Background HIV home sampling offers an acceptable and convenient method for HIV testing and may provide a practical solution for increasing testing in high risk groups. However, we are unaware of any data comparing the effectiveness of different sampling methods. From August 2013 users of our online HIV testing service were offered an informed choice between blood and saliva HIV sampling.

Method We interrogated the database of all HIV home sampling requests and analysed any differences in demographics and return rates for both blood and saliva samples.

Results Between 15.8.13 and 31.11.14, 14312 home tests were requested. Blood tests were preferentially chosen (9532, 66.6% vs 4780, 33.4%). 7257 samples (50.7%) were returned, this encompassed 4758 blood samples and 2499 saliva samples (49.9% of requested blood samples vs 52.2% of requested saliva samples p = 0.01). The service is predominantly aimed at men who have sex with men and of the returned samples the majority were from men (6416, 84.7%) Men were significantly statistically more likely to request blood samples than women (67% vs 51%, p < 0.00001). In total there were 123 reactive samples (1.7%, 116 men, 7 women), 82 from blood samples (77 men, 5 women) 41 from saliva (39 men, 2 women). The average age of all requests was 30.3 years, 30.8 years in persons who returned samples and 29.7 years for those who did not (p < 0.00001). There was a significant difference in the ages of people requesting saliva versus blood samples (29.7 years vs 30.6 years p < 0.0001). The average age of persons with negative samples was 30.8 years vs. 33.0 years in those with positive samples (p < 0.05). The median number of days from when the sample was ordered to when it was collected back was 6 days in all groups (negative samples, reactive samples, men, women, blood and saliva).

Discussion Despite being more invasive when given an informed choice, more people chose blood over saliva sampling. However saliva samples were more likely to be returned. Women were statistically more likely than men to choose saliva sampling. There was no difference in the length of time it took to return reactive and negative samples.