Reader Comments

Post a new comment on this article

Limitations of the conclusions of this study

Posted by plosmedicine on 31 Mar 2009 at 00:27 GMT

Author: Nimalan Arumainayagam
Position: GPVTS ST2
Institution: Green Lane Hospital
Submitted Date: July 06, 2008
Published Date: July 8, 2008
This comment was originally posted as a “Reader Response” on the publication date indicated above. All Reader Responses are now available as comments.

Dear Sir / Madam,

I find this study to be of high importance and relevance to prescribing neuroleptics in dementia patients. However, I feel that there are some limitations to the results and hence the conclusions that can be drawn.

The initial recruitment and subsequent groups which were analysed were considerably underpowered, and hence any conclusions drawn are liable to type 2 error. Thus the conclusion of no difference in SIB and NPI scores between the two groups at 6 months may be erroneous.

Furthermore the authors find that at 12 months the group given placebo had a significantly greater deterioration in NPI scores, thereby suggesting that 6 month follow-up may have been too short to detect a difference.

The results section is a little unclear as initially the authors state that analysis was to be conducted on 51 patients in each group (Figure 1). However, in table 6 within the results section when comparing NPI changes at 6 months 56 patients in the continue treatment group vs 53 patients in the placebo group are compared. When comparing BADLS score changes, 54 in continue treatment and 52 patients in the placebo group are compared for statistical analysis. These numbers contradict the initial maximum numbers for analysis in Figure 1.
There is no 'intention to treat' analysis as would normally be expected in a randomised controlled trial, and again I am unsure as to why this has not been done.

Hence future studies will need to have greater numbers recruited given the high loss to follow up (in part due to comorbidity and mortality), observed in this particularly frail patient group. A longer duration of follow-up with initial larger recruitment numbers would allow for confident confirmation of the conclusions alluded to by the authors.

Yours Faithfully,

Dr Nimalan Arumainayagam

No competing interests declared.