Interventions to increase physician efficiency and comfort with an electronic health record system.

OBJECTIVE: To determine comfort when using the Electronic Health Record (EHR) and increase in documentation efficiency after an educational intervention for physicians to improve their transition to a new EHR. METHODS: This study was a single-center randomized, parallel, non-blinded controlled trial of real-time, focused educational interventions by physician peers in addition to usual training in the intervention arm compared with usual training in the control arm. Participants were 44 internal medicine physicians and residents stratified to groups using a survey of comfort with electronic media during rollout of a system-wide EHR and order entry system. Outcomes were median time to complete a progress note, notes completed after shift, and comfort with EHR at 20 and 40 shifts. RESULTS: In the intervention group, 73 education sessions averaging 14.4 (SD: 7.7) minutes were completed with intervention group participants, who received an average of 3.47 (SD: 2.1) interventions. Intervention group participants decreased their time to complete a progress note more quickly than controls over 30 shifts (p < 0.001) and recorded significantly fewer progress notes after scheduled duty hours (77 versus 292, p < 0.001). Comfort with EHRs increased significantly in both groups from baseline but did not differ significantly by group. Intervention group participants felt that the intervention was more helpful than their standard training (3.47 versus 1.95 on 4-point scale). CONCLUSION: Physicians teaching physicians during clinical work improved physician efficiency but not comfort with EHRs. More study is needed to determine best methods to assist those most challenged with new EHR rollouts.

The vibration quantitation scale (VQS): a simple, reproducible bedside measure of sensory function in multiple sclerosis.

OBJECTIVES: To assess the utility of a bedside measure of sensation (the Vibration Quantitation Scale (VQS)) in patients with multiple sclerosis (MS) and in normal controls. To correlate the VQS with the Kurtzke Expanded Disability Severity Score (EDSS) and sensory abnormalities in these patients. METHODS: We developed the VQS and tested its performance in patients with MS of various ages, MS types, and EDSS scores. We compared this with controls (normal volunteers or patients with other neurological diseases) who did not have sensory symptoms. In a subgroup, two examiners measured VQS independently at the same patient visit. A standard C-128 tuning fork was used for the VQS measurement. RESULTS: The VQS had a good inter-observer reproducibility (r=0.920, p<0.001). The VQS fell with increasing age in normals consistent with declining sensory function. The VQS was significantly lower in the multiple sclerosis patients compared with age - matched controls (p<0.001). Abnormalities in VQS were present in patients with brief duration of MS (<5 years) and low EDSS scores, correlating with the presence of sensory abnormalities early in the disease course in some patients. There was a strong correlation between the VQS and EDSS (r=-.509). The VQS correlated with abnormal sensation in the hands (r=0.310), but did not meet statistical significance for abnormal sensation in the feet or face. A second cohort of MS patients was studied using a modified VQS measure (single stimulation, omitting forehead measurement). This reconfirmed the correlation between the modified VQS and EDSS as well as with age. The modified VQS may be useful in clinical practice since it takes little time and is strongly correlated with the EDSS (r=0.578). CONCLUSIONS: The VQS provides a continuous sensory scale applicable in most patients with MS, which is measurable with standard bedside equipment, and which may avoid some of the pitfalls of sensory scoring in MS.

Preinjury warfarin does not impact outcome in trauma patients.

