Observational retrospective study calculating health service costs of patients receiving surgery for chronic rhinosinusitis in England, using linked patient-level primary and secondary care electronic data

Objectives Chronic rhinosinusitis (CRS) symptoms are experienced by an estimated 11% of UK adults, and symptoms have major impacts on quality of life. Data from UK and elsewhere suggest high economic burden of CRS, but detailed cost information and economic analyses regarding surgical pathway are lacking. This paper estimates healthcare costs for patients receiving surgery for CRS in England. Design Observational retrospective study examining cost of healthcare of patients receiving CRS surgery. Setting Linked electronic health records from the Clinical Practice Research Datalink, Hospital Episode Statistics and Office for National Statistics databases in England. Participants A phenotyping algorithm using medical ontology terms identified ‘definite’ CRS cases who received CRS surgery. Patients were registered with a general practice in England. Data covered the period 1997–2016. A cohort of 13 462 patients had received surgery for CRS, with 9056 (67%) having confirmed nasal polyps. Outcome measures Information was extracted on numbers and types of primary care prescriptions and consultations, and inpatient and outpatient hospital investigations and procedures. Resource use was costed using published sources. Results Total National Health Service costs in CRS surgery patients were £2173 over 1 year including surgery. Total costs per person-quarter were £1983 in the quarter containing surgery, mostly comprising surgical inpatient care costs (£1902), and around £60 per person-quarter in the 2 years before and after surgery, of which half were outpatient costs. Outpatient and primary care costs were low compared with the peak in inpatient costs at surgery. The highest outpatient expenditure was on CT scans, peaking in the quarter preceding surgery. Conclusions We present the first study of costs to the English healthcare system for patients receiving surgery for CRS. The total aggregate costs provide a further impetus for trials to evaluate the relative benefit of surgical intervention.


REVIEWER
Yong, Michael The University of British Columbia, Otolaryngology -Head and Neck Surgery REVIEW RETURNED 15-Oct-2021

GENERAL COMMENTS
Thank you for this thorough review of the costs related to CRS in the UK. It is an important addition to the literature which will allow further subsequent research. The methodology is well-done and the discussion is appropriate.
I was wondering if you might add, as an additional issue in the limitations in the discussion section, the absence of included costs related to more severe forms of CRS like AERD, and how medical therapies such as mono-clonal antibodies and allergy/asthmarelated costs are not included in this analysis. Since it is known that about 10% of patients with CRSwNP have AERD, additional costs related to aspirin desensitization or mono-clonal antibody medical therapy such as with dupilumab may be sizeable, but do not seem to be captured in this study. Given the nature of the dataset used, it is understandable that these costs are not captured -however, it may be helpful to mention these gaps in the discussion.
In addition, I was wondering about the costs that do not seem to be captured regarding specialist visits. Reading through the article, it seems that costs of healthcare visits are only associated with primary care. However, I assume that when a patient receives a surgery, they will follow-up with a specialist for surgery-related follow-up and complications. Was this included in the analysis?

Gill, Amarbir
The University of Utah School of Medicine REVIEW RETURNED 21-Oct-2021

