Gender-specific differences in care-seeking behaviour among lung cancer patients: a systematic review

Abstract

Background

In the literature, men are often described as unwilling to use healthcare services, whereas women as frequent users. We conducted a systematic literature review to examine the gender differences in healthcare utilisation of lung cancer patients. Our aim was to synthesise evidence to assess whether men and women utilise cancer diagnosis and treatments differently.

Methods

The databases of PubMed, Scopus, Web of Science, EBSCO Host, Ovid nursing, and Cochrane was systematically searched. We used pre-defined eligibility criteria to identify peer-reviewed published literature that reported healthcare use of lung cancer patients. Two reviewers independently screened the title, abstract, full texts and retrieved relevant data.

Results

A total of 42 studies met the eligibility criteria from 1356 potential studies. In these studies, the most commonly measured healthcare utilisation is surgery (n = 19), followed by chemotherapy (n = 13). All the studies were from developed countries and had a higher percentage of male participants. Substantial evidence of heterogeneity in the use of treatments by gender were found. In relation to diagnosis interval and stage of cancer diagnosis, it was found that women had longer diagnostic intervals. Nonetheless, women tend to get diagnosed at an earlier stage. Furthermore, women had a higher probability of using inpatient cancer-care services and surgical treatments. Conversely, men had greater risks of readmission after surgery and longer length of stay. Lastly, there were no significant gender differences in the likelihood of receiving chemotherapy and radiation therapy.

Conclusion

This study synthesised evidence of disparities in the use of lung cancer treatments based on gender in developed countries, with no evidence available from least-developed and developing countries. Further studies are required to understand this gender-specific inequality and to design interventions to improve the survival rate of lung cancer patients.

Electronic supplementary material

The online version of this article (10.1007/s00432-020-03197-8) contains supplementary material, which is available to authorized users.

Introduction

Lung cancer is the most frequently diagnosed cancer in both men and women globally (11.6% of total cancer cases in 2018). It is the principal cause of cancer-related mortality (18.4% of total cancer deaths in 2018) (Bray et al. 2018); 1.18 million deaths in men and 0.58 million in women worldwide in 2018 (International Agency for Research on Cancer 2018; Thun et al. 2018). There is evidence in the last few decades of a reducing trend in age-adjusted lung cancer mortality rate among men but increasing in women mainly due to the increasing smoking rate among women (Australian Institute of Health Welfare 2018; Siegel et al. 2014). Most recent reviews of the literature suggest that the peak of lung cancer incidence in women is yet to appear, attributed to a myriad of factors. The later peak in the uptake of nicotine smoking amongst women and a subsequent later peak in passive smoking exposure are important factors (Henley et al. 2018; Torre et al. 2016). Moreover, increasing lung cancer incidence is placing an additional burden on the healthcare system, and it is becoming a major component of overall health expenditure in developed countries (Sullivan et al. 2011). Therefore, lung cancer has been a major area of scientific study (both quantitative and qualitative) in recent years, especially in developed countries.

Lung cancers are commonly classified into two types: small cell (SCLC) and non-small cell lung cancer (NSCLC) and with clinical, therapeutic and pathophysiological implications (Patel et al. 2007). National Institute for Clinical Excellence (NICE) and the American Society of Clinical Oncology (ASCO) are continuously updating their evidence-based guidelines on lung cancer treatments. These guidelines, as well as the latest literature, recommend surgical resection for early-stage (Stage I and II) patients with minimal comorbidities and medical complications, while chemotherapy alone, radiotherapy alone, or a combination of both, is advised for advanced-stage (Stage III and IV) patients with good prospects of survival as determined by their functional status and other health comorbidities (Patel et al. 2007; Rich et al. 2011). Despite these closely followed guidelines, significant heterogeneities in men’s and women’s lung care mortality and 5 year survival rates are well documented. There is limited evidence, and it is not clear whether the differences are biological or social in nature, or due to psychological differences driven by the gender-specific characteristics (Bird and Rieker 1999).

The question of discrepancies in healthcare utilisation by lung cancer patients differentiated by gender is a source of heightened debate. Previous literature concluded that lung cancer in women contains specificities that distinguish it from lung cancer in men (Mennecier et al. 2003). For example, adenocarcinoma is the most frequent histologic subtype, and it is more common in women than men (Levi et al. 1997). Several advances in molecular subtype determination of lung cancers have occurred. For example, in adenocarcinomas, the discovery of the epidermal growth factor (EGFR) mutation, or similar ROS, ALK and K-RAS mutations, has led to the subsequent utilisation of targeted immunotherapy in certain populations yielding much better prognoses. Another example, amongst several such advances in the last decade, lung adenocarcinomas in never-smoker females, especially those of south-east Asian descent, have been shown to have immensely positive prognoses when treated with EGFR inhibitors in targeted immunotherapy, even in advanced-stage cases (Grigoriu et al. 2015). Furthermore, Schreiber et al. (1997) found that the likelihood of Allele loss was higher, and Lai et al. (2005) concluded higher frequency of oestrogen receptor hypermethylation among male lung cancer patients. Such discoveries resulting in paradigm shifts in lung cancer treatments raise the question of whether gender-specific biological differences may drive the choice of lung cancer treatments irrespective of the stage of diagnosis. In a recent study, Hunt et al. (2010) stated the importance of the further investigation to compare the gender-specific differences in healthcare utilisation of lung cancer patients which will assist in developing modulated policies and practices concerning critical cancer care. Understanding the influence of gender on the pattern of healthcare utilisation or care-seeking behaviour of lung cancer patients is pivotal to improving the survival rates, of both genders.

Importantly, there are several social determinants that affect health and healthcare utilisation. Adverse health outcomes are often associated with income, education level, race, gender, rurality, health behaviour, lifestyles, environmental factors and minority populations (Woolf and Laudan 2013). Past studies have found increased lung cancer risk due to socioeconomic position (Strand and Kunst 2012; De Kok et al. 2008), along with disparities in smoking  (Menvielle et al. 2009) and occupational exposures (Menvielle et al. 2010). Noticeably for women, the risk of cancer was strongly associated with education and housing characteristics, yet to the contrary, all the indicators of socioeconomic position have an equal and independent impact on the risk of cancer for men (Merletti et al. 2011).

The available literature supports evidence of inequalities in lung cancer incidence and mortality by gender (due to biological and socioeconomic factors). Nonetheless, according to the best of the authors’ knowledge, no study has yet attempted to systematically review the literature to examine the gender-specific variations in the healthcare utilisation of lung cancer patients. Health policymakers recognise reducing disparity (e.g. gender specific) in cancer mortality as a critical priority; however, designing appropriate interventions is often impeded by incomplete or inadequate evidence (Chirikos et al. 2008). In this systematic review, the aim is to provide a comprehensive overview of the available evidence and obvious omissions in the current literature on the gender-specific differences in healthcare utilisation of lung cancer patients. This will involve summarising features such as study design and measurement aspects of treatment, assessment of the quality of the reported results, and to provide a narrative synthesis of the key findings. In this review, seven categories of healthcare utilisation (from diagnosis to radiation therapy) have been included. The narrative synthesis of the current evidence from the peer-reviewed published papers presented in this review will assist in understanding the current knowledge gaps and to better understand why men and women have differing adoption and utilisation rates of the range of lung cancer treatments. These are important issues to identify and potentially influence healthcare provision, policies and procedures in developing and implementing interventions to reduce lung cancer mortality.

Methods

This study used the PRISMA (preferred reporting items for systematic reviews and meta-analysis) guidelines to develop the systematic literature review (McInnes et al. 2018). A review protocol was registered with the International Prospective Register of Systematic Reviews (PROSPERO) on March 11, 2019 (registration number CRD42019124672) (Online Appendix A).

