Mammography screening: an incremental cost effectiveness analysis of double versus single reading of mammograms

OBJECTIVE

To compare mammography reading by one radiologist with independent reading by two radiologists.

DESIGN

An observational non-randomised trial at St Margaret's Hospital, Epping.

SUBJECTS

33 734 consecutive attenders for breast screening in the main trial and a sample of 132 attenders for assessment who provided data on private costs.

INTERVENTIONS

Three reporting policies were compared: single reading, consensus double reading, and non-consensus double reading.

MAIN OUTCOME MEASURES

Numbers of cancers detected, recall rates, screening and assessment costs, and cost effectiveness ratios.

RESULTS

A policy of double reading followed by consensus detected an additional nine cancers per 10 000 women screened (95% confidence interval 5 to 13) compared with single reading. A non-consensus double reading policy detected an additional 10 cancers per 10 000 women screened (95% confidence interval 6 to 14). The difference in numbers of cancers detected between the consensus and non-consensus double reading policies was not significant (95% confidence interval -0.2 to 2.2). The proportion of women recalled for assessment after consensus double reading was significantly lower than after single reading (difference 2.7%; 95% confidence interval 2.4% to 3.0%). The recall rate with the non-consensus policy was significantly higher than with single reading (difference 3.0%; 2.5% to 3.5%). Consensus double reading cost less than single reading (saving 4853 pounds per 10 000 women screened). Non-consensus double reading cost more than single reading (difference 19 259 pounds per 10 000 women screened).

CONCLUSIONS

In the screening unit studied a consensus double reading policy was more effective and less costly than a single reading policy.

Evaluation of diffuse technologies: the case of digital imaging networks

There have been significant developments in recent years in the methodologies and methods for the evaluation of a wide range of health technologies. There remain, though, many technologies which are difficult to evaluate. Often the difficulty stems from the complexity of the technologies themselves, which are in effect hybrids, comprising combinations of several distinct elements. In this paper these are termed 'diffuse' technologies, because the different elements exert different costs and effects, often across several different services. Computer networks are one, increasingly important, example of such technologies in health care. While it is possible to evaluate individual elements of such technologies, it is not clear how to evaluate the technology as a whole, where the whole may be greater (or less) than the sum of the parts. The paper outlines a seven-stage framework for the evaluation of diffuse technologies. The general principles of evaluation are illustrated using the example of picture archiving and communication systems (PACS), which are computer systems designed to capture, store and distribute electronic radiological images within a hospital.

Variations in the incidence, management and outcome of stroke in residents under the age of 75 in two health districts of southern England

BACKGROUND

The aim of the study was to determine the incidence, outcome and health service resources consumed by stroke care in defined populations.

METHODS

Patients under the age of 75 experiencing their first stroke between August 1989 and July 1991 were assessed at the onset, and at three and 12 months after their stroke. The settings were West Lambeth (WL) and Tunbridge Wells (TW) health authorities in southern England. The main outcome measures used were: age- and sex-specific incidence rates, hospital admission rates, length of stay and use of rehabilitation services. Functional disability was assessed using the Barthel scale.

RESULTS

Four hundred and fifty-six strokes were registered. The annual incidence rates/1000 population aged under 75 years old [with 95% confidence interval (CI)] were 0.77 (0.67-0.87) in WL and 0.66 (0.58-0.75) in TW. The age- and sex-standardized incidence ratios were significantly higher in WL (126; 95% CI 110-144) than in TW (84%; 95% CI 74-95) (p < 0.001). There were independent associations of incidence with age group (p < 0.001), sex (p < 0.001) and ethnic group (p < 0.001). One year case-fatality was 36% (80/225). At one year, 11% (14) of surviving patients were moderately to severely disabled and 23% (28) mildly disabled. Seventy-one percent (326) of patients were admitted to hospital and the average health service cost per case was pound 3800 in WL and pound 2650 in TW, 93% of the cost being for in-patient care.

CONCLUSION

The study has demonstrated a significantly increased incidence of stroke in an inner-city district compared with a district in rural southern England. It has also established ethnic group as a significant independent risk factor for stroke in the United Kingdom. The cost of care to the health services is considerable, and largely reflects nursing costs in hospital rather than effective treatment packages for stroke.

