SEARCH

SEARCH BY CITATION

Keywords:

  • Cessation;
  • inequalities;
  • prospective cohort study;
  • smokeless tobacco;
  • South Asians

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Aim

To evaluate smokeless tobacco cessation in communities of South Asian origin.

Design

Multi-centre prospective cohort study.

Setting

Three tobacco cessation services offering specialist smokeless tobacco cessation outreach clinic support to South Asians (Bangladeshi, Indian and Pakistani) resident in England.

Participants

A total of 239 South Asian participants seeking to stop smokeless tobacco use between November 2010 and December 2011.

Measurements

Socio-demographics, tobacco use and dependence, self-reported abstinence at 4 weeks and satisfaction measures.

Findings

Participants' mean age was 45 [standard deviation (SD) = 13] years, were predominantly female (76%), of Bangladeshi origin (74%), either home carers (53%) or not working (29%). Sixty-three per cent were recruited from community locations, 21% through a clinical contact and 16% through friends and family. Mean daily number of smokeless tobacco intakes was 10 (SD = 7) and the mean dependence score was 4.5 (SD = 1.9). Sixty-three per cent of participants achieved continuous abstinence 4 weeks after quitting. Using nicotine replacement therapy (NRT) versus not using it [OR = 3.47, 95% confidence interval (CI): 1.25, 9.62] and below median (≤8) daily smokeless tobacco intakes (OR = 1.91, 95% CI: 1.07, 3.40) predicted successful abstinence.

Conclusion

South Asian smokeless tobacco users resident in England accessing services to help them stop appear to have short-term success rates comparable with smokers attending stop-smoking services, with higher success rates being reported by those using nicotine replacement therapy.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Smokeless tobacco is defined as any product containing tobacco that is placed in the mouth or nose and not burned [1]. The smokeless tobacco products used most often in England are the paan quid with tobacco (PQT), gutkha and khaini [2]. These products contain the core ingredients of tobacco leaf or processed tobacco (zarda), slaked lime as an alkalinity regulator and areca nut, which has itself been classified by the International Agency for Research on Cancer (IARC) as carcinogenic [3].

England is a member of the European Union (EU) and is required to follow its tobacco regulation requirements [4]. These regulations distinguish between ‘tobacco products’ (whether smoked or chewed) and ‘tobacco for oral use’, which is sucked. Swedish snus is considered a tobacco ‘for oral use’ and is sold only in Sweden and other non-EU Scandinavian countries. As a tobacco ‘product’ the use of smokeless tobacco in England is allowed, although largely restricted to South Asians from the Indian, Pakistani and especially Bangladeshi communities, which together make up 4.5% of the population [5]. More members of the UK resident Pakistani and Bangladeshi communities suffer poor health than the general population, as almost two-thirds of these communities live in low-income households [6].

Smokeless tobacco products used by these communities in England are widely available, despite not meeting the current EU regulatory requirements for tobacco products [4], most particularly with respect to ingredient disclosure, pack identification codes and health warnings. It is widely recognized that these smokeless tobacco products have adverse effects on health with a specific increased risk of oral cancer, especially among South Asian women [7], and tobacco dependency [8]. Oral pain is also commonly reported following a cessation attempt, suggesting that chewing smokeless tobacco acts to mask symptoms of oral disease [9].

Current English practice guidelines suggest that relevant health professionals, particularly dentists, should ask about smokeless tobacco use in their South Asian patients, update the patient's notes and advise of potential health risks. Dentists should examine smokeless tobacco users' mouths for any potentially malignant change. Specialist smoking cessation services should provide counselling for smokeless tobacco use where there is a demand [10]. English National Health Service (NHS) Stop Smoking Service (SSS) guidance recommends the use of behavioural support (BS) alone in helping smokeless tobacco users to quit, excluding nicotine replacement therapy (NRT) use. Providing any NHS SSS quit support for smokeless tobacco users has been a discretionary activity and outcomes are excluded from target monitoring [11].

