Characteristics of adults with potential refractory chronic cough identified using an algorithm designed for administrative claims databases: A descriptive study

Abstract

Objective As population-based studies describing the characteristics of patients with refractory chronic cough (RCC) are sparse, the objective of this descriptive study was to identify and describe such patients using an algorithm developed for administrative claims databases and requiring validation in future. Methods We identified adults with chronic cough (N = 782,121) from Optum Clinformatics™ Data Mart as individuals with a ‘cough event’ (primary cough event; based on ICD codes/relevant prescriptions) and ≥2 cough events in the 56–180 preceding days. We applied several exclusion criteria to identify potential RCC cases and stratified them into probable, possible, and unlikely RCC cohorts by the number of cough events during 1-year follow up (≥3, 1–2 or 0 events, respectively). Patient characteristics were described during the year before the primary cough event and follow up. Results 16.8% (n = 131,772) of patients with chronic cough were potential RCC cases: 25.8% probable, 35.9% possible and 38.3% unlikely. The majority were female (66.4–70.5%); median age was 53–60 years. The most common comorbidities and cough-associated complications at baseline were: allergic rhinitis (30.7–39.1%), hypertension (37.3–47.7%), gastro-oesophageal reflux disease (23.7–34.3%), asthma (18.1–27.3%), insomnia (6.3–8.3%) and stress incontinence (2.5–3.9%). Among probable RCC cases, use of several medications was higher during follow up versus baseline: 52.7% versus 49.0% (cough treatments), 73.3% versus 69.0% (respiratory drugs), 40.5% versus 34.2% (gastrointestinal drugs) and 58.8% versus 56.1% (psychotherapeutics). Conclusion Our algorithm requires validation but provides a starting point to identify patients with RCC in claims databases in future studies.

Keywords

Cough database epidemiology algorithm

Introduction

Chronic cough in adults is defined as a cough that lasts for ≥8 weeks.^1,2 It is a feature of many common respiratory conditions such as chronic obstructive pulmonary disease, asthma, rhinitis/sinusitis, upper airway cough syndrome and bronchitis, as well as many common non-respiratory conditions, such as gastro-oesophageal reflux disease (GERD).³ It has a prevalence of ∼10% in the general adult population.⁴ The condition can lead to exhaustion, sleep disruption, social embarrassment, urinary incontinence, interference with speech and depression, significantly lowering quality of life.^5–10 It also results in a considerable economic burden due to a large number of associated healthcare visits from multiple investigations and often referral to secondary care specialists.^11,12 Refractory chronic cough is a cough that persists despite appropriate treatment of the underlying comorbid condition(s), and idiopathic or unexplained chronic cough is a cough where the underlying cause is unidentified.^13,14 These definitions have been used to identify participants in clinical trials,¹ and together they constitute refractory/unexplained chronic cough, hereafter referred to as RCC.

Refractory chronic cough is seen in around 20–46% of patients presenting to specialist cough clinics.^15–17 There is considerable unmet need among this patient population due to the lack of specific treatments, and the ineffectiveness of over-the-counter cough medications.^14,18 Patients with RCC in real-world clinical practice may differ from the selected patients enrolled in clinical trials or those referred to specialised clinics¹⁹; however, identifying them is challenging because chronic cough is considered a symptom, not a distinct medical condition/disease, and, until recently (October 2021), there was no identifying International Classification of Disease (ICD) code. There are limited population-based studies describing the characteristics of patients with chronic cough in clinical practice,^20–22 and those among the RCC population are particularly sparse.¹⁵ Furthermore, no algorithm exists to identify patients with RCC from administrative healthcare databases. Therefore, we aimed to identify patients with potential RCC using a purposefully developed operational algorithm, based on clinical guidelines,^2,23 for application to administrative claims databases, and to describe their characteristics when applied to a specific US claims database. This was a descriptive study; validation of the algorithm was not performed during the course of the study for reasons detailed in the discussion of this article.

Materials and methods

Study design and data sources

The overall study design is depicted in Figure 1. We conducted a retrospective cohort study using data from the Optum Clinformatics™ Data Mart,^24,25 a large administrative claims database that contains de-identified patient-level information from privately insured and Medicare advantage beneficiaries across all 50 US states. During our study period (1 January 2009 to 31 December 2019), the database covered approximately 58 million patients. In addition to member details such as age, sex and date of eligibility, it captures medical claims (ICD-9/ICD-10 coded),²⁵ pharmacy claims, laboratory test results and information on inpatient stays. All Optum CDM data were de-identified and are thus in compliance with the Health Insurance Portability and Accountability Act of 1996 to preserve patient anonymity and confidentiality. Further, in terms of ethical approval, the present study was determined by Bayer's institutional review board to not constitute research involving human subjects according to 45 CFR 46.102(f) and was deemed exempt from board oversight. As patient data were anonymised, informed consent was not applicable.

Figure 1.

