| Exam component |
Evidence base |
| Visual fields |
When compared with formal visual field testing, these bedside maneuvers have the following sensitivity and specificity, respectively:134
Description of face—35 and 99% Finger counting—25 and 100% Finger comparison—71 and 57% Red comparison—77 and 27% Static finger wiggle—44 and 97% Kinetic finger wiggle—39 and 97% Kinetic 5-mm red target—74 and 93% The combination of static finger wiggle and kinetic 5-mm red target has 78% sensitivity and 90% specificity134 |
| Funduscopy |
Training medical students in direct ophthalmoscopy has attracted debate both for and against:135,136
For: essential nature for the ophthalmic examination and diagnostic capacity when taught/used properly135 Against: lack of confidence (43% of UK general practitioners lack confidence137) and superseded by technology like fundus photography (a study with fundus photography showed 33 new ocular findings in an ED; none were identified by emergency department physicians and only 6 were identified in ophthalmologic consultations138) |
| Pupillary light reflex |
Medical students consider it the component of the ophthalmologic examination that they are most confident with139 Neurocritical care nurses diagnose only half of the cases of anisocoria compared with a pupillometer140 Interobserver agreement for anisocoria is 0.50 for neurologists and 0.22 for trainees (kappa)141 |
| Pursuit extraocular movements |
Medical students consider it the component of the ophthalmologic examination that they are second-most confident with after pupil function139
Interobserver agreement for jerky eye movements is 0.67 for neurologists and 0.26 for trainees (kappa)141 |
| Facial sensation |
In patients with injury to the inferior alveolar nerve (from V3) during bilateral sagittal split osteotomy, bedside sensory tests have sensitivity of 17–59% and specificity of 73–91% on the lip, and sensitivity of 40–58% and specificity of 56–100% on the chin. Out of the tests studied, orientation discrimination of square-wave gratings is the most sensitive on the lip and touch detection using different monofilament strengths the most sensitive on the chin142 |
| Facial muscles |
Facial paresis has the strongest correlation with stroke out of NIH Stroke Scale items. When combined with arm weakness and dysarthria, it has a sensitivity of 100% and specificity of 92% for determining stroke143
Facial paresis is one of the three clinical examination items on both the Cincinnati Prehospital Stroke Scale and Los Angeles Prehospital Stroke Screen144, 145
Interobserver agreement for facial palsy is 0.57 for neurologists and 0.81 for trainees (kappa)141 |
| Sternocleidomastoid |
In patients with hemiparetic strokes, 70% have decreased strength of the sternocleidomastoid ipsilateral to the lesion146
In a study of healthy individuals (104 medical students and 102 people aged above 60), all had normal sternocleidomastoid function147 |
| Tongue |
In patients with hypoglossal palsy, the following signs are seen:148
In healthy individuals, tongue deviation occurs with protrusion in 2% of medical students and 12% of people aged above 60 years147
Speech pathologists use eyeball/clinical judgment (88%) or a ruler (9%) to measure tongue range of motion (3% did not specify an answer when surveyed)149 |
| Gait |
Eighty-nine percent of general practitioners and 98% of medical students feel they have difficulty analyzing a gait disorder; 80 and 50%, respectively, feel there is insufficient teaching during medical school150
There are many scales used to grade gait with many having >0.75 inter- and intraobserver agreement (ICC).151 |
| Tandem gait |
Abnormal tandem gait has odds ratio of 4.07 for falls over the next 6 mo in patients with Parkinson's disease152
Abnormal tandem gait has sensitivity of 55% and specificity of 95% for identifying patients with a structural cerebral lesion (although the control group included only patients with a normal neurological examination, so there is likely bias causing a falsely high specificity)153 |
| Pronator drift |
Pronator drift is present in 94% of patients with an acute motor stroke and was the most sensitive of the upper motor neuron signs studied154
Pronator drift has sensitivity of 92% and specificity of 90% in patients with a proven lesion in the motor pathways155 In a comparison of patients with upper extremity weakness, drift without pronation was present in all of the 26 patients with conversion disorder (sensitivity of 100%) but only 2 of the 28 patients with organic lesions (specificity of 93%)156 |
| Rapid alternating movements |
Rapid alternating movement abnormalities have positive predictive value of 0.75 and negative predictive value of 0.57 among people with cerebellar disorders, Parkinson's disease, and upper motor neuron disorders.157 Of note, this study did not perform subgroup analysis by condition for these abnormalities nor specify exactly what constituted abnormal (e.g., rhythm or speed). As predictive values depend on the proportion of the study population affected by conditions, we calculated sensitivity and specificity; they are 62 and 72%, respectively
Interobserver agreement for rapid alternating movements is 0.43 between a neurologist and senior trainee (kappa)157 |
| Finger–nose |
Finger–nose abnormalities have positive predictive value of ~1 and negative predictive value of ~0.5 for cerebellar disorders, Parkinson's disease, and upper motor neuron disorders grouped together. As with rapid alternating movements, this study did not perform subgroup analysis by condition nor specify exactly what constituted abnormal. As predictive values depend on the proportion of the study population affected by conditions, we calculated sensitivity and specificity; they are ~26 and ~100%, respectively.