Explaining variation in radiologists' reporting times

This paper describes an investigation into the reasons for variation in the time taken by senior radiologists to complete radiological reports. An observational study of the reporting process at one UK hospital was undertaken for a 25 day period. An independent health service researcher observed the radiology reporting process and collected data on a variety of factors including the time taken to produce the report, the number and nature of all images viewed, the experience of the radiologist, and the number of disturbances that occurred. The nature of the variation in reporting time was explored using both simple comparative statistics and more sophisticated multiple regression techniques. Data were collected on 2345 report observations and the median report time was 117 s. This research provides the first empirical evidence for systematic variation in reporting time. The results confirm the importance of certain factors that were expected to explain report time variation. For example, the results indicate that report time tended to be significantly shorter in reporting sessions that were busy, and significantly longer when the radiologist was disturbed during the reporting process or was training juniors during a reporting session. More surprising were the results indicating that there was no significant difference in report time for reports categorized as urgent or "hot" and those categorized as less urgent or "cold", and that report time appeared to vary systematically depending on the day of the week and on the time of day.

Issues in the evaluation of picture archiving and communication systems

Picture archiving and communication systems (PACS) are an example of the application of computer technology in the medical field. PACS automates image handling in a hospital and has the potential to transform the way radiology is currently performed. This paper focuses on the evaluation of the PACS technology, and considers the claims that have been made for PACS, how these claims might be turned into questions to be addressed by evaluation and the appropriate methods for the evaluation of PACS. A distinction is drawn between evaluation questions for which the hospital is the appropriate focus and those for which the patient is the appropriate focus. The preferred research design is different for hospital focused PACS evaluation and patient-focused evaluation of small scale PACS systems. A contemporaneous experimental comparison within hospitals is the preferred design for the patient-focused evaluation of small scale PACS systems. The patient-focused evaluation of large scale systems and the hospital-focused evaluation of all PACS systems could feasibly be conducted as contemporaneous experimental comparisons between hospitals but the large research costs implied by such a design almost certainly mean that non-contemporaneous, non-experimental comparisons within hospitals are more realistic. The current situation for the PACS technology is that it has potential, but as yet unproven, benefits and a large capital cost. Thus, the primary purpose of funding additional PACS implementations must be to add to the currently small body of evaluation evidence.

Private costs associated with abdominal aortic aneurysm screening: the importance of private travel and time costs

OBJECTIVES

To assess the importance of the private costs incurred by patients when making a judgment on the economics of screening for abdominal aortic aneurysm (AAA), and to explore the variation in such costs depending on screening location.

SETTING

A district general hospital and general practitioner surgeries.

METHODS

Four hundred and ninety nine consecutive subjects attending for AAA screening completed a questionnaire asking about travel arrangements for the journey to and from the clinic, the distance travelled, the time taken, the mode of transport, and any out-of-pocket expenses incurred. In addition, at the clinic each subject was asked what activities they had forgone in attending the clinic. Time was valued differently depending on whether work or leisure activities were forgone. The total private cost for each attender was calculated and comparison was made between attenders at hospital and at general practice.

RESULTS

A significantly greater proportion of subjects were accompanied when attending hospital than when attending general practitioner (GP) surgeries. Most attenders travelled by car, but the journey time was significantly longer for those visiting hospital. The expected total private cost associated with attendance for AAA screening was 5.47 pounds. Attendance at GP surgeries had a lower private cost (4.21 pounds) than attendance at hospital (6.87 pounds). Only 7.3% of all men surveyed, and 6.5% of all companions, would have been taking part in some form of paid occupation if they had not attended for screening.

CONCLUSION

Despite the fact that most attenders for AAA screening will be retired, the associated private costs are appreciable and should be considered in assessing the economics of such screening programmes. The level of private costs varied depending on the location of screening; clinics held at GP practices had lower private costs than those held at hospital.

Mammography screening: an incremental cost effectiveness analysis of two view versus one view procedures in London

STUDY OBJECTIVE

To compare the costs and effects of routine mammography screening by a single mediolateral-oblique view and two views (mediolateral-oblique plus craniocaudal) of each breast.