Overall study design. RCC: refractory chronic cough.

Algorithm to identify potential RCC

Our operational algorithm was developed based on treatment guidelines from the European Respiratory Society Guidelines²³ and CHEST Expert Cough Panel.² It was based on firstly identifying patients with ‘chronic cough’, and then narrowing down to patients with potential RCC by applying inclusion/exclusion criteria (Figure 2). The algorithm was developed to focus on specificity over sensitivity. Patients with chronic cough were defined as individuals with a ‘cough event’ and with at least two earlier cough events recorded in the 56–180 preceding days during the study period. Any cough event within the 55 days preceding a qualifying cough event was disregarded (Supplemental Figure 1). The first event during the study period that met these criteria was designated as the primary cough event (index date). The 56-day window was stipulated, firstly, to ensure an 8-week chronicity of the cough condition, and secondly to minimise potential misclassification of the primary cough event (i.e. to ensure it was a separate cough event) and not a repeated database entry made at a follow-up visit for an earlier event. A ‘cough event’ was defined as either an ICD code for cough (ICD-9-CM: 786.2 or ICD-10-CM: R05) or a prescription for the following (based on the fact that these medications are primarily indicated for cough): either benzonatate, dextromethorphan (excluding combinations with quinidine) or codeine, excluding combinations with acetaminophen (for pain), butalbital, a barbiturate, aspirin, caffeine, and a methylxanthine (for tension headaches), carisoprodol, and aspirin (for acute, painful musculoskeletal conditions). We intentionally did not include prescriptions for other drugs, such as specific codeine-combinations, gabapentin, amitriptyline and pregabalin, in the definition of a cough event because these drugs have other main indications and so were likely not prescribed for cough for patients who did not have recorded cough event.

Figure 2.

Identification of potential RCC cases from Optums Clinformatics Claims database, 2009–2019 (N = 782,121). ^a Excluding combinations with quinidine. ^b Excluding codeine combinations with acetaminophen (for pain), butalbital, a barbiturate, aspirin, caffeine, and a methylxanthine (for tension headaches), carisoprodol, and aspirin (for acute, painful musculoskeletal conditions). ^cA total of 100 adults were excluded due to missing data on sex. ^d In the year before the index date. ^e A prescription for roflumilast was used as a proxy for the disease in the absence of a diagnostic code. ^f A prescription for nintedanib or pirfenidone was used as a proxy for the disease in the absence of a diagnostic code. ^g A prescription for isoniazid, pyrazinamide, ethambutol, bedaquiline or delamanid was used as a proxy for the disease in the absence of a diagnostic code. ACE: angiotensin-converting enzyme; COPD: chronic obstructive pulmonary disease; RCC: refractory chronic cough.

To identify the subset of adults with potential RCC, we applied the following exclusion criteria: aged <18 years, missing data on sex, <365 continuous enrolment before the index date, a prescription for an angiotensin-converting enzyme (ACE) inhibitor in the 6 months before the index date, a respiratory tract infection in the 12 weeks before the index date, or any of the following in the 365 days before the index date: idiopathic pulmonary fibrosis (or nintedanib/pirfenidone prescription), bronchiectasis, chronic bronchitis, cavity lesions (i.e. solitary pulmonary nodule), unstable rib fracture, tuberculosis (or prescription for isoniazid, pyrazinamide, ethambutol, bedaquiline or delamanid), pleural disease (i.e. pneumothorax, pleural effusion/plaque, pleurisy or pleural mesothelioma) or a reported abnormal spirometry test. Information on chest X-rays or chest computed tomography were not available in the database, patients were followed-up for 1-year from the index date; those with <1 year of follow up after the index date were excluded to enable sufficient data for determining patient characteristics after potential RCC diagnosis.

Probable, possible and unlikely RCC cohorts

Based on the number of subsequent cough events during the 365 days after the primary cough event, patients with potential RCC were stratified into three cohorts – ‘probable RCC’ (3 or more cough events, indicating cough remains unresolved), ‘possible RCC’ (1–2 cough events; inconclusive of RCC but could be) or ‘unlikely RCC’ (zero cough events). For purposes of our descriptive analysis, we identified those who had cough variant asthma, GERD or sinusitis, asthma, or upper airway cough syndrome, based on ICD codes entered in the year before the index date.

Patient characteristics

Patient demographics, comorbidities/possible cough-associated complications and medication use were determined in the year before the index date using searches for relevant ICD codes. Demographics included age, sex and ethnicity. Comorbidities included respiratory and non-respiratory disorders (see Supplemental Methods for a complete list), and potential cough complications (insomnia, stress incontinence, costochondritis, subconjunctival haemorrhage, vomiting and rib fracture). Medications included (potential) cough treatments, such as antitussives, amitriptyline, pregabalin, gabapentin, codeine and inhaled/oral respiratory medications (nasal/inhaled corticosteroids, short-acting beta-antagonists, long-acting beta-antagonists, short-acting muscarinic antagonists, long-acting muscarinic antagonists, leukotriene modifiers and H1 antihistamines), and gastrointestinal drugs (e.g. proton pump inhibitors and H2 blockers), psychotherapeutics (e.g. antidepressants, anxiolytics and neuromodulators) and other medications (including potential respiratory antibiotics, oral corticosteroids and opioids other than codeine; see Supplemental Methods).

