Trends in the Timeliness of Spinal Muscular Atrophy Detection in US Infants, 2016–2023

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The authors analyzed an exceptionally large dataset, leveraging nationally representative CMS Medicaid claims data covering more than half of all U.S. births (over 25 million enrollees), with supplemental analyses using MarketScan Commercial data. The methodology is clearly and rigorously described in the Materials and Methods section, including thoughtful and appropriate discussion of the inherent limitations of claims-based analyses.

The results and Discussion are well balanced and appropriately interpreted, demonstrating substantial improvements in the timeliness of claims-based SMA detection following the implementation of newborn screening, while also acknowledging other contributing factors. Overall, this is an excellent and carefully conducted study, and I commend the authors for their comprehensive and transparent approach.

I have only a few optional comments that may further improve clarity for readers; these are offered purely as suggestions and do not affect the validity or interpretation of the study’s findings:

- In the Abstract, the terms “Medicaid infants” and “Medicaid-covered newborns” appear to be used interchangeably. If these refer to the same population, the authors may consider standardizing the terminology for consistency, although the current wording is acceptable.

- The Abstract appropriately reports proportions of timely SMA detection, and the denominator is clearly defined in the main text. If space permits, readers might benefit from a brief clarification in the Abstract that these proportions are based on cases identified within a fixed follow-up window (e.g., by 24 months of age). That said, I recognize that the authors may have intentionally avoided additional detail to maintain brevity, which is also reasonable.

- In the 2022 Medicaid cohort, the proportion of cases detected by 1 month of age (76.4%) differs slightly from the ratio calculated using rounded prevalence estimates reported in Table 1 (6.6 per 100,000 by 1 month and 8.5 per 100,000 by 12 months). This minor difference is likely attributable to the use of unrounded case counts for percentage calculations, and no revision appears necessary; a brief confirmation would suffice if the authors deem it helpful.

- The citation of German data demonstrating earlier treatment initiation following SMA newborn screening is highly appropriate and effectively underscores the clinical importance of timely detection. For additional context, readers might benefit from noting that the referenced nonrandomized controlled trial (Schwartz et al., JAMA Pediatrics 2024) evaluated children with SMA born between 2018 and 2021. This temporal context may help readers interpret the reported mean age at treatment initiation (1.3 months) as reflecting an early implementation phase of screening programs. Inclusion of this detail is entirely optional, as the current discussion is already well supported.

In summary, this manuscript represents a high-quality and impactful contribution to the literature on SMA newborn screening, and I fully support its publication in its current form.

Author Response

Comments 1:  In the Abstract, the terms “Medicaid infants” and “Medicaid-covered newborns” appear to be used interchangeably. If these refer to the same population, the authors may consider standardizing the terminology for consistency, although the current wording is acceptable.

Response 1: Neither term is exact, since the Medicaid database includes enrollees in both Medicaid and CHIP plans. Also, having Medicaid or CHIP coverage at a point in time is not an attribute of a child, which can change over time. We have substituted the term “publicly insured” in the Abstract when referring to children included in the Medicaid database.   

Comments 2: The Abstract appropriately reports proportions of timely SMA detection, and the denominator is clearly defined in the main text. If space permits, readers might benefit from a brief clarification in the Abstract that these proportions are based on cases identified within a fixed follow-up window (e.g., by 24 months of age). That said, I recognize that the authors may have intentionally avoided additional detail to maintain brevity, which is also reasonable.

Response 2: We agree that clarity is important, but we were constrained by the word limit for the Abstract.

Comments 3: In the 2022 Medicaid cohort, the proportion of cases detected by 1 month of age (76.4%) differs slightly from the ratio calculated using rounded prevalence estimates reported in Table 1 (6.6 per 100,000 by 1 month and 8.5 per 100,000 by 12 months). This minor difference is likely attributable to the use of unrounded case counts for percentage calculations, and no revision appears necessary; a brief confirmation would suffice if the authors deem it helpful.

Response 3: We included in the text the ratios of numbers in parentheses to show how the percentages were calculated, e.g., “76.4% (133/174) in 2022.”

Comments 4: The citation of German data demonstrating earlier treatment initiation following SMA newborn screening is highly appropriate and effectively underscores the clinical importance of timely detection. For additional context, readers might benefit from noting that the referenced nonrandomized controlled trial (Schwartz et al., JAMA Pediatrics 2024) evaluated children with SMA born between 2018 and 2021. This temporal context may help readers interpret the reported mean age at treatment initiation (1.3 months) as reflecting an early implementation phase of screening programs. Inclusion of this detail is entirely optional, as the current discussion is already well supported.

Response 4: Thank you for that observation and suggestion. We have added the time period and appended the following sentence, “The magnitude of reduction in the age at initiation of treatment during the early phase of implementation of SMA screening in Germany may be conservative if timely detection and intervention continued to increase over time.”

Comments 5: In summary, this manuscript represents a high-quality and impactful contribution to the literature on SMA newborn screening, and I fully support its publication in its current form.

Response 5: We thank the reviewer for their support and helpful comments. 

Reviewer 2 Report

Comments and Suggestions for Authors

Dear authors,

This report described preliminary data regarding the trends in timely SMA detection from the newborn screening through a period of time (2016-2023) in the US. This paper included 2 independent databases to present results.

Please provide justification or rationale to the following points for clarification:

1. Please provide a brief rationale for the purpose of data stratification into 1, 3, 12, and 24 months.
2. What is the reason to include results from 2 databases? Does it help gain more insights on “the earlier, the better” for the NBS of SMA detection in addition to the similar trend in data?
3. On Line 172-174, you stated that the percentage of cases present at 12 months detected by 1 month of age, a secondary measure of timing, further increased from 63.9% (101/133) in 2021 to 76.4% (133/174) in 2022. Please confirm if the value of 63.9% was deduced from “101/158” instead.

Author Response

Comments 1: Please provide a brief rationale for the purpose of data stratification into 1, 3, 12, and 24 months.

Response 1: The selection of 1 month as an endpoint for timely detection reflects recommended practice to minimize the delay of treatment initiation. Our selection of 3, 12, 24, and 48 months as additional temporal endpoints was somewhat arbitrary. As we do not reference the 3-month cutoff numbers in the text, the 3-month observations could be dropped from each of the four tables if requested. 

Comments 2: What is the reason to include results from 2 databases? Does it help gain more insights on “the earlier, the better” for the NBS of SMA detection in addition to the similar trend in data?

Response 2: Reporting results from two independent databases can increase confidence in the findings of RWE analyses. Our primary focus was on the overall similarity in trends. We have now added a note that the apparently more rapid increase in early detection of SMA in the private insurance database is suggestive of potential disparities in access to timely diagnosis, but the numbers are too small for a statistical hypothesis test. Finally, the MarketScan Commercial database allowed us to the extend the complete analysis to 2022 compared to 2021 for the Medicaid data. 

Comments 3: On Line 172-174, you stated that the percentage of cases present at 12 months detected by 1 month of age, a secondary measure of timing, further increased from 63.9% (101/133) in 2021 to 76.4% (133/174) in 2022. Please confirm if the value of 63.9% was deduced from “101/158” instead.

Response 3: We thank the reviewer for catching that mistake; the correct denominator from Table 1 at 12 months in 2021 is 158.