DESIGN

A cost effectiveness analysis of a prospective non-randomised trial comparing one and two view mammography screening was carried out at St Margaret's Hospital, Epping. All women in the study had two view mammography. The mediolateral-oblique view was always the first image read by the radiologist. After reading the films for a clinic session, the same radiologist then went back and read both the mediolateral-oblique and craniocaudal views together. Each set of films was read by two radiologists. The main outcome measures were recall rates, number of cancers detected, screening and assessment costs, and cost effectiveness ratios.

SUBJECTS

A total of 26,430 women who attended for breast screening using both one and two view mammography participated. A sample of 132 women attending for assessment provided data on the private costs incurred in attending for assessment.

RESULTS

There was a reduction in the recall rate from 9.1% (2404 of 26,430) after one view screening to 6.7% (1760 of 26,430) after two view screening. The results also suggest that for every 10,000 women screened an additional five cancers would be detected earlier with two view screening. The additional health service screening cost associated with two view screening was estimated to be 3.63 pounds: the costs associated with one and two view screening policies were estimated to be 41.49 pounds and 32.99 pounds respectively. Private costs incurred were estimated to be 0.35 pounds per woman screened and 32.75 pounds per woman assessed. Two cost effectiveness ratios were calculated: an incremental health service cost per additional cancer detected of 4129 pounds and an incremental health service plus private cost per additional cancer detected of 2742 pounds. The sensitivity analysis suggested that the results were sensitive to relatively large changes in a number of parameters. These included screening costs, assessment costs, equipment life, and recall rates.

CONCLUSIONS

Use of two view screening increased early cancer detection and also costs. The reduction in the recall rate with two views was not sufficiently large to make the cost of two view screening neutral. While these results are not completely generalisable, a framework is provided to allow other centres to estimate the cost effectiveness of two view screening in their locality.

Nursing care of a child with meningococcemia

This case presentation will discuss the pathophysiology of a child in septic shock due to Neisseria meningitidis. The most prevalent nursing care concerns of this case encountered during the pediatric intensive care unit (PICU) and during the general floor stay will be addressed. The nursing skill required for identifying problems and planning care that clearly fall under the nursing domain also will be covered. In addition, the complexities of this case demonstrate that collaboration between the PICU nurse and the general pediatric nurse is imperative for successful patient outcome. A.W. was a 5 1/2-month-old infant transported to our PICU from a referral hospital in the state. Diagnosis at time of admission was meningococcemia, disseminated intravascular coagulopathy, septic shock, respiratory failure, and purpura fulminans. There was a 2- to 3-day history of a runny nose, cough, and vomiting. On the day of admission, A.W. had three seizures and developed a fever and a purpuric rash.

An economic evaluation of transcervical endometrial resection versus abdominal hysterectomy for the treatment of menorrhagia

OBJECTIVE

To evaluate the relative health service cost of endometrial resection versus abdominal hysterectomy for the treatment of menorrhagia and the value women attach to their health state before and after surgery.

DESIGN

A prospective economic evaluation running alongside a randomised controlled trial.

SETTING

The gynaecology department of a teaching hospital.

SUBJECTS

200 women requiring surgical treatment of menorrhagia between January 1990 and May 1991; after withdrawals, 97 women underwent hysterectomy and 99 underwent endometrial resection.

MAIN OUTCOME MEASURES

The total health service cost of managing women in the two arms of the trial until 4 months after their operation. The change in women's valuation of their health state a fortnight after and a minimum of 4 months after surgery relative to that 1 month prior to their operation.

RESULTS

Total health service costs are significantly higher amongst abdominal hysterectomy patients (mean 1059.73 pounds) than amongst endometrial resection patients with a mean difference of 499.68 pounds (95% CI 432 pounds-567 pounds). This significant difference exists under alternative assumptions about the difference in lengths of stay in hospital between the two treatment groups and the hotel cost per in-patient day. On a scale of 0 to 100, relative to a month before surgery, there is a statistically significant difference in favour of endometrial resection between the two groups in the increase in value women attach to their health state at a fortnight after surgery (mean difference 11.2; 95% CI 0.6-21.7), but not at a minimum of 4 months after surgery (mean difference 7; 95% CI -17.4 to 3.4).