Statistical analysis

Characteristics of the patients with potential RCC were described using frequency counts and percentages and stratified by cohort; average age was summarised as median with inter-quartile range (IQR). The analysis of possible cough-associated complications and medication use during the 1-year follow up was not stratified by cohort due to low event numbers in each stratum. All analyses were undertaken using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

Results

Of 782,121 patients identified with ‘chronic cough’, 16.8% remained (N = 131,772) after applying the algorithm criteria. Distribution was as follows: 25.8% probable RCC (N = 33,936), 35.9% possible RCC (N = 47,310), 38.3% unlikely RCC (N = 50,526; see Supplemental Figure 2 for the distribution of recorded cough events during the 1-year follow up). Among probable, possible and unlikely RCC cases, respectively, 29.0%, 27.1% and 24.8 had either cough variant asthma/GERD/sinusitis, 13.1%, 13.1% and 12.5% had either asthma/upper airway cough syndrome, 17.8%, 13.7% and 10.2% had both groups of conditions, and 40.1%, 46.1% and 52.5% had neither.

Baseline characteristics

Demographics of patients with potential RCC stratified by cohort are shown in Table 1. Most patients were female; 70.5% (probable RCC), 67.9% (possible RCC) and 66.4% (unlikely RCC), and more than two-thirds were of white ethnicity. The most common age group at the index date was 60–69 (probable RCC) and 50–59 years (possible and unlikely RCC) (see also Supplemental Figure 3); median age was 60 (probable RCC), 56 (possible RCC) and 53 years (unlikely RCC).

Table 1.

Demographics of patients with potential RCC by cohort (probable, possible and unlikely RCC).

	Probable RCC (≥3 cough events during follow up) N = 33,936	Possible RCC (1–2 cough events during follow up) N = 47,310	Unlikely RCC (0 cough events during follow up) N = 50,526
Sex
Male	10,009 (29.5)	15,209 (32.1)	16,989 (33.6)
Female	23,927 (70.5)	32,101 (67.9)	33,537 (66.4)
Age (years) ^a
18–29	1927 (5.7)	3923 (8.3)	5406 (10.7)
30–39	2882 (8.5)	5437 (11.5)	7065 (14.0)
40–49	4891 (14.4)	7906 (16.7)	9169 (18.1)
50–59	6959 (20.5)	9412 (19.9)	10,021 (19.8)
60–69	7204 (21.2)	8974 (19.0)	8299 (16.4)
70–79	6415 (18.9)	7559 (16.0)	6791 (13.4)
≥80	3658 (10.8)	4099 (8.7)	3775 (7.5)
Median (IQR)	60 (47–72)	56 (43–69)	53 (40–67)
Ethnicity
White	23,540 (69.4)	32,508 (68.7)	34,582 (68.4)
Hispanic/Latino	3750 (11.1)	5548 (11.7)	6219 (12.3)
African- American	3363 (9.9)	4495 (9.5)	4677 (9.3)
Asian	1904 (5.6)	2906 (6.1)	3014 (6.0)
Missing	1379 (4.1)	1853 (3.9)	2034 (4.0)

Data are n (%) unless otherwise specified.

At baseline (index date).

IQR: inter-quartile range; RCC: refractory/unexplained chronic cough.

Insomnia was the most recorded possible cough-associated complication at baseline, occurring in 8.3% (probable RCC), 7.2% (possible RCC) and 6.3% (unlikely RCC) of patients, followed by stress incontinence (2.5–3.9%), vomiting (1.8–2.4%), costochondritis (0.8–0.9) and subconjunctival haemorrhage (0.2–0.4%); there were no cases of rib fracture at baseline (see Table 2; Supplemental Figure 4). The most recorded respiratory comorbidities at baseline were allergic rhinitis (30.7–39.1%), followed by asthma, chronic sinusitis and pneumonia/influenza or other acute lower respiratory tract infection. The most recorded non-respiratory comorbidity at baseline was hypertension (37.3–47.7%), followed by GERD, depression, anxiety and obesity (Table 2; Supplemental Figures 5 and 6). The majority of potential RCC cases had used an inhaled/oral respiratory drug and approximately 50% had used a ‘cough medication’. Use of psychotherapeutics was also high (44.2–56.1%) (Table 2, Supplemental Figure 7).

Table 2.

Baseline possible cough-associated complications, comorbidities (respiratory and non-respiratory) and medications (within 365 days before the index date among patients with potential RCC).

