my sugestion is to change HO THRESHOLDS LEV UL RX LEVEL. Make a trial with -100 dBm and -105 dBm.
Because UL is limiting factor for MS-BS comunication. If you have same HO settings for UL and DL, there will be 95% of imperative HO with cause UL. If you change this parameter to lower value, there will be slightly increase of DL cause (and may be slightly increse number of HO due to Quality – UL and DL). We are using parameters mutch more closer to the threshold -110.