CONCLUSIONS

On the basis of health service resource cost up to 4 months after surgery, endometrial resection has a cost advantage over abdominal hysterectomy. However, given the fact that a subgroup of women requires retreatment due to resection failure and that this study considers a relatively short period of follow up, the long term costs and benefits of endometrial resection need to be evaluated before widespread diffusion is justified.

Economic evaluation of treatments for respiratory disease

Considerable quantities of health service resources are being devoted to tackling the problems posed by respiratory diseases and this can be expected to continue as the prevalence of such diseases increases. This paper provides an assessment of the most efficient use of these healthcare resources by reviewing the literature on economic evaluation relating to interventions in the field of respiratory medicine. Currently, this literature largely comprises cost-minimisation studies of both management and educational interventions. Asthma educational interventions, whether targeted at adults or children, appear to be effective in improving patient self-management and adherence to medications, and appear to be associated with a lower overall use of healthcare resources. In terms of management interventions the overall picture is rather less clear although there is some support for the greater efficiency of patient administration of beta-adrenergic agonists by metered dose inhalers over therapist-administered up-draft nebulisation. Two features of respiratory disease make evaluation in this field somewhat unusual: there are alternative methods of delivering therapy to patients, which makes patient compliance an important issue, and since most respiratory diseases are chronic conditions the long-term effectiveness of interventions must be assessed. The scarcity of cost-effectiveness and cost-utility studies in this field may, in part, reflect the difficulties of measuring outcomes in respiratory disease.

Chiropody and the QALY: a case study in assigning categories of disability and distress to patients

Quality adjusted life years (QALYs) are claimed to be a universal means of measuring output from health care interventions. However, existing QALY research has been carried out mainly in 'high-tech', life extending areas of health care. This paper presents an application of QALY measurement to a 'low-tech' life-quality enhancing area of health care, chiropody. Information on changes in quality of life following chiropody interventions was elicited from both practitioners and patients. We found the apparently low benefit, but low cost service of chiropody to be a potentially cost-effective use of NHS resources. Methodological issues are also addressed relating to the assignment of patients to health states, and whether practitioners' or patients' assessments of changes in quality of life should be used.

Complications of pregnancy in infertile couples: routine treatment versus assisted reproduction

Ninety pregnancies conceived by infertile couples using assisted reproductive technologies and 86 pregnancies conceived by infertile couples with routine infertility treatment were analyzed to determine the outcome of and the complications experienced during the pregnancies. Pregnancies ending after 24 weeks' gestation were evaluated for the following complications: pregnancy-induced hypertension, diabetes mellitus, preterm labor, premature rupture of membranes, placenta previa, and fetal growth retardation. A matched control group of normal fertile patients admitted to the obstetric service at Vanderbilt University Medical Center was used to compare the incidence of pregnancy complications among the groups. In the group treated by assisted reproduction, 81 pregnancies were singleton and nine were multiple gestations, whereas in the routine group, 84 were singleton and two were multiple gestations. In the group treated by assisted reproduction, 29 of 90 gestations (32%) ended before 24 weeks, compared with 18 of 86 (21%) in the routine group, a nonsignificant difference. Mean birth weight and gestational age were similar among the three groups for singleton gestations. Among multiple gestations, the mean (+/- standard error of the mean [SEM]) birth weights were 2513 +/- 115, 724 +/- 57, and 2282 +/- 132 g in the group treated by assisted reproduction, the group receiving routine methods, and the control group, respectively (P less than .001 when those treated by routine methods were compared with the other two groups). The mean (+/- SEM) gestational ages were 36 +/- 1.2, 26.5 +/- 2.0, and 35.5 +/- 1.2 weeks, respectively (P less than .01 comparing those treated by routine methods and the other two groups).(ABSTRACT TRUNCATED AT 250 WORDS)

Randomized clinical trial comparing mitoxantrone with doxorubicin in previously treated patients with metastatic breast cancer

