Increase uptake of Smoking cessation services in deprived areas

Background

Tobacco is a major killer. About half of the smokers die prematurely due to smoking related diseases. The most recent estimates show that around 114,000 people in the UK are killed by smoking every year, accounting for one fifth of all UK deaths (Peto, R. et al). People under 35 who never stop smoking may lose 10 years of life expectancy.

In the UK about 30% of smokers attempt to quit, of whom 3% to 6% are successful (1% to 2% of all smokers). Almost all smokers try to quit without professional support from NHS, private clinics or workplace schemes.

An analysis of quit attempts made by Scottish Index of Multiple Deprivation (SIMD) category shows the largest numbers to be in the most deprived categories and the smallest in the least deprived, Scottish Household Survey estimates (2005/2006) of smoking prevalence by deprivation quintiles show a smoking prevalence rate of 41% in the most deprived areas and 13% in the least deprived areas. People from deprived background are underprivileged in terms of income, employment, housing, education and most importantly access to health services. So while planning smoking cessation intervention for people from deprived background these factors need to be taken into consideration. For instance costly habit of smoking causes burden on this poor man’s household budget. One survey found that if both partners smoke, they could be spending as much as 15% of their income on tobacco. There are many smoking cessation services in action such as one to one counselling, group counselling, pharmacotherapy, anti-smoking mass media campaign, telephone counselling. Doctor can give advice on how to stop smoking and prescribe nicotine replacement therapy (NRT), Zyban or Champix. In some areas, local pharmacists are part of the stop smoking services and can also offer one-to-one support along with NRT.
‘Telephone mediated advice and counselling to assist smoking cessation efforts has become increasingly a subject of interest to tobacco control programmes, in the search to provide accessible and affordable methods of quitting smoking’( Wakefield, et al). Another study conducted in United States gives evidence of effectiveness of telephone help lines. ‘Low income earners may not be able to afford some quit smoking methods such as nicotine replacement therapies, but helpline services which have a free call number are accessible to all’. (Pierce JP, Anderson DM). This study is supported by smoking cessation study done in California. This study emphasizes on one of the attractive qualities of telephone help lines. Smokers who live in rural areas, where community based smoking cessation options are few, will benefit from help lines. (Zhu SH, Rosbrook B, Anderson C,et al.)
One to one counselling is also one of the effective interventions. ‘Approximately 70% of smokers visit health care provider. This provides them good opportunity to provide personal counselling to smokers to quit’. (Campion P).
Therefore help lines and doctor’s counselling being more suitable interventions for people from deprived background will be evaluated in this study. Here experimental group is considered in context of its adjacent factors.
Objectives:
Current smokers of both sexes above age of 20 years who are motivated to quit from clinics located in Glasgow city (35% of the most deprived data zones according to SMID category) are regarded as Patient population. To increase uptake of smoking cessation service by people from deprived background, will be quantifying effects of two interventions i.e. telephone help lines and one to one counselling by doctor. These two experimental groups will be compared with ‘no intervention group’. Outcomes are measured in terms of number of people entering in service and number of people who successfully quit smoking. Research question: ’ Does provision of one to one counselling or telephone counselling to smokers in deprived areas result in increase uptake of Smoking cessation services when compared with no intervention?’
In the current study we will be quantifying the effectiveness of two interventions. We will be tasting hypothesis that either one or both interventions would: 1. Increase the uptake of smoking cessation service in deprived areas, 2. Increase rate of successful quitting. Primarily, number of current smokers entering in any one of interventions is recorded. Secondary outcome is to measure number of patients who quit smoking.
Design:
Complex randomised controlled trial is preferred over simple RCT. Since, this study is comprised of two intervention groups and one control group. Multi arm trial method is conducted where study population from selected deprived data zones will be receiving different smoking cessation interventions. It is a simple design and allows observing all patients in every group. It is a pragmatic trial which will be quantifying effectiveness of interventions. Outcomes of this study are resource oriented and are patient centred. Such pragmatic trials give long term outcomes. Though simple RCT is considered as gold standard of research designs, it has limitation of lack of generalisability. With complex RCT result can be generalisable to the population with more accuracy as participants are selected from routine clinical practice. Clinics are randomised to interventions as in this study we are non pharmacologic treatments. Blinding either single or double with individual randomisation is not possible. Confounding variables related to geographical variation or socioeconomic differences are reduced by selecting from deprived data zones with same baseline.
In addition factorial design is alternative to multi arm trail but it is more difficult to operationalise. Patient preference trail can be thought about, to maximise motivational factors. To quit smoking most important factor is strong motivation to quit. But this design may cause unequal distribution across the groups.
Although multi arm trial may face with few biases such as attrition bias. Smoking cessation attempt gives rise to withdrawal symptoms which contribute to systematic difference in withdrawals from interventions. So this study requires large sample size to achieve same level of power. Also multiple comparisons make study difficult. Cluster randomising clinics to different interventions minimises risk of contamination bias between groups. But at the same time it raises question over intra cluster correlation. Also performance bias cannot take place in this study as clinics are randomised to different intervention. Each performer was randomised to different clinics.
Data is extracted by either doctor or health personnel for help line but is analysed by third person who is blinded for intervention allocation. Firstly, we analysed the results with all patients who entered the trial. Then odds ratio for abstinence from smoking after 6 and 12 months follow up is taken. We selected the strictest measure of abstinence, using biochemically validated rates. We assessed statistical heterogeneity amongst sub groups of clinically comparable studies using the chi square statistic.