OBJECTIVE: The objective of this study was to determine whether the preinjury condition of anticoagulation had an adverse impact on patients sustaining injury. METHODS: A retrospective analysis was performed for prospectively collected registry data from 1995-2000 from all accredited trauma centers in Pennsylvania. The registry was queried for all trauma patients who had anticoagulation therapy as a preinjury condition (PIC). This group served as our experimental cohort. A control cohort (not having warfarin therapy as a PIC) was developed using case-matching techniques for age, sex, Glasgow Coma Scale (GCS), Injury Severity Score (ISS), A Severity Characterization of Trauma (ASCOT) score, and in the head injured patients, International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnoses. Head and non-head injured patients were evaluated separately. The cohorts were examined for 28-day mortality, intensive care unit length of stay (ICU-LOS), hospital length of stay (HOS-LOS), PICs, occurrences, discharge destinations, and functional status at discharge. Chi2 and Student's t test were used to evaluate the data; p values < 0.05 were considered significant. RESULTS: Two thousand nine hundred forty-two patients were available for analysis. The prevalence of PICs was significantly greater in the warfarin group for both the head and non-head injured populations (p < 0.003 and p < 0.0001, respectively). The incidence of occurrences in the non-head injured population was statistically higher for the warfarin patients (p < 0.001), but showed no difference in the head injured group regardless of warfarin use (p = 0.15). Functional status at discharge demonstrated no clinically significant difference between the warfarin and non-warfarin groups in both head and non-head injured populations. There was no difference in discharge destination in the head injured population; however, in the non-head injured population a greater percentage of non-warfarin patients was discharged to home when compared with the warfarin patients. CONCLUSION: Our data suggest that the PIC of anticoagulation with warfarin does not adversely impact mortality or LOS outcomes in both head and non-head injured patients. In non-head injured patients, however, the occurrence rates and discharge destination were different. More research needs to be done to determine whether this is related to anticoagulation or other reasons (i.e., number of PICs). These data should be used when weighing risk/benefit ratios of prescribing chronic anticoagulation.

Outcome analysis of Pennsylvania trauma centers: factors predictive of nonsurvival in seriously injured patients.

BACKGROUND: The purpose of this study was to evaluate the impact of five trauma center characteristics on survival outcome in nine serious injury categories. METHODS: A retrospective analysis of prospectively collected data from 1992 to 1996 on patients older than 14 years of age from 24 accredited trauma centers in Pennsylvania was performed. Trauma center characteristics selected for evaluation were level of accreditation, volume of trauma admissions, presence of in-house trauma surgeons, presence of a surgical residency program, and presence of an on-site medical school. Each of these characteristics was evaluated to determine its impact on survival in the selected serious injuries. A logistic regression model was then created to evaluate the most seriously injured patients as defined by A Severity Characterization of Trauma score of < 0.50. On the basis of the logistic regression model, odd ratios were calculated treating low volume as a significant risk factor for mortality. RESULTS: Of the 88,723 patients meeting registry criteria, 13,942 met the serious injury criteria. Independent analysis suggested that accreditation was beneficial regardless of level, volume of patients treated had a direct impact on survival outcome, and the presence of a surgical residency program may confer survival benefit. Of the 13,942 patients with serious injuries, those with A Severity Characterization of Trauma score of < 0.5 were selected for evaluation by logistic regression (n = 3,562). The logistic regression model, however, showed that only volume of patients treated had a consistent association with improved survival. Odds ratio analysis revealed low volume as a significant risk factor for mortality in seven of the nine injuries studied. CONCLUSION: In this analysis, only volume of patients treated had a direct impact on survival outcome. Accreditation, regardless of level, appears to be beneficial.

Establishing reliability and validity of the critical care family satisfaction survey.