GENERAL COMMENTS
The authors should be commended on their investigation focused on research area that has no data to date specific to the UK and thus represents a critical knowledge gap. Overall, this is a very well-written manuscript; it is clear, concise, and flows well. It represents a good first step in identifying potential variables that should be further considered when analyzing cost and management of CRS. I have a few concerns/suggestions that are outlined below.
1. Consider shortening/combining the first two paragraphs of the introduction as there is information here that is not critical to the narrative. This would allow the reader to get to the third paragraph sooner, which does a great job of setting up the question/objective of the study and how it is going to be answered in the rest of the paper 2. Within the limitations, please also acknowledge a potential for error in selection of procedure codes for correctly capturing surgeries for CRS only. For example, inpatient costs are much higher in the "unknown polyp" group compared to the "polyp" groupthis observation calls into question whether non-CRS surgeries (ie skull base resections) were potentially included in the procedures done in the "unknown polyp," as inpatient stays are more likely in skull base resections. I understand this may not be easy to uncover, but it should at least be acknowledged as a limitation of the study.
3. How did the study address potential revision surgeries that occurred during the f/u period? This is briefly alluded to in the discussion section but would be better placed in the methods section. For example, when revision surgeries did occur, were they simply overlooked/ignored in calculating costs? 5% (those that needed revision surgery during the time period) is not insignificant if it is not being incorporated into costs. This is a limitation of the study, as it prevents the study from portraying the true cost of CRS care. If not able to incorporate the costs of the revision surgeries, it may be advisable to remove all patients who had revision surgery and only includ those who required one surgery during the time period analyzed (this would allow for more homogeneity within the analyses). Thanks for this comment. The purpose of the Poisson regression was to obtain estimated rates and 95% confidence intervals for service use, and yes it was used for each of the services, and this latter point has been clarified in section 2.3.
Regarding the query about "confounding factors", the authors note that we were not estimating a treatment effect as such, so there would be nowhere to explore the confounding of any treatment effect by way of some intermediary/confounding variable. It could be that the reviewer is suggesting that the service use costs might vary by demographic group, which is a possibility and is something that could potentially be explored in future work. We were interested here in calculating a mean overall populationcost, and have not split this by e.g. age group or sex. The population that we were working with comes from data that are considered to be representative of the population at large, and so we consider that the mean costs we have calculated are appropriate for the population in England. This point is mentioned in the Discussion and a note has been added in the first paragraph of the Discussion for further clarity. We have also modified the third bullet point in the Article Summary to remove any hint or suggestion of a comparison potentially being made, as no comparison was made, and no treatment effect was estimated, and we hope that this helps to clarify these issues.

8
Reviewer 1 It would be helpful to include a flow chart describing how participants were identified and included in the study.
Participants were identified using combinations of the code lists that are given in the Appendix. We have constructed a flow chart as suggested and added this to the Supplementary Materials (referring to it in section 2.1 of the manuscript), thanks very much for the suggestion. 9 Reviewer 2 I was wondering if you might add, as an additional issue in the limitations in the discussion section, the absence of included costs related to more severe forms of CRS like AERD, and how medical therapies such as monoclonal antibodies and allergy/asthma-related costs are not included in this analysis.
Since it is known that about 10% of patients with CRSwNP have AERD, additional costs related to aspirin desensitization or monoclonal antibody medical therapy such as with dupilumab may be sizeable, but do not seem to be captured in this study. Given the nature of the dataset used, it is understandable that these costs are not captured -however, it may be helpful to mention these gaps in the discussion.
Thanks for this interesting suggestion that addresses the generalisability of our work outside the UK context. Monoclonal antibodies are not available in the English NHS for the management of CRSwNP, and therefore no patient in our study will have received dupilumab or any other mAbs. Aspirin desensitisation has very restricted availability in the UK and is only offered in a small number of centres (Guy's, Wythenshaw and a few others) so will not have been captured, but applies to only very small numbers. We have added some text on this in the Discussion, thanks.

10
Reviewer 2 In addition, I was wondering about the costs that do not seem to be captured regarding specialist visits. Reading through the article, it seems that costs of healthcare visits are only associated with primary care. However, I assume that when a patient receives a surgery, they will follow-up with a specialist for surgery-related follow-up and complications. Was this included in the analysis?
Yes, hospital care of this type was also included (this is referred to in the manuscript as inpatient and outpatient secondary care), and specialist visits would be included in here. We have adjusted some of the wording around hospital/health service care to hopefully clarify this for an international audience.

11
Reviewer 3 Consider shortening/combining the first two paragraphs of the introduction as there is Thanks for this helpful comment, we have re-information here that is not critical to the narrative. This would allow the reader to get to the third paragraph sooner, which does a great job of setting up the question/objective of the study and how it is going to be answered in the rest of the paper worded in the Introduction as suggested.

