MatthewBarrett, Assistant Professor, University of Virginia[email protected]
Matthew James Barrett, Charlottesville, VA; Guofen Yan, Charlottesville, VA; Robert D. Abbott, Otsu, Shiga, Japan
Submitted March 11, 2014
As Willis et al. recognized, one of the major limitations of Medicare data is that ICD-9 diagnosis coding may not accurately identify individuals with PD. [1] Atypical parkinsonian syndromes, drug-induced parkinsonism, and vascular parkinsonism may be coded as parkinsonism or PD (332 or 332.0) and not secondary parkinsonism (332.1) or other degenerative diseases of the basal ganglia (333 or 333.0). [2] While this is an accepted limitation of PD research using Medicare data, the way in which the DBS and non-DBS groups were allocated in the recent study likely resulted in differences in the false positive rate of PD diagnosis between groups. As DBS is performed in those with advanced PD when the diagnosis is more certain and DBS subjects were more likely to be under the care of a neurologist, there was likely a lower false-positive rate of PD in this group. The non-DBS group, on the other hand, probably included a larger number of subjects without PD who would not be candidates for DBS. This could have altered the distribution of sociodemographic predictors in the non-DBS group. While the prevalence of atypical parkinsonian syndromes is low, vascular parkinsonism and drug- induced parkinsonism comprise a significant proportion of elderly individuals with parkinsonism. [3-5]
It is possible that findings from this study would not change if the non-DBS group had a false-positive rate of PD similar to the DBS group, but a comparison of more homogenous groups would strengthen the present findings. A partial solution to achieving this would be to use medication records to include only subjects prescribed antiparkinsonian medications (e.g., levodopa) and exclude those prescribed medications associated with parkinsonism.
1. Willis AW, Schootman M, Kung N, Wang XY, Perlmutter JS, Racette BA.
Disparities in deep brain stimulation surgery among insured elders with parkinson disease. Neurology 2014;82:163-171.
2. Wright Willis A, Evanoff BA, Lian M, Criswell SR, Racette BA.
Geographic and ethnic variation in parkinson disease: A population-based study of US medicare beneficiaries. Neuroepidemiology 2010;34:143-151.
3. Bower JH, Maraganore DM, McDonnell SK, Rocca WA. Incidence and distribution of parkinsonism in olmsted county, minnesota, 1976-1990.
Neurology 1999;52:1214-1220.
4. Benito-Leon J, Bermejo-Pareja F, Morales-Gonzalez JM, et al. Incidence of parkinson disease and parkinsonism in three elderly populations of central spain. Neurology 2004;62:734-741.
5. Baldereschi M, Di Carlo A, Rocca WA, et al. Parkinson's disease and parkinsonism in a longitudinal study: Two-fold higher incidence in men.
ILSA working group. italian longitudinal study on aging. Neurology 2000;55:1358-1363.
For disclosures, please contact the editorial office at [email protected].
As Willis et al. recognized, one of the major limitations of Medicare data is that ICD-9 diagnosis coding may not accurately identify individuals with PD. [1] Atypical parkinsonian syndromes, drug-induced parkinsonism, and vascular parkinsonism may be coded as parkinsonism or PD (332 or 332.0) and not secondary parkinsonism (332.1) or other degenerative diseases of the basal ganglia (333 or 333.0). [2] While this is an accepted limitation of PD research using Medicare data, the way in which the DBS and non-DBS groups were allocated in the recent study likely resulted in differences in the false positive rate of PD diagnosis between groups. As DBS is performed in those with advanced PD when the diagnosis is more certain and DBS subjects were more likely to be under the care of a neurologist, there was likely a lower false-positive rate of PD in this group. The non-DBS group, on the other hand, probably included a larger number of subjects without PD who would not be candidates for DBS. This could have altered the distribution of sociodemographic predictors in the non-DBS group. While the prevalence of atypical parkinsonian syndromes is low, vascular parkinsonism and drug- induced parkinsonism comprise a significant proportion of elderly individuals with parkinsonism. [3-5]
It is possible that findings from this study would not change if the non-DBS group had a false-positive rate of PD similar to the DBS group, but a comparison of more homogenous groups would strengthen the present findings. A partial solution to achieving this would be to use medication records to include only subjects prescribed antiparkinsonian medications (e.g., levodopa) and exclude those prescribed medications associated with parkinsonism.
1. Willis AW, Schootman M, Kung N, Wang XY, Perlmutter JS, Racette BA. Disparities in deep brain stimulation surgery among insured elders with parkinson disease. Neurology 2014;82:163-171.
2. Wright Willis A, Evanoff BA, Lian M, Criswell SR, Racette BA. Geographic and ethnic variation in parkinson disease: A population-based study of US medicare beneficiaries. Neuroepidemiology 2010;34:143-151.
3. Bower JH, Maraganore DM, McDonnell SK, Rocca WA. Incidence and distribution of parkinsonism in olmsted county, minnesota, 1976-1990. Neurology 1999;52:1214-1220.
4. Benito-Leon J, Bermejo-Pareja F, Morales-Gonzalez JM, et al. Incidence of parkinson disease and parkinsonism in three elderly populations of central spain. Neurology 2004;62:734-741.
5. Baldereschi M, Di Carlo A, Rocca WA, et al. Parkinson's disease and parkinsonism in a longitudinal study: Two-fold higher incidence in men. ILSA working group. italian longitudinal study on aging. Neurology 2000;55:1358-1363.
For disclosures, please contact the editorial office at [email protected].