OBJECTIVE: To develop and validate the Critical Care Family Satisfaction Survey as a proxy for patient satisfaction. DESIGN: Instrument validation study. SETTING AND TIME FRAME: The Medical Intensive Care, Shock Trauma, Acute Coronary Care, Central Nervous System, Surgical Intensive Care, and Special Care units of Lehigh Valley Hospital (Allentown, PA), for the period December 1997 through September 1998. PATIENTS/PARTICIPANTS: One family member for each of 237 critical care patients. INTERVENTION(S): Content and construct validity were examined on 37 items and 6 constructs thought to measure family satisfaction with the quality of critical care in hospitals. Initially, 14 items and 1 construct were removed from the questionnaire based on this analysis. It was then administered to 237 family members. MEASUREMENTS AND MAIN RESULTS: Factor analysis and confirmatory factor analysis using path models were performed. Internal consistency using Pearson correlations and Cronbach's alpha, and discriminant validation were also calculated. Factor analysis yielded a single eigenvalue >1 (3.712), whereas confirmatory factor analysis led to the final instrument being reduced to 20 items and 5 subscale constructs. One subscale ("Comfort") performed poorly, indicating the possible need for a four-factor model. Subsequently, internal consistency assessed by Cronbach's alpha was 0.9101 for the five-factor model and 0.9327 for the four-factor model. Subscale correlations were no lower than 0.750 for the five-factor model and 0.856 for the four-factor model. CONCLUSIONS: This study provides support that the Critical Care Family Satisfaction Survey-which yields five subscales, "Assurance," "Information," "Proximity," "Support," and "Comfort"--is reliable and valid. Using five constructs rather than four is recommended because of the following: a) the internal consistency loss of 0.0226 for the "Comfort" subscale is not enough to warrant its removal, b) a four-factor questionnaire can be administered and totaled independently of this subscale, c) the need for the fifth construct is indicated by this study's results, and d) including the extra data may allow for more detailed analysis.

Influence of restraining devices on patterns of pediatric facial trauma in motor vehicle collisions.

In the Commonwealth of Pennsylvania, it is required that all children under the age of 4 years be restrained by an infant seat or car seat appropriate for their age and weight. Furthermore, all individuals riding in the front seat must be restrained by a seatbelt. This study examined the relationship between patterns of facial injuries and the use of restraining devices in the pediatric population. A retrospective analysis was performed on motor vehicle collision data submitted to the Pennsylvania Trauma Outcome Study database from 1990 through 1995. Criteria for submission included trauma patients who were admitted to the Intensive Care Unit, those who died during hospitalization, those who were hospitalized for more than 72 hours, or those who were transferred in or out of the receiving hospital. A subset of 412 pediatric patients, 15 years of age or younger, was analyzed for patterns of facial injury and the presence or absence of restraining devices. Restraining devices were categorized as a car seat or a seatbelt. Statistical analysis was performed using chi-square and Fisher's exact tests. Of the 412 pediatric patients, only 17 children were restrained with a car seat and 121 were wearing a seatbelt. A total of 30 children sustained facial fractures, and 50 children suffered facial lacerations. There was a statistically significant increase in the incidence of facial fractures with increasing age of the child (p < 0.001). Of children with facial fractures, 70 percent of those 5 to 12 years old and 90 percent of those 13 to 15 years old were unrestrained (p = 0.166). In conclusion, despite legislation mandating the use of restraints, a large proportion of children involved in motor vehicle collisions were unrestrained. Furthermore, there seems to be a direct relationship between the age of a child and the incidence of facial fractures sustained in motor vehicle collisions.

Comparison of a new method for computer analysis with standard techniques for measuring survival rates in the rat transverse rectus abdominis musculocutaneous flap.

Many methods have been used to measure experimental flap survival in animal models. These have previously included planimetry, computer programs, and (routinely by the authors) the simple measurement of the weight of a template traced according to the dimensions of a given flap. To test the validity of their method, an alternative technique was developed by the authors' biostatiticians. Initially, in a study of rat transverse rectus abdominis musculocutaneous (TRAM) flaps, all templates were also scanned electronically to form bitmaps. Using commonly available software programs, each bitmap was colorized corresponding to viable and nonviable areas of the flaps. A summation of individual pixels by color could then be used to calculate the percentage of flap survival. A comparison of both methodologies in 64 rat TRAM flaps found that the survival rate was slightly greater overall when a computer analysis had been performed (3.1 +/- 1.7%), but this was not a significant difference (p = 0.479). Thus, the previous standard method was proven reliable, but just as precise calculations are now possible using the convenience of any personal computer.

Measurement of endotracheal tube cuff leak to predict postextubation stridor and need for reintubation.

