Search

In older adults, nutritional health is essential for good quality of life and living independently at home. Especially in cancer patients, malnutrition is common and known to complicate treatment. This study aims to evaluate the nutritional status and its associated factors in community-dwelling older adults with and without cancer. This is an observational study. This study focuses on older community-dwelling people. This study included older people with and without cancer (ae70 years). Cancer patients included patients with a new diagnosis of breast, lung, prostate, or colorectal cancer. Data collection included measures of nutritional status, quality of life, depression, fatigue, distress and functional status. We used multivariate logistic regression analysis to assess the association between personal characteristics and malnutrition. Data were available for 657 people; 383 people without cancer and 274 with a cancer diagnosis. Overall, malnutrition was detected in 245 (37.5%) people; in cancer patients this was 66.1%. Multivariate analysis showed that having cancer (OR 14.4, 95% CI: 8.01 - 23.3), being male (OR 2.38, 95% CI: 1.49-3.70), having depression (OR 13.5, 95% CI: 6.02-30.0), distress (OR 2.60, 95% CI: 1.55-4.37) and impaired instrumental activities of daily living (IADL) (OR 2.63, 95% CI: 1.63-4.24) were associated with a higher risk of malnutrition. The prevalence of malnutrition in community-dwelling older people is high, particularly in patients with cancer. Benchmarking and routine screening of older patients may be helpful strategies to increase awareness of (risk of) malnutrition among professionals.

Inleiding Een CRP-sneltest via een vingerprik zou de huisarts kunnen helpen bij het uitsluiten van een ernstige infectie bij kinderen. Wij onderzochten of we de sneltest bij alle kinderen moeten aanbieden of alleen bij kinderen met klinische risicofactoren. Methode We voerden een clustergerandomiseerd onderzoek uit onder acuut zieke kinderen die zich aanmeldden bij 133 huisartsen in 78 huisartspraktijken in België (3147 ziekte-episodes). We randomiseerden praktijken om een CRP-sneltest uit te voeren bij alle kinderen of alleen bij kinderen met klinische risicofactoren (kortademigheid, temperatuur ≥ 40 °C, diarree bij kinderen tussen 12 tot 30 maanden of een niet-pluisgevoel van de arts). De uitkomstmaat was ziekenhuisopname met een ernstige infectie binnen vijf dagen. Resultaten Als kinderen een CRP-test krijgen op basis van klinische risicofactoren, is zo’n test slechts nodig bij 20% van de kinderen. Er was tussen de twee onderzoeksgroepen geen verschil in het aantal kinderen met een ernstige infectie dat de huisartsen dadelijk naar het ziekenhuis verwezen (0,16% versus 0,14%, p = 0,88), of misten bij het eerste contact (0,29% versus 0,14%, p = 0,35). Elf kinderen hadden een ernstige infectie, van hen hadden er zes een CRP van minder dan 20 mg/L. Slechts een kind met een CRP ≺ 5 mg/L had een ziekte die een opname rechtvaardigde. Conclusie De CRP-sneltest is alleen nuttig bij kinderen met klinische risicofactoren. Een CRP ≺ 5 mg/L sluit een ernstige infectie uit en kan de huisarts helpen om onnodige verwijzingen te vermijden. Het klinische oordeel van de arts blijft essentieel, zelfs als een lage CRP-waarde een ernstige infectie lijkt uit te sluiten.

Questionnaire. Original questionnaire given to patients for this study (French language). (PDF 1004 kb)

