[Evaluation of the effects of a complex intervention]

Appraising the effects of a complex intervention is one step in a more broadly based research process, from the construction or modelling of the intervention to its actual deployment. It consists in measuring the effectiveness or impact of the intervention, i.e. analyzing its capacity to produce change. The aim of this article is to obtain unbiased measurement of the average effects of an intervention, based on a panel of predetermined parameters and on the assumption of a causal link between the intervention and the measured result. This article is consequently devoted to evaluation of the effects of a complex intervention and focuses on the methodological challenges of its three key stages : 1) modelling of the intervention and feasibility analysis, essential prerequisites ; 2) the choice of study design and of the effects to be measured, that is to say the methodological premises ; and 3) process analysis, carried out in parallel with the evaluation of effects, leading to an indispensable appraisal of the intervention implementation and of the context into which it is integrated. The article is illustrated by five intervention impact assessment projects. A specific objective when evaluating the effects of a complex intervention consists in (a) moving away from a simple search for causality involving the intervention and its effects and (b) toward understanding of the effectiveness mechanisms, once again taking into account the context and the actual conditions of implementation. The challenge is to embrace rather than limit the complexity of the intervention, this being an essential prerequisite for its successful deployment and eventual generalization.

Clinical Prediction of Lower Limb Deep Vein Thrombosis in Symptomatic Hospitalized Patients

We evaluated two clinical scores for the prediction of deep venous thrombosis (DVT) in hospitalized patients (Wells’ and Kahn’s). We included 273 patients referred to the vascular exploration unit for the suspicion of DVT. A clinical questionnaire was filled in by the practitioner and the scores were calculated from this form. 66 of the 273 patients had a DVT. When Wells’ score was 3, a DVT was found by duplex echography in 51% patients ; when the score was 0, a DVT was found in 9%. Kahn’s score was not adapted to this population. We then developed a new simple score (cancer, palsy or plaster immobilization, warmth, superficial venous dilation, unilateral pitting edema, other diagnosis). A DVT was found in 76% patients with a score of 3 and in 11% in those with a score of 0. We therefore propose a 6-item score whose main advantages are simplicity and usefulness in routine practice.

The recent dramatic decline in road mortality in France: how drivers' attitudes towards road traffic safety changed between 2001 and 2004 in the GAZEL cohort

A very significant decline in the number of road casualties has been observed recently in France, concomitantly with a dramatic increase in law enforcement. The aim of this study was (i) to assess changes in attitudes about road traffic accident (RTA) prevention initiatives in France from 2001 to 2004 and (ii) to identify factors associated with an increase in positive attitudes towards RTA prevention initiatives. In 2001 and 2004, 9216 participants reported their attitudes towards traffic safety using the same self-administered Driving Behaviour and Road Safety Questionnaire. Sociodemographic, psychological and behavioural data were also available. The mean change in scores analysis showed that support for relaxing existing regulations decreased significantly during this period, while support for heightened enforcement and stricter regulations showed some decrease but remained high overall, especially concerning blood alcohol content and speed controls. Multivariate analyses suggest that highly educated drivers changed their attitudes towards road safety regulations more than other categories. Our results suggest that increased traffic law enforcement measures led to increasing support for current restrictions. Even if support for additional traffic law enforcement began to wane slightly in 2004, a large part of our population remained in favour of strengthening law enforcement related to speeding and drunk driving.

Survey of European programmes for the epidemiological surveillance of congenital toxoplasmosis

The objective of this investigation was to describe systems for the epidemiological surveillance of congenital toxoplasmosis implemented in European countries. In September 2004, a questionnaire, adapted from the evaluation criteria published by the United States Centers for Disease Control and Prevention, was sent to a panel of national correspondents in 35 countries in the European geographical area with knowledge of the epidemiological surveillance systems implemented in their countries. Where necessary, we updated the information until July 2007. Responses were received from 28 countries. Some 16 countries reported routine surveillance for toxoplasmosis. In 12 countries (Bulgaria, Cyprus, Czech Republic, England and Wales, Estonia, Ireland, Latvia, Lithuania, Malta, Poland, Scotland and Slovakia), surveillance was designed to detect only symptomatic toxoplasmosis, whether congenital or not. Four countries reported surveillance of congenital toxoplasmosis, on a regional basis in Italy and on a national basis in Denmark, France and Germany. In conclusion, epidemiological surveillance of congenital toxoplasmosis needs to be improved in order to determine the true burden of disease and to assess the effectiveness of and the need for existing prevention programmes.

[The fate of the abstracts presented to the biannual meetings of the French National Society of Internal Medicine]

PURPOSE

Medical meetings give the opportunity to present oral communications or posters to the attending participants. However, the peer-reviewed publication of a full article allows to reach a wide readership.

KEY POINTS

The survey that was performed on the oral communications and posters presented at the 43rd meeting of the French National Society of Internal Medicine, December 2000, showed that amongst the 303 selected podium presentations and posters, only 82 (27%) were published during the five following years. Podium presentations were more likely to be published than posters (36% versus 22%).

CONCLUSION

Many oral communications and posters that are presented in medical meeting are not followed by the publication of a peer-reviewed full article despite the modern means of communication. However, this issue is of paramount importance as beyond the legitimate personal satisfaction of a publication, the scientific and academic recognition are the ground of medical career achievement for many physicians.