	Probable RCC (≥3 cough events during follow up) N = 33,936n (%)	Possible RCC (1–2 cough events during follow up) N = 47,310n (%)	Unlikely RCC (0 cough events during follow up) N = 50,526n (%)
Possible cough-associated complications
Insomnia (sleep disturbance)	2830 (8.3)	3389 (7.2)	3208 (6.3)
Stress incontinence	1320 (3.9)	1435 (3.0)	1265 (2.5)
Vomiting	811 (2.4)	1022 (2.2)	921 (1.8)
Costochondritis	285 (0.8)	416 (0.9)	445 (0.9)
Subconjunctival haemorrhage	141 (0.4)	157 (0.3)	125 (0.2)
Rib fracture	0 (–)	0 (–)	0 (–)
Respiratory comorbidities
Allergic rhinitis	13,280 (39.1)	16,520 (34.9)	15,521 (30.7)
Asthma	9280 (27.3)	10,898 (23.0)	9154 (18.1)
Chronic sinusitis	6100 (18.0)	7671 (16.2)	7209 (14.3)
Pneumonia/influenza/other acute lower respiratory tract infection	6043 (17.8)	8203 (17.3)	8488 (16.8)
Obstructive sleep apnoea	3029 (8.9)	3815 (8.1)	3408 (6.7)
Chronic rhinitis	2993 (8.8)	3382 (7.1)	2874 (5.7)
Upper airway cough syndrome	2724 (8.0)	3378 (7.1)	3473 (6.9)
Haemoptysis	632 (1.9)	857 (1.8)	726 (1.4)
Nasal polyp disease	374 (1.1)	442 (0.9)	334 (0.7)
Inducible laryngeal obstruction	292 (0.9)	274 (0.6)	248 (0.5)
Sarcoidosis	200 (0.6)	198 (0.4)	165 (0.3)
Non-respiratory comorbidities
GERD	11,624 (34.3)	13,532 (28.6)	11,963 (23.7)
Hypertension	16,176 (47.7)	20,037 (42.4)	18,852 (37.3)
Depression	5522 (16.3)	6598 (13.9)	6406 (12.7)
Anxiety	5377 (15.8)	6894 (14.6)	7065 (14.0)
Obesity	5020 (14.8)	6971 (14.7)	7346 (14.5)
Coronary artery disease	3671 (10.8)	4413 (9.3)	3895 (7.7)
Congestive heart failure	2164 (6.4)	2540 (5.4)	2205 (4.4)
Atrial fibrillation/flutter	1709 (5.0)	2031 (4.3)	1827 (3.6)
Peripheral vascular disease	1339 (3.9)	1496 (3.2)	1388 (2.7)
Respiratory, intrathoracic, laryngeal, or breast cancer	1158 (3.4)	1214 (2.6)	1047 (2.1)
Medications
Cough treatments^a	16,636 (49.0)	24,095 (50.9)	26,765 (53.0)
Respiratory drugs (inhaled/oral)^b	23,429 (69.0)	30,590 (64.7)	30,309 (60.0)
Gastrointestinal drugs^c	11,616 (34.2)	13,197 (27.9)	11,447 (22.7)
Psychotherapeutics^d	19,054 (56.1)	23,258 (49.2)	22,365 (44.2)
Other^e	23,945 (70.6)	33,453 (70.7)	35,108 (69.5)

Antitussives, amitriptyline, pregabalin, gabapentin and codeine.

Corticosteroids (nasal/inhaled), short-acting beta-agonists, long-acting beta-agonists, short-acting muscarinic antagonists, long-acting muscarinic antagonists, leukotriene modifiers and H1 antihistamines.

Proton pump inhibitors, H2 blockers.

Antidepressants, anxiolytics and neuromodulators.

Antibiotics used for potential respiratory infection, oral corticosteroids and opioids other than codeine.

GERD: gastro-oesophageal reflux disease; RCC: refractory/unexplained chronic cough.

Characteristics during 1-year follow up

Insomnia remained the most common possible cough-associated complication during the 1-year follow up (6.8%), followed by stress incontinence (3.0%), vomiting (2.0%), costochondritis (0.6%), rib fracture (0.4%) and subconjunctival haemorrhage (0.3%) (Table 3, Supplemental Figure 8) Among probable RCC cases, medication use was generally higher during the 1-year follow up compared with baseline; 52.7% versus 49.0% (cough treatments), 73.3% versus 69.0% (respiratory drugs), 40.5% versus 34.2% (gastrointestinal drugs) and 58.8% versus 56.1% (psychotherapeutics). Among possible and unlikely RCC cases, use of cough treatments and respiratory drugs was notably lower during follow up versus baseline, while use of gastrointestinal drugs and psychotherapeutics was slightly lower/similar (Table 4, Supplemental Figures 9).

Table 3.

