Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “missing data”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

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

At least 1,171 records · Page 65Linked to original sources
PubMed2003

Effect of maxillary sinus augmentation on the survival of endosseous dental implants. A systematic review.

BACKGROUND: Grafting the floor of the maxillary sinus has become the most common surgical intervention for increasing alveolar bone height prior to the placement of endosseous dental implants in the posterior maxilla. Outcomes of this procedure may be affected by specific surgical techniques, simultaneous versus delayed implant placement, use of barrier membranes over the lateral window, selection of graft material, and the surface characteristics and the length and width of the implants. RATIONALE: The primary objective of this systematic review was to determine the efficacy of the sinus augmentation procedure and compare the results achieved with various surgical techniques, grafting materials, and implants. FOCUSED QUESTION: In patients requiring dental implant placement, what is the effect on implant survival of maxillary sinus augmentation versus implant placement in the non-grafted posterior maxilla? SEARCH PROTOCOL: MEDLINE, the Cochrane Oral Health Group Specialized Trials Register, and the Database of Abstracts and Reviews of Effectiveness were searched for articles published through April 2003. Hand searches were performed on Clinical Oral Implants Research, International Journal of Oral and Maxillofacial Implants, and the International Journal of Periodontics & Restorative Dentistry and the bibliographies of all relevant papers and review articles. In addition, researchers, journal editors, and industry sources were contacted to see if pertinent unpublished data that had been accepted for publication were available. SELECTION CRITERIA INCLUSION CRITERIA: Human studies with a minimum of 20 interventions, a minimum follow-up period of 1-year loading, an outcome measurement of implant survival, and published in English, regardless of the evidence level, were considered. EXCLUSION CRITERIA: Studies involving multiple simultaneous interventions (e.g., simultaneous ridge augmentation) and studies with missing data that could not be supplied by the study authors were excluded. DATA COLLECTION AND ANALYSIS: Where adequate data were available, subgroups of dissimilar interventions (e.g., surgical techniques, graft materials, implant surfaces, membranes) were isolated and subjected to meta-regression, a form of meta-analysis. MAIN RESULTS: 1. Forty-three studies, 3 randomized controlled clinical trials (RCTs), 5 controlled trials (CTs), 12 case series (CS), and 23 retrospective analyses (RA) were identified. Thirty-four were lateral window interventions, 5 were osteotome interventions, 2 were localized management of the sinus floor, and 2 involved the crestal core technique. 2. Meta-regression was performed to determine the effect of the variables of block versus particulate grafting techniques, implant surface, graft material, and the use of a membrane over the lateral window. 3. The survival rate of implants placed in sinuses augmented with the lateral window technique varied between 61.7% and 100%, with an average survival rate of 91.8%. For lateral window technique: 4. Implant survival rates reported in this systematic review compare favorably to reported survival rates for implants placed in the non-grafted posterior maxilla. 5. Rough-surfaced implants have a higher survival rate than machine-surfaced implants when placed in grafted sinuses. 6. Implants placed in sinuses augmented with particulate grafts show a higher survival rate than those placed in sinuses augmented with block grafts. 7. Implant survival rates were higher when a membrane was placed over the lateral window. 8. The utilization of grafts consisting of 100% autogenous bone or the inclusion of autogenous bone as a component of a composite graft did not affect implant survival. 9. There was no statistical difference between the covariates of simultaneous versus delayed implant placement, types of rough-surfaced implants, length of follow-up, year of publication, and the evidence level of the study. REVIEWERS' CONCLUSIONS: Insufficient data were present to statistically evaluate the effects of smoking, residual crestal bone height, screw versus press-fit implant design, or the effect of implant surface micromorphology other than machined versus rough surfaces. There are insufficient data to recommend the use of platelet-rich plasma in sinus graft surgery.

Bone Transplantation↗
PubMed1997

Logbook data as a source of birth information. Are they valid?

OBJECTIVE: To determine if the data recorded in the labor and delivery logbook in two hospitals are complete and consistent with the information in the patient's medical record. STUDY DESIGN: We performed a retrospective comparison of the information in the labor and delivery logbook to the content of 110 patients' hospital charts in each of two hospitals. RESULTS: The logbooks of both hospitals had erroneous entries and missing data as compared to the antepartum complications, intrapartum events and immediate newborn care recorded in the patients' medical records. The range of error in these three categories was 61%, 31% and 60% in one hospital and 37%, 29% and 30% in the other hospital. CONCLUSION: The results of this study do not support the use of logbooks as a source of data for individual or collated departmental, hospital or agency reports containing antepartum, intrapartum and newborn information.