Literature search

The online databases of PubMed, Scopus, Web of Science, EBSCO Host, Ovid nursing, and Cochrane were searched from inception to April 2019 to find articles that estimated the healthcare utilisation of lung cancer patients, based on gender. The following categories of care were considered: diagnosis interval and screening for lung cancer, physician visits, emergency department visits, hospital admissions and readmissions (including the length of stay), and life-extending treatments such as surgery, chemotherapy and radiation therapy. In the clinical context for stage I and II lung cancer patients, all therapies including surgery, neoadjuvant or adjuvant chemotherapy and radiotherapy, are offered with a curative intent. Conversely, for advanced-stage (stage  IV) lung cancer patients, chemotherapy and radiotherapy when offered is with a palliative intent, where the primary aim is symptom control, and may or may not be associated with the by-product of small-to-modest increase in the duration of survivability. With this clinical context in mind, studies including chemotherapy and/or radiotherapy for advanced-stage lung cancer patients were included. When exploring studies of radiotherapy treatment, all radiotherapy offered to lung cancer patients were considered part of life-extending therapies, including extra-thoracic radiation treatment in stage  IV metastatic disease.

Only peer-reviewed published articles written in the English language were considered. No time/date restriction was applied. Additional literature was identified by scanning the references of the selected articles.

The complete search strategy is provided in the study protocol. Search terms included gender, healthcare use and lung cancer. The search strategy was planned using the help of a research librarian and a clinical librarian. Two reviewers independently conducted the search, based on the adopted strategy and the inclusion and exclusion criteria. EndNote (X8) software was used to organise and manage the references.

Eligibility criteria

Journal articles included had healthcare utilisation or treatment of lung cancer patients with the following criteria: (1) the published article was a novel work (including review papers of all kind), (2) the study included adult (18 +) human participants or cohorts, and (3) the study explicitly showed quantitatively specified healthcare use of lung cancer patients by gender. Studies of all types and stages of lung cancer were within the scope of the review, and no limitations were applied to study design.

Excluded studies included those with children and patients suffering from other acute long-term health conditions, as their intensity of care utilisation will differ. Studies with a primary focus on the utilisation of palliative, nursing home and social care were also omitted. The excluded studies of palliative care service utilisation were those focusing solely on end-of-life care and did not include studies investigating palliative intent chemo or radiotherapy. In addition, quantitative studies that reported gender-based differences in lung cancer incidence, risk factors, mortality or survivorship, disease management and patient satisfaction were not included. All studies that did not present gender-specific differences in care-seeking events were excluded. Finally, qualitative studies that did not document the occurrences of healthcare use were also deemed outside the scope of this study.

Study selection and data extraction

Studies identified from the database search were independently evaluated by two investigators (RHR and FA) to assess their eligibility for inclusion using the title, abstract, and screening of full text (if required). Disagreements were resolved by discussion between the two reviewers, and KA was involved when no agreement could be reached, and KA resolved the conflict. Four studies were not included due to disagreements, where RHR was in favour of including those studies but FA and KA disagreed. No other major issues arose. Full-text versions of the selected articles were further examined by KA and JG. Lastly, the reference lists of the included articles were also searched for potential additional studies by RHR.

The following data from each selected study were extracted: author, year of publication, journal, study design, study population and setting, sample size, distribution of gender and age, stage and type of lung cancer, types of treatment used and main outcome measures and key findings. Data extraction was conducted by RHR using the guidelines provided by the PRISMA statement (McInnes et al. 2018). KA and JG verified the extraction of data from the selected studies.

Assessment of study quality

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement checklist was used to evaluate the quality of the included studies (Von Elm et al. 2007). Nineteen key items from the checklist were used: background, objective, setting, participants, data source, study size, quantitative variables, statistical method, missing data, sensitivity analysis, descriptive data adjusted and unadjusted results, limitations, interpretation of findings and funding sources of the study (items 2, 3, 5–8, 10–14, 16,18–20, 22). The quality appraisal was conducted by RHR and FA independently, which was rechecked by JG. Each item was coded Y = present, N = not present, P = partially present and N/A = not applicable followed by the calculation of the percentage positive judgement.

Data synthesis

Significant heterogeneity was observed in study design, methods, measures of outcomes and key findings among the selected studies, and therefore, a meta-analysis was not conducted. Instead, a descriptive analysis of the characteristics of the included studies was performed, and a qualitative synthesis of the principal outcomes was developed. The main focus was to identify both qualitative and quantitative estimates that reported variations in healthcare use of lung cancer patients by gender. The different categories of treatment use reported were also evaluated as to whether the papers concluded any significant differences in care usage or care-seeking behaviour by gender.

Results

Identification of studies

The title and abstract of 1356 articles were screened, and 190 full texts were reviewed, out of which 42 studies met the inclusion criteria and reported gender-specific differences in healthcare use of lung cancer patients (Fig. 1). Forty-one of these were population-based observational studies, and there was a single literature review. No randomised control trials which investigated the primary question of this study were identified.

Fig. 1.

Framework of the systematic literature review process

Study characteristics

In total, the main objective of 38 (out of 42) selected articles were to examine healthcare utilisation of lung cancer patients, 1 evaluated health expenditure, and 3 papers studied both (Table 1). The articles were published from 1982 to 2018 and 21 of these were published after the year 2010, while only 4 were published before 2000. Noticeably, the study setting and population of all the included papers were primarily from high-income countries (North America = 17; Europe = 20; Australia and New Zealand = 2, Japan = 1 and South Korea = 1). The median time gap between final observation year and publication year is in the range of 3–5 years. The minimum age for the inclusion of lung cancer patients in these papers were 18 years. In total, 11 papers (26.9%) specifically looked at older age groups (60 +) and 5 did not specify the age range of interest. The studied population size ranged from 271 to 186,741.

Table 1.

Characteristics of included studies (n = 42)

Study objective	All studies (%)	Recentness (the gap between final observation year and publication year)	All studies (%)
Healthcare utilisation	38 (90.5)	1–3 years	10 (23.8)
Health expenditure	1 (2.4)	3–5 years	17 (40.5)
Both	3 (7.1)	> 5 years	14 (33.3)
Publication year		Not reported	1 (2.4)
≤ 1990	2 (4.8)	Number of care utilisation reporteda
1991–2000	2 (4.8)	Single measure	28 (68.3)
2001–2010	17 (40.4)	Multiple measures	13 (31.7)
2011 ≤	21 (50)	Type of lung cancer studied
Study settinga		Small cell	1 (2.4)
North America	17 (41.4)	Non-small cell	19 (46.4)
Europe	20 (48.8)	Both	8 (19.5)
Australia and New Zealand	2 (4.9)	Not reported	13 (31.7)
Japan and Korea	2 (4.9)	Reported outcome (gender difference)
Minimum age for inclusion (years)		Significant difference in care use	20 (47.6)
Any age	5 (11.9)	No significant difference in care use	20 (47.6)
18 +	5 (11.9)	Mixed findings	2 (4.8)
30 +	6 (14.3)	Percentage of male participantsa
40 +	3 (7.1)	50%	3 (7.3)
50 +	7 (17)	50–60%	19 (46.3)
60 +	11 (25.9)	> 60%	17 (41.5)
Other/not specified	5 (11.9)	Not reported	2 (4.9)

^aIndicates n = 41 as we selected one systematic literature review, some of the criteria were not applicable. Year of publication and lag was used

There are significant heterogeneities in the types and number of treatments measured. Thirteen studies reported healthcare utilisation data for multiple lung cancer treatment usages and the rest reported a single measure of healthcare use. Nineteen studies considered patients with non-small cell only, and 1 study focused on only small cell lung cancer, while 8 included both types of lung cancer. The gender distribution of the population of the selected studies was predominantly male. Thirty-six of them had more than 50% of the participants as male. Finally, 20 of the included articles reported statistically significant differences in healthcare utilisation of lung cancer patients by gender and 2 reported mixed findings. The rest of the studies concluded that there was no significant gender-based differences.