Possible cough-associated complications among patients with potential RCC during the 1-year follow up after the index date.^a

Possible cough-associated complications	All potential RCC cases (any number of cough events during follow up) n (%)
Insomnia (sleep disturbance)	8962 (6.8)
Stress incontinence	3907 (3.0)
Vomiting	2662 (2.0)
Costochondritis	848 (0.6)
Rib fracture	526 (0.4)
Subconjunctival haemorrhage	388 (0.3)

^aNot stratified by RCC cohort due to low numbers of events by strata.

RCC: refractory/unexplained chronic cough.

Table 4.

Medication use among patients with potential RCC during the 1-year follow up after the index date.

Medication	Probable RCC (≥3 cough events during follow up) N = 33,936n (%)	Possible RCC (1–2 cough events during follow up) N = 47,310n (%)	Unlikely RCC (0 cough events during follow up)N = 50,526n (%)
Cough treatments^a	17,872 (52.7)	18,546 (39.2)	4171 (8.3)
Respiratory drugs (inhaled/oral)^b	24,887 (73.3)	27,615 (58.4)	17,954 (35.5)
Gastrointestinal drugs^c	13,738 (40.5)	13,695 (28.9)	10,466 (20.7)
Psychotherapeutics^d	19,947 (58.8)	23,673 (50.0)	22,005 (43.6)
Other^e	23,516 (69.3)	28,958 (61.2)	23,017 (45.6)

^aAntitussives, amitriptyline, pregabalin, gabapentin and codeine.

Proton pump inhibitors and H2 blockers.

Antidepressants, anxiolytics and neuromodulators.

Antibiotics used for potential respiratory infection, oral corticosteroids and opioids other than codeine.

RCC: refractory/unexplained chronic cough.

Discussion

In this study, we identified and characterised patients with potential RCC, including treatments received, through the implementation of an algorithm specifically designed to identify patients with RCC from claims databases. Describing this largely uncharacterised subset of patients with chronic cough is important to help evaluate current management strategies, and guide future investigations.

We found that 16% of patients identified with chronic cough had potential RCC,²⁶ and a quarter of these were deemed to be probable cases. The demographics of RCC cases in our study – a high preponderance of females and white ethnicity, and the age (∼60 years) at RCC ascertainment – are in line with findings from smaller studies of patients with RCC from the United States^15,27 and Spain.²⁸ They also align with demographics of the broad spectrum of patients with chronic cough identified from referrals to specialist clinics,^17,29–31 electronic health records (EHRs),²⁰ administrative claims data^21,32 and internet surveys,³³ although the populations described by Zeiger et al. were more ethnically diverse (less than half were white and up to a third were hispanic).^21,32

Insomnia and stress incontinence were the most common possible cough-associated complications among potential RCC cases, as also seen by Zeiger et al.²¹ among their chronic chough population using claims data from the Kaiser Permanente Southern California claims database.^21,22 As expected, respiratory comorbidities were common, particularly allergic rhinitis (39% of probable cases) and asthma (27% of probable cases), as similarly reported by Zeiger et al.^21,32 and smaller chronic cough populations.^30,34,35 Prevalence of other common respiratory disorders were also in line with findings by Zeiger et al.,³² except for chronic sinusitis and rhinitis, which were notably lower (18% and 9% of probable cases, respectively).^21,32 Reasons for this are unclear but could be due to differences in the ethnic composition or other aspects of the study population, coding, or ascertainment periods. Hypertension, GERD and depression were the most commonly recorded non-respiratory conditions (48%, 34% and 16% of probable cases), and we observed high use of the groups of medication investigated, as also shown previously in chronic cough populations.^{20,21,30,32,34} There is no approved medication for the treatment of chronic cough, and use of a wide range of medications could reflect a trial and error approach to patient management, as found in a recent qualitative study of American primary care physicians.³⁶ Additionally, the generally increased use of medications in the 1-year follow-up period among probable RCC (which was not seen for possible/unlikely cases) could suggest that trial and error treatment strategies are common in patients with RCC. The mostly similar characteristics of our potential RCC cases with other RCC and broader chronic cough populations, goes some way to support our algorithm as a method of identifying RCC from claims databases. However, our algorithm requires validation in other datasets including manual review of patient records to test the performance and refine it further before implementing more widely.

