Improving Satisfaction with Assistive Technology Through Better Service Delivery: Evidence from the WHO rATA Survey in Italy

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Points to be improved
- The abstract is very good and cohesive, but it could be enhanced by opening with a quote about the study relevance: what gap it intends to fill in? Before the “This study investigates…”.
- A very minor detail - in the abstract and in the rest of the paper, make sure the right elements of statistical notation are italicised - like p or β.  This needs to be done in the whole paper.
- Introduction is very complete and covers most relevant aspects. Personally, I thing in paragraph 95-104 (about person-centred approaches), I would add the secondary effects of such approaches. Besides best suited technologies and increased satisfaction, are their other impacts in well-being, empowerment, inclusion, self-determination? If yes, summarised evidence here.
- As the article is specifically about the Italian context, it would also be good to have particular data about AT in Italy in the introduction, so readers can culturally situate the study.
- In “2.2. Measures”, it would be good to have psychometric properties (internal consistency, reliability, external validity) of rATA, if available.
- This bit of results make me question some aspects: “One-way ANOVAs revealed significant group differences in product satisfaction by sex (F(1,986) = 5.85, p = .016) and proxy respondent status (F(1,986) = 4.93, p = .027). Satisfaction with pre-delivery services also differed significantly by proxy use (F(1,986) = 5.42, p = .020), while a similar trend was observed for post-delivery services (F(1,986) = 4.76, p = .029). (Lines 232-236)”
- If you are using sex as your IV, and assuming it as only two groups (otherwise it would be gender), why are you doing an ANOVA and not a t test or its non parametric equivalent? ANOVA should only be used for IV with more than 2 groups.
- For the other 2 ANOVAs - satisfaction with pre-delivery and post-delivery - this is not fully reported. If there are differences, you need to report, through appropriate post hoc analysis what are those differences and between which groups, illustrated with descriptives.
- Above all this, is data for this variables normally distributed? It needs to be for you to do ANOVA, otherwise, you need to perform kruskal-wallis.
- Both in discussion and conclusions, you should narrow a bit your claims. Your findings are relevant, but some sentences should clearly stated this is only about Italian context.

Author Response

Dear Reviewer 1,

Thank you for your thoughtful and constructive comments on our manuscript, “Improving Satisfaction with Assistive Technology through Better Service Delivery: Evidence from the WHO rATA Survey in Italy.” In response, we have carried out a thorough revision of the manuscript. All changes are highlighted in yellow in the revised manuscript. 

Your comments are presented together with those from the other reviewers so that all revisions, cross-references, and their rationale can be appreciated in a consolidated way.

We hope that the revised version addresses your concerns and meets your expectations. We sincerely appreciate your engagement with our work and look forward to any further feedback you may have.

Kind regards,
The Authors

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

Below are Key Suggestions to Improve the Manuscript: 

• Fix Table 2. The pre-delivery and post-delivery question text are identical, which looks like a copy-paste error and completely confuses your measures.

• You can't measure complex ideas like "satisfaction" with a single question. This is a major limitation and you need to state it more forcefully.

• Averaging scores across different products is a problem. You're losing all the detail—a user might love their wheelchair but hate their communication aid.

• Your data is cross-sectional. You must tone down the causal language—you found associations, not proof that one thing causes another.

• Why assume satisfaction leads to usefulness? The reverse is just as likely. You need to justify your mediation path or admit it's just one possibility.

• The 80% "mixed" AT category is a black box. It's too varied to be a useful group, and it's hiding all the important differences between product types.

• Your policy claims are too strong. You can't call pre-delivery "decisive" based on this data. Soften your claims to match the exploratory nature of the findings.

• Did you account for the rATA's complex survey design (like sampling weights) in your analysis? If not, your stats might be off, and you need to state this as a limitation.

Author Response

Dear Reviewer 2,

Thank you for your thoughtful and constructive comments on our manuscript, “Improving Satisfaction with Assistive Technology through Better Service Delivery: Evidence from the WHO rATA Survey in Italy.” In response, we have carried out a thorough revision of the manuscript. All changes are highlighted in yellow in the revised manuscript. 

Your comments are presented together with those from the other reviewers so that all revisions, cross-references, and their rationale can be appreciated in a consolidated way.

We hope that the revised version addresses your concerns and meets your expectations. We sincerely appreciate your engagement with our work and look forward to any further feedback you may have.

Kind regards,
The Authors

Author Response File: Author Response.pdf

Reviewer 3 Report

Comments and Suggestions for Authors

Overall
(1) The operational definition of the primary variable (predelivery) is vague, leading to unstable conclusions.

