The study took place at two academic family practice units and a community health centre with diverse practice profiles in Toronto. We included all patients who received regular injections (i.e., every 1 to 3 months), regardless of age and etiology of B₁₂ deficiency. We excluded patients if their last injection occurred more than three months before the start of patient recruitment, if they had left the practice, if their B₁₂ therapy had been discontinued or if they had already switched to oral therapy, if they did not speak English and did not have access to a translator, or if they had been recently diagnosed with a serious illness. After applying these criteria, 133 patients were included in the study (Figure 1).

The recruitment cover letter included one sentence describing the equivalency of oral B₁₂ therapy to injections: "Studies have shown that vitamin B₁₂ pills are just as effective and safe as B₁₂ injections." Educational sessions on the effectiveness of oral therapy were held with the nursing staff (since they administer the injections) at all sites, and they were encouraged to pass this information on to their patients. As well, contact information for the study investigators was provided to the patients in case they had any questions.

We placed the initial questionnaire [see Additional file 1 ] in patients' charts to be completed at their next visit for a B₁₂ injection. The initial questionnaire elicited demographic data, medication history, logistical aspects of B₁₂ related visits, history of past B₁₂ therapy, and attitudes about injection and oral therapy. Non-responders received up to two telephone reminders.

Patients willing to try oral therapy were given a six-month supply of B₁₂ tablets (one 1000 μg tablet daily), and were offered testing of serum B₁₂ levels at baseline (one month after their last injection) and after the six-month trial. Patients had the option to withdraw from the trial and return to injections at any time for any reason. A follow-up questionnaire was placed in the charts of patients who tried the pills. This questionnaire re-assessed patients' attitudes about the different forms of B₁₂ therapy and asked whether they would continue on oral supplementation or return to injection therapy (see Additional file 2).

Where variables were not already dichotomous (e.g., for satisfaction with injections, the questionnaire listed "very satisfied," "satisfied," "neutral," "unsatisfied," and "very unsatisfied" as choices), we dichotomized the variable of interest (e.g., "satisfied" vs. "neutral/unsatisfied"). To ascertain the relationship between questionnaire responses and preference for oral therapy (as "willingness to switch" and "permanently switching") we performed bivariate and multivariate analyses. We used a p-value of 0.20 (for the continuity-adjusted χ² and Fisher's exact tests) as the cut-off for inclusion of individual patient factors into multivariate logistic regression models. We then performed backwards stepwise regression to determine statistically significant relationships after adjustment. We generated both crude and adjusted odds ratios with 95% confidence intervals.

One investigator (D.C.) conducted 17 semi-structured interviews in a private meeting room at the family practice unit. A purposive sample was selected amongst those willing to be interviewed to reflect diversity in terms of sex, age, willingness to switch, and final choice of therapy after the trial of B₁₂ pills. These interviews were performed at various stages of the study (before, during, and after the trial of oral therapy). Amongst the 17 participants, four were selected from those not willing to switch, while of the 13 who were willing to try oral therapy, six were interviewed prior to, two during, and five after the trial of oral therapy.

The interviews were audiotaped and transcribed. They lasted 20–30 minutes, with questions about the patient's knowledge and history of their B₁₂ therapy, his or her relationship with health care providers, perceived advantages and disadvantages of both pills and injections, and attitudes about past and current B₁₂ therapy. Analyses of the transcripts were performed independently by two investigators (J.K. and D.T-K.), using a three-step content analysis approach to identify and collate relevant themes 16. Through a process of clarification, confrontation, and consensus, we reached agreement regarding the themes and sub-themes.

Sixty-three of our 86 respondents reported a willingness to switch to oral therapy. A large number of patient factors had no clear statistical relationship to a "willingness to switch" to oral therapy (i.e., p > 0.20) and were excluded from the multivariate model: age, education, income, drug insurance coverage, study site, number of prescription medications, mode of travel to the clinic, years of past B₁₂ therapy, number of non-B₁₂ related visits per month, improvement in perceived well-being since initiating B₁₂ injections, several perceived disadvantages of injections (risk of complications, travel/parking costs), and several perceived disadvantages of pills (take too many pills already, would have to pay for them, won't work as well as injections).

