The Cost of Endometriosis and Chronic Pelvic Pain Burden in New Zealand (Aotearoa): Results from a Nationwide Survey

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The study analyzes the national economic impact of endometriosis and chronic pelvic pain (CPP) in New Zealand using a cost-of-illness model adapted from the WERF EndoCost protocol. This is the first national cost-of-illness analysis of endometriosis and CPP in New Zealand. Its originality lies in adapting the EndoCost model to the local context and introducing an attribution correction model to avoid double-counting between CPP and endometriosis. Its relevance is high, as these conditions impose a significant but often underestimated socioeconomic impact. The article provides a robust macroeconomic estimate for New Zealand, filling a gap due to the lack of a national model. It adds methodological innovation by applying the attribution correction logic and highlights productivity losses as the main cost driver. Compared to the 2019 Australian study, the results show substantially higher per capita and macroeconomic costs. While encouraging, it would have been useful to further address potential sampling bias, as recruitment relied heavily on support networks and online channels, likely overrepresenting individuals with severe symptoms. Explore alternative prevalence scenarios (e.g., using higher international estimates of endometriosis prevalence of 18%) to test sensitivity. Include health utility measures (e.g., QALYs) directly in the model rather than relying on secondary literature. Stratify costs by ethnicity, given the known disparities affecting Māori and Pacific women, even if only for exploratory purposes.

Although the manuscript addresses an important and underexplored issue, several limitations warrant further evaluation prior to publication. First, the recruitment strategy, which relies heavily on advocacy networks, social media, and clinical contacts, introduces a risk of sampling bias, likely overrepresenting individuals with severe symptoms and more intense care-seeking behaviors. This could inflate cost estimates and limit generalizability to the broader population. Second, the economic modeling does not incorporate direct measures of health utility (such as QALYs), which would strengthen the assessment of non-financial burden and facilitate the study's provision of valuable, original data. However, revisions are needed to improve the representativeness of the sampling, incorporate health utility outcomes, and expand equity-focused analyses. These improvements would increase the robustness and policy relevance of the findings.

Author Response

The study analyzes the national economic impact of endometriosis and chronic pelvic pain (CPP) in New Zealand using a cost-of-illness model adapted from the WERF EndoCost protocol. This is the first national cost-of-illness analysis of endometriosis and CPP in New Zealand. Its originality lies in adapting the EndoCost model to the local context and introducing an attribution correction model to avoid double-counting between CPP and endometriosis. Its relevance is high, as these conditions impose a significant but often underestimated socioeconomic impact. The article provides a robust macroeconomic estimate for New Zealand, filling a gap due to the lack of a national model. It adds methodological innovation by applying the attribution correction logic and highlights productivity losses as the main cost driver. Compared to the 2019 Australian study, the results show substantially higher per capita and macroeconomic costs.

While encouraging, it would have been useful to further address potential sampling bias, as recruitment relied heavily on support networks and online channels, likely overrepresenting individuals with severe symptoms.

Sampling bias was definitely a feature and significant limitation in our manuscript. This has been outlined under the limitations section in the discussion. We have added a qualifying statement as per below.

“Given the advocacy-based recruitment strategy and known bias toward individuals with greater symptom severity and diagnostic engagement, we acknowledge that our sample is more likely to reflect the higher end of the burden distribution. While this does not allow us to assert definitively that these are maximum values, they should be interpreted as representative of higher-burden cases rather than population averages.”

Explore alternative prevalence scenarios (e.g., using higher international estimates of endometriosis prevalence of 18%) to test sensitivity.

There exists data around higher endometriosis prevalence internationally. Higher prevalence estimates have not been used in the presentation of our data given the absence of local data reflecting up to date prevalence estimates. The most comparable available data for prevalence is Australian data which states a 14% prevalence rate for endometriosis. In lieu of up to date local New Zealand prevalence estimates, it is felt that this number should be used to not over represent symptom burden. This is a focus for future research from this research team. I have added a statement to this effect in the discussion to outline this.

Include health utility measures (e.g., QALYs) directly in the model rather than relying on secondary literature.

Whilst it is noted that QALYs are an important measure to include in health economic data, we did not collect data to represent this in our questionairre. This can be noted for future iterations of this model. We have included a paragraph in the discussion outlining to this effect. "While this study did not directly model health utility loss, previous research using EQ-5D has estimated utility scores of 0.72 to 0.81 among women with endometriosis, corresponding to an annual decrement of approximately 0.19 quality-adjusted-life-years (QALYs)¹⁵. This equates to a cumulative loss of up to 6.9 QALYs over the working lifespan, reinforcing the profound non-financial burden associated with chronic pelvic pain and endometriosis."

Stratify costs by ethnicity, given the known disparities affecting Māori and Pacific women, even if only for exploratory purposes.

