Scientific protocols should be page numbered and include the following:
Further details can be obtained from the following 'Checklist for the Development of a Study Protocol' and accompanying guidance. It is important, however, that investigators note that this document contains information relating to all aspects of a study protocol - namely the HSE justification, the scientific protocol and the ethical considerations. The Research Ethics Committee is primarily concerned with the latter two elements of the study protocol; the first, relating specifically to project management systems, should be discussed and cleared through line management before making any ethical submission.
1. Introduction
(a) Reason for the study
(b) Historical background and literature review
2. Study objectives
3. Survey methods
(a) Population to be studied
(b) Controls
(c) Design - cross sectional: longitudinal
(d) Measurements of exposure:
(e) Measurements of response:
(f) Other variables: eg atopy, smoking
(g) Definitions: eg bronchitis, work related asthma
4. Agreements
(a) Management
(b) Unions
(c) Ethical
(d) Other parts of HSE
5. Tools
(a) Check lists for field visits
(b) Questionnaires
(c) Data collection and summary forms
(d) Consent forms and Subject Information Sheets/Letters
6. Data analysis
(a) Statistical advice ((a) pre and post study)
(b) Computer support
7. Target dates
(a) for agreement of protocol (including ethical consideration)
(b) for survey
(c) for statistical analysis
(d) for completion of report
(e) for publication
8. Publication
(a) authors
(b) media
(a) Reason for the study
This should explain in both professional and HSE terms why a study is needed. It may be that several workers are suffering from the same clinical problem and there is the suggestion that it is work related. There may be the possibility to study a workplace situation because of its scientific interest, eg asthma to a completely new type of pharmaceutical. There may be the need to investigate the way that employers or employees interpret regulations in different work situations and relate them to risk. This short paragraph at the beginning of the protocol should make it clear to any reader (lay or medical) that there is a valid professional reason for the study.
(b) Historical background and literature review
This will vary in depth and length according to the size of the proposed investigation. A field investigation that will only take a few days work does not require a full and exhaustive literature review. However if the project needs the commitment of considerable other resources eg laboratory time, statistical analysis, outside consultants, then you have to show that you know what has gone before and that you are not re-inventing the wheel. Time spent reviewing the literature is not wasted. It may suggest additional questions to ask for little extra effort. It will have to be done eventually if you want to publish or write a thesis. Your paper might not be acceptable at the end of a whole block of work if you have not taken into account some critical piece of published literature.
It is quite reasonable and sometimes elegant to ask one simple question. Too often it is possible to ask an inexperienced investigator "what scientific question are you asking" only to be met by a blank look. Study objectives can be presented in different ways. The major question can be posed first eg -
"Is there a relationship between smoking and lead absorption?"
"Does atopy predispose to allergy to xyz in this work environment?"
"Which component of the new cutting oils is responsible for the dermatitis?"
"What is the incidence of RSI in this particular trade and what are the tasks that cause it?"
"What are the factors that stop people using the available PPE?"
Thus the major objective is to answer this question.
Sub-objectives can be added but as one asks more and more questions the chances of answering anyone go down significantly.
(a) Population to be studied
Preliminary agreement from an employer or employers to study their workforce lends credibility to a set of research proposals. Clearly the population must be exposed to the appropriate working conditions. A major problem is deciding the necessary size of the population. If you are investigating the occurrence of an established problem in a factory the population is defined for you. However what ever you do you have to consider the "power" of your study, eg
"Do enough employees fall off the scaffolding in a year to enquire about their drug and alcohol intake?"
"Are there enough epileptics in this workforce to subdivide them according to the medication they are on?"
"Do enough people become sensitized to this dye per year to establish dose-response relationships?"
This problem of power is crucial. You may need professional statistical advice here. It also interacts with your need for ...
(b) Controls
Often investigators have established a study population only to be horrified when told they need a control group!
If you are studying musculoskeletal injuries in a working population, you must know the prevalence of sport and domestic injuries in a comparable population. It is difficult sometimes to find a matched unexposed group of workers who will accept multiple venepunctures! However for some immunological or biochemical studies it may be possible to use blood samples from another population going through the laboratory as controls. But for some studies you will have to match for age, gender and smoking. In other studies, these effects can be dissected out during the statistical analysis.