Table 2 presents the various aspects of healthcare utilisation evaluated in the selected studies. Surgery (n = 19) was the most common nature of healthcare utilisation investigated, followed by chemotherapy (n = 13) and radiation therapy (n = 9) (Table 2). Many of these studies examined all three categories of treatments (Berglund et al. 2012; Mahmud et al. 2003; Smith et al. 1995; Visbal et al. 2004; Wouters et al. 2010).

Table 2.

Specialisation of healthcare utilised in the included studies

Category of care	n	References
Diagnosis/delays in seeking care	4	Din et al. (2015); Lyratzopoulos et al. (2012); Marshall et al. (1982); Neal and Allgar (2005)
Physician visit	2	Kurtz et al. (2006); Shugarman et al. (2008)
Emergency department visit	6	Abel et al. (2015); Beatty et al. (2009); Kurtz et al. (2006); Mitchell et al. (2015); Raine et al. (2010); Sikka and Ornato (2010)
Hospitalisation	9	Kurtz et al. (2006); McDevitt et al. (2013); Mennecier et al. (2003); Ogawa et al. (2015); Nebreda et al. (2016); Puri et al. (2015); Shugarman et al. (2008); Skaug et al. (2009); Wright et al. (2008)
Surgery	19	Berglund et al. (2012); Chirikos et al. (2008); Currow et al. (2014); de Perrot et al. (2000); Mahmud et al. (2003); McMahon et al. (2011); Mehta et al. (2012); Mennecier et al. (2003); Nilssen et al. (2016); Ouellette et al. (1998); Potosky et al. (2004); Raine et al. (2010); Rich et al. (2011); Starr et al. (2013); Smith et al. (1995); Strand et al. (2012); Tammemagi et al. (2004); Visbal et al. (2004); Wouters et al. (2010)
Chemotherapy	13	Berglund et al. (2012); Lairson et al. (2015); Lee et al. (2018); Mahmud et al. (2003); Mennecier et al. (2003); Noonan et al. (2015); Patel et al. (2007); Potosky et al. (2004); Ramsey et al. (2004); Smith et al. (1995); Tammemagi et al. (2004); Visbal et al. (2004); Wouters et al. (2010)
Radiation therapy	9	Berglund et al. (2012); Chirikos et al. (2008); Hayman et al. (2007); Koning et al. (2012); Mahmud et al. (2003); Nilssen et al. (2016); Smith et al. (1995); Visbal et al. (2004); Wouters et al. (2010)

Generally, the methodological quality and comprehensiveness of reporting of the studies were in the range of moderate to good using the STROBE guidelines (Table 3). Background/rationale, study setting, eligibility of the participants, data sources and management, statistical methods with control variables, category of continuous variables and a discussion on main results were reported by all studies either partially or whole. In addition, 19 of the included articles declared funding information. Eighteen of these published papers presented information about the methods used for handling missing data, and only 1 in 5 included a detailed sensitivity analysis.

Table 3.

Methodological quality assessment and depth of reporting

Strobe items	2	3	5	6	7	8	10	11	12A	12C	12E	13A	14A	16A	16B	18	19	20	22
Paper names
Abel et al. (2015)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	Y
Beatty et al. (2009)	Y	P	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	N	Y	Y	N	Y	Y
Berglund et al. (2012)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	Y
Chirikos et al. (2008)	Y	Y	Y	Y	Y	Y	P	Y	Y	Y	N	P	Y	Y	N/A	Y	Y	Y	N
Currow et al. (2014)	Y	Y	Y	Y	P	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	P	N
de Perrot et al (2000)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	P	Y	Y	N	Y	N
Din et al. (2015)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	Y
Hayman et al. (2007)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	N
Koning et al. (2012)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	P	Y	Y	Y	Y	N
Kurtz et al. (2006)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	P	Y	Y	Y	Y	Y
Lairson et al. (2015)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	P	Y	Y	Y	Y	Y
Lee (2018)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	N	Y	Y	Y	Y	Y	Y	Y
Lyratzopoulos et al. (2012)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
Mahmud et al. (2003)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	N
Marshall et al. (1982)	Y	Y	Y	Y	P	Y	N	P	Y	N	N	N	N	N	N/A	Y	Y	Y	N
McDevitt et al. (2013)	Y	P	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y
McMahon et al. (2011)	Y	P	Y	Y	N/A	Y	Y	Y	Y	N	Y	Y	Y	N	Y	Y	P	Y	N
Mehta et al. (2012)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	N
Mennecier et al. (2003)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	P	Y	Y	P	Y	N
Mitchell et al. (2015)	Y	Y	N/A	N/A	Y	Y	Y	N/A	N/A	N/A	N/A	Y	N/A	N/A	N/A	Y	Y	Y	Y
Neal and Allgar (2005)	Y	Y	P	P	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N/A	Y	Y	P	Y
Nebreda et al. (2016)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	P	Y	Y	Y	Y	N
Nilssen et al. (2016)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	N
Noonan et al. (2015)	Y	Y	Y	Y	P	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	P	Y	Y
Ogawa et al. (2015)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	P	Y	Y	Y	Y	N
Ouellette et al. (1998)	Y	Y	Y	Y	P	Y	Y	P	Y	N	N	Y	P	N	N/A	Y	N	Y	N
Patel et al. (2007)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	N	Y	N
Potosky et al. (2004)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	N
Puri et al. (2015)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	P	Y	Y	Y	Y	Y	N
Ramsey et al. (2004)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	N
Raine et al. (2010)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	Y
Rich et al. (2011)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	Y
Shugarman et al. (2008)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	N
Sikka et al. (2010)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N
Skaug et al. (2009)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	Y	Y	N
Smith et al. (1995)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	P	N	Y	Y	P	Y	Y
Starr et al. (2013)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	Y	Y	Y	Y	Y	Y	Y	Y
Strand et al. (2012)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	P	Y	Y	Y	Y	N
Tammemagi et al. (2004)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	P	Y	Y	Y	P	Y	Y
Visbal et al. (2004)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	Y	Y	Y	Y	Y	N	Y	Y
Wright et al. (2008)	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	Y	N
Wouters et al (2010)	Y	Y	Y	Y	Y	Y	Y	Y	Y	N	N	N	Y	Y	Y	Y	Y	Y	Y
% positive judgements	100%	93%	97.50%	97.50%	90.4%	100%	95%	95%	100%	45%	19%	93%	87.50%	67.50%	100%	100%	73%	95%	45%

2. background/rationale, 3. objective, 5. setting, 6. eligibility of the participants, 7. variables, 8. data sources/measurement, 10. study size, 11. quantitative variables, 12a. statistical methods with control variables, 12c. addressing missing data, 12e. sensitivity analysis, 13a. participant number, 14a. descriptive data, 16a. main results, 16b. category of continuous variables, 18. key results, 19. limitations, 20. interpretation, 22. funding. Items 1, 4, 9,12b, 12d, 13b, 13c, 14b, 14c, 15, 16b, 16c and 17 were not applicable for assessing the papers included in this study

Y present, N not present, P partially present, N/A not applicable, % positive judgements total /total number of papers

Summary of the key findings

The measures of healthcare utilisation during active lung cancer treatment period were structured into seven categories. Table 4 summaries the characteristics of the 42 included studies. Most of the studies (n = 28) investigated a single modality of treatment use.

Table 4.