Study population:
Study site and population: as the study involves non pharmacological treatments, eligibility criteria for settings and locations affect the external validity of a trial. Study centres are located in Glasgow city (35% of the most deprived data zones according to SMID category). They should be easily accessible by study population. Telephone help lines are centralised to one fully equipped call centre. Number of counsellors should be enough to call and reply to proactive or reactive callers. Counselling was most commonly provided by professional counsellors or trained healthcare professionals. Trained postgraduate students can also perform as counsellor. The eligibility criteria do not affect the internal validity of a trial, but they do affect the external validity. The schedule of the telephone counselling was up to 12 calls for up to four months and access to hot line. Current smokers of both sexes above age of 20 years who are motivated to quit from clinics located in Glasgow city are included as study population. This inclusion criterion decides about generalisability of result to population. Smokers can current smoker with or without previous attempts to quit. Both sexes are included as consequences of smoking in both genders are distinct and equally hazardous. Pregnancy may act as confounding factor. It is a multi centre trail involving ‘n’ number clinics located in Glasgow city and one centralised call centre for telephone counselling. Clinics are located at similar deprived data zones, therefore enables to give similar geographical distribution and socio- economic back ground. This reduces differences between participants. Smokers who are lost to follow up are not excluded but are considered as patients continuing smoking.

Sample size:

Sample size depends on study design, outcome measure and the statistical analysis. In this study multi arm trial with cluster sampling to measure continuous outcome of attempt and quit rates with Chi square test is performed. All the three categories under this multi arm trial with cluster sampling gives intra cluster correlation which can reduced by increase in sample size. Smoking cessation attempt gives rise to withdrawal symptoms which contribute to systematic difference in withdrawals from interventions. So this study requires large sample size to achieve same level of power. Mean difference in attempt and quit rates with 80 % power and 5 % significance will give difference in effectiveness of intervention.

Study recruitment:
This study is conducted in multiple centres involving ‘n’ number clinics located in Glasgow city and one centralised call centre for telephone counselling. Clinics are located at similar deprived data zones, therefore enables to give similar geographical distribution and socio- economic back ground. This reduces differences between participants. These deprived data zones are identified with the help of Scottish Index of Multiple Deprivation (SIMD) category.

‘n’ number of clinics are allocated to different interventions( Arm ‘A’, ‘B’ or ‘C’). A computer-generated randomisation list is drawn up by the statistician. In this trial double blinding is difficult or impossible. However, blinded assessment of outcome can be achieved. Patients and care providers are not blinded. Person accessing the outcome and who is not involved in the running of the trial is blinded to the allocation of centers to different intervention. The code will be revealed to the assessment team once recruitment, data collection, and analyses are complete. In Glasgow city most deprived data zones are selected and splitting is done into clusters. For survey computer generated list of cluster sampling of areas is used to recruit clinics under respective area to different interventions. This strategy reduces contamination between groups and is easy to organise practically. Stratified sampling requires homogenous strata which not possible in this Trials of Non pharmacologic Treatments. Multi stage sampling can be used in place of cluster sampling if sub population is heterogeneous. Eligible smokers are told all Information of possible withdrawal symptoms, access to management were told to eligible smokers and were consented before entering into study. The trial received ethical approval from the multicentre and local research ethics committees

