Computer Prescribing System and Impaired Renal Function

Context: The incidence of drug side effects increases with patient age, and the frail old subjects hospitalized in geriatric units are particularly exposed to this risk. Optimizing drug prescription is an important issue to prevent the side effects. Among the various possible actions, it is important to pay attention to impaired renal function which requires careful dose adjustment for drugs eliminated by renal way.

Objective: To evaluate the quality of orders for target drugs in old patients with impaired renal function, hospitalized or living in long stay units, and to improve these practices by means of counselling system coupled to a computer prescribing system. The computer prescribing system allow (a) to detect in real-time the order of drugs contraindicated in case of renal failure or prescribed in excessive amount taking into account the creatinine clearance (ClCr), and (b) to counsel the physician in these situations.

Participants: A coordinating centre (team of geriatrics 1, hospital Charles Foix, Ivry-sur-Seine, France), and 12 hospital investigator centres (hospital services of geriatrics working already in network, including the co-ordinating centre), and the department of Biostatistics (hospital Pitié-Salpétrière, Paris)

Type of study: Intervention study, cluster randomised, open, parallel.

General design: The investigator centres will be equipped with the electronic prescribing system (EPS) developed for the project. For each patient, the software requires the seizure of weight and plasma creatinine and calculates ClCr using Cockcroft formula, from which the result is accessible to the physician. After a phase of training of the EPS, the quality of the orders given to old subjects having renal failure (defined by ClCr\<60 ml/mn) will be studied for 3 months under natural conditions of prescribing, without use of the counselling system (sequence A). After this phase, the counselling system will be activated for 6 months (sequence B) in 6 investigatory centres chosen by a cluster randomisation process. The centres were paired by the numbers of orders during the sequence A. For each pair, the centres were allocated to either intervention or control group by a randomisation table.

Reference frame to evaluate the quality of target drug orders:This reference frame aims to analyze the respect of the mentions of the "summaries of characteristics of products" (or RCP) relative to the impaired renal function. Drug RCP is a health authorities-approved document which describes in detail the information for prescribing. It corresponds in the US to the FDA-approved Full prescribing information. Using an automated analysis of the RCP, we will identify target drugs, defined as drugs which are contraindicated in case of renal failure or are cited in precautions for use for patient with impaired renal function. For each speciality thus identified, the possible thresholds of ClCr implied to define the counter-indication or the adjustment of dose will be sought by reading and analyzes CCP. An order line will be regarded as inappropriate if the drug is contraindicated in renal failure or if the dose prescribed exceeds that indicated in RCP for the level of patient's ClCr.

Counselling system: This system will operate in patients with ClCr \<60 ml/mn. It will comprise three functionalities: (a) detection of orders of a drug contraindicated in case of renal failure and production of an alert message; (b) detection of excessive dose taking into account patient's ClCr and production of an alert message; (c) counsel to monitor plasma concentration of drugs having a renal elimination in the event of impaired renal function. This system will be very largely based on the data obtained to work out the reference frame.

Intervention: The investigator centres will be allocated to the group Intervention (6 centres) or Controls, without intervention (6 centres). The intervention applied for 6 months (sequence B), consists in activating the counselling system. An alert message indicates to the physician, according to the case, that the drug is contraindicated in the event of impaired renal function or that dose must be adjusted, by indicating the amounts recommended adjusted to ClCr when it are mentioned in the CCP. The physician keeps the final choice of his order and the prescribed dose. If the final choice differs from the counsel of the system of council the reason for the maintenance of the regulation is asked.

Data collection: Information concerning the investigator centres and the patients (type of stay, age, sex, weight, score of dependence of Activities off daily living room (ADL), index of co morbidity, plasma creatinine); occurred of adverse drug reactions and their description; lines of orders. All the data necessary for the study will be recorded through the EPS and seizures by the physician will be transmitted to the co-ordinating centre after encoding via Internet network.

Statistical analysis: The analysis will relate on the patients having ClCr lower than 60 mL/mn during the duration of the study and to the orders which are made to them. The characteristics of the patients of each group (Control/Intervention) and the proportion of target drugs will be studied and compared during the sequence A, then during the sequence B. We will compare the reductions in the proportions of inappropriate target drug orders between the sequences B and A between the Intervention group and Control (principal judgement criterion). In addition will be required in the Intervention units the existence during the 6 months of use of a possible effect of training of the physicians (progressive transfer of the expertise to the physician) which could result in a time-related reduction in the number of counsels generated by physician. Lastly, the adverse drug reactions which have occurred during the study period will be collected to analyze their nature, their incidence and their possible bonds with impaired renal function in order to guide future studies on this set of themes.

Expected results: If such a system proves to be effective to optimize prescribing of the drugs among in-patients having impaired renal function and if it is well accepted by the physicians, diffusion could be considered in centres of geriatrics. In the same way, similar studies in hospital medical wards with no geriatricians in their staff or also in family medicine settings could be considered. The algorithms will be made available for the developers which would wish to integrate it into other applications of regulations.