Delivery, Obstetric↗
PubMed2006

A comparison of several methods for haplotype frequency estimation and haplotype reconstruction for tightly linked markers from general pedigrees.

Haplotype inference for tightly linked markers from general pedigrees remains a challenging problem. Only a few methods are available to efficiently and accurately estimate haplotype frequencies and reconstruct haplotypes for a large number of tightly linked markers from general pedigrees in the presence of missing data, and their performance has not been carefully and extensively evaluated. In this paper, we compare four published methods for haplotype reconstruction and frequency estimation for tightly linked markers from general pedigrees, including HAPLORE, GENEHUNTER, PedPhase, and MERLIN. We review these methods and discuss the differences between them in terms of the models and computational strategies employed. We assess their performance based on simulations using pedigrees and haplotypes on tightly linked single nucleotide polymorphisms from real studies. We investigate the effect of several factors, including the missing rate, the departure from Hardy-Weinberg Equilibrium, and the sample size, on the accuracy for haplotype inference. We also compare these methods with a widely used method for haplotype inference from unrelated individuals, PHASE, by treating individuals within a pedigree as unrelated samples. This comparison allows us to investigate the relative efficiency in haplotype inference using pedigree data. Our results indicate that incorporation of pedigree information can improve the precision for haplotype frequency estimation and the accuracy for haplotype reconstruction. Among four haplotyping methods capable of analyzing general pedigrees, HAPLORE and MERLIN have comparable performance and outperform the other two methods in almost all situations.

Computer Simulation↗
PubMed2004

Validation of the Taiwan Birth Registry using obstetric records.

BACKGROUND AND PURPOSE: Birth certificates are a major source of population-based data often used in maternal and child health research, but their value depends upon the level of accuracy of the data. The purpose of this study was to assess the validity of the birth registry data in Taiwan. METHODS: Obstetric records of a total of 2779 infants born at a municipal hospital in Taipei between 1995 and 1997 were abstracted and linked to the Taiwan Birth Registry (TBR). Concordance was presented by agreement percentages for various birth characteristics including gender, birth order, birth weight, and gestational age. Concordance was further assessed using Cohen's k and sensitivity/specificity for both birth weight and gestational age, with both variables categorized according to their customary definitions. RESULTS: Missing data for specific items in the TBR ranged from 0% (gender) to 1.6% (birth order). The birth registry accurately reported all 4 selected items, with the highest and lowest concordance percentage noted for gender (99.0%) and gestational age (88.3%), respectively. The level of agreement indicated by the Cohen's k statistic ranged from 0.92 to 0.96 for birth weight and gestational age when both variables were treated as categorical variables. CONCLUSIONS: The TBR showed a low rate of missing information and high levels of validity for the elements frequently used in maternal and child health research in Taiwan.

Birth Certificates↗
PubMed1993

Evaluating the effects of tubal sterilization on menstrual function: selected issues in data analysis.

We examined selected issues in data analysis in the Collaborative Review of Sterilization (CREST). CREST is a multicentre, prospective, observational study of women undergoing tubal sterilization. We analysed menstrual function after sterilization in over 5000 women who were enrolled in the period 1978-1983 and followed for 5 years with yearly follow-up interviews. To take into account the dependency among repeated responses from the same individuals, we used the generalized estimating equations (GEE) approach to longitudinal data analysis. Marginal modelling resulted in a statistically significant increase in the odds of menstrual dysfunction at 5 years after tubal sterilization. Transitional modelling produced rates of menstrual dysfunction given a woman's menstrual function at baseline, after adjusting for other baseline characteristics such as method of contraception before sterilization. To examine the direction of the bias that could result from non-random missing data, we refitted our models using imputed values. The models with imputed values showed the same trends as the original models.

Bias↗
PubMed2002

Handling missing marker coordinates in 3D analysis.

The use of mathematical formulations to retrieve missing data is not usually implemented as an automated computer procedure. The purpose of this paper is to present a simple mathematical formulation that can be used to retrieve the coordinates of markers hidden during a task. A strategy based on the smoothness of the variation of distance between points is devised, which leads to an automated computer procedure for retrieving hidden skin markers or missing joint coordinates (missing 3D points). This approach has the advantage of not being restricted to the usual rigid-body assumption.