A key strength of our study is its novelty. Previous studies have developed algorithms to identify the chronic cough population as a whole from claims³⁷ or EHR²⁰ databases, but we are unaware of other algorithms specifically designed to identify patients with RCC. Another strength is the use of a large population-based database covering claims data from individuals across 50 US states, which enabled the identification of a large and geographically diverse sample of potential RCC cases. Our algorithm applied the principle that identifying RCC is based on determining patients with chronic cough and exclusion of possible causal factors and/or treatments such as ACE inhibitor use and smoking. The application of several exclusion criteria to maximise specificity of our RCC definition is a study strength; however, we acknowledge that this may have reduced sensitivity. The algorithm by Bali et al.³⁷ designed to identify patients with chronic cough from claims data was found to have 99% specificity and 15.5% sensitivity. Validation of our algorithm is key to evaluating these parameters, as well as to help refine the algorithm if necessary; however, this would be challenging. It would be difficult to identify a gold standard against which to measure the sensitivity and specificity of the algorithm – we have previously conducted a feasibility analysis whereby a review of a random sample of physician's notes was undertaken, and this showed that RCC (or similar terminology) is seldom used, and that substantial heterogeneity exists in how RCC is recorded. Furthermore, RCC can be undiagnosed/undetected for many years thus an individual with potential RCC identified from our algorithm may not necessarily have been picked up in clinical care and recorded in a physician's notes. It is also noteworthy that previous publication around this topic have focused on algorithms for chronic cough rather than RCC,^11,20,22,32 which is unsurprising due to the difficulties in its identification. As no single ICD code (or groups of ICD codes) exists to conclusively identify RCC, we built an algorithm based on multiple types of cough that could be mild, chronic, seasonal, chronic or caused by underlying conditions, and this proved to be a challenging endeavour. We believe that it is important that we share our algorithm with others in the scientific community, in the hope that others can use it as a starting point for developing a more refined algorithm with the opportunity for validation. Another study limitation is that there may have been some selection bias towards capturing the more severe cases of chronic cough because the database only captured individuals who sought medical care for their symptoms and not those who used over-the-counter medications because this is not captured in the database. Personal perspective of the seriousness/triviality of the cough and levels of interference with social roles have been found to be influencing factors in this regard.³⁸ Another study limitation is possible misclassification of cough events and patient characteristics due to coding errors. Some cough-associated complications may also have been under-recorded; for example, stress incontinence and insomnia if patients did not seek medical attention for these. Indeed, in surveys of patients attending cough clinics, 63–65% have been reported to suffer from stress incontinence^5,8; for insomnia, previous estimates have ranged from 26% to 34% based on self-reports among chronic cough patients.^7,9 Also RCC is not an acute event so ascertaining an exact date of onset is challenging, especially in secondary data sources. Lastly, although the database covers individuals in both private or Medicaid Advantage plans, most are the former and therefore our results should be considered more generalisable to privately insured rather than low-income individuals. The results are also not generalisable to populations outside the United States where prescribing practices may differ.

Conclusions

In conclusion, our algorithm provides a starting point to identify patients with RCC in claims databases, although it requires validation to determine whether further refinement is necessary. Findings from our study could potentially support a better understanding of the characteristics of patients with RCC, in turn helping guide decisions around their management. Further research could also explore cough patterns in patients with comorbidities such as GERD, respiratory conditions, insomnia, and obstructive sleep apnoea.

Supplemental Material

sj-docx-1-sci-10.1177_00368504241238080 - Supplemental material for Characteristics of adults with potential refractory chronic cough identified using an algorithm designed for administrative claims databases: A descriptive study

Supplemental material, sj-docx-1-sci-10.1177_00368504241238080 for Characteristics of adults with potential refractory chronic cough identified using an algorithm designed for administrative claims databases: A descriptive study by Sascha van Boemmel-Wegmann, Kenneth W. Altman, Ron Herrera, Philippe Vieira Pires, Priscilla F.A Pichardo and Pareen Vora in Science Progress

Footnotes

Acknowledgements

We thank Susan Bromley, EpiMed Communications Ltd (Abingdon, UK), for medical writing assistance in development manuscript drafts and in accordance with Good Publication Practice.

Author contributions

All authors (SvB-W, KA, RH, PVP, PP, and PV) contributed to the design of the study and protocol development. SvB-W extracted the data and performed the statistical analysis. All authors contributed to the interpretation of the data and the critical review of manuscript drafts, and all approved the final manuscript.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: SvBW and PV are employees of Bayer AG (Germany). RH and PVP were employees of Bayer AG (Germany) at the time the study was carried out. KA has served as a consultant and speaker for Merck and Vindico, Axdev, Lyra Pharmaceuticals, and Bellus Pharmaceuticals. PP declares no conflicts of interest.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was funded by Bayer AG.

ORCID iDs

Priscilla F.A Pichardo

Pareen Vora

Supplemental material

Supplemental material for this article is available online.

Author biographies

Sascha van Boemmel-Wegmann is an integrated evidence generation researcher at Bayer AG. In his role he is responsible for leading the development of observational research activities using real-world clinical practice data. His research focuses on real-world treatment patterns, healthcare resource utilization and drug safety/effectiveness in diverse therapeutic areas such as chronic kidney disease, chronic pulmonary diseases, and advanced cancers.