Points of criticism
`Abstract` (: "Results showed that satisfaction with predelivery services (β = 0.571, p < .001) was the strongest predictor..."
`Limitations` (: "...in the rATA, assessment, and training are jointly evaluated under a single item... the aggregated nature of the rATA data precludes separate analysis..."
`Table 2` In the user-provided text, "Service satisfaction (postdelivery)" appears twice, and the "predelivery" item is missing. This is an obvious transcription error, complicating the peer review process.

Points of criticism
The most important conclusion of this study is that "predelivery services (before delivery) have a significantly stronger impact on satisfaction than postdelivery services (after delivery) (β = 0.571 vs. 0.280)."

However, the authors themselves As acknowledged in the `Limitations` section (lines 392396), the rATA questionnaire used breaks down "Predelivery" into two completely different elements: "assessment" and "training." These elements are conceptually and practically distinct.

High satisfaction with "Assessment" means satisfaction that "an expert selected the perfect product for me."

High satisfaction with "Training" means satisfaction that "I was able to fully understand how to use the product."

The strong influence of β=0.571 could be due to a strong influence from one of these two factors (e.g., insufficient training but perfect product selection), or it could be due to a combination of both. From a policy perspective, this could lead to completely different conclusions: "Should we allocate budget to train assessment experts?" or "Should we increase training time at the time of delivery?" - but the results of this study make no conclusions.

The authors explain this by stating `Limitations` While briefly mentioned as one of the limitations, this is not merely a limitation; it fundamentally undermines the validity of the study's core assertion ("predelivery services ... plays a decisive role").

In addition to mentioning this in the Limitations section, the policy dilemma implied by this confusion should be actively discussed in the Discussion section (Section 4.1 or 4.3). It should be clearly stated that "The rATA data in this study cannot determine whether 'appropriate product selection' or 'appropriate training' contributes to satisfaction. This is a critical issue that should be separated and verified in future research."

(2) Risk of "ecological fallacy" due to heterogeneity and averaging of the analyzed data

Points to be noted
`Table 1` (line 170): "Mixed 803 (80.9%)"
`Methods`: "For this analysis, satisfaction and utility scores were averaged across all rated products for each respondent, resulting in one composite score per dimension per person."
`Limitations`: "Notably, around 80% of the respondents reported using a combination of assistive products... this heterogeneity... may obscure more subtle variations..."

Points to Note
Of the sample (N=992) in this study, a whopping 80.9% (803 people) used products across multiple categories ("Mixed") rather than a single assistive technology (AT).

The analysis method then averaged the satisfaction scores for these different products (up to three) and treated them as a single score per person ("one composite score") (lines 183184).

This can lead to a serious flaw in statistical analysis (ecological fallacy). For example, if a user:

Product A (e.g., wheelchair): Predelivery satisfaction 5 points, product satisfaction 5 points

Product B (e.g., communication device): Predelivery satisfaction 1 point, product satisfaction 0 points 1 point
If a user's experience is 1 point, their data will be entered into the analysis as "Average Predelivery Satisfaction: 3 points" and "Average Product Satisfaction: 3 points."

This "average" does not reflect any real-world experience. Because 80.9% of the data is this "artificial average," the regression analysis results (β = 0.571) cannot accurately reflect the relationship between individual products and their services. This completely ignores the possibility that different products (e.g., mobility vs. cognitive assistance) require completely different service delivery processes (e.g., in-hospital rehabilitation vs. home-based settings).

The authors also mention this point in the `Limitations` section, but this is not a "future topic" but a fundamental issue that affects the validity of this analysis itself.

(3) Errors in the temporal and logical order of the mediation analysis model (RQ2)

Points to Note
`Abstract` "...satisfaction with predelivery services partially mediated the relationship between product satisfaction and perceived usefulness..."
`Statistical Analysis` "The model specified product satisfaction as the independent variable (X), perceived utility as the dependent variable (Y), and one service satisfaction dimension as the mediator (M).

Figure 1: Diagram path (X > M > Y)

Implications
RQ2 (mediation analysis) is another pillar of this study, but its model contains a serious logical error.
This is an impossible chronological sequence.
(Time 1) The user experiences "predelivery."
(Time 2) As a result, "product satisfaction" is formed.
(Time 3) Then, by continuing to use the product with satisfaction, the user realizes "perceived utility."

