Can we find the missing men in clinics? Clinic attendance by sex and HIV status in rural South Africa

Background: HIV-negative men are over-represented in tuberculosis (TB) prevalence surveys including the first South African national TB prevalence survey in 2018. Traditionally, TB screening is focused in clinics. We aimed to determine the frequency of primary healthcare clinic (PHC) attendance among HIV-negative men in a TB-prevalent setting. Methods: Since January 2017, PHC attendees in a rural South African demographic surveillance area (DSA) were asked their reason for attendance. HIV status was defined as positive if tested positive in a DSA sero-survey or attended clinic for HIV care; negative if tested negative between January 2014—December 2017 and no HIV-related visits; and HIV-unknown otherwise. Results: Among 67124 DSA residents (≥15 years), 27038 (40.3%) were men; 14196 (21.2%) were classified HIV-positive, 18892 (28.1%) HIV-negative and 34036 (50.7%) HIV-unknown. Between April 2017 and March 2018, 24382/67124 (36.3%, 95% confidence interval [CI] 36.0–36.7) adults made ≥1 PHC visit, comprising 9805/40086 (24.5%, 95%CI 23.6–25.3) of HIV-negative or unknown women and 3440/27038 (12.7%, 95%CI 11.6–13.8) of HIV-negative or unknown men. Overall, HIV care accounted for 37556/88109 (42.6%) of adult PHC visits. Conclusion: In this rural population, HIV-negative and -unknown men rarely attend PHCs. Improving TB screening in clinics may not reach a key population with respect to undiagnosed TB. Additional strategies are needed to diagnose and treat TB earlier.


Introduction
Based on data from 2013, South Africa was estimated to have 160000 "missing" people with tuberculosis (TB), that is individuals with active TB disease who were not on treatment 1 , who may contribute to continuing transmission.South Africa is committed to finding the missing 160000 1 ; however, how best to do this is uncertain.
The World Health Organization (WHO) has traditionally recommended active case finding for TB among individuals attending health facilities 2 .However, this approach will only reduce transmission if people with undiagnosed, infectious TB are identified at health facilities and start appropriate treatment early, and therefore reduce their duration of infectiousness.This is particularly a concern for men, who are generally perceived to attend health care facilities less often than women 3,4 .Modelling work has suggested that the delay from symptom onset to treatment initiation in men is a year longer than women 5 .In South Africa in particular, TB accounts for a large proportion of the difference in life expectancy between men and women 6 .In 2018, the male: female TB notification ratio was 1.7 globally and for the period of 1970-2016, the age standardised TB incidence rate for men was nearly 1.8 times that of women (154.4 vs 86.3 per 100 000 people).The TB mortality rate for men was twice as high as women (21.9 vs 10.8 per 100 000) 7 .
National surveys in Tanzania, Rwanda, Zambia and Kenya have found that TB prevalence was significantly higher in men than women [8][9][10] .In addition, in Zambia and Kenya (where the prevalence surveys offered HIV testing as well as TB screening) over 80% people with undiagnosed active TB were HIV-negative or HIV-unknown.In South Africa the first national TB prevalence survey took place in 2018.Among the survey participants who reported at least one TB symptom, more men (71.3%) than women (63.4%) did not seek care 11 .TB prevalence was higher in men, with a prevalence almost 1.6 times that of women 11 .77% of participants who screened positive for TB were HIV negative or unknown 11 .It is therefore likely that a priority group among the "missing 160000" are men with negative or unknown HIV status 11 .
The aim of this study was to describe the frequency of primary healthcare clinic (PHC) attendance by sex and HIV status in 11 clinics in rural KwaZulu-Natal, South Africa.

Study area and population
The study was conducted in the Africa Health Research Institute (AHRI) demographic surveillance area (DSA), in uMkhanyakude district, KwaZulu-Natal, South Africa.The AHRI DSA covers 845km 2 , with approximately 25000 homesteads and over 60000 residents aged 15 years or above.AHRI has undertaken population-based demographic surveillance since 2000 with annual HIV sero-surveys since 2003.This district had an annual notification rate of all TB cases of 394 per 100,000 population in 2018 and 64.4% of those notified for TB were HIV positive 12 .In addition, in 2017-8, 9.9% of Xpert MTB/RIF results from uMkhanyakude showed rifampicin resistance, among the highest prevalence in the country 12 .
The population of interest in our analysis were all individuals over the age of 15 years who were a resident member (defined as intending to spend the majority of nights at a household within the study area) of a household in the AHRI DSA on 1st July 2017.