Included studies on gender difference in healthcare utilisation of lung cancer patients

First author, year	Sample characteristics	Years	Gender distribution	Stage of cancer studied	Outcome measured or type of treatment estimated	Key statistically significant findings	Key statistically negative findings
Abel et al. (2015)	National Cancer Data repository (England) N = 162,543	2006–2010	Not reported	Not specified	Emergency presentation	1. Women had a significantly greater risk of emergency presentation than men for lung cancer: (odds ratio) women 1.06 vs men 1.00, p < 0.05	N/A
Beatty et al. (2009)	Auckland-Northland regional oncology service and ED data of District Health Boards (New Zealand) N = 478	2004	M = 55.85; F = 44.15	Localised Locally advanced Metastatic	Emergency department presentation	N/A	1. Gender was not associated with ED presentation for lung cancer patients Noticeably, women’s ED presentations were often associated with GP referrals, but men’s ED visits were mostly self-referral 2. ED presentations were associated with GP referral in women 48% vs men 37% ED presentations were associated with no GP referral in women 52% vs men 63%
Berglund et al. (2012)	Thames Cancer Registry in South East England (England) N = 15,582	2006–2008	M = 57.1; F = 42.9	I; II; III; IV	The likelihood of receiving surgical resection, radiation therapy and chemo therapy	1. Women showed a lower likelihood of receiving chemotherapy in advanced-stage lung cancer or SCLC. The adjusted odds ratio: men 1.0 vs women 0.90, p < 0.05	1. There were no significant differences in the probability of having surgical resection in early-stage NSCLC and radiation therapy in stage III disease
Chirikos et al. (2008)	Surveillance, Epidemiology, and End Results. N = 73,771 (SEER) cancer registry files linked to Medicare claim files (USA) N = 73,771	1986–1999	M = 54.61; F = 45.39	I; II; III; IV	Use of surgery and radiation therapy	N/A	1. No significant difference in treatment use between men and women 2. Women enjoy a slight advantage to be candidates for surgery. Descriptive statistics (% of patients): (surgery) white men = 27.7 vs white women = 29.0; (radiation therapy) white men = 44.9 vs white women 40.6
Currow et al. (2014)	The NSW Central Cancer Registry (NSWCCR) and hospital patient data were linked by the Centre for Health Record Linkage (Australia) N = 3040	2003–2007	M = 62.5; F = 37.5	I; II	Receiving surgical resection	N/A	1. Women with NSCLC had a higher resection rate than men. However, the result was not statistically significant after adjusting for age, histology type and comorbidity 2. The odds of resection for localised NSCLC were: women 1.05 vs men 1.0
Din et al. (2015)	The Clinical Practice Research Datalink—CPRD and General Practice Research Database-GPRD. It is a longitudinal general practice database of (England and Wales) N = 6552	2007–2010	M = 57.2; F = 42.8	Not specified	Variations in diagnostic intervals (in days)	1. Longer diagnostic intervals were associated with women. They had a longer diagnostic interval than men: Adjusted mean difference (women vs men) = 8 days, p < 0.05	N/A
de Perrot et al. (2000)	The Thoracic Surgery Unit, University Hospital of Geneva (Switzerland) N = 1046	1977–1996	M = 80; F = 20	I; II; III; IV	Use of surgical treatment	1. Type of surgical procedure performed varied significantly by gender 2. Higher proportion men received Pneumonectomy: men 32% vs women 22%, p < 0.05 More women underwent segmentectomy: women 11% vs men 5%, p < 0.05	N/A
Hayman et al. (2007)	Surveillance, Epidemiology and End Results (SEER)-Medicare data (USA) N = 11,084	1991–1996	M = 59; F = 41	IV	Use of radiation therapy	N/A	1. Gender was not associated with receiving of radiation therapy among patients with NSCLC 2. The adjusted odds ratio of receiving radiation therapy were: men 1.0 vs women 0.96
Koning et al. (2012)	Regional cancer registries of four of nine Dutch comprehensive cancer centres (CCCs): Amsterdam (A), Stedendriehoek Twente (B), West (C) and South (D) (The Netherlands) N = 24,185	1997–2008	M = 71; F = 29	I; II; III	Utilisation rate of radiotherapy	N/A	1. Women had a lower probability of utilizing radiotherapy at all stages of lung cancer, although, the results were not statistically significant 2. Probability of receiving radiotherapy (odds ratio) Stage I and II: men 1.00 vs women 0.95 Probability of receiving radiotherapy (odds ratio) Stage III: male 1.00 vs female 0.97
Kurtz et al. (2006)	Data were obtained by a combination of patient interview and patient self-administered questionnaire at four intervals: baseline (wave 1), 3 months (wave 2), 6 months (wave 3), and 12 months (wave 4) (USA) N = 277	N/A	M = 62.5; F = 37.5	Early; late	Number of physician visits, emergency department visits and hospital nights	1. Women reported fewer hospital nights and emergency department visits Hospital nights: men 2.08 vs women 1.16, p < 0.05 Emergency department visits: men 0.19 vs women 0.06, p < 0.05	1. Women reported lower frequency of emergency department visits than men during the active treatment period 2. The mean treatment use during the active treatment period: Physician visits: men 4.43 vs women 3.37
Lairson et al. (2015)	SEER (Surveillance, Epidemiology, and End Results) program- and Medicare-linked data (USA) N = 4884	2006–2009	M = 54.61; F = 45.39	III; IV	Platinum-based chemotherapy and platinu M + targeted therapy	N/A	1. No significant difference in the use of chemotherapy between men and women 2. Descriptive statistics: (platinum-based chemotherapy) men 54.61% vs women 45.39%; (platinu M + targeted therapy) men 53.56% vs women 46.44%
Lee et al. (2018)	National Health Insurance Service Database (South Korea) N = 7298	2007–2012	M = 76.25; F = 23.75	Advanced NSCLC	Use of combination of chemotherapy	1. Lower use of a combination of chemotherapy for women compared to men (odds ratio: men 1.0 vs women 0.71, p < 0.05)	N/A
Lyratzopoulos et al. (2012)	The Eastern Cancer Registration and Information Centre (ECRIC) (UK) N = 16,714	2006–2010	M = 58; F = 42	I; II; III; IV	Lung cancer diagnosis	1. Women were less likely to be diagnosed in advanced stage compared with men for lung cancer 2. The adjusted odds ratio of advanced-stage lung cancer diagnosis: men 1.0 vs women 0.88, p < 0.05	N/A
Mahmud et al. (2003)	The database of the National Cancer Registry (Ireland) N = 7286	1994–1998	M = 65.55; F = 34.45	Not specified	Use of chemotherapy, radiotherapy and surgery	N/A	1. There were no significant differences in patterns of treatment for SCLC and NSCLC by gender
Marshall et al. (1982)	The Great Lakes Area Tumour Service Registry, which re- corded all cancers diagnosed at major hospitals in Western New York and Western Pennsylvania (USA) N = 1976	1970–1977	M = 76.9; F = 23.1	Local disease; regional disease; distant disease	Delay in seeking treatment	N/A	1. Women delay longer in obtaining treatment for lung cancers. However, the result is not statistically significant 2. Mean reported delay (in months) between first symptom notice and cancer diagnosis: Men 2.6 vs women 2.7
McDevitt et al. (2013)	The National Cancer Registry (NCR) and the Hospital In-Patient Enquiry (HIPE) database (Ireland) N = 1284	2002–2008	M = 58; F = 42	I; II; III	Length of stay/ Factors predicting longer LOS (upper quartile, > 20 days), and emergency readmission within 28 days	N/A	1. No significant gender difference in hospital readmissions following surgery 2. Median inter-decile range LOS (days): men = 13 vs women 12 Prolonged LOS in patients having resection for NSCLC. Prolonged LOS (> 20 days): (risk ratio) women = 1 vs men = 0.99 3. Emergency readmission in patients having a resection. Readmitted within 28 days after surgery: (risk ratio) women = 1 vs male = 1.06
McMahon et al. (2011)	Eastern Cancer Registration and Information Centre (ECRIC) (UK) N = 18,813	1995–2006	M = 65.42; F = 34.58	Not specified	Use of surgery	N/A	1. No gender-related differences in the percentage of patients treated with surgery 2. The adjusted odds ratio of patients treated with surgery: men 1.0 vs women 0.95
Mehta et al. (2012)	The Surveillance, Epidemiology, and End Results (SEER) database (USA). N = 51,938	1988–2002	M = 53.7; F = 46.2	I; II	Receipts of surgical treatment	1. Women had lower odds of refusing surgical treatment than men 2. The adjusted odds ratio of refusing treatment: men 1.17 vs women 1.0, p < 0.05	N/A
Mennecier et al. (2003)	The population-based cancer registry of the Bas-Rhin, an administrative sub division in eastern France (France) N = 787	1981–1994	M = 88.4; F = 11.6	Limited; extensive	Use of Chemotherapy, surgery, number of hospital admissions and hospital stay	1. Women had significantly longer stay at hospitals than men. The mean length of stay (days): women 74.6 vs men 63.2, p < 0.05	1. Men and women generally underwent the same number and type of tests 2. The mean use of chemotherapy, surgery and hospital admission were not significantly different by gender. Women used chemotherapy more. Conversely, men had a higher number of surgery and hospital admissions
Mitchell et al. (2015)	MEDLINE, EMBASE, CINAHL, EBM Reviews, Science and Social Sciences Citation Indexes, Conference Proceedings Citation Index-Science and Conference Proceedings Citation Index-Social Science and Humanities. Number of studies included = 22	1996–2014	Not applicable	Not applicable	Risk factor for emergency presentations	N/A	1. Women are more at risk of emergency presentation for lung cancer compared to men 2. Women showed a higher probability of being diagnosed for lung cancer following emergency presentation
Neal and Allgar (2005)	Secondary analysis of patient-reported data from the ‘National Survey of NHS patients (UK)	N/A	Not reported	Not specified	Delays in seeking care. Pre-hospital delays, referral delays and secondary care delay	N/A	1. Women cancer patients had longer delays in seeking care than men. However, the findings were not statistically significant 2. The mean delays (in days) for primary care: women 82 vs men 77 The mean delays (in days) for referral: women 35 vs men 32 The mean delays (in days) for secondary care: women 13 vs men 11
