SEARCH · PubMed Health

Results for “Logistic Models”

Explore indexed PubMed citations for clinical trials, systematic reviews and public health research. Read source abstracts and follow each citation to its original PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,027 records · Page 57Linked to original sources

PubMed1996

Variations in familial neonatal mortality risks in four countries.

This paper investigates variations in the strength and structure of familial association in neonatal mortality risks in four populations; Bolivia, Kenya, Peru, and Tanzania. Exploratory analyses of the structure of the familial association are presented for each population. Random effects logistic models are then used to estimate the strength of familial association in neonatal mortality risks using a standard set of control variables. The results suggest that the strength of familial association in neonatal mortality risks is quite similar in these four populations which would be consistent with a biological explanation for the association. However, some differences were found, particularly in the form of the association in Peru, which may suggest at least a small role of other factors.

Africa, Eastern↗

PubMed2003

The impact of rural-urban migration on under-two mortality in India.

This paper examines the impact of rural-urban migration on under-two mortality in India, using data from the 1992/93 Indian National Family Health Survey. Multilevel logistic models are fitted for mortality in three age groups: neonatal, early post-neonatal, and late post-neonatal and toddler. Migration status was not a significant determinant of mortality in any of the three age groups. Further analysis shows that a relationship between migration status and mortality exists when socioeconomic and health utilization variables are omitted from the models. The relationship between migration and mortality is thus explained by differences in socioeconomic status and use of health services between rural-urban migrant and nonmigrant groups. The selectivity of rural-urban migrants on socioeconomic characteristics creates mortality differentials between rural-urban migrants and rural non-migrants. Problems faced by migrants in assimilating into urban societies create mortality differentials between rural-urban migrants and urban non-migrants. These results highlight the need to target migrants in the provision of health services, and demonstrate that rural areas continue to have the highest levels of infant-child mortality. Further research is needed to understand the health care needs of rural-urban migrants in order to inform the provision of appropriate health care.

Adolescent↗

PubMed1997

Risk-adjusted outcome measures and quality of care in nursing homes.

OBJECTIVES: This study examines the use of the patient review instrument (PRI), a reimbursement reporting system used by the New York State Department of Health, to assess quality of care based on risk-adjusted outcomes. METHODS: Data for all residents in approximately 550 nursing homes in upstate New York are used to develop five risk-adjusted outcome measures. The five measures are rates of decline in functional status activities of daily living (ADL), rates of increases in severity of decubitus ulcers, physical restraints rates, dehydration rates, and rates of major accidents. Logistic models are used to adjust for individual patients' risk factors. The face validity, content, construct, and criterion validity of these measures is examined. RESULTS: Measures based on ADL decline, deterioration in decubiti, and physical restraints rates met all validity criteria and were correlated significantly with deficiency citations. Measures based on accident rates and rates of dehydration did not perform as well. There was significant variation in these quality measures across regions and between for-profit and nonprofit nursing homes. CONCLUSIONS: Information about quality of care is important to the efficient operation of competitive markets. Such information, however, often is costly to obtain and not available to individual patients. This study demonstrates that valid risk-adjusted outcome measures of quality can be developed based on data collected for reimbursement purposes.

Accidents↗

PubMed2006

Preventable hospitalization and medicaid managed care: does race matter?

This study examines the preventable hospitalization patterns of Medicaid patients by race/ethnicity to determine whether Medicaid managed care (MMC) has been more effective in some subgroups than others. It uses logistic models for three states, comparing preventable hospitalizations with marker admissions (urgent admissions, insensitive to primary care). Hospital discharge data from the Healthcare Cost and Utilization Project State Inpatient database of the Agency for Health Care Research and Quality for New York, Pennsylvania, and Wisconsin residents aged 20-64 years is used. In a more urban state, New York, MMC was effective for Whites but not for minorities. In a more rural state, Wisconsin, MMC was effective for minorities. Overall, the evidence is not strong that any particular racial group consistently benefited from MMC, or that any state consistently showed a favorable impact of MMC across racial groups. However, racial/ethnic disparity associated with the risk of preventable hospitalization is significantly lower among Medicaid patients than among private fee-for-service patients.

Adult↗

PubMed2003

[Study of delaying hospital admission after acute stroke].