Kenneth W. Altman, MD, PhD, is chair of Otolaryngology – Head & Neck Surgery for the Geisinger Neuroscience Institute and a professor at Geisinger Commonwealth School of Medicine. He earned a doctorate in biomedical engineering and his medical degree from Duke University. He completed a residency in Otorhinolaryngology – Head & Neck Surgery at the University of Pennsylvania and a fellowship in Laryngology & Care of the Professional Voice at Vanderbilt University. He is a board certified, nationally known authority on laryngology, care of the professional voice, cough, and dysphagia. He has published articles and chapters in over 140 peer-reviewed journals and books, edited 2 books and delivered presentations at more than 260 conferences and courses. He is secretary/treasurer of the American Academy of Otolaryngology – Head and Neck Surgery and past president of the American Laryngological Association (ALA). He is also a fellow of the ALA, the American Academy of Otolaryngology, the American College of Surgeons, the American Rhinologic Society and the Triological Society. His research interests include pre-malignant vocal fold disease, chronic cough, and dysphagia in the presence of neurodegenerative disease.

Ron Herrera, PhD - Ron is a Statistician and Epidemiologist working as RWE Analyst at Boehringer Ingelheim. He has a PhD in Epidemiology and Public Health from the Ludwig-Maximillian University of Munich- Germany. His research interests include statistical methods in epidemiology, phenotype building, and causal inference. He has participated in many network studies using the OMOP CDM in areas like treatment patterns in CKD patients, women's health, cardiovascular disease, and prostate cancer. He is an active contributor to EHDEN and has contributed to other consortiums like PIONEER while working for Bayer Pharmaceuticals.

Philippe Vieira Pires is a board-certified cardiologist. Before joining Pharma Industry, he worked for 15 years as assistant physician of the Heart Failure and Heart Transplantation Team at Hospital Israelita Albert Einstein in São Paulo, Brazil. Having joined Pharma almost ten years ago, he has worked as a Global Clinical Lead for both early and late development, with different assets for cardiovascular, coagulation, pulmonary hypertension, and pulmonology indications. Currently, he leads a Phase 3 mega cardiovascular outcome trial at Johnson and Johnson Innovative Medicine.

Priscilla F.A Pichardo, is a fifth-year resident physician completing an Otolaryngology – Head and Neck Surgery Residency Program. Throughout her training, she developed a passion for treating head and neck cancer and will be completing a Head and Neck Oncology and Microvascular Surgery Fellowship. Her love for the field has inspired her to pursue innovation and contribute to research, which she hope to continue throughout her career.

Pareen Vora, PhD is a senior epidemiologist and a science fellow at Bayer AG. He earned his doctorate in Epidemiology and Public Health from the Ludwig-Maximillian University of Munich- Germany. He has been working in the industry for approximately 10 years. He has been responsible for conducting observational research using real world data in multiple therapeutic areas such as cardiovascular, pulmonology, and digital epidemiology. Currently leading real world evidence activities in critical care space.

References

Morice

Dicpinigaitis

McGarvey

, et al. Chronic cough: new insights and future prospects. Eur Respir Rev 2021; 30: 210127.

Irwin

French

Chang

et al. Classification of cough as a symptom in adults and management algorithms: CHEST guideline and expert panel report. Chest 2018; 153: 196–209.

Chung

Pavord

. Prevalence, pathogenesis, and causes of chronic cough. Lancet 2008; 371: 1364–1374.

Song

W-J

Chang

Y-S

Faruqi

, et al. The global epidemiology of chronic cough in adults: a systematic review and meta-analysis. Eur Respir J 2015; 45: 1479–1481.

Dicpinigaitis

. Prevalence of stress urinary incontinence in women presenting for evaluation of chronic cough. ERJ Open Res 2021; 7: 00012–2021,.

Dicpinigaitis

Tso

Banauch

. Prevalence of depressive symptoms among patients with chronic cough. Chest 2006; 130: 1839–1843.

Domínguez-Ortega

Plaza

, et al. Prevalence of chronic cough and patient characteristics in adults in Spain: a population-based cross-sectional survey. Chron Respir Dis 2022; 19: 14799731221098722.

Zoglmann

Nguyen

Engberts

, et al.

Do patients with stress incontinence cough or do cough patients suffer from urinary incontinence?

Eur Respir J 2015; 46: PA713.

French

Irwin

Curley

, et al. Impact of chronic cough on quality of life. Arch Intern Med 1998; 158: 1657–1661.

10.

Bali

Adriano

Byrne

, et al. Chronic cough: more than just a persistent cough: a systematic literature review to understand the impact of chronic cough on quality of life. Qual Life Res 2023; doi:https://doi.org/10.1007/s11136-023-03556-1.

11.

Birring

Currie

Holden

, et al. The financial burden of treating patients presenting with acute and chronic cough. Curr Med Res Opin 2021; 37: 2175–2184.

12.

Visca

Beghè

Fabbri

, et al. Management of chronic refractory cough in adults. Eur J Intern Med 2020; 81: 15–21.

13.

