A prospective controlled trial of computerized decision support for lipid management in primary care

• HIT Description: Computerized decision support. More info...
  
• Purpose of Study: To assess the uptake and effect in primary care of computerize decision support of hyperlipidaemia.
  
• Years of study: 1992
  
• Study Design: Controlled before/after.
  
• Outcomes: Prescribing of lipid lowering agents, use of laboratory tests and referrals for the investigation of hyperlipidaemia.
  

• Settings: 25 general practices covering a population of 150,000 people in a city of Burmingham, England.
• Intervention: 21 of 25 practices were given the decision support system and their data were compared to a historical control period.
• Evaluation Method: Measures of laboratory tests and prescribing activity and referrals along with measures of usage of the system.
• Description: PRIMED system which is described as a Òrule based system written in C and running on DOSÓ. This study evaluated the hyperlipidaemia decision support module in use in general practice. The software comprises of an initial screen prompting data capture for sociodemographic details and other cardiovascular risk factors. Further prompts remind the clinician to enquire on other significant current or past medical and family history and specific aspects of examination.Ó The program offers advice on appropriate management of the patient based on a protocol developed by a lipid specialist.
• Strategy: All practices in Burmingham were written to about the study and of the 22% expressing interest, 25 practices with computers and the resources to collect data were selected.
• Barriers: Several barriers to implementation were discussed including the citing of computers within practices, limiting access to clinicians or nurses consulting with patients. Speed of the program was also cited as a problem. The inability of the software to import and export data successfully from the practice medical system led to users being disillusioned by the time wasted inputting information already available in a host database. In an analysis of usage based on only 14 practices, most practices used the system fairly infrequently with a mean of 12 patients entered. 55% of the practices used the system on less than 12 out of a possible of 100 working days.
• Healthcare Utilization: There was no effect of having a system on the gross rate of requests for lipid tests. However, there was a small but significant increase in the number of patients receiving a full lipid profile in the study period. There were no differences in the cost of lipid lowering drugs prescribed between the intervention and control period. There was a 55% decrease in expected referrals from the practices to a specialist in the intervention groups, but this decrease did not achieve statistical significance.