Algorithms↗
PubMed2004

Research fundamentals: follow-up of subjects in clinical trials: addressing subject attrition.

Many published clinical trials have less than adequate follow-up. When conducting a clinical trial, researchers attempt to minimize data loss; however, some data may not be collected, particularly when subjects are lost to follow-up. Careful planning of research protocols, including comprehensive initial data collection, identification of locators, flexible scheduling, systematic subject tracking, monitoring subject loss, and systematically approaching problem cases can ensure high follow-up rates. This article presents a compendium of techniques and procedures that researchers can use to enhance follow-up and address attrition in their studies. Finally, this article outlines statistical techniques that can be used to address the effects of missing data, particularly when patients are lost to follow-up.

Appointments and Schedules↗
PubMed2005

Using patients' and rheumatologists' opinions to specify a short form of the WOMAC function subscale.

BACKGROUND: The WOMAC (Western Ontario and McMaster Universities) function subscale is widely used in clinical trials of hip and knee osteoarthritis. Reducing the number of items of the subscale would enhance efficiency and compliance, particularly for use in clinical practice applications. OBJECTIVE: To develop a short form of the WOMAC function subscale based on patients' and experts' opinions (WOMAC function short form). METHODS: WOMAC function subscale data (Likert version) were obtained from 1218 outpatients with painful hip or knee osteoarthritis. These patients and their rheumatologists selected the five items that they considered most in need of improvement. The rheumatologists were asked to select the five items for which patients in general are the most impaired. Items that were least important to patients and experts, those with a high proportion of missing data, and those with a response distribution showing a floor or ceiling response were excluded, along with one of a pair of items with a correlation coefficient >0.75. RESULTS: The WOMAC function short form included items 1, 2, 3, 6, 7, 8, 9, and 15 of the long form. The short form did not differ substantially from the long form in responsiveness (standardised response mean of 0.84 v 0.80). CONCLUSIONS: A short form of the WOMAC function subscale was developed according to the views of patients and rheumatologists, based on the responses of 1218 patients and 399 rheumatologists. The clinical relevance and applicability of this WOMAC function subscale short form require further evaluation.

Activities of Daily Living↗
PubMed2002

Gender differences in life expectancy free of impairment at older ages.

This article uses data from the United Kingdom Medical Research Council Cognitive Function and Ageing study (MRC CFAS) to analyze morbidity associated with three areas of impairment. We use cognitive status, functional status, and physical illness to examine differences in the proportion of time that older women and men will spend with co-morbidity. We also analyze differences among various impairments, and investigate the relationship between missing data and sex. Women have a larger burden of impairment than men, and, by including cognitive impairment together with functional impairment, a very large impairment burden is highlighted at all ages. Policy implications of the findings from the perspective of older women in the United Kingdom are discussed.

Activities of Daily Living↗
PubMed2003

Sample size estimation for GEE method for comparing slopes in repeated measurements data.

Sample size calculation is an important component at the design stage of clinical trials. Controlled clinical trials often use a repeated measurement design in which individuals are randomly assigned to treatment groups and followed-up for measurements at intervals across a treatment period of fixed duration. In studies with repeated measurements, one of the popular primary interests is the comparison of the rates of change in a response variable between groups. Statistical models for calculating sample sizes for repeated measurement designs often fail to take into account the impact of missing data correctly. In this paper we propose to use the generalized estimating equation (GEE) method in comparing the rates of change in repeated measurements and introduce closed form formulae for sample size and power that can be calculated using a scientific calculator. Since the sample size formula is based on asymptotic theory, we investigate the performance of the estimated sample size in practical settings through simulations.

Data Interpretation, Statistical↗
PubMed1991

Practice databases and their uses in clinical research.

A few large clinical information databases have been established within larger medical information systems. Although they are smaller than claims databases, these clinical databases offer several advantages: accurate and timely data, rich clinical detail, and continuous parameters (for example, vital signs and laboratory results). However, the nature of the data vary considerably, which affects the kinds of secondary analyses that can be performed. These databases have been used to investigate clinical epidemiology, risk assessment, post-marketing surveillance of drugs, practice variation, resource use, quality assurance, and decision analysis. In addition, practice databases can be used to identify subjects for prospective studies. Further methodologic developments are necessary to deal with the prevalent problems of missing data and various forms of bias if such databases are to grow and contribute valuable clinical information.