Reliability of the pulp nail bone (PNB) classification for fingertip injuries

Evans and Bernardis proposed the 'PNB classification', by which a fingertip injury is classified according to each structure: pulp P, nail N, bone B. The objective of this study was to assess the inter-observer reliability, repeatability and accuracy of PNB. One hundred patients presenting with a fingertip injury were included prospectively, photographed, then classified in randomly chosen orders by nine independent observers. A third were drawn randomly and classified a second time to measure repeatability. A reference classification was also provided by one of the authors of the PNB system. Classifications agreed with the reference in 59% of injuries for P, 55% for N and 54% for B. The Kappa values for inter-observer agreement were 0.520 for P, 0.512 for N, and 0.504 for B; for intra-observer agreement, they were 0.616 for P, 0.658 for N, and 0.577 for B. Although levels of agreement are comparable with results found for other classifications, they are insufficient for use of the PNB classification without improvement.

Metabolic syndrome components in Indian migrants with type 2 diabetes. A matched comparative study

OBJECTIVE

To study the metabolic syndrome (MS) in Indian subjects with type 2 diabetes (T2D) in comparing them with controls from the Indian community and from the general population.

METHOD

An adapted definition of MS by the Third report of the National Cholesterol Education Program's Adult Treatment Panel III was used. We defined three groups matched for sex and age (+/-5 years). Non parametric tests for comparison of matched samples and conditional logistic regression were used.

RESULTS

We selected 71 Indians with T2D (group 1) and two control groups with fasting blood glucose<6.1 mmol/L: 71 Indians (group 2) and 213 subjects from the general population (group 3). Patients were 24 to 76 years-old and each group contained 56% men. Globally, MS was identified in 77% of the group 1 when diabetes was taken into account. When diabetes was excluded there were 47% of MS in group 1, 18% in group 2 and 16% in group 3. The clusters of four factors (hypertension, large waist circumference, hypertriglyceridemia and Low HDL-C) were more common in Indians. The most frequent factors were hypertriglyceridemia and large waist circumference in Indians. Indians with T2D had a 5-fold higher risk of MS than the general population group, OR (95% CI): 4.93 (2.71 - 8.97); P<0.001.

CONCLUSION

The high frequency of MS and of hypertriglyceridemia in Indians with T2D highlights the need for screening and management of MS in this population facing a high cardiovascular risk.

Pathways to care of first-admitted subjects with psychosis in South-Western France

BACKGROUND

A limited number of studies have assessed the pathways to care of patients with first-episode psychosis. The aim of the study was to describe the pathways to care of subjects with psychosis between onset of psychosis and first admission, and to examine the demographic and clinical factors influencing access to care.

METHOD

Number and type of helping contacts since onset of first psychotic symptoms were assessed using multiple sources of information in 86 subjects with psychosis first-admitted in two hospitals of South-Western France. Characteristics independently associated with long delays between onset of symptoms and first helping contact, first treatment and first admission were explored using logistic regressions.

RESULTS

Twelve per cent of subjects were first admitted without any previous helping contact. The patients were seen by a median of two helpers (maximum 7). For most patients (70%), the first helping contact was a health care professional, and the same proportion of patients had a first contact with a GP or a psychiatrist. The type of first contact was not predicted by demographic or clinical characteristics. Subjects with poor pre-morbid functioning or at-risk behaviour were more likely to have a delayed access to care.

CONCLUSIONS

The delay in access to care may not be totally attributed to inadequate management by health professionals, but may be a characteristic of the disease itself, at least in part independent of the organization of the health system.

Multivariate analysis of determinants of bacterial contamination of whole-blood donations

BACKGROUND AND OBJECTIVES

Introduction of bacteria into blood components at the collection stage seems to be a frequent occurrence. We therefore assessed determinants of bacterial contamination of whole-blood donations to gain insight into contamination mechanisms and direct prevention.

MATERIALS AND METHODS

A cross-sectional study was carried out on donors accepted for whole-blood donation in four French blood banks. Each blood bank used its own two-stage procedure for phlebotomy site preparation. Contamination was identified by culturing two 15-ml samples (collected aseptically at the outset of donation) in a BacT/Alert 240 system. Determinants were assessed by logistic regression analysis.

RESULTS

Bacterial contamination, mainly by skin flora, occurred in 76 (2.2%) out of 3385 donations. Significant determinants were as follows: the blood bank (odds ratio [OR] range = 3.0-5.6, P < 0.001); lack of repetition of scrub (OR = 2.7, P = 0.032); and donor age > 35 years (OR = 1.8, P = 0.036).

CONCLUSION

Systematic scrub repetition should be implemented to reduce bacterial contamination by skin flora at the collection stage. Further research is required to clarify the role of different antiseptic agents and of donor age.

Factors associated with nurses' poor knowledge and practice of transfusion safety procedures in Aquitaine, France

OBJECTIVES

To (1) describe knowledge, attitudes, and reported practice of blood transfusion of nurses in Aquitaine's hospitals; (2) measure the potential threat for patient safety of poor transfusion-related knowledge and practice; and (3) identify factors associated with poor knowledge and practice.

DESIGN

Survey conducted in 14 hospitals in Aquitaine (one university and 13 general hospitals).

SETTING

Hospitalized care.

PARTICIPANTS

A random sample of nurses.

MATERIALS

Data were collected anonymously by investigators through structured individual interviews. The questionnaire contained mainly knowledge and practice questions about blood transfusion regulation.