A sub-analysis of ethnicity data will be presented separately and is currently in development by the current research team.

Although the manuscript addresses an important and underexplored issue, several limitations warrant further evaluation prior to publication. First, the recruitment strategy, which relies heavily on advocacy networks, social media, and clinical contacts, introduces a risk of sampling bias, likely overrepresenting individuals with severe symptoms and more intense care-seeking behaviors. This could inflate cost estimates and limit generalizability to the broader population.

Addressed above.

Second, the economic modeling does not incorporate direct measures of health utility (such as QALYs), which would strengthen the assessment of non-financial burden and facilitate the study's provision of valuable, original data. However, revisions are needed to improve the representativeness of the sampling, incorporate health utility outcomes, and expand equity-focused analyses. These improvements would increase the robustness and policy relevance of the findings.

Addressed above.

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript addresses an important and underexplored area of women’s health by quantifying the economic burden of endometriosis and chronic pelvic pain (CPP) in New Zealand using a modified WERF EndoCost protocol. The topic is highly relevant given the paucity of national-level cost-of-illness (COI) models and the significant health, social, and economic implications of these conditions. The paper is well written, methodologically innovative, and provides robust estimates that have direct policy implications.Nevertheless, several aspects require clarification or further refinement before publication.

1)The decision to allocate 60% of CPP costs to endometriosis is reasonable but requires stronger justification. While references are cited, the rationale for selecting this exact proportion should be better contextualized with supporting prevalence or diagnostic overlap data. Sensitivity analyses using alternative attribution rates (e.g., 50% and 70%) would strengthen the robustness of the conclusions.

2)The reported per capita costs are substantially higher than those from the 2019 Australian study and other international COI reports. The authors attribute this to the attribution correction and productivity multiplier inclusion, but a more detailed comparative analysis is warranted. Presenting a standardized table contrasting key findings across countries would enhance interpretability and credibility.

3)Recruitment via advocacy networks and online platforms likely resulted in an overrepresentation of highly symptomatic individuals. While acknowledged as a limitation, the extent to which this may inflate estimates should be discussed more explicitly. Could weighting or post-stratification have been applied to mitigate selection bias?

4)The manuscript notes disparities affecting Māori and Pacific women but does not provide stratified analyses. Given the acknowledged inequities in New Zealand healthcare, even partial subgroup analysis (where sample sizes allow) would significantly enrich the findings and strengthen the policy relevance.

5)The methodology for presenteeism and WEP relies on modeled proxies rather than direct measures. Although the conservative nature of these assumptions is emphasized, more detail on the validation of these proxies is necessary. Additionally, providing alternative estimates using only absenteeism (directly reported) would allow readers to assess the incremental effect of modeled components.

6)The conclusions appropriately highlight the need for investment in diagnostics, fertility care, and culturally responsive services. However, the recommendations would benefit from greater specificity. For example, which diagnostic innovations (e.g., imaging, biomarkers) or workplace interventions should be prioritized? Explicit links between the quantified costs and potential cost-saving strategies would enhance the translational value of the study.

7)The abstract is dense and could be streamlined for clarity. Phrasing such as “structural inefficiencies in care delivery” should be briefly illustrated with concrete examples.

Comments on the Quality of English Language

Small refinements would enhance clarity and flow.

Occasional excessively long sentences that reduce readability.

Redundancy in terminology (e.g., alternating between chronic pelvic pain and persistent pelvic pain).

Dense bureaucratic phrasing in some sections, which could be simplified for clarity without compromising academic tone

Examples for Improvement:

• Original (Abstract):
  “This burden reflects not only clinical delays and underdiagnosis, but also structural inefficiencies in care delivery, disability recognition, and labour-force participation.”
  Suggested:
  “This burden reflects clinical delays, underdiagnosis, and systemic inefficiencies in care, disability recognition, and workforce participation.”

• Original (Discussion):
  “This contrast highlights the impact of attribution correction, productivity multiplier inclusion, and workforce-calibrated wage scaling on national burden estimates.”
  Suggested:
  “This contrast highlights how attribution correction, productivity multipliers, and wage scaling substantially influence national burden estimates.”

 

Author Response

This manuscript addresses an important and underexplored area of women’s health by quantifying the economic burden of endometriosis and chronic pelvic pain (CPP) in New Zealand using a modified WERF EndoCost protocol. The topic is highly relevant given the paucity of national-level cost-of-illness (COI) models and the significant health, social, and economic implications of these conditions. The paper is well written, methodologically innovative, and provides robust estimates that have direct policy implications.Nevertheless, several aspects require clarification or further refinement before publication.

1)The decision to allocate 60% of CPP costs to endometriosis is reasonable but requires stronger justification. While references are cited, the rationale for selecting this exact proportion should be better contextualized with supporting prevalence or diagnostic overlap data. Sensitivity analyses using alternative attribution rates (e.g., 50% and 70%) would strengthen the robustness of the conclusions.