BACKGROUND: The purpose of this study was to determine the predictive value of an endotracheal tube cuff leak for the development of postextubation stridor and the need for reintubation. STUDY DESIGN: Consecutive trauma patients who required intubation at a level I trauma center from July 1997 to July 1998 were studied prospectively. Pediatric patients and those who did not meet the standard weaning protocol criteria established by the Division of Trauma and Surgical Critical Care were excluded. Injury Severity Score, endotracheal tube size, reason for intubation, and the number of days intubated before the initial extubation attempt were recorded. At the time of extubation, the difference in exhaled tidal volume from before to after endotracheal tube cuff deflation was calculated. This number was then divided by the exhaled tidal volume before cuff deflation and was recorded as the percent cuff leak. Patients were followed for 24 hours after extubation for the development of stridor or need for reintubation. Statistical analysis to compare subgroups of patients was performed using ANOVA with Scheffé post hoc analysis. RESULTS: Among the 110 patients analyzed, the most common reason for intubation was closed-head injury. Seven patients (6.4%) developed stridor alone and had a mean cuff leak of 5 8 mL (8.4% of tidal volume before cuff deflation). Six patients (5.5%) experienced stridor that required reintubation and had a mean cuff leak of 68 mL (9.2% of tidal volume before cuff deflation). Patients who developed stridor or needed reintubation had been intubated for a significantly greater length of time than those not developing stridor or requiring reintubation (2.6 versus 3.0 days, p < 0.001). There were no differences in Injury Severity Score, endotracheal tube size, or reason for intubation between these groups. CONCLUSIONS: A cuff leak of less than 10% of tidal volume before cuff deflation is useful in identifying patients at risk for stridor or reintubation (96% specificity). It appears that the amount of cuff leak decreases after intubation for more than 3 days, increasing the risk of stridor and need for reintubation. This information may be helpful in identifying those patients who need treatment for laryngotracheal edema, ie, use of steroids or anesthesia during extubation, the efficacy of which remains to be determined.

The influence of air bags and restraining devices on extremity injuries in motor vehicle collisions.

The influence of air bags and other restraining devices on injury after motor vehicle collisions is not well defined. This study examined the relationship between the use of restraining devices and the incidence of extremity injuries in motor vehicle collisions. A retrospective analysis was performed on motor vehicle collision data submitted to the Pennsylvania Trauma Outcome Study database from 1990 through 1995. Criteria for submission included trauma patients who were admitted to the intensive care unit, who died during hospitalization, who were hospitalized for more than 72 hours, or who were transferred in or out of the receiving hospital. A total of 21,875 patients met these criteria. These patients were analyzed for the presence or absence of upper and lower extremity injuries and were compared based on their use of restraining devices. Restraining devices were categorized into four groups: air bag alone, air bag and seat belt, seat belt or carseat without air bag, and no restraining device. Statistical analysis was performed using the chi-squared test of association. For contingency tables with small expected frequencies, Fisher's exact test was used. Study participants included 11,688 men and 10,185 women with a mean age of 38 +/- 20 years. There were 16,033 drivers and 5,842 passengers. Air bags were deployed in 472 instances. In 297 of these cases, additional restraint was provided with a seat belt. In 6,632 cases, air bags were not deployed; however, patients were restrained with either a seat belt or a carseat. In 14,771 cases, patients were not restrained. When comparing restraining devices as a group vs. no restraint, there was a significant decrease in the incidence of upper (p = 0.018) and lower (p < 0.001) extremity injuries. Air bags, however, were associated with an increased incidence of both upper (p = 0.033) and lower (p = 0.002) extremity injuries when compared with no restraint or when compared among patients who were restrained. As a group, restraining devices decrease the incidence of upper and lower extremity trauma sustained by patients injured in motor vehicle collisions. Air bags, however, are associated with an increased incidence of upper and lower extremity injuries when compared with seat belts alone or when no restraining devices are used.

