Journal of Medical Screening

Previous randomised studies of mammography screening have shown a significant effect on breast cancer mortality, particularly in women aged 50–Q69 at randomisation. Breast cancer mortality has traditionally been studied by judgments on causes of death, either from cause of death registers or from medical records. In this study an alternative method was used, estimating the excess mortality associated with breast cancer.

In 1990 two counties of northern Sweden started population based mammography screening of women aged 40–74. The unscreened population in the two other counties of the same region were selected as controls.

Excess mortality associated with breast cancer was lower in the screened population, and was discernible three to four years after the start of screening. The relative risk estimate, based on the cumulative excess number of deaths from breast cancer during 1990–95 in the screened versus the control population aged 40–74 (at diagnosis of breast cancer), was 0.72 (95% confidence interval (CI) 0.53 to 0.99). For women aged 50–69 it was 0.67 (95% CI 0.46 to 0.99). In the 50–69 age group the estimated excess number of deaths from breast cancer during 1995 was 17.0 per 100 000 women (95% CI 5.0 to 29.0) in the screened counties and 51.1 per 100 000 (95% CI 30.2 to 71.9) in the unscreened counties.

Population based routine screening has substantial effects on breast cancer mortality in women aged 50–69. Estimation of excess mortality can be used in future studies to evaluate the effects of mammography screening on breast cancer mortality.

To compare the UK breast screening programme with the Forrest Report recommendations of 1986.

The UK breast screening programme.

A postal survey of 97 local breast screening programmes in the United Kingdom. The main outcome measures were the frequency of screening, the use of two view screening on incident screens, reading of screening mammograms, assessment procedures and visits, staffing levels, and the use of building and equipment.

Eighty two (85%) of the questionnaires were completed and returned. All programmes screen every three years, as Forrest intended, with the exception of one health region which screens more often. The national policy is to use two views on incident screens where there is a clinical indication. None the less, 14% of programmes are using, or intending to use, two views on all women. Double reading of mammograms is not recommended in the United Kingdom outside Scotland, but is used by 88% of programmes. All programmes have access to the equipment required for the assessment techniques recommended by Forrest. Variation exists between programmes in the procedures women can expect to receive at their initial assessment visit and in the total number of assessment visits. Sixty eight per cent of programmes' breast screening budgets cover the staff required for a multidisciplinary team as denned by the Forrest Report. Ninety three per cent of screening programmes are organised around static sites, with 86% of these also using mobile vans.

The national programme is following recommendations about the frequency of screening, but there seems to be some divergence from policy as regards the use of double reading, two views at incident screening, and the multidisciplinary team covered by the programmes' breast screening budget. Further research is needed on the effectiveness and cost effectiveness of two view incidence screening, double reading, and non-radiologists as readers. Investigation is also needed of the costs and effects of the variation between programmes in the number of assessment visits a woman may have.

To determine the effectiveness of follow up letters to non-attenders for screening on the breast screening uptake in practices with a low preliminary uptake of screening.

Observational study of two groups of general practices. In 40 of these practices, the preliminary uptake of screening was less than 60%These 40 practices were offered help from a clerical officer to check names and addresses of non-attenders, and to send non-attenders a reminder letter.

93 general practices in South West London in 1995–96.

Preliminary and final uptake of breast screening.

Breast screening uptake increased by an average of 4.6% in the 40 intervention practices compared with 1.6% in the 53 control practices (difference 3.0%, P < 0.0001). However, the absolute increase in the uptake of screening in the intervention group was small (from 53.8% to 58.5%). The marginal cost for each additional women screened was £7 (compared with an average cost for each women screened of £27).

Reminder letters can help increase the uptake of screening in practices with a low preliminary uptake of breast screening. However, they have a limited role in improving the uptake of breast screening in inner city areas, and other methods of increasing uptake therefore need to be developed and evaluated.

To document current practice of communicating screen negative results to pregnant women undergoing a test for Down's syndrome.

169 British NHS hospital antenatal clinics currently offering multiple marker serum screening for Down's syndrome and giving results directly to women.

All 169 clinics were sent a letter asking about the method and form of communicating screen negative results.

In only 29% of programmes were specific arrangements made to inform women of screen negative results, and in 5% these results were not given at all. Screen negative results were given as a verbal phrase in 44% of programmes, as a risk figure in 16% of programmes and as both in 40% of programmes.

These results highlight a gap between screening policy guidelines and practice in the case of Down's syndrome serum screening.

To assess neonatal screening for cystic fibrosis using immunoreactive trypsin, either alone or in conjunction with DNA analysis for the AF508 mutation. A novel three-stage screening protocol was compared with the previously introduced two-stage immunoreactive trypsin-DNA protocol.

(a) Collection of data from a 4¹/2 year period (phase 1) of two-stage immunoreactive trypsin screening. The initial dried blood samples were obtained at 6 days of age and repeat samples at 27 days of age from babies with results above the 99.5th centile. Babies with persistent hypertrypsinaemia were referred for a diagnostic sweat test. (b) Retrospective DNA analysis: Patients with cystic fibrosis diagnosed in phase 1 were genotyped and most samples from babies with increased initial immunoreactive trypsin but normal results in the second sample were analysed for the δF508 mutation, (c) Phase 2, a prospective study of a three-stage neonatal screening protocol, in which only babies heterozygous for the δF508 cystic fibrosis mutation progressed to the second immunoreactive trypsin test.

The Trent neonatal screening programme.

437 859 babies born between August 1989 and March 1996.

Proportions of unaffected babies requiring a second blood sample or a sweat test. Overall sensitivity for the detection of cystic fibrosis.