You were right to flag that the prior 60% attribution scalar for endometriosis within CPP was under-justified. We’ve now aligned with the midpoint of international estimates from Ghiasi et al. (referenced in text), placing our corrected Diagnostic Attribution Correction (DAC) at 43.4%.

This change was made throughout the Methods and Results sections of the manuscript.

2)The reported per capita costs are substantially higher than those from the 2019 Australian study and other international COI reports. The authors attribute this to the attribution correction and productivity multiplier inclusion, but a more detailed comparative analysis is warranted. Presenting a standardized table contrasting key findings across countries would enhance interpretability and credibility.

We agree the reported NZD per-capita estimates are materially higher than those in Armour et al. or the 2019 Australian study. This is addressed explicitly in the Discussion section.

• We cite three structural drivers of the difference:
1. Application of the attribution model (not used in the Australian study 2019).
2. Use of WEP-adjusted productivity metrics (not raw),
3. Lifetime horizon = 36 years, vs annual or short-horizon views.

Additionally, we now report values in both NZD and INT$, allowing easier benchmarking with global estimates.

Furthermore, the prior Australian is a top-down study. This leads to a large underreporting. The current research team are currently working on applying this model to update the underreported numbers used in the Australian study in order to change the methods to be bottom-up in the next iteration of the Australia study.

3)Recruitment via advocacy networks and online platforms likely resulted in an overrepresentation of highly symptomatic individuals. While acknowledged as a limitation, the extent to which this may inflate estimates should be discussed more explicitly. Could weighting or post-stratification have been applied to mitigate selection bias?

Whilst we agree that this is significant limitation, the research team disagrees that these methodologies could be used to mitigate this as this would add further complexities to interpretation. This is acknowledged and could be verified in future studies whereby a more representative sample could be utilised to reflect generalisability more so. We have amended the discussion to explicitly state:

We will review the manuscript to explicitly state:

“Given the advocacy-based recruitment strategy and known bias toward individuals with greater symptom severity and diagnostic engagement, we acknowledge that our sample is more likely to reflect the higher end of the burden distribution. While this does not allow us to assert definitively that these are maximum values, they should be interpreted as representative of higher-burden cases rather than population averages.”

4)The manuscript notes disparities affecting Māori and Pacific women but does not provide stratified analyses. Given the acknowledged inequities in New Zealand healthcare, even partial subgroup analysis (where sample sizes allow) would significantly enrich the findings and strengthen the policy relevance.

This subgroup analysis will be presented separately and is currently in development with the research team. Publication will be sought after this analysis and manuscript drafting is completed.

5)The methodology for presenteeism and WEP relies on modeled proxies rather than direct measures. Although the conservative nature of these assumptions is emphasized, more detail on the validation of these proxies is necessary. Additionally, providing alternative estimates using only absenteeism (directly reported) would allow readers to assess the incremental effect of modeled components.

We acknowledge that both the WEP scalar and presenteeism estimates are modelled rather than directly measured. The derivation is now described more explicitly in the Methods section:

• The WEP adjustment applies a scalar of 1.72 to employer-attributable productivity loss, following Stromberg et al. This is applied only to the combined absenteeism/presenteeism subtotal and does not inflate the base wage or total productivity pool.
• Presenteeism is calculated using bounded scalar estimates derived from international survey methodology, weighted conservatively.
• Absenteeism-only estimates are already included in the results table and supplementary materials. We have updated the manuscript to make this more explicit.

Regarding the reviewer’s note on linking quantified costs to cost-saving strategies: we agree this is an important translational goal, but we view that as a separate analytic frame. These strategy-level and intervention-modifiable elements will be addressed in an upcoming Budget Impact Model (BIM) manuscript, which builds directly on the outputs of this COI study.

6)The conclusions appropriately highlight the need for investment in diagnostics, fertility care, and culturally responsive services. However, the recommendations would benefit from greater specificity. For example, which diagnostic innovations (e.g., imaging, biomarkers) or workplace interventions should be prioritized? Explicit links between the quantified costs and potential cost-saving strategies would enhance the translational value of the study.

I have amended the wording in the conclusion to address this. "This analysis provides a fiscal case for investment in early access to medical assessment and prioritizing diagnosis (e.g. improvement in skilled practitioners in women’s health, improved access to validated imaging modalities such as dynamic ultrasound imaging or MRI, and improved access to specialists who are experienced in diagnosing endometriosis and CPP which includes but not limited to laparoscopic surgery. Further considerations for equitable fertility care access, and culturally responsive, symptom-centered care should be a priority. Addressing this burden is not merely a clinical necessity, it is a national economic imperative."

