Experiences from patients in mental healthcare accessing their electronic health records: results from a cross-national survey in Estonia, Finland, Norway, and Sweden | BMC Psychiatry

Main findings

The aim of the present study was to provide a more extensive exploration of MHC patients’ experiences with ORA in four European countries where PAEHRs has been long implemented. In total, a greater proportion of participants reported positive experiences with patient ORA as opposed to negative ones. This tendency was found in all four countries. Approximately one-third of the MHC respondents recounted distinctly negative experiences with the health record, the proportion ranging from 26 to 32% across countries. Notably, around one-third of the patients in MHC who participated in our study expressed feeling offended upon reviewing the contents of their health records, ranging from 18 to 38%. Half of the respondents detected errors in their medical documentation, countries ranging from 21 to 51%. In a similar fraction, one-third, reported the presence of omissions, ranging from 28 to 38%.

Positive experiences

The majority of the respondents indicated that having access to the EHR improved their trust and communication with the HCP, echoing the results of earlier studies [38, 39]. A majority also trusted that only authorised personnel were accessing their EHR, with only 8–15% disagreeing. We did not specifically investigate the respondents’ overall satisfaction with patient ORA, but in all countries more respondents reported positive experiences than negative experiences, a tendency also found in all included studies in a recent scoping review on patients in MHC’s experiences with ORA [19], and the international Delphi study on the harms and benefits of patient access to mental health notes [30].

Errors and omissions

In total, almost half of the respondents reported identifying errors in their documentation. Notably, the highest proportion were found in Sweden, more than twice than in Estonia. This is likely influenced by the fact that in Estonia, only epicrisis is shared with MHC patients, while a more extensive suite of documentation is provided in Sweden. In our study, we differentiated between incorrect information and missing information by asking separate questions about error and omissions. It is possible that omissions can be reported as errors and vice-versa. Consequently, it is appropriate to approach errors and omissions as related concepts, and previous qualitative studies can give an indication to drivers behind the commonly reported errors and omissions in our data. A study that included military veterans’ ORA pointed out that interviewees did not appreciate when the notes did not reflect what had occurred during the session, or they identified wrong information or outdated copy-pasted sections [39]. Similarly, in a study with PAEHR users in Finland, analysis of free text answers indicated that many users found that the notes differed from what they had experienced, while others wished for either more or less detailed notes [36]. In a qualitative study [18] among users who identified as Sámi who reported mental healthcare experience in Norway, a respondent stated that they had spent the large part of a consultation emphasising his extended family relations, but it was not reflected in the notes, and this led to disappointment for the patient. This exemplifies the possibility that similar documentation procedures might yield different appraisal from patients, depending on both individual and cultural factors.

Action taken after noticing errors and omissions

In previous studies, HCPs have voiced concern that patient ORA might increase workload. A qualitative study suggested that HCP were concerned that patient ORA might require them to allocate time to handle phone calls, discussion with patients and amending notes [40]. In a review that summarised HCP perspectives on patient ORA in 2015 [41], half of the included studies expressed concerns that patient ORA increased HCP workload. The results from the present study suggest that patients with ORA frequently identify errors and omissions in their notes, and out of those who found errors, half of them report to have taken action following the discovery. It is likely that action from patients after reading their notes may convert to a reaction in healthcare, at least to handle the request but possibly also explaining or negotiating. Supporting this assumption, a previous study in Norway found that more than a third of HCP and administrative staff had received questions from patients or relatives related to the use of PAEHR [24]. Consequently, and in line with previous research, our results suggest that patient ORA can affect HCP workload. On the other hand, it is likely that errors in the documentation have the potential to introduce clinical errors and patient safety issues, either throughout the treatment, or afterwards. It is therefore likely that patients identifying errors in their documentation and taking action to have it explained or amended, can help avoid treatment errors. In a study among HCP in Norway after the implementation of PAEHR, 25% agreed that they noticed patients were better informed about diagnosis, treatment or follow-up than before implementation [28]. The medical record serves the purpose of a work tool for HCP, a legal documentation of clinical assessments and provided health and, after the introduction of PAHER, increasingly a source of information and empowerment for the patient.

Negotiating these sometimes conflicting requirements is not a trivial task; decisions about whether to amend perceived patient errors, or failures to modify records may lead to legal consequences that have not yet been fully explored [42]. Furthermore, sometimes HCP may make clearly relevant and decisive observations that they are mandated to document, but in doing so risk offending or even introducing conflict in the HCP-patient relationship, which when unresolved can negatively impact the treatment. The discourse on how to balance the different requirements is not settled, illustrated by the Chair of the Norwegian Psychiatric Associations recent statement that the medical record is primarily a work tool for HCP, and that ORA does not imply that the patient should be able to understand everything in it [43]. It is likely that the risk of offence or mismatch between patient expectations and documentation is to some degree a built-in price to pay for transparency, particularly in the MHC field, where the EHR usually contains substantial amounts of qualitative and emotionally loaded content. Still, it is likely that a move towards reimagining the EHR as a multi-purpose tool, and incorporating patient feedback, both to adapt in ongoing individual treatment contacts, as well as patient-informed HCP training can mitigate the dilemma between patient transparency and potential harm [33]. Similarly, HCP can be trained to document healthcare with emphatic and supportive language, and avoid clearly paternalistic, offensive or derogatory formulations [44].