Impact of preinjury warfarin use in elderly trauma patients.

BACKGROUND: This study examined the hypothesis that elderly trauma patients on warfarin before injury will have increased morbidity and mortality compared with elderly trauma patients not on warfarin. METHODS: From 1993 to 1995, trauma patients were grouped by age and presence or absence of warfarin use before injury. Groups were analyzed with respect to Injury Severity Score, Trauma Registry and Injury Severity Score, Glasgow Coma Scale score, Intensive Care Unit days, hospital days, units of blood transfused, and mortality rates. Statistical analysis was completed by using the Student's t test. RESULTS: Records of 61 patients administered warfarin and 800 patients not administered warfarin were available for analysis. There were no statistically significant differences between patients on prehospital warfarin and those not on prehospital warfarin. CONCLUSION: This study indicates that elderly trauma patients on warfarin before injury do not have increased morbidity and mortality compared with elderly trauma patients not on warfarin.

The influence of airbag and restraining devices on the patterns of facial trauma in motor vehicle collisions.

According to the National Highway Traffic Safety Administration (1990), there were more than 3 million motor vehicle collisions severe enough to lead to significant injury or fatality. Airbags may prevent brain and facial injury caused by these accidents. To date, however, no study has focused primarily on the correlation between facial injuries and the use of airbags and restraining devices. A retrospective analysis was performed on motor vehicle collision data submitted to the Pennsylvania Trauma Outcome Study database from 1990 through 1995. Criteria for submission to the database included admission to the intensive care unit, death during hospitalization, hospitalization for >72 hours, or transfer to or from the receiving hospital. There were 15,450 patients who sustained facial trauma (identified by ICD-9 codes) and were analyzed for patterns of injury and the presence or absence of protective devices. Protective devices were categorized into four groups: airbag alone, airbag with seatbelt, seatbelt or car seat without airbag, and no restraining devices. Statistical analysis was performed using chi-squared test of association. For contingency tables with small expected frequencies, Fisher's exact test was used. There were 9408 male and 6042 female subjects, with a mean age of 38 years (range, 3 to 98 years). There were 11,672 drivers and 3778 passengers. Airbags were deployed in 429 instances. In 276 of these cases, additional restraint was provided with a seatbelt. Airbags were not deployed in 4866 cases when a seatbelt or a car seat was used. In 10,155 cases, no restraining device was employed. There was significantly more facial trauma in patients without protective devices (p < 0.001). Drivers sustained significantly fewer facial fractures when airbags were used, either alone or in combination with a seatbelt (p < 0.001); however, there was no difference in the number of facial lacerations. Among passengers, airbags provided protection from lacerations (p < 0.001) but had no impact on the incidence of facial fractures. In collisions in which airbags were deployed, the use of a seatbelt provided no additional protection from facial fractures or lacerations. In summary, the use of any protective device decreased the incidence of facial fractures and lacerations sustained in motor vehicle collisions (p < 0.001). Airbags provided the best protection of all currently available devices.

Evidence for the Will Rogers phenomenon in migration of employees to managed care plans.

Employees have increasing opportunities to enroll in managed care plans, and employers tend to favor these plans because of their lower costs. However, lower costs may be the result of selection of healthier patients into managed care plans. This study measured differences in health care utilization across an indemnity plan and a managed care plan, and for all employees together. We found that apparent increases in utilization in both indemnity and managed care plans disappeared when the plans were viewed together, reflecting the migration of sicker patients from indemnity plans to managed care plans.

A software program to calculate Goodman and Kruskal's gamma: a method to monitor surgical-site infection rates.

Several studies have described a need for a valid statistical methodology to facilitate interhospital and intrahospital infection rates for their operative procedures. Physician awareness of postoperative surgical-site infection rates has been demonstrated as an effective means to reduce infection rates in both high- and low-risk surgical patients. This article presents a new software program that allows for simple data entry and provides results including the gamma statistic, sample size, variance, standard error, and Z value and P value of the gamma result. This software also provides a data interpretation table, allowing infectious disease department physicians or staff to determine if a significant relation exists between operative procedure infection rates and patient risk factors.