Data collection
Since January 2017, individuals who sought health care at any one of the 11 PHC's serving the AHRI DSA on weekdays between 7am and 7pm have been registered by a member of AHRI staff, and their self-reported reason for attending clinic recorded, using an electronic system known as ClinicLink 13 .For this analysis, we used ClinicLink data to determine the number of PHC visits made by AHRI DSA residents between 1 April 2017 and 31 March 2018.Visit data from AHRI DSA residents were retrospectively linked to their demographic surveillance data and HIV sero-survey data.

Case definitions
Participants were considered to be HIV negative if they tested negative in a sero-survey between 1st January 2014 and 31st December 2017 and had no HIV-related PHC visits recorded in ClinicLink; HIV-positive if they tested HIV-positive in a DSA sero-survey or had an HIV-related visit recorded in ClinicLink; or HIV-unknown otherwise.

Statistical analysis
The main outcome for this analysis was the proportion of DSA residents visiting PHCs between April 2017 and March 2018, stratified by sex, age and HIV status.
We categorized PHC visits into three subgroups: 1. HIV visits, including antiretroviral therapy (ART) start or follow-up; 2. acute visits, including family planning, minor ailments, maternity, reproductive health, circumcision, or emergency care; and 3. 2 Participants were considered to be HIV-positive if they tested HIV-positive in a demographic surveillance area sero-survey or had an HIV-related visit recorded.
3 Participants were considered to be HIV negative if they tested negative in a sero-survey between 1st January 2014 and 31st December 2017 and had no HIV-related PHC visits recorded.
4 Participants were considered to be HIV unknown if neither 2 nor 3 was true.
other chronic (non-HIV), including care for TB, diabetes or hypertension. All

Discussion
By nesting this study within a demographic surveillance area, we have been able, for the first time, to quantify PHC visits based on a population denominator and stratified by HIV status.As anticipated, we found that men who are HIV-negative or with unknown HIV status rarely visit PHCs for any reason.Younger men with an unknown HIV status were the least likely to attend clinic.Reducing TB transmission requires that people with active TB are identified and start effective treatment early, in order to reduce their duration of infectiousness.The traditional approach of "passive case finding" depends on people who are symptomatic with active TB seeking care in clinics and being successfully diagnosed and treated 15 .The rationale for "intensified case finding" among clinic attendees is that it is far more efficient for health workers to screen 2 n represents the number of visits.
3 The percentages represent n/N, where N is the total visits made by all adults during study period.for TB among people attending clinics than in the general population 2 .However, if people with TB have relatively mild or intermittent symptoms 16 , particularly if they are HIV-negative or unaware of their status, they may not prioritise seeking care.Our data show that HIV-negative or -unknown men rarely visit PHCs, and therefore attempts to reduce the duration of infectiousness of HIV-negative men with active TB will need to reach outside health facilities.These data illustrate that HIV care has become a dominant reason for PHC attendance among adults in this setting of very high HIV prevalence, accounting for nearly half of all daytime visits by adults.This finding is unlikely to be generalisable to settings where HIV prevalence is less high.However, as we move towards 90:90:90 and then 95:95:95 HIV care cascade targets, and HIV prevalence increases as ART prolongs survival, PHCs in many settings will need to provide care for increasing numbers of people on ART.Our data underline the importance of efforts to simplify HIV care by reducing visit frequency and strengthening systems to support further decentralisation of ART delivery to community level 17 .
Several studies both in South Africa and the wider African region have quantified reasons for visits to health facilities, but we are not aware of similar studies based in demographic surveillance areas or other settings with a population denominator and comprehensive data on HIV status 18 .
A study in the Western Cape in 2016 used home-based surveys to determine if respondents had attended a PHC facility in the past six months.Similar to our finding, they found that men were less likely to attend clinic that women 19 .Furthermore, fewer than 20% of all men aged 18-25 years, or men aged 26-45 who attended bars, attended a clinic 19 .
A study in rural Tanzania in 2019 examined the quality of outpatient care of 2002 women seen in PHCs by asking women about their most recent outpatient experience in a household survey in 2002 20 .The study found that the most common reasons for seeking care were fever or malaria (33.9%), vaccination of infants (33.6%) and non-emergency check-ups (13.4%) 20 .This difference is expected given that malaria is rare in South Africa 20 with 9478 malaria cases and 76 malaria deaths in 2017 21 , whereas Tanzania had 4241364 malaria cases and 6313 deaths in 2015 22 .
In South Africa, a study in 2010 looked specifically at chronic non-communicable diseases in PHC facilities in the Western Cape, North West, Northern Cape and Limpopo provinces in South Africa.Health care workers recorded the age and gender of each patient and the reasons for each encounter.They found that hypertension was the commonest non-communicable disease (NCD) diagnosis encountered (13.1%), followed by type 2 diabetes (3.9%) 23 .Only 1.1% of respondents with an NCD also had HIV by self-report 23 .Women accounted for 12526 (66.6%) of consultations and men for 6288 (33.4%) 23 .However, this study was focused on multimorbidity and did not include HIV care as an independent reason for clinic attendance.Furthermore, it did not consider acute care as a primary reason for consultation.
In the past 30 years, the number of individuals receiving care for HIV in clinics has grown significantly with improved and comprehensive ART treatment services 24 .However in South Africa, men feel alienated by clinics; view illness as weakness which may result in loss of masculinity, and fear being diagnosed with HIV 5 .Common concerns and barriers among men with TB accessing clinic services, many of whom were informal workers who lost their job due to their symptoms or TB-associated stigma, included community gossip, and clinic operations and environments not conducive to men's needs 25 .Men with TB experience challenges accessing healthcare unless the illness is severe, due to cultural norms, perceived male-unfriendliness of clinics, fear of ridicule from peers, women and clinic staff and fear of mistreatment by HCW 5,26,27 .Men's prioritisation of work over healthcare delays seeking and engaging in care 28 .Given our finding that HIV-negative men rarely visit PHCs, improved coverage of TB screening in PHCs is likely to be inadequate to achieve earlier TB diagnosis and treatment among this important group.Additional research is needed to determine how to promote early diagnosis of TB among men.Facilitators of men's access and utilisation of PHC clinics including improved models of male-centred care, and novel approaches need to be explored 29 .Barriers that prevent men testing and accessing health services, including stigma, need to be better understood 19,29 .Novel approaches are necessary for men to access TB screening services 29 ; these could include mobile services, flexible clinic opening times, shorter waiting times and male-friendly services which may need evaluation or simply implementation 5 .
Public health has traditionally given more attention to mother and child health but innovative strategies are needed to access this consistently underserved key population, interrupt TB transmission and find the "missing 160000".
Limitations of this analysis include that clinic attendance was only captured in the eleven PHCs serving the AHRI DSA, therefore any visits by residents in our study population to more distant clinics, hospitals and private institutions would have been missed.The number of PHC visits we recorded should therefore be regarded as a minimum estimate.In addition, no data capture occurred at night and over weekends, which means that some acutely ill people, and those seeking care for trauma will have been missed, particularly at the one PHC which is open 24 hours a day.This will have resulted in nonrandom missingness of emergency visits.We were not able to explore his further because of the lack of data.
Some misclassification of HIV status is likely.Participants were considered to be HIV negative if they tested negative between the 1 st of January 2014 and the 31st December 2017 and had no evidence of HIV-positive status.This may have resulted in misclassification of HIV-negative people as HIVunknown.A participant was considered to be HIV-positive if the last HIV test recorded by AHRI's DSA was positive or if they have had any HIV-related PHC visit in the DSA.Therefore, anyone who tested HIV positive outside the DSA and did not attend for HIV care in the eleven DSA clinics or References became HIV-positive after testing negative will have been misclassified as HIV-negative or -unknown 30 .

