Yes, we need to factor in age. But who has the data? Otherwise we need to adopt an indirect approach.

Lethbridge et al. looked at death certificates issued in Nova Scotia.

For PwP the main co-morbidities were:
Alzheimer's/dementia at 26.2%
Pneumonia, 22.3%
Stroke, 14.2%

Does this disprove the 80% figure after 8 years? Not quite. First, we have the issue of under-reporting: there may be a reluctance to report dementia. Second, we don't know what proportion of the people had had PD for at least 8 years before dying. A clue to this is given by Isihara et al. who report [2] "a Dutch study, ... median survival after PD diagnosis at 71 years or older was 9 years".

At this stage I get stuck. Can anyone progress this?

References

[1] Prog Palliat Care. 2013 July; 21(3): 140–145. .
doi: 10.1179/1743291X12Y.0000000037
PMCID: PMC3703207
"Co-morbidities of persons dying of Parkinson's disease"
Lynn Lethbridge, Grace M Johnston, and George Turnbull
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3703207/

[2] J Neurol Neurosurg Psychiatry 2007;78:1304-1309 doi:10.1136/jnnp.2006.100107
"Estimated life expectancy of Parkinson’s patients compared with the UK population"
Lianna S Ishihara, Anne Cheesbrough, Carol Brayne1, Anette Schrag
http://jnnp.bmj.com/content/78/12/1304.full

John