MAIN OUTCOME MEASURES

Hazardous knowledge and practice scores have been constructed, reflecting the levels of potential danger in the answers to the questionnaire. Factors associated with these scores have been studied using a random-effect linear regression.

RESULTS

In our sample of 1090 nurses, poor knowledge and practice concerned mainly (1) the bedside blood compatibility test [proportion of responses (PR) with potential life threat between 12.7 and 35.5%]; (2) pre-transfusion compatibility check when receiving blood units (PR = 34.5%); (3) delay between screening of red cell antibodies and transfusion (PR = 20.5%); (4) delay in preservation of blood unit in the ward (PR = 33.4%); and (5) recognition of abnormal reactions after transfusion (PR = 47.1%). Frequency of transfusion and training were the factors most strongly associated with hazardous knowledge and practice scores.

CONCLUSION

Low training and transfusion activity were key determinants of poor transfusion-related knowledge and practice.

Clinical prediction of lower limb deep vein thrombosis in symptomatic hospitalized patients

We evaluated two clinical scores for the prediction of deep venous thrombosis (DVT) in hospitalized patients (Wells' and Kahn's). We included 273 patients referred to the vascular exploration unit for the suspicion of DVT. A clinical questionnaire was tilled in by the practitioner and the scores were calculated from this form. 66 of the 273 patients had a DVT. When Wells' score was 3, a DVT was found by duplex echography in 51% patients; when the score was 0, a DVT was found in 9%. Kahn's score was not adapted to this population. We then developed a new simple score (cancer, palsy or plaster immobilization, warmth, superficial venous dilation, unilateral pitting edema, other diagnosis). A DVT was found in 76% patients with a score of 3 and in 11% in those with a score of 0. We therefore propose a 6-item score whose main advantages are simplicity and usefulness in routine practice.

[Evaluation of the immediate transfusion reaction incident reporting system at the Brest University Hospital Center]

The Haemovigilance Unit of Brest University Hospital has had a reporting system of transfusion reactions since october 1994. Reporting "any unexpected or undesirable effect due or likely to be due to the administering of blood cell components" must be done on an answering machine immediately or in the next eight hours. The main goal of the evaluation of this epidemiological surveillance system was to assess its sensitivity, its positive predictive value, its acceptability, its timeliness and its simplicity, according to the Centers for Disease Control criteria. An exhaustive monitoring of the immediate transfusion reactions (ITR) occurring within the 24 hours following the procedure was conducted from April 1, to June 30, 1998. Two sources of information were used, the spontaneous notification to the Haemovigilance Unit using the answering machine, and a telephone survey of the nurse responsible for the transfusion or post-transfusion follow-up. During the survey, 19 ITR, among which 12 were reported to the Haemovigilance Unit on the answering machine, were recorded. The incidence rate of the I.T.R. was estimated at 5@1000 transfused blood cell components. The sensitivity of the notification system was estimated at 63% (95% confidence interval: 41-85) and the positive predictive value at 70% (95% confidence interval: 48-92). This notification system is operational. The function of sanitary alert is ensured at the primary level of the system surveillance. The undernotification of the ITR (37% of false negative) must be corrected by specific recommendations.

Determinants of transfusion-associated bacterial contamination: results of the French BACTHEM Case-Control Study

BACKGROUND

Transfusion-associated bacterial contamination (TABC), probably the most frequent transfusion-transmitted infection, may induce serious adverse events. Systematic information and documentation on determinants are lacking.

STUDY DESIGN AND METHODS

The BACTHEM Study is a French matched case-control study assessing TABC determinants. Included were cases of TABC reported in France in a 2-year period, as determined from uniform definitions. Information on recipient-, blood component-, and donor-related potential determinants was collected on site. ORs were estimated by conditional logistic regression.

RESULTS

Of the 158 cases of suspected TABC reported, 41 that involved transfusion with 25 RBCs and 16 platelet concentrates were included. Gram-negative rods accounted for nearly half of the bacteria species involved and for all six deaths. In comparison with the risk of TABC for patients receiving RBCs for anemia, the risk was higher for patients receiving RBCs for pancytopenia (OR, 7.3; 95% CI, 1.3-41.0) and for those receiving platelets for thrombocytopenia (OR, 5.3; 95% CI, 1.2-24.1). Other potential determinants were platelet transfusion for pancytopenia (OR, 4.5; 95% CI, 0.5-40.0), immunosuppressive treatment (OR, 2.8; 95% CI, 0.7-10.6), shelf-life of more than 1 day for platelets or 8 days for RBCs (OR, 2.6; 95% CI, 0.7-9.6), and more than 20 previous donations by donors (OR, 1.9; 95% CI, 0.7-5.3).

CONCLUSION

This first comparative study revealed TABC determinants that suggest approaches for prevention.

What should be included in meta-analyses? An exploration of methodological issues using the ISPOT meta-analyses

OBJECTIVE

To explore the impact of methodologic issues on the results of meta-analyses. The following issues were examined: the type of literature search strategy used; inclusion or exclusion of non-peer-reviewed studies; the inclusion or exclusion of non-English language publications; the effect of trial quality; and the inclusion or exclusion of non-placebo-controlled studies.

METHODS

The International Study of Perioperative Transfusion (ISPOT) meta-analyses were used to evaluate each of the methodologic issues. The 10 meta-analyses consisted of technologies to reduce the need for perioperative red blood cell transfusion. The number of trials for each of the meta-analyses varied from 2 to 45. Both EMBASE and MEDLINE searches were conducted, including the use of systematic search strategies.