Conclusion
In rural South Africa in 2017-18, only 15% HIV-negative or unknown men attended a PHC for any reason.Therefore, improving TB screening in clinics as an isolated strategy will not reach a key population with respect to undiagnosed TB.Additional strategies are needed to diagnose and treat TB earlier, particularly in men: some should be implemented based on existing evidence, and others need well-designed trials.
Thank you for this manuscript that casts a needed spotlight on a population group that requires attention in efforts to address the TB burden.The manuscript is well written.
Did the authors analyze the data by age group?Stratification by age will also be helpful since it likely affects the decision and ability to attend PHC facilities:-for example the SA TB prevalence survey found the largest prevalence to notification ratio in younger people (15-24 years) and older people 65 years+.This may also be true by sex and could direct targeting and differentiation of interventions for different segments of the male population.
Did the authors consider separating the HIV-ve and HIV status-unknown groups?Those whose status is unknown group could be different from those LWH and those HIV-ve since they have never accessed testing and could require further and different attention and interventions -did they even attend PHCs at all even for chronic conditions?
Please add more literature including some of the more recent literature that explores men's views on TB and healthcare facility attendance.It is conventional wisdom to assert that men do not seek health care at the same frequency as women but the evidence behind the claim has always been weak which is why this is an important and cleverly designed study that will be widely cited.Researchers working on gender and TB have been searching for just this parameter for their models, so this study fills an unmet need.Taking place in a HDSS, the investigators obtained hard-to-find denominators for their estimates and stratified by HIV status, a key predictor of health seeking in many settings.It is great to have an estimate of the magnitude of the difference in clinic attendance both for the modellers but also for the funders of different kinds of active case finding efforts.
This methodology should be emulated in other settings and WHO should recommend this approach to estimation of health seeking as the starting point for all national TB care cascade and patient pathway studies instead of the flawed proxy currently recommended (i.e.use of the entirely hypothetical child fever health seeking question from demographic and health surveys.The use of health seeking for child fever is invalid for TB symptom health seeking behavior and obscures the important gender differences shown by the authors.The authors are giving us a great new way to work in a gender disaggregated way and it would be handy if they provided an SOP to help other DHSS with the linking aspect.
To be thorough, I offer some critiques: The authors assert that because HIV-negative men are less likely to seek formal public health care during day time hours, that non-passive case finding may be needed to reach this population with TB screening services.However, they do not mention exactly what that would look like.I infer that they are well aware that actually, men are also less likely to participate in generic community TB screening activities (and the South African TB prevalence survey is an example).Therefore the authors were strategically vague as to possible solutions to this challenge. 1.
In the abstract the proportion of HIV-negative men seeking any care was (12.7%, 95%CI 2.
In conclusion, this is a methodologically important paper and I hope it is widely replicated.Reviewer Expertise: TB epidemiology, stigma and gender I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