There is no evidence that pharmacotherapy assists cessation of South Asian smokeless tobacco. A Cochrane Review [12] reported eight trials of NRT used as an aid to smokeless tobacco cessation which were of varying effectiveness in assisting successful cessation [odds ratio (OR) = 1.14, 95% confidence interval (CI): 0.91, 1.42]. In addition, none of these studies examined NRT use in English South Asian communities. Behavioural interventions with an oral examination or telephone follow-up were therefore recommended. However, the included trials reported a low prevalence of withdrawal symptoms and NRT-related adverse events. NRT patch and gum use was observed to reduce withdrawal symptoms both in range and intensity compared to placebo. NRT-related adverse events were reported as minimal, more likely to be heartburn or gastroesophageal reflux if lozenges were used [13-19].

A higher proportion of the general population have quit smoking successfully compared to the proportion of English resident Bangladeshi women who have quit smokeless tobacco use successfully [20]. There are two reports of a specialist outreach service [the Bangladeshi Stop Tobacco Project (BSTP)] using NRT to support smokeless tobacco cessation activity with Bangladeshi women resident in Tower Hamlets, England [21, 22]. The most recent of these reports demonstrated that those using NRT in addition to BS to support their quit attempt were nearly five times more likely to report continuous abstinence over 4 weeks. Participants recruited in the community were nearly twice as likely to report continuous abstinence over 4 weeks as those recruited from health clinics. A third significant variable predicting successful self-reported cessation was a relatively superior socio-economic position. These two reports provided no data on the quitting process, such as withdrawal symptoms, adverse events or participant satisfaction. A third study has compared the outcomes of an outreach project for Pakistani and Bangladeshi male smokers with the standard NHS SSS [23]. The potential to increase cessation attempts and a small increase in the number of 4-week abstinent smokers was reported. No significant differences in participant satisfaction between either the outreach intervention or standard programme were identified.

The aim of this study was to evaluate the process and outcomes of smokeless tobacco cessation in three English NHS SSS settings with communities of South Asian origin, with particular emphasis upon client withdrawal symptoms, NRT-related adverse events and satisfaction. It was hypothesized that South Asian smokeless tobacco users would be more likely to make a successful cessation attempt if they were recruited using an outreach clinic model, had access to NRT to support their cessation attempt, had lower levels of withdrawal symptoms and NRT-related adverse events and that successful cessation would result in positive satisfaction levels. The primary outcome was self-reported continuous abstinence from smokeless tobacco use during the cessation attempt.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Study design and setting

This multi-centre prospective cohort study took place in the London borough of Tower Hamlets and the cities of Leicester and Bradford. These areas were chosen because of their large South Asian origin populations and their use of different smokeless tobacco products [5]. The most common form of smokeless tobacco use in Tower Hamlets is the paan quid with tobacco (PQT), while chewers in Bradford and Leicester were more likely to chew gutkha and khaini. One full-time cessation support worker (full- or part-time equivalents) worked within each location for a consecutive period of no more than 12 months. One male and two female workers were employed. Recruited from the communities they would serve, they spoke a wide range of community languages, including Bengali, Urdu, Hindi and Gujarati.

The training needs of the workers were assessed and appropriate training was given on the prevalence of South Asian smokeless tobacco use in England, the range of interventions and resources available to support a smokeless quit attempt and the opportunities available for implementing a specialist outreach clinic model of service delivery.

The principles proposed for accessing populations not recruited readily through mainstream, conventional settings such as primary care services were adopted [24, 25]. Workers identified key community informants who, in turn, enabled their access to informal, otherwise overlooked client recruitment locations. Venues identified included community and faith centres, shops, libraries and schools as well as local primary health-care clinics. Workers were able to gain privileged access to the community by their language skills, gender and age. In some communities participants expressed a preference for a gender-matched worker. Similar participant preferences might be expressed with respect to the age of the worker. Workers adapted their approach proactively to raise awareness of the health and social impacts of smokeless tobacco use, gaining the trust of potential participants and inviting a cessation attempt if there was interest and readiness to make such an attempt. Participants starting a cessation attempt received weekly one-to-one support for up to 12 weeks. Each session, lasting up to 30 minutes, incorporated strategies identified as supporting tobacco cessation [26]. Self-help materials and leaflets were also made available to participants at the initial visit. Access to NRT on prescription for up to 12 weeks in the cessation attempt was negotiated jointly. Nicotine replacement therapy, including gums, patches and inhalers, is widely available from pharmacists with a prescription. Nicotine replacement therapy is also available on general sale over-the-counter from pharmacists and a wide range of other retail outlets [27], but study participants were not expected to purchase their NRT. Weekly one-to-one follow-up contacts included assessment of tobacco use, identifying and addressing barriers to successful cessation and adjustment of NRT dosage.