RESULTS

MEDLINE identified the vast majority of trials. Alone, MEDLINE would have missed 8 studies compared to 10 for EMBASE. Use of the systematic search strategies greatly reduced the number of articles to be reviewed compared to open searches. Type of publication, country of study origin, inclusion of non-English publications, and trial quality had very little impact on the estimates of effect. The use of placebo versus open-label control affected the magnitude of the odds ratio for two of the meta-analyses. The results of the two meta-analyses were not statistically significant if only placebo-controlled trials were included.

CONCLUSIONS

While methodologic issues had very little impact on the ISPOT meta-analyses, further studies are needed in a variety of other clinical settings. Because MEDLINE, coupled with a review of the references in the identified trials, identified the vast majority of trials, one needs to consider the costs and benefits of searching EMBASE and the pursuance of unpublished and unindexed trials.

Publication pattern of medical theses, France, 1993-98

CONTEXT AND OBJECTIVE

In many countries, medical training must be completed by presentation of a thesis. We report publication patterns of French medical theses.

MATERIALS

We drew a random sample of theses presented in the 36 French medical universities between 1 January 1993 and 31 December 1997. The sample was stratified according to size of university, and drawn from the CD-ROM Doc-Thèse.

METHODS

We recorded the research area (medicine, surgery, biology) and study type (clinical, epidemiological, laboratory). We used the name of the student and supervisor to assess whether the thesis resulted in a publication indexed in MEDLINE.

RESULTS

Most of the 300 theses included were from medicine (79.3%) and were clinical studies (69.3%). A total of 51 theses (17. 0%) resulted in publication. The proportion of theses which were published, the median impact factor of the journals in which they were published, the proportion of publications in English and the proportion of publications in which the name of the student was missing varied with the research domain and type of study.

CONCLUSIONS

Most French medical theses are not made available to the scientific community. In the European context, where medical training and qualification have to be standardized, our study provides a simple method of assessing that publication objectives of thesis research are met. Further research is needed to explore the educational value of medical professional theses.

Efficacy of a new collection procedure for preventing bacterial contamination of whole-blood donations

BACKGROUND

Transfusion-related bacterial contamination is a serious problem. The introduction of bacteria into donations at the collection stage seems frequent, despite well-conducted phlebotomy site preparation. Additional preventive measures are required.

STUDY DESIGN AND METHODS

The aim of this study was to assess the potential efficacy of excluding the first 15 mL of blood to reduce the bacterial contamination of donations. A special device allowed the aseptic collection of two samples at the beginning of donation: S1 (first 15 mL) and S2 (next 15 mL). Bacteriologic cultures of S1 and S2 were performed by using an automated system. The procedure's efficacy was measured by the proportion of positive donations in S1 that were then negative in S2.

RESULTS

S1 and/or S2 were positive in 76 (2.2%) of 3385 donations. In about three-fourths of the culture-positive donations, contamination was detected in the first 15-mL sample only. Gram-positive cocci accounted for 81 percent of species, gram-positive bacilli for 14 percent, and gram-negative bacilli for 5 percent. The new procedure would have prevented the introduction of bacteria in 55 donations, reducing to 0.6 percent the risk of contamination from the first 15 mL collected.

CONCLUSION

Although the final effect on blood component bacterial contamination rates cannot be derived from the study, excluding the first 15 mL of blood may reduce the rate of bacterial contamination in donations.

[Which health information systems for the evaluation of health actions?]

The evaluation of health actions requires large amounts of information allowing an assessment of relevance, coherence, efficacy, efficiency and impact of these interventions. Information systems should support evaluation processes by helping to obtain pertinent indicators, tracers and standard operating procedures. To reach this objective, specific functions need to be implemented, including traceability, documentation, investigation and scientific awareness. But this supposes that health information systems respect quality criteria that this article defines. The following criteria are considered: simplicity, validity, acceptability, informative value, representativeness, continuity, reactivity, fluidity, flexibility of the system, and also confidentiality of information. The historical developments of medical informatics have induced the creation of independent information systems, answering to specific objectives. This lack of integration is an obstacle to the evaluation of health actions because of the difficulty to view transversally and longitudinally the sequence of actions for a same patient. Thus integrated health information systems, organised around patient care episodes, are necessary to support evaluation and to contribute effectively to decision making in public health, because the evaluation of health actions implies the availability of information about the individuals who are the targets of these actions.

[Process assessment: importance and current limits]

Assessment of processes in health care is needed for the measurement of six domains of quality of care (appropriateness, security, respect and caring, availability, continuity and timeliness). We propose here a typology of care, organisational and management processes. Process indicators can be classified according to objectives of the evaluation, type of processes, concerned healthcare professionals, and stage of the process. Process assessment is hindered by two main difficulties. Due to the lack of data about processes in information systems and in medical records, specific data collection is often needed and impede the feasibility of routine evaluation. The second difficulty is related to the lack of information about the properties of the indicators, their appropriateness (capacity to infer improvement actions from the result of the indicators), their operational properties (cost of data collection and acceptability by the concerned professionals), and their measurement properties (reliability and validity). Process assessment is strongly needed because outcome assessment can be not possible or not appropriate. The lack of information about the process-outcome relationship is the major obstacle and the main contra-indication to its use. Validation of process indicators is therefore an important research area for the future.

Blood transfusion in a random sample of hospitals in France

BACKGROUND

Representative information on blood use is scarce. A large-scale study of blood recipients and blood use in France was conducted.