OBJECTIVE: To explore the association between hospital admission after acute stroke and delaying factors, and try to determine the factors that delay hospital admission and treatment, in order to find good solution. METHODS: We studied 300 patients with acute stroke from Sep. 2002 to Mar. 2003 in our emergency department and used chi 2 test and logistic model to analyze the collected data. RESULTS: The stroke subtypes were confirmed by CT or MRI scan in all the patients, including 210 cases of cerebral infarction, 80 cases of cerebral hemorrhage, and 10 cases of subarachnoid hemorrhage. The most important factor associated with a delay in reaching the hospital was the pattern of arriving hospital. Other factors associated with a delay were the referred hospital, stroke subtype and distance from the hospital. CONCLUSION: The majority of patients who are candidates for acute stroke arrive at the hospital after prolonged delays for multiple reasons. The main reason for hospital admission delay is that they do not pay enough attention. Public and medical personnel education could result in significant reduction of these delays.

Adult↗

PubMed1995

Effectiveness of bystander cardiopulmonary resuscitation and survival following out-of-hospital cardiac arrest.

OBJECTIVE: To examine the independent relationship between effectiveness of bystander cardiopulmonary resuscitation (CPR) and survival following out-of-hospital cardiac arrest. DESIGN: Prospective observational cohort. SETTING: New York City. PARTICIPANTS: A total of 2071 consecutive out-of-hospital cardiac arrests meeting Utstein criteria. INTERVENTION: Trained prehospital personnel assessed the quality of bystander CPR on arrival at the scene. Satisfactory execution of CPR required performance of both adequate compressions and ventilations in conformity with current American Heart Association guidelines. MAIN OUTCOME MEASURE: Adjusted association between CPR effectiveness and survival. Survival was defined as discharge from hospital to home. RESULTS: Outcome was determined on all members of the inception cohort--none were lost to follow-up. When the association between bystander CPR and survival was adjusted for effectiveness of CPR in the parent data set (N = 2071), only effective CPR was retained in the logistic model (adjusted odds ratio [OR] = 5.7; 95% confidence interval [CI], 2.7 to 12.2; P < .001). Of the subset of 662 individuals (32%) who received bystander CPR, 305 (46%) had it performed effectively. Of these, 4.6% (14/305) survived vs 1.4% (5/357) of those with ineffective CPR (OR = 3.4; 95% CI, 1.1 to 12.1; P < .02). After adjustment for witness status, initial rhythm, interval from collapse to CPR, and interval from collapse to advanced life support, effective CPR remained independently associated with improved survival (adjusted OR = 3.9; 95% CI, 1.1 to 14.0; P < .04). CONCLUSION: The association between bystander CPR and survival in out-of-hospital cardiac arrest appears to be confounded by CPR quality. Effective CPR is independently associated with a quantitatively and statistically significant improvement in survival.

Cardiopulmonary Resuscitation↗

PubMed2005

Trends in cannabis use and its determinants among teenagers in the French-speaking community of Belgium.

BACKGROUND: Cannabis consumption among teenagers has undergone dramatic changes in Europe since the beginning of the 1990s. A number of behaviors associated with cannabis consumption, such as tobacco smoking, excessive drinking and truancy are developing too, each in their own way. METHODS: To assess the evolution over time of the various types of cannabis consumption (both ever and weekly consumption) in relation to these determinants (age, sex, studies chosen, truancy, tobacco smoking and recurrent intoxication), we have analyzed the cross-sectional study on Health Behaviour in School-Aged Children in the French-speaking Belgian Community (12-17 years) since 1994. We used logistic models to analyze the evolution of the various types of cannabis consumption and to identify the associated factors. Finally, in order to demonstrate time trends, we tested for each type of consumption in the interactions between the significant predictive variables in each model and the survey year (1994-1998-2000). RESULTS: Rates of ever use, past 30-day use and weekly use among the ever users have been increasing from 1994 to 2002 and reached, respectively, 22.0%, 11.6%, 6.8% and 32.9%. Cannabis ever use rose more noticeably among the general education students (adjusted OR (95%CI)): 3.08 (2.66-3.57) and among the truants: 4.57 (3.39-6.14). Weekly cannabis smoking rose most especially among the truants: 1.92 (1.34-2.78). CONCLUSION: Truants should constitute a priority target for the prevention of cannabis consumption, while the phenomenon of truancy must be moreover examined in depth in order to more thoroughly identify the appropriate prevention programs organized both in and outside of the school environment.