The primary outcome was continuous self-reported abstinence from smokeless tobacco use, defined as no tobacco use at all between the quit date and a defined follow-up time [28]. This paper reports on the initial 4-week follow-up, chosen for theoretical reasons [29] and also because this is the outcome period used within the English NHS SSS [11].

Study participants

Members of the South Asian communities residing in these three settings were recruited between November 2010 and December 2011. Participants were aged 18 years and older and able to give informed consent. Those who were pregnant or with serious systemic illnesses were excluded. A power calculation based on previous outcome data [22], that 24.7% of those receiving behavioural support (BS) alone compared to 67.3% of those receiving BS and NRT would make a successful cessation attempt, was calculated assuming 90% power and alpha set at 0.05. This indicated that the minimum recruitment numbers would be 16 in the BS alone group and 63 in the BS and NRT group.

Variables

A pre-piloted data collection form incorporated routinely collected NHS SSS items [11], with additional data recording nicotine dependence, withdrawal symptoms and NRT-related adverse events. Routine items included gender, postcode, date of birth and age, ethnicity, occupational status, exemption from prescription costs, type of NRT prescribed, clinic visits and self-reported abstinence at 4 weeks. NRT-related adverse events and withdrawal symptoms were collected from all clients during the following weeks of the cessation attempt. The routine items and dependence data were collected at the start of the quit attempt, while the data on NRT-related adverse events and withdrawal symptoms were collected six times (weeks 1, 2, 3, 4, 6 and 12). A six-item nicotine dependence measure [30] (the Fagerström Test for Nicotine Dependence—Smokeless), eight-item withdrawal symptom measure [31] (the Minnesota Nicotine Withdrawal Scale) and 12-item adverse events related to NRT use measure (generated from an evidence review [32]) were used. This latter measure included one item seeking reports of oral discomfort, tooth pain and/or throat irritation.

Satisfaction with the cessation process was measured using a NHS SSS Client Satisfaction Questionnaire [33]. A subsample of participants was selected randomly in proportion to recruitment numbers in each location. Those selected were interviewed by telephone, using the appropriate community language.

Data analysis

All data, appropriately anonymized, were transferred to Queen Mary University of London (QMUL) for data cleaning and analysis. The continuous variables were grouped into categories as appropriate. According to variable distribution, responses were manipulated operationally into two or three categories for statistical analysis according to the distribution of values around the mean or median or as reported in the literature. Overall dependence scores were calculated from individual participant's responses to questions with a range of 0 (least dependent) to 10 (most dependent). Participants who were lost to follow-up were assigned ‘not quit’ status.

Descriptive statistics for individual variables were performed initially. Cross-tabulation statistics were performed comparing different categorical variables. Variables were categorized around the median value for dependence scores, daily smokeless tobacco intakes, initial withdrawal symptoms and initial NRT-related adverse events. Relative risk for cessation differences between BS alone and BS and NRT together was estimated. Binary logistic regression models were developed using backward stepwise elimination and the log-likelihood ratio test with client abstinence as the dependent variable. Those independent variables with significance values up to 0.20 (P < 0.20) [34] were entered into this modelling with the aim, after adjusting for participant age, of identifying significant predictors of first, cessation and secondly, client satisfaction. Results are reported as OR with significance levels (P-values) and 95% CI. A significance level of P = 0.05 was adopted for all analyses.

