Box 1.

Observations by a relative

Hospital admission and (emergency) triage represent a significant opportunity for the patient–carer dyad to explore and agree a common version of history; specifically those events and associated timeline leading to the present clinical episode. Basic disagreements over dates, the relevance of symptoms and the precision of recall must all be negotiated.
Data are generated within the dyad from multiple standpoints, not simply the primary perspectives of patient and carer but also those potentially arising from multiple roles played by the relative as carer, advocate and/or attorney, or indeed by the patient as both parent/child and sick person. The simulated elderly patient may wish to leave hospital as quickly as possible without too much attention being paid to their chronic constipation problem, whereas the simulated relative as carer may be more interested in a period of respite and a complete purge; while at the same time, as the patient's advocate or attorney, may consider that prolonged hospitalisation could result in further loss of independent living skills.
The personal health record (PHR) can facilitate the dyad in completing this information-gathering task precisely because it obliges both members to focus on achieving unanimity. The logical design inherent in any associated data collection tool(s) and underlying database systems can also assist in imposing the use of a common jargon-busting language on all participants, both data contributors (patients and carers) and all subsequent users (patients, healthcare practitioners and hospital administrators). However, some thought may need to be given to the fundamental unobservability of many variables of interest in the PHR (eg progress of disease) and that recording a single version of the truth may actually represent a loss of information, in particular the ability to assess degree of correlation between differing versions at the same timepoint.
Among the most important additional roles played by the carer when the PHR is computerised and the patient lacks dexterity, is visually impaired or simply does not belong to the digital era is that of the ‘touchscreen operator’ (cf medical scribe) controlling data input. While there may be few initial surprises (name, age, gender etc), the general influence of the touchscreen operator grows as time passes and symptom/medication lists and associated sub-menus become more interminable, which can result in an abbreviated approach to creating the record, whose ownership is then uncertain. The simulation was realistic in this respect.