No data have been reported yet on treatment outcome in persons who inject drugs (PWID) infected with hepatitis C virus treated with boceprevir or telaprevir in combination with peginterferon (Peg IFN) and ribavirin (RBV). Additionally, there are concerns about the safety of boceprevir and telaprevir in some subgroups of patients with hepatitis C (HCV). In a cohort of HCV patients infected with genotype 1 in Belgium, treatment outcome of patients infected due to IV drug use was analyzed and compared with patients who have no history of substance use. The study population consisted of 179 patients: 78 PWID and 101 controls treated with boceprevir (n = 79) or telaprevir (n = 100) additional to Peg IFN and RBV; 53 (30%) had advanced disease (F3, F4) and 79 (44%) had an antiviral therapy previously. There were no significant differences in the baseline characteristics between both groups, except that PWID patients were more frequently infected with genotype 1a (67% vs 21%), were younger and were predominantly male. Psychiatric complaints during follow-up occurred more frequently in the PWID patients: 24% versus 11% (P = .02). Treatment failure for other reasons than absence of viral response was 70% and 64% in PWID and non-PWID respectively. The sustained viral response (SVR) rates were similar in both groups (71% in PWID vs 72% in non-PWID); with a non-inferiority test with -5% margin there is a difference of -1% (95% CI [-15%, 13%]) and P = 0.30. There are no reasons to exclude PWID from treatment with boceprevir, telaprevir and novel antiviral therapies.

We evaluated the performance of the IMMIDIET food frequency questionnaire (FFQ) used to collect dietary data for the Belgian case-control study on bladder cancer. Thirty-seven men and women aged 50 years and older were recruited from the University Hospital in Leuven, Belgium. Participants completed the IMMIDIET FFQ, a 7-day diet diary and a 24-hour diet recall. Median intakes and inter-quartile ranges were calculated for 27 foods and nutrients from each dietary assessment method. All dietary factors were log-transformed and adjusted for energy using the nutrient density method. Pearson correlation coefficients were used to compare the different dietary assessment methods. Bland-Altman plots were also used to assess levels of agreement between the dietary methods. Energy, fruit and vegetable intake estimates were higher from the IMMIDIET FFQ compared with the two reference methods.The highest deattenuated correlations between the FFQ and 7-day diary were meat (0.58), bread (0.44), fruit (0.38) and fish (0.38). The highest deattenuated correlations between the FFQ and 24-hour recall were for fruit (0.72), fat (0.48), alcohol (0.44), cholesterol (0.42), monounsaturated fatty acid (0.42) and polyunsaturated fatty acid (0.41). Generally, correlation was lower for the micro-nutrients except for phosphorus (0.42), vitamin C (0.41) and calcium (0.40). The IMMIDIET FFQ is an appropriate instrument to measure usual dietary intake for the Belgian case-control study on bladder cancer risk. Further investigation of nutritional assessment methods is necessary.

Background: Do doctors really need to establish an etiological diagnosis each time a patient presents? Or might it often be more effective to treat simply on the basis of symptoms and signs alone, relying on research and on our experience of outcomes for patients who presented in similar ways in the past? Discussion: At a time of increase health care costs especially in pharmaceuticals and expensive diagnostic tests, this article uses examples from recent research to address this question. Our examples come from general practice, because that is where doctors frequently see patients presenting with a yet undifferentiated disease which is consequently difficult to diagnose. The examples include respiratory tract infections, low back pain and shoulder pain. Finally we discuss the 'something is wrong' feeling. Summary: We conclude that, in addition to diagnostic research, a renewed focus on prognostic research is needed.

In the first two articles of this series devoted to Evidence-Based Medicine, we have shown how to translate a clinical problem in a well formulated question and how to derive search terms from the PICO. In this article, we describe in more detail the different sources of information and how to use them. The strategy will be illustrated by answering the question whether it should be recommended to treat hypercholesterolaemia in a young woman without cardiovascular risk factors. In a following article, we will show how to adapt a search strategy to a specific question or PICO.

We evaluated the accuracy of a disposable, sterile thermometer that is practical in use and holds no risk of cross-infections.a cross-sectional study was set up in which we compared the Tempa Dot with the mercury thermometer in adults and children. Subjects were recruited from general practice and a paediatric ward. The mercury thermometer was used orally in subjects over 2 years of age and rectally in children up to 2 years old. The Tempa Dot was used either orally or axillary.The total population consisted of 212 patients, of which 131 children were younger than 16 years old. Their mean age was 17.3 years old, ranging from 1 month to 76 years. The mean difference between the mercury thermometer and the Tempa Dot, used orally or axillary, was 0.04 degrees C. For children between 0 and 16 years old, the mean difference was 0.08 degrees C. Agreement between the two methods as assessed with regression analysis and Bland and Altman plots was very good. ROC curve analysis suggests cut-off points of 37.2 and 37.6 degrees C to detect fever for the Tempa Dot at the oral and the axillary site respectively. Sensitivity and specificity were 100.0% and 79.0% for the total population, measuring orally and 100.0% and 95.9% axillary. In children, sensitivity and specificity were 100.0% and 83.1% orally, and 100.0% and 95.4% axillary. Using a single cut-off point for both measuring sites, namely 37.5 degrees C, sensitivity dropped and specificity increased for the oral site. For the axillary site, sensitivity remained unchanged and specificity was somewhat less.the Tempa Dot is a reliable alternative for the mercury thermometer. In clinical use, a cut-off point of 37.5 degrees C for both the oral and axillary site is most appropriate.

