A carer-held record (CHR) was developed to enable carers to contribute to care plans for their relatives with dementia, alongside other professionals. The aim of this study was to establish whether CHRs provide any measurable benefits for carers, in the areas of health, stress, knowledge of dementia and locus of control. This was a controlled parallel-group trial involving 84 carer/patient pairs, drawn from all new referrals to two community mental health teams (CMHTs). Subjects from one CMHT were given CHRs. The control group from the other CMHT were not offered the initiative. Carers in both groups completed instruments measuring knowledge of dementia, carer stress and strain, carer health, and locus of control (internal, external, chance, shared care) at three time points: baseline, 6 months and 1 year. Findings are reported for outcome at 6 months and 1 year on 51 carer/patient pairs. Having a CHR was associated with significantly reduced scores on the Carer Strain Index at 1 year (mean between-group difference: -1.34, 95% CI -2.67 to -0.02). Regarding locus of control, external control scores were significantly lower in the intervention group, by 2.5 points at 6 months (95% CI -4.82 to -0.18) and 3.43 points at 1 year (95% CI -5.57 to -1.29) (p = 0.01). There was a significant group by time interaction in chance control, with the difference in chance control between groups widening between 6 and 12 months. In the shared care dimension of locus of control, there was no difference at 6 months, but a significant group by time interaction, with the intervention group maintaining a higher level of internal control between 6 and 12 months, while the control group reduced. No significant differences were identified in the domains of carer health or knowledge of dementia. Although not a randomized controlled trial, and limited by smaller sample size than planned, this study indicates that use of CHRs provides some benefits for carers in the domains of carer strain and locus of control.