Are patients referred to rehabilitation diagnosed accurately?

BACKGROUND: An accurate diagnosis of the leading health condition and comorbidities is a prerequisite for safe and effective rehabilitation. The problem of diagnostic errors in physical and rehabilitation medicine (PRM) has not been addressed sufficiently. The responsibility of a referring physician is to determine indications and contraindications for rehabilitation.

AIM: To assess the rate of and risk factors for inaccurate referral diagnoses (RD) in patients referred to a rehabilitation facility. We hypothesized that inaccurate RD would be more common in patients: 1) referred by non-PRM physicians; 2) waiting longer for the admission; 3) older patients.

DESIGN: Retrospective observational study.

POPULATION: One thousand randomly selected patients admitted between 2012 and 2016 to a day-rehabilitation center (DRC).

SETTING: University DRC specialized in musculoskeletal diseases.

METHODS: On admission all cases underwent clinical verification of RD. Inappropriateness regarding primary diagnoses and comorbidities were noted. Influence of several factors affecting probability of inaccurate RD was analyzed with multiple binary regression model applied to 6 categories of diseases.

RESULTS: The rate of inaccurate RD was 25.2%. Higher frequency of inaccurate RD was noted among patients referred by non-PRM specialists (30.3% vs. 17.3% in cases referred by PRM specialists). Application of logit regression showed highly significant influence of the specialty of a referring physician on the odds of inaccurate RD (joint Wald Test χ2 (6)=38.98, P value =0.000), controlling for the influence of other variables. This may reflect a suboptimal knowledge of the rehabilitation process and a tendency to neglect of comorbidities by non-PRM specialists. The rate of inaccurate RD did not correlate with time between referral and admission (joint Wald Test of all odds ratios equal to 1, χ2 (6)=5.62, P value =0.467), however, mean and median waiting times were relatively short (35.7 and 25 days respectively). A high risk of overlooked multimorbidity was revealed in elderly patients (all odds ratios for variable age significantly higher than 1). Hypotheses 1 and 3 were confirmed.

CONCLUSIONS: Over 25% of patients referred to DRC had inaccurate RD. Risk factors for inaccurate RD include referral by a non-PRM specialist and elderly age.

CLINICAL REHABILITATION IMPACT: Verification of RD should be routinely introduced to PRM practice.