OBJECTIVE Strategies for overcoming alloimmune refractoriness to random donor platelets are based on the use of compatible platelets selected from large panels of HLA-typed donors or cross-matching (XM). The aim of this study was to review the effectiveness of a platelet XM programme for treating refractory haematological patients at Milan's Policlinico Hospital (PHM) 2002-2014 and Spedali Civili in Brescia (SCB) 2013-2016. METHODS A commercially available solid-phase antibody detection system was used for platelet antibody detection and XM. Forty-nine alloimmune refractory patients at PHM and 13 at SCB, respectively, received a median [IQR] of 12 [6-13] and 18 [13-15] XM compatible platelet transfusions after the detection of refractoriness. The absolute increases in post-transfusion platelet counts obtained using random, and XM platelets were retrieved from the patients' hospital records. RESULTS The critical review at SCB showed that the median [IQR] 1 h post-transfusion increase in platelet counts was 3 × 109 /L [1-5] after 47/47 random platelet transfusions, and 10 × 109 /L [2-25] after 325/326 XM compatible platelet transfusions. The documentation concerning the outcomes of XM platelet transfusions at PHM was incomplete, and so the findings of the review were inconclusive. CONCLUSIONS This retrospective analysis confirmed the effectiveness of the XM programme at SCB, but revealed defective data collection and retrieval methods at PHM, thus underlining the importance of such methods. The literature review accompanying this retrospective analysis identified a recently described algorithm for ensuring platelet support in refractory patients that optimally integrates the combined use of XM and HLA typing.