Ethics

The Consortium projects were not considered to be research by the National Research Ethics Service, but rather service evaluation and development. As such, they did not require mandatory Research Ethics Committee review. This pilot also followed the governance requirements required in individual locations. All participating participants gave full written consent, confirming not only that their quit attempt was voluntary but also their recognition that their anonymized data would be used for analysis.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Two hundred and thirty-nine participants who had completed their cessation attempt between November 2010 and December 2011 and provided demographic information and data on outcomes, withdrawal effects and NRT-related adverse events were included into the analysis. Twelve (5%) of these participants were lost to follow-up. An overview of the participant flow through the study is presented in Fig. 1.

figure

Figure 1. An overview of the client flow through the study

Download figure to PowerPoint

Demographics, tobacco use and dependence

Participants were predominantly female (76%), of Bangladeshi origin (74%), either home carers (housewives) (53%) or unemployed (29%). The mean sample age was 45 (SD = 13) years. Those in paid work outside the home were more likely to have manual occupations; 95% were exempt from prescription charges (Table 1). The mean daily number of smokeless tobacco intakes was 10 (SD = 7) and the mean dependence score was 4.5 (SD = 1.9).

Table 1. Characteristics of client sample, cessation process and outcome.
Variablen (%)
  1. GP: general practitioner; NRT: nicotine replacement therapy.

Age (n = 239)
<44 years93 (39%)
>44 years146 (61%)
Gender (n = 239)
Male58 (24%)
Female181 (76%)
Ethnicity (n = 238)
Bangladeshi176 (74%)
Indian61 (26%)
Pakistani1 (0.4%)
Occupation (n = 233)
Home carer (housewife)123 (53%)
Not working68 (29%)
Routine42 (18%)
Recruitment location (n = 237)
GP/dentist/pharmacy50 (21%)
Friends and family37 (16%)
Community150 (63%)
Exempt from prescription (n = 239)
Yes226 (95%)
No13 (5%)
Intervention (n = 239)
Behavioural support alone20 (8%)
Behavioural support + NRT219 (92%)
Self-reported quit success (n = 239)
Quit150 (63%)
Not quit89 (37%)

Tobacco cessation recruitment, NRT prescription and abstinence

Sixty-three per cent of participants were recruited from community locations, 21% through a clinical contact (general medical or dental practitioner, pharmacist) and the remaining 16% through friends and family. Female participants were more likely to be recruited in community settings (P = 0.0005). Behavioural support and NRT was used by 92% of the participants. There were no significant differences between those receiving NRT and BS compared with BS alone with respect to mean age (P = 0.190), mean daily smokeless tobacco intake (P = 0.474), gender (P = 0.532) and level of nicotine dependence (P = 0.195) (Table 1).

The mean number of contacts each participant had with their cessation adviser was 4 (range: 1–7). Of those receiving NRT, 65% received a combination of two or more products. Seventy-three per cent of these received both a patch with gum, usually a 15 mg/16-hour patch, and either 2- or 4-mg gum. Of those receiving one NRT product alone, 81% received either 2- or 4-mg gum. There were no significant differences between locations with respect to cessation success, type of intervention adopted, type of NRT used, withdrawal symptoms and NRT-related adverse events after 1 week.

At 4 weeks, 63% self-reported abstinence, 40% of whom had quit with BS alone, compared to 65% who quit with BS and NRT, a relative risk of 1.62. Four-week continuous abstinence was reported by 59%. Participants who used NRT and behavioural support were more likely to report successful abstinence (P = 0.028), although number of NRT products used (single or combination) in the cessation attempt was not significant (P = 0.244). Demographic factors were not linked significantly to abstinence (gender: P = 0.151; ethnicity: P = 0.23; age: P = 0.358; occupation: P = 0.07; prescription exemption: P = 0.620; and method of recruitment: P = 0.903). While nicotine dependence was not related to abstinence (P = 0.687), those participants with a below median (≤8) daily number of smokeless tobacco intakes were more likely to make a successful cessation attempt (P = 0.012). While fewer (≤2) initial withdrawal symptoms indicated a greater likelihood of abstinence (P = 0.0005), fewer (≤2) initial NRT-related adverse events were not related significantly to abstinence (P = 0.148).

Table 2 reports the results of the binary logistic regression model predicting abstinence. Using nicotine replacement therapy (NRT) versus not using it (OR = 3.47, 95% CI: 1.25, 9.62) and below median (≤8) daily smokeless tobacco intakes (OR = 1.91, 95% CI: 1.07, 3.40) predicted successful abstinence. These two variables explained 66–90% of the variance in the model.