STUDY DESIGN AND METHODS

Based on a random sampling, this study was carried out in teaching and other hospitals between March and December 1997. In each hospital, a patient was included if he or she received an allogeneic or an autologous transfusion during the observation period for that hospital. For each recipient, product and patient characteristics for 24 hours after inclusion were collected.

RESULTS

From the 175 hospitals that had given a transfusion to at least one patient during the observation period, 3206 patients were included. Most transfusion recipients (57%) were over 65 years old; 42 percent were in teaching hospitals and 53 percent in medical wards. Among the 3044 adults, 91 percent received an allogeneic transfusion. Fifty-three percent of allogeneic units were WBC reduced. The indications most frequently reported for allogeneic transfusion were neoplasms (48%) and those for autologous transfusion were disorders of musculoskeletal (63%) or circulatory (15%) systems. The patients in nonteaching hospitals were more often transfused during surgery and were more likely to be aged and to have a musculoskeletal disorder than were patients in teaching hospitals.

CONCLUSION

General collection of such data, within a system of traceability, could provide relevant denominators from which to interpret adverse-reaction data.

Errors in interpreting the pretransfusion bedside compatibility test: an experimental study

BACKGROUND AND OBJECTIVE

Analysis of reports of incidents, involving ABO incompatibility suggests that the main problem is poor interpretation of the pretransfusion bedside compatibility test (PBCT). We studied sources of error as experienced by nurses as to the blood groups of donor blood and of the recipient.

MATERIALS AND METHODS

According to their seniority in the profession and on the ward, 48 nurses were randomly selected from four transfusion sectors of the University Hospital of Grenoble, France. Each nurse interpreted 24 photos of PBCTs, including some with procedural irregularities, and was asked to assess the compatibility of the blood types of the donor and the recipient. At random, half the nurses were provided with a diagram to facilitate interpretation.

RESULTS

The overall frequency of errors was 39.8%. Errors were fewer when the tests were interpreted as compatible (7.3%) or incompatible (6.3%), and when the nurse had been in the profession between 3 and 5 years and in the ward less than 3 years (25.5%), or worked in hematology (34.7%) or anesthesia (36.5%). Use of the diagram limited the number of errors, provided the test was interpretable (22.2%).

CONCLUSION

PBCTs cannot be considered a valid safety procedure. We need other, more effective methods to reduce the risk of incompatibility accidents.

Capture-recapture method to determine the best design of a surveillance system. Application to a thyroid cancer registry

The capture-recapture method is often confronted, when assessing completeness of surveys, to problems of dependence of data sources. The objective of this paper is to discuss the application of capture-recapture methods to choose the optimal combination of sources for a surveillance system. Our approach is based on: (1) using multiple sources, (2) assessing dependence between sources and between pools of dependent sources and other sources, (3) ruling out combinations that yield biased estimates, and (4) choosing the combinations of sources that have the best ratio between precision and applicability. We studied the independence for each pair of sources by computing the capture-recapture odds ratio. We characterized all combinations of sources by their sensitivity, coefficient of variation of the estimated number of cases, and level of resources needed to ascertain cases. Application of the approach is illustrated by data from a survey of thyroid cancer in New Caledonia, where five sources were used to estimate the incidence. The five sources provided 119 cases; the exhaustivity of sources and combinations of sources varied from 27.1 to 99.2%. Determination of dependence revealed ten dependencies out of 22 combinations. Coefficients of variation of the estimated number of cases varied from 0.83 to 27.79. The preferred combination included four sources and had a sensitivity of 97.5 and a coefficient of variation of 0.94. An assessment of dependence, based on simple criteria, can be used to choose the best combination of sources for a registry or a surveillance system.

[Effect of dapsone on survival in HIV infected patients: a meta- analysis of finished trials]

BACKGROUND

The aim of the study was to estimate the effect of dapsone on survival in HIVinfected patients.

METHOD

The method was a metaanalysis. Data searches used MEDLINE, AIDS TRIALS, and AIDS DRUGS databases from 1983 to January 1996, clinical trials registries of appropriate collaborative research groups, abstract books of International Conferences on AIDS and infectious diseases between 1988 and 1996, references listed within selected articles and active experts in HIV infection. Were considered as eligible: randomized clinical trials, conducted in adults, with one arm evaluating dapsone as prophylactic agent for Pneumocystis Carinii Pneumonia (PCP). Each primary investigator was asked to provide the most recent aggregated study data by completing a standardized questionnaire and to provide files of individual patient data whenever possible.

RESULTS

Overall, 17 trials (4343 patients) were eligible for the metaanalysis. The analysis of all available aggregated data included 16 trials (4267 patients) and showed no deleterious effect of dapsone on survival: OR=1.11, 95% Confidence Interval (CI)=0.961.29. There was no evidence of heterogeneity among studies (p=0.50). The analysis of individual data included 10 trials (3115 patients) (OR for aggregated data from those trials=1.10, CI=0. 931.29) and confirmed the absence of deleterious effect of dapsone on survival: stratified Hazard Ratio=1.12, CI=0.991.27 (logrank test: p=0.08). In this subsample, there was evidence of a deleterious effect of dapsone used as secondary prophylaxis. However, this result did not remain when the trial reporting the greatest negative effect of dapsone on survival was omitted.

