We conducted a single-centre cohort study of the largest Adult Congenital Heart Disease (ACHD) service, embedded in a tertiary referral centre in New South Wales, Australia.

The outpatient clinic receives medical referrals for adults with ACHD, from paediatric and adult generalists and specialists, encompassing a population of 8.16 million people over a geographical area of 801,150 km².

All adults registered to the ACHD clinic database with moderately or severely complex anatomy were included in the study. ACHD anatomical complexities were categorized according to the 2018 AHA Guidelines. Researchers examined database attendance records in May 2019 to identify patients who had not attended follow-up at the clinic for three years or more. Follow-up attendance was determined by the frequency of care plans documented in the medical records. Patients were excluded if they were non-English speaking, aged 16 years or younger, intellectually disabled or had ACHD anatomy classified as simple complexity. We also excluded people over 75 years as that group were confounded by other issues such as nursing home placement and difficulty travelling to appointments. This study was approved by the Human Research and Ethics Committee Sydney Local Health District (Protocol X18-0189 HREC/11/RPAH/625). Additionally, all people surveyed provided informed verbal consent at the time of telephone contact.

Loss to follow-up was defined as absence from any form of ACHD cardiology care for greater than 3 years or being unable to be located. Patients were telephoned at various times of the day and evening, across all weekdays, including weekends. Voicemail or text messages were left on valid numbers to locate and engage missing patients. Further hand searching of electronic and paper medical records for alternate contacts was undertaken for patients with invalid contact details.

Patients who gave verbal consent were invited to participate in the survey which was conducted by an experienced clinician not working at the ACHD clinic to reduce response bias. Telephone questionnaires were administered between May 2019 and May 2020 to understand current care, barriers to clinic attendance and to offer specialty care reconnection if desired. Participants without a current ACHD specialist were invited to return to the adult CHD clinic or another appropriate specialist centre. The inclusion process is shown in Fig. 1.

Data were collected directly from the database, electronic medical record (eMR) and survey tool. Demographic and clinical characteristics of the patients were collected, including gender, age at last clinic review, socio-economic status, geographical location and ACHD anatomy. Socio-economic status was sourced from the Socio-Economic Indexes for Areas (SEIFA) by the Australian Bureau of Statistics that ranks areas in Australia according to relative socio-economic advantage and disadvantage [20]. An area with a high score on this index has a relatively high incidence of advantage and a relatively low incidence of disadvantage. Geographical remoteness was derived from the Accessibility/Remoteness Index of Australia (ARIA) which categorizes areas according to their distance from “service centers” [21]. ARIA defines five categories of remoteness from very remote, remote, outer regional, inner regional to major city by utilizing postcodes. Remote and very remote categories were condensed.

The survey tool was adapted from the HEART-ACHD trial published by Gurvitz et al. [18], to assess reasons for returning to or abstaining from specialized ACHD care. The self-reported survey of approximately ten (10) minutes duration was administered over the telephone to consenting people. The survey examined demographics (7 fields; gender, nationality, ethnicity, education level, English fluency, language spoken at home, independence in ADL), ACHD care arrangements (5 fields with yes, no or numeric responses); reasons for absence from care (Likert 6 points scale with 19 items). People were asked to rate their reasons for ceasing ACHD care from 5 strongly agree to 1 strongly disagree and 0 as not relevant if no choices were appropriate. Patients who had returned to general cardiology care were asked what prompted this to occur (Likert 6 points scale with 14 items). The results are reported as frequencies.

Data were analyzed using IBM SPSS Statistics version 27. Data are presented as means and standard deviation (SD) or frequency and proportions. Characteristics of the subjects were compared using t-tests for continuous variables if normally distributed, Mann-Whitney if non-normally distributed and categorical data were analyzed using Pearson’s χ2 test and two sample z tests (https://epitools.ausvet.com.au/ztesttwo).

The predictors of loss to follow-up were examined with univariable and multivariable analyses using Cox Hazard regression. The Cox Regression was modelled with the event LTF from the last clinic review date, time as the years in care and covariates included gender, geographical remoteness, socio-economic disadvantage and ACHD lesion. Age was not included as a covariate as it linked to the “event” and forms part of the model and consequently every iteration of analysis shows age as having a very strong relationship. All variables were retained in the model to identify the multivariate predictors.