Table 2. Binary logistic regression model reporting predictors of abstinence (n = 233).
Explanatory variableaUnadjusted OR (95% CI)P-valueAdjusted OR (95% CI)P-value
  1. OR: odds ratio, 95% CI: 95% confidence interval; NRT: nicotine replacement therapy; ST: smokeless tobacco. Variance explained in final model (Cox & Snell R2–Nagelkerke R2): 66–90%. aAdjusting for age.

Gender
Male1 1 
Female0.63 (0.33–1.18)0.1510.96 (0.39–2.36)0.944
Occupation
Home carer (housewife)10.07910.136
Never worked/retired0.94 (0.51–1.72)0.8570.82 (0.41–1.66)0.598
Routine2.42 (1.06–5.51)0.0342.23 (0.75–6.61)0.145
Number of ST/day
>8 ST per day1 1 
<8 ST per day1.97 (1.16–3.37)0.0121.91 (1.07–3.40)0.027
Intervention
Behavioural support alone1 1 
Behavioural support + NRT2.76 (1.08–7.05)0.0283.47 (1.25–9.62)0.017

Withdrawal symptoms and NRT-related adverse events

Initial withdrawal symptoms were reported by 78% of participants and NRT-related adverse events were reported by 74% of participants. The mean numbers of initial withdrawal symptoms and NRT-related adverse events were 2.3 (SD = 2.1) and 1.9 (SD = 1.8), respectively. The most common initial withdrawal symptom was ‘craving’, reported by 64%, while ‘headache’ was the most common NRT-related adverse event, reported by 35%. Participants who were less dependent were less likely to report initial withdrawal symptoms (P = 0.037), while participants receiving a single NRT product were less likely to report NRT-related adverse events compared to participants receiving a combination of two NRT products (P = 0.003). Both mean withdrawal symptoms and NRT-related adverse events reduced significantly in the following 2 weeks, withdrawal symptoms from 2.3 to 1.6 (P = 0.0005) and NRT-related adverse events from 1.9 to 1.0 (P = 0.0005). In the continuously abstinent group the mean number of withdrawal symptoms reduced from 2.7 (SD = 2.41) to 0.7 (SD = 0.64) 2 weeks later (P = 0.002). Oral pain (discomfort, pain and/or irritation) after 1 week was reported by 37% of participants, which reduced to 22% in the following 2 weeks (P = 0.0005). Participants initially reporting oral pain were more likely to be aged more than 44 years (P = 0.008).

Participant satisfaction with the quit process

One hundred and four participants (42 from Tower Hamlets, 42 from Leicester and 20 from Bradford) reported patient satisfaction data. Participants were either ‘satisfied’ (40.5%) or ‘very satisfied’ (59.5%) with the support they had received for their cessation attempt and 98% would recommend the service to other smokeless tobacco users. The service was reported as ‘easy to contact’, with acceptable appointment waiting times and ‘convenient’ venue by 98% of respondents. Forty per cent of these respondents had stopped smokeless tobacco use, while 60% were unsuccessful. Male rather than female participants were more likely (OR = 4.90, 95% CI: 1.24, 19.34) and those reporting a successful self-reported cessation attempt were more than twice as likely (OR = 2.72, 95% CI: 1.02, 7.22) to report higher satisfaction levels.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This multi-centre prospective study, using an adequately powered sample, investigated the process and outcomes of a specialist outreach smokeless tobacco cessation service for participants of South Asian origin using a range of validated data collection instruments and measures. Sixty-three per cent of participants reported a successful quit attempt at 4 weeks. The results of binary logistic regression modeling showed that this quit success was predicted by receiving an intervention including BS and NRT. Participants were highly satisfied with the support they received.

What is already known

Using NRT within a specialist outreach clinic model to support successful smokeless tobacco cessation in participants of South Asian (Bangladeshi) origin resident in England has been reported in two other studies [21, 22]. A cessation trial with male smokers of South Asian (Pakistani, Bangladesh) origin has also indicated the benefits of providing a specialist outreach service for otherwise ‘hard-to-reach’ participants [23].