CONCLUSION

Dapsone may be used safely as a primary prophylactic regimen for PCP or toxoplasmosis. However, no definitive recommendation can be made for the use of dapsone as secondary PCP prophylaxis.

[Writing and publication of a medical article]

To advance in their strategies to manage patients, clinicians need new research results. To be accessible, medical research must be published. Writing and publishing medical articles should respect principles that are described in this article. Good writing is based on a logical organization and the application of scientific style. Organization according to the IMRD structure (Introduction, Methods, Results, Discussion) allows one to present the reasons for and objectives of the study (Introduction), details on whatever has been done to answer the question (Methods), data on the actual study population and answers to the main question (Results), and a critical appraisal of these results, given the limits of the study and current knowledge (Discussion). The main elements of scientific style are precision, clarity, fluidity and concision. Finally, submitting a paper to a scientific journal implies presenting the work in a covering letter and respecting rules for formatting a manuscript (order of presentation, typography, etc.).

A meta-analysis of the effectiveness of cell salvage to minimize perioperative allogeneic blood transfusion in cardiac and orthopedic surgery. International Study of Perioperative Transfusion (ISPOT) Investigators

Concern about risks of allogeneic transfusion has led to an interest in methods for decreasing perioperative transfusion. To determine whether cell salvage reduces patient exposure to allogeneic blood, we performed meta-analyses of randomized trials, evaluating the effectiveness and safety of cell salvage in cardiac or orthopedic elective surgery. The primary outcome was the proportion of patients who received at least one perioperative allogeneic red cell transfusion. Twenty-seven studies were included in the meta-analyses. Cell salvage devices that do not wash salvaged blood were marginally effective in cardiac surgery patients when used postoperatively (relative risk [RR] = 0.85, 95% confidence interval [CI] = 0.79-0.92). Devices that wash or do not wash salvaged blood considerably decreased the proportion of orthopedic surgery patients who received allogeneic transfusion (RR = 0.39, 95% CI = 0.30-0.51 and RR = 0.35, 95% CI 0.26-0.46, respectively). No studies of cell savers that wash salvaged blood during cardiac surgery were included. Cell salvage did not appear to increase the frequency of adverse events. We conclude that cell salvage in orthopedic surgery decreases the risk of patients' exposure to allogeneic blood transfusion perioperatively. Postoperative cell salvage in cardiac surgery, with devices that do not wash the salvaged blood, is only marginally effective.

IMPLICATIONS

This meta-analysis of all published randomized trials provides the best current estimate of the effectiveness of cell salvage and is useful in guiding clinical practice. We conclude that cell salvage in orthopedic surgery decreases the proportion of patients requiring allogeneic blood transfusion perioperatively, but postoperative cell salvage is only marginally effective in cardiac surgery.

Assessing the quality of randomized trials: reliability of the Jadad scale

An instrument was developed and validated by Jadad, et al. to assess the quality of clinical trials using studies from the pain literature. Our study determined the reliability of the Jadad scale and the effect of blinding on interrater agreement in another group of primary studies. Four raters independently assessed blinded and unblinded versions of 76 randomized trials. Interrater agreement was calculated among combinations of four raters for blinded and unblinded versions of the studies. A 4 x 2 x 2 repeated measures design was employed to evaluate the effect of blinding. The interrater agreement for the Jadad scale was poor (kappa 0.37 to 0.39), but agreement improved substantially (kappa 0.53 to 0.59) with removal of the third item (an explanation of withdrawals). Blinding did not significantly affect the Jadad scale scores. A more precise description of how to score the withdrawal item and careful conduct of a practice set of articles might improve interrater agreement. In contrast with the conclusions reached by Jadad, we were unable to demonstrate a significant effect of blinding on the quality scores.

A meta-analysis of the effectiveness of cell salvage to minimize perioperative allogeneic blood transfusion in cardiac and orthopedic surgery. International Study of Perioperative Transfusion (ISPOT) Investigators

Concern about risks of allogeneic transfusion has led to an interest in methods for decreasing perioperative transfusion. To determine whether cell salvage reduces patient exposure to allogeneic blood, we performed meta-analyses of randomized trials, evaluating the effectiveness and safety of cell salvage in cardiac or orthopedic elective surgery. The primary outcome was the proportion of patients who received at least one perioperative allogeneic red cell transfusion. Twenty-seven studies were included in the meta-analyses. Cell salvage devices that do not wash salvaged blood were marginally effective in cardiac surgery patients when used postoperatively (relative risk [RR] = 0.85, 95% confidence interval [CI] = 0.79-0.92). Devices that wash or do not wash salvaged blood considerably decreased the proportion of orthopedic surgery patients who received allogeneic transfusion (RR = 0.39, 95% CI = 0.30-0.51 and RR = 0.35, 95% CI 0.26-0.46, respectively). No studies of cell savers that wash salvaged blood during cardiac surgery were included. Cell salvage did not appear to increase the frequency of adverse events. We conclude that cell salvage in orthopedic surgery decreases the risk of patients' exposure to allogeneic blood transfusion perioperatively. Postoperative cell salvage in cardiac surgery, with devices that do not wash the salvaged blood, is only marginally effective.

IMPLICATIONS

This meta-analysis of all published randomized trials provides the best current estimate of the effectiveness of cell salvage and is useful in guiding clinical practice. We conclude that cell salvage in orthopedic surgery decreases the proportion of patients requiring allogeneic blood transfusion perioperatively, but postoperative cell salvage is only marginally effective in cardiac surgery.