Qualitative comments from the participant survey were analyzed thematically and patterns reported.

Of people with care gaps in our database, 407 were lost to follow-up. This translates to 22% of our cohort.

We compared the patients with LTF to those in current care and found differences between the groups in age, socio-economic status, remoteness and disease complexity (Table 1). Patients LTF were younger (30.9, SD 11.4 years) than those in current care (35.0, SD 12.65 years) based on age at their last clinic visit (p < 0.001). There were no differences in gender between the groups; males with LTF and those in current care were 53.8% and 54.5% (p = 0.87), respectively. Geographical remoteness scores showed some differences between the groups, with the majority of people in both groups residing close to a major city (80.1 % vs. 83.5%, p = 0.16). A higher proportion of LTF compared to current care resided in remote/very remote areas (1.4% vs. 0.2%, p = 0.003). We found that patients LTF had lower relative socio-economic advantage and disadvantage indices (6.0 SD 2.98) compared to those in the current care group (6.6 SD 2.77, p = 0.002). In the moderate ACHD category, more people were LTF than in current care (84% vs. 74.8%, p < 0.001), while in the severe category, more were in current care than LTF (25.2% vs. 16%, p < 0.001).

Predictors of loss to follow-up in our cohort are detailed in Table 2. The same variables were predictive in both the univariable and multivariable analyses. People who lived in remote/very remote areas were 3.12 (95% CI: 1.15, 8.43) times more likely to be lost to follow-up than those living in a major city. The risk of becoming lost to follow-up reduced by 6% (95% CI: 0.90, 0.98) for every one-point increase in the level of socio-economic advantage (p = 0.002). People with moderately complex lesions were 1.46 (95% CI: 1.06, 1.99) times more likely to be LTF than those with severely complex lesions. Table 2 reports results calculated from cases with complete data only to enable comparison between univariate and multivariate analyses. Appendix 1 provides supplementary univariate analysis of all cases, demonstrating consistent predictors of LTF, irrespective of missing data for gender, SES and geographical remoteness.

Telephoning, digital medical record and hand searches determined the remaining care status and survey inclusions accordingly as shown in Fig. 1. More than 1500 phone calls were placed in search of lost patients and to engage the survey participants, with 3–5 phone attempts required to connect with 75% of people, 4–9 call attempts to reach 23%, and 10–15 call attempts for the remaining 1%. Fifty-six eligible participants were invited to complete the survey. Further analysis was undertaken in the survey sub-group who responded to our phone calls. Call duration averaged 10 minutes, however 39% of calls were longer due to participant comments (20–30 minutes).

We stratified the survey responses into “no ACHD care” (n = 37) and “general cardiology” (n = 19) summarized in Table 3.

Thirty-seven people were absent from ACHD care citing positive health status as a reason; 62% stated they felt well, 35% had lost track of time and 24% believed they no longer needed follow-up. A further 24% stated their GP did not recommend any ongoing cardiac care and 19% perceived their condition to be fixed, as seen in Fig. 2. Several people were worried about receiving bad news (11%), had moved away from the ACHD centre (11%) and had been recommended to have care every 3 years (19%). There were several statements that did not resonate with the cohort and deemed “not applicable” outlined in Appendix 2.

% Agree and Strongly Agree with reasons for loss to follow-up statements.

General cardiology care was motivated by several factors shown in Fig. 3 with recommendation from other health care providers (32%) and a desire to prevent potential problems (32%) being the top responses for statement agreement. The onset of new symptoms or health problems (16%), an interest in getting pregnant (11%), concern about deterioration (5%) and an emergency department visit (5%) were the only other affirmative responses.

% Agree and Strongly Agree with the return to care statements.

As with the “no care” group, a number of statements did not resonate with the cohort and deemed “not applicable” outlined in Appendix 3.

Self-reports showed that 34% were being seen by a general cardiologist, the remaining 66% were not under any care. People reported independently finding a general cardiology specialist, with some selected because they were local or offered private rooms consultations which were preferable for some to a public clinic. Several people shared their fears of surgery, previous bad experiences, lack of understanding, the difficulty of navigating their own way back to care, re-engaging in generalist care rather, as an alternative to no ongoing cardiac care in Appendix 4.