This study used cross-sectional data from the first wave of the Podiatrists in Australia: Investigating Graduate Employment (PAIGE) survey collected in 2017. The PAIGE study was designed to explore different elements of the podiatry workforce to understand recruitment and retention. Approval for PAIGE was provided by the Monash University Human Research Ethics Committee (7871). Details of the PAIGE study methods are published in detail elsewhere [15]. The researchers used the CHERRIES (Checklist for Reporting Results of Internet E-Surveys) to guide collected data reporting [16].

All podiatrists and podiatric surgeons working in Victoria, Australia, were invited to participate in Wave 1 of the survey. The survey was open between 21st of February 2017 to 27th September 2017. When the survey closed, there were an estimated 1,585 podiatrists registered in Victoria [17]. Participants were recruited through direct emails to podiatry alumni university graduate lists, the Australian Podiatry Association, at Victorian podiatry conferences and via social media accounts of the research team (Facebook, LinkedIn, and Twitter). Podiatrists were also encouraged to share the survey link with podiatry colleagues. Participants were given the option at the end of the survey to leave their contact details to participate in the draw for one of five $100AUD Coles Myer gift cards, participate in a focus group, and receive results following the survey closure.

The PAIGE survey questions were based on the Medicine in Australia: Balancing Employment and Life (MABEL) study [18]. Data were collected as per the PAIGE study methodology [19] and details related to data domains collected in each wave are published elsewhere [15]. Wave 1 survey is provided as Supplementary File 1.

Demographic data collected in Wave 1 included information about age, gender, years practicing, employment profile, rural background, and overall health rating. Data relating to podiatrists’ finances included information about gross earnings, gross household income, whether they received any ‘in kind’ benefits or subsidies as part of their employment, debt related to podiatry education and training (including HECS debt or other debt associated with podiatry training expenses), financial investment in private practice, superannuation contributions, and professional medical liability and insurance premiums. Participants were asked to indicate their level of agreement as to whether they would have enough to live on when they retire through a 5-point Likert scale (1 = strongly agree, 2 = agree, 3 = neutral, 4 = disagree, 5 = strongly disagree) [20].

Survey data were collected through Qualtrics® software (Qualtrics, Provo, UT, USA) [21]. Participating podiatrists were asked to create their own unique identifier code so that subsequent wave responses could be linked. Podiatrists could withdraw at any time by closing their internet browser and forced or requested prompts were used to minimise missing data. Due to the sensitive nature of financial information, questions in this section of the survey were optional and podiatrists could answer all, some or none of the financial questions and still progress through the survey. Cookies were used to save responses for up to four hours for partial completion and Internet Protocol (IP) addresses are routinely collected by Qualtrics® as the de-identified metadata in the survey responses.

Data were initially cleaned to remove any responses not including core demographics (age, gender, recency of practice, work setting, postcode where the participant lived). These data points were deemed essential for the broader research aim of the PAIGE study. Data were reviewed and further cleaned if podiatrists had not provided the suburb and postcode of their work location. Data management and cleaning decisions are outlined in Fig. 1. Data were analysed using Stata 15 software (StataCorp, College Station, TX, USA). Descriptive statistics of all variables of interests were grouped for the entire cohort and then categorised based on postcode data into metropolitan responses (MMM 1) or rural responses (MMM 2, 3, 4, 5, 6 and 7) using the Modified Monash Model (MMM) [22]. This grouping method was used to explore contextual differences. Likert scale data relating to agreement was reduced to a 3-point scale (Agree (combined Strongly Agree and Agree), Neutral, Disagree (combined Strongly Disagree and Disagree)). Univariate regression analysis was used initially to determine if demographic and financial/remuneration factors were associated with work location (metropolitan or rural). Subsequently, data from all podiatrists regardless of working location was combined and multivariate analysis was applied to explore factors associated with financial situation and prospects, the perception they will have enough to live on at retirement. The multivariate model was based on backwards stepwise multiple ordered logistic regression. Factors within this model were chosen where univariate analysis revealed a value of p ≤ 0.20. The variable with the least significant fit was then removed in a backward stepwise procedure until all remaining variables were significant at p < 0.05. The research team built both the univariate and multivariate model based on variables known to impact financial security from the MABEL study [18]. Results are reported in Odds Ratios (OR), adjusted Odds Ratios (adj OR) and 95% confidence intervals (95% CI).