Patient offence at online health records

The present study is to our knowledge the largest study examining perceived offence among MHC users accessing their PAEHR. Only a few studies have previously asked respondents to indicate if they took offence from reading notes in MHC. Previously, a mixed-methods study used a graded response (yes, somewhat/a little, not at all, don’t know) to indicate whether readers of psychotherapy notes, isolated from general psychiatry notes, had taken offence and found that 7 out of 85 (8.2%) responded somewhat/ a little or yes to the item [29]. Similarly, a pilot study [45] that included respondents based on HCP clinical assessment found that 8.9% either agreed or strongly agreed that they felt offended when reading mental health notes. Compared to previous studies that used either isolated psychotherapy notes or included respondents based on HCP clinical assessment, our study employed a broader recruitment strategy, providing all patients that accessed the PAEHR service and had received MHC the opportunity to respond. Therefore, our sample contains users from all care levels of MHC, including primary care, outpatient specialised care, hospitalised patients and patients that have received emergency care. Thus, the sample included in the present study is likely to tap into clinical sub-populations more susceptible to experiencing offence.

Furthermore, we observed differences between the Nordic countries included in the present study, with respondents from assumed culturally similar countries (Sweden and Norway) reporting comparable levels of offence, twice the proportion of that observed in the Baltic country of Estonia. It is possible that sampling methods, cultural differences, different documentation practices, or differences in the information provided through the PAEHR affect whether patients report feeling offended. Consequently, sampling methods, clinical sub-populations and cultural differences should be taken into account when comparing the proportion of offended patients in our results with previous studies. Nevertheless, as indicated by a large thematic analysis of free-text answers from patients in a general outpatient healthcare setting [46] the feeling of being offended can be sorted into three thematic domains: errors and surprises, labelling, and disrespect. The large proportions of reported errors and omissions in our data may have constructed overlap with the concept of offence (i.e. an error or omission in the notes is also causing offence), and contribute to the large proportion of offended respondents.

Perceptions of PAEHR’s security and privacy

Although the number of respondents who experienced breaches in the security and privacy of their records was very small in our study, it is a significant number. Between 2.5% and 6% of MHC participants reported someone demanding access to their records and between 8% and 16% reported that someone accessed against their will. Finland had the highest proportion of such incidents. The reason for this is uncertain but may be related to a recent hacking of the PAEHR service [47]. A recent qualitative study from Sweden, found that some respondents were concerned about who was able to access their medical information [48], and in the US a survey reported that 14% of the respondents were extremely concerned about privacy [29]. Another US survey found that 41% of MHC patients in primary care expressed concerns about privacy [49]. In our study we found a lower rate, potentially due to asking about personal experiences and not opinions. We did not explore these experiences further in terms of underlying causes. It is possible that severe experiences such as an MHC patient being threatened by a family member about releasing information to others, and milder instances such as an unknown name appearing in the read log listing who had accessed the document, are both included in our results. It is important to note that patient users in general are in a disadvantaged position when determining whether the security and privacy of their EHRs has been breached. This is particularly important for MHC users who experience more unwanted access than other patients [50].

Limitations

The present study reports from the largest cross-nationally distributed survey aimed to gather experiences from PAEHR users with MHC experience. Due to the lack of demographic characteristics of the service user populations, it was not possible to estimate whether our samples are representative. Further, the tendency of respondents to provide favourable responses must be considered when employing self-reported measurements. In our study, we used an anonymous survey that were distributed digitally. This method is known to produce less social desirability bias than pan-and-paper and open-identity measurements [51]. The four participating countries, Estonia, Finland, Norway and Sweden, are similar in terms of predominantly public healthcare systems that have provided residents with patient ORA on a national level, for at least a decade. Still, country-specific implementations of patient ORA caution drawing broad conclusions about MHC users’ experiences. For example, users in Estonia could only read their epicrisis which limits the amount of content they see and thus reduces the opportunities to identify errors or omissions, or indeed taking offence. Further research is needed into analysing the relationship between available content in the records and rates of errors and offence. There were differences in survey distribution strategies. The survey was designed in English, and then translated into Norwegian, Swedish, Finnish, Estonian and Russian, possibly introducing lingual nuance differences in the distributed versions. Additionally, the way users were introduced to the survey, as well as the visual presentation varied slightly between the countries [34]. Our data trends could have been affected by whether or not the respondents accessed their EHR immediately before answering the survey, and whether or not the questions were mandatory. For example, the Estonian survey was placed externally from the PAEHR as a post alongside other news, and thus some Estonian respondents could have responded to the survey without ever reading their records. In Norway and Sweden, the questions were mandatory compared to Estonia and Finland, where the questions were optional, so respondents could choose not to respond to some questions. Finally, the medical information provided in the PAEHR service varies between countries. Due to these differences in the national ORA services, survey distribution and administration, the underlying assumptions for each sample and prerequisites for calculating weather the observed differences were due to chance were not present. Consequently, the results were presented in descriptives.