In other words, the logical causal chain should be `Predelivery Satisfaction (X) > Product Satisfaction (M) > Perceived Utility (Y)`.

However, in this study's model (`Product Satisfaction (X) > Predelivery Satisfaction (M)...`), This is logically flawed, as it assumes that satisfaction with the product (Time 2) influences satisfaction with the pre-delivery service received in the past (Time 1).

Although RQ1 correctly concludes that "pre-delivery predicts product satisfaction," the reason for reversing this order in RQ2 is completely unclear, making the mediation analysis result (accounting for 29% of the total effect) completely meaningless.

(4) Lack of important contextual information (reason for excluding "spectacles only" and insufficient digging into the "comparison with China")

Points to be noted
`Methods` "...the exclusion of users of spectacles only is justified by the different service delivery procedures to obtain this product in the Italian context..."
`Discussion` (comparison with China by Chen et al. [35])

Points to be noted
This study critically lacks an explanation of the "context," which is essential for interpreting the results.

a) Regarding the exclusion of "glasses-only" users: The authors excluded "glasses-only" users, the most common and widely used type of AT, from their analysis simply because "the eyeglass service provision procedures are different in the Italian context."
However, they did not explain how "different" they were.

If the eyeglass provision process were (for example) more simplified or based on market principles (optician's shop) and resulted in higher or lower satisfaction rates than the other AT provision processes analyzed in this study (presumably based on public benefits), then that would be the most important comparison point for discussing Italy's overall AT policy.

This exclusion not only significantly reduces the generalizability of the study, but also appears to intentionally forego the opportunity for potentially most insightful comparative analysis.

b) Shallow Comparison with China and India: One of the most interesting findings of this study is that while the impact of "post-delivery" service satisfaction was nearly zero in studies of China [35] and India [38] (LMICs), it had a significant impact (β = 0.280) in Italy (HIC).

The authors conclude that this difference is due to "the Italian system being 'mature' and having 'consistent perceived quality.'"

This is mere speculation and a very shallow argument.

Why doesn't post-delivery work in LMICs?
(Hypothesis 1) Since repair and maintenance services don't exist in the first place, satisfaction or dissatisfaction isn't an issue.
(Hypothesis 2) The product is simple and doesn't need to be repaired (e.g., a cane or prosthetic limb).
(Hypothesis 3) Users' expectations are focused on "getting the product" and they don't expect after-sales service.

Why does post-delivery work in Italy?
(Hypothesis A) Because the system is mature (the authors' argument).
(Hypothesis B) Because products are advanced and complex (e.g., electric wheelchairs, advanced hearing aids) and require frequent maintenance, their quality affects satisfaction.

If Hypothesis B is correct, then it's simply a difference in the "complexity of the products involved," not the "maturity of the system." It is highly likely that the data in this study (80.9% "Mixed" in Table 1) include a more complex product group than data from LMICs. Without discussing this point (differences in product types), we cannot conclude that "the Italian system is mature."

Author Response

Dear Reviewer 3,

Thank you for your thoughtful and constructive comments on our manuscript, “Improving Satisfaction with Assistive Technology through Better Service Delivery: Evidence from the WHO rATA Survey in Italy.” In response, we have carried out a thorough revision of the manuscript. All changes are highlighted in yellow in the revised manuscript. 

Your comments are presented together with those from the other reviewers so that all revisions, cross-references, and their rationale can be appreciated in a consolidated way.

We hope that the revised version addresses your concerns and meets your expectations. We sincerely appreciate your engagement with our work and look forward to any further feedback you may have.

Kind regards,
The Authors

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

Comments and Suggestions for Authors

The authors have revised and improved the manuscript.

Author Response

We thank the Reviewer for his/her insightful comments which we believe have greatly contributed to improve the quality of our manuscript. 

Kind regards,

The Authors

Reviewer 3 Report

Comments and Suggestions for Authors

Overall
The approach utilising WHO's new global standard tool, “rATA (rapid Assistive Technology Assessment)”, to statistically elucidate the impact of the “service provision process (Pre vs Post)” on “user satisfaction” – rather than merely surveying adoption rates – represents a valuable perspective.

Evaluation: High validity, but caution is required in interpreting causality

The analytical framework is robust, employing standard and appropriate statistical methods: hierarchical multiple regression analysis and mediation analysis.
The results (predominant influence of pre-delivery factors) are logical, and comparisons with the literature (Chen et al., 36) are appropriately conducted.
As a cross-sectional study, it cannot strictly prove the “direction of causality” implied by terms like “predictor” or “mediate”. While the authors acknowledge this in the Limitations section, some expressions in the Discussion are somewhat overstated.
The sample size of $N=992$ is sufficient to ensure statistical power for this type of investigation.
   