Adolescent↗

PubMed1996

Influence of patient characteristics, socioeconomic factors, geography, and systemic risk on the use of breast-sparing treatment in women enrolled in adjuvant breast cancer studies: an analysis of two intergroup trials.

PURPOSE: To investigate the frequency of breast-sparing treatment among breast cancer patients subsequently enrolled in national cooperative group studies of adjuvant chemotherapy. PATIENTS AND METHODS: A data base was formed of 5,172 patients randomized onto two intergroup trials. Lumpectomy rates were analyzed within study-defined risk strata and across geographic regions. Significant predictors of lower lumpectomy usage were determined in multivariate analyses with variables that described patient and disease characteristics, systemic risk strata, geographic region, and socioeconomic indicators based on zip code of residence. RESULTS: Breast-conservation rates were 30% in the node-negative and 15% in the node-positive trials, with a wide geographic variation within each study (range, 14% to 49% and 9% to 31%, respectively). Lumpectomy use declined with increasing tumor size and did not exceed 40% even for tumors < or = 1 cm with negative nodes. With increasing risk of systemic relapse, frequency of lumpectomy declined (rates for five strata in order of increasing systemic risk: 41%, 33%, 24%, 18%, and 11%), even though these strata were not known at the time of the surgical decision. A logistic model confirmed the joint significance of geographic region and systemic risk. An exploratory model that adjusted for all important variables identified the following significant predictors of lower lumpectomy use: positive nodes; many positive nodes, increased systemic risk; tumor size > or = 2.0 cm; older age; South, Central or non-New England regions; and either lack of college degree or lower income levels. CONCLUSION: Breast-sparing therapy was used in the minority of women subsequently accrued to two national adjuvant breast cancer studies, even though this cohort and their referring surgeons represented a select population. Although multiple concrete factors were independent predictors of lower lumpectomy rates, prospective research is needed into how patients and their physicians approach the mastectomy versus lumpectomy decision.

Adult↗

PubMed2005

Prognostic significance of estrogen receptor Beta in epithelial hyperplasia of usual type with known outcome.

The prognostic significance of ER-alpha expression in benign proliferative breast disease has been confirmed in epithelial hyperplasia of usual type (HUT). However, little is known about the role of ER-beta in these lesions. Therefore, this study was performed to test the hypothesis that, in HUT lesions, the ratio of ER-alpha:ER-beta is an accurate determinant of breast cancer risk and of predicting subsequent progression to invasive breast cancer. This case-control study analyzed a cohort of benign proliferative breast lesions and foci of ductal HUT in 117 patients with long follow-up (20 years). These foci were analyzed by morphometric image analysis together with immunohistochemistry using monoclonal antibodies to ER-beta1 and to ER-alpha. The data were compared with ER-beta expression in all breast carcinomas that subsequently developed in the same patients as well as to ER-alpha expression in the corresponding tissues. In cases that progressed to carcinoma, the ratio of ER-alpha to ER-beta in HUT was significantly higher (P < 0.001) than in those that did not progress. None of the HUT foci from patients who progressed to breast cancer were simultaneously ER-alpha negative and ER-beta positive. Using both ER-beta and ER-alpha in a logistic model demonstrated a 75% correct classification rate for the cohort studied. These findings confirm the diagnostic and prognostic value of defining the ER-alpha and ER-beta status of HUT lesions identified morphologically. The data support the hypothesis that high ER-alpha:ER-beta levels characterize those cases within HUT likely to progress to breast cancer. The data also reveal that a reduced level of ER-beta relative to ER-alpha is an accurate predictor of individual cases of HUT likely to progress to invasive breast carcinoma, thus supporting the concept that ER-alpha transcriptional activity is directly modulated by ER-beta.

Breast↗

PubMed2002

Shortening the NIH Stroke scale for use in the prehospital setting.