Nilssen et al. (2016)	Cancer Registry, Statistics Norway and the Norwegian Patient Register (Norway) N = 24,324	2002–11	M = 58.47; F = 41.53	Localised; regional; metastatic	Treatment using surgery and radiotherapy	1. Women were significantly more likely to undergo surgery. The odds ratio of the multivariate analysis (surgery): women 1.0 vs men 0.84, p < 0.05	1. No significant gender difference in the probability of receiving radiotherapy among the SCLC patients. The odds ratios (radiotherapy): women 1.0 vs men 0.93
Noonan et al. (2015)	British Columbia Cancer Agency database (Canada) N = 744	2009	M = 52; F = 48	IIIB; IV	Wait and watch approach for chemotherapy	1. Men remained significant predictors of not receiving chemotherapy or missed opportunity for chemotherapy 2. Patient characteristics for ‘wait and watch missed’ and ‘wait and watch lost to follow-up: men 58% vs women 42%, p < 0.05	1. No significant difference in ‘immediate’ and ‘wait and watch’ chemotherapy: (odds ratio) men 1.0 vs women 0.88
Ogawa et al. (2015)	Patients underwent curative resections for NSCLC from January 2000 to September 2012 at the Kitasato University Hospital (Japan) N = 969	1998–2012	M = 63; F = 37	IA; IB; IIA; IIB; IIIA; IIIB; IV	Readmission at the hospital after surgery	1. Women had a significantly lower rate of readmission after surgery. Out of 969 lung cancer patients, 33 had readmission after surgery, and 28 of them were men, p < 0.05	N/A
Ouellette et al. (1998)	Patient charts at a university hospital (Canada) N = 208	1988–1990	M = 50; F = 50	I; II; IIIA; IIIB; IV	Utilisation of surgical resection	N/A	1. Men and women received similar treatments (surgical resection) for their disease. However, more women than men refused treatment
Nebreda et al (2016)	National Epidemiological Surveillance System for hospital data (minimum basic data set [MBDS]) managed by the Ministry for Health, Social Affairs and Equality (Spain). N = 298,435	2001–2011	M = 86; F = 14	Not specified	Incidence rate of hospitalisation	1. Women had a higher incidence rate of hospital admissions than men. The incidence rate ratio (IRR) of hospital admission for lung cancer were: men 1.0 vs women 0.133, p < 0.05 2. The hospitalisation rate of men with lung cancer fell significantly from 112.5 in the year 2001 to 107.7 in 2011, p < 0.05 The hospitalisation rate of women with lung cancer increased significantly from 11.8 in 2001 to 23.6 in 2011, p < 0.05	N/A
Patel et al. (2007)	The Thames Cancer Registry (UK) N = 11,215	1994–2003	M = 62; F = 38	I; II; III; IV	Receiving chemotherapy	1. Women were less likely to receive chemotherapy than male lung cancer patients 2. The adjusted proportion of patients receiving chemotherapy: men 16.7% vs women 15.4%, p < 0.05	N/A
Potosky et al. (2004)	National Cancer Institute’s Surveillance, Epidemiology and End Results (SEER) registries (USA) N = 898	1996	M = 56.2; F F = 43.8	I; II; III; IV	Receiving surgical resection and chemotherapy as part of initial therapy	N/A	1. Higher percentage of women received surgical resection or chemotherapy as part of initial therapy than men, although, the differences were not statistically significant 2. Patients receiving surgical resection (adjusted percent): women 72% vs men 67% Patients receiving chemotherapy (adjusted percent): women 44% vs men 39%
Puri et al. (2015)	The National Cancer Database (NCDB) established by the American College of Surgeons and the American Cancer Society N = 5624	1998–2010	M = 49.47; F = 50.53	I; II; III	Hospital readmission after receiving surgery	1. Women had a lower probability of unexpected hospital readmission after surgery compared to men 2. The odds ratio of the risk of postoperative readmission within 30 days of surgery: men 1.16 vs women 1.0, p < 0.05	N/A
Raine et al. (2010)	Hospital episode statistics (HES) dataset. Inpatient treatment delivered by NHS hospitals (England) N = 186,741	1999–2006	M = 55; F = 45	I; II	Emergency admission and surgical treatment	1. Women were more likely than men to undergo surgery for lung cancer and emergency admissions 2. The adjusted odds ratio for emergency admission: women 1.12 vs men 1.0, p < 0.05 The adjusted odds ratio of receiving surgery: women 1.22 vs men 1.0, p < 0.05	N/A
Ramsey et al. (2004)	National Cancer Institute database of the SEER cancer registry linked to Medicare claims (USA) N = 14,875	1994–1999	M = 57; F = 43	IIIb; IV	Chemotherapy use	1. Women were significantly less likely to receive chemotherapy: (odds ratio) men 1.0 vs women 0.87, p < 0.05	N/A
Rich et al. (2011)	The National Lung Cancer Audit (NLCA) linked to hospital episode statistics in (UK) N = 34,513	2004–2008	M = 60; F = 40	IA; IB; IIA; IIB; IIIA; IIIB; IV	Likelihood of receiving surgery		1. Women have a higher likelihood of having surgery but not the finding was not statistically significant 2. The adjusted odds ratio of the likelihood of receiving surgery:men 1.0 vs women 1.06
Shugarman et al. (2008)	Medicare claims file (USA) N = 13,120	1996–1999	M = 57.4; F = 42.6	Not specified	Inpatient, outpatient and physician visits	N/A	1. Women’s adjusted odds of using inpatient care were 1.2 times that of men (95% confidence interval, 1.07–1.33) 2. Gender was not associated with the use of outpatient services: (odds ratio), women 1.00 vs men 1.00 3. Use of physician services did not differ significantly by gender
Sikka et al. (2010)	The Michigan Tumour Registry (inpatient and outpatient claims file), a state-wide, population-based registry (USA). N = 11,281	1996–2000	M = 56.20; F = 43.8	Early; late	A diagnosis associated with the ED visit	1. Women were more likely to have a diagnosis associated with an ED visit: (odds ratio) women 1.13 vs men 1.0, p < 0.05	N/A
Skaug et al. (2009)	Lung cancer patients in the Haugalandet area and Norwegian Cancer Registry and hospital records of Haugesund hospital (Norway) N = 271	1990–1996	M = 79; F = 21	I; II; III; IV	Number of hospital admissions and length of stay	N/A	1. No significant differences between the type of hospitalisation and hospital days between men and women 2. The adjusted hazard ratio of fewer hospital days: men 1.0 vs women 1.2
Smith et al. (1995)	Virginia Cancer Registry (VCR), Medicare Health Insurance Master File (HIM), the Medicare Annual Demographic Files, the Medicare Provider Analysis and Review, Medicare Automated Data Retrieval System (MADRS) file, the Area Resource File (USA). N = 4999	1985–1989	M = 68.6; F = 31.4	Localised; distant	Opting for no therapy. Use of surgery, radiation therapy, surgery + radiation therapy and chemotherapy	1. For patients with loco-regional disease, women were significantly less likely to receive radiation therapy: women 41% vs men 50%, p < 0.05 2. For patients with loco-regional disease, women were significantly more likely to receive surgery: women 34% vs men 26%, p < 0.05	1. No significant variations in treatment utilisation among men and women were found for patients with distant disease
Strand et al. (2012)	Cancer Registry (Norway) N = 2201	1994–1995 2000–2007	M = 68; F = 38	I; II; III; IV	Utilisation of surgical resection	1. Women had a significantly lower rate of surgical resection than men 2. The mean percentage of patients receiving surgical resection in three different time periods were: women 38% vs men 62%, p < 0.05	N/A
Starr et al. (2013)	The Danish Lung Cancer Register, the Central Population Register, the Integrated Database for Labour Market Research and the Danish Hospital Discharge Register (Denmark) N = 5,538	2001–2008	M = 56.24; F = 43.76	I; II; IIIA	Likelihood of receiving surgery	N/A	1. Women with stages I-III NSCLC had a higher probability of not receiving surgery than men 2. The adjusted odds ratio of not using surgical treatment: women 1.19 vs men 1.0
Tammemagi et al. (2004)	Josephine Ford Cancer Centre Tumour Registry and medical records (USA). N = 1155	1995–1998	M = 59; F = 41	I; II; III; IV	Surgery in localised disease and chemotherapy in advance disease	N/A	1. A higher percentage of women had surgery in localised NSCLC and chemotherapy in advanced NSCLC and SCLC 2. In the multivariate analysis, the odds ratio of surgery in Stage I and II of NSCLC is: women 1.0 vs men 0.86 The odds ratio of chemotherapy in Stage III and IV NSCLC and SCLC is: women 1.0 vs men 0.95
Visbal et al. (2004)	Patients diagnosed or confirmed for lung cancer in Mayo Clinic (USA) N = 4618	1997–2002	M = 59; F = 41	IA; IB; IIA; IIB; IIIA; IIIB; IV	Use of radiation therapy, chemotherapy and surgery	N/A	1. There were no significant differences in treatment use between genders 2. Rate of treatment use: (surgical resection) women 51% vs men 48%; (chemotherapy) women 33% vs men 32%; (radiation therapy) women 30% vs men 31%
Wouters et al. (2010)	The population-based Cancer Registry (The Netherlands) N = 43,544	2001–2006	M = 69; F = 31	I; II; III; IV	Resection rate, radiation therapy and chemotherapy or combined modality therapy (chemoradiation)	N/A	1. No significant gender differences were found in the probability of, receiving surgical resection for stage I and II and receiving combined modality therapy for stage III NSCLC 2. The adjusted odds ratio of receiving resection: men 1.0 vs women 1.0 The adjusted odds ratio of receiving combined modality therapy: men 1.0 vs women 0.92
Wright et al. (2008)	The Society of Thoracic Surgeons (STS) general Thoracic surgery database (USA). N = 4979	2002–2006	M = 50; F = 50	Not specified	Hospital length of stay	1. Prolonged length of stay (> 14 days) at hospital were significantly associated with men, p < 0.05 2. The estimated odds ratio of prolonged length of stay after lobectomy for lung cancer: men 1.45 vs women 1.0, p < 0.05	N/A