Outcome of blunt thoracic aortic injury in a level I trauma center: an 8-year review.

BACKGROUND: The purpose of this study was to evaluate our experience with blunt thoracic aortic injury and identify factors predictive of outcome. METHODS: Hospital charts, trauma registry data, and autopsies of 64 patients with blunt thoracic aortic injury from 1988 to 1995 were reviewed. RESULTS: Patients were identified and segregated based on admission physiology. Group 1 patients (n = 19) arrived in arrest. Group 2 patients (n = 10) arrived in shock with systolic BP 90. Group 3 patients (n = 35) arrived with systolic BP>90. All patients in groups 1 and 2 expired. Injury Severity Scores for nonsurvivors in group 3 (n = 12) were significantly higher than survivors. There were no significant differences when comparing time of injury to repair or arrival between groups, or in mortality or paralysis comparing repair techniques or clamp/bypass times. Double lumen endotracheal tubes caused significant operative delays compared to single lumen tubes. CONCLUSIONS: Predictors of survivability were hemodynamic stability on arrival and lower Injury Severity Scores. In thoracic aortic injury patients arriving hemodynamically stable, Injury Severity Score correlated with mortality but not paralysis.

Ultrasonography in the diagnosis of acute appendicitis.

The purpose of this study was to assess our results of using graded compression ultrasonography (US) to confirm the diagnosis of acute appendicitis. Graded compression US was performed on 94 patients who presented at the Lehigh Valley Hospital, Allentown, Pennsylvania with an equivocal clinical picture of acute appendicitis. When used to diagnose acute appendicitis, US provided a specificity of 93.7%, sensitivity of 74.2%, and accuracy of 87.2%. We conclude that graded compression US was useful to rule out the diagnosis of acute appendicitis and helped to avoid unnecessary appendectomies and reduced negative laparotomies. Our results proved to be comparable to previously reported studies.

The S factor--a new derived hemodynamic oxygenation parameter--a useful tool for simplified mathematical modeling of global problems of oxygen transport.

We describe a new derived hemodynamic oxygenation parameter, the S factor (S). The factor is based on oxygen delivery and oxygen consumption and can range from -3 to 1. It allows simplified mathematical modeling of clinical problems of oxygen transport and can be applied to many clinical situations. A new hemodynamic oxygenation parameter, the S factor (S), is introduced as an aid to mathematical modeling. It is defined as follows: [formula: see text] (DO2 = oxygen delivery, VO2 = oxygen consumption) S can theoretically vary from -3 (DO2 = VO2) to +1 (VO2 = 0). When DO2/VO2 = 4 (ie. OER = 0.25), S = 0. An S < 0 implies utilization of reserve oxygen transport capacity. An S > 0 implies increased oxygen delivery in relation to oxygen consumption (ie. "shunted oxygen delivery"). By algebraic manipulation and substitution of the components of DO2 into Equation 1: DO2 = Q x Ca x 10 DO2 = Q [(Hb)(Sat)(1.36) + PaO2(.0031)] 10 (2) the following equations can be derived: [formula: see text] [formula: see text] Ca - Cv (Ca = arterial content, Cv = venous content) can be determined by substituting components of oxygen consumption: VO2 = Q (Ca - Cv) x 10 (5) into equation 1 and solving for Ca - Cv. [formula: see text] Equation 6 can be simplified to: [formula: see text] A previously defined relationship between mixed venous PO2 (PvO2) and DO2/VO2 (where calculated P50 is 26.6 +/- 1.0) can be used to modify S in a clinically relevant manner. PvO2 = 5.44D O2/VO2 + 18.16 (8) The relationship between S and PvO2 can be defined by substituting Equation 4 into Equation 1 and solving for PvO2 PvO2 = [21.76/(1-S)] + 18.16 (9) As an example, at a PvO2 of 28 torr (anaerobic threshold), S = -1.2. The relationship between PvO2 and S is shown in Figure 1. S, which can also be defined as 1-4(VO2/DO2) or 1-4(OER), is a useful tool for mathematical modeling of global problems of oxygen transport because the previously derived equations with the S value allow the components of oxygen transport to be interrelated in a clinically relevant manner. Additional advantages of using S in mathematical modeling are: 1. Conceptually it 'fits' in that in regards to the sign (+ or -), as a -S implies utilization of reserve oxygen transport capacity and a +S implies wasted or excess oxygen delivery (shunted). 2. These concepts are easily quantified using the S factor. 3. It 'spreads out' the difference between values for parameters (OER or S) integrating components of oxygen transport, ie. in the 'normal state' regarding oxygen transport, OER = 0.25 and S = 0. At the anaerobic threshold (PvO2 = 28 torr), OER = 0.55 and S = -1.2. Thus, the change in OER from 'normal state' to anaerobic threshold is 0.3 (0.55-0.25) and the change in S is 1.2. This represents a four-fold increase. Four examples of mathematical modeling of global problems of oxygen transport using the S factor are described below.