Heroin addiction is a chronic relapsing disease that is difficult to cure, but stabilisation and harm reduction can importantly increase the life time expectancy and the quality of life of the patient, his immediate vicinity and society in general. Currently, no proven effective pharmacological interventions are available for cocaine addiction, and treatment has to rely on existing cognitive behaviour therapies combined with contingency management strategies. Substitution therapy, however, is effective in caring for heroin addicts. Methadone is a synthetic opioid that counteracts withdrawal symptoms of heroin. Buprenorphine is a derivative of the morphine alkaloid, thebaine, and is a partial opioid agonist at the micro opioid receptor in the nervous system. A substitution treatment program effectively reduces and often eliminates heroin injection behaviour, rendering patients more socially stabilised. Reduction in the number of viral co-infections can be observed. Methadone undergoes oxidative biotransformation in the liver, but is also stored in the liver and released into the blood in unchanged form. The usual dose can be continued in patients with stable chronic liver disease, including advanced cirrhosis. In acute liver disease or acute decompensation of chronic liver disease, close clinical observation for signs of narcotic overdose or withdrawal is necessary. A modest alteration in methadone dose may be appropriate for some patients. Buprenorphine can cause liver dysfunction after sublingual and even more after intravenous administration. It is advised to follow the liver function during buprenorphine treatment and to warn the clients for intravenous use of buprenorphine. Neither methadone nor buprenorphine do influence the effect of interferon and ribavirin during the treatment of chronic hepatitis C patients. It may be necessary to increase the dosage of methadone during interferon treatment.

To examine the evidence for excluding chronic hepatitis C (CHC) patients with substance abuse from treatment with interferon (IFN) and ribavirin.We reviewed clinical trials focussing on the treatment of chronic hepatitis C of patients with substance abuse between 2001 and 2004. Ten clinical trials concerning antiviral treatment in substance abusers were described of which six were controlled ones. There were no randomised trials. There was one controlled multi-centre trial. One trial used pegylated IFN.In the total group of substance abusers the sustained viral response (SVR) and the adherence was not different from control groups. In former drug users, active drug users and patients taking substitution therapy for opioid dependence the sustained viral response and adherence was not different from control populations. However, non-substituted active drug users seemed more likely to be lost to follow-up. Discontinuation of treatment occurred most frequently during the first 8 weeks of therapy. Neurobehavioural changes leading to depression started in the first 8 weeks of treatment. Although follow-up periods after SVR were short, the currently described re-infection rate occurring in active intravenous drug users remains low.There is no evidence to withhold antiviral treatment against HCV in active substance abusers. It seems important to advise to start substitution therapy in non-substituted active drug users, increase substitution therapy dose in substituted patients and treat depression as early as possible. More prospective controlled trials on HCV treatment in active and difficult-to-reach substance users are needed.

Substance use is emerging in Belgium. The 'typical user' starts at a young age. Polydrug use has become very common. Cannabis and alcohol are the most frequently used substances among the school population. Heroin and cocaine are the most frequently injected drugs. Sharing of injecting material and paraphernalia is reported to happen in half of the subjects injecting drugs. Substance use causes a lot of adverse organic, social and psychiatric events. Management of substance use consists of information, drug-free treatment and harm reduction, including substitution and maintenance programs. The management of care in Belgium differs between the regions.