We appreciate this observation and agree that connecting cost burden to modifiable drivers is critical for policy and investment decision-making. However, this manuscript is intentionally limited to cost-of-illness estimation. Intervention modelling, comparative strategy evaluation, and scenario-based cost-offset analysis are beyond the scope of this paper.

A dedicated Budget Impact Model (BIM) is currently in development using this dataset as its economic base. That model will simulate the potential cost savings associated with earlier diagnosis, improved access to care, and workplace accommodation strategies and will include sensitivity-adjusted scenarios across a range of policy levers.

7)The abstract is dense and could be streamlined for clarity. Phrasing such as “structural inefficiencies in care delivery” should be briefly illustrated with concrete examples.

I have attempted to make this clearer " Endometriosis and chronic pelvic pain (CPP) impose substantial economic burden in New Zealand, yet no national cost-of-illness model currently exists. This study provides the first nationwide estimate using a modified World-Endometriosis-Research-Foundation (WERF) EndoCost protocol incorporating direct healthcare, productivity and carer cost. An attribution correction model was applied to account for diagnostic overlap, assigning 43.4% of CPP cases to endometriosis. Attribution-adjusted annual per capita costs were INT$97,497 (NZD$174,130) for endometriosis and INT$33,262 (NZD$59,406) for CPP. Macroeconomic costs ranged from INT$12.7B-17.7B INT (NZD$22.6B–$31.7B) per annum, depending on prevalence. Productivity losses were the primary cost driver, accounting for 65% of endometriosis and 75% of CPP cost. The unattributed lifetime burden was INT$1.96M (NZD$3.50M NZD) per person for endometriosis and INT$1.54M (NZD$2.74M) per person with CPP. This reflects total economic burden over a 34.5-year working lifespan, adjusted for labour-force participation. Diagnostic delays and health system inefficiencies such as poor healthcare access and suboptimal symptom management is likely to be the most significant modifiable contributor to this burden. Addressing this will require investment in healthcare provision and symptom management alongside equitable access to fertility care."

Reviewer 3 Report

Comments and Suggestions for Authors

I find the study analyzed important because it visualizes the economic impact of a disease as prevalent as endometriosis.

It is the first study conducted in New Zealand, following the same methodology as in Australia. Regarding the cost evaluation methodology, I am not qualified to evaluate it, But if we understand the impact of its socioeconomic impact

The only doubt I have is when comparing costs with cases of pelvic pain that:

.We don't know the causes or diagnoses

.The comparison of all cases, regardless of their intensity. Mild or minimal cases may have other, minor repercussions.

.We don't know the cost of physiotherapy (the cornerstone) of chronic pelvic pain treatment

For all these reasons, my overall assessment of the study was high and suggested some recommendations:

.You should specifically mention pelvic pain diagnoses (in a table).

.I'm missing the cost of physical therapy, which is the cornerstone of pelvic pain treatment.

.The results could change if we only analyzed moderate-severe disease (at least clarify this in the discussion).

 

Author Response

I find the study analyzed important because it visualizes the economic impact of a disease as prevalent as endometriosis.

It is the first study conducted in New Zealand, following the same methodology as in Australia. Regarding the cost evaluation methodology, I am not qualified to evaluate it, But if we understand the impact of its socioeconomic impact

The only doubt I have is when comparing costs with cases of pelvic pain that:

.We don't know the causes or diagnoses

.The comparison of all cases, regardless of their intensity. Mild or minimal cases may have other, minor repercussions.

.We don't know the cost of physiotherapy (the cornerstone) of chronic pelvic pain treatment

For all these reasons, my overall assessment of the study was high and suggested some recommendations:

.You should specifically mention pelvic pain diagnoses (in a table).

This has been included and referenced in the introduction. See Table 2.

.I'm missing the cost of physical therapy, which is the cornerstone of pelvic pain treatment.

As this study data was collected using a validated and standardised questionaire, we did not specifically quantify the cost of physical therapy. This therapy may be captured amongst the "other treatments" category and will be included in estimates. This total "other treatments" figure is available within the submitted documents.

.The results could change if we only analyzed moderate-severe disease (at least clarify this in the discussion).
The research maintains the approach that analysing the data by symptom severity rather than by disease severity should be prioritised given that international literature supports this. We have presented our justification for this in our previous work which is well referenced within the manuscript. https://www.nature.com/articles/s41598-022-08464-x

Round 2

Reviewer 1 Report

Comments and Suggestions for Authors

The revised work is now more concise, internally consistent, and terminologically aligned with the current health economics literature. Redundant wording and informal constructions have been eliminated, while the key numerical results have been retained and harmonized. The attribution model and macroeconomic data are now reported more clearly, and the conclusion presents a stronger causal link between diagnostic delay, system inefficiencies, and preventable costs. Overall, the revision improves readability, scientific accuracy, and suitability for publication in health policy or health economics journals.