 weeks to recover in the standard-of-care arm. Has this risk been addressed in the RCT in a way that could be explained in this paper?
Authors' response: We thank the reviewer for this important observation and agree with the reviewer that the greater spacing between visits in the intervention arm (e.g. monthly follow up) could indeed result in a longer length of stay in the program before nutritional recovery is declared, compared to the standard care with weekly follow-up. This effect was indeed addressed in the primary analysis of the RCT that examined time to programmatic recovery as a secondary outcome. The primary analysis, as suggested by the reviewer, showed that the average time to programmatic recovery was in fact longer in the monthly follow-up arm, compared to weekly follow-up. As the reviewer notes, this longer time to recovery in the monthly arm may be due to the longer spacing between measurements. We do not consider this as a bias but rather a true representation of the programmatic outcome of interest. We confirm that the differential length of stay in the program noted here is included as a secondary outcome in the primary analysis and is accounted for in the costs of therapeutic foods in this cost effectiveness analysis. For transparency, and to respond to the reviewer's comment, we have also included the table of primary and secondary outcomes from the parent trial (reproduced from Hitchings et al 2022) as a supplementary table (Table S1).
5. Line 362: I find it surprising that monthly approach required so many more sachets of therapeutic food. Your explanation about the extended period between visits is not totally satisfying to me. Is this because there is also a longer time-to-recovery in the monthly treatment arm, given the lower recovery rates, or simply because of wastage after recovery since the family has been supplied with a months worth of food? Is it perhaps not only that more food is distributed, but also that more food is required and/or consumed in the treatment arm of the trial? Is there more wastage or loss, e.g. as the family shares with relatives and neighbors when they have a month's worth of sachets in their house. Any expansion on theorizing why this surprising finding has happened would be a welcome addition to the discussion. Even if they are common to the treatment and control arm, it would be a missed opportunity if readers of your paper could see the scale of these essential parts of a successful program in context. It would be great to also include them in Table 1, for busy readers who might miss them in the text.
Authors' response: We agree with the reviewer that these estimates would be of value to some readers. However, as the study was designed to estimate the outcomes required for an incremental cost-effectiveness analysis, we did not collect data on costs that did not differ between study arms, and therefore cannot report this result.
8. Line 88: typo, should be "enrollment" Authors' response: Thank you for spotting this, we have corrected this typo.
9. Line 91: Why does monthly follow-up include a visit at 10 weeks? Perhaps "monthly" is a confusing term for this less frequent visit schedule.  [lines 193-199] 14. Line 565: typo, should be "not conducted" instead of "no conducted".
Authors' response: Thank you for identifying this typo, which we have now corrected.
15. Figure 1A, 1B, and 2A: Consider anchoring axes at zero to contextualize the difference between study arms relative to the absolute magnitude of each study arm ---this might tell a different story about the value of monthly follow-up in a setting where weekly follow-up is simply infeasible.
Authors' response: Thank you for this suggestion. We created this figure for comparison (shown below), but prefer the original version as it provides a clearer visualization of the incremental difference between the study arms. Moreover, we feel that the inclusion of (0,0) in the alternative figure implicitly invites comparison to this intercept, yet this is not an appropriate comparison as there is no realistic scenario in which one should expect zero programmatic recovery and zero costs.
16. Figure 2B: I recommend excluding this figure, since the treatment is never cost-effective.

Reviewer #2
17. Overall this is an important paper and well worthy of publication in a high profile journal like PLoS Global Health. It is a well written report of a well done and robustly analysed study.
Authors' response: Thank you for these kind words.
18. I have a few minor comments which I hope will help future readers make the most of the study: it's not very clear (in title, abstract or final sentence of intro), whether this paper presents in-depth economic evaluation of a RCT that's reported elsewhere (as suggested by reference 14 and by line 106) OR whether the cost-effectiveness was the major RCT outcome and this is the main paper reporting all outcomes. Please clarify. If the former, it'd be better to say this was an "economic evaluation OF a (separately reported) RCT". It'd also be good in this case to briefly summarise the main RCT results in the methods. The word "alongside" is what leads to the ambiguity -change that. 20. Also consider citing the increasing evidence (e.g. two recent systematic reviews in BMJ Global Health) about the long term adverse effects of childhood SAM. This would also be worth including in the discussion as better treatment might have long term benefits: hence the costs of treatment might in fact have much more long lasting benefit and be even more cost-effective than appears in short-term analysis such as this. 21. Lines 90-95. Did participants attend until 12 weeks irrespective of growth/progress or could they be discharged earlier if meeting specified anthro criteria (e.g. WLZ >-2 z-score on 2 consecutive visits -as is standard in many CMAM programmes worldwide. Please give details for both arms as again this could make a big difference to final costs.