Clinical Medicine↗
PubMed1998

Neonatal circumcision. Randomized trial of a sucrose pacifier for pain control.

OBJECTIVE: To assess the effectiveness of oral sucrose via a nipple compared with no treatment and dorsal penile nerve block (DPNB) for alleviating pain in neonatal circumcision. DESIGN: Randomized control trial. Data analysis performed by investigators blinded to the 3 treatment groups. SETTING: University teaching hospital, General Care Nursery. PATIENTS: One hundred nineteen full-term male, normal birth weight neonates, 12 hours old or older. INTERVENTIONS: No treatment (our standard care), DPNB, or oral sucrose prior to circumcision. MAIN OUTCOME MEASURES: Differences between groups in heart rate and oxygen saturation changes from baseline during specified intervals of the circumcision procedure. Differences between groups in loss of data due to episodes of excessive motion. RESULTS: Sucrose gave significant (P < .001) pain relief compared with the no treatment control throughout most of the circumcision and particularly in the early stages of the procedure. Overall, the average difference in the elevation of heart rates during the circumcision operative procedure among the 3 groups and the 95% confidence intervals (CIs) were as follows: control vs DPNB, 27.1 beats/min (17.6, 36.6) and control vs sucrose, 9.7 beats/min (0.1, 19.3). Furthermore, newborns who received either DPNB or sucrose had less loss of oxygen saturation data due to excessive motion during the procedure than the no-treatment controls. The total percentages of lost data due to excessive motion in the 3 groups were 31% for control, 10% for DPNB, and 8% for sucrose. Relative risk and 95% CIs were: DPNB vs control, 0.32 (0.23, 0.43); sucrose vs control, 0.26 (0.18, 0.36). Differences in oxygen saturation among the 3 groups during the circumcision operative procedure were statistically (P < .001), but perhaps not clinically, significant. However, the analysis did not include missing data due to excessive motion, which occurred predominantly in the no-treatment control group. CONCLUSION: Sucrose on a pacifier is an inexpensive and effective method for pain relief in neonatal circumcision when DPNB is not desirable.

Analgesics↗
PubMed1984

A note on data monitoring, incomplete data and curtailed testing.

A question often asked in a fixed sample size clinical trial is whether the final outcome has been determined before the data has been completely collected. Recent papers in this journal and elsewhere have considered this question from the viewpoint of curtailed testing. In this paper, a related concept, "the probability of reversal" of a test procedure, is described and associated least upper bounds are given that are pertinent to the issue. The results are also of relevance in assessing, at the end of the trial, whether missing data are of consequence in the test of significance.

Clinical Trials as Topic↗
PubMed2007

At what age can children report dependably on their asthma health status?

OBJECTIVE: This study examined psychometric properties and feasibility issues surrounding child-reported asthma health status data. METHODS: In separate interviews, parents and children completed 3 visits. Child questionnaires were interviewer administered. The primary instrument was the Children's Health Survey for Asthma-Child Version, used to compute 3 scales (physical health, activities, and emotional health). The following were assessed: reliability (internal consistency and test-retest reliability), validity (general health status, symptom burden, and lung function), and feasibility (completion time, missing data, and inconsistent responses). RESULTS: A total of 414 parent-child pairs completed the study (mean child age: 11.5 years). Reliability estimates for the activities and emotional health scales were > .70 in all but 1 age category; 5 of 9 age groups had acceptable internal consistency ratings (> or = .70) for the physical health scale. Cronbach's alpha tended to increase with child age. In general, test-retest correlations between forms and intraclass correlation coefficients were strong for all ages but tended to increase with child age. Correlations between forms ranged from .57 (7-year-old subjects, physical health) to .96 (14-year-old subjects, activities). Intraclass correlation coefficients ranged from .76 (13-year-old subjects, emotional health) to .94 (15-16-year-old subjects, physical health). Children with less symptom burden reported higher mean Children's Health Survey for Asthma-Child Version scores (indicating better health status) for each scale, at significant levels for nearly all age groups. Children's Health Survey for Asthma-Child Version completion times decreased from 12.9 minutes at age 7 to 6.9 minutes at age 13. CONCLUSIONS: This research indicates that children with asthma as young as 7 may be dependable and valuable reporters of their health. Data quality tends to improve with age.

Adolescent↗
PubMed2004

Delayed antibiotics for symptoms and complications of respiratory infections.

