The authors stated that the objective of the report was to present evidence-based guidelines in the prevention of oral cancer and pre- cancer among asymptomatic patients. The realised objective was to summarise evidence pertaining to recommendations for practice regarding oral cancer mortality.

MEDLINE and Cancerlit were searched from 1966 to 1999 using the following MeSH terms and textwords: 'mouth neoplasms', 'oral cancer', 'precancer', 'screening', 'population surveillance', 'therapy', 'smoking cessation', 'alcohol reduction' and 'evaluation studies'. In addition, relevant journals (unspecified) were searched manually and the references from appropriate studies were reviewed. Only studies published in the English language were considered.

Study designs of evaluations included in the review

All study designs apart from case reports, reviews, expert opinions and abstracts appear to have been included. Studies of smoking and alcohol cessation included randomised controlled trials, cohort and case-control studies.

Specific interventions included in the review

The specific interventions were: counselling of patients to modify (cease or reduce) smoking and alcohol consumption; screening for oral cancer; subsequent treatment (surgery or chemotherapy) of pre-cancerous lesions; subsequent treatment (surgery and/or radiotherapy) of early-stage cancerous lesions.

Participants included in the review

All studies of screening or treatment had to have included patients with oral cancer (International Classification of Diseases, ICD-9 code: 140 to 149). For studies on cancer therapy, patients had to have early-stage disease of the oral cavity of oro-pharynx. No information was given regarding the participants included in the review. No inclusion or exclusion criteria relating to the characteristics of the patients were mentioned.

Outcomes assessed in the review

The following outcomes were assessed.

For screening: sensitivity, specificity, positive predictive value, and the yield of lesions found (% of suspicious lesions and % of confirmed lesions).

For the effectiveness of counselling: the rates or incidence of cancerous or pre-cancerous lesions.

For the effectiveness of treatment: the recurrence rates of pre- malignant lesions; the reversing and stabilising of oral leukoplakia; progression to malignancy; decrease in lesion size; and 5-year survival rate. Specific outcome results had to be reported for stage I and stage II cancer.

How were decisions on the relevance of primary studies made?

The authors do not state how the papers were selected for the review, other than the fact that the lead author made all decisions on the relevance of identified studies. These studies were then considered by the Canadian Task Force on Preventive Health Care.

The authors do not describe the methods used to assess validity, although they did refer to the quality of the studies being rated using the recommendations of the Canadian Task Force on Periodic Health Examination (see Other Publications of Related Interest). These recommendations related only to study design.[A:The quality of studies was rated using the "levels of evidence hierarchy" of the Canadian Task Force. The quality of studies was not rated using the recommendations]. The authors do not state how the papers were assessed for validity. The lead author provided ratings on the quality of the papers, but these were not presented in the paper and appeared to be based on study design alone. These ratings were considered by the Task Force.

The lead author extracted data from the primary studies. No further information was provided on the method used to extract the data. The following information was tabulated.

For screening studies: study name, number of participants, number and type of examiner, 'gold' standard, sensitivity, specificity, and the type of programme.

For yield of lesions from screening studies: study name, number of participants, country where study conducted, percentage yield of suspicious lesions, percentage yield of confirmed cancerous lesions, and the number of confirmed cancerous lesions.

For effectiveness of treatment: study name, number of participants, intervention, site of lesion, 5-year survival rate.

Data from the other types of studies were not tabulated.

How were the studies combined?

The studies were combined using narrative accounts for some outcomes. For other outcomes, the data were presented but no attempt was made to synthesise the results.

How were differences between studies investigated?

Differences between the studies were discussed narratively, but as the studies were not synthesised, no formal assessment of heterogeneity was performed.

Screening for oral cancer: 5 population-based studies with 40,069 patients were included. No randomised study has been completed, although one screening intervention study was begun in India in 1995.

Yield of lesions from screening programmes: 9 studies with over 13 million patients were included.

Effectiveness of counselling for smoking cessation: 3 case-control studies, one cohort study, 3 school-based studies, and one case series. There was no information relating to the number of participants.

Effectiveness of counselling for alcohol cessation or reduction: there was no information relating to the number of studies or participants.

Effectiveness of treatment: 9 studies with 2,524 participants were included.

Screening for oral cancer: population-based studies showed high specificity (98 to 99%) and variable sensitivity (56 to 94%). The positive predictive values also varied from 15 to 91%.

Yield of suspicious lesions from screening programmes: the yield from population-based screening studies was under 6% overall, and under 2% in 5 of the studies. The yield of confirmed cancerous lesions was under 0.05% in all but one study, which included high-risk patients in Italy. In that study, the yield was 2.4%.