     Common Method Variance (CMV) exists as all variables are self-reported by the same respondents and measured simultaneously. This may inflate observed correlations (e.g., the “halo effect” where respondents in a generally positive mood rate both products and services highly).
The survey sample is limited to ‘current AT users’. Individuals who abandoned the service due to poor quality may not be included in this analysis.

 

 2. Specific Proposed Corrections and Improvements

A. Introduction / Discussion (Toning Down Causality)
Although mediation analysis is performed, this is a theoretical model and does not measure the actual temporal sequence (service → satisfaction → usefulness). Given all data is self-reported, respondents' subjective tendencies (such as a propensity for positive responses) may be strengthening the correlation between service satisfaction and product satisfaction.
As this study only targets “current users”, it may not reflect the dissatisfaction of those who discontinued use due to poor service. This represents a significant bias affecting the results (overall high satisfaction levels).

References (Correction of bibliographic information)
The reference list contains clear formatting errors. These must be corrected prior to submission.

DOI: The ‘DOI:’ prefix and the DOI itself (e.g., `DOI: 10.1016/...`) are redundant.
  Verification of publication year: Numerous references include ‘Technologies 2025’ or ‘2025’. Please confirm whether these are correct as special issue titles or planned publication years (as they represent future dates).

Author Response

Dear Reviewer,

Thank you for your comments. In response, we have carried out a further revision of the manuscript. All changes are highlighted in GREEN in the revised manuscript. 

We hope that the revised version addresses your concerns and meets your expectations. We sincerely appreciate your engagement with our work and look forward to any further feedback you may have.

Kind regards,

The Authors

Author Response File: Author Response.pdf

Round 3

Reviewer 3 Report

Comments and Suggestions for Authors

Overall 

This work appears to be an original approach to utilizing existing rATA (WHO Rapid Assistive Technology Assessment) data.

However, while revisions have been made in response to reviewers' comments, there are still areas of inadequacy in the areas of "logical soundness" and "interpretation of bias." In particular, there remains a contradiction in the analytical model's assumption of causality while denying causality.

1. Evaluation of Soundness and Rigor (Logic & Rigor)

In response to reviewers' comments, the authors revised their statement to "noncausal." However, the mediation analysis method employed is essentially a model for verifying causal mechanisms (X → M → Y).
"Product Satisfaction (X)" → "Service Satisfaction (M)" → "Perceived Usefulness (Y)"

Logical Question: Wouldn't it be more natural for user experiences to flow from "usefulness" to "satisfaction," or from "received good service" to "satisfied with the product"?
Please provide additional rationale for why utility was included as the final outcome variable. For example, you might explain that utility is treated as a proxy for continued use in the ISO definition and existing AT models (e.g., the MPT model).

2. Data Support and Bias

Reviewers 3 and 4 point out that "only current users" were included; this is not simply a limitation, but may fundamentally distort (inflate) the results.

The data completely omits people who stopped using the service due to poor service, which likely results in high "service satisfaction" and "product satisfaction." The author added this to the "Limitations" section, but the interpretation in the "Discussion" section is still optimistic.

> Original: "satisfaction scores... were numerically very similar (i.e., M = 4.01 and M = 4.05)"
> Criticism: The high scores may simply be due to the exclusion of dissatisfied users who left the service.
> Suggested revision: Including a brief note in the interpretation of the results (Section 4.1) that "high satisfaction scores may be influenced by survivorship bias due to the exclusion of defectors, and therefore actual service quality may be lower than this figure" would be a sign of sincerity.

3. Other specific revisions

The conclusion that predelivery is important is reasonable, but the beta value (0.280) for postdelivery is not low at all. A more balanced conclusion would be achieved by not overemphasizing "predelivery is important" and instead emphasizing the complementary nuance of "predelivery is the foundation and postdelivery complements it." Please also double-check that the "Following... post delivery services" part of the Abstract does not downplay postdelivery services.

Author Response

 

Dear Reviewer,

Thank you for your comments. In response, we have carried out a further revision of the manuscript. All changes are highlighted in BLUE in the revised manuscript.

We hope that the revised version addresses your concerns and meets your expectations. We sincerely appreciate your engagement with our work and look forward to any further feedback you may have.

Kind regards,

The Authors

Author Response File: Author Response.pdf