Defining "dead on arrival": impact on a level I trauma center.

OBJECTIVE: To determine the potential impact of defining criteria for "dead on arrival" (DOA) on a Level I trauma center. METHODS: From 1990 to 1994, trauma patients having cardiopulmonary resuscitation (CPR) performed by certified prehospital personnel were reviewed for time of CPR, outcome, and costs to determine whether any benefit would have been realized had DOA criteria been followed. RESULTS: A total of 106 patients had prehospital CPR; 20 did not meet DOA criteria and underwent resuscitation, three survived (15%). Eighty-six patients met DOA criteria; 16 were pronounced dead without further resuscitative efforts (in-hospital costs of $200/patient), while 70 (81%) had continued resuscitation with no survivors (in-hospital costs of $4150/patient). The positive predictive value for criteria was 100%. Had criteria been implemented, total cost savings over the 5-year period would have been $290,000. CONCLUSIONS: National DOA criteria could dramatically reduce the burden on trauma centers with an estimated minimum annual savings of $14 million.

Role of magnetic resonance imaging in the diagnosis of osteomyelitis in diabetic foot infections.

PURPOSE: The role of magnetic resonance imaging (MRI) in the diagnosis of osteomyelitis in foot infections in diabetics was investigated. The accuracy, sensitivity, and specificity of MRI, plain radiography, and nuclear scanning were determined for diagnosing osteomyelitis, and a cost comparison was made. METHODS: Twenty-seven patients with diabetic foot infections were studied prospectively. All patients underwent MRI and plain radiography. Twenty-two patients had technetium bone scans, and 19 patients had Indium scans. Nineteen patients had all four tests performed. Patients with obvious gangrene or a fetid foot were excluded. RESULTS: The diagnosis of osteomyelitis was established by pathologic specimen (n = 18), bone culture (n = 3), or successful response to medical management (n = 6). Osteomyelitis was confirmed in nine of the pathologic specimens. The diagnostic sensitivity, specificity, and accuracy for MRI was 88%, 100%, and 95%, respectively, for plain radiography it was 22%, 94%, and 70%, respectively, for technetium bone scanning it was 50%, 50%, and 50%, respectively, and for Indium leukocyte scanning it was 33%, 69%, and 58%, respectively. The data were analyzed statistically with the two-tailed Fisher's exact test. MRI was the only test that was statistically significant (p < 0.01). CONCLUSIONS: MRI appeared to be the single best test for the diagnosis of osteomyelitis associated with diabetic foot infections. It had a better diagnostic accuracy than conventional modalities and appeared to be more cost-effective than the frequently used Indium scan.