Three hundred twenty-five women with metastatic adenocarcinoma of the breast who had failed one prior chemotherapeutic regimen for advanced disease were randomized to receive 14 mg/m2 of mitoxantrone or 75 mg/m2 of doxorubicin intravenously (IV) every 3 weeks. Enrollment was closed on October 31, 1984, after 165 patients were randomized to mitoxantrone and 160 patients to doxorubicin. Patients randomized to the two treatment groups were compared for response rate, duration of response, time to progression or death, time to treatment failure (TTF), and survival. The response rate to mitoxantrone was 20.6%, to doxorubicin 29.3% (P = .07). The median response duration was 151 days for the mitoxantrone group and 126 days for the doxorubicin group (P = .16). The median TTF was 70 days in the mitoxantrone group and 104 days in the doxorubicin group (P = .36). The median survival of patients initially randomized to receive mitoxantrone was 273 days; for doxorubicin 268 days (P = .40). There were three responses among 77 patients crossed over to mitoxantrone after initial treatment with doxorubicin. The major dose-limiting toxicity for both drugs was leukopenia. There was significantly less severe and less frequent toxicity with mitoxantrone administration. Severe nausea and vomiting occurred in 9.5% of mitoxantrone patients and 25.3% of doxorubicin patients (P less than .001). The incidence of severe stomatitis and mucositis was 0.6% in the mitoxantrone group and 8.4% in the doxorubicin group (P = .001). Severe alopecia occurred in 5.1% of mitoxantrone patients and 61.0% of doxorubicin patients (P less than .001). A life-table comparison of the cumulative dose to the development of a cardiac event showed that mitoxantrone had significantly less cardiotoxicity than doxorubicin (P = .0005). This study demonstrates that mitoxantrone is active as a single agent in the treatment of metastatic breast cancer. Compared with doxorubicin it appears to be marginally less active and significantly less toxic. We conclude that mitoxantrone can be used alone or with other standard drugs to palliate the symptoms of metastatic breast cancer, especially in settings where drug toxicity is an important consideration.

Deadenylation of alpha 1-acid glycoprotein mRNA in cultured hepatic cells during stimulation by dexamethasone

During acute phase induction in rats, alpha 1-acid glycoprotein (AGP) mRNA is modified by a reduction in poly(A) tail size (Shiels, B.R., Northemann, W., Gehring, M.R., and Fey, G.H. (1987) J. Biol. Chem. 262, 12826-12831). In the studies presented here, we analyzed AGP mRNA induction and poly(A) tail modification in both primary rat hepatocytes and in the rat hepatoma cell line HTC. Poly(A) tail shortening occurred during stimulation by both glucocorticoids and hepatocyte stimulating factor. Induction of AGP by the glucocorticoid dexamethasone resulted in an approximately 50-fold increase in transcription by 4 h, which was followed by an equally rapid decrease. The large mRNA pool that resulted from this early burst of transcription was very stable, having a half-life of well over 24 h, and the individual molecules in this pool had an average poly(A) tail length of 200-250 bases. This stable pool of AGP mRNAs was then deadenylated to form a pool with an average tail length of 20-50 bases; the time course of this shortening followed that seen in the liver. Ongoing RNA synthesis, but not ongoing protein synthesis, was required for shortening of the tail. In contrast to the conclusion of Shiels et al. (1987), our data indicate that deadenylation occurs in the cytoplasm rather than the nucleus. Our data also suggest that shortening of AGP mRNA represents a specific example of the general deadenylation seen in earlier studies of total cellular RNA.

Alternate pathways of DNA replication in Escherichia coli

We have described the pcbA1 mutation which enables E. coli cells to replicate DNA in the absence of a functional dnaE gene product if DNA polymerase I (the polA gene product) is present. The pcbA1 mutation phenotypically suppresses multiple dnaEts and dnaEam alleles. The pcbA1/PolI replication pathway differs from normal in sensitivity to certain DNA-damaging agents such as methylmethane sulfonate (MMS) and a lack of damage-directed mutagenesis. We report here cloning of the pcbA1 gene in a multicopy plasmid. The pcbA1 mutation is detected only in cis; therefore, cloning necessitated gene eviction. The pcbA1 gene lies closely- linked to gyrB. We have demonstrated the physical presence of DNA polymerase I in the replicating holoenzyme complex by immunoblotting using dnaEam strains. We conclude that E. coli has two alternate replisome structures: REP-A, in which DNA polymerase I is the functional synthetic subunit; and REP-E, in which the alpha-subunit, product of the dnaE gene, is functional. To investigate further the role of individual DNA polymerases in replication, we have isolated the polB gene on multicopy plasmids.