BACKGROUND AND PURPOSE: Prehospital stroke scales should identify stroke patients and measure stroke severity. The goal of this study was to identify a subset of the 15 items in the National Institutes of Health Stroke Scale (NIHSS-15) that measures stroke severity and predicts outcomes. METHODS: Using 2 distinct data sets from acute stroke clinical trials, we derived and validated shortened versions of the NIHSS (sNIHSS). Stepwise logistic regression and bootstrap techniques were used in selection of NIHSS-15 items. Areas under the receiver operator characteristic curve (C statistics) were used to compare predictive performance of logistic models incorporating differing versions of the NIHSS. RESULTS: The derivation analyses suggested the 8 NIHSS-15 items that were most predictive of "good outcome" 3 months after stroke, in order of decreasing importance: right leg item, left leg, gaze, visual fields, language, level of consciousness, facial palsy, and dysarthria. The sNIHSS-8 comprises all 8 and the sNIHSS-5, the first 5. In the validation models, C statistics were NIHSS-15=0.80, sNIHSS-8=0.77, and sNIHSS-5=0.76. Statistical comparisons suggested that the NIHSS-15 had better predictive performance than the sNIHSS-8 or the sNIHSS-5; the absolute difference in C statistics was small. There was no significant difference between the sNIHSS-8 and the sNIHSS-5. CONCLUSIONS: Much of the predictive performance of the full NIHSS-15 was retained with a shortened scale, the sNIHSS-5. Shortening the NIHSS-15 will facilitate its use during prehospital evaluations. The sNIHSS severity information may be useful to triage acute stroke patients in communities and to provide a baseline stroke severity for prehospital acute stroke trials.

Acute Disease↗

PubMed2001

Injury mortality risk assessment and targeting the subpopulations for prevention in the Northwest Territories, Canada.

BACKGROUND AND OBJECTIVE: Prevention of injury mortality is a top priority for public health in the Northwest Territories, Canada. However, the frequency of injury mortality may not be evenly distributed in subgroups. Assessment of population relative risk of injury mortality can assist in better targeting of prevention programs. METHODS: A population-based case-control study consisting of 197 cases was used to quantify the relative risks of injury mortality for variables age, sex, ethnicity, geographic location and remoteness. RESULTS: A multivariate logistic model showed that the odds ratios were 2.90 (95% CI 2.09-4.03) for males, 3.39 (95% CI 2.55-4.52) for age group 15-64 years, 11.50 (95% CI 6.47-20.42) for age group over 64 years, 1.51 (95% CI 1.13-2.02) for people in remote communities, and 2.04 (95% CI 1.50-2.79) for people above 66 degrees latitude. Aboriginal people of Dene and Inuit presented increased risks in the stratified analysis by community remoteness and latitude. CONCLUSIONS: Males, over age 14, living in remote communities, living in the far north, and being aboriginal were at higher risk of injury mortality. Based on the relative risks of injury mortality and the population size in each subpopulation, injury prevention investments should be extended to remote communities.

Adolescent↗

PubMed1999

One-year predictive factors for various aspects of neck disorders.

STUDY DESIGN: A longitudinal epidemiologic study conducted over 12 months among active workers in different occupations. The study was primarily designed to evaluate intervention for prevention of low back and other spinal disorders. OBJECTIVE: To determine factors that predict incidence, recurrence, and persistence of neck disorders (ND), taking into account various dimensions of ND. SUMMARY OF BACKGROUND DATA: Neck pain is often supposed to have essentially the same risk factors as back pain; however, there is comparatively little data relevant to this issue. Moreover, there is a lack of prospective studies that take into account a diversity of predictive factors. METHODS: The Nordic questionnaire for the analysis of musculoskeletal symptoms was completed twice at a 12-month interval by 568 workers. Predictive factors were studied with logistic models for four dimensions of ND from the second questionnaire: 1) any ND in the past 6 months; 2) ND for more than 30 days; 3) treatment for ND; and 4) visit to a health care professional for ND. The predictive factors were obtained from the first questionnaire and included gender, age, occupational group, level of psychological distress and psychosomatic problems, and ND at baseline. RESULTS: Female gender and older age were predictors of ND. Headaches or pain in the head, psychological distress, and psychosomatic problems were predictors for all dimensions of ND. These effects were observed for both incidence and persistence of ND. CONCLUSIONS: The results emphasize the role of psychosomatic and psychological factors in the occurrence and course of ND for various dimensions of the disorder.

Adult↗

PubMed2001

Locus of control and health behaviour revisited: a multivariate analysis of young adults from 18 countries.