The evidence regarding the co-morbidity of chronic hepatitis C, psychiatric illness and intravenous drug abuse is reviewed from the literature. Also the occurrence and the treatment of psychiatric side effects during treatment with interferon in patients with a history of drug abuse are reviewed. There is insufficient evidence for a specific hepatitis C induced depression or fatigue, but a direct link between hepatitis C and cerebral dysfunction is not excluded. Immune system activation rather than drug use may explain cerebral symptoms. In HCV positive substance users anxiety and depression are more prevalent than in HCV negative substance users. During treatment with regular or pegylated (PEG) interferon depression is a frequent side effect (ca 30%) and occurs independently from pre-existing psychiatric disorders or drug abuse. A history of drug abuse per se does not increase the risk of depression as a side effect of interferon treatment. It is extremely important to monitor symptoms of depression in the early weeks of treatment and to start antidepressant treatment as early as possible. Antidepressants should be continued throughout the interferon treatment period. There are insufficient data to assess these situations in which preventive antidepressant treatment should be started before interferon treatment. Clinical judgement can, however, lead to preventive antidepressant treatment, even at subclinical levels of depression. A cut off score of10 on the Beck Depression Inventory before interferon treatment is associated with a higher risk of depression during treatment. Both selective serotonin reuptake inhibitors and other classes of antidepressants can be used.

In the first article of this series, we have shown how to translate a clinical problem into a well built question, by creating a PICO (Patient--Intervention--Comparison--Outcome). In this second article, we will explain how to transform the PICO in English search terms for use on the internet. We use these terms in the different databases to find the answer to the following clinical problem: "In patients aged 65 or over with hip osteoarthritis, and a history of peptic ulcer, is the risk of a new peptic ulcer less using a cox-2 inhibitor than with a classic nonsteroidal anti-inflammatory drug?"

This series of articles begins with a problem frequently encountered in children: acute otitis media. Under the insistence of parents, antibiotics are frequently prescribed. The physician knows that the value of that therapy is debatable, but what for a child of 6 years of age? The physician whishes to compare his/her opinion with what is published in the litterature or available on the web. How to start his/her research? In this first article the physician will find out how to translate his/her practical question into a question which leads to a practical answer, and then to a PICO question. This will allow the physician to find sources of directly usable and relevant data which will offer a rational and wellfounded answer to his/her clinical question.

Intravenous drug use is a major route of hepatitis C virus (HCV) transmission. In Belgium, more than 70% of the intravenous drug users (IVDUs) are HCV seropositive. In the past, medical treatment of HCV-positive IVDUs has been controversial. However, current studies support that the anti-HCV therapy of IVDUs should be the same as in other HCV-infected patients. In prison populations, HCV screening and therapy has to be performed. Patients should be counseled about the benefits of alcohol abstinence, should be educated about safer injection techniques to avoid reinfection, and should be vaccinated to avoid hepatitis A or B co-infections. Treatment of HCV infections should not be withheld from patient populations with complicated social problems. Physicians should rather develop individual treatment and follow-up plans in order to optimize compliance in IVDUs.

The aim of this paper is to provide a theoretical background for performing and reading systematic reviews of diagnostic studies. We first discuss items for assessment of methodological quality in diagnostic studies and then present methods on how to incorporate these quality measures in systematic reviews. The items of internal validity determine whether the presented results of the individual studies are unbiased and can be trusted. Items of external validity determine to what extent the results are applicable outside the population in which the study was performed. The issues concern the adequacy of the study population, the performance and interpretation of the diagnostic tests and the presentation of the results. Several methods exist for incorporation of issues of methodological quality into systematic reviews, such as subgroup analyses, meta-regression analysis, and methodological scores. Publications of diagnostic studies should provide sufficient information to enable assessment of the methodological quality. Furthermore, publication of results of subgroup analyses should be promoted. Methodological criteria lists might help to improve the quality of systematic reviews of diagnostic research. With the items of methodological quality in mind the general practitioner might be better equipped to critically read and interpret diagnostic reviews.