N/A not available

Diagnosis

Four studies which investigated the diagnostic interval and delays in care seeking were included. Two of these studies investigated gender-specific factors, and the rest assessed other sociodemographic factors (including gender) that influenced cancer screening and diagnostic intervals. Diagnosis interval was defined as the time gap between the first incidence of a lung cancer symptom and the date of cancer diagnosis. Two studies reported that longer diagnostic intervals were significantly associated with women (Table 4). It is noteworthy that both studies (Din et al. 2015; Neal and Allgar 2005) used clinical data from the UK. Conversely, Lyratzopoulos et al. (2012) compared the stage of diagnosis between men and women with data from the UK and concluded that men are more likely to be diagnosed at a more advanced stage of lung cancer. However, this result was not statistically significant.

Physician visit

Two studies focused explicitly on outpatient medical services and physician visits. These studies concluded that no significant gender differences were observed in the number of outpatient services used or number of physician visits during the active and continuing period of treatment.

Emergency department visit

Six papers examined gender-specific differences in emergency department visits. Among these papers, four examined the likelihood of emergency department visits, presentation or admissions and two studied whether gender influences the odds of a lung cancer diagnosis being associated with an emergency department visit. Three studies concluded that women patients have a significantly greater risk of having an emergency presentation than men. Conversely, 1 study obtained data from 277 elderly (65 +) patients in the USA and reported that men had a significantly higher frequency of emergency department visits during the first year following a lung cancer diagnosis. On the other hand, a study conducted with more than 20,000 older (65 +) patients in the USA concluded that women lung cancer patients had a higher probability of a diagnosis associated with an emergency department visit. However, a similar study with data from New Zealand (n = 478) stated that gender was not associated with emergency department presentation.

Hospitalisation and length of stay

Nine studies reported different results on the association of gender and hospitalisation and length of stay. Among these, five papers studied hospital admission and readmissions, three papers studied night stays in hospital, and one reported both categories of health services. Shugarman et al. (2008) using the US data concluded that women’s use of inpatient care services was higher than men and similarly, Nebreda et al. (2016) also identified that the incidence of hospital admissions are higher for women using Spanish data. Conversely, Skaug et al. (2009) with Norwegian data and Mennecier et al. (2003) with French data showed that the hospitalisation rates were comparable in both genders. Two studies established that being male was significantly associated with unanticipated hospital readmissions after surgery (Ogawa et al. 2015; Puri et al. 2015). Furthermore, two US-based studies concluded significant differences in hospital length of stay by gender. Meanwhile, another study employed data from the National Cancer Registry of Ireland and found no statistically significant gender difference in hospital readmissions following surgery and in prolonged (> 20 days) length of stay in patients having resection for NSCLC (McDevitt et al. 2013).