It may be important to take advice from the laboratory about controls at an early date and from statisticians about power and matching.
(c) Design - cross sectional or longitudinal
Cross sectional studies are easier to set up and do BUT may be very difficult to interpret. Are you just studying a survivor population?
Longitudinal studies are difficult to manage but are essential to answer some questions. If you set one up, will you be around to analyse the results: will the firm still be operating the same process in 3 years time?
Think carefully with a cross sectional study about the numbers of leavers from the job you are studying. The affected subjects may have left leaving you a very selected set of tough survivors to study.
(d) Measurements of exposure - current
- historical
Exposure today may be under control. Are you interested in this or for a chronic condition do you need to reconstruct past exposures. Are there records of occupational hygiene measurements going back for some years?
Current exposures may be easy to measure; ie by the use of passive samplers. Others may need extensive and sophisticated laboratory studies just to develop the method (eg complex fumes, protein antigens). If you need laboratory measurements that are anything other than routine you may need to enlist other specialists as co-workers early on and will need them to help with the protocol. The validity of your study may depend critically on the quality control of the lab method. It is better to establish "equal colleague" relationships with a scientist than have him withdraw his results as not being suitable to answer your question at the end of the study.
(e) Measurements of response - symptoms
An analysis of work related symptoms is a major part of many field investigations. There is a range of questionnaires available. The most well known is the MRC Chronic Bronchitis questionnaire. Others have been constructed for work-related respiratory disease. But be careful if you are thinking of modifying one for your own circumstances. Changes in the wording or order of questions can invalidate the use of a well established questionnaire. It is also important to design forms so that the information can be easily entered into a data system for analysis.
- clinical
Make sure you use the appropriate measures. Is spirometry really worthwhile? Would metacholine challenge be better? How do you record the extent and severity of VWF?
- laboratory
Do you really want to measure everything (LFTs, renal function, IgEs) in the hope that something will turn up? Have you asked the lab if they have seen anything new in the literature or heard anything at recent scientific meetings that is particularly relevant to your study? Can they offer you a useful control population.
- psychological
Again there is a range of questionnaires to test mood, present psychological state, alertness etc.
(f) Other variables
eg - atopy
- smoking
(g) Definitions
eg bronchitis, work related asthma
Frequently the use of "case" definitions has played an important role in investigating a disease. Find out if there is an appropriate definition for the condition you are investigating. If there is not consider defining a "case" at the beginning of your study.
(a) Management
(b) Unions
(c) Ethical
(d) Other parts of HSE
Although each Directorate and HSE as a whole have formalised planning systems through which one bids for other people's time, much can be gained by informal approaches to potential collaborators at the stage of protocol development. The laboratory measurements you need may take time to set up or there may be a similar approach that would be improved by relating to yours.
(a) Check lists for field visits
Field work involving hygiene and medical laboratory measurements as well as questionnaires and perhaps clinical assessment needs very careful planning and co-ordination. A master plan and detailed checklists are essential to get the right people and equipment on site at the same time.
(b) Questionnaires
(c) Data collection and summary forms
(a) Statistical advice (before and after)
Nothing annoys a statistician more than someone arriving out of the blue with a pile of completed questionnaires expecting elegant statistical analysis, when there's been no consultation on the questions to be asked, the controls, the power, the use of questionnaires, the design of forms, whether he has time to do the analysis....TALK TO THE STATISTICIAN FIRST.
BUT do not assume that because you have a full suite of statistical programs on your micro you can do it yourself!
(b) Computer support
(a) for agreement of protocol
(b) for survey
(c) for statistical analysis
(d) for completion of report
(e) for publication
Clearly if you are dependent on other professionals you will need to agree a timetable with them. This will obviously have to take account of the work programme of the target population, especially if the work (exposure) is intermittent. Target dates in a protocol shows that you are serious and intend to do a professional job. People tend to underestimate the time needed for analysis and writing
(a) authors
(b) media
When a colleague in another part of HSE puts time and effort into the study you have set up, he is NOT producing work that will directly advance his career. So if he is going to invest his skills into your work he needs something back. Agreement to be joint authors costs little but does buy the commitment you need.
Social media
Javascript is required to use HSE website social media functionality.
Follow HSE on Twitter:
Follow @H_S_E