A randomized trial comparing mitoxantrone with doxorubicin in patients with stage IV breast cancer

Mitoxantrone (Novantrone; dihydroxyanthracenedione) is an anthraquinone previously shown to be active in human breast cancer. It appears to have less toxicity than doxorubicin. Results of this phase II-III randomized cross-over trial to determine the relative efficacy and toxicity of mitoxantrone in comparison to doxorubicin, are presented. Patients with measurable, recurrent breast cancer with limited prior chemotherapy with or without radiotherapy for metastatic disease, and who had not been exposed to prior doxorubicin, were randomized to receive either mitoxantrone or doxorubicin every three weeks with cross-over on progression. Response rates, duration of remission, time to treatment failure, and drug toxicity, including cardiac toxicity evaluated with serial radionuclide angiocardiography, were evaluated. Differences in the response rates for the two groups were not statistically significant. Neither time to treatment failure nor duration of response are significantly different (p greater than 0.05). With respect to toxicity, mitoxantrone treated patients consistently exhibited a lower incidence and less severe drug toxicity as compared to their doxorubicin-treated counterparts. Cardiac toxicity was carefully monitored and thus four patients on doxorubicin have had drug related congestive heart failure, as compared to none on mitoxantrone. In summary, mitoxantrone appears to be as active as doxorubicin in patients with stage IV breast cancer previously treated with chemotherapy; however, mitoxantrone causes significantly less nausea, vomiting, stomatitis and alopecia at doses which induce equal or greater myelosuppression than doxorubicin, and appears to be less cardiotoxic.

Mitoxantrone in malignant lymphoma

Two phase II trials of mitoxantrone (Novantrone; dihydroxyanthracenedione) in refractory malignant lymphoma have been conducted. In the first of these, mitoxantrone, 5 mg/m2, was given weekly for six weeks and in the second, 14 mg/m2 was administered every three weeks. The first trial was conducted by the Southeastern Cancer Study Group (SECSG) and the second was a multicenter trial sponsored by Lederle Laboratories. Of the 51 patients entered in the SECSG trial, 28 could be evaluated for response and 43 for toxicity. WBC nadirs below 4.0 X 10(9)/litre were recorded in 25 patients. Three partial responses and no complete responses were obtained. These results contrast with those of the single dose every three weeks study in which 96 patients were entered and 69 of these were evaluated for response. Responses were obtained in 30 patients (4 complete, 26 partial). Side-effects on this three-weekly dose regimen were minimal. WBC nadirs below 4.0 X 10(9)/litre occurred in 85 patients. Twenty-three patients experienced at least mild nausea and vomiting and 15 had at least mild alopecia. These preliminary data indicate that mitoxantrone has significant activity in malignant lymphoma. All of the responding patients had received extensive prior therapy, many of them with anthracyclines in combination or as single agents. The higher response rate to mitoxantrone given at 14 mg/m2 every three weeks suggests that careful consideration should be given to dose schedule when this drug is examined further in phase III trials.

Studies on two novel angiotensin II actions

Animal data suggest that angiotensin II may directly affect renal sodium retention independent of aldosterone (4,6). Additionally there is evidence to suggest that the hormone can stimulate protein synthesis in a variety of tissues and that indeed it may be a vasculotoxin (5,7). We describe here experiments designed to elucidate the role of angiotensin II in renal sodium retention in normal man. Additionally we present preliminary evidence suggesting that unlike the rapid sodium retaining effect, some delayed (trophic) actions of the hormone may be generated by a hitherto unappreciated mechanism.

Effect of oral contraceptives on rhythmic urinary metal excretion

Thirteen women, six of whom were using oral contraceptive agents, between the ages of 23 and 34 years, free from illness, and without a previous clinical history of menstrual abnormality, were studied with chronobiological methods for circadian (about 24 hour) patterns of metal and creatinine excretinine excretion. Circadian rhythms were detected for the excretion of lithium, potassium, iron, sodium, copper, and creatinine. Statistically significant by lower 24 hour excretion (mesors) of lithium, potassium, and iron was detected for the subjects using oral contraceptive agents. Although all women adhered to nearly identical activity-rest (synchronization) schedules, the peak times (acrophases) of the metal excretory rhythms for oral contraceptive users were phase delayed by about two hours.