The inconsistent and small associations between health locus of control and health behaviour found in previous studies may be due to the use of small samples, and an overreliance on correlations as measures of association. We assessed relationships between internal powerful others and chance health locus of control, health values, and ten health-related behaviours (physical exercise, smoking, alcohol consumption, breakfast, tooth-brushing, seat belt use, and consumption of fruit, fat, fibre and salt) in 4358 female and 2757 male university students from 18 European countries. Multivariate logistic modelling, assessing the odds of engaging in healthy behaviour with graded changes in locus of control, identified substantial associations. For five behaviours, the odds of healthy behaviour were more than 40% greater among individuals in the highest vs. lowest quartile of internal locus of control after adjustment for sex, age, health value and other locus of control scales. High chance locus scores were associated with more than 20% reductions in the likelihood of healthy options for six behaviours, while powerful others scores showed more variable associations with healthy actions. Inclusion of health value within the analyses did not change the nature of the relationships observed between variables. Associations between health locus of control and health behaviour are of a similar magnitude to other psychosocial factors when appropriate statistical tests are employed.

Adolescent↗

PubMed1991

Risk indicators for future clinical attachment loss in adult periodontitis. Patient variables.

We studied patient-derived variables to identify individuals at risk for future clinical attachment loss (CAL). Seventy-five patients with chronic adult periodontitis were followed for 6 months and clinical and epidemiological parameters collected at baseline were related to CAL. Clinical parameters were obtained from 6 sites per tooth and whole-mouth averages were calculated. Epidemiologic parameters were obtained by questionnaire and interview. After the baseline examination, patients were treated with root planing and scaling. Thirty-one patients (41.3%) demonstrated greater than or equal to 1 site with CAL of greater than or equal to 2.5 mm, while 16 patients (21.3%) demonstrated CAL at greater than or equal to 2 sites. Epidemiological factors such as gender, health status, marital status, education, and occupation were not associated with CAL. In contrast, baseline mean attachment level, age, baseline mean probing depth, baseline mean recession, percentage of sites exhibiting bleeding on probing, and the number of missing teeth were related to CAL. Using logistic modelling, we found that baseline attachment level was the primary risk indicator for post-treatment CAL. Nineteen percent of the patients with baseline attachment levels less than 4.0 mm, 50% of the patients with 4.0 to 4.9 mm, and 85% (P less than .005) of the patients with greater than or equal to 5.0 mm exhibited CAL. The age of the patient was also a major risk indicator for CAL, and was independent of baseline attachment levels. Eighty-nine percent of the 60 to 69 year old patients demonstrated CAL, compared to only 35% of patients between the ages of 30 and 59 (P less than or equal to .005).(ABSTRACT TRUNCATED AT 250 WORDS)

Adult↗

PubMed2001

Risk factors for nosocomial transmission of multidrug-resistant tuberculosis due to Mycobacterium bovis among HIV-infected patients.

OBJECTIVE: To identify risk factors for transmission of multidrug-resistant tuberculosis (MDR-TB) among hospitalized human immunodeficiency virus (HIV) infected patients exposed during a nosocomial outbreak. DESIGN: Case control study. Cases were HIV-infected patients with MDR-TB due to Mycobacterium bovis (MDR-TBMb) who acquired the disease after exposure to an MDR-TBMb patient in an hospital ward. Controls were HIV-infected patients exposed to a case of MDR-TBMb in an hospital ward but who did not develop MDR-TBMb during the follow-up. RESULTS: Nineteen cases and 31 controls were included. CD4 cell counts were significantly lower among cases. Exposure in the infectious diseases ward or exposure to the index patient were associated with development of MDR-TBMb, while exposure during a single-room hospital stay and exposure in the respiratory isolation ward were associated with non-development of MDR-TBMb. A multiple regression logistic model showed that only a CD4 cell count below 50/microL and exposure to the index patient increased the risk of developing MDR-TBMb (P < 0.05). Hospitalization in a single room seemed to protect HIV-infected patients from developing nosocomial MDR-TBMb (P = 0.052). CONCLUSIONS: Over classic risk factors, such as length of exposure or sharing a room with a case, severe immunosuppression independently increases the risk of MDR-TB transmission in the context of a nosocomial MDR-TB outbreak among HIV-infected patients. This information must be considered in the management of tuberculosis outbreaks. Patients with CD4 cell counts below 50/microL should be the principal group targeted for prevention strategies in nosocomial outbreaks.

Adult↗

PubMed1991

Ultrasound examination in jaundiced patients. Is computer-assisted preclassification helpful?