The aim of this study was to get a profile of patients who are vulnerable to get multiple chronic, recurrent or high-impact diseases in a limited time period. We studied the incidence rates of morbidity and multimorbidity, and the influence of psychosocial characteristics on their occurrences.Cohort study with 3551 subjects. Baseline measurement of psychosocial characteristics and a 2-year follow-up period for morbidity. The relations were evaluated using multiple logistic regression analysis.After adjustment for basic socio-demographic variables, a high internal locus of control belief was found to be protective [odds ratio (OR)=0.82] for the occurrence of morbidity, negative life events increased the risk (OR=1.22). Characteristics specifically protective for the occurrence of multimorbidity as compared to monomorbidity were: a high internal locus of control belief (OR=0.73), living as a couple or in a family as compared to living alone (OR=0.68) and a large social network (OR=0.41).It appears that certain patient characteristics are specifically related to the occurrence of multimorbidity. This provides opportunities for the future development of preventive interventions.

The objective of this study was to identify possible risk factors associated with open-angle glaucoma (OAG). A case-control study included patients seen at an ophthalmologic clinic. Cases were all consecutive new patients with either unilateral or bilateral OAG detected during the study period. Controls were a random sample of all other patients aged 30 or more, seen in the same department in the same period. Data on demographic, anthropometric and diet habits as well as medical characteristics were collected from 144 patients by medical examination and interview. The study took place at the University Department of Ophthalmology and general private practice of ophthalmology, both in the city of Kinshasa. Forty consecutive patients with OAG and 104 controls were chosen randomly between all consecutive non OAG patients. Odds ratio (OR) are presented for the relation between OAG and age, sex, ethnicity, family history of glaucoma, the length of stay in Kinshasa, body mass index, hypertension, diabetes mellitus, cigarette smoking, alcohol, diet habits. Adjusted odds ratio resulting from stepwise logistic regression was employed. Results indicate: family history of glaucoma (OR, 18; 95% CI, 5.80-59.00; P0.001), age (OR, 1.05; 95% CI, 1.01-1.09; P = 0.025), body mass index (OR, 1.09; 95% CI, 1.01-1.18; P = 0.05), hyperopia (OR, 2.9; 95% CI, 1.05-7.08; P = 0.03), Mongo ethnic group (OR, 3.5; 95% CI, 1.11-12.20; P = 0.03) and consumption of rice (OR, 4.6; 95% CI, 1.65-12.20; P = 0.004) conferred a significantly greater risk of OAG. This study seems confirm that Mongo ethnic group is associated with an increased risk of OAG.

To compare the quality of cervical smears taken by practice assistants with the quality of smears taken by general practitioners.Retrospective.Data were collected on the 1672 smears taken as part of the population screening for cervical cancer and recorded 13 general practices in Limburg in the year 1997. The results of smears taken by practice assistants were compared with those made by general practitioners with regard to the quality indicators 'presence of endocervical cells' and 'absence of Pap-o'.In 12.3% of the 481 cervical smears taken by the 17 practice assistants endocervical cells were absent. This is statistically significantly more than to the results (7.2%) in 1191 smears taken by the 22 general practitioners (relative risk: 1.70: 95% confidence interval: 1.23-2.70). Similar differences, but not statistically significant, were found with regard to Pap-o results and pathological findings. Most of the practice assistants had 1-2 years' experience. The proportion of smears made by practice assistants in which endocervical cells were lacking was reasonable compared with the average of 10% for the Netherlands as a whole.The cervical smears taken by practice assistants were of lower quality than those taken by the general practitioners but almost as good as the nationwide mean.