Surgery

Surgery was the most frequent treatment reported in 19 studies. These studies reported significant gender differences. Among these studies, six used data from the USA, four from the UK, two from Norway and one study each from France, Switzerland, Denmark, Netherlands, Ireland, Canada and Australia. Eight out of these 19 studies concluded that women had a higher likelihood of undergoing surgery; however, 1 paper from Norway concluded the opposite, and the rest found no significant gender-specific variations in surgery as a treatment option.

Raine et al. (2010) concluded that men were less likely than women to use lung cancer resection, Smith et al. (1995) found that for patients with loco-regional diseases, women were more likely to have surgery. Nilssen et al. (2016) argued that women had a higher probability of undergoing surgery, Mehta et al. (2012) reported that male patients had a greater probability of refusing surgery, and Currow et al. (2014) demonstrated that women with localised non-small cell lung cancer had higher resection rates than men. Similarly, Chirikos et al. (2008), Rich et al. (2011) and Berglund et al. (2012) also concluded that women patients are more likely to undergo surgery than men, but their findings were not statistically significant. In another study, de Perrot et al. (2000) found that patients in either gender were treated similarly; however, pneumonectomy was more frequently performed on men and women had a higher probability of undergoing a segmentectomy.

Chemotherapy

Thirteen papers included chemotherapy as treatment and the majority of them (8) found no statistically significant difference by gender in receiving it. Five of these eight papers used data from the USA and the remaining three were from France, Ireland and Netherlands. Four studies found that male lung cancer patients were significantly more likely to receive chemotherapy but one concluded that being male remained a significant predictor of not receiving chemotherapy or having a missed opportunity to receive chemotherapy. Berglund et al. (2012) and Patel et al. (2007) estimated that women lung cancer patients were less likely to receive chemotherapy than men. Both of these studies used data from the UK. Lee et al. (2018) (Korea) concluded that female patient’s use of chemotherapy is lower than males. Likewise, another study using the USA data also came to a similar conclusion (Ramsey et al. 2004).

Radiation therapy

Nine studies examined gender differences in undergoing radiation therapy. Only one study (from the USA), Smith et al. (1995), showed that male lung cancer patients (loco-regional disease) had a statistically significant higher probability of receiving radiation therapy than women. Another six studies found similar outcomes, but their findings were not statistically significant. By comparison, two studies (from the UK and Norway) expressed opposite conclusions; nonetheless, their findings were also statistically insignificant (Berglund et al. 2012; Nilssen et al. 2016).

Discussion

This study is the first systematic review examining gender-specific differences in healthcare utilisation among lung cancer patients. There has been an increasing number of studies that investigated the treatment use of lung cancer patients in the last decade. This has increased the prospect of conducting a population-based retrospective of the factors that influence these patients’ care-seeking behaviours or patterns of healthcare use. This is evident because half of the articles selected for this review was published within the last 10 years.

When investigating for gender-specific differences in healthcare utilisation, one must be mindful of the inherent biological basis of gender differences in lung cancer prognoses and outcomes. There are known differences in nicotine metabolism, enzyme-dependent activation and detoxification of tobacco carcinogens, DNA repair mechanisms and hormonal effects, which pose gender-specific differences in lung cancer trajectory and outcomes. Certainly, there are many current studies investigating the link between oestrogen and propensity for lung adenocarcinomas, and exploring whether a similar relationship exists to that of oestrogen and gynaecologic adenocarcinomas (Olak and Coulson 2004). Furthermore, higher incidence of certain genetic mutations, such as K-RAS in male smokers with lung cancers and EGFR in female never smokers with lung adenocarcinomas, has been found (Grigoriu et al. 2015, Olak and Coulson 2004). These discoveries, along with a higher incidence of certain histologic subtypes of lung cancers in one gender over the other, and ongoing current research into proposed theories for why these differences exist must be kept in mind when analysing for differences in healthcare utilisation among men and women with lung cancer.

Included studies in this systematic review measured seven categories of treatment use: diagnosis interval and stage of diagnosis, physician visits, emergency presentation, hospital admission and length of stays, surgical resection, chemotherapy and radiation therapy. The preliminary findings show that all the included articles have employed data from developed countries and the study populations were male dominated. However, the objectives, population size, types and stages of lung cancer, measures of treatment used, statistical method and study design were heterogeneous in the selected studies. Surgery, chemotherapy and radiation therapy dominated the types of healthcare use investigated. Some contradictions in the types of treatment used have been found in the selected studies. These contradictions may have arisen owing to heterogeneity in the characteristics of the study design, study populations and healthcare systems and policies of the studied countries.

In relation to diagnosis interval and stage of cancer diagnosis, studies suggest that women tend to have longer diagnostic intervals; nonetheless, they often get diagnosed with lung cancer at an earlier stage. At first glance, these findings seem contradictory. These results should be interpreted based on gender differences in awareness, detection and reporting of symptoms and the willingness to take appropriate healthcare actions. Previous studies have reported that women are more incisive in reporting symptoms (Warner and Procaccino 2007), receiving health information (Khakbazan et al. 2014) and seeking cancer-related information (Manierre 2015), compared to men. Henceforth, if women are identifying symptoms earlier than men, a couple of things seem likely. In early stages, health practitioners may not accurately judge symptoms as critical (Marshall et al. 1982) especially since, occupational and nicotine exposure risk factors of lung cancer are more easily attributed to men than women. Furthermore, doctors perhaps may take a wait and see approach (symptom normalization) (Brindle et al. 2012) before prescribing screening tests. This may result in a longer diagnosis interval. On the other hand, the ‘masculinity effect’ or a reductionist approach by males to their healthcare needs may lead them to ignore and/or remain silent about their symptoms (Galdas et al. 2005). There are also the possibilities that men have fewer available hours to have regular contact with health practitioners owing to full-time and longer hours of work (Brittle and Bird 2007). These delays result in men with lung cancer symptoms triggering an immediate screening test. This may explain why men are seen to have shorter diagnosis intervals but are more often diagnosed with an advanced stage of lung cancer, compared with women.

Several articles reported emergency department visits of lung cancer patients and the results were diverse. Three papers indicated that women were more likely to visit the emergency department; however, one of these papers (Abel et al. 2015) concluded that the odds of emergency presentation of women reduced significantly after a case-mix adjustment. Another study found that men are more likely to visit the emergency department. The evidence concerning lung cancer diagnosis through an emergency department presentation is not conclusive. These studies did not contain a suitable explanation for the reported gender difference in emergency department visits. Previous studies showed that decreased physical functioning (Kurtz et al. 2006), patients with a prior history of emergency visits (before cancer diagnosis) (Sikka and Ornato 2010), old age (Mitchell et al. 2015), experience of pain (chest and abdominal) and respiratory symptoms (nausea and vomiting, shortness of breath) (Gorham et al. 2013; Hsu et al. 2018) were strongly associated with emergency department presentations by cancer patients. Furthermore, it is unclear from the included studies how much of emergency department visits and ER diagnosis of lung cancer was driven by inadequate access to primary health care facilities or appropriate prospective screening in those with symptoms and histories of risk factor exposure, which may inherently differ between men and women in certain geographic locations. Future studies are needed to investigate the demographic risk factors of emergency department presentation for lung cancer patients during the active and continuing treatment period.

Contradictory results were found for the likelihood of hospital admission, readmission after surgery and length of stay by gender. The findings indicate that women had a higher probability of using inpatient cancer-care services (n = 2). Conversely, men had higher risks of readmission after surgery (n = 2) and longer length of stay (n = 2). Several papers, however, found that the hospitalisation rate (n = 2) and hospital length of stay (n = 1) were similar for both genders. Henceforth, it is difficult to determine if there are heterogeneities in the use of hospital care for lung cancer patients by gender. The differences in findings related to hospital admission may have arisen from other demographic or socioeconomic factors. For example, men’s higher likelihood of readmission after surgery could be associated with factors such as (1) advanced-stage diagnosis and (2) being a more suitable candidate for pneumonectomy, due to the locally advanced nature of their cancer, compared to segmentectomy which is less radical. Previous studies have identified these two factors as primary risk factors for readmission after surgery among lung cancer patients (Handy et al. 2001; Ogawa et al. 2015). The re-hospitalisation rate of lung cancer patients might also be influenced by other comorbidities (Tammemagi et al. 2004), which partly explains the variation in the findings.