12
Reviewer 3 Within the limitations, please also acknowledge a potential for error in selection of procedure codes for correctly capturing surgeries for CRS only. For example, inpatient costs are much higher in the "unknown polyp" group compared to the "polyp" groupthis observation calls into question whether non-CRS surgeries (ie skull base resections) were potentially included in the procedures done in the "unknown polyp," as inpatient stays are more likely in skull base resections. I understand this may not be easy to uncover, but it should at least be acknowledged as a limitation of the study.
Thanks for this comment. There is indeed the potential for coding errors, and limitations regarding coding difficulties and possible errors are discussed in the Discussion section, and some additions have been made here. Procedures that were included are listed in the Supplementary Materials. We note however that the inpatient (day case and elective surgery) costs are in fact higher in the positive polyps group (£2284.63) than in the unknown polyps group (£1117.37) as seen in Table 2. 13 Reviewer 3 How did the study address potential revision surgeries that occurred during the f/u period? This is briefly alluded to in the discussion section but would be better placed in the methods section. For example, when revision surgeries did occur, were they simply overlooked/ignored in calculating costs? 5% (those that needed revision surgery during the time period) is not insignificant if it is not being incorporated into costs. This is a limitation of the study, as it prevents the study from portraying the true cost of CRS care. If not able to incorporate the costs of the revision surgeries, it may be advisable to remove all patients who had revision surgery and only includ those who required one surgery during the time period analyzed (this would allow Revision surgeries would have appeared in the data after the initial surgery, but were not specifically extracted and labelled as revision surgery, rather they were just included as downstream inpatient costs. The codes used to identify and cost revision surgery would have been the same surgery codes as those for which the patients were selected into the cohort, and as such all events within the time period with those codes would have been identified and included in the costs. There was no obvious peak in the inpatient costs at a time after surgery, which suggested that there was no preferred timing for any subsequent surgery, so any revision surgeries would have been for more homogeneity within the analyses).
spread out over the time period after the surgery date, as expected. We have added some words in the Results and Discussion around this.
The comment in the Discussion was aiming to explain that if a patient had their first surgery before 1997, followed by a revision surgery during our time period of 1997-2016, then we would only have found their revision surgery and would unknowingly have considered that to be their first surgery. This has been moved to the Results section (section 3.3) and some extra words have been included there to clarify this.
14 Reviewer 3 The importance of separate analyses focusing of unknown polyp patients are difficult to interpret, given the ambiguity surrounding which patients may actually be present in this cohort. Consider having a polyp only cohort and a total cohort, and removing unknown polyp cohort, or otherwise provide justification for why it is important to consider the unknown polyp cohort separately, what this analysis adds, and how to interpret it given the ambiguity of its definition Thanks for this comment. It illustrates the difficulty of using observational data instead of prospectively collected trial data. If we were in a trial (indeed as we are in the MACRO RCT), we would ask all patients to be screened to see definitively whether or not they had polyps, but in observational data we are not routinely told that patients definitely do not have polyps, there is merely the absence of a polyp code or polypectomy code (see code list in the supplementary materials). The best approximation is to say that if there was no positive reporting of polyps, then either the patient didn't have any or they were too small to warrant investigation, hence using the term 'unknown-polyps' instead of 'without-polyps' for the remaining patients who do not fall into the positive-polyps group. Some additional references to our earlier work where this categorisation has been used have been added in Section 2.1, and it has been added in the Article Summary bullets as well.
The positive polyp group contains CRS patients who definitely had had polyps at some point (notwithstanding coding errors), and the unknown polyp group contains CRS patients who either have no polyps or had very small ones that never gave cause for the doctor to look specifically and find them (or remove them), and therefore no polyps were recorded. The prognosis in the latter group is homogeneous. This is however a limitation of the analysis, and we discuss this in the Discussion section. The results are presented throughout as positive-polyps, unknownpolyps, and total cohort, and we would rather not remove information on the unknown-polyps group as that would seem to leave a gap.
(Incidentally, we wonder if these reviewer's comments may have arisen in part due to geographical differences in the use of terminology. In Europe, CRS is used to describe both CRS with and without polyps (CRSwNP and CRSsNP, respectively), while in North America perhaps they are seen and therefore coded as more distinct entities, where CRS is assumed to be CRS without NP, and having nasal polyps (NP) is not necessarily referred to as CRS with NP, rather is referred to as NP alone?)