Article Figures & Data
Tables
Characteristic Development Cohort (n = 203) Validation Cohort (n = 415) Unfavorable dizziness course,a No. (%) 150 (73.9) 181 (43.6) Female, No. (%) 127 (62.6) 305 (73.5) Age, mean (SD), y 77.5 (7.0) 78.5 (7.1) Number of chronic diseases, mean (SD) 1.8 (1.4) 2.1 (1.4) Number of drugs used, mean (SD) 5.5 (3.6) 4.4 (3.0) Dizziness characteristics DHI score,b mean (SD) [range] 46.2 (13.4) [30–88] 36.1 (19.9) [0–88] Time since onset of dizziness, No. (%) 0–4 weeks 3 (1.5) 29 (7.0) 1–6 months 34 (16.7) 98 (23.6) 6–48 months 49 (24.1) 109 (26.3) 2–10 years 74 (36.5) 120 (28.9) >10 years 42 (21.2) 59 (14.2) Description of dizziness, No. (%)c Instability or unsteadiness 162 (79.8) 280 (67.5) Loss of balance 165 (81.3) 287 (69.2) Lightheadedness 145 (71.4) 247 (59.5) Rotational sensation 137 (67.5) 228 (54.9) Tendency to fall 135 (66.5) 225 (54.2) Giddy 115 (56.7) 176 (42.2) Environment spinning 84 (41.4) 137 (33.0) Becoming unwell 60 (29.6) 114 (27.5) Near faint 59 (29.1) 96 (23.1) Everything turning black 39 (19.2) 68 (16.4) DHI = Dizziness Handicap Inventory.
↵a At 6 months of follow-up.
b Score ranges from 0 to 100, with higher scores indicating more severe dizziness-related impairment.
c Adds up to more than 100% because more than 1 answer was allowed.
Predictor B OR (95% CI) P Value Intercept −6.385 – – DHI-S score,a per point 0.127 1.14 (1.07–1.21) <.001 Age, per year 0.063 1.06 (1.00–1.13) .04 History of arrhythmiab 0.661 1.94 (0.81–4.63) .14 Looking up as provoking factorb 0.690 1.99 (0.92–4.32) .08 B = regression coefficient; OR = odds ratio; DHI-S = screening version of Dizziness Handicap Inventory.
Note: Adjusted model on internal validation, linear predictor = −5.687 + (0.115 × DHI-S) + (0.056 × age) + (0.597 × history of arrhythmia) + (0.623 × looking up as provoking factor).
↵a Score ranges from 0 to 40, with higher scores indicating more severe dizziness-related impairment.
↵b Yes = 1, no = 0.
Predictor Development Cohort Validation Cohort Age, mean (SD), y 77.5 (7.0) 78.5 (7.1) DHI-S score, mean (SD) 17.4 (6.9) 13.7 (9.3) History of arrhythmia 59 (29.1) 39 (22.4) Looking up as provoking factor 129 (63.5) 214 (51.6) DHI-S = screening version of Dizziness Handicap Inventory.
Predictor at Baseline Score Age in years 1 × age DHI-S score in points 2 × DHI-S score History of arrhythmia Yes 11 No 0 Looking up as provoking factor Yes 11 No 0 DHI-S = screening version of the Dizziness Handicap Inventory.
Note: Risk score is derived by adding all scores. Example: A man aged 78 years with a DHI-S score of 14, a history of arrhythmia, and looking up not being a provoking factor for dizziness will have a score of (78 + [2 × 14] + 11 + 0) = 117.
A score of ≥134 corresponds to a high risk of an unfavorable course of dizziness. The risk of an unfavorable course of dizziness is ≥77.1% for patients with a score ≥134 (see Table 5).
- Table 5
Model Predictive Performance in the Validation Cohort at Various Risk Score Thresholds
Risk Score Threshold (Percentage of Total Sample) True Positive, No. True Negative, No. False-Positive, No. False-Negative, No. Sensitivity, % Specificity, % Positive Predictive Value, % Negative Predictive Value, % ≥69 (100) 181 0 234 0 100.0 na 43.6 na ≥86 (90) 175 35 199 6 96.7 15.0 46.8 85.4 ≥92 (80) 169 71 163 12 93.4 30.3 50.9 85.5 ≥98 (70) 159 102 132 22 87.8 43.6 54.6 82.3 ≥105 (60) 152 137 97 29 84.0 58.5 61.0 82.5 ≥113 (50) 137 163 71 44 75.7 69.7 65.9 78.7 ≥119 (40) 116 184 50 65 64.1 78.6 69.9 73.9 ≥126 (30) 91 201 33 90 50.3 85.9 73.4 69.1 ≥134 (20) 64 215 19 117 35.4 91.9 77.1 64.8 ≥150 (10) 34 227 7 147 18.8 97.0 82.9 60.7 >176 (0) 0 234 0 181 na 100.0 na 56.4 na = not applicable.
Additional Files
Supplemental Appendixes 1-5
Supplemental appendixes 1-5
Files in this Data Supplement:
- Supplemental data: Appendixes - PDF file
The Article in Brief
Predicting an Unfavorable Course of Dizziness in Older Patients
Hanneke Stam , and colleagues
Background Dizziness, which strongly affects daily functioning in older adults, can refer to several different sensations. Defining the type of dizziness might help the physician better understand the cause of a patient"s dizziness. The aim of this study is to develop and externally validate a prediction model for an unfavorable course of dizziness in older people in primary care, with dizziness-related impairment as starting point.
What This Study Found A new tool can help primary care physicians identify older patients at risk of an unfavorable course of dizziness (i.e., six months of substantial dizziness-related impairment). The risk score, based on a validated prediction model, consists of four easily obtained predictors of dizziness: age, history of arrhythmia, score on the Dizziness Handicap Inventory (screening version), and looking up as a trigger for dizziness.
Implications
- The authors suggest that the tool can activate primary care physicians to target potential contributing factors for high risk of an unfavorable course of dizziness, even when the cause of dizziness is unknown.
In this issue
Jump to section
Related Articles
Cited By...
- Treatment success of internet-based vestibular rehabilitation in general practice: development and internal validation of a prediction model
- Prognosis and Survival of Older Patients With Dizziness in Primary Care: A 10-Year Prospective Cohort Study
- Internet based vestibular rehabilitation with and without physiotherapy support for adults aged 50 and older with a chronic vestibular syndrome in general practice: three armed randomised controlled trial
- Clinical Prediction Rules: Challenges, Barriers, and Promise
- In This Issue: Nothing Simple