The results of this study also confirm that the mean number of daily smokeless tobacco intakes is approximately 10 in English resident South Asians [21, 22]. Results from other smokeless tobacco trials have reported the value of NRT in reducing withdrawal symptoms and NRT-related adverse events during quit attempts [13-19]. Participant satisfaction with English NHS SSS has been reported previously to be high [23, 33]. Oral pain was reported commonly by clients at the start of the quit attempt, as in a previous study [9].

What the study adds

Withdrawal symptoms and NRT-related adverse events have not been reported previously for South Asian smokeless tobacco users making quit attempts. Both of these are demonstrated to be at a low level and to reduce significantly between initial and later weeks of the quit attempt. Participant satisfaction (100%) was comparable to that reported previously in the English NHS SSS (93%) [33], with similar (98%) high levels of service recommendation to potential participants. The use of telephone interviewing in this study to collect these participant satisfaction data has, however, resulted in a much higher participation rate.

Limitations

Limitations should be acknowledged arising from factors not explored in this study. First, higher abstinence rates were reported by participants receiving NRT with BS. These participants may have been more motivated to quit or less nicotine-dependent. Data on participant motivation to quit and NRT prescribing compliance was not collected, while the mean level of dependence reported was relatively low. While there was no relationship between nicotine dependence score and a successful quit attempt, participants with fewer daily smokeless tobacco intakes were more likely to make a successful quit attempt. The cessation support workers were not blinded to the different interventions negotiated with participants, while participants were self-selected with no random allocation to treatment options. As this was a new service in two (Leicester, Bradford) of the three study locations, the possibility cannot be discounted that the cessation workers working there altered their performance due to their awareness that their performance was being observed. The limited period of research funding provided meant that participants with incomplete data at the completion of the funding period were excluded from the study analysis. This constraint also required the adaptation of current outcome criteria for cessation trials with respect to the proposed period of follow-up, although a more stringent continuous abstinence definition was adopted than that currently used in the English NHS SSS, which allows a 2-week grace period [35]. The lack of a validated NRT adverse events measure should also be noted.

Secondly, this study aimed to select participants from the three South Asian communities resident in England. It should be noted that the major participant group was of Bangladeshi origin, and that there were relatively fewer Pakistani participants recruited than expected if the national smokeless tobacco prevalence data are used [5]. Future multi-centred studies should seek to address this sampling shortfall as well as to explore any variation that might exist between settings.

Thirdly, South Asian smokeless tobacco products present particular challenges. A wide range of products are available for use which may be used singly or in combination, while many also contain areca nut, which is itself a recognized carcinogen [3]. Furthermore, dual smoked and smokeless tobacco use behaviours should be considered. While this behaviour is uncommon in women, it has been reported as relatively prevalent among Bangladeshi men [36].

Future research

These findings indicate the need for a randomized trial to examine the effectiveness of NRT to support smokeless tobacco cessation in English resident South Asians. This study has adopted as a key principle that NRT can be used to address the biological aspects of nicotine addiction. Secondly, alternative medication such as varenicline could be researched, because a randomized trial in Swedish snus users shows it be effective for supporting smokeless tobacco cessation [37]. The data related to participant satisfaction have not been reported before and require replication and further extension. Oral pain as a barrier to a successful quit attempt requires further exploration, including possible oral factors such as advanced decay and tooth wear [38], poor gum health [39, 40] or oral mucosal lesions [9]. The recognized analgesic role of nicotine, accessed either in smokeless tobacco or as NRT during the quit attempt, may facilitate management of any oral pain symptoms and awaits further investigation.

In conclusion, South Asian smokeless tobacco users resident in England accessing services to help them stop appear to have short-term success rates comparable with smokers attending stop-smoking services, with higher success rates being achieved by those using nicotine replacement therapy.

Declarations of interest

This research was carried out as part of a programme of inequalities projects implemented by the UK Centre for Tobacco Control Studies and funded by the Department of Health. There are no constraints on publishing or competing financial and non-financial interests to declare.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The Bangladeshi Stop Tobacco Project is funded by NHS Tower Hamlets.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References