References 1 .
Chikovore J, Pai M, Horton KC, Daftary A, et al.: Missing men with tuberculosis: the need to address structural influences and implement targeted and multidimensional interventions.BMJ Glob Health.5 (5).PubMed Abstract | Publisher Full Text 2. Daniels J, Medina-Marino A, Glockner K, Grew E, et al.: Masculinity, resources, and retention in care: South African men's behaviors and experiences while engaged in TB care and treatment.Soc Sci Med.270: 113639 PubMed Abstract | Publisher Full Text 3. Gibbs A, Sikweyiya Y, Jewkes R: 'Men value their dignity': securing respect and identity construction in urban informal settlements in South Africa.Glob Health Action.2014; 7: 23676 PubMed Abstract | Publisher Full Text 4. Van Heerden A, Msweli S, Van Rooyen H: "Men don't want things to be seen or known about them": A mixed-methods study to locate men in a home based counselling and testing programme in KwaZulu-Natal, South Africa.African Journal of AIDS Research.2015; 14 (4): 353-359 Publisher Full Text 5. Reihling H: Affective Health and Masculinities in South Africa.Routledge.2020.Reference Source Is the work clearly and accurately presented and does it cite the current literature?Yes Is the study design appropriate and is the work technically sound?Yes Are sufficient details of methods and analysis provided to allow replication by others?Yes If applicable, is the statistical analysis and its interpretation appropriate?Yes Are all the source data underlying the results available to ensure full reproducibility?Yes Are the conclusions drawn adequately supported by the results?Yes Competing Interests: No competing interests were disclosed.

Table 2
3504 (73.7%, 95%CI 72.2-75.1)ofHIV positive men visited a PHC for any reason at least once during the one-year period.In contrast, among the HIV-negative adult residents, 21.8% (95% CI 20.9-22.8) of men as opposed to 45.6% (95% CI 44.7-46.5) of women visited a PHC at least once during the one-year period of the study.
shows the proportion of adults making one or more PHC visits (for any reason) during the study year, by sex and HIV status.Among all 67124 residents, 36.3% (95% confidence interval [CI] 36.0-36.7)visited one of the eleven PHCs serving the AHRI DSA during the study period.The median number of PHC visits (among the 24382 residents who visited a PHC) was two (range: 1-26) per person among men and three (range: 1-22) visits per person among women.Of the 67124 adult DSA residents, among HIV-positive individuals, 8554/10692 (80%, 95%CI 79.3-80.8) of HIV-positive women and 2593/

Table 3 . Number and type of primary healthcare visits made by adults resident in the Africa Health Research Institute demographic surveillance area 1 during the study year by sex (N=88109).
1 Africa Health Research Institute (AHRI) undertakes population-based demographic surveillance in a rural district of KwaZulu Natal, South Africa with annual HIV sero-surveys.Individuals were included if they were resident in the AHRI demographic surveillance area on 1st July 2017.

the work clearly and accurately presented and does it cite the current literature? Partly Is the study design appropriate and is the work technically sound? Yes Are sufficient details of methods and analysis provided to allow replication by others? Yes If applicable, is the statistical analysis and its interpretation appropriate? Yes Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes
Chikovore J et al have published extensively on this subject (see reference list for example) 1 .Issues around male attendance at health facilities are also documented in the SA CRG report: South African Community Rights and Gender Assessment: Exploring the impact of gender, key population membership, and the legal environment on TB vulnerability, treatment access, and quality of care 2 .
Ellen M. H. MitchellDepartment of Public Health, Institute for Tropical Medicine, Antwerp, Belgium