This study has identified significant gender differences in the probability of receiving surgical treatment for lung cancer patients. Approximately half of the studies that reported on surgery as the treatment modality indicated women as the probable nominee for the treatment. Several explanations are possible in support of this outcome. Sitas et al. (2014) suggested that men with lung cancer are less likely to undergo surgery, because they present at the later stages of cancer and with more comorbidities, where surgical resection is no longer the recommended mode of treatment. In addition, differences in smoking habits is another likely cause of differential cancer treatment between men and women. Squamous cell carcinoma (a type of cancer tumour) which is more common in heavy smokers (often men) is challenging to detect at early stages through chest radiography (de Perrot et al. 2000). Conversely, female lung cancer patients have a higher proportion of adenocarcinomas which is commonly located in the periphery of the lung, and therefore, surgical resection, such as segmentectomy (if the disease is localised) is comparatively more frequent (Currow et al. 2014; de Perrot et al. 2000).

A substantial number of the studies (75%) in this review concluded there were no significant gender differences in the odds of receiving chemotherapy and radiation therapy. However, a few of the studies suggested that men are more likely than women to undergo this treatment. These types of treatments are more generic at the advanced stage of lung cancer and with tumours that manifest in an immediate threat to survival (Ramsey et al. 2004). Consequently, if more women are being diagnosed at an early stage, it is natural to find a lower proportion of women receiving chemo and radiation therapy. Furthermore, there is a growing body of evidence that men and women experience different physiological reactions after chemotherapy. Women often show severe adverse effects (e.g. nausea and vomiting, cardio and neuro toxicities) which has a potential impact on their decision to receive or continue chemotherapy treatment (Benchetrit et al. 2019; Wang and Huang 2007).

This systematic literature review has some limitations. It was not possible to perform a meta-analysis with data from the selected studies due to substantial heterogeneity in the study design, statistical estimation and categories of treatment measured. Additionally, the included studies did not provide enough data to allow consistent correlation of surgery, chemotherapy and radiotherapy treatment options offered with lung cancer disease stage, hence introducing an added layer of heterogeneity in the analysis of healthcare utilisation behaviours. Furthermore, no grey literature or articles (all types including peer reviewed) published in a non-English language, have been included, and studies with null findings are less likely to be published in peer-reviewed journals. Henceforth, as with many reviews, it is not possible to avoid the likelihood of publication bias. Despite the thorough and accepted approach (e.g. back referencing), it difficult to judge if all potential studies have been included in the review. Finally, no randomised controlled trial study was identified, which is the ‘gold standard’ of research evidence.

Gaps in the evidence and future research

A key section of gender-based differences in lung cancer treatment utilisation that remains to be studied is that of targeted and immune therapies. These agents are novel technologies and their full efficacy profile is yet to be exploited. We are already aware of certain immunotherapies, such as EGFR inhibitors, which are more efficacious in certain populations of females with the known molecular profile. It remains to be seen in future studies whether there are distinct gender-based immune therapy utilisation differences beyond that simply driven by the biological molecular targets.

Despite an extensive search of the literature, no peer-reviewed published studies were identified that had a study setting of a least-developed and developing country (e.g. South and South-east Asia, Africa, South America) though nearly 58% of lung cancer incidences (2012) occur in the least-developed countries (Ferlay et al. 2015). These countries suffer from lack of adequate medical resources, lung cancer awareness and experience high gender-specific health inequalities. Hence, it can only be speculated that there is a greater magnitude of gender-specific inequalities in utilisation (women using healthcare substantially less than men) of cancer treatments in these settings. Future research to address this gap in the evidence is needed. Policymakers in these countries need to develop and ensure the availability of appropriate databases to enable credible research to be conducted. Furthermore, confounding evidence was found that men are diagnosed at later and more advanced stages of lung cancer. After establishing the causes of this phenomenon, appropriate interventions (if possible, through a randomised controlled trial) should be developed to address this problem. This could substantially reduce the cost of treatment and improve the survivorship of lung cancer patients. Further investigation is required to understand how general practitioners evaluate the symptoms of lung cancer for men and women. Also noticeable is the lack of an adequate explanation in these past studies regarding the discrepancies in lung cancer patients emergency department visits, hospital admissions and hospital night stays, by gender. It is still unclear why studies in different settings found opposing results. The question remains: what contributes to this heterogeneity? Is it patient preferences or medical practitioner’s behaviour and attitudes towards processing the treatment needs of a particular subgroup or cohort? This warrants the collection of in-depth and specific data so that further research could help to understand what drives care-seeking behaviour of lung cancer patients. Is it purely biological factors (e.g. genetic, hormonal) or social (Bird and Rieker 1999), purely patient preference, or is it the health system and practices? Future research should focus on understanding these questions.

Conclusions

This systematic review contributes to understanding the gender-based differences in care-seeking behaviour of lung cancer patients. Substantial evidence of heterogeneity in the seven categories of treatment was found. Nonetheless, some clear trends are apparent from this review of the literature. Men appear to be diagnosed with more advanced-stage lung cancer compared to women. Possible explanations for this may include a reductionist approach to self-health concerns, a higher incidence of exposure to risk factors such as nicotine smoking and subsequent higher incidence of certain types of lung cancers in men than women. Other possibilities including that of a differing preference of utilisation of screening and primary health care options for those with family histories or early symptoms, compared to women. This impacts upon the availability of cancer treatment for men, including a lower candidacy level for surgical resection due to the more advanced nature of the malignancy, which may also explain higher rates of unexpected hospital readmissions after surgery because of the complex nature of the surgery and a higher propensity to receive radiotherapy (offered during advanced malignancy). These findings from this systematic review raise significant concern that inherent gender-based lung cancer-care-seeking and utilisation behaviours in men may be significant contributors to their lung cancer-related morbidity and mortality. The gender-specific inequality in lung cancer survivorship may be explained in part by men’s tendency towards inaction for ill health-related symptoms and a likely underutilisation of evidence-based therapies for treatment.

Within the limitations of the included studies in this systematic review (male-dominated populations), there appear some general trends in women’s care-seeking behaviour. The higher rate of emergency department lung cancer diagnosis and longer length of inpatient hospital stay may indicate that women are taking earlier initiatives to address their health concerns in an urgent and holistic fashion. This trend notes a lower rate of women being diagnosed with advanced-stage lung cancer than men and higher uptake of proposed therapies such as surgery or chemotherapy. However, conversely, a higher rate of ER presentation may signify less access to screening or prospective planned healthcare assessments, and a longer inpatient hospitalisation duration may signify a higher rate of side effects to therapy or disease complications, which were not clarified in the included studies. Furthermore, the included studies did not stratify the disease characteristics in detail, hence, some of the contradictory results are expected. Therefore, future studies can look to elucidate these trends further to guide better understanding of gender-specific differences to drive positive changes in healthcare utilisation behaviour for the broader population.

Lung cancer treatment is a rapidly evolving area of oncology, with changing treatment paradigms and with the latest evidence-based clinical practises continuing to challenge care providers and patients alike. Ongoing research into gender-specific differences in treatment-seeking and healthcare utilisation of lung cancer patient (in developed and developing countries) will prove highly beneficial in establishing a guide for policymakers and clinical care providers to optimise healthcare management by promoting care utilisation amongst all lung cancer patients.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Funding

No funding was received by the authors for this research study.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflict of interest.