Ambulatory blood pressure (ABP) monitoring is used increasingly in clinical practice, but how it affects treatment of blood pressure has not been determined.To compare conventional blood pressure (CBP) measurement and ABP measurement in the management of hypertensive patients.Multicenter, randomized, parallel-group trial.Family practices and outpatient clinics at regional and university hospitals.A total of 419 patients (or =18 years), whose untreated diastolic blood pressure (DBP) on CBP measurement averaged 95 mm Hg or higher, randomized to CBP or ABP arms.Antihypertensive drug treatment was adjusted in a stepwise fashion based on either the average daytime (from 10 AM to 8 PM) ambulatory DBP (n=213) or the average of 3 sitting DBP readings (n=206). If the DBP guiding treatment was above (89 mm Hg), at (80-89 mm Hg), or below (80 mm Hg) target, 1 physician blinded to the patients' randomization intensified antihypertensive treatment, left it unchanged, or reduced it, respectively.The CBP and ABP levels, intensity of drug treatment, electrocardiographic and echocardiographic left ventricular mass, symptoms reported by questionnaire, and cost.At the end of the study (median follow-up, 182 days; 5th to 95th percentile interval, 85-258 days), more ABP than CBP patients had stopped antihypertensive drug treatment (26.3% vs 7.3%; P.001), and fewer ABP patients had progressed to sustained multiple-drug treatment (27.2% vs 42.7%; P.001). The final CBP and 24-hour ABP averaged 144.1/89.9 mm Hg and 129.4/79.5 mm Hg in the ABP group and 140.3/89.6 mm Hg and 128.0/79.1 mm Hg in the CBP group. Left ventricular mass and reported symptoms were similar in the 2 groups. The potential savings in the ABP group in terms of less intensive drug treatment and fewer physician visits were offset by the costs of ABP monitoring.Adjustment of antihypertensive treatment based on ABP monitoring instead of CBP measurement led to less intensive drug treatment with preservation of blood pressure control, general well-being, and inhibition of left ventricular enlargement but did not reduce the overall costs of antihypertensive treatment.

Six observers (three physicians and three nurses) assessed 27 wounds of different types in 20 elderly patients, using five different classification systems. Using classification by colours, inter-observer agreement was moderate when six colours were used (70% agreement, group kappa 53%), and good if this was reduced to three (80% agreement, group kappa 64%). Results were only moderate using the Shea scoring system (68% agreement, group kappa 42%). Assessment for signs of infection was fair to moderate (75-85% agreement, group kappa 29-55%), but there was high inter-observer agreement for size and area (analysis of variance, p≥0.88).

The U.S. Food and Drug Administration designed the trough-to-peak ratio as an instrument for the evaluation of long-acting antihypertensive drugs, but the ratios are usually reported without accounting for interindividual variability. This study investigated how the trough-to-peak ratio would be affected by interindividual and intraindividual variability and by smoothing of the diurnal blood pressure profiles. The ambulatory blood pressure was recorded on placebo in 143 hypertensive patients (diastolic pressure on conventional measurement95 mm Hg). After 2 months, the recordings were repeated on 10 mg (n = 66) or 20 mg (n = 77) lisinopril given once daily between 7 and 11 PM. The baseline-adjusted trough-to-peak ratios were determined from diurnal blood pressure profiles with 1-hour precision. Lisinopril reduced (+/- SD) the 24-hour pressure by 16 +/- 17 mm Hg for systolic and 10 +/- 10 mm Hg for diastolic (P.001). According to the usual approach, disregarding interindividual variability, the trough-to-peak ratio was 0.72 for systolic pressure and 0.67 for diastolic pressure. In the 143 patients the ratios were not normally distributed. They were the same on both lisinopril doses. When interindividual variability was accounted for, the median trough-to-peak ratio was 0.34 (P5 to P95 interval, -0.46 to 0.87) for systolic pressure and 0.26 (-0.44 to 0.84) for diastolic pressure. In 66 patients examined twice on 10 mg lisinopril at a median interval of 32 days, the trough-to-peak ratios were characterized by large intraindividual variability.(ABSTRACT TRUNCATED AT 250 WORDS)

Determining the role of the general practitioner (GP) in the management of the postinfarction patient, the factors that influence this role, and the relation between GP and cardiologist.Descriptive.Registration Network of Family Practice (RNF) of the University of Limburg, the Netherlands.Data were obtained from a sample of 284 patients from 30 GPs, all participating in the RNF. The GPs completed a questionnaire about their own patients while the data concerning the cardiac state of the patients were obtained from their problem lists.Postinfarction care was given in only 50% of the patients. Relatively more patients with a more severe cardiac condition were under regular medical follow-up than those with a less severe cardiac condition. However, the severity of the disease had no effect on the specialty of the doctor performing the follow-up (GP or cardiologist), nor had the time expired after the myocardial infarction. In almost 50% of the cases the GPs found continuation of the follow-up by the cardiologist not necessary, in 25% of the cases they were not satisfied with the complementary roles of both disciplines.Postinfarction care by the GPs was not uniform. They need guidelines for the management of such patients, based on the patient's risk profile. There is a need for better coordination between GP and cardiologist.