[Arterial blood pressure in homozygote patients with drepanocytosis]

BACKGROUND

Relative hypotension has been reported in sickle cell patients. The aim of this study was to compare blood pressure in patients with SS disease and subjects with normal hemoglobin genotype AA and to assess whether the same clinical, biological and socio-demographic variables are associated to the mean arterial pressure in patients with sickle cell disease and normal subjects.

METHOD

Blood pressure was measured with a standardized automated oscillometric method in 88 SS patients et 88 AA control subjects seen in the University Hospital of Pointe-à-Pitre (Guadeloupe). A multiple linear regression analysis for mean arterial pressure was done including type of hemoglobin (forced variable), age, sex, body mass index, pulse rate, hemoglobin concentration and interaction terms between type of hemoglobin and other variables. A regression was also fitted separately for each population. A downward stepwise strategy was used to simplify the models.

RESULTS

The two groups were similar for age, height and gender ratio and pulse rate. Mean arterial pressure was significantly lower in sickle cell patients (81.6 mmHg in SS patients vs 89.9 mmHg in AA subjects, p < 10(-4)). The final model included type of hemoglobin, age, sex, body mass index, pulse rate and an interaction between type of hemoglobin and age (global F = 22.04, adjusted R2 = 42%). The separate models indicated that sex was associated with mean arterial pressure only in patients with sickle cell disease and that age and hemoglobin concentration was associated with mean arterial pressure only in normal subjects.

CONCLUSION

Blood pressure determinants are not similar in the two populations. The effect of age, especially, is not the same in patients with sickle cell disease and in normal subjects. These results confirm that specific patho-physiological models should be defined in sickle cell disease.

Is there an association between Helicobacter pylori infection and coronary heart disease?

In addition to gastroduodenal diseases, Helicobacter pylori infection has been associated with extradigestive diseases, including coronary heart disease. Different studies have failed to demonstrate a clear association. The aim of this work was to carry out a meta-analysis of these studies' results, with a special emphasis on heterogeneity. A MEDLINE search of all studies published in English from 1994 to 1998 was conducted. Five criteria for eligibility of studies were defined. The quality of each study was assessed on a five-point scale adapted from studies by the Quebec Task Force. The final results are reported based on a fixed-effects model. Pooled odd ratios were calculated for subgroups of studies (defined on quality of adjustment, on confounding factors, and on the type of control population). Sixty-nine articles and abstracts fulfilling the defined criteria were analyzed. They were all case-control or cross-sectional studies. H. pylori diagnosis was carried out by serology, urea breath test or both; coronary heart disease was diagnosed mainly by coronary angiography. The total sample size was 6603. The quality score varied from 4 to 10. The pooled odds ratios of the 24 articles included in the model was 1.55 (95% confidence interval (CI): 1.38-1.74). It ranged from 1.07 to 2.34 when the quality of adjustment for confounding factors was considered and from 1.25 to 1.99 when the type of control group was considered. In any case there was significant heterogeneity (p < 0.001). The present study demonstrated a possible weak association but the high degree of heterogeneity in the studies impedes a clear demonstration.

[Cost-benefit analysis of screening strategies by human immunodeficiency virus in French blood donors]

BACKGROUND

The residual risk of human immunodeficiency virus (HIV) infection from screened blood transfusion was estimated to be 1.7/10(6) between 1993 and 1995 in France. To orient blood safety policies, we have evaluated what would be, from the perspective of blood banks, the best screening strategy in terms of gain in effectiveness and added costs.

METHODS

A cost-effectiveness analysis compared 20 HIV-testing protocols using (1) available data for performances of the current screening tests; and, (2) national insurance estimates for the cost of tests. Results were expressed as the number of false negative donations that would be avoided and the cost by avoided false negative donation.

RESULTS

For 3 million donated blood units a year and a prevalence of 24 per million, there would be 72 infectious donated blood units, 70.56 of which would be detected by the current screening strategy. The number of additional donated blood infections avoided in all other strategies would be low (between 0.25 and 1.28) with a very high cost (280 million French francs per added false negative avoided or more).

CONCLUSION

A change in screening strategies for blood donations in France is not currently justified. If such a change was to be done, adding p24 antigen detection to the current screening strategy would be one of the worst solutions.

[Factors associated with survival in Kaposi's sarcoma in patients infected with the human immunodeficiency virus. Clinical Epidemiological Group for AIDS in Aquitaine]

BACKGROUND

To study behavioral risk factors of Kaposi's sarcoma (KS) among HIV infected homosexuals in Bordeaux, southwest France.

METHODS

A case-control study was performed within the Aquitaine Cohort. Cases of KS surviving in 1995 and homosexuals were systematically enrolled. For each case, two controls were selected among homosexuals surviving in the cohort. Cases and controls were matched on year of diagnosis of HIV infection. Data collection was based on a self administered questionnaire focusing on use of recreational drugs, detailed sexual practices and sexually transmitted diseases in the year preceeding the diagnosis of HIV infection, in the year after the HIV diagnosis and in the year preceeding the diagnosis of KS (or an equivalent period of time for controls).

RESULTS

Twelve cases were matched to 2 controls, 15 cases to one control and 13 cases remained unmatched. Matched analysis identified an association between KS and regular sexual partner (odds ratio = 0.07; 95% confidence interval: 0.01-0.52 and p < 0.001) and active and passive oro-anal intercourse before HIV diagnosis and before KS diagnosis (p = 0.01). In the unmatched analysis including all cases, we found an association between KS and the overall number of sexual partners (p < 0.03) for all periods of interest.

