After viewing a narrowPeak file in IGV I found that some regions had overlapping records, each of which has a different name and different associated peak statistics. For example:

chr17    35164351    35165393    p300_EpiLC_minusActivin_peak_17078a    49    .    3.62557    7.39305    4.98814    174
chr17    35164351    35165393    p300_EpiLC_minusActivin_peak_17078b    122    .    5.43658    15.04741    12.24058    420
chr17    35164351    35165393    p300_EpiLC_minusActivin_peak_17078c    247    .    6.42027    27.98737    24.78897    823

Is this normal behaviour for MACS2? I would have thought it wrong to report three separate records for the same peak region? I called the peaks using the following command:

macs2 callpeak \
  -t p300_EpiLC_minusActivin.filtered.bam \
  -c p300_EpiLC_minusActivin.filtered.bam \
  -n p300_EpiLC_minusActivin \
  -f BAM \
  -g mm \
  -s 36 \
  -q 0.01 \
  -B \
  --call-summits 2> p300_EpiLC_minusActivin.log

My apologies, I have found my answer. See this excerpt from the MACS2 manual:

--call-summits

MACS will now reanalyze the shape of signal profile (p or q-score depending on cutoff setting) to deconvolve subpeaks within each peak called from general procedure. It's highly recommended to detect adjacent binding events. While used, the output subpeaks of a big peak region will have the same peak boundaries, and different scores and peak summit positions.