The two-stage screen failed to identify six out of 94 cases of cystic fibrosis (without meconium ileus). The introduction of the DNA analysis step would have resulted in one additional case being missed. With the three-stage screen there was a 92% reduction in babies requiring a second blood sample and an 80% reduction in negative sweat tests, results close to the predictions of the retrospective study.

The three-stage screening protocol is a marked improvement on the two-stage immunoreactive trypsin strategy and on the two-stage immunoreactive trypsin-DNA strategy recently introduced in some other screening programmes.

To survey screening practices, knowledge, and attitudes towards screening among first degree relatives of people with colon cancer.

A random sample of people with colon cancer listed on the New South Wales (NSW) Cancer Registry were mailed a questionnaire to be passed on to an appropriate first degree relative.

Two hundred and twenty five first degree relatives completed a self administered questionnaire.

Although there were high levels of awareness about colorectal cancer, and attitudes towards colorectal cancer were generally positive, screening rates were low, and only three relatives had been screened in accordance with current Australian recommendations. Factors associated with previous participation in any type of screening test (usually once) included receiving a medical recommendation to screen, having more than one relative with colorectal cancer, being a sibling of the relative with colon cancer, the relative with cancer being female, and perceiving screening as messy, but not painful.

Strategies to enhance screening awareness and participation among relatives need to be considered. This study provides some insight into factors to be considered in developing awareness programmes. Further research is required to explore these factors, and to identify ways to overcome barriers.

To test the hypothesis that a personal invitation for cervical screening by a woman's own general practitioner (GP) achieves a higher attendance of women with an increased risk for cervical cancer.

Two general practices and the local health authority screening programme for cervical cancer, Nijmegen, The Netherlands.

Attendance rates of women with an increased risk of cervical cancer were compared for two invitation strategies: (a) invitation by the woman's own GP, and (b) invitation by a national call system through the local health authority. Data on risk profiles were gathered by questionnaire. Two hundred and thirty eight women eligible for screening were invited by their GPs (GP group), and 235 women by the local health authority (control group) in 1992.

The personal invitation by the GP resulted in an 18% higher overall attendance, and a 28% higher attendance of women with greater risk because of sexual behaviour and smoking.

Greater involvement of the GP in inviting women for cervical cancer screening results in a higher attendance, particularly among women with increased risk, than a less personal health authority call system.

To determine trends in incidence of invasive adenocarcinoma of the uterine cervix in East Anglia.

Cervical cancer incidence data for both squamous cell carcinomas and adenocarcinomas were obtained from the East Anglian Cancer Registry for the period 1971–94. Similar data were obtained for England and Wales. European age standardised rates (ASRs) were used for comparisons.

The mean incidence (ASR) of cervical adenocarcinoma was 0.85 per 10^s in 1971–76, rising to 2.54 per 10⁵ in 1989–94. There has been a marked age shift, with the main increase in incidence occurring in younger women aged 30–39. The mean incidence (ASR) of squamous cell carcinoma of the cervix has decreased from 9.78 to 8.74 per 10^s over the periods 1971–76 and 1989–94. Again there has been an age shift, moving from a single incidence peak in the 4S-S9 age band in earlier years to incidence peaks in both the 30–39 and 55–69 age bands in more recent years. Similar trends were noted when data for England and Wales were analysed. Birth cohort analyses show that both tumours are occurring progressively earlier (about five years earlier in each five year birth cohort).

Although the overall incidence of cervical carcinoma is declining, this study has shown an increased incidence of cervical adenocarcinoma, particularly in the younger age groups. In future it would seem advisable to publish separate incidence and mortality data for squamous cell carcinoma and adenocarcinoma of the uterine cervix. All practitioners involved in the cervical cancer screening programme would then be aware of the very real significance of this tumour.

To study the trends in the incidence of invasive cervical cancer in East Anglia.

Statistical analysis of age specific incidence rate for the period 1971–93 using East Anglian Cancer Registry data.

All cases of invasive cervical cancer registered with the East Anglian Cancer registry, diagnosed in the period 1971–93.

Changing incidence of cervical cancer.

For the 20 years 1971–90, trends varied widely by district and by age group, with little discernible overall effect of the increasing screening activity. Since 1990, rates have fallen sharply in the age groups targeted for screening, with a reduction of 34% (95% confidence interval 26% to 42%) from that expected based on 1971–90 trends. This fall was preceded by a rapid rise in the national uptake of screening. A shift to more favourable stage at diagnosis has also occurred.

Changes in the organisation and management of the national screening programme introduced in 1988 and 1989 seem to have led to substantial improvements in effectiveness.

To assess quality and variability in measurements of the infrarenal aortic diameter by ultrasound, and to recommend quality control measures to improve consistency in measurements of the infrarenal aortic diameter (IAD) in a long running screening programme.

An aneurysm screening programme in Huntingdon.

Quality of the ultrasound image was subjectively assessed by each ultrasonographer. Quality of the measurements was assessed by analysing the frequency of measurements that were outside the normal variability of the estimated true diameter. The interobserver variability was measured by analysing repeated measurements of the IAD in the same patient by two ultrasonographers, using the same scanner. The variability between different scanners was measured by analysing repeat measurements of the IAD in the same patient by the same ultrasonographer, using two scanners. The intraobserver variability was estimated by analysing all patients with three consecutive measurements of the IAD, carried out by the same ultrasonographer.

Although the subjective assessment of the quality of the ultrasound image of the aorta varied, there were no statistically significant differences in the likelihood of obtaining measurements outside the limits of agreement between the ultrasonographers. The interobserver, intraobserver, and between scanner variability of ultrasound measurements of the IAD were all around 6 mm.

The variability in ultrasound measurements of aortic diameters is acceptable for clinical decision making, and the interobserver variability is very similar to the long term intraobserver variability. Quality control measures are suggested to maintain long term consistency of ultrasound measurements of the IAD.