CONCLUSIONS

This case-control study identified sexual practices in favor of a sexually transmitted agent of KS.

Epidemiology of blood collection in France

The objectives of the cross-sectional study (EpiCoS) were to describe, at different stages, volunteers offering their blood, and to characterize various ways of collecting blood. From 15 September 1996 to 31 December 1996, individuals presenting at fixed or mobile sessions in one of 11 randomly selected blood banks were included after they had a medical examination. Variables studied were relative to type of collection, individuals, medical examination, patterns of blood letting, use of collected donations and if unused, reasons for discarding. Sixty four thousand and ninety two volunteers, aged 17-66 years old were included. The proportion of exclusion during medical examination was 10.8% (95% confidence interval (CI): 10.6-11.0%). Exclusions were more frequent among new volunteers and were mostly related to the safety of recipients. Most of the 57,003 donations were whole blood (94.0%) and collected in mobile sessions (89.9%). Five percent of collected donations were discarded; 3.5% (95% CI: 3.4-3.7%) of donations discarded for biological abnormalities, including 1.5% only for initial screen reactions to infectious disease markers (HBs antigen, anti-HBc antibodies, anti-HCV antibodies, anti-HIV antibodies, anti-HTLV antibodies, malaria antibodies and anti-syphilitic antibodies). The most frequent biological abnormality was a high alanine aminotransferase level. A follow-up of these indicators, within the French haemovigilance system, should allow further identification of risk factors and high-risk contexts, and planning means of optimizing blood collection in France.

Road accident statistics: discrepancies between police and hospital data in a French island

In most developed countries, information on road crashes are routinely collected by the police. However, comparison of police records and hospital data underlines a deficit of the number of road accidents in the routine statistics. In La Réunion, a French overseas dependency, an epidemiological study of injuries leading to hospitalisation or deaths has been performed from June 1993 to June 1994. The comparison between hospital data and police records showed that only 37.3% of non-fatally traffic-injured in-patients were recorded by the police. Length of stay in hospital, physician in charge of the first aid, urban place of the crash, type of vehicle involved, day and time of the crash and blood alcohol concentration were significantly associated with the presence in the police file. Police overestimated the severity of the injuries. Police notified 100 deaths on the 115 counted by the study. In France, non-fatally traffic-injured should be followed 30 days to improve quality of police death records. A capture-recapture method was used to estimate the total number of injured people. The capture-recapture method consists in merging information from several sources of notification to determine the real number of cases in the population and the exhaustivity of each source. We estimated that 346 subjects were injured in one month whereas police data recorded only 87 and hospital data 137. This method seems interesting to use in routine after validation when unique personal identifiers are available.

[Evaluation of the actual traceability of labile blood products using medical records]

The traceability of blood products is an essential part of haemovigilance and transfusion safety. A pilot survey assessed the actual traceability by analysing transfusion information collected from medical records of a representative sample of 390 labile blood products transfused in a French university hospital. Transfusion and distribution forms were missing in 2.3% and 6.9% respectively. Availability and validity of transfusion information varied according to the nature of the expected information, elements of patients' records and types of wards. The location where the transfusion was performed was false or ambiguous in 38% of cases in surgery. Crude traceability, evaluated by the feedback of validated distribution forms, was estimated at 85.2% whereas actual traceability was estimated at 81.9% (SD 1.7%). High availability (98.7%) of at least one of the two sheets of the distribution form in medical records, or in the blood bank, revealed that a significant improvement of traceability should come from a better compliance to the rules of information transmission. The actual traceability differed significantly according to clinical services (worse in surgery) and was lower in case of autologous or absence of previous transfusion. An analysis of markers of good traceability should suggest efficient evolution of organization and information systems. This pilot study shows the relevance and feasibility of this kind of survey which could interestingly be performed on a large national representative random sample.

Reliability of the pretransfusion bedside compatibility test: association with transfusion practice and training

BACKGROUND

The utility of a pretransfusion bedside blood compatibility protocol to decrease immunohemolytic accidents has been questioned for years.

STUDY DESIGN AND METHODS

The reliability of a standard bedside ABO compatibility test was evaluated with a stratified random sample of 48 nurses who performed agglutination testing by using Bristol cards, interpreted compatibility, and decided whether to transfuse red cells for 12 randomly and blindly selected donor-and-recipient blood sample pairs. An expert judged technical performance and the interpretation of each card.

RESULTS

Erroneous decisions occurred in 18.2 percent of 576 tests, including 12 decisions to transfuse incompatible blood. Errors involved both testing protocols and the interpretation of compatibility. Anti-A and anti-B were detected with 92.8-percent sensitivity and 95.9-percent specificity. The expert judged 17.7 percent of tests to be technically inadequate, most often because of the application of excess blood to the card and a lack of rotation of the card. Testing errors (16.1% of tests) were significantly linked to infrequent transfusion activity by the nursing service, inexperience, and insufficient training. Compatibility misinterpretation occurred in 14.6 percent of the tests and was significantly linked to the nurses' infrequent transfusion activity, inexperience, insufficient training, lack of practical experience, and confusion regarding the use of ABO-compatible but not identical blood.

CONCLUSION

Bedside pretransfusion compatibility determination should not be considered a reliable supplemental safety procedure in the hands of inexperienced and insufficiently trained operators.