Identifying Parkinson’s disease and parkinsonism cases using routinely-collected healthcare data: a systematic review

Abstract

BackgroundPopulation-based, prospective studies can provide important insights into Parkinson’s disease (PD) and other parkinsonian disorders. Participant follow-up in such studies is often achieved through linkage to routinely-collected healthcare datasets. We systematically reviewed the published literature on the accuracy of these datasets for this purpose.MethodsWe searched four electronic databases for published studies that compared PD and parkinsonism cases identified using routinely-collected data to a reference standard. We extracted study characteristics and two accuracy measures: positive predictive value (PPV) and/or sensitivity.ResultsWe identified 18 articles, resulting in 27 measures of PPV and 14 of sensitivity. For PD, PPVs ranged from 56-90% in hospital datasets, 53-87% in prescription datasets, 81-90% in primary care datasets and was 67% in mortality datasets. Combining diagnostic and medication codes increased PPV. For parkinsonism, PPVs ranged from 36-88% in hospital datasets, 40-74% in prescription datasets, and was 94% in mortality datasets. Sensitivities ranged from 15-73% in single datasets for PD and 43-63% in single datasets for parkinsonism.ConclusionsIn many settings, routinely-collected datasets generate good PPVs and reasonable sensitivities for identifying PD and parkinsonism cases. Further research is warranted to investigate primary care and medication datasets, and to develop algorithms that balance a high PPV with acceptable sensitivity.

Identification of Drugs Associated with Lower Risk of Parkinson’s Disease Using a Systematic Screening Approach in a Nationwide Nested Case–Control Study

“…[38][39][40] Diagnosis of PD and its differential diagnostics is challenging, and false diagnoses are common in the early phase. 41,42 The proportion of excluded persons in FINPARK (25.9%) is in line with estimated proportion of misdiagnosed PD, [41][42][43][44] supporting the validity of outcome. It should be noted that the PD diagnoses were from nearly 20 consecutive years (1998-2015), and it is possible that there may have been variations in clinical diagnostic process.…”

Alcohol Intake and Parkinson's Disease Risk in the Million Women Study

“…In addition, in a recent systematic review that evaluated the accuracy of routinely collected health care data for identifying PD cases, the investigators found that the positive predictive value was >70% for a majority of hospital studies. 34 To our knowledge, this is the largest prospective study to investigate the association between alcohol intake and PD risk. The results suggest that alcohol intake does not materially influence the risk of PD in UK women.…”

“…In addition, in a recent systematic review that evaluated the accuracy of routinely collected health care data for identifying PD cases, the investigators found that the positive predictive value was >70% for a majority of hospital studies …”

Routinely collected data as real-world evidence for physiotherapy practice

“…A third development relates to the rapidly increasing availability of large amounts of clinical data generated from electronic health records, registries, insurance claims, and even social media and wearable smart devices, and the ability to access, process, link, and analyze these data in fairly efficient ways, as compared to conducting RCTs (Jarrow, LaVange, and Woodcock, 2017). In medicine, many examples of such studies have been published across several domains and health problems, for instance, for estimating accuracy of diagnosis of: motor neuron disease (Horrocks et al, 2017); dementia (Wilkinson et al, 2018); and Parkinson's disease and parkinsonism (Harding et al, 2019); for prognosis of undiagnosed chest pain (Jordan et al, 2017); and to evaluate treatment on survival in metastatic melanoma (Van Zeijl et al, 2018). Routinely collected data are generated and collected during the course of health-care delivery, not primarily for research purposes, and therefore have the potential to minimize costs and effort and maximize representativeness and generalizability by capturing information in large populations over long periods from large databases that are continually updated as opposed to establishing treatment effectiveness using RCTs (Hemkens, Contopoulos-Ioannidis, and Ioannidis, 2016).…”