The following factors were included in the multivariate model: gender, time to clinic, satisfaction with past B₁₂ injections, several perceived disadvantages of injections (shots are painful, frequent visits to see MD/nurse, cost to health care system), and one perceived disadvantage of pills (won't see MD/nurse as often). Factors associated with willingness to switch after multivariate adjustment were: being able to get to the clinic more quickly (i.e., in less than 30 minutes) (OR 9.29, 95% CI 2.16–39.97), and believing that injection therapy is disadvantageous due to the need for frequent visits to health care provider (OR 5.41, 95% CI 1.10–26.56) and the increased costs to the health care system (OR 16.68, 95% CI 1.51–184.22) (Table 2).

Of the 63 patients who were willing to switch to oral B₁₂ therapy, eight changed their minds before the trial started. Therefore, 55 patients were started on oral B₁₂. Five dropped out and three were lost to follow-up, leaving 47 patients who completed the six-month trial. Reasons cited for discontinuing oral therapy included fatigue, neurological symptoms, and gastrointestinal intolerance.

Fifty-two patients returned the follow-up questionnaire. Over three-quarters (39 of 51 respondents) reported being satisfied or very satisfied with oral therapy. Only 8% of patients (4/51) perceived that they felt worse with pills, while 23.5% (12/51) felt better and the rest felt the same. Self-reported compliance was good; 48 patients or 92% reported forgetting to take the pills two times or less per month. Thirty-nine patients (71% of the 55 who actually switched) stated that they wished to permanently switch to B₁₂ pills. Of the 35 patients who reported feeling the same with the pills as with injections, 28 (80%) chose pills.

Again, many patient factors had no association with a desire to permanently switch to oral B_12. Patient factors that on bivariate analyses had a p-value of less than 0.20 and were consequently included in the multivariate model were: patient beliefs that frequent visits to see the doctor or nurse and the associated travel/parking costs are disadvantages of injections, and the belief oral B₁₂ would add unnecessarily to an already large number of prescribed oral medications. Table 3 presents the factors significantly associated, after multivariate adjustment, with permanently switching to oral therapy: agreeing that disadvantages of injections included frequent visits to see the doctor/nurse (OR 35.41, 95% CI 2.90–432.70) and travel/parking costs (OR 8.66, 95% CI 1.15–65.30).

Baseline and post-intervention serum B₁₂ levels were obtained for 39 of the 55 patients who switched to pills (Figure 2). The mean serum level increased from 387 to 698 pmol/L (p < 0.0001) (reference range: ≥ 180 pmol/L, unlikely to have B₁₂ deficiency). Only one patient's serum level decreased but even she remained within the normal range.

Some patients felt that the B₁₂ injections they had been receiving in the past were effective: "I find it gives me more energy." Patients who expressed this view were often not willing to switch to oral therapy, or if they did try the pills, they eventually went back to injections. Others felt that the injections hadn't helped much: "There is a little feeling that, 'Oh, I'm taking this. I should feel better,' but I'm not sure if I do." These patients tended to switch permanently to oral therapy.

When asked about the anticipated efficacy of oral therapy, some patients, particularly those who were not willing to switch, were sceptical: "Well, I don't think it will work. If the doctors suggested that I needed to go on the injections, I'm certainly pretty sure that the pills are not going to work. Otherwise he'd have put me on pills, would he not?" Many others were unsure about the efficacy of oral therapy.

Nearly every patient cited convenience as an advantage of switching to B₁₂ pill. Other advantages of oral therapy included savings to the health care system and ease of travel. Cited disadvantages of injections included decreased compliance due to the need for frequent visits and potential complications associated with injections. A number of patients decided to switch because they were interested in participating in a research study. Some patients acknowledged that oral therapy would benefit those who are averse to needles, however, needles did not bother most of the patients interviewed: "...needles don't bother me, because – well, I don't know. They just don't seem to bother me. I know some people are very concerned about it."

For both patients interviewed during their trial of oral B₁₂ and for three of the five patients interviewed afterward, all of whom permanently switched to oral therapy, there was a high level of satisfaction with oral B₁₂. One patient even reported feeling better on oral B₁₂: "Well, I'm more level. I don't feel at the end of the month that I'm running out of energy. I'm quite well aware of that." (see Additional file 3).

Disadvantages of switching to oral therapy included having to take an additional oral medication, fear of side effects, concern about swallowing difficulties, the inconvenience of taking a medication daily, and the potential for losing contact with health care providers and the opportunity for minor drop-in consultations. Some patients were under the impression that injections would work more quickly and directly.

Two patients who switched back to injections after trying the pills were interviewed. These patients switched back due to side effects – one cited neurological and gastrointestinal symptoms (her post-trial serum B₁₂ level was significantly increased from baseline) and the other cited decreased energy (his post-trial serum B₁₂ level was not available). (see Additional file 4).