This was a cohort study using RRR and TT methodology, which encompassed 600 randomly selected healthcare records from ten home healthcare organizations across Sweden, included by convenience sampling. The identification of both AEs and no-harm incidents in the home healthcare context was conducted at the same time when reviewing the records. Sample size calculation was based on the identification of AEs . The method and the identification of AEs is described in detail elsewhere [19, 20].
Setting and sample
The Swedish Health and Medical Services Act,  designed to ensure that everyone living in Sweden has access to good healthcare, gives the county councils/regions and municipalities considerable freedom in organizing care. The organization of home healthcare services differs within the country and the provision of home healthcare in an area can be the responsibility of either the county council or the municipality. Physicians are always employed by the county councils, with which the municipalities collaborate. There are both public and private home healthcare facilities, but they are generally publicly funded. Home healthcare usually means care provided by licensed healthcare professionals and does not include home care organizations with unlicensed staff administering social care.
In this study, seven of the review teams were employed by municipalities and three worked in home healthcare for county councils. Each review team consisted of one to three registered nurses and one or two physicians, in total 28 clinicians. The ten review teams represented nine different regions across Sweden.
In the RRR process, a no-harm incident was an event caused by healthcare or social care that reached the patient and could have led to an AE, but resulted in no discernible harm . A preventable no-harm incident was defined as an event that could have been prevented if adequate actions had been taken during the patient’s contact with healthcare or social care. No-harm incidents related to acts of either omission or commission were included.
Inclusion and exclusion criteria
All patients aged 18 years or older admitted to home healthcare at the studied units during 2015 were included in the sample from where the randomly selected records were retrieved. The record review covered a maximum of 90 days from the date of a patient’s admission to home healthcare (index admission). If a patient had been discharged from home healthcare and readmitted within the 90-day period, the review of the record continued.
To be included as a no-harm incident, the following criteria had to be met; the no-harm incident occurred and was detected in the home healthcare records during the index admission, i.e., within 90 days after enrolment in home healthcare, regardless of caregiver.
Randomization was performed by one of the researchers (MU), using an online randomizer, to ensure it was carried out in the same way for all review teams. Oversampling was performed by 10 records per team. If a patient in the random sample was receiving limited home healthcare, e.g., only blood pressure measurement, this patient was replaced by another randomly selected admission.
Before review start, each review team member underwent a mandatory one-day education in the TT methodology, including practical training in reviewing healthcare records and discussion of detailed examples. A trigger manual and a study manual were used as the basis for decision-making during the review process.
Each review team reviewed 60 randomly selected home healthcare records from their own organization. The RRR was conducted in two stages, a primary and a secondary review. Usually, the registered nurses in each team performed the primary and secondary reviews and then discussed their findings with the physicians until consensus was reached. In some teams, the physicians carried out some of the primary and/or secondary reviews. In the primary review, the reviewer documented demographic data and screened for the presence of the 38 predefined triggers in each healthcare record . A trigger is a word or a sentence in a record that could indicate that a no-harm incident had occurred, such as “fall” or “insufficient planning, coordination, communication, and/or information.” For each identified trigger, in parallel to screening for AEs,  the reviewer assessed whether or not it reflected the presence of a potential no-harm incident. Ten percent of the records in the primary review process were evaluated concerning inter-rater reliability. The inter-rater reliability of the primary reviewers’ judgements concerning if a record was to be forwarded to secondary review was κ=0.801 (substantial) . If a potential no-harm incident was found, the record went forward to secondary review.
The reviewer sorted the different triggers into potential no-harm incidents, as several triggers can be related to a specific no-harm incident. Each event was then reviewed separately. Initially, the reviewer assessed whether the event was associated with healthcare/social care using a 4-point scores; 1) the event was not related to healthcare/social care; 2) the event was probably not related to healthcare/social care; 3) the event was probably related to healthcare/social care; 4) the event was related to healthcare/social care. A score of 3 or 4 on the Likert scale was required for the event to went forward in the review process. A score of 1 or 2 was for example an event that was assessed as a consequence of the disease's development or considered to be socially acquired. A similar 4-point scale was used to assess if the event was considered preventable or not and was deemed from the patient perspective. For example, a "fall without harm" was considered to be preventable if no risk assessment, or risk identification was carried out, or if preventive measures of predictable risks in the physical care environment (i.e home environment) had not been discussed with the patient.
“Unplanned hospitalization” due to inability from home healthcare personnel to “handle medical-technical equipment” in the patient’s home was considered preventable through adequate education. For each trigger the study manual gave examples to guide decisions for preventability.
The National Coordinating Council for Medication Error Reporting and Prevention criteria (NCC MERP) were used to assess severity . NCC MERP consists of nine categories (A–I). In the present study, category C (an event that reached the patient but did not cause harm) and category D (an event that reached the patient and required monitoring or intervention to ensure that no harm occurred) were included, i.e., were seen as no-harm incidents.
Data were also collected regarding interventions and potential contributing causes linked to the no-harm incidents. Several interventions and potential contributing causes per no-harm-incident could be selected from a predefined list. The contributing causes we used have been evolved over time. They are originally derived from HMPS studies and other classification systems as the Swedish clinical incident reporting system.
All no-harm incidents that were detected in the home healthcare records during the review were included, regardless of if they originated from home healthcare or other healthcare or social care organizations. The record systems were generally computerized, but documentation routines and access to the full records or parts thereof differed between the teams.
The data were analyzed descriptively (presenting cumulative incidence, range and percent). The chi-squared (χ2) test was used to analyze differences between men and women and between patients aged 80 years or older and patients younger than 80 years. Differences between groups were considered to be statistically significant if the p values were < 0.05. The SPSS statistics program version 24 and Statistica 64 V.13 were used to perform calculations.