Study intervention:
Primary care centres within selected deprived data zone are randomised to two interventions or control group. Here every clinic has equal opportunity to get randomised to interventions. Smoker patients who are motivated to quit will be getting respective intervention or no intervention depending upon status of their clinics allotted to. In participating practices, doctors were asked to recruit all eligible smokers seen in clinic who are ready to quit.
Here clustering of clinics is done. And we are comparing relative effectiveness three interventions: Arm ‘A’, Control; Arm ‘B’, one to one counselling; Arm ‘C’, telephone help line.
Intervention arm ‘A’ is regarded as control group. Patients attending clinics under arm ‘A’ are supposed to receive only general information of smoking cessation along with the other complaint with which they attended the clinic. There is no special ‘one to one counselling’ with patient as it will be taking place in clinics under arm ‘B’ i.e. clinics with ‘one to one counselling’ intervention. In clinics under arm ’B’ doctors with specialised knowledge listen to patient’s reasons of failure to quit and discuss health problem arising from smoking and also supports patients to quit. It is time consuming intervention but has strong influence on smoker due to direct contact. It may give high response rate. In clinics under arm ’C’ eligible patients for this study are provided information regarding smoking free help line service. It is an easily administered as can access many smokers easily. Information is provided by trained health care personnel. They received training related to all aspects of smoking, its effects on body and how to quit. Provision of proactive or reactive telephone counselling assists smoking cessation. These telephone services may offer information, recorded messages, personal counselling or a mixture of components according to the need of patient. In arm ‘B’ and ‘C’ smoker gets support in making a quit attempt or avoiding relapse during the initial period of abstinence. But doctor or counsellor on telephone are not supposed to nag or give lecture to smokers. Because smoker’s own decision to stop smoking has string influence on quit attempt.
Outcome measurement:
To measure attempt rate i.e. number of smokers entering into study is the primary outcome. And further needs to calculate quit rate amongst participants. This outcome guides towards effectiveness of intervention. It is a smoking group specific outcome. To measure outcome one can either use existing published scale or can device a new scale. In this study, we propose six standard criteria comprising the 'Russell Standard' (RS). These criteria are applicable to trials of cessation aids where participants have a defined target quit date and there is face-to-face contact with researchers or clinic staff, as follows.
‘(1) Follow-up for 6 months (RS6) or 12 months (RS12) from the target quit date or the end of a predefined 'grace period'
(2) Self-report of smoking abstinence over the whole follow-up period allowing up to five cigarettes in total
(3) Biochemical verification of abstinence at least at the 6-month or 12-month follow-up point
(4) use of an 'intention-to-treat' approach in which data from all randomized smokers are included in the analysis unless they have died or moved to an untraceable address (participants who are included in the analysis are counted as smokers if their smoking status at the final follow-up cannot be determined)
(5) following-up 'protocol violators' and using their true smoking status in the analysis.’
These criteria are practical and ensure long term cessation and also enable meaningful comparison between interventions. For telephone counselling where there may not be specified quit date; self reporting of quitting, continued follow up to intervention and biochemical verification of abstinence at least at the 6-month or 12-month follow-up point will be criteria for assessment of outcome.
One needs to access reliability of outcome. Performers i.e. doctors and trained counsellors in this study have influence on reliability of outcome. It is multi centre trial. Therefore it can give rise to both intra observer and inter observer reliability assessment.
Timing of outcome assessment:
Primarily, number of current smokers entering in any one of interventions is recorded at the beginning of study. Each group is given specified quit date. Secondary outcome is to measure number of patients who quit smoking. This assessed at the interval of six and twelve months. At these specific timings point prevalence of successful quit rate is assessed. At this time all smokers in the study are need to undergo biochemical verification of abstinence.

(2502 words excluding references)
References:
Ann Bowling, Research methods in health, Buckingham, Open University press, 1997.
Black N., Brazier J., Health services research methods, London, BMJ books, 2000.
Lecture notes of Health service research course http://www.consort-statement.org http://www.consort-statement.org/mod_product/uploads/CONSORT%20Extension%20for%20Non-Pharmacologic%20Treatments%202008.pdf http://www.cochrane.org/reviews http://tobaccocontrol.bmj.com/cgi/content
http://www3.interscience.wiley.com/journal