Three different forms of feedback designed to improve the sampling quality of cervical smears were tested in a randomized controlled trial among 179 physicians. The reactions of the doctors to the various interventions are reported as they appeared from their answers in a questionnaire. The influence of the interventions on the employed sampling technique is described on the basis of their answers to a telephone inquiry. The inclusion in laboratory protocols of an evaluation of the sampling quality of the submitted smears was largely appreciated, as was specific advice offered on indication. Monthly overviews comparing the performance of the individual doctor to that of his peers were esteemed by three out of four. Further information, help or training was requested more frequently than average by physicians obtaining inadequately sampled smears relatively often. The number of doctors using the combined spatula and Cytobrush method as sampling technique doubled during the study period.

Meta-analysis is a way of reviewing literature in a structured and systematic format. A formalized protocol for selection of the reviewed papers and a qualitative as well as quantitative evaluation are essential. Mathematical pooling of the results of the individual studies can be part of the process. A number of pitfalls are possible. However, when performed in a correct way and evaluated critically by the reader, a meta-analysis can add consistently to the medical knowledge concerning a specific subject.

Three different forms of feedback designed to improve the sampling quality of cervical smears were tested in a randomized controlled trial among 179 physicians. The reactions of the doctors to the various interventions are reported as they appeared from their answers in a questionnaire. The influence of the interventions on the employed sampling technique is described on the basis of their answers to a telephone inquiry. The inclusion in laboratory protocols of an evaluation of the sampling quality of the submitted smears was largely appreciated, as was specific advice offered on indication. Monthly overviews comparing the performance of the individual doctor to that of his peers were esteemed by three out of four. Further information, help or training was requested more frequently than average by physicians obtaining inadequately sampled smears relatively often. The number of doctors using the combined spatula and Cytobrush method as sampling technique doubled during the study period.

It is generally accepted that cervical smears are adequately sampled when they contain endocervical and/or metaplastic cells. An extensive literature search via medline and famli for papers on cervical smears yielded 33 articles providing explicit information both on the country where the study was executed and on the relationship between the adequate sampling of smears and the sampling method used. The correlation between the geographic origin of cervical smears and the proportion of those smears that were adequately sampled was examined. The results were controlled for the sampling method. It was found that the proportion of adequately sampled smears reported in studies in the U.K. is exceedingly low, in contrast with the proportion reported from The Netherlands. The differences between the various countries studied are stable over different sampling methods. It is suggested that, in order to ameliorate the sampling quality of cervical smears, factors associated with the working conditions in particular regions also need be considered.

Since the introduction of the Cytobrush for sampling the uterine cervix, some practitioners have ceased taking a concomitant cervical scraping using a spatula. To examine whether Cytobrush sampling alone is adequate for the diagnosis of cervical lesions, the Cytobrush and spatula samples in 444 smears (most with original diagnoses of at least mild dysplasia) were analyzed separately for the presence of diagnostic cells, endocervical cells and squamous cells. Of the 412 smears showing pathologic findings (mild to severe dysplasia or worse), diagnostic cells were present in 400 Cytobrush samples and in 369 spatula samples; the combination of both samples thus gave a 3% gain in correct diagnoses as compared to use of the Cytobrush samples alone. Another 18 smears would have been underdiagnosed based only on the Cytobrush samples. Endocervical cells were present in 95.3% of the Cytobrush samples and 83.8% of the spatula samples; squamous cells were present in 93.9% of the Cytobrush samples and 96.8% of the spatula samples. Analysis confirmed that it is important that the smear should contain both endocervical and squamous cells. A positive relationship between the absence of squamous cells in the Cytobrush sample and the probability of a false-negative assessment was suggested. It thus seems inadvisable to replace the combination sampling method by Cytobrush sampling alone, which may lead to a false-negative diagnosis.