In this study we attempted to determine the diagnostic accuracy and reproducibility of ultrasonography (US) for jaundice and to see how US can best be combined with preliminary clinical-biochemical diagnoses to plan the invasive work-up. US proved reproducible in two diagnostic departments (127 agreements in 135 cases). But, since obstruction was underdiagnosed (15 double-false negatives), the predictive value of a negative result was only 0.83. By adding a term which represents the US conclusion, obstruction or not, to the Copenhagen pocket diagnostic chart score (based on the logistic model) we found that an obstructive conclusion increases the odds of obstruction by a factor of 25, and a non-obstructive conclusion decreases the odds by a factor of only 1.9. We conclude that the preliminary diagnosis is frequently sufficiently certain to be unalterable by US. This leaves only 40% of the jaundice cases in which US is necessary to plan invasive work-up. The US workload can even, it appears, be reduced to about 22% without appreciable penalty in terms of unrewarding invasive procedures. Using these strict indications, four US examinations seem to suffice to avoid one such error. Relying on either US or clinical-biochemical data alone is inferior to the combined strategy.

Algorithms↗

PubMed2003

Advanced soft-tissue sarcoma: a disease that is potentially curable for a subset of patients treated with chemotherapy.

Adult patients with metastatic or locally advanced irresectable soft-tissue sarcoma (ASTS) are generally considered as incurable. Whether some of these patients achieve long-term survival after first-line treatment with chemotherapy is not known. Patients with ASTS still alive 5 years after initial treatment with a doxorubicin-containing regimen, i.e. long-term survivors, were analysed among the 2187 patients included in first-line chemotherapy protocols between 1976 and 1990 in seven trials of the European Organization for Research and Treatment of Cancer Soft Tissue and Bone Sarcoma Group (EORTC STBSG) group. 1888 patients were followed for at least 5 years. The initial clinical characteristics and the outcome of the long-term survivors were investigated. 66 of the 1888 patients were alive at 5 years and the projected 5-year survival was 8% in this series. Age or histological subtypes were similar in the long-term survivors compared with the other patients. The percentages of females (69%), of grade 1 tumours (35%), of patients with an initial performance status (PS) of 0 (63%) were significantly higher in the long-term survivors while liver metastasis (6% versus 21%) were significantly less frequent. Long-term survivors were observed in all subgroups of patients. 31, 31, 31 and 6% of the long-term survivors were in complete response (CR), partial response (PR), stable disease (SD), and progressive disease (PD), respectively, after the first-line regimen. A CR to a doxorubicin-containing regimen was the major parameter correlated to 5-year survival. In multivariate analysis using a logistic model, independent parameters correlated to 5-year survival were PS, female gender, grade I tumours, and the achievement of a CR after first-line treatment, which was retained as the most powerful predictor for 5-year survival. 10 of the 66 patients died after 5 years in this series, including 8 patients in PD or SD after first-line treatment versus 2 patients in PR or CR (P=0.01). 8% of patients with ASTS are alive 5 years after first-line chemotherapy with a doxorubicin-containing regimen. Long-term survivors are observed in all prognostic subgroups of patients, in particular those achieving a CR to first-line chemotherapy.

Adult↗

PubMed2002

Human immunodeficiency virus type 1 Western blot: revised diagnostic criteria with fewer indeterminate results for epidemiological studies in Africa.

BACKGROUND: Western blot (WB) criteria in epidemiological studies in Africa exhibit an unacceptably high proportion of indeterminate results. New diagnostic criteria are urgently needed. METHODS: From 1989 to 1998, WB confirmatory tests were performed after weakly positive or discordant results of two enzyme immunoassays in a large Ugandan population. Enzyme immunoassays (EIA) on new sera taken prospectively from the same individuals one year later were used to assess the human immunodeficiency virus (HIV) status of these people. A logistic model was used to determine which set of WB bands was the most predictive of HIV status. Diagnostic criteria were then established, based on the likely HIV status determined using the predictive values and the intensity of the bands. RESULTS: Using 1109 WB tests, the best diagnostic criteria were based on only two bands (gp160 and p31). These criteria were validated on an independent sample of 587 WB tests, giving a high sensitivity and specificity (90.3% and 97.0%, respectively) and few indeterminate results (2.7%). These criteria classified correctly 96.3% of the sera. CONCLUSION: Our diagnostic criteria gave far better results in our population than the existing published criteria. This suggests that new criteria could be developed to improve WB interpretation in African settings.

Africa↗