Interobserver agreement is 0.12–0.14 between a neurologist and senior trainee (kappa).157 A different study reports 0.55 for neurologists and 0.50 for trainees (kappa)141 |
| Tone arms Tone legs |
In patients with a focal cerebral hemisphere lesion, spasticity was found in the upper extremities in 2 out of 46 patients and 0 out of 19 controls (sensitivity of 4%, specificity of 100%); in the lower extremities in 4 patients and 0 controls (sensitivity of 9%, specificity of 100%).158 Of note, this study did not detail the chronicity of the neurological conditions nor were the lesions necessarily in the motor cortex
Spasticity has interobserver agreement of >0.75 for elbow flexors and hamstrings, but <0.50 for hip adductors, quadriceps, gastrocnemius and soleus between a physical therapist and physical therapy students (ICC)159 |
| Power arms Power legs |
In patients with a focal cerebral hemisphere lesion (although not necessarily in the motor cortex), upper motor neuron weakness was found in the upper extremities in 14 out of 46 patients and 0 out of 19 controls (sensitivity of 30%, specificity of 100%); in the lower extremities in 9 patients and 0 controls (sensitivity of 20%, specificity of 100%).158
Segmental motor weakness has sensitivity of 39% and specificity of 98% in patients with a proven lesion in the motor pathways155
Interobserver agreement for elbow extension force is 0.61 for neurologists and 0.53 for trainees (kappa)141
The Medical Research Council (MRC) scale of 0 to 5 is commonly used to grade power.160 Multiple studies have examined the inter- and intraobserver agreement:
Interobserver agreement between physical medicine and rehabilitation physicians is 0.77–0.78 for finger extension, wrist extension, and grip strength after radial nerve injury (weighted kappa)161 Interobserver agreement between a senior trainee and attending critical care physician is 0.23–0.64 for individual upper and lower extremity muscle groups in critically ill patients (weighted kappa)162 Intraobserver agreement by physical therapists is 0.65–0.93 for individual upper and lower extremity muscle groups in Duchenne's muscular dystrophy; more proximal muscles have higher agreement (weighted kappa)163 |
| Biceps reflex Brachioradialis reflex Triceps reflex Patellar reflex Achilles reflex |
In patients with a focal cerebral hemisphere lesion, hyperreflexia was found in the upper extremities in 5 out of 46 patients and 1 out of 19 controls (sensitivity of 11%, specificity of 95%)158 Of note, this study did not detail the chronicity of the neurological conditions and the lesions were not necessarily in the motor cortex.
Hyperreflexia has sensitivity of 69% and specificity of 88% in patients with a proven lesion in the motor pathways155
Interobserver agreement for knee jerk is 0.40 for neurologists and 0.27 for trainees (kappa)141
The National Institute of Neurological Disorders and Stroke (NINDS) scale of 0 to 4 is commonly used to grade reflexes.164
Multiple studies have examined the inter- and intraobserver agreement:
Interobserver agreement between neurologists is 0.43–0.74 for different reflexes; lower extremity reflexes have higher agreement. Intraobserver agreement is 0.77–0.89 (weighted kappa)165 Interobserver agreement between neurologists and trainees is 0.11–0.35 for different reflexes (kappa)166 |
| Plantar |
The extensor plantar response has been evaluated in many conditions:
It has sensitivity of 35%, specificity of 77%, and interobserver agreement of 0.30 between neurologists and primary care providers (kappa) in lower extremities with upper motor neuron weakness.167 It has sensitivity of 51%, specificity of 99% and inter-observer agreement of 0.73 between neurologists (kappa) in patients with pyramidal tract dysfunction.168 It is present in 64% of early-acute and 43% of late-acute patients with motor strokes154 Interobserver agreement for extensor plantar reflex is 0.54–0.55 for neurologists and 0.59 for trainees (kappa)141,169 |
| Light touch |
When physicians or physical therapists assessed individual dermatomes with light touch in patients with spinal cord injury, 14 dermatomes had no variation, 26 had interobserver agreement of 0.61–0.91, 11 had agreement of 0.41–0.60, and 5 had agreement of ≤0.40 (kappa)170 |
| Vibration |
Quantitative vibration measurements (with a Rydel-Seiffer tuning fork) but not qualitative vibration measurements can distinguish between people with abnormal and normal nerve conduction study responses. Reduced quantitative measurements have a 3.3-fold higher likelihood (p < 0.001) of abnormal sensory nerve action potentials, compared with reduced qualitative measurements that have only a 1.4-fold higher likelihood (p = 0.12)171 |
| Pinprick |
When physicians or physical therapists assessed individual dermatomes with pinprick in patients with spinal cord injury, 10 dermatomes had no variation, 4 had interobserver agreement of 0.61–0.69, 25 had agreement of 0.41–0.60, and 17 had agreement of ≤0.40 (kappa)170 |
| Romberg |
When people are tested using different versions of the Romberg test, they can stand longest with their feet 15 cm apart, then feet together, and then feet in line (heel to toe). Controls can stand longer than patients with cerebellar ataxia, who can stand longer than patients with sensory ataxia172
Patients' risk of falling increases as their time of standing for the modified Romberg test (standing on a foam pad with eyes closed) decreases; those who stand less than 20 s have more than a 3-fold higher risk of having had a fall over the preceding 12 mo173 |
Comments