Effectiveness of counselling for smoking cessation: no results were presented, but the authors stated that smoking cessation decreases the risk of developing oral cancer. School-based programmes had mixed results, whereas smokeless tobacco cessation programmes have been assessed in only small case series. The authors also stated that randomised controlled trials indicate that counselling by trained health professionals is effective in promoting smoking cessation.

Effectiveness of counselling for alcohol cessation or reduction: the authors stated that studies have failed to show that alcohol reduction leads to a decrease in the risk of oral cancer and pre- cancer. They also stated that no studies have examined whether alcohol reduction reduces the risk of oral cancer or pre-cancer. From these two statements, it is unclear whether there is no evidence, or whether the evidence shows no difference. [A:The author regrets using the statement "studies have failed to show". This statement should have read "no studies have assessed" or "no evidence was available". At the time of this review, there was no evidence concerning the association between alcohol reduction and decreased risk of oral cancer.]

Effectiveness of treatment: surgical excision does not eliminate the risk of pre-cancerous lesions progressing to malignancy. The effectiveness of laser removal has yet to be assessed in clinical trials. No attempt was made to synthesise the results of the effectiveness of chemotherapy (13-cis-retinoic acid, beta-carotene, bleomycin) for pre-cancerous lesions, or radiotherapy versus surgery for cancerous lesions.

The following recommendations were made on the basis of the results found in this review. Each recommendation was assigned a level of evidence from A to E: the recommendation was level A if there was good evidence to suggest that the intervention should be adopted, and level E if there was good evidence to suggest the intervention should not be adopted.

Smoking cessation counselling by a trained health professional should be adopted as part of the periodic health examination (level A).

Population screening for oral cancer should probably not be adopted as part of the periodic health examination (level D).

Opportunistic screening for oral cancer: there was insufficient evidence to support a recommendation (level C).

[A:Recommendations were assigned a grade, not a level of evidence, from A to E. In this section, the term "level" should be "grade".]

The literature search was probably inadequate. The searches were restricted to two databases, and only studies published in the English language were included. It is possible that this review is incomplete, although since not all the primary studies were referenced, it is difficult to judge the impact of this flawed methodology on the results.

The description of the review methodology was poor. There was no information on how the data were extracted. In addition, there were no details of the patients' characteristics, and there were no data for some outcomes. Furthermore, there was no indication of the quality of the studies, although the papers appear to have been judged relevant by the lead author, and were presented to and deliberated on by the Canadian Task Force on Preventive Health Care. [A: Levels of evidence were provided in table 4 and these indicate the study design used in the studies - an indication of the quality of the studies.]

The results of the review were muddled. The review did not provide the reader with a clear idea of the evidence-base upon which the recommendations were made. [A: Table 4 provides a summary of the preventive intervention, its effectiveness, the level of evidence of studies with references, and the recommendations.] For one outcome (effectiveness of alcohol reduction counselling on oral cancer risk), no studies were found. [A:A recommendation was not made regarding alcohol reduction counselling on oral cancer risk. However, it was stated that a recommendation regarding alcohol reduction counseling could be made on other grounds (e.g. prevention of heart disease), and the CTFPHE has previously made such a recommendation.] However, the authors stated that studies have failed to show that alcohol reduction leads to a decrease in the risk of oral cancer or pre-cancer. [A: For the effectiveness of alcohol reduction counselling on oral cancer risk, no studies were provided because none existed at the time of the review. The author regrets using the statement "Studies have failed to show". This statement should have read "no studies have assessed" or "no evidence was available"]. A recommendation made in relation to heart disease was referred to in this section. In summary, although there appears to be a number of useful studies that could inform strategies to reduce oral cancer mortality, these need to be identified and synthesised in a systematic way if they are to be of use.

Practice: The authors state that smoking cessation counselling by a trained health professional should be adopted as part of the periodic health examination. They also state that population screening for oral cancer should probably not be adopted as part of the periodic health examination, and that there is insufficient evidence to support a recommendation for opportunistic screening for oral cancer in Canada.

Research: The authors state that prospective evaluation of screening programmes for high-risk individuals is required. Consideration should also be given as to which health professionals should perform the screening and counselling.

The authors state that studies are needed to establish a causal link between alcohol reduction and the reduced risk of cancer and pre-cancer.

Reviewer's statement: It is the opinion of this abstract reviewer that a well-conducted systematic review of this area is warranted, prior to any research or practice recommendations being made. This does not negate the clear beneficial effect of smoking cessation on health. Furthermore, the presence or absence of an association between alcohol reduction and the reduced risk of cancer and pre-cancer is required before possibilities of the 'causal link' are investigated.

Woolf SH, Battista RN, Anderson GM, Logan AG, Wang E. Assessing the clinical effectiveness of preventive manoeuvres: analytic principles and systematic methods in reviewing evidence and developing clinical practice recommendations. A report by the Canadian Task Force on the Periodic Health Examination. J Clin Epidemiol 1990;43:891-905.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.