Irwin

French

Madison

. Managing unexplained chronic cough in adults: what are the unmet needs? Lancet Respir Med 2020; 8: 745–747.

14.

Mazzone

McGarvey

. Mechanisms and rationale for targeted therapies in refractory and unexplained chronic cough. Clin Pharmacol Ther 2021; 109: 619–636.

15.

Sundar

Stark

, et al.

Is laryngeal hypersensitivity the basis of unexplained or refractory chronic cough?

ERJ Open Res 2021; 7: 00793–2020.

16.

Gibson

Vertigan

. Management of chronic refractory cough. Br Med J 2015; 351: h5590.

17.

van den Berg

JWK

Baxter

Edens

, et al. The demographics, clinical characteristics and quality of life of patients with chronic cough from the Isala cough clinic in the Netherlands. ERJ Open Res 2022; 8: 00232–2022.

18.

Andersson

Bonvini

Horvath

, et al. Research highlights from the 2017 ERS International Congress: airway diseases in focus. ERJ Open Res 2018; 4: 00163–2017.

19.

Niimi

. Natural language processing: a promising research tool of chronic cough for the big data era. Chest 2021; 159: 2149–2150.

20.

Weiner

Dexter

Heithoff

, et al. Identifying and characterizing a chronic cough cohort through electronic health records. Chest 2021; 159: 2346–2355.

21.

Zeiger

Xie

Schatz

, et al. Prevalence and characteristics of chronic cough in adults identified by administrative data. Perm J 2020; 24: 1–3.

22.

Zeiger

Schatz

Xie

, et al. Identification and burden of chronic cough in adults in a large managed care organization. D104 phenotypes, prognostic tools, and adjunct therapies in cough, CF and non-CF bronchiectasis. American Thoracic Society International Conference abstracts. American Thoracic Society; 2020, p. A7775.

23.

Morice

Millqvist

Bieksiene

, et al. ERS guidelines on the diagnosis and treatment of chronic cough in adults and children. Eur Respir J 2020; 55: 1901136.

24.

Optum. Clinformatics Data Mart. https://www.optum.com/content/dam/optum/resources/productSheets/Clinformatics_for_Data_Mart.pdf. Accessed 31 January 2024

25.

CMS.gov Centers for Medicare & Medicaid Services. ICD-10 Resource. https://www.cms.gov/medicare/coding-billing/icd-10-codes/icd-10-resources. Accessed 31 January 2024

26.

Song

Morice

. Cough hypersensitivity syndrome: a few more steps forward. Allergy Asthma Immunol Res 2017; 9: 394–402.

27.

Slovarp

Jetté

Gillespie

, et al. Evaluation and management outcomes and burdens in patients with refractory chronic cough referred for behavioral cough suppression therapy. Lung 2021; 199: 263–271.

28.

Dávila

Puente

Quirce

, et al. Characteristics and management of patients with refractory or unexplained chronic cough in outpatient hospital clinics in Spain: a retrospective multicenter study. Lung 2023; 201: 275–286.

29.

Morice

Jakes

Faruqi

, et al. A worldwide survey of chronic cough: a manifestation of enhanced somatosensory response. Eur Respir J 2014; 44: 1149–1155.

30.

Kastelik

Aziz

Ojoo

, et al. Investigation and management of chronic cough using a probability-based algorithm. Eur Respir J 2005; 25: 235–243.

31.

Campi

Noale

Fabbrizzi

, et al. The demographic and clinical characteristics of an Italian population of adult outpatients with chronic cough. Aging Clin Exp Res 2020; 32: 741–746.

32.

Zeiger

Schatz

Butler

, et al. Burden of specialist-diagnosed chronic cough in adults. J Allergy Clin Immunol Pract 2020; 8: 1645–1657.e7.

33.

Chamberlain

Garrod

Douiri

, et al. The impact of chronic cough: a cross-sectional European survey. Lung 2015; 193: 401–408.

34.

Weiner

Weaver

Gowan

, et al. Health-related experiences of adults with chronic cough. D37 A pulmonary and critical care medley. American Thoracic Society International Conference abstracts. American Thoracic Society; 2020, p. A6606.

35.

Lätti

Pekkanen

Koskela

. Defining the risk factors for acute, subacute and chronic cough: a cross-sectional study in a Finnish adult employee population. BMJ Open Jul 2018; 8: e022950.

36.

Gowan

Huffman

Weiner

, et al. Management of chronic cough in adult primary care: a qualitative study. Lung 2021; 199: 563–568.

37.

Bali

Weaver

Turzhitsky

, et al. Development of a claims-based algorithm to identify patients with chronic cough. TP65 TP065 environmental exposures and lung disease. American Thoracic Society International Conference abstracts. American Thoracic Society; 2021, p. A3146.

38.

Cornford

. Why patients consult when they cough: a comparison of consulting and non-consulting patients. Br J Gen Pract Nov 1998; 48: 1751–1754.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.19 MB