BACKGROUND: The use of antibiotics for upper respiratory tract infections is controversial. Any benefits have to be weighed against common adverse reactions (including rash, abdominal pain, diarrhoea and vomiting), cost and antibacterial resistance. There has been interest in ways to reduce antibiotic prescribing for acute respiratory infections. One is delaying the use of prescribed antibiotics by more than 48 hours for acute upper respiratory tract infections. Such methods have been shown to reduce prescribing. This review asks what effect this practice has on the clinical course of the illness. OBJECTIVES: To evaluate the clinical effect of delayed antibiotic use in acute upper respiratory tract infections compared to immediate use of antibiotics SEARCH STRATEGY: The following electronic databases were searched: the Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library Issue 1, 2004) which includes the Acute Respiratory Infection Groups' specialised register; MEDLINE (January 1966 to January Week 1 2004), EMBASE (1990 to September 2003) and Current Contents (1998 to 2003). The search was carried out by an expert librarian. Abstracts of identified articles were used to determine which studies were trials. SELECTION CRITERIA: Randomised controlled trials involving patients of all ages defined as having acute otitis media, acute pharyngitis, sore throat, common cold, a viral upper respiratory tract infection, acute sinusitis, and acute bronchitis were included in which delayed antibiotics are compared to antibiotics used immediately. Delayed antibiotic use was defined as the use of or advice to use antibiotics more than 48 hours after the initial consultation. 'Immediate antibiotic use' was defined as the immediate use of oral antibiotics given at the initial consultation. Clinical outcomes measured included: the presence or absence of fever, cough, pain, duration and severity of illness, complications of the disease, adverse effects from the antibiotics. Trial quality was assessed independently by two reviewers who were blinded to the author, journal and results of each study. DATA COLLECTION AND ANALYSIS: Data was collected by two reviewers who were blinded to the author and journal. Data were analysed and reported using RevMan. MAIN RESULTS: Seven trials were eligible on the basis of design and all reported patient-centred outcomes. Methodological quality of included trials was generally high. There was no difference between immediate and delayed antibiotic groups for symptoms on day one and day seven. For most symptom measures there was no significant difference between the immediate and delayed antibiotic groups. Missing data and marked heterogeneity between study outcomes prevented pooling of results as a meta-analysis. Three studies out of six reporting fever, all involving patients with sore throat, indicated that there was more fever in the delayed antibiotic group. The remaining three studies showed no difference. There was no significant symptom difference for patients with cough or the common cold between the two intervention groups. Pain and malaise severity scores at day three significantly favoured the immediate antibiotic group in children with acute otitis media (Little 2001). In this study by Little 2001 of children with otitis media proxies for other malaise related outcomes were reported, including 'last day of crying' which favoured the immediate antibiotic group by approximately 16 hours (0.69 days; 95% CI 0.31 to 1.07). In the same study, just over half a spoon of paracetamol a day less was used in the immediate antibiotic group (0.59; 95% CI 0.25 to 0.93). There was no significant difference between the intervention groups for the adverse outcome of rash. Two studies reported the outcome of vomiting which was reduced in the immediate antibiotic group in children with suspected streptococcal pharyngitis in El-Daher 1991 but there was no difference in children with sore throat in Little 1997. Diarrhoea was reported by three studies of which two showed no difference Little 1997; Arroll 2002a while Little 2001 reported less diarrhoea in the delayed antibiotic group in children with otitis media. REVIEWERS' CONCLUSIONS: When considering treatment options for upper respiratory tract infections, the option of delayed antibiotics has been used in an attempt to reduce the use of antibiotic prescriptions. This review shows that for all symptom scores the evidence varies between trials. Most symptom outcomes show no difference between immediate and delayed antibiotic groups. Three of the six studies, all involving patients with sore throat, indicated that patients in the delayed antibiotic group had significantly more fever that their counterparts in the immediate antibiotic group. The other three showed no difference for the outcome of fever. There is evidence indicating that for children with otitis media, pain and malaise scores are worse in the delayed antibiotic group compared to the immediate antibiotic group. This price must be weighed up against the benefits of reduced antibiotic prescribing. Future randomised controlled trials of delaying antibiotics as an intervention should fully report symptoms as well as changes of prescription rates.

Anti-Bacterial Agents↗
PubMed2007

Augmented generalized SENSE reconstruction to correct for rigid body motion.

The correction of motion artifacts continues to be a significant problem in MRI. In the case of uncooperative patients, such as children, or patients who are unable to remain stationary, the accurate determination and correction of motion artifacts becomes a very important prerequisite for achieving good image quality. The application of conventional motion-correction strategies often produces inconsistencies in k-space data. As a result, significant residual artifacts can persist. In this work a formalism is introduced for parallel imaging in the presence of motion. The proposed method can improve overall image quality because it diminishes k-space inconsistencies by exploiting the complementary image encoding capacity of individual receiver coils. Specifically, an augmented version of an iterative SENSE reconstruction is used as a means of synthesizing the missing data in k-space. Motion is determined from low-resolution navigator images that are coregistered by an automatic registration routine. Navigator data can be derived from self-navigating k-space trajectories or in combination with other navigation schemes that estimate patient motion. This correction method is demonstrated by interleaved spiral images collected from volunteers. Conventional spiral scans and scans corrected with proposed techniques are shown, and the results illustrate the capacity of this new correction approach.

Algorithms↗
PubMed2005

Mortality from venous thromboembolism in young Swedish women and its relation to pregnancy and use of oral contraceptives--an approach to specifying rates.

BACKGROUND: Pregnancy and use of combined oral contraceptives (COCs) are major risk factors for venous thromboembolism (VTE) in young women and we wanted to obtain accurate VTE mortality data overall, by age, associated with the use of COCs and pregnancy. METHODS: From the Swedish Cause of Death Register (CDR) we identified women aged 15-44 with VTE as underlying or contributory cause of death during the period 1990-1999. We scrutinized medical records and included verified VTE cases without active cancer or terminal disease. Pregnancy statistics and COC utilization data were obtained from national databases. RESULTS: Of the 120 cases included, 9 (8%) were associated with pregnancy and 28 (23%) with current COC use. The overall refined VTE mortality rate in current COC users was 7.5[4.7; 10.3] per million user-years and the corresponding pregnancy-related rate was 8.9[4.1;17.0] per million pregnancy years, rates increasing with age. For ages 15-24, the rate was significantly higher in current COC users than in non-pregnant women not using COCs: 6.0[3.1; 10.5] per million user-years vs. 0.3[0.0; 1.2] per million woman years. Underlying cause mortality data included 82% of VTE deaths associated with COCs, and 56% of maternal deaths had a pregnancy-related code. CONCLUSION: Mortality figures from VTE associated with the use of COCs and pregnancy were similar. COC use had an important impact on the total VTE mortality in the youngest age group. Standard mortality statistics do not allow accurate monitoring of VTE mortality in young women due to missing data, misdiagnoses and coding rules.

Adolescent↗
PubMed1994

Discrimination between stages of Alzheimer's disease with subsets of Mini-Mental State Examination items. An analysis of Consortium to Establish a Registry for Alzheimer's Disease data.

OBJECTIVE: To identify minimal sets of Mini-Mental State Examination (MMSE) items that can distinguish normal control subjects from patients with mild Alzheimer's disease (AD), patients with mild from those with moderate AD, and those with moderate from those with severe AD. DESIGN: Two randomly selected equivalent half samples. Results of logistic regression analysis from data from the first half of the sample were confirmed by receiver operating characteristic curves on the second half. SETTING: Memory disorders clinics at major medical centers in the United States affiliated with the Consortium to establish a Registry for Alzheimer's Disease (CERAD). PARTICIPANTS: White, normal control subjects (n = 412) and patients with AD (n = 621) who met CERAD criteria; nonwhite subjects (n = 165) and persons with missing data (n = 27) were excluded. MAIN OUTCOME MEASURES: Three four-item sets of MMSE items that discriminate, respectively, (1) normal controls from patients with mild AD, (2) patients with mild from those with moderate AD, and (3) patients with moderate from those with severe AD. RESULTS: The MMSE items discriminating normal controls from patients with mild AD were day, date, recall of apple, and recall of penny; those discriminating patients with mild from those with moderate AD were month, city, spelling world backward, and county, and those discriminating patients with moderate from those with severe AD were floor of building, repeating the word table, naming watch, and folding paper in half. Performance on the first two four-item sets was comparable with that of the full MMSE; the third set distinguished patients with moderate from those with severe AD better than chance. CONCLUSIONS: A minimum set of MMSE items can effectively discriminate normal controls from patients with mild AD and between successive levels of severity of AD. Data apply only to white patients with AD. Performance in minorities, more heterogeneous groups